1 GENDER BASED VIOLENCE IN GEORGIA Links Among Conflict, Economic Opportunities and Services GENDER BASED VIOLENCE IN GEORGIA Links Among Conflict, Economic Opportunities and Services © 2017 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington, D.C. 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpre- tations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guaran- tee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. This work was made possible by the financial contribution of the State-and Peacebuilding Fund (SPF). SPF is a global multi-donor trust fund administered by the World Bank to finance critical develop- ment operations in situations of fragility, conflict, and violence. Over 150 grants and transfers have been approved in 39 countries for a total of US$300 million. The SPF also supports global analysis, expert learning, and multi-stakeholder partnerships. The SPF is kindly supported by: Australia, Den- mark, France, Germany, The Netherlands, Norway, Sweden, Switzerland, The United Kingdom, as well as IBRD. Rights and Permissions The material in this work is subject to copyright. Because the World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, D.C. 20433, USA; fax: 202-522- 2625; e-mail: pubrights@worldbank.org. Cover design and text layout: Duina Reyes-Bakovic Acknowledgements T he World Bank Group and Analysis and Con- sulting Team (ACT) Georgia carried out re- search and analysis for this report. This report was modation, and Refugees (MRA), the Office of the Prime Minister, the Office of the Public Defender, the National Statistics Office (GEOSTAT), and the prepared by a team led by Rebecca Lacroix, Task State Fund for Protection and Assistance of (Stat- Team Leader and Social Development Special- utory) Victims of Human Trafficking (ATIPFUND) ist, and Holly Benner, Co-Task Team Leader and also provided support and strategic direction to Senior Social Development Specialist. The core the World Bank Task Team. team members were Maximillian Ashwill, Consul- tant; Chiara Broccolini, Consultant; Cesar Cancho, The research benefited from consultations with Economist; Phil Crehan, Consultant; Jonathan many national and international partners. These Karver, Research Analyst; Nora Dudwick, Con- partners include NGO Women, Union Sapari, sultant; Sophia Georgieva, Social Development Youth Voices Georgia, Institute for Policy Studies, Specialist; Diana Jimena Arango, Senior Gender Municipal Development Fund, Women’s Infor- Specialist; and Jocelyn Kelly, Consultant. Maxi- mation Centre, Anti-Violence Network of Geor- millian Ashwill and Rebecca Lacroix authored this gia, IDP Women Association ‘Consent’, Georgian report. Peer reviewers included Joanna de Berry, Young Lawyers’ Association, Women’s Initiatives Senior Social Development Specialist; Maria Be- Supportive Group, European Union Monitoring atriz Orlando, Lead Social Development Special- Mission, Organization for Security and Co-op- ist; and Julia Vaillant, Economist. Additional ad- eration in Europe, and the United Nations (UN) vice was received by Ana Maria Munoz Boudet. agencies, including UN Women, United Nations We are grateful for the support and guidance Population Fund, and United Nations Devel- from the World Bank Group’s Country Manage- opment Programme. The team acknowledges ment Unit (CMU) in Georgia. This includes the Duina Reyes-Bakovic for design and formatting CMU’s Country Director, Mercy Miyang Tembon, support. Finally, we especially thank all the Geor- and Program Leader, Sarah Michael. gians who participated in this study’s survey, in- terviews, and focus group discussions for sharing The Georgian Ministry of Internally Displaced their valuable time and experiences with us. Persons from the Occupied Territories, Accom- 5 Abbreviations ACT Analysis and Consulting Team IRB Institutional Review Board ALP Administrative Line Person KII Key Informant Interview ATIPFUND State Fund for Protection and LGBT Lesbian, Gay, Bisexual, and Assistance of (Statutory) Victims Transgender of Human Trafficking LGBTI Lesbian, Gay, Bisexual, CEDAW Convention on the Elimination Transgender, and Intersex of All Forms of Discrimination LiTS Life in Transition Survey against Women MPI Multidimensional Poverty Index CMU Country Management Unit MRA Ministry of Internally Displaced DHS Demographic and Health Persons from the Occupied Survey Territories, Accommodation, EBRD European Bank for and Refugees Reconstruction and NAP Nonconflict-affected Person Development OHCHR Office of the United Nations EU European Union High Commissioner for Human FFV Four Forms of Violence Rights FGD Focus Group Discussion OSCE Organization for Security and FRA European Union Agency for Co-operation in Europe Fundamental Rights PRIO International Peace Research GBV Gender-based Violence Institute, Oslo GDP Gross Domestic Product SGBV Sexual and Gender-based Violence GEOSTAT Georgia’s National Statistics Office SIDA Swedish International Development Cooperation GBVIMS Gender-Based Violence Agency Information Management System SPF State and Peacebuilding Fund IASC Inter-Agency Standing TSA Targeted Social Assistance Committee UN United Nations IDMC Internal Displacement UNDP United Nations Development Monitoring Centre Programme IDP Internally Displaced Person UNFPA United Nations Population Fund IHS Integrated Household Survey UNHCR Office of the United Nations ILGA International Lesbian, Gay, High Commissioner for Bisexual, Transgender, and Refugees Intersex Association USAID United States Agency for IMAGES International Men and Gender International Development Equality Survey WHO World Health Organization 6 Table of Contents Acknowledgements.................................................................................................................................... 5 Abbreviations............................................................................................................................................... 6 Executive Summary...................................................................................................................................11 Key Findings................................................................................................................................................................................12 Policy Implications and Recommendations.............................................................................................................................14 შემაჯამებელი მიმოხილვა.................................................................................................................17 ძირითადი შედეგები...................................................................................................................................................................18 პრაქტიკული დასკვნები და რეკომენდაციები.........................................................................................................................22 1 Introduction............................................................................................................................................26 2 Methodology..........................................................................................................................................28 2.1 Overview of Research Components..................................................................................................................................29 2.1.1 Quantitative Survey Sampling Design and Questionnaire...................................................................................30 2.1.2 FGD Sampling Design ..............................................................................................................................................30 2.1.3 Forms of Violence Included in the Survey ..............................................................................................................32 2.1.4 Structure of the Survey and of the Violence Module ............................................................................................32 2.2 Ethical Considerations and Ethics Review.........................................................................................................................32 2.3 Uniqueness and Limitations of Methodology..................................................................................................................33 3 Georgian Context and Respondents’ Profiles...................................................................................35 3.1 History of Conflict and Displacement................................................................................................................................ 35 3.2 Georgian Economy..............................................................................................................................................................36 3.3 Sample Group Details.........................................................................................................................................................36 4 Gender Norms and Reported Violence..............................................................................................42 4.1 Gender Norms in Georgia Today.......................................................................................................................................42 4.2 Changes in Attitudes over Time ........................................................................................................................................46 4.3 Gender Norms Related to Physical and Sexual Violence................................................................................................47 4.4 Women’s Experiences of Violence ....................................................................................................................................49 4.5 Underreporting of Violence................................................................................................................................................ 54 4.6 Trends in Women’s Experiences of Violence.................................................................................................................... 55 4.7 Trends in Men’s Experience of Violence............................................................................................................................ 55 4.8 Trends in Violence Against Sexual Minorities................................................................................................................... 57 5 GBV, Conflict, and Economic Stressors...............................................................................................59 5.1 Conflict-affectedness........................................................................................................................................................... 59 5.2 Conflict’s Economic Legacy................................................................................................................................................63 5.3 Economic Stress and GBV...................................................................................................................................................66 5.4 Women’s Lack of Financial Independence Contributes to GBV Risks...........................................................................68 7 6 Services and GBV...................................................................................................................................71 6.1 Survivor Responses to GBV.................................................................................................................................................71 6.2 Lack of Awareness of GBV Services...................................................................................................................................74 7 Policy Implications and Recommendations.......................................................................................79 7.1 National Policy and Legislation on GBV............................................................................................................................79 7.2 Stakeholder Mapping..........................................................................................................................................................82 7.3 Recommendations and Looking Ahead............................................................................................................................84 8 Bibliography...........................................................................................................................................86 ANNEXES...................................................................................................................................................91 Annex 1: Methodology .............................................................................................................................................................93 Annex 2: Organizations Addressing GBV in Georgia..........................................................................................................103 Annex 3: Quantitative Analysis...............................................................................................................................................106 Annex 4: Socio-economic Characteristics of the Sample Groups......................................................................................167 List of Figures Figure E.1: Research Questions...............................................................................................................................................12 ნახ. E.1: კვლევის დროს დასმული კითხვები .......................................................................................................................18 Figure 1: Implementation Phases............................................................................................................................................28 Figure 2: Conceptual Framework for Population’s Life Experiences in Georgia Survey, 2016........................................29 Figure 3: Distribution of IDP Population, 2016.......................................................................................................................36 Figure 4: GDP Per Capita (at current prices), US$ .................................................................................................................37 Figure 5: Share of Population Under 60 percent of the Median Consumption (relative poverty) (%)............................37 Figure 6: Percentage of Respondents Who Feel Gender Equality Has Come Far Enough, by Age Groups..................43 Figure 7: Percentage of Urban and Rural Men and Women Who Feel Gender Equality Has Come Far Enough.........43 Figure 8: Gender Role Indicators with the Smallest Discrepancy Between Male and Female Responses, ‘Do you agree with the following?’..........................................................................................................................44 Figure 9: Gender Role Indicators with a Larger than 10 Percent Difference in Male and Female Responses, ‘Do you agree with the following?‘.......................................................................................................................... 45 Figure 10: Percentage of Respondents Who Agree with the Following Norms Statements, by Gender....................... 45 Figure 11: Stigma toward Various Social Groups in Georgia (2010–2016)........................................................................47 Figure 12: Stigma toward Homosexuals in Georgia (2010–2016), by Gender and Age...................................................48 Figure 13: Responses to ‘A Man has a Good Reason to Hit his Wife if’, by Gender...........................................................48 Figure 14: Percentage of Respondents Who Agree Women Can Refuse Sex for the Following Reasons, by Gender................................................................................................................................................................49 Figure 15: Top Four Contributing Factors for the Most Significant Partner Incident, by Gender.................................... 53 Figure 16: Percentage of Respondents Who Indicate That They or Their Partner Have Directly Witnessed or Experienced Conflict in the 1990s or 2008, by Gender and Population Groups...........................................60 Figure 17: Comparing Levels of GBV before the Conflict versus after the Conflict, by Sample Group..........................62 Figure 18: Respondents Who Indicated That They Worried That Soldiers or Police Might Physically or Sexually Attack Them or Their Partner during the Conflict or Displacement, by Population Group............63 Figure 19: Changes in Livelihoods for Survey Sample Groups before and after the 2008 Conflict...............................64 Figure 20: Individuals Not Working but Seeking a Job over the Last Four Weeks, pre and post the 2008 Conflict, by Gender and Population Group......................................................................................................... 65 Figure 21: Stability of Employment, by Population Group....................................................................................................66 8 Figure 22: Response to the Question ‘If GBV Increased from Conflict, What Do You Think Is the Most Important Cause?’..................................................................................................................................................67 Figure 23: Responses to the Question ‘Do You Possess Any of the Following Items?’, by Gender...............................69 Figure 24: Most Important Type of Support to Get a Job, by Gender...............................................................................70 Figure 25: Responses to the Question ‘If a Man Mistreats his Partner, Should Others Outside of the Family Intervene?’..................................................................................................................................................71 Figure 26: Where Respondents Went to Escape GBV.........................................................................................................71 Figure 27: Responses to the Question ‘Did the Police Learn of the Incident?’ (women)................................................72 Figure 29: Perspectives on the Presence of More Female Police Officers, by Gender....................................................72 Figure 30: Responses to the Question ‘Are There Female Police Officers in your Community?’, by Gender ..............72 Figure 28: Reasons for not Reporting an Incidence (women)............................................................................................72 Figure 31: Perspectives on GBV Legal Protections, by Gender..........................................................................................74 Figure 32: Availability of Services for GBV Survivors, by Location, Population Group, and Gender............................. 75 Figure 33: Respondents Who Have Seen a GBV Advertisement or Announcement, by Gender..................................77 Figure 34: Relationship between Gender Perspectives and Awareness of GBV-related Campaigns...........................77 Figure 35: Timeline of Legislation and Policy Developments Related to GBV.................................................................81 Figure 36: Main GBV-related Stakeholders in Georgia.......................................................................................................82 Figure 37: Mapping Stakeholder Support of GBV Survivors .............................................................................................83 Figure 1A.1: Implementation Phases.......................................................................................................................................93 Figure 1A.2: Conceptual Framework.......................................................................................................................................94 List of Tables Table E.1: Women’s Vulnerability Profiles to GBV..................................................................................................................13 ცხრილი E.1: ქალთა მოწყვლადობის/დაუცველობის მოდელები და გენდერული ნიშნით ძალადობა/GBV.............19 Table 1: Basic Research Components.....................................................................................................................................29 Table 2: Number of Survey Interviews and Size of Sample (in Brackets)...........................................................................30 Table 3: Survey Modules and Themes....................................................................................................................................32 Table 4: Ethnic Diversity in Georgia Among the Survey’s Sample Groups........................................................................37 Table 5: Socioeconomic Characteristics of the Sample Groups.........................................................................................38 Table 6: Gender Views Profiles................................................................................................................................................46 Table 7: Percentage of Women Reporting At Least One Form of Violence, by Type of Violence and Individual Characteristics........................................................................................................................................... 51 Table 8: Women’s Vulnerability Profiles to GBV..................................................................................................................... 52 Table 9: Average Number of Experiences of Violence for Women, by Population Group.............................................. 52 Table 10: Percentage of Men Reporting At Least One Form of Violence.......................................................................... 56 Table 11: Conflict-affected Persons as a Percentage of the National Population, by Gender........................................60 Table 12: Percentage of Women Reporting at Least One Form of Violence, by Population Group...............................61 Table 13: Average Number of Violent Events Experienced by Women, by Population Group .....................................61 Table 14: When the Most Significant Non-partner Event Occurred, by Gender..............................................................62 Table 15: Percentage of Respondents Who Feel Shame If Poor or Unemployed, by Gender and Population Groups.67 Table 16: Reasons for Returning Home after GBV Incidents (Women)..............................................................................68 Table 17: Reasons Cited for Not Securing a Job, by Gender..............................................................................................69 Table 18: Perception of the Availability and Usefulness of Services to Protect Against GBV, by Gender.....................76 Table 1A.1: Basic Research Components...............................................................................................................................94 Table 1A.2: Number of Survey Interviews and Size of Sample (in brackets)...................................................................... 95 9 Table 1A.3: National Sample Calculation..............................................................................................................................96 Table 1A.4: IDP Sample Calculation.......................................................................................................................................97 Table 1A.5: ALP Sample Calculation......................................................................................................................................97 Table 1A.6: Focus Group Composition..................................................................................................................................98 Table 1A.7: Survey Modules and Themes.............................................................................................................................99 Table 3A.1: Explanatory Variables in the Basic Model.......................................................................................................107 Table 3A.2: Predictors of Number of Total Violence - Basic Model..................................................................................107 Table 3A.3: Predictors of Number of GBV - Basic Model..................................................................................................110 Table 3A.4: Predictors of Number of Sexual Abuse - Basic Model..................................................................................112 Table 3A.5: Predictors of Number of Physical Violence.....................................................................................................114 Table 3A.6: Predictors of Number of Sexual Violence.......................................................................................................116 Table 3A.7: Predictors of Number of Sexual Harassment.................................................................................................119 Table 3A.8: Predictors of Number of Controlling Behavior..............................................................................................121 Table 3A.9: Predictors of Violence Number Including Gender Views Variables............................................................124 Table 3A.10: Predictors of Violence Number Including Indicators of Economic Stress.................................................125 Table 3A.11: Predictors of Number of Total Violence - Including Indicators of Partner Issues - Only Married or Cohabiting Respondents............................................................................................................127 Table 3A.12: Predictors of Number of GBV Including Indicators of Partner Issues - Only Married or Cohabiting Respondents................................................................................................................................130 Table 3A.13: Predictors of Number of Sexual Abuse Including Indicators of Partner Issues - Only Married or Cohabiting Respondents...........................................................................................................................132 Table 3A.14: Predictors of Number of Physical Violence Including Indicators of Partner Issues - Only Married or Cohabiting Respondents............................................................................................................134 Table 3A.15: Predictors of Number of Sexual Violence Including Indicators of Partner Issues - Only Married or Cohabiting Respondents............................................................................................................136 Table 3A.16: Predictors of Number of Sexual Violence Including Indicators of Partner Issues - Only Married or Cohabiting Respondents............................................................................................................139 Table 3A.17: Predictors of Number of Controlling Behavior Including Indicators of Partner Issues - Only Married or Cohabiting Respondents...................................................................................................141 Table 3A.18: Predictors of Number of Total Violence – Including Partner’s Job Stability.............................................144 Table 3A.19: Predictors of Number of Sexual Abuse - Including Partner’s Job Stability Female ................................146 Table 3A.20: Predictors of Number of Sexual Abuse - Including Partner’s Job Stability..............................................148 Table 3A.21: Predictors of Number of Physical Violence - Including Partner’s Job Stability........................................150 Table 3A.22: Predictors of Number of Sexual Violence - Including Partner’s Job Stability...........................................153 Table 3A.23: Predictors of Number of Sexual Harassment - Including Partner’s Job Stability.....................................155 Table 3A.24: Predictors of Number of Controlling Behavior - Including Partner’s Job Stability..................................157 Table 3A.25: Factors Related to Norms and Beliefs toward Gender...............................................................................159 Table 3A.26: Factors Related to the Probability to Have Reported the FFV or More Forms of Violence....................161 Table 3A.27: Awareness of Services.....................................................................................................................................163 Table 4A.1: Socioeconomic Characteristics of the Sample Groups.................................................................................167 List of Boxes Box1: Study Rationale.............................................................................................................................................................27 Box 2: Violence Indicators by Category of Violence Used in Survey of Population’s Life Experiences in Georgia Survey, 2016..................................................................................................................................................31 Box 3: Domestic Violence Shelters and Crisis Shelters.......................................................................................................81 Box 1A.1: Violence Indicators by Category of Violence Used in This Study...................................................................100 10 Executive Summary This report summarizes research undertaken as part of the World Bank State- and Peace-building Fund (SPF) financed grant, Strengthening Capacity for Prevention and Response to Sexual- and Gender-Based Violence in Georgia (GBV). The goal of the grant is to build knowledge and capacity on prevention and response to GBV in Georgia, with a particular focus on conflict- and displacement-affected populations, economic opportunity and services. The project is part of the World Bank’s glob- and impacts on their living conditions and liveli- al initiative on conflict and Gender-based Vi- hoods. The research aims to complement exist- olence (GBV). The Global Initiative, financed by ing initiatives by the Government, international the SPF, includes pilot projects across East Asia partners and the NGO community on GBV in and the Pacific, Africa, and South Asia (nearly 10 Georgia and to explore openings for additional million total in project financing). The aim of the progress. Specifically, filling research gaps on the initiative is to increase understanding regarding potential links between GBV and conflict and in- the development dimensions of GBV and poten- ternal displacement, economic opportunity, and tial areas of enhanced World Bank and develop- services. ment partner programming. While the global ini- tiative supports operational projects in the other The research included a mix of qualitative and regions, the Georgia pilot, representing the Eu- quantitative methods including a stakehold- rope and Central Asia Region (ECA), is unique in er analysis, key informant interviews, nation- its focus on deepening knowledge and promot- wide focus group discussions, and a nation- ing capacity building. ally representative quantitative survey. (see Figure E.1). The survey also sought to understand Research was undertaken in Georgia recog- gender norms as they might influence experienc- nizing the country’s legacy of conflict and dis- placement challenges as well as recent steps es and opportunities in each of the research ar- taken by the Government of Georgia on gen- eas. The research compares the experiences of der action and GBV response. Given conflicts two groups affected by armed conflict—internally in Abkhazia and South Ossetia regions in the displaced persons (IDPs) and administrative line 1990s and 2008, Georgia has many conflict-af- persons (ALPs)—with that of nonconflict-affected fected people. This includes between 190,000 persons (NAPs) and analyzes four forms of vio- and 275,000 IDPs, who have been displaced by lence (FFV), namely controlling behavior/emo- conflict and make up almost 6 percent of the tional abuse, sexual harassment, sexual violence, population, among the highest relative propor- and physical violence. The stakeholder analysis tions in the world. Also, people living near former and institutional mapping conducted was uti- conflict zones in Abkhazia and South Ossetia re- lized to help guide recommendations for future gions continue to experience periodic insecurity development action. 11 12 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Figure E.1: Research Questions Conflict ƒƒ Are conflict-affeced people more at risk from GBV? ƒƒ Can displacement induced stress increase risks? Violence Service Livelihoods & economic opportunities ƒƒ What are the service needs of ƒƒ How have conflict-affected people’s livelihoods conflict-affected people? been impacted by the conflict? ƒƒ What are the service needs of the ƒƒ Does household economics impact the general population? experience of violence? ƒƒ What is the level of knowledge and ƒƒ What economic options and coping mechanisms access to services? are available to survivors? KEY FINDINGS of their income to a spouse against their will or to give up a job because their partner wanted them The following are key findings based on the to. research questions: Specific sub-groups are more susceptible to Over a quarter of all survey respondents have violence. The study created ‘vulnerability pro- experienced some form of violence. Twen- files’ to illustrate which groups within the survey ty-seven percent of women report having experi- respondents report the most types of violence enced at least one type of emotional abuse, sex- (Table E.1). Urban women are most likely to re- ual harassment, sexual abuse, or physical abuse. port experiencing violence, while women with Among respondents, 16.3 percent of women worse housing conditions report higher sexual reported being a survivor of at least one type harassment and sexual violence. Compared to of sexual abuse (sexual harassment, sexual vio- the others, the survey finds that women who got lence, or both) and 8 percent of women report- married before the age of 18 were much more ed experiencing physical violence. It should be likely to suffer emotional abuse, sexual violence, noted that these are not incidence rates1 and that and physical violence. The analysis demonstrates women’s actual experience of violence and their that respondents with stressed partners, who are reporting of violence often differs. Furthermore, unemployed or drink and stay away from home, nearly 5 percent of women, respectively, indicate report higher levels of violence. Female respon- that they have been forced to give up all or part dents whose partners have an unstable job re- 1 While all surveys on GBV likely see a difference between reported violence and actual violence, the primary objective of this survey is to look at links between violence, conflict, economic opportunities and services. As such, the methodology and questionnaire differs from surveys that aim to collect incidence rates. 13 Table E.1: Women’s Vulnerability Profiles to GBV Controlling Behavior/ Sexual Harassment Sexual Violence Physical Violence Emotional Abuse Young Young Young Less educated Less educated Single or separated Less educated Separated Urban Urban Separated Urban Living in smaller household Less adequate housing Urban Living in smaller household Less adequate housing Less affluent Source: Population’s Life Experiences in Georgia Survey, 2016. port higher levels of physical violence and sexual However, conflict-affected individuals per- harassment.2 Finally, former combatants were al- ceive economic stress resulting from the most 2.5 times as likely to experience these forms conflicts to have caused more GBV. While it is of violence compared to those without military not possible to ascertain historic rates of GBV, experience. Men and women with military ser- most IDPs and ALPs surveyed felt that GBV has vice were also twice as likely to experience sexual increased since conflict and displacement. Re- violence or sexual harassment. The analysis also spondents pointed to a decline of economic examined which individuals suffered multiple conditions resulting from the conflict as the driv- forms of violence. Specifically, by looking at in- er of increased GBV risks. They also believe these dividuals who reported more than one incidence economic difficulties led to shame and stress that of violence. These survivors overwhelmingly live manifested as domestic abuse. in Tbilisi or belong to the ALPs group and do not have stable employment. IDPs continue to experience difficulties, though they are economically better-off than The study found few differences in reported rural people or the national population. Re- violence between conflict-affected and non- search demonstrated that IDPs have in many cas- conflict-affected people, though impacts are es transitioned from their previous livelihoods in more acute for specific subgroups. In general, agriculture and manufacturing into retail and ser- people affected by conflict and displacement do vice sectors, as they fled from rural areas to the not report higher experience of GBV in Georgia. cities. IDPs surveyed display both higher unem- However, there are important exceptions that ployment and less secure employment than the warrant more research and attention. Internal- national population. On average, IDPs live in the ly displaced women living in collective centers most cramped dwellings. Although three out of report more sexual harassment and sexual vio- four IDPs are urban and live in more expensive lence. Administrative line women report more cities, their per capita consumption rate is quite physical violence, though this could also be an high, higher than the national average. Based on expression of more patriarchal gender norms in the multidimensional poverty index (MPI) utilized those areas. Men who have direct conflict experi- in this report, IDP poverty fell back below the na- ence report more experience with sexual harass- tional average. ment and sexual violence than men with no such experience. This suggests that individuals’ expe- A lack of economic resources limits women’s riences with conflict and consequences of con- choices in cases of GBV. Over half of all wom- flict, such as difficult living conditions can affect en declared they had no personal income, com- their experiences of GBV, though a straightfor- pared to less than a quarter of men. Men also ward relationship between conflict and the FFV control most household assets. Even in the cases included in this study was not found. where women are the legal owners of jewelry or even land, it is often the man who controls that 2 From Table 36 to Table 40 in Annex 3 14 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES wealth. Although women are considered equally tional experience suggests the risk for GBV is capable as men, they often lack the job skills or lower when roles inside and outside the home networking capabilities to find work and pursue are more equal (Heise and Kotsadam 2015). economic independence. This means that many women are almost fully reliant on their partners POLICY IMPLICATIONS AND or families for subsistence. Further, it means that women cannot easily report abusive behav- RECOMMENDATIONS ior without risking their own economic survival. The stakeholder analysis and institution- This risk may also contribute to underreporting al mapping conducted for the study under- among women. In GBV cases in which the wom- scores that Georgia has a relatively robust po- en fled the home, of those that returned, 36 per- cent of responses were related to not possessing litical and institutional framework to combat enough financial resources to stay away. GBV, but additional capacity building, financ- ing and better implementation are needed. In Women feel without recourse for reporting 2010, Georgia adopted the Gender Equality Law GBV and have limited knowledge and expe- which defines gender equality as “a part of hu- rience of services. Although many laws and man rights referring to equal rights and obliga- services exist to protect women, either they are tions, responsibilities and equal participation of not implemented effectively or women have little men and women in all spheres of personal and confidence in them. Over 70 percent of female public life.” This led to corresponding National respondents feel that existing laws do not offer Action Plans in 2011, 2014 and 2016. Georgia is sufficient protection. Besides a lack of confidence also a signatory to key international conventions in the police and the legal system, survivors are on protecting women from violence, including often unaware of the services available to them. on women and conflict. In 1994, the Georgian For example, only 4 percent of respondents in- Parliament ratified the United Nations (UN) Con- dicate they are aware of domestic shelters. The vention on the ‘Elimination of All Forms of Dis- society and even their own families believe GBV crimination against Women’ (CEDAW). In 2011, should remain a private matter. As such, women the Parliament of Georgia adopted the first Ac- continue to endure abuse while remaining un- tion Plan for implementation of UN Security willing or unable to talk about it. Council Resolution 1325 on ‘Women, Peace and Security’. Furthermore, in 2014, Georgia signed Gender norms may contribute to continued ‘the Council of Europe Convention on prevent- GBV, but a shift is taking place among the ing and combating violence against women and youth. Expectations placed upon how women and men should behave are stark in Georgia. domestic violence’, better known as the Istanbul Most Georgians believe the man’s main role in Convention. the family is being the primary breadwinner, and Major milestones in combating GBV in Geor- the woman’s is to take care of the home. Although women and men do not condone violence, vio- gia were the 2006 Anti-Domestic Violence lence inside the family is not often talked about Law and the 2009 National Referral Mecha- and is sometimes seen as a fact of life, particu- nism on Domestic Violence as well as a series larly when the violence is sporadic. Rural, older of National Action Plans related to combating men hold the most patriarchal gender views, domestic violence and on women in conflict. while younger Georgians, especially women, These have included proposed whole-of-gov- have the most egalitarian views regarding gen- ernment responses to counter GBV related chal- der. Although roles within the home are sharply lenges, including actions by law enforcement defined, Georgians also overwhelmingly believe agencies and departments of health and so- that women are as capable as men in leadership cial services. However, key informant interviews roles both within the community and at the na- pointed to gaps in policy action and a lack of tional level. This is interesting, because interna- dedicated budgets for their implementation. 15 The Government of Georgia has also provided over capacity building for implementation across a range of services to survivors of domestic vi- relevant agencies and departments. This would olence. The State Fund for Protection and Assis- also include a robust monitoring plan to estab- tance of (Statutory) Victims of Human Trafficking lish implementation milestones and reporting (ATIPFUND) was tasked with providing services requirements, and to provide updates to GoG to domestic violence survivors in 2009 and is the leadership and the public. main institution providing assistance to survivors of domestic violence and human trafficking. The Consider women’s economic security as a key range of services include shelter, medical and dimension of the GBV challenge and a focus psychological assistance, legal aid, and rehabil- in GBV prevention and response efforts. Study itation services to the victims/survivors of sexu- findings point to the potential links between con- al violence (Government of Georgia 2014c). A flict, economic stress and GBV risk as well as the number of nongovernmental organizations also value of economic security in women accessing provide services to survivors and support aware- support and GBV services. Law enforcement and ness raising on GBV. However, survey findings social service agencies have traditionally taken point to a lack of awareness and utilization of ser- the lead on GBV prevention and response in vices offered. Georgia; however, a more holistic response, in- cluding a focus on employment opportunities There is a complex network of stakeholders and economic independence could help in ad- working on GBV issues in Georgia. Because no dressing the challenge. single entity is responsible for all government ac- tivities related to GBV, many separate stakehold- Invest in strategic communication, recogniz- ers focus on different aspects of GBV. The closest ing the gaps in knowledge particularly on GBV entity to a coordinating body is the Interagency services. Survey research also indicated a poten- Council for the Prevention of Domestic Violence, tial link between awareness-raising campaigns but they do not provide services, which fall with- and more egalitarian views on gender. However, in the realm of eight different government agen- less than half of the population has even seen a cies. The rapid policy reforms to protect survivors GBV campaign. While general awareness cam- from GBV mean that many agencies still lack the paign can play a role in shifting popular percep- capacity for proper implementation. tions around GBV, they should focus more on advertising available services for GBV survivors, Based on findings of the survey research and the particularly shelters. Information campaigns results of the stakeholder analysis and institution- should be tailored to target audiences, including al mapping, the following are recommendations to the vulnerable groups identified in the study— for a development response to challenges of e.g. urban women, those in challenging housing GBV. environments, former combatants, and youth/ young brides. To the Government of Georgia (GoG): To international development partners, includ- Focus additional resources and attention on ing the World Bank: implementation of existing GBV-related legis- lation and policy reforms. The recent national Provide capacity building support to insti- action plans provide an enabling policy environ- tutions involved in prevention and response ment for robust action on GBV prevention and to GBV in Georgia. Support could include for response. However, individual agency roles and training, expansion of services, mainstreaming responsibilities need to be further defined and attention to GBV in line agencies, and analytics/ associated with dedicated budgets and quali- tracking of GBV trends. For example, the State fied personnel. The mandate of the Interagency Fund for Protection and Assistance of (Statutory) Council for the Prevention of Domestic violence Victims of Human Trafficking (ATIPFUND) and the could be further expanded to ensure oversight Social Service Agency are required to provide 16 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES social services to survivors of domestic violence. Invest in housing and support for IDP collec- However, there is limited capacity within the tive centers while taking into account GBV Agencies to implement these new mandates. De- risk. Study findings point to the potential link be- velopment partners could also support the GoG tween housing challenges and increased prev- through bringing the experiences of other coun- alence of GBV, including for IDPs living in often tries and region in successful implementation overcrowded collective housing. Development of GBV-related legislation and policy, including partners can consider GBV prevention and re- coordinated cross-agency responses. Partners sponse strategies in designing housing solutions can also encourage oversight and monitoring by for these vulnerable groups. non-governmental organizations to hold GoG to account in implementation of existing legislation Deepen knowledge and track progress in re- and policy reforms. ducing GBV. Leveraging information provided by this survey or upcoming surveys could help Mainstream attention to GBV in large-scale gauge GBV trends. It would also help understand development projects, including a focus on the pervasiveness of underreporting and any im- those that could advance economic opportu- provements in that regard. Future research could nities and employment for vulnerable wom- further explore profiles of those most vulnerable en. Study results point to vulnerable populations to GBV and allow for better targeting of preven- that could be targeted through development tion efforts, information campaigns and services. programming. In particular, projects that focus This would include a future focus on men’s ex- on employment-generating activities and skills perience of violence and risk factors to adapt development to help increase options to women future services and support. In addition, apply- affected by GBV. Gains may be even higher for ing longitudinal studies over time regarding the conflict-affected people, who reported the high- same subjects would help track the longer-term est levels of unemployment and unstable work. impacts of conflict and displacement. Improve- They also consistently believed these economic ments, such as these, to GBV’s knowledge-gen- stressors lead to GBV. Employment focused pro- erating mechanisms would allow organizations gramming for IDPs and conflict-affected people to create benchmarks and better target their in- should include attention to potential GBV risks vestments. and investments to increase economic security. გენდერული ძალადობა საქართველოში: კავშირი კონფლიქტებს, ეკონომიკურ შესაძლებლობებსა და მომსახურებას შორის შემაჯამებელი მიმოხილვა ამ ანგარიშში შეჯამებულია კვლევა, რომელიც დაფინანსებულია მსოფლიო ბანკის ფონდის (სახელმწიფოს მშენებლობისა და მშვიდობის მშენებლობის ფონდი SPF) გრანტით - საქართველოში სექსუალური და გენდერული ძალადობის თავიდან აცილებისა და რეაგირების შესაძლებლობების გაძლიერება. გრანტის მიზანია საქართველოში გენდერული ძალადობის თავიდან აცილების/პრევენციის და რეაგირების შესახებ ცოდნისა და შესაძლებლობების განვითარება, კერძოდ, ყურადღება გამახვილებულია კონფლიქტისა და იძულებითი გადაადგილების შედეგად დაზარალებულ მოსახლეობაზე, ეკონომიკურ შესაძლებლობებსა და მომსახურებაზე. ეს პროექტი არის კონფლიქტისა და საქართველოში ჩატარდა კვლევა, სადაც გენდერული ნიშნით ძალადობის (GBV) გათვალისწინებულია როგორც ქვეყანაში შესახებ მსოფლიო ბანკის გლობალური არსებული პრობლემები კონფლიქტებთან ინიციატივის ნაწილი. გლობალური და ადგილნაცვალ/გადაადგილებულ ინიციატივა, რომელიც დაფინანსებულია მოსახლეობასთან მიმართებაში, ასევე სახელმწიფოს მშენებლობისა და მშვიდობის საქართველოს მთავრობის მიერგადადგმული მშენებლობის ფონდის SPF-ის მიერ, მოიცავს ბოლოდროინდელი ნაბიჯები გენდერულ საპილოტე პროექტებს აღმოსავლეთ აზიასა ღონისძიებებთან და გენდერულ და ოკეანეთში, აფრიკასა და სამხრეთ აზიაში ძალადობაზე რეაგირებასთან დაკავშირებით. (ჯამურად დაახლოებით 10 მილიონია აფხაზეთისა და სამხრეთ ოსეთის რეგიონების გამოყოფილი პროექტების დაფინანსებაზე). ამ 1990 და 2008 წწ კონფლიქტების ინიციატივის მიზანია გაიზარდოს გენდერული გათვალისწინებით, საქართველოში ძალადობის განვითარების ასპექტებისა და კონფლიქტის შედეგად დაზარალებული მსოფლიო ბანკისა და განვითარების საკითხებში მოსახლეობის დიდი რაოდენობაა. იგი მისი პარტნიორების გაფართოებული მოიცავს დაახლოებით 190 000-დან 275 პროგრამირების სფეროების გაგება. მაშინ, 000-მდე იძულებით გადაადგილებულ პირს, როდესაც გლობალური ინიციატივა მხარს უჭერს რომლებიც კონფლიქტის შედეგად არიან ოპერატიულ პროექტებს სხვა რეგიონებში, გადაადგილებულები და რომლებიც შეადგენენ საქართველოს საპილოტე პროექტი, რომელიც მთელი მოსახლეობის თითქმის 6 პროცენტს, წარმოადგენს ევროპისა და ცენტრალური რაც მსოფლიოში შედარებითი პროპორციის აზიის რეგიონს (ECA), უნიკალურია, რადგანაც ერთერთი ყველაზე მაღალი მაჩვენებელია. განსაკუთრებულ ყურადღებას აქცევს ცოდნის გარდა ამისა, აფხაზეთისა და სამხრეთ გაღრმავებასა და შესაძლებლობების ოსეთის რეგიონების ყოფილ კონფლიქტის განვითარების მხარდაჭერას. ზონებთან ახლოს მცხოვრები ადამიანები კვლავ 17 18 გენდერული ძალადობა საქართველოში: კავშირი კონფლიქტებს, ეკონომიკურ შესაძლებლობებსა და მომსახურებებს შორის ნახ. E.1: კვლევის დროს დასმული კითხვები კონფლიქტი ƒƒ უფრო მეტად ემუქრებათ თუ არა გენდერული ძალადობის საშიშროება კონფლიქტის შედეგად დაზარალებულებს? ƒƒ რამდენად შასაძლებელია, რომ იძულებითი გადაადგილებით გამოწვეულმა სტრესმა გაზარდოს ეს საშიშროება? ძალადობა მომსახურება საარსებო წყაროები და ეკონომიკური შესაძლებლობები ƒƒ რა ტიპის მომსახურება სჭირდებათ ƒƒ კონფლიქტის შედეგად ƒƒ როგორი გავლენა მოახდინა კონფლიქტმა დაზარალებულებს? ƒƒ კონფლიქტის შედეგად დაზარალებულების ƒƒ რა ტიპის მომსახურება სჭირდება საარსებო წყაროებზე? ზოგადად მოსახლეობას? როგორია ƒƒ ახდენს თუ არა გავლენას ძალადობაზე მომსახურების შიდამეურნეობების ეკონომიკა? ƒƒ შესახებ ინფორმაციის დონე და ƒƒ რა ეკონომიკური შესაძლებლობები და ადაპტაციის მომსახურებაზე წვდომა? მექანიზმებია ხელმისაწვდომი გადარჩენილთათვის? განიცდიან პერიოდულ დაუცველობასა და რადგანაც მათ შეეძლოთ გავლენა მოეხდინათ მათ საცხოვრებელ პირობებსა და საარსებო გამოცდილებასა და შესაძლებლობებზე საშუალებებზე ზემოქმედების შედეგებს. კვლევის თითოეულ სფეროში. მოცემული კვლევის მიზანია შეავსოს მთავრობის, კვლევა ადარებს შეიარაღებული კონფლიქტის საერთაშორისო პარტნიორებისა და შედეგად დაზარალებული ორი ჯგუფის, არასამთავრობო ორგანიზაციების არსებული ქვეყნის შიგნით იძულებითი გადაადგილებული ინიციატივები საქართველოში გენდერულ პირებისა/ (IDPs) და ადმინისტრაციული ძალადობასთან/ GBV მიმართებაში და საზღვრისპირას მცხოვრები პირების/ (ALPs) შეისწავლოს შესაძლებლობები ამ სფეროში გამოცდილებას და იმ პირების გამოცდილებას, შემდგომი პროგრესისათვის. კერძოდ, შეავსოს რომლებიც არ არიან კონფლიქტის შედეგად ხარვეზები პოტენციურ კავშირებზე როგორც დაზარალებულები/ (NAPs) და აანალიზებს გენდერულ ძალადობასა/ GBV და კონფლიქტსა ძალადობის ოთხ ფორმას/ (FFV), კერძოდ, და შიდა გადაადგილებას შორის, ასევე მაკონტროლებელი ქცევა/ემოციური ხარვეზები ეკონომიკურ შესაძლებლობასა და შევიწროვება, სექსუალური შევიწროვება/ მომსახურებასთან დაკავშირებით. დევნა, სექსუალური ძალადობა და ფიზიკური ძალადობა. დაინტერესებულ მხარეთა ანალიზი კვლევა მოიცავდა ხარისხობრივი და და ინსტიტუციური ასახვა/მეპინგი გამოყენებულ რაოდენობრივი მეთოდების შერწყმას, იქნა განვითარების სფეროში სამომავლო მათ შორის, დაინტერესებულ მხარეთა მოქმედებებისათვის სახელმძღვანელო ანალიზს, ინტერვიუებს ძირითად რეკომენდაციების მისაცემად. ინფორმანტებთან, ფოკუს ჯგუფების სახალხო დისკუსიებსა და ეროვნულ- ძირითადი შედეგები წარმომადგენლობით რაოდენობრივ კვლევას. (იხ. ნახ. E.1). კვლევა აგრეთვე კვლევის დროს დასმული კითხვებიდან შეეცადა გაერკვია გენდერული ნორმები, გამომდინარე, გამოიკვეთა შემდეგი 19 ცხრილი E.1: ქალთა მოწყვლადობის/დაუცველობის მოდელები და გენდერული ნიშნით ძალადობა / GBV სექსუალური მაკონტროლებელი ქცევა სექსუალური ძალადობა ფიზიკური ძალადობა შევიწროვება/დევნა ახალგაზრდა; ახალგაზრდა; ახალგაზრდა; უმაღლესი განათლების უმაღლესი განათლების დაუოჯახებელი ან უმაღლესი განათლების გარეშე; გარეშე; გაშორებული; გარეშე; გაშორებული; ქალაქში მცხოვრები; ქალაქში მცხოვრები; გაშორებული; ქალაქში მცხოვრები; მცხოვრები მცირე ზომის ნაკლებად ადეკვატური ქალაქში მცხოვრები; მცხოვრები მცირე ზომის ოჯახში საცხოვრებელი ნაკლებად ადეკვატური ოჯახში; პირობების მქონე საცხოვრებელი ნაკლებად შეძლებული პირობების მქონე ოჯახის მქონე წყარო: მოსახლეობის ცხოვრებისეული გამოცდილების კვლევა საქართველოში, 2016. ძირითადი შედეგები: მოწყვლადები/ დაუცველები არიან ძალადობის საკითხში. კვლევის პროცესში კვლევაში მონაწილე რესპონდენტების შეიქმნა „მოწყვლადობის/ დაუცველობის მეოთხედზე მეტს აქვს ძალადობის მოდელები“ რათა ეჩვენებინათ გამოკითხვის გარკვეული ფორმის გამოცდილება. მონაწილეთა რომელი ჯგუფები ავლენენ გამოკითხულ ქალთა 27 პროცენტი აცხადებს, ძალადობის ყველაზე მეტ ტიპებს. (იხ. რომ გამოცდილი აქვთ ძალადობის ოთხი ცხრილი E.1). სავარაუდოდ, ქალაქში ფორმიდან - ემოციური შევიწროვება, მცხოვრები ქალები უფრო მეტად ძალადობის სექსუალური დევნა, სექსუალური შევიწროვება, შემთხვევებზე საუბრობენ, მაშინ, როდესაც ფიზიკური შეურაცხყოფა - სულ მცირე, ერთი უარეს საცხოვრებელ პირობებში მცხოვრები ფორმა მაინც. გამოკითხულთა შორის 16.3 ქალები უფრო მეტად სექსუალურ პროცენტმა განაცხადა, რომ მსხვერპლია შევიწროვებასა და სექსუალურ ძალადობაზე ძალადობის ერთი ფორმის მაინც (სექსუალური საუბრობენ. როგორც გამოკითხვამ აჩვენა, დევნა/ შევიწროება, სექსუალური ძალადობა სხვებთან შედარებით ქალები, რომლებიც 18 ან ორივე), ხოლო ქალთა 8 პროცენტმა წლამდე ასაკში დაქორწინდნენ უფრო მეტად განაცხადა, რომ ფიზიკური ძალადობის განიცდიან როგორც ემოციურ შეურაცხყოფას, მსხვერპლი იყო. აქვე უნდა აღინიშნოს, ასევე სექსუალურ და ფიზიკურ ძალადობას. რომ მოცემული პროცენტები არ აღნიშნავს ანალიზი გვიჩვენებს, რომ ისეთი პარტნიორების შემთხვევების რაოდენობას1 და რომ ქალების რესპონდენტები საუბრობენ ძალადობის უფრო მიერ რეალურად გამოცდილი ძალადობა მაღალ დონეზე, რომელთა პარტნიორებიც და მათ მიერ ამ ძალადობის აღრიცხვიანობა/ იმყოფებიან სტრესის ქვეშ, უმუშევრები არიან ან შეტყობინება ხშირად განსხვავდება სვამენ და სახლში არ ათევენ ღამეს. ქალები, ერთმანეთისაგან. მეტიც, შესაბამისად, ქალთა რომელთა პარტნიორებს არასტაბილური თითქმის 5 პროცენტი აღნიშნავს, რომ მათ სამსახური აქვთ ასევე აცხადებენ, რომ ისინი ძალა დაატანეს, რომ მთელი მათი შემოსავალი განიცდიან უფრო მეტ ფიზიკურ ძალადობასა ან შემოსავლის ნაწილი საკუთარი სურვილის და სექსუალურ შევიწროვებას.2 და ბოლოს, გარეშე დაეთმოთ მეუღლისათვის ან თავი ყოფილი მებრძოლები, სავარაუდოდ, დაენებებინათ სამსახურისათვის, რადგანაც ეს დაახლოებით 2,5 ჯერ უფრო მეტად მათი პარტნიორის სურვილი იყო. განიცდიდნენ ძალადობის ამ ფორმებს, ვიდრე ისინი, რომელთაც სამხედრო გამოცდილება კონკრეტული ქვეჯგუფები უფრო 1 თუმცა გენდერულ ძალადობასთან დაკავშირებულ ყველა გამოკითხვაში ჩანს განსხვავება ძალადობის ფაქტის შეტყობინებასა და რეალურ ძალადობას შორის, ჩვენი კვლევის ძირითად ამოცანას მაინც იმ კავშირების კვლევა წარმოადგენს, რომელიც არსებობს ძალადობას, კონფლიქტს, ეკონომიკურ შესაძლებლობებსა და მომსახურებას შორის. აქედან გამომდინარე, ჩვენი მეთოდიკა და კითხვარი განსხვავდება იმ გამოკითხვებისაგან, რომელთა მიზანსაც შემთხვევების რაოდენობის შეგროვება წარმოადგენს. 2 დანართი 3, ცხრილი 36-40 20 გენდერული ძალადობა საქართველოში: კავშირი კონფლიქტებს, ეკონომიკურ შესაძლებლობებსა და მომსახურებებს შორის არ გააჩნდათ. სამხედრო სამსახურში მყოფი დაკავშირებით, თუმცა, ამავე დროს, უნდა მამაკაცები და ქალები სავარაუდოდ, ორჯერ აღინიშნოს, რომ ამ კვლევაში არ მოიძებნა მეტად განიცდიდნენ სექსუალურ ძალადობასა პირდაპირი კავშირი კონფლიქტსა და და სექსუალურ შევიწროვებას/ დევნას. ძალადობის ოთხ ფორმას/ FFV შორის. ანალიზმა ასევე ცხადყო, რომ არსებობდნენ ცალკეული პირები, რომლებმაც ძალადობის თუმცა კონფლიქტის შედეგად მრავალი ფორმა გამოსცადეს. კონკრეტულად, დაზარალებული პირები მიიჩნევენ, რომ გაანალიზდნენ ის პირები, რომლებმაც ისაუბრეს უფრო მეტი გენდერული ძალადობის/ GBV ძალადობის ერთზე მეტი შემთხვევის შესახებ. შემთხვევების მიზეზი არის კონფლიქტის ამ გადარჩენილთა უმეტესობა თბილისში შედეგად მიღებული ეკონომიკური ცხოვრობს ან მიეკუთვნება ადმინისტრაციულ სტრესი. მიუხედავად იმისა, რომ გენდერულ საზღვრისპირას მცხოვრებთა რიცხვს და არ ძალადობასთან დაკავშირებული ისტორიული არის სტაბილურად დასაქმებული. მონაცემები ძნელი დასადგენია, გამოკითულთა უმეტესობას, როგორც ქვეყნის შიგნით კვლევამ გამოავლინა, რომ ძალადობის იძულებით გადაადგილებულ პირებს/ IDPs, თვალსაზრისით არსებობს მცირე განსხვავება ასევე საზღვრისპირა მცხოვრებ პირებს/ ALPs იმ ადამიანებს შორის, რომლებიც მიაჩნიათ, რომ კონფლიქტისა და იძულებითი კონფლიქტის შედეგად დაზარალდნენ და გადაადგილების გამო გენდერული ძალადობის/ მათ შორის, ვინც კონფლიქტის შედეგად GBV შემთხვევების რიცხვი გაიზარდა. არ დაზარალებულა, თუმცა კონკრეტული რესპონდენტებმა აღნიშნეს, რომ კონფლიქტის ქვეჯგუფებისათვის ზემოქმედება უფრო შედეგად ეკონომიკური პირობების გაუარესებამ მწვავეა. ზოგადად, კონფლიქტის შედეგად წამყვანი როლი შეასრულა გენდერული დაზარალებული და ქვეყნის შიგნით იძულებით ძალადობის რისკების ზრდის საკითხში. მათ გადაადგილებული პირები საქართველოში ასევე აღნიშნეს, რომ მოცემულმა ეკონომიკურმა გენდერული ძალადობის უფრო მეტ სიძნელეებმა გამოიწვია სირცხვილი და სტრესი, შემთხვევაზე არ საუბრობენ. თუმცა, არსებობენ რომელმაც თავის მხრივ, გამოხატვა ჰპოვა მნიშვნელოვანი გამონაკლისებიც, რომლებიც ოჯახურ ძალადობაში. უფრო მეტ კვლევასა და ყურადღებას მოითხოვენ. ქვეყნის შიგნით იძულებით ქვეყნის შიგით იძულებით გადაადგილებული გადაადგილებული ქალები, რომლებიც პირები/ IDPs კვლავ განიცდიან დროებითი განსახლების ცენტრებში ცხოვრობენ სირთულეებს, თუმცა ეკონომიკურად მათი უფრო მეტად განიცდიან სექსუალურ მდგომარეობა უკეთესია, ვიდრე სოფლად შევიწროვებასა და სექსუალურ ძალადობას. მცხოვრები ან ეროვნული მოსახლეობისა. ადმინისტრაციულ საზღვრისპირას მცხოვრები კვლევამ ცხადყო, რომ დევნილები უმეტეს ქალები კი უფრო მეტად ფიზიკურ ძალადობას შემთხვევაში ტოვებდნენ ადრინდელი უჩივიან, თუმცა ეს შესაძლოა დასახლების საარსებო წყაროებს სოფლის მეურნეობასა ადგილებში პატრიარქალური გენდერული და მრეწველობაში და გადადიოდნენ საცალო ნორმის გამოხატულების ბრალი იყოს. ვაჭრობასა და მომსახურებაში, რადგანაც მამაკაცები, რომელთაც კონფლიქტის უშუალო ისინი მასიურად სახლდებოდნენ სოფლებიდან გამოცდილება გააჩნიათ აცხადებენ, რომ ქალაქებში. დანარჩენ მოსახლეობასთან გააჩნიათ უფრო მეტი გამოცდილება სექსუალურ შედარებით იძულებით გადაადგილებულ შევიწროვების/ დევნის და სექსუალური პირებში გამოვლინდა როგორც უმუშევრობის ძალადობისა, ვიდრე იმ მამაკაცებს, უფრო მაღალი დონე, ასევე უფრო ნაკლებად რომელთაც კონფლიქტთან პირდაპირი შეხება უსაფრთხო დასაქმებულობა. ზოგადად, არ ჰქონიათ. ეს გულისხმობს, რომ ცალკეულ დევნილები ყველაზე ვიწროდ ცხოვრობენ. პირთა გამოცდილებამ კონფლიქტთან და მიუხედავად იმისა, რომ ოთხი დევნილიდან კონფლიქტის შედეგებთან დაკავშირებით, სამი საკმაოდ ძვირ ქალაქში ცხოვრობს, მათი როგორიცაა მძიმე საცხოვრებელი პირობები, მოხმარების მაჩვენებელი ერთ სულ მოსახლეზე შეიძლება გავლენა მოახდინოს მათ საკმაოდ მაღალია, გაცილებით მაღალი, გამოცდილებაზე გენდერულ ძალადობასთან ვიდრე მოხმარების ეროვნული საშუალო 21 მაჩვენებელი. ამ ანგარიშში გამოყენებული აღსრულება არ ხდება ეფექტიანად, ან თავად მრავალგანზომილებიანი სიღარიბის ქალები ნაკლებად ენდობიან მათ. გამოკითხულ მაჩვენებლის (MPI) მიხედვით, იძულებით ქალთა 70 პროცენტზე მეტი აცხადებს, რომ გადაადგილებულ პირთა სიღარიბის დონე არსებული კანონები ვერ უზრუნველყოფენ ეროვნულ საშუალო მაჩვენებელზე დაბლა მათთვის საჭირო შესაბამის დაცვას. გარდა დაეცა. პოლიციისა და იურიდიული სისტემის მიმართ უნდობლობისა, დაზარალებულებმა ეკონომიკური რესურსების ნაკლებობა ხშირად არ იციან რა მოსახურება არის ზღუდავს ქალის არჩევანს გენდერული მათთვის ხელმისაწვდომი. მაგალითად, ძალადობის შემთხვევაში. გამოკითხული გამოკითხულთა მხოლოდ 4 პროცენტმა იცის ქალების ნახევარზე მეტმა განაცხადა, რომ საოჯახო თავშესაფრების არსებობა. მეტიც, მათ პირადი შემოსავალი არ გააჩნიათ, საზოგადოებაც და მათი საკუთარი ოჯახებიც შედარებისათვის, მამაკაცების მხოლოდ მიიჩნევენ, რომ გენდერული ძალადობა პირად მეოთხედზე ნაკლებმა განაცხადა იგივე. საკითხად უნდა დარჩეს, აქედან გამომდინარე, მამაკაცები ასევე აკონტროლებენ ქალები კვლავ აგრძელებენ შევიწროებისა და შიდამეურნეობის აქტივების უმეტეს ნაწილს. შეურაცხყოფის მოთმენას და არ სურთ ან არ იმ შემთხვევაშიც კი, როდესაც ქალები შეუძლიათ მასზე საუბარი. იურიდიულად არიან ძვირფასეულობის ან მიწის მფლობელები, ხშირად მაინც ისევ მამაკაცები გენდერულ ნორმებს შესაძლოა თავისი განკარგავენ მათ ქონებას. მიუხედავად წვლილი შეაქვთ გენდერული ძალადობის იმისა, რომ ქალი მამაკაცის თანასწორია, მას საქმეში, მაგრამ ახალგაზრდობაში მაინც ხშირად არ გააჩნია დასაქმებისათვის უკვე შეინიშნება ამ მხრივ გარკვეული საჭირო უნარ-ჩვევები ან არ იცის ინტერნეტის ცვლილებები. საქართველოში საკმაოდ გამოყენება სამსახურის მოსაძიებლად, რომ მკაფიოდ არის განსაზღვრული ქალისა და დაიწყოს მუშაობა და მოიპოვოს ეკონომიკური მამაკაცის ქცევის ნორმები. ქართველთა დამოუკიდებლობა. ეს კი, თავის მხრივ ნიშნავს, უმეტესობას სწამს, რომ მამაკაცის უპირველესი რომ ბევრი ქალი არსებობისათვის თითქმის როლი ოჯახში გამოიხატება მის მიერ მთლიანად არის დამოკიდებული საკუთარ უპირველესად ოჯახის რჩენაში, ხოლო პარტნიორზე ან ოჯახზე. გარდა ამისა, ეს ასევე ქალმა კი ოჯახს უნდა მიხედოს. მიუხედავად გულისხმობს, რომ ქალებს არ შეუძლიათ იმისა, რომ არც ქალები და არც მამაკაცები თავისუფლად განაცხადონ მათ მიმართ მიესალმებიან ძალადობას, ოჯახურ განხორციელებული შეურაცხმყოფელი ქცევის ძალადობაზე მაინც ხშირად არ საუბრობენ შესახებ, რადგანაც რისკის ქვეშ დგება მათი და ზოგჯერ მას ჩვეულებრივ, ცხოვრებისეულ ეკონომიკური არსებობის საკითხი. ეს რისკი ფაქტად აღიქვამენ, განსაკუთრებით მაშინ, თუ გარკვულწილად განაპირობებს ქალების იგი/ ძალადობა არარეგულარული ხასიათისაა. მიერ ძალადობის შემთხვევების მიჩქმალვას. სოფლად მცხოვრებ მოხუც მამაკაცებს ისეთი გენდერული ძალადობის შემთხვევისას, ყველაზე პატრიარქალური გენდერული როდესაც ქალები სახლიდან გარბიან, კვლავ მსოფლმხედველობა აქვთ, მაშინ, როცა ოჯახში დაბრუნებული ქალების 36 პროცენტის ახალგაზრდა ქართველებს, განსაკუთრებით დაბრუნების მიზეზი ფინანსური რესურსების ქალებს, გენდერთან დაკავშირებით ყველაზე უქონლობა ხდება. თანასწორი ხედვა აქვთ. მიუხედავად იმისა, რომ ქალისა და მამაკაცის როლები ქალები თავს დაუცველად გრძნობენ ოჯახში მკაცრად არის განსაზღვრული, გენდერული ძალადობის შესახებ ქართველებს მკაფიოდ მიაჩნიათ, რომ განცხადების შემთხვევაში და შეზღუდული ქალს ისევე შეუძლია იტვირთოს ლიდერის ინფორმაცია და გამოცდილება გააჩნიათ როლი, როგორც მამაკაცს, იქნება ეს თემში/ არსებული მომსახურების გამოყენების საზოგადოებაში, თუ ეროვნულ დონეზე. ეს თაობაზე. მიუხედავად იმისა, რომ ქვეყანაში კი ფრიად საინტერესოა, რადგან როგორც მრავალი კანონი და სამსახური არსებობს საერთაშორისო გამოცდილება გვიჩვენებს, ქალთა დასაცავად, ან მათი განხორციელება და გენდერული ძალადობის რისკები მცირდება, 22 გენდერული ძალადობა საქართველოში: კავშირი კონფლიქტებს, ეკონომიკურ შესაძლებლობებსა და მომსახურებებს შორის როცა როლები როგორც ოჯახში, ასევე მის რეფერალურ მექანიზმს ოჯახური ძალადობის გარეთ თანასწორუფლებიანია. (Heise and Kot- შესახებ. ამას მოჰყვა ეროვნული სამოქმედო sadam 2015). გეგმების სერია ოჯახურ ძალადობასთან ბრძოლისა და საქართველოში კონფლიქტის პრაქტიკული დასკვნები და ზონაში ქალების შესახებ. ეს მოიცავდა მთელი მთავრობის მიერ შეთავაზებულ საპასუხო ზომებს რეკომენდაციები გენდერულ ძალადობასთან დაკავშირებულ კვლევისათვის ჩატარებული დაინტერესებულ გამოწვევებთან საბრძოლველად, მათ მხარეთა ანალიზი და ინსტიტუციური ასახვა/ შორის, სამართალდამცავი ორგანოებისა და მეპინგი ხაზს უსვამს, რომ საქართველოს ჯანმრთელობისა და სოციალური სამსახურების გააჩნია შედარებით ძლიერი პოლიტიკური დეპარტამენტების მიერ გასატარებელ და ინსტიტუციური ჩარჩო გენდერულ ღონისძიებებს. თუმცა ინტერვიუები ძირითად ძალადობასთან საბრძოლველად, მაგრამ ინფორმანტებთან მიუთითებდნენ სტრატეგიულ ამავდროულად, დამატებით აუცილებელია ხარვეზებსა და მათი განხორციელებისათვის შესაძლებლობების შემდგომი გაძლიერება, საჭირო გამოყოფილი ბიუჯეტის არარსებობაზე. დაფინანსება და უფრო ეფექტიანი საქართველოს მთავრობამ ასევე განხორციელება. 2010 წელს საქართველოში უზრუნველყო ოჯახური ძალადობის მიღებულ იქნა კანონი გენდერული მსხვერპლთა მომსახურების მთელი თანასწორობის შესახებ, რომელიც გენდერულ სპექტრი. 2009 წელს ადამიანით ვაჭრობის/ თანასწორობას განსაზღვრავს, როგორც ტრეფიკინგის მსხვერპლთა, დაზარალებულთა „ადამიანის უფლებათა ნაწილს, რომელიც დაცვისა და დახმარების სახელმწიფო ფონდს გულისხმობს ქალისა და მამაკაცის თანასწორ (ATIPFUND), დაევალა ოჯახური ძალადობის უფლება-მოვალეობებს, პასუხისმგებლობასა მსხვერპლთათვის დახმარების უზრუნველყოფა. და თანასწორუფლებიან მონაწილეობას ეს არის ძირითადი ორგანიზაცია, რომელიც პირადი და საზოგადოებრივი ცხოვრების ეხმარება ოჯახური ძალადობისა და ყველა სფეროში“. მას მოჰყვა შესაბამისი ტრეფიკინგის მსხვერპლებს. ორგანიზაციის ეროვნული სამოქმედო გეგმა 2011, 2014 და მიერ გაწეული მომსახურების სპექტრი მოიცავს 2016 წწ. საქართველოს აგრეთვე ხელი აქვს სექსუალური ძალადობის მსხვერპლისათვის/ მოწერილი ძირითად ევროპულ კონვენციებზე, დაზარალებულისათვის თავშესაფრის რომელიც გულისხმობს ქალთა დაცვას გამოყოფას, სამედიცინო და ფსიქოლოგიურ ძალადობისაგან, მათ შორის, ხელი მოეწერა დახმარებას, იურიდიულ და სარეაბილიტაციო კონვენციას „ქალებისა და კონფლიქტის მომსახურებას (საქართველოს მთავრობა 2014). შესახებ“. 1994 საქართველოს პარლამენტის მთელი რიგი არასამთავრობო ორგანიზაციები მიერ რატიფიცირებული იყო გაეროს კონვენცია ასევე უწევენ მომსახურებას გენდერული „ქალთა წინააღმდეგ დისკრიმინაციის ყველა ძალადობის შედეგად დაზარალებულებს და ფორმის აღმოფხვრის შესახებ“ (CEDAW). 2011 ეხმარებიან გენდერული ძალადობის შესახებ წელს საქართველოს პარლამენტმა მიიღო ცნობიერების ამაღლებაში. თუმცა, როგორც პირველი სამოქმედო გეგმა გაროს უშიშროების გამოკითხვის შედეგები მიუთითებენ, კვლავ საბჭოს რეზოლუციის N1325 „ქალები, მშვიდობა შეინიშნება ინფორმაციის ნაკლებობა და და უსაფრთხოება“, განხორციელების მიზნით. შეთავაზებული მომსახურების გამოუყენებლობა. უფრო მეტიც, 2014 წელს საქართველომ ხელი მოაწერა ევროპის საბჭოს კონვენციას საქართველოში არსებობს დაინტერესებულ „პრევენცია და ბრძოლა ქალებისა და ოჯახური მხარეთა განვითარებული ქსელი, რომელიც ძალადობის წინააღმდეგ“, რომელიც ცნობილია მუშაობს გენდერული ძალადობის სტამბოლის კონვენციის სახელით. პრობლემებზე. იმის გამო, რომ არც ერთი ცალკე აღებული ორგანიზაცია არ არის საქართველოში გენდერულ ძალადობასთნ პასუხისმგებელი მთელი მთავრობის მიერ ბრძოლის ძირითადი ეტაპები მოიცავდა გენდერულ ძალადობასთან დაკავშირებულ 2006 წლის კანონს ოჯახური ძალადობის გასატარებელ ღონიძიებებზე, მრავალი წინააღმდეგ და 2009 წლის ეროვნულ 23 ცალკეული დაინტერესებული მხარეე გათვალისწინებულ იქნას ქალთა ეკონომიკური მუშაობს გენდერული ძალადობის სხვადასხვა უსაფრთხოება, რაც თავის მხრივ გენდერულ ასპექტებზე. მაკოორდინებელ ორგანოსთან ძალადობასთან დაკავშირებული გამოწვევის ყველაზე ახლოს დგას ოჯახური ძალადობის ძირითადი ასპექტია და ყურადღება პრევენციის უწყებათაშორისი კომისია, თუმცა გამახვილდეს გენდერული ძალადობის ისინი დაზარალებულებს არ უზრუნველყოფენ თავიდან აცილებისა და რეაგირების მომსახურებით - მომსახურებით უზრუნველყოფა ძალისხმევაზე. კვლევის შედეგები მიუთითებენ, შედის რვა სხვადასხვა სახელმწიფო რომ არსებობს პოტენციური კავშირი როგორც ორგანიზაციის მოქმედების სფეროში. სწრაფი კონფლიქტს, ეკონომიკურ სტრესსა და პოლიტიკური რეფორმები გენდერული გენდერული ძალადობის რისკებს შორის, ძალადობის მსხვერპლთა დაცვის საკითხებში ასევე ქალთა ეკონომიკურ უსაფრთხოებასთან მიუთითებს იმაზე, რომ ბევრ ორგანიზაციას და მათ მიერ დახმარებისა და გენდერული კვლავ არ აქვს ამ რეფორმების სათანადოდ ძალადობის სერვისებთან ხელმისაწვდომობას გატარების რესურსი და შესაძლებლობა. შორის. სამართალდამცავმა და სოციალური მომსახურების სააგენტოებმა საკუთარ თავზე გამოკითხვის კვლევისა და დაინტერესებული აიღეს საქართველოში გენდერული ძალადობის მხარეების ანალიზისა და ინსტიტუციური პრევენციისა და მასზე რეაგირების საკითხების ასახვის/მეპინგის შედეგების საფუძველზე გადაწყვეტა, თუმცა, უნდა აღინიშნოს, რომ ქვემოთ მოცემულია გენდერული ძალადობის უფრო კომპლექსური მიდგომა, რომელიც გამოწვევებზე რეაგირების შემდეგი ყურადღებას გაამახვილებს დასაქმების რეკომენდაციები. შესაძლებლობებსა და ეკონომიკურ დამოუკიდებლობაზე უფრო მეტად დაეხმარება რეკომენდაციები საქართველოს ამ გამოწვევის გადაჭრაში. მთავრობისათვის (GoG): საჭიროა სტრატეგიული კომუნიკაციების ყურადღება მიექცეს დამატებით ინვესტირება, ინფორმირებულობის რესურსებსა და გენდერულ საკითხებში ხარვეზების აღიარება - ეს ძალადობასთან დაკავშირებული არსებული განსაკუთრებით ეხება გენდერული კანონმდებლობისა და სტრატეგიული ძალადობასთან დაკავშირებული რეფორმების განხორციელებას. მომსახურების შესახებ ინფორმირებულობას. ბოლოდროინდელი ეროვნული სამოქმედო გამიკითხვების შესწავლამ ასევე მიუთითა, გეგმები უზრულველყოფენ შესაფერის რომ არსებობს პოტენციური კავშირი პოლიტიკურ გარემოს გენდერული ძალადობის ცნობიერების/ინფორმირებულობის ამაღლებასა თავიდან აცილებისა და რეაგირების და გენდერთან დაკავშირებულ უფრო პრობლემების სწრაფი გადაჭრისათვის. თანასწორუფლებიან შეხედულებების შორის. თუმცა საჭიროა დამატებით განისაზღვროს ამასთანავე უნდა აღინიშნოს, რომ მოსახლეობის ინდივიდუალური სახელმწიფო სააგენტოების მხოლოდ ნახევარზე ნაკლებს აქვს ნანახი როლი და პასუხისმგებლობა, გამოიყოს გენდერული ძალადობის წინააღმდეგ გაწეული მათთვის ბიუჯეტი და დაკომპლექტდეს კამპანია, მაშინ, როცა ზოგადად ამ საკითხთან კვალიფიცირებული კადრებით. უნდა დაკავშირებული ცნობიერების ამაღლების გაფართოვდეს ოჯახური ძალადობის კამპანიას შეუძლია შეცვალოს პოპულარული პრევენციის უწყებათაშორისი კომისიის მანდატი, სტერეოტიპები გენდერულ ძალადობის რაც უზრუნველყოფს შესაძლებლობების შესახებ, უფრო მეტი ყურადღება უნდა მიექცეს განვითარების განხორციელებაზე გენდერული ძალადობის მსხვერპლთათვის ზედამხედველობას შესაბამისი სააგენტოებისა ხელმისაწვდომი მომსახურების რეკლამირებას, და დეპარტამენტების მეშვეობით. განსაკუთრებით კი თავშესაფრებს. აუცილებელია მონიტორინგის საიმედო საინფორმაციო კამპანიები უნდა მორგებული გეგმა, რომელიც შექმნის განხორციელების იყოს მიზნობრივ აუდიტორიაზე, მათ შორის ეტაპებს და ანგარიშგების მოთხოვნებს და მოწყვლად ჯგუფებზე, რომლებიც ჩვენს ასევე განახლებულ ინფორმაციას მიაწვდის კვლევაში გამოიკვეთა - მაგ. ქალაქში საქართველოს მთავრობასა და საზოგადოებას. 24 გენდერული ძალადობა საქართველოში: კავშირი კონფლიქტებს, ეკონომიკურ შესაძლებლობებსა და მომსახურებებს შორის მცხოვრები ქალები, ყოფილი მებრძოლები, დაუცველ/მოწყვლად ქალებს ეკონომიკური მძიმე საცხოვრებელი პირობების მქონე შესაძლებლობებისა და დასაქმების კუთხით. ადამიანები და ახალგაზრდები/ ნაადრევად ჩვენი კვლევის შედეგები მიუთითებს, რომ გათხოვილი გოგონები (ახალგაზრდა არსებობს მოსახლეობის სოციალურად პატარძლები) დაუცველი ნაწილი, რომელიც უნდა გახდეს განვითარების პროგრამების ძირითადი სამიზნე. რეკომენდაციები საერთაშორისო ეს განსაკუთრებით ეხება იმ პროექტებს, განვითარების პარტნიორებს, მათ შორის, რომლებიც ყურადღებას ამახვილებენ მსოფლიო ბანკს: დასაქმებასა და უნარების განვითარებაზე, რომ გენდერული ძალადობის მსხვერპლ ქალებს უზრუნველყონ იმ დაწესებულებების დაეხმარონ შესაძლებლობების გაფართოვებაში. შესაძლებლობების განვითარების უფრო მეტი სარგებელი ექნებათ კონფლიქტის მხარდაჭერა, რომლებიც ჩართული არიან გამო დაზარალებულ ადამიანებს, რომელთაც საქართველოში გენდერული ძალადობის აღნიშნეს, რომ მათთან უმუშევრობისა და თავიდან აცილების/პრევენციისა და არასტაბილური სამუშაოს ყველაზე მაღალი რეაგირების საქმეში. ეს მხარდაჭერა დონე შეიმჩნევა. ისინი ასევე დარწმუნებული შესაძლოა მოიცავდეს ტრენინგებს, იყვნენ, რომ ეს ეკონომიკური სტრესები არიან მომსახურების სფეროს გაზრდას, დარგობრივ მიზეზი გენდერული ძალადობისა. ქვეყნის სამინისტროებში ყურადღების აქცენტირებას შიგნით იძულებით გადაადგილებულ და და გენდერული ძალადობის ტენდენციების კონფლიქტის შედეგად დაზარალებულ პირთა ანალიტიკას/ თვალყურის მიდევნებას. დასაქმებაზე ფოკუსირებული პროგრამა უნდა მაგალითად, ადამიანით ვაჭრობის/ტრეფიკინგის მსხვერპლთა, დაზარალებულთა დაცვისა და მოიცავდეს და ყურადღებას ამახვილებდეს დახმარების სახელმწიფო ფონდი (ATIPFUND) გენდერული ძალადობის პოტენციურ რისკებსა და სოციალური მომსახურების სააგენტო და ეკონომიკის უსაფრთხოების ზრდისათვის ვალდებული არიან უზრუნველყონ ოჯახური აუცილებელ ინვესტიციებზე. ძალადობის მსხვერპლთათვის სოციალური უნდა მოხდეს დევნილთა დევნილთა მომსახურების გაწევა. თუმცა, ამავდროულად, ამ ორგანიზაციებს შეზღუდული შესაძლებლობები დროებითი განსახლების ცენტრებში გააჩნიათ ახალი მანდატების განხორციელების დევნილების საცხოვრებელი პირობებისა საკითხში. საერთაშორისო განვითარების და დევნილთა ხელშეწყობის ინვესტირება პარტნიორებმა უნდა ასევე მხარი დაუჭირონ გენდერული ძალადობის რისკების საქართველოს მთავრობას სხვა ქვეყნებსა და გათვალისწინებით. კვლევის შედეგები რეგიონში ძალადობასთან დაკავშირებული მიუთითებენ, რომ არსებობს პოტენციური კანონმდებლობისა და სტრატეგიის კავშირი რთულ საცხოვრებელ პირობებსა წარმატებული განხორციელების გაზიარების და გენდერული ძალადობის გავრცელების საქმეში, მათ შორის, უწყებათაშორისი ზრდას შორის, ეს ასევე ეხებათ დევნილებს, სამინისტროების კოორდინირებულ რომლებიც ხშირად ცხოვრობენ დევნილთა რეაგირებაში. პარტნიორებს აგრეთვე დროებით განსახლების გადატვირთულ შეუძლიათ ხელი შეუწყონ არასამთავრობო ცენტრებში. საერთაშორისო განვითარების ორგანიზაციების მხრიდან ზედამხედველობასა პარტნიორებს შეუძლიათ განიხილონ და მონიტორინგს, რომ საქართველოს გენდერული ძალადობის თავიდან აცილების მთავრობა ანგარიშვალდებული იყოს და რეაგირების სტრატეგიები ამ სოციალურად არსებული კანონმდებლობისა და პოლიტიკური დაუცველი/მოწყვლადი ჯგუფებისათვის საბინაო რეფორმების განხორციელების საქმეში. საკითხის შემუშავების საშუალებით. ძირითადი ყურადღება დაეთმოს ცნობიერების გაღრმავება და პროგრესის ფართომასშტაბიან პროექტებში გენდერული გათვალისწინება გენდერული ძალადობის ძალადობის საკითხებს, განსაკუთრებული შემცირების საკითხში. ამ გამოკითხისა და ყურადღება მიექცეს იმ პროექტებს, მომავალი გამოკითხვების შედეგად მიღებული რომლებიც ეხმარებიან სოციალურად ინფორმაციის გამოყენება შესაძლოა დაეხმაროს 25 გენდერული ძალადობის ტენდენციების შემდგომში მომსახურებისა და მხარდაჭერის შეფასებაში. ეს ინფორმაცია ასევე დაეხმარება ადაპტაციისთვის იქნება საჭირო. გარდა გენდერული ძალადობის არარეგისტრირებული ამისა, ხანგრძლივი კვლევების გამოყენება შემთხვევების უკეთესი გაგებისა და ამ დროთა განმავლობაში დაგვეხმარება იმავე მიმართულებით სიტუაციის გაუმჯობესებაში. საკითხებთან დაკავშირებით დავუკვირდეთ მომავალი კვლევა უფრო ღრმად გამოიკვლევს კონფლიქტისა და იძულებითი გადაადგილების გენდერული ძალადობის მსხვერპლთა გრძელვადიან შედეგებს. გენდერული პროფილებს/მოდელებს და საშუალებას ძალადობის შესახებ ცნობიერების ამაღლების მოგვცემს გავატაროთ მიზნობრივი სამუშაოები მექანიზმების ასეთი გაუმჯობესება საშუალებას პრევენციული ღონისძიებების, საინფორმაციო მისცემს საერთაშორისო ორგანიზაციებს კამპანიებისა და მომსახურების მიმართულებით. შექმნან შედარებითი ანალიზის კრიტერიუმები მომავალში მასში უნდა შევიდეს აქცენტი და მიზნობრივად მიმართონ საკუთარი ძალადობის გამოცდილებაზე მამაკაცთა ინვესტიციები. შორის და რისკის ფაქტორებზე, რომელიც 1 Introduction This report explores the links among conflict, development, and gender-based violence (GBV) in Georgia. Specifically, it looks at how three social groups—Internally Displaced Persons (IDPs), Administrative Line Persons (ALPs), and Nonconflict-affected Persons (NAPs)—are affected by the interaction among gender norms, conflict impacts, access to services, economic opportunities, and GBV. IDPs are people who were dis- placed by conflicts in the 1990s and 2008 in the Abkhazia and South Ossetia regions. ALPs are people who live along the former conflict zones. Both IDPs and ALPs are con- sidered as conflict-affected populations. NAPs are the rest of the Georgian population. In this research, GBV is measured through four forms of violence (FFV), which are con- trolling behavior/emotional abuse, sexual harassment, sexual violence, and physical violence. This study utilizes the definition of GBV pro- harm or suffering, threats of such acts, co- vided by the Inter-Agency Standing Commit- ercion and other deprivations of liberty, tee (IASC). This includes acts that inflict “phys- whether occurring in public or in private ical, mental or sexual harm or suffering, threats life. The term is also used by some actors of such acts, coercion and other deprivations of to describe some forms of sexual violence liberty, whether occurring in public or in private against males and/or targeted violence life” (see full definition below). Given structural against LGBTI populations, in these cases power differentials between women and men, when referencing violence related to gen- women are primarily at risk from GBV, though der-inequitable norms of masculinity and/ some forms of GBV can also be experienced by or norms of gender identity.” (IASC 2015) men (for example, sexual violence or targeting of gay or transsexual men because of their gender This report is the culmination of nearly two identity). The IASC definition of GBV reads: years of research, which is a mix of qualita- tive and quantitative methods. These include “An umbrella term for any harmful act that a stakeholder analysis, key informant interviews is perpetrated against a person’s will and (KIIs), nationwide focus group discussions (FGDs), that is based on socially ascribed (i.e. gen- and a nationally representative quantitative sur- der) differences between males and fe- vey. It includes inputs from international experts males. The term ‘gender-based violence’ and all levels of Georgian society. Box 1 explains is primarily used to underscore the fact the rationale behind this research. that structural, gender-based power dif- ferentials between males and females The survey research has several key findings. around the world place females at risk for Conflict-affected people, although not reporting multiple forms of violence. As agreed in more GBV compared to the general population, the Declaration on the Elimination of Vio- can be more vulnerable to GBV. This is mainly be- lence against Women (1993), this includes cause conflict and displacement cause economic acts that inflict physical, mental or sexual hardships. These economic stressors, like unem- 26 27 Box1: Study Rationale This research is part of the World Bank’s Global Platform on Gender-based Violence funded by the State and Peacebuilding Fund (SPF). The platform’s objectives are to provide services to GBV survivors, contribute to prevention, raise public awareness, and build capacity of member countries through South-South knowledge sharing. Pilot projects were approved for several countries in different re- gions of the world. Each of these countries—including Nepal, Sri Lanka, Papua New Guinea, and the Democratic Republic of the Congo—face challenges related to conflict and GBV. Georgia, as the Europe and Central Asia representative in this global platform, faces challenges related to conflict and GBV. Given conflicts in Abkhazia and South Ossetia regions in the 1990s and 2008, Georgia has many conflict-affected people. This includes between 190,000 and 275,000 IDPs,a who have been displaced by conflict and make up almost 6 percent of the population, among the highest relative proportions in the world. Also, people living near former conflict zones in Abkhazia and South Ossetia regions continue to experience periodic insecurity and impacts on their living conditions and livelihoods. There is an increasing awareness of the effects of GBV within Georgia, and the Government of Georgia has undertaken noteworthy efforts to address domestic violence. This includes strengthen- ing of legal protections, elaboration of National Action Plans, provision of services, and training of law enforcement personnel (Government of Georgia 2014c). These ongoing efforts are described in more detail in the section on national GBV legislation. The Georgian research is unique within the SPF in that it focuses on capacity building and deepening knowledge. Because of this focus, the study covers an expansive set of themes. This includes gender norms, conflict, welfare, services, economic opportunities, and both male and female perspectives from conflict-affected and nonconflict-affected communities—all areas the research team identified as relevant for understanding the myriad of factors contributing to response and prevention of GBV in Georgia. There are many modern studies on GBV and many more on conflict issues, but the amount of research that examines the link between the two is not nearly as com- mon (Spinelli 2014). Adding additional themes, as in this research, limits the existing literature base even more. This report fills some of these research gaps. The research findings will also be used, in partnership with other stakeholders, to build capacity within the Georgian government to prevent and respond to GBV. a The national census estimates the IDP population at 190,000 (GEOSTAT, 2017), while the Internal Displacement Monitor- ing Centre (IDMC) estimates that there are 206,000 IDPs in Georgia (IDMC, 2016). As of January 2017, 274,013 IDPs have registered with the Ministry of Internally Displaced Persons from the Occupied Territories, Accommodation, and Refugees (MRA, 2017a). This registry number is higher because it includes the children of people with IDP status. ployment and shame from not being able to pro- efficacy and there is great scope for aligning ex- vide for a family, are believed by respondents to isting services with awareness of services. increase GBV. Overall, GBV levels in Georgia are on par with global and regional averages (WHO The next section of the report presents the 2013), but evidence suggests that survey respon- research methodology. Section 3 provides con- dents may have underreported their experienc- text to the study by briefly looking at the history es with GBV. This may have been especially true of conflict and displacement and the makeup of in communities that believe the least in gender the main groups in the survey sample. Section 4 equality and the most in maintaining traditional examines the country’s prevailing gender norms gender roles, such as rural areas. Young people and reported levels of the FFV. After analyzing tend to support gender equality the most, which these aspects, Section 5 looks at the links those bodes well for the future. Another positive sign aspects have with Georgia’s history of conflict for the future is that Georgia has a relatively ro- and current economic conditions. Section 6 re- bust legal and institutional framework in place to views the country’s awareness of, and confidence protect survivors of GBV.3 Because of problems in, GBV service delivery, or lack thereof. Section in implementation or little awareness of these 7 provides concluding remarks and recommen- services, people largely lack confidence in their dations. 3 According to IASC (2015), “The terms ‘victim’ and ‘survivor’ can be used interchangeably. ‘Victim’ is a term often used in the legal and medical sectors, while the term ‘survivor’ is generally preferred in the psychological and social support sectors because it implies resilien- cy.” Both terms are used freely throughout the report. 2 Methodology The methodology is a combination of qualitative and quantitative research meth- ods. The research implementation lasted 22 months from the first stages of design in November 2015 to the delivery of the final report in September 2017. It followed five phases, as outlined in Figure 1. The official name of the survey is ‘Population’s Life Experiences in Georgia Survey, 2016’. A more detailed methodology is provided in Annex 1. Figure 1: Implementation Phases Phase I Phase II Phase III Phase IV Phase V Research Design Initial FGDs Survey Implementation Second round Data Analysis and Stakeholder and KIIs of FGDs and and report Analysis KIIs writing NOVEMBER 2015 SEPTEMBER 2015 The conceptual framework and main research groups—NAPs, IDPs, and ALPs—and the men and questions are presented in Figure 2. The re- women within each group. The study also asked search looked at links between GBV and con- these questions of members of the Lesbian, Gay, flict, economic opportunities, and services. It Bisexual, and Transgender (LGBT) community also sought to understand gender norms as they but only through FGDs as LGBT people were not might influence experiences and opportunities included in the survey. This was done to add an in each of these areas. Research questions are extra layer of privacy for LGBT people. asked of each of the study’s three main social 28 29 Figure 2: Conceptual Framework for Population’s Life Experiences in Georgia Survey, 2016 Conflict ƒƒ Are conflict-affeced people more at risk from GBV? ƒƒ Can displacement induced stress increase risks? Violence Service Livelihoods & economic opportunities ƒƒ What are the service needs of ƒƒ How have conflict-affected people’s livelihoods conflict-affected people? been impacted by the conflict? ƒƒ What are the service needs of the ƒƒ Does household economics impact the general population? experience of violence? ƒƒ What is the level of knowledge and ƒƒ What economic options and coping mechanisms access to services? are available to survivors? 2.1 OVERVIEW OF RESEARCH the information and data from the other com- COMPONENTS ponents, consisting of: a stakeholder analysis, a quantitative survey, Focus Group Discussions, This study applies a triangulated research ap- and Key Informant Interviews. Table 1 shows the proach. Results from each of the research com- study’s basic research components. ponents were contrasted and validated against Table 1: Basic Research Components Data Collection Stakeholder Analysis Quantitative Survey Focus Group Discussions Key Informant Interviews Method 1. Male NAPs 2. Female NAPs 1. Government officials 1. Male NAPs 3. Male IDPs 2. Academics 1. Institutions 2. Female NAPs 4. Female IDPs 3. Civil society organization engaged in GBV 3. Male IDPs 5. Male ALPs members Target Group 2. GBV-related policies 4. Female IDPs 6. Female ALPs 4. International 3. Background note 5. Male ALPs 7. Young women organization officials on LGBT people 6. Female ALPs 8. Elderly women 5. Thematic and regional 9. LGBT people specialists 10. GBV service practitioners Sample Size — 3,014 interviews 35 FGDs (280 participants) 10 in-depth interviews Random stratified Purposive sampling and Sampling Method Desktop review Purposive sampling sampling snowball sampling Source: World Bank team. Note: For the FGDs, LGBT people and GBV service practitioners are represented in italics to signify that these groups were not included in the quantitative survey. 30 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES 2.1.1 Quantitative Survey Sampling Design and Table 2: Number of Survey Interviews and Size of Questionnaire Sample (in Brackets) The survey targets men and women over 18   Women Men Total years old in all three target groups—NAPs, NAP 706 (700) 700 (700) 1,406 (1,400) IDPs, and ALPs. Given the specificity of the three target groups, the research team followed three IDP 506 (500) 502 (500) 1,008 (1,000) different methods to build samples for these ALP 301 (300) 299 (300) 600 (600) groups: Total 1,513 (1,500) 1,501 (1,500) 3,014 (3,000) (1) For NAPs, Georgia’s National Statistics Of- Source: World Bank team. fice (GEOSTAT) provided a list of census districts with populations broken down by Certain subpopulations were analyzed with- age, gender, region, and settlement type. in the survey sample—disaggregated by age, ALP communities that bordered conflict education, welfare level, housing condition, zones and individuals who reported them- urban-rural location, conflict experience, and selves as IDPs were excluded from this sam- more. Often, these subpopulations did not have ple. NAPs live in all areas of Georgia. enough survey responses to achieve these con- fidence levels. In these cases, the study explicitly (2) For IDPs, the study used the MRA’s official states that the results are not representative. registry of all IDPs in Georgia. This infor- mation was kept in complete confidentiality 2.1.2 FGD Sampling Design and only accessed by the local research team to build a sample and identify respondents. The FGD target groups included NAPs, IDPs, IDPs live in all areas of Georgia. ALPs, LGBT people, and GBV service providers like shelter workers. Most focus groups includ- (3) For ALPs, the research team used a govern- ed participants between the ages of 25 and 60, ment list of the communities where barbed but a select number of focus groups included in- wire fences were erected after the 2008 dividuals between the ages of 18 and 25 to gain conflict. The list includes 88 settlements, 86 a better insight into the perspectives of the youth. of which were verified by GEOSTAT. We built Of the conflict-affected FGDs, only IDPs and ALPs the sample and selected respondents from with direct experience of conflict in the 1990s or these 86 settlements. All settlements border 2008 participated. LGBT participants were re- Abkhazia and South Ossetia. cruited using LGBT-focused nongovernmental organizations. This was done to ensure the safe- The total survey sample size was 3,000 peo- ty and comfort of participants. Direct quotations ple, which ensures a 2–3 percent margin of from FGD were selected and articulate well-for- error, at a 95 percent confidence level for the mulated expressions of views that were shared total sample. In total, 3,014 interviews were suc- within the focus group, and where noted, across cessfully carried out. The number of interviews is many of the focus groups. shown in Table 2. For data weighting purposes, different poststratification criteria were used for FGDs were performed over three rounds. The each target group, and for each division, popu- first round of FGDs informed the survey ques- lation information delivered by GEOSTAT was tionnaire. The second round deepened the un- used.4 derstanding of survey results. The third round included new demographics that the research team felt were underrepresented in the survey and the previous two rounds of FGDs. These de- 4 For a detailed discussion of the weighting strategy, see Annex 1. 31 Box 2: Violence Indicators by Category of Violence Used in Survey of Population’s Life Experiences in Georgia Survey, 2016. Controlling behavior/emotional abuse is threats or acts committed by a person to exert or maintain control over another person. These are psychological in nature and refer to the infliction of mental or emotional pain or injury such as threats or humiliation (adapted from IASC 2015). The questions included in this survey are: 1. *Insulted you or made you feel bad about yourself? 2. Belittled or humiliated you in front of other people? 3. Threatened to hurt you or someone you care about? 4. *Harassed you by repeatedly sending you messages or calling you? Sexual harassment includes unwelcome sexual advances and other verbal or physical conduct of a sexual nature (IASC 2015, from US Department of State, n.d.). The questions included in this survey are: 1. Touched, hugged, or kissed you against your will? 2. Stared or leered at you inappropriately so that it made you feel intimidated? 3. Made sexually suggestive comments or jokes that made you feel offended? 4. Sent or showed you sexually suggestive photos or pictures, e-mails, or texts? 5. Exposed themselves to you indecently? Sexual violence refers to any sexual act committed against the person’s will (IASC 2015, from WHO 2002). Specific forms of sexual violence such as forced pregnancy, sexual slavery, sexual trafficking, or sexual exploitation are not part of this survey. The questions included in this survey are: 1. Physically forced you to have sexual intercourse when you did not want to? 2. Had sexual intercourse with you when you were unable to refuse (for example, too drunk)? 3. Had sexual intercourse with you against your will, because you were afraid of what they might do? 4. Forced you to do something sexual that you found degrading or humiliating? Physical violence refers to a physical assault that is not sexual in nature. This includes hitting, slapping, choking, cutting, shoving, burning, and using weapons against a person (IASC 2015, from GBVIMS 2010.) The questions included in this survey are: 1. Slapped you or thrown something at you that could hurt you? 2. Dragged, pushed, or shoved you? 3. Kicked you or hit you with their fist or with something else that could hurt you? 4. Choked or burnt you on purpose? 5. Threatened to use or actually used a gun, knife, or other weapon against you? Source: World Bank team. Questions are derived from a range of questionnaires indicated in footnote 4. Note: Indicators with * in the controlling behavior/mental abuse section were not included in the violence index that served for the main analysis and regressions in this report. This was because one of the indicators (Insulted you or made you feel bad about yourself?) had results that were atypical, that is, markedly higher than the other indicators and that skewed the reported violence upwards. Conversely, the indicator ‘Harass you by repeatedly sending you messages or calling you?’ was also excluded because it had results that were markedly lower than the other indicators and that consequently added little value to the measure. 32 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES mographics included youth, LGBT people, and questionnaire was modelled after questionnaires GBV service providers. that have been previously used and tested.5 2.1.3 Forms of Violence Included in the Survey 2.1.4 Structure of the Survey and of the Violence Module The research analyzed four forms of violence, henceforth referred to as FFV. These include Given the focus of this survey on exploring controlling behavior/emotional abuse, sexual the links between GBV, conflict, economic op- harassment, sexual violence, and physical vio- portunities, and services, the questionnaire lence. For the analysis of the population group’s modules collected data in these respective general experience of violence, this report does areas, in addition to pertinent background not distinguish between partner or non-partner information. The modules and themes are de- violence, though relevant information pertaining tailed in Table 3. to either partner/non-partner violence is point- ed out where this is known and relevant. FFV is 2.2 ETHICAL CONSIDERATIONS AND typically interpersonal violence or violence per- ETHICS REVIEW petrated by one person against another, but it can include more than one perpetrator. Each of Answering questions on GBV and conflict ex- the four forms of violence, as used in this survey, periences can be traumatic. Sharing such ex- are defined in turn, and the specific indicators or periences with strangers can be difficult and po- questions that were used to measure each form tentially cause problems with partners who may of violence are included in Box 2. The survey prefer that information stays private. As such, this Table 3: Survey Modules and Themes Module Themes Covered by the Module 1. Conflict status IDP affected by conflict in 1990s or 2008; ALP affected by conflict in 1990s or 2008; NAP not affected by conflict 2. Basic household data Household members and relationship between them; educational levels; marital status; ethnicity; disability status 3. Welfare Dwelling; land; livestock; durable goods; income; social assistance 4. Employment, skills, and Employment status, sector and characteristics (before and after conflict); unemployment; financial livelihoods autonomy; jobs training 5. Gender norms and attitudes Gender norms; norms and violence; violence and the law 6. Violence Controlling behavior /emotional abuse; sexual harassment; sexual violence; physical violence; most serious partner incident; most serious non-partner incident; health consequences; reporting of violence 7. Conflict and displacement Pre- and post-conflict situation; perception of GBV and conflict; access to services 8. Services Awareness of GBV campaigns; access to services; female police officers Source: World Bank team. 5 The following tested survey questionnaires were used to inform the eight modules of the survey: the multi-country study on wom- en’s health and domestic violence against women of the World Health Organization (WHO) (http://www.who.int/reproductivehealth/ publications/violence/24159358X/en/), EU-wide Violence Against Women Survey of the European Union Agency for Fundamental Rights (FRA) (http://fra.europa.eu/en/publication/2014/violence-against-women-eu-wide-survey-main-results-report), the International Men and Gender Equality Survey (IMAGES) (https://www.icrw.org/publications/international-men-and-gender-equality-survey-images/), Demo- graphic and Health Survey (DHS) Domestic Violence Module of the United States Agency for International Development (USAID) (https:// dhsprogram.com/pubs/pdf/DHSQMP/domestic_violence_module.pdf.pdf), the World Bank’s Socio-economic Impact Assessment of the Presence of Syrians under Temporary Protection on Turkish Hosting Communities, The European Union (EU) and Handicap International’s Livelihoods Assessment questionnaire, the Rural Health Research Center GBV Toolkit, the London School of Hygiene and Tropical Medi- cine questionnaire on Women’s Health, the Kinyanda War Trauma module, GEOSTAT’s Integrated Household Survey (IHS) for 2015, and George Washington University’s Database of Questionnaires on Violence Against Women and Girls in Humanitarian Settings. 33 research followed WHO Guidelines (2005) for and Pham 2013) rather than assessing multiple researching violence against women to ensure forms of violence like this study. However, some the safety and well-being of respondents and re- studies also look at physical violence outside of searchers and undertook several activities to sen- the home (Falb et al. 2013b; Hossain et al. 2014). sitize the data collection mechanisms. The meth- This study represents a contribution to the ex- odology included a gender sensitivity training for isting literature by drawing on nationally repre- all fieldwork team members to ensure the safety sentative samples, with both male and female and well-being of participants and research- respondents, that report conflict and GBV expe- ers. Because of the sensitivity of GBV questions, riences. only women or only men were surveyed in each sample site. The methodology passed an inde- This survey draws from other violence-related pendent ethics review by an Institutional Review survey questionnaires but may not be directly Board (IRB) managed by the Health Research comparable. It is similar because it asks about Union in Georgia. respondents’ experiences with specific types of behaviors drawing upon tested questionnaires. 2.3 UNIQUENESS AND LIMITATIONS It differs in other important respects. Given the OF METHODOLOGY broader scope of this survey and out of con- cern for respondents’ time, the questions about This study’s methodology was designed to gender-based violence had to be limited (see complement existing GBV research gaps. The Table 3). The survey was thus not able to collect survey assesses respondents’ experiences with as much information on each violent incident as different forms of violence, including controlling standard surveys like Demographic and Health behavior, sexual harassment, sexual violence, Survey (DHS) or the WHO Women’s health and and physical violence. Previous studies have rare- domestic violence against women survey. Per- ly included analyses of harassment as part of vio- petrator information was not systematically col- lence assessments,6 but harassment can be wide- lected for each type of violence as it would have spread and meaningful to measure. Because this made the violence module too long, although type of violence has been included, and because the distinction between partner and non-partner this survey seeks to add to, rather than replicate violence was included for the single most serious other violence surveys,7 it is important to note incident (along with location and impact of this that these measures are not directly compara- incident). Unfortunately, response rates to this ble to studies, such as the DHS, which have been question are not representative and did not allow done in other countries. The survey also focuses for robust analysis of this aspect. Finally, the cur- on economic outcomes and experiences from rent survey asked these questions of all respon- Georgian conflicts. This approach adds to a lim- dents—men and women—and did not limit the vi- ited literature base that examines how conflict olence modules to ever-partnered women or to correlates to different forms of violence. In cur- intimate partner violence—a common approach rent literature, many studies ask survey questions in some violence-focused surveys. Therefore, about conflict and violence experiences but most these results may not be directly comparable to draw on small samples with idiosyncratic popula- other nationally representative surveys. tions such as refugees (Falb et al. 2013a, 2013b; Gupta et al. 2012; Saile et al. 2013). Moreover, The study also carries with it a few limitations. previous studies often focus only on women and First, the sensitive nature of some questions emphasize intimate partner violence (Falb et al. could lead to underreporting of certain experi- 2013a; Gupta et al. 2012; Saile et al. 2013; Vinck ences to researchers. Certain gender-inequitable 6 This is generally true, though some notable studies have included harassment and controlling behavior in their measures of GBV. See FRA (2014). 7 The research team understands that a multi-country survey on Violence against Women that will collect internationally comparable data is being planned by the Organization for Security and Co-operation in Europe (OSCE) and that UN Women is collecting national data on GBV in Georgia. 34 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES attitudes may be underreported because they historical rates. The survey simply ask about for- are socially undesirable. Certain groups, such as mer levels of violence and well-being, but this is men or those living in very rural areas, may have still a measure of the respondent’s current think- been more likely than other groups to underre- ing, which includes memory biases. The reported port experiences like sexual violence because incident refers to ‘ever experienced’ and as such of the highly stigmatizing nature of these expe- it does not allow to establish trends over time. In riences. Similarly, women currently in a relation- the future, a longitudinal study following subjects ship may prefer not to disclose experiences that over time could track how changes in econom- have happened in the context of that relationship ic status, exposure to services, and awareness of to researchers, despite the confidential nature of GBV campaigns are associated with changes in those exchanges. Second, this research presents reported rates of violence. a snapshot in time and does not allow the re- searchers to compare current rates of violence to 3 Georgian Context and Respondents’ Profiles This section provides contextual information for the report’s subsequent analyses. To understand discussions on conflict and GBV in Georgia, one must understand Geor- gia’s conflict history, its economic growth since then, and the sociodemographic char- acteristics of each sample group. This includes descriptive indicators for the three conflict-defined populations—NAPs, IDPs, and ALPs—and the national male and female populations. 3.1 HISTORY OF CONFLICT AND 65 years (MRA 2017b). The national census esti- DISPLACEMENT mates the IDP population at 190,000 (GEOSTAT 2017). The vast majority of IDPs in Georgia are Georgia suffered through two main waves ethnic Georgians, as illustrated in the research of conflicts that led to large numbers of in- sample in Table 4. ternally displaced populations. The first was the South Ossetian and Abkhazian conflict in Most IDPs live in Tbilisi and areas adjacent to 1991–1993. The second was a relatively shorter the conflict zones. Seventy-five percent of IDPs armed conflict with Russia in South Ossetia in live in cities compared to 57 percent of the gen- 2008, with spillover effects in Abkhazia. These eral population. Many IDPs from Abkhazia settled conflicts led to the permanent displacement of in Imereti and Samegrelo regions, while many between 190,000 and 275,000 people. Most IDPs from South Ossetia settled in the Shida Kartli IDPs were displaced in the 1990s, while a lesser region. According to the MRA (2011), 44 percent number remain displaced since the 2008 con- of IDPs live in Tbilisi, and another 26 percent live flict. The 1990s conflict led to about 300,000 dis- in the Samegrelo-Zemo Svaneti region (Figure 3). placed people, of which about 220,000 remain Many IDPs from the earlier conflict were placed displaced today. The 2008 conflict led to another in collective centers, such as state-owned hotels, 138,000 displaced, of which 30,000 remain dis- and unused public buildings, where close to 40 placed (UNHCR 2009). The IDMC estimates that, percent remain to this date.8 overall, 206,000 Georgians remain displaced The Government of Georgia has taken de- (IDMC 2016). As of January 30, 2017 there were monstrable steps toward improving the socio- 274,013 IDPs in the MRA registry (MRA 2017a). economic conditions of IDPs. Newly registered This higher number is because it includes chil- IDPs are entitled to a one-off cash assistance for dren with at least one IDP parent who are also newly displaced persons, followed by a month- entitled to the status. Also, according to the MRA, ly cash allowance. A one-off rental assistance is 54 percent of IDPs are women, 27 percent of also available for the most vulnerable IDP house- IDPs are under 18 years, and 13 percent are over 8 According to UNHCR (2015), 16.7 percent of IDPs live in rehabilitated collective centers and 21.5 percent remain in non-rehabilitated collective centers. 35 36 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Figure 3: Distribution of IDP Population, 2016 Source: World Bank, utilizing 2016 MRA data on regional distribution of IDPs. Based on a concept by IDMC http://www.inter- nal-displacement.org/europe-the-caucasus-and-central-asia/georgia/2011/internal-displacement-in-georgia-2011. holds. IDPs may also apply for Targeted Social percent. Moreover, GEOSTAT (2017) reports a Assistance (TSA) program, though this requires 15 percentage point reduction in the number them to give up their IDP status as these support of Georgians living in absolute poverty, from 36 systems cannot be combined (World Bank 2016). percent in 2010 to 21 percent in 2016. 3.2 GEORGIAN ECONOMY 3.3 SAMPLE GROUP DETAILS Georgia’s economy has grown and poverty The Republic of Georgia is a small country in rates have declined since 2010. In 2008, Geor- the South Caucasus of about 3.7 million peo- gia experienced economic declines associat- ple. Of these, 57 percent live in urban areas and ed with the global economic downturn and the 43 percent live in rural areas (GEOSTAT 2017).10 South Ossetia conflict. But since 2010, when the It has been estimated that up to 80 ethnic groups impacts of those events began to subside, Geor- live in Georgia (Cornell 2001). The survey sample gia has steadily recovered. Figure 4 shows that indicates that a clear majority are ethnic Geor- annual per capita gross domestic product (GDP) gians (Table 4). There also exists a significant has increased by more than US$1,0009 over the number of Armenians and Azeri, with the two past six years. Figure 5 shows that the popula- representing 10 percent of the Georgian popu- tion living in the bottom 60 percent of median lation. consumption line has decreased by about 2.5 9 All dollar amounts are in U.S. dollars. Georgian currency amounts are in Georgian lari (GEL). 10 The sample for this research survey found 55 urban population and 45 percent rural. This is comparable to GEOSTAT 2017 census data when considering the margin of error. 37 Figure 4: GDP Per Capita (at current prices), US$ Figure 5: Share of Population Under 60 percent of the Median Consumption (relative poverty) (%) $4,000 24% $3,500 22% $3,000 20% $2,500 $2,000 18% 2010 2011 2012 2013 2014 2015 2016* 2010 2011 2012 2013 2014 2015 2016 Source: GEOSTAT 2017. Source: GEOSTAT 2017. Note:* indicates data for 2016 are preliminary. Adjusted data will be published by GEOSTAT in November 2017. Table 4: Ethnic Diversity in Georgia Among the Survey’s Sample Groups   NAPs (%) IDPs (%) ALPs (%) Females (%) Males (%) Total Population (%) Georgian 87.8 98.8 94.8 90.6 85.9 88.4 Armenian 6.4 0.1 0.2 4.7 7.5 6.0 Azeri 4.5 0.8 4.7 3.1 5.4 4.2 Ossetian 0.1 0.1 0.3 0.3 0.2 0.2 Russian 0.6 — — 0.8 0.2 0.5 Yazidi/Kurd 0.2 — — 0.1 0.3 0.2 Greek 0.1 — — 0.1 0.1 0.1 Other 0.4 0.2 — 0.3 0.4 0.4 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: N = 3,014. NAPs = Non conflict-Affected Persons, IDP= Internally Displaced Persons, and ALP=Administrative Line Persons The sample group has distinct socioeconom- when examining the data, one should consider ic traits, which must be considered when an- that, for example, ALPs not only live near conflict alyzing the data. Each group—whether NAPs, zones, but they also tend to be poor and rural. IDPs, or ALPs—is not characterized simply by their IDPs, by contrast, have had to leave their areas level of conflict exposure. For example, IDPs are of origin and often live in Tbilisi and other cit- three-quarters urban, while ALPs are completely ies. These are factors that could influence their rural. ALPs are also the poorest. They have the gender views, vulnerability to GBV, or access to least adequate housing and overpopulate the services as much, or more, than their experience lowest consumption categories. IDPs have the of conflict. Table 5 uses survey data to summa- most vocational and university graduates. So, rize each of the survey’s representative popula- Age 18-24 18-24 18-24 12% 13% 11% 38 65+ IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES GENDER BASED VIOLENCE 65+ AND SERVICES 65+ 18% 16% 25-34 25-34 25-34 22% 16% 55-64 19% 55-64 20% 16% 16% 55-64 17% Table 5: Socioeconomic Characteristics of the Sample Groups 35-44 35-44 35-44 16% 45-54 45-54 18% 17% 17% 18% 45-54 18% Non conflict-Affected Person Internally Displaced Person Administrative Line Person Education ƒƒ Any non-IDP or non-ALP Primary ƒƒ A person forced to flee his or her home Primary ƒƒ Person residing in areas Master bordering Primary 4% Masterof the 1% ƒƒ 93.8 percent national population Master within but who remains 1% his or her current or former Bachelor 1% zones conflict 16% 17% 5% ƒƒ From all regions of Georgia country’s borders ƒƒ 1.1 percent of national population Bachelor ƒƒ 5.1 percent of national population ƒƒ Defined as areas where barbed wire Vocational Bachelor 11% 14% from Abkhazia and South ƒƒ Displaced fences were installed in 2008 by the 20% Ossetia during conflicts in the 1990s government along former conflict zones Vocational Secondary and 2008; now living in all Vocational parts of Secondary Secondary 25% 47% 46% 70% Georgia 22% Rural-urban Rural 25% Rural Rural 100% 45% Urban 55% Urban 75% Welfare Top 60 Bottom 40 16% 26% Bottom 40 Top 60 40% 60% Top 60 74% Bottom 40 84% Housing Marginal Marginal 30% Marginal 31% 31% Adequate 33% Adequate Adequate 58% 62% Inadequate Inadequate 7% 12% Inadequate 36% Source: Population’s Life Experiences in Georgia Survey, 2016. Note: Chart data weighted by sampling weights. 39 tions regarding location (rural/urban), welfare tom 40 percent.14 and housing. More detailed characteristics can be found in Annex 411). Welfare and housing There are three explanations for the higher are composite indices based on several factors. welfare rates for IDPs. First, the IHS and the Welfare12 is a measure of per capita consump- survey sample define IDPs differently. The IHS tion based on observable characteristics of the samples households using national census data household. Housing conditions13 are based on and asks respondents if they are IDPs; if they re- whether the household is overcrowded, has its ply ‘yes’, they are included in the IDP sample. By own heating system, and has access to sewage, contrast, this survey samples IDPs using the offi- which indicates access to basic services. cial IDP registry of the government, which is con- fidential and includes all people who have ever According to consumption-based measure- registered as an IDP. As a result, in the IHS, wealth- ment of welfare conducted as part of data ier families who no longer self-identify as IDPs or analysis, IDPs are better-off than the national do not receive social assistance for being an IDP, population. In the charts of Table 5, it should be often do not report themselves as displaced. In noted that IDPs have fewer people living in the this survey sample, these better-off families are bottom 40 percent of per capita consumption included. Second, consumption metrics like the than any other population group. They also have one we used in this survey, carry with it a natu- the most university graduates. While it could be ral urban upward bias. Even after controlling for expected that IDPs, largely urban residents, are other sociodemographic characteristics, living better-off than ALPs who are completely rural, it in urban poverty leads to higher imputed con- is unexpected that they would be better-off than sumption. Whether it is because of higher food the national sample. To have a better sense of prices or costlier utilities, living in cities requires this result, IDPs were compared to the national higher consumption than living in rural areas. sample, limiting the analysis to those respon- And, from the charts of Table 5, we see that IDPs dents residing in urban areas: in this case, while have the highest urban residency rate of all sam- the national sample has the highest share of uni- ple groups, by far. Also, previous studies have versity graduates (38 percent for NAP versus 36 found that while the share of income from hired percent for IDPs), IDPs still have fewer people in employment is similar for both IDPs and non-IDPs the bottom 40 percent of per capita consump- sources of income indicate a higher dependence tion (17 percent for IDPs versus 29 percent for of IDPs on social transfers such as pensions, NAP). According to the survey, only 26 percent scholarships, and social assistance, and on re- of IDPs are in the country’s bottom 40 percent mittances (World Bank 2016). Therefore, to sum of consumption. This differs from previous mea- up, the survey finds that IDPs are wealthier than sures on the topic, most notably GEOSTAT’s 2017 previous studies because they are largely urban, IHS, which shows 42 percent of IDPs in the bot- social transfers make up a more significant share 11 Specifically, Annex 4 includes information on the sample’s rural-urban divide, age demographics, education levels, per capita welfare levels, and housing conditions by population group and comparing men and women respondents. 12 Welfare at the household level was imputed using observable characteristics of the household. The decision to impute the welfare was based on time considerations at the moment of data collection, as consumption modules (the variable used for measuring welfare in Georgia) require extensive amount of time. A combination of machine learning techniques and expert opinion were applied to generate a model to predict consumption per capita using the national household survey, focusing on the geographic areas of interest and using out-of-sample prediction through cross-validation. Once the model was estimated, it was used to predict consumption per capita at the household level in the project’s survey data. Multiple imputation were used to track the uncertainty of the imputations throughout the process. 13 A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwell- ing has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indicator their housing was considered adequate, marginal, and inadequate (in descending order). 14 Other surveys looking at incomes, rather than consumption, report starker differences between IDP and non-IDPs. For example, the Intentions survey on durable solutions: voices of internally displaced persons in Georgia (UNHCR 2015) indicate that 4-5 members of IDP households have an average monthly income of 453 GEL, whereas an average income of non-IDP population amounts to 887 GEL. 40 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES of their incomes, and because the study includes al sample (58 percent to 61 percent), own their wealthier IDPs in the survey sample. home less often (66 percent of the time com- pared to 90 percent of the time), and are more This is not to say that IDPs are better-off by reliant on social assistance (45 percent of IDP all measures. To test this the analysis included families who applied received social assistance a multidimensional poverty index (MPI). This in- compared to 33 percent of the national sample). dex measures poverty across four non-consump- These metrics also do not account for the psy- tion-based dimensions, including demographics, chological impacts of conflict or the difficulties education levels, employment status, and access inherent to displacement. IDPs described some to infrastructure.15 According to this measure, the of those impacts in the following terms: national sample is substantially better-off than both IDPs and ALPs.16 It also shows that IDPs in “Ten of my family members, including my- private dwellings are much better-off in per cap- self, reside under the same roof; yet we are ita consumption than IDPs in collective centers. only equipped with 2 bedrooms and a sin- This seems to confirm that the sampling strate- gle living room. My brother, his wife and gy of using the IDP database over census data child, my mother, father, and our grand- includes a greater number of affluent IDPs who parents, all inhabit the same cottage.” (IDP do not live in collective centers and probably do woman, Gori) not self-identify as IDPs in the IHS. The analysis also demonstrates that 1990s IDPs, who have had “It is not surprising that our population is over 20 years to adjust to their displacement, are on nerves; most of them died because of somehow better-off than 2008 IDPs. In particular, nerves and because no one lacked any- 22 percent of IDPs displaced in the 1990s are thing there; and here they can only dream in the bottom 40 percent of the per capita con- about everything…” (IDP man, Tbilisi) sumption against 47 percent of IDPs displaced in 2008. On the other hand, the share of respon- ALPs, who are completely rural and live in ar- dents who may be considered ‘multidimension- eas bordering former conflict zones, are the ally poor’ according to the MPI is quite similar for poorest sample group according to the con- the two groups (23 percent for IDPs displaced in sumption-based welfare measure used in this the 1990s and 21 percent for those displaced in survey. Nearly 84 percent of ALPs live in the bot- 2008). tom 40 percent of per capita consumption. They are the most likely to live in inadequate housing IDPs are also worse off than the national sam- and have the lowest higher education gradua- ple in other non-consumption-based mea- tion rates. A higher percentage of ALPs (14 per- sures. They live in the smallest dwellings (59 m2 cent) receive government assistance than IDPs compared to 104 m2 for the national sample), (10 percent) and NAPs (5 percent). Access to ba- have less adequate housing than the nation- sic services has generally improved for all sam- ple groups, except ALPs. Moreover, they are the 15 The four sets of deprivation indicators used to create the MPI are defined as follows: (a) Demographic: the respondent is deprived if the ratio between dependent members (younger than 15 or older than 64)/working age adult (15–64 years) in the household is higher than 1; (b) Education: (i) the respondent is deprived if at least one adult member of the household has not completed a secondary lower education degree; (ii) the respondent is deprived if no adult member of the household has completed a college degree; (c) Labor: the re- spondent is deprived if s/he (and her/his partner) is unemployed; and (d) Services and Infrastructures: (i) Access to water - the respondent is deprived if the respondent’s dwelling has no access to water—either central or individual; (ii) Access to gas - the respondent is deprived if the respondent’s dwelling has no access to gas—either central system or liquid; (iii) Access to heating - the respondent is deprived if the respondent’s dwelling has no access to heating system; (iv) Access to bathroom - the respondent is deprived if the respondent’s dwelling has no access to bathroom either individual or shared. Each dimension is equally weighted and within each dimension, each indicator is also equally weighted. A respondent is defined as ‘multidimensionally poor’ when s/he is deprived on 33 percent or more of the weighted indicators in the index. A respondent is defined as ‘multidimensionally vulnerable’ when s/he is deprived of 20 percent to 33 percent of the weighted indicators in the index. 16 Only 18 percent of the national sample is ‘multidimensionally poor’ compared to 23 percent among the IDPs and 31 percent among the ALPs. 41 poorest sample group according to the MPI.17 river where children swim and people fish. As one ALP woman in Gori described, ALPs explained that they experience the re- alities of the conflict differently than other “They cross the line and take people, cat- sample groups. Unemployment is a pervasive tle…We have an orchard near the border, theme of all the ALP FGDs, with participants em- Russians stand and stare at us, we are al- phasizing the problems that prevent them from ready used to it but it’s still scary. They can farming as they did before the conflict. In FGDs, cross the line and…” a preponderance of men and women com- plained of having lost land, pastures, access to Men and women respondents in the survey irrigation canals, and deteriorated infrastructure. are comparable with regard to age, educa- Male ALPs living near Gori reported people los- tion, location, welfare, and housing, but this ing garden plots each year as the administrative masks some differences between the social line moves. They say they can no longer visit the groups studied in this report. For example, few- cemetery where relatives are buried on Easter as er ALP women had attained higher education (18 they once had. People can be fined, seized, and percent) than the national or IDP population (35 detained for three days if they wander across the percent and 45 percent respectively). Fewer IDP administration line. Women in Gori and Tsilkani women were married (52 percent) than IDP men complained about the daily psychological pres- (61 percent) or NAP or ALP women (63 percent sure from helicopters flying overhead and seeing and 65 percent, respectively). Other inequalities armed Russian troops on the “border.” Since the pertain to gender roles, access to resources, vul- conflict, ALPs say municipal services have com- nerability to violence, and economic opportuni- pletely deteriorated; water sources are contami- ties. These themes are explored throughout this nated with people now dumping garbage in the report. 17 ALPs also report the highest share of respondents who are ‘multidimensionally vulnerable’ 44 percent (the share is 26 percent for NAPs and 23 percent for IDPs). 4 Gender Norms and Reported Violence “Attitudes matter...because they are a barometer of people’s potential behavior...they can determine how society treats these groups, how these groups engage with soci- ety, and how the policies that aim to improve their status are implemented.” (World Bank 2013) This section looks at pervading gender attitudes different and is in the majority of cases not an ex- and norms18 and reported levels of violence. It il- pression of GBV.21 Although men report high lev- lustrates that communities with the most unequal els of interpersonal physical violence outside the views on gender also report the lowest levels of home this is most often by a non-partner. Emo- GBV. This is likely because women may be less tional abuse is reported by both men and wom- likely to report GBV in communities with the most en, although it takes different forms. patriarchal19 gender norms where violence may not be recognized as such, and women may feel Overall, this research depicts a country struggling less at ease reporting violence. Generally, young- to reconcile highly patriarchal gender norms with er urban individuals have the most egalitarian new realities toward gender equality. Women, views on gender. While older rural individuals particularly more highly educated urban women, have the least. In terms of reported levels of GBV, seem to be at the forefront of change. Men, in- young urbanites report more and older rural vil- cluding those in the younger generations, seem lagers report less violence.20 to continue to struggle to embrace gender equi- table notions, though there are signs of change. Furthermore, reported experiences of GBV in this survey are in line with what was reported in the 4.1 GENDER NORMS IN GEORGIA last nationwide gender survey in 2010 (Chitash- TODAY vili and others 2010). Women endure the highest levels of GBV, including sexual violence, sexu- Gender norms are dynamic and are changing al harassment, and physical violence inside the with the younger generation. Generally, older home. Experience of violence for men looks very generations resist change and younger gener- 18 Social norms refer to patterns of behavior that flow from socially shared beliefs and are enforced by informal social sanctions (World Development Report 2012). Beliefs, or attitudes, thus precede and help determine social norms. 19 Egalitarian gender views refer to perspectives that prefer men and women to have equal rights, roles, and access to resources. These views have also been referred to as ‘progressive’ or ‘feminist’. By contrast, patriarchal gender views are non-egalitarian and are sometimes referred to as ‘traditional’ or ‘misogynist’. These views prefer strongly defined gender roles, usually to the advantage of men. 20 With the data at hand, it is not possible to disentangle the effects of reporting versus norms. To the extent that younger people are more likely to report violence because they recognize violence when they experience it, then higher levels of reported violence among youth would be expected. However, more egalitarian norms should also translate into lower levels of gender-based violence, assuming that youth experience of violence is caused by other youth. 21 According to the definition of GBV used in this report: “The term ‘gender-based violence’ is primarily used to underscore the fact that structural, gender-based power differentials between males and females around the world place females at risk for multiple forms of vio- lence. [...] The term is also used by some actors to describe some forms of sexual violence against males and /or targeted violence against LGBTI populations, in these cases when referencing violence related to gender-inequitable norms of masculinity and/or norms of gender identity.” (IASC 2015) 42 43 Figure 6: Percentage of Respondents Who Feel Figure 7: Percentage of Urban and Rural Men and Gender Equality Has Come Far Enough, by Age Women Who Feel Gender Equality Has Come Far Groups Enough 47.0% 47.0% 46%46% 50%50% 43.5% 43.5% 38.5% 38.5% 44%44% 40%40% 33.2%32.9% 33.2% 32.9% 30.1% 30.1% 42%42% 25.6% 30%30% 25.6% 40%40% 37.8% 37.8% 20%20% 38%38% 10%10% 36%36% 34%34% 32.7% 32.7% 0% 0% 32.0% 32.0% 32.1%31.8% 32.1% 31.8% 18 18 -24 -24 25 25 -34 -34 35-44 35-44 45 45 -54 -54 55 55 -64 -64 65+65+ 32%32% 30%30% Source: Population’s Life Experiences in Georgia Survey, WomenMenMen Urban Women Rural Urban Urban Rural Rural Urban Rural 2016. womenwomen women women menmen menmen Note: Data include both male and female respondents. Data are weighted by sampling weights. Source: Population’s Life Experiences in Georgia Survey, 2016. ations, who are more exposed to travel, the In- Note: Data are weighted by sampling weights. ternet, and social media, are more welcoming of change. While the survey measured a range of is seen for women and men overall. However, gender attitudes, women and men’s perception when broken down further, 32–33 percent each of whether gender equality has come far enough of urban men, urban women, and rural women in Georgia will first be explored. Figure 6 shows agree that gender equality has come far enough, that nearly 50 percent of elderly women and men while 43.5 percent of rural men agree (Figure 7). believe gender equality has come far enough, This large percentage of men, in rural areas, ac- but barely 25 percent of young adults (18 to 24 counts for the urban-to-rural difference on this years) feel the same. perspective. This discrepancy was seen for sever- al other indicators as well. Urban-rural differences are also strong, but only for men. For example, when ALP respon- Both men and women believe in patriarchal dents were asked the same question—has gender gender roles inside the home, while women equality come far enough—nearly twice as many are seen as equally capable as men outside men (41 percent) answered ‘yes’ than women (18 the home. Figure 8 shows that men should be percent).22 This is, by far, the largest discrepancy the household’s ultimate decision maker. An ALP among men and women of the sample groups. woman in Zugdadi confirmed, “It is absolutely Because all administrative line areas are rural, this true that a family cannot have two homemakers.” reflects an urban-to-rural dynamic. When looking The survey further indicates that women and at the national sample’s urban and rural popula- men also agree that women are equally capable tions (men and women combined), 32 percent as men outside the home. Figure 8 shows the in- of urban respondents and 38 percent of rural dicators with the smallest discrepancy between respondents say gender equality has come far male and female views. All of these indicators re- enough. The same 32 percent to 38 percent split late to outside-of-the-home, where both men and 22 It should be noted from the sample group profiles that ALPs are on average slightly older than IDPs or NAPs. 44 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Figure 8: Gender Role Indicators with the Smallest Discrepancy Between Male and Female Responses, ‘Do you agree with the following?’ Women are as competent as men to be president, prime minister, 89.5% or any other high office 93.4% 92.5% Women make good local leaders 95.6% 90.5% When women work they do not take jobs from men 94.9% 85.1% A man’s most important role is to provide for the family 81.6% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Men Women Source: Population’s Life Experiences in Georgia Survey, 2016. Note: Data are weighted by sampling weights. 90% 76.0% women 80% feel that women 69.0% are equal. Over 90 per- ion is taken into account in decisions made by 70% cent of men and women feel that women make the household (Figure 10). However, the data on 58.0% 56.0% equally good leaders as men. Neither sex feels 60% gender attitudes and norms are consistent with that women 50% are a threat to men’s jobs, but both other studies in Georgia that a woman’s first duty 40% to agree that a man’s primary duty is sexes tend is at home and that she should be obedient to 30% for the family. To sum up, Figure 8 and to provide her26.0%husband. A United Nations Development 9 show that men simultaneously believe in tradi- 16.0% Programme (UNDP) report (2013) found that 63 20% tional gender roles in the home while also believ- percent of respondents believe women should 10% ing that women are equally capable as men. be obedient. The focus group results confirm this 0% finding; as one ALP woman from Gori affirmed, should bewhen A womanacute Attitudes are more able to choose it comesA woman to au- should tolerate violence in A man should have the final her own friends even if her partner “Almost every woman order to keep her family together will say that at least once word about decisions in his tonomy and decision making. disapproves Figure 9 shows she has been forced by her husband to stop household/family the gender norm indicators with a discrepancy talking when he argues that she is a woman and between male and female respondents larger Women knows Men nothing.” This comment reflects views that than 10 percent.23 These indicators relate to a were expressed by a range of women in IDP and woman’s agency and well-being. Although dif- NAP FGDs. These marked differences in gender ferences are marked between women and men, roles inside and outside the home described ear- three-quarters of men and more than half of lier, as well as different opinions between women women agree that men should be the final de- and men when it comes to questions related to cision maker inside the household. This is not agency, are significant because international ex- to say that women’s opinions are not valued: in perience suggests GBV risk is lower when gender the 2016 Life in Transition Survey (LiTS) survey, norms, inside and outside the house, are more 89 percent of women indicated that their opin- equitable (Heise and Kotsadam 2015). 23 7 percent is roughly the required difference between indicators to make them statistically significant with the survey sample’s 3 per- cent margin of error. 85.1% A man’s most important role is to provide for the family 81.6% 45 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Men Women Figure 9: Gender Role Indicators with a Larger than 10 Percent Difference in Male and Female Responses, ‘Do you agree with the following?‘ 90% 76.0% 80% 69.0% 70% 58.0% 56.0% 60% 50% 40% 26.0% 30% 16.0% 20% 10% 0% A woman should be able to choose A woman should tolerate violence in A man should have the final her own friends even if her partner order to keep her family together word about decisions in his disapproves household/family Women Men Source: Population’s Life Experiences in Georgia Survey, 2016. Note: Data are weighted by sampling weights. Figure 10: Percentage of Respondents Who Agree with the Following Norms Statements, by Gender 100 88.7 88.6 88 85.7 90.3 90 82.9 76.2 74.5 76.7 80 68.3 70 64.1 59 60 45.6 49 45.9 50 44 37.3 36.4 40 30 20 10 0 A woman should do most of the husband is not working My opinions are taken into account in decisions made by the household Co-habiting partners should be married It is better for everyone involved if the man earns the money and the woman takes care of the home and children achieves university education Women are as competent as men to be business Men make better political leaders than women do Equal rights for women as citizens existin the country Equal rights for women as citizens is important for the country household chores even if the It is important that my daughter executives Women Men Source: LiTS. 2016. Background note on LGBT People to the Population’s Life Experiences in Georgia Survey. 46 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Table 6: Gender Views Profiles Less equal gender views More equal gender views ƒƒ Men ƒƒ Women ƒƒ Over 55 years ƒƒ 18−24 years old ƒƒ Rural ƒƒ Urban ƒƒ Bottom 40% income distribution ƒƒ Highest 60% income distribution ƒƒ Worse housing conditions ƒƒ Better housing conditions ƒƒ Non-college educated ƒƒ College educated Source: Population’s Life Experiences in Georgia Survey, 2016. Gender profiles prepared by the World Bank team based on data cross-tabulations and confirmed by regression analysis in Annex 3. These findings are aligned with other surveys more tolerant attitudes but does not allow us to measuring social attitudes and norms in Geor- compare trends over time. To better understand gia. Similarly, the LiTS for 2016 reveals that Geor- the drivers and scope of change, this section gians support more traditional gender norms draws upon other quantitative surveys to explore inside the house while supporting women’s how the norms have changed between 2010 and equality and roles outside the home. Similar to 2016. This subsection examines data on social this survey, responses between Georgian women norms from the LiTS developed by the Europe- and men vary more when it comes to decision an Bank for Reconstruction and Development making and autonomy (Figure 10). However, it is (EBRD), which provides nationally representa- important to note that the two surveys are not di- tive estimates among adults 18 years and older. rectly comparable given that the questions, sam- In this section, tolerance and aversion to people pling, and methodology used differ. who may be perceived to look or behave differ- ently from the majority population (for example, Overall, data suggest certain groups of people immigrants, ethnic minorities, antisocial behav- hold more egalitarian gender views than oth- ior, or sexual minorities) are considered. This ers in Georgia. When comparing the responses section indicates that tolerance for some forms of all surveyed individuals across several demo- of difference has increased while simultaneously graphic groups, it appears that young, urban, decreasing in other areas. affluent, educated women hold the most equal gender views. By contrast, older, rural, poorer, The LiTS’ efforts to measure public opinion of less-educated men hold the least equal gender marginalized groups show an increase in intol- views. This is generally true for all the study’s 25 erance related to people from other cultures or gender norm indicators. Table 6 shows the pro- countries, of low economic means, or who may files of those with the least and most egalitarian need more assistance (for example, the elder- gender views.24 ly and families with children) (Figure 11). At the same time, it shows a drastic decrease in stigma 4.2 CHANGES IN ATTITUDES OVER toward same-sex attracted people or people with antisocial behavior (for example, heavy drinkers TIME or drug users). However, the stigma for these This survey provides a snapshot in time of norms groups remains overall higher with more than 50 and attitudes of Georgian men and women. The percent of Georgians declaring not wanting, as dataset suggests that younger people display neighbors, homosexuals, people living with AIDS, 24 For details on how these profiles were calculated, see factors correlated to traditional views in Annex 3. This section seeks to identify what characteristics are correlated with the prevalence of norms and attitudes toward GBV. The indicators described in Section 1 Accept- ability of violence are used for this purpose (Table 50). 47 Figure 11: Stigma toward Various Social Groups in Georgia (2010–2016) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Different religion Unmarried couples Immigrants Different language Jews Different race Families w/children Gypsies AIDS Heavy drinkers Homosexuals Drug addicts Paedophiles Elderly Poor % adults 2016 % adults 2010 Source: LiTS. 2016. Background note on LGBT People to the Population’s Life Experiences in Georgia Survey. Note: The terminology in this graph reflects reflects the questions/wordings that LiTS respondents were asked. They do not imply a judgement on the part of the authors and do not align with official World Bank terminology. heavy drinkers, drug addicts, or pedophiles.25 status, children in household, and geographical 100% Tolerance for Roma people or pedophiles has residence). Moreover, the observed decrease in remained 80% constant between 2010 and 2016. stigma relative to homosexuals (or increase in tol- This seems to suggest a general trend toward a erance) is relatively the largest among the young- 60% muting of stigmatizing attitudes in Georgia for est cohorts (who coincidentally stigmatized the 40% that were heavily stigmatized against in groups most in 2010) and is even significant among the past, with a rising intolerance for people from older cohorts and larger among males, both of other20% cultures or with specific needs. which tend to hold more stigmatizing views (Fig- 0% ure 12). Part of the observed decrease in stigma (or rise 20%- in tolerance) is likely due to the composition of 4.3 GENDER NORMS RELATED TO the sample: the 2010 sample was older (and a 40%- higher share were retired) and more represented PHYSICAL AND SEXUAL VIOLENCE 60% by women—the - former of which generally tends Most Georgians feel it is not justified for a man to represent more conservative views. Using the Male, 18 - 24 Female, 18 - 24 Male, 45 - 54 Male, 25 - 34 Female, 55 - 64 Female, 45 - 54 Female, 35 - 44 Male, 65+ Female, 25 - 34 Female, 65+ Male, 55 - 64 Male, 35 - 44 to hit his spouse, except if notions of ‘honor’ - - - - - example of stigma against homosexuals, we see are at stake. Figure 13 shows that, on average, that a drop in stigma is significant among near- less than 10 percent of sample groups feel phys- ly all sociodemographic and socioeconomic ical abuse is justified for reasons like not com- groups, and the reduction over time%holds even Change (2010 - 2016) pleting housework. % of adults in 2010 However, tolerance is higher when accounting for various respondent charac- when ‘honor’ is at stake; the two most frequently teristics (gender, age, education, employment cited reasons justifying physical violence relate 25 The terminology in this section and in the graph utilizes LiTS terminology and reflects the questions/wordings that respondents were asked. They do not imply a judgement on the part of the authors and do not align with official World Bank terminology. F % adults 2016 % adults 2010 48 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Figure 12: Stigma toward Homosexuals in Georgia (2010–2016), by Gender and Age 100% 80% 60% 40% 20% 0% 20%- 40%- 60%- Male, 18 - 24 Female, 18 - 24 Male, 45 - 54 Male, 25 - 34 Female, 55 - 64 Female, 45 - 54 Female, 35 - 44 Male, 65+ Female, 25 - 34 Female, 65+ Male, 55 - 64 Male, 35 - 44 - - - - - % Change (2010 - 2016) % of adults in 2010 Source: LiTS. 2016. Background note on LGBT People to the Population’s Life Experiences in Georgia Survey. to faithfulness and obedience. In almost all cases, Figure 13: Responses to ‘A Man has a Good Reason men are more likely than women to justify abuse to Hit his Wife if’, by Gender (Figure 13). Again, rural men—ALP men—display a higher tolerance to hitting one’s wife than the 30% other groups with over 10 percent saying abuse 25% is justified for four of the five categories. The no- table exception is IDP men, where more women 20% than men feel abuse is justified, though data did not achieve representativeness for this category. 15% Therefore, although the sample suggests that 10% more women than men in IDP communities feel abuse is justified, it may not be true for the coun- 5% try’s total IDP population. Regressions also find that older respondents are more likely to consid- 0% er GBV acceptable. Generally, rural, less-educat- Women Men IDP Women IDP Men ALP Women ALP Men NAP Women NAP Men ed males are more accepting of GBV toward a partner. Although tolerance for violence is low, both She does not complete She refuses to have sex her household work to with him women and men often indicated that women his satisfaction She asks him if he has other are somehow responsible for the violence. In She disobeys him girlfriends several focus groups, respondents suggested He finds out she was unfaithful that women were partly to blame for abuse, with one participant noting, “I think a woman should Source: Population’s Life Experiences in Georgia Survey, not drive the man to the point that he raises a 2016. hand against her.” Across focus groups, wom- Note: Data are weighted by sampling weights. 49 en agreed that in cases of violence, society was fewer than half of ALP males believe a woman has quicker to blame women and excuse men. In one the right to refuse if she does not feel like having of the exercises in the FGDs concerning a young sex. By contrast, most feel the woman can refuse female teenager who feels harassed by a young sex if she is sick, he is drunk, or he mistreats her. man, many respondents noted that if she were to To sum up, in the most patriarchal communities, it complain to her parents, they might blame her seems that women need an excuse to refuse sex for provoking the unwanted attention and beat with their partners. her. Participants in FGDs also draw a distinction depending on the frequency and severity of the 4.4 WOMEN’S EXPERIENCES OF violence. Except for ALP respondents, no one ex- cused GBV or suggested that women should stay VIOLENCE with a chronically abusive husband. Rather, the We measure violence according to 16 indicators great majority of respondents felt that although in four categories that were selected from test- one episode of violence might be excused, sys- tematic violence could not, and not only the wife ed and commonly used questionnaires that have would suffer if she remained but also the chil- measured violence against women (see Box 3 dren. for the specific questions). The four categories include controlling behavior/emotional abuse, Most Georgians, except in rural communities, sexual harassment, sexual violence, and physi- also feel that women can refuse sex with a cal violence. We will refer to these categories as partner. Figure 14 shows that most Georgians ‘the four forms of violence’, or FFV. Contrary to feel a woman can refuse sex with her husband some surveys on GBV, this survey includes sex- or partner under most scenarios. However, for ual harassment as a distinct category. While it the poorest and most rural communities, ALPs, is a form of violence exerted onto another per- the percentage of respondents who feel refusal son against their will, we also included sexual is acceptable is startlingly low. From Figure 15, harassment because of what it reveals in social you can see ALPs lag behind each of the other attitudes. Controlling behavior/emotional abuse sample groups in sexual autonomy. For example, was included in the violence index for similar Figure 14: Percentage of Respondents Who Agree Women Can Refuse Sex for the Following Reasons, by Gender 90% 85% 80% 75% She does not want to 70% Her partner is drunk 65% She is sick 60% Her partner mistreats her 55% Her partner refuses to 50% use condoms 45% 40% Men Women IDP Women IDP Men ALP Women ALP Men NAP Women NAP Men Source: Population’s Life Experiences in Georgia Survey, 2016. Note: Data are weighted by sampling weights. 50 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES reasons given what it can reveal about the social ‘vulnerability profiles’ was tested. Specifically, a context in which violence occurs. The finding that set of regressions27 investigated, in a multivariate emotional abuse can be severe and have long setting, the relationship between reported level lasting impacts upon respondents’ well-being of the FFV and (a) sociodemographic variables emerged clearly in the focus group discussions. (age, level of education, marital status, household As one female ALP from Gori said, “It depends size); (b) geographic variables (urban or rural lo- on the beating, but sometimes physical wounds cation); (c) conflict status (NAPs, IDPs, ALPs); (d) heal faster than damage inflicted on the soul.” measures of empowerment (employment status Both controlling behavior/emotional abuse and before and after the conflict); and (e) economic sexual harassment can also create an enabling status (belonging to the bottom 40 percent of environment for other types of violence to occur the per capita consumption, housing indicator by blurring the lines between what is permissi- and ownership of assets). Table 8 shows, for each ble/non-permissible behavior by perpetrators, as type of violence, the profiles of the most vulner- well as affecting the victim’s sense of self-worth. able female respondents. Detailed results of the regression analysis are reported in Annex 3. It is Over a quarter of all Georgian women have worth repeating that these ‘vulnerability profiles’ experienced some form of violence. We find may be more indicative of who is open to dis- that 27 percent of women report having expe- cussing violence than who is most vulnerable to rienced at least one type of emotional abuse, violence. sexual harassment, sexual abuse, or physical abuse. Reported violence for women by differ- ƒƒ Controlling behavior/emotional abuse. ent individual characteristics is indicated in Table Younger, less educated women, living in 7. Among the total population, 16.3 percent of smaller households and IDP women not re- women reported being a survivor of at least one siding in collective centers most often report type of sexual abuse (sexual harassment, sexual emotional abuse, or controlling behavior. violence, or both) and 8 percent of women re- Controlling behavior is also more frequent ported experiencing physical violence. It should among women residing in Tbilisi compared be noted that these are not incidence rates26 and to other urban areas and rural areas. that women’s actual experience of violence and their reporting of violence differs (see section 4.5 ƒƒ Sexual harassment. Young, urban, single, or on underreporting of violence). We also collect- separated women report the most sexual ha- ed some information regarding economic coer- rassment. This is especially the case in Tbilisi, cion. Nearly 5 percent of women, respectively, which has the highest reported rates of sex- indicate that they have been forced to give up ual harassment. There seems to be a link be- all or part of their income to a spouse against tween sexual harassment and poor housing their will or to give up a job because their partner conditions, as both women living in inade- wanted them to. quate housing and women living in IDP col- lective centers report higher experience with From the descriptive statistics, overall, urban sexual harassment. women are most likely to report experiencing ƒƒ Sexual violence. Young and urban women violence, while women with worse housing are most often the targets of sexual violence. conditions report higher sexual harassment And, again, poor housing conditions play a and sexual violence. To better understand which role. Lower levels of education are also cor- groups of women are more likely to report differ- related with increased reporting of sexual ent types of violence, the relationships between violence. This type of violence, together with the FFV and several variables to create violence 26 While all surveys on GBV likely see a difference between reported violence and actual violence, the primary objective of this survey is to look at links between violence, conflict, economic opportunities and services. As such, the methodology and questionnaire differs from surveys that aim to collect incidence rates. 27 For a detailed presentation of the regression analysis, see Annex 3. 51 Table 7: Percentage of Women Reporting At Least One Form of Violence, by Type of Violence and Individual Characteristics Types of Violence Respondent Emotional Abuse/ Sexual Harassment (%) Sexual Violence (%) Physical Violence (%) Characteristics Controlling Behavior (%) Age 18–24 38* 23 3 6 25–34 20 15 5 8 35–44 26 18 5 11 45–54 27 16 3 8 55–64 22 11 5 7 65+ 16* 9* 5 6 Education Lower secondary 33 19 8 15* Upper secondary 20* 12* 3* 6* Vocational 25 17 5 10 University Degree 29 18 6 6 Location Tbilisi 40* 25* 7* 15* Urban 16* 9* 4 4* Rural 17* 11* 3* 5* Marital Status Single 26 17 4 7 Married-cohabiting 18* 12* 4 6 Separated 63* 34* 7 27* Housing conditions Adequate 27* 14 4 7 Marginal 19* 16 6 8 Inadequate 14* 17 1* 7 Welfare Bottom 40 19* 13 3 7 Top 60 26* 16 5 8 TOTAL 24 15 4 8 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: Data are weighted by sampling weights. Data report the share of female respondents who report at least one form for each type of violence, by individual characteristics. As mentioned in Box 3, in the violence index that served for the main analysis and regressions in this report, we included only two indicators of controlling behaviors: (a) ‘Belittled or humiliated you in front of other people?’ and (b) ‘Threatened to hurt you or someone you care about?’ For each characteristic, the table reports the results of the two-tail mean comparison test across each category and the rest of the respondents within the fe- male sample; * indicates a p-value < 0.10. 52 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Table 8: Women’s Vulnerability Profiles to GBV Controlling Behavior / Sexual Harassment Sexual Violence Physical Violence Emotional Abuse ƒƒ Young ƒƒ Young ƒƒ Young ƒƒ Less educated ƒƒ Less educated ƒƒ Single or separated ƒƒ Less educated ƒƒ Separated ƒƒ Urban ƒƒ Urban ƒƒ Separated ƒƒ Urban ƒƒ Living in smaller ƒƒ Less adequate housing ƒƒ Urban ƒƒ Living in smaller household ƒƒ Less adequate housing household ƒƒ Less affluent Source: Population’s Life Experiences in Georgia Survey, 2016. Note: Columns report only those variables which are significantly correlated with each type of violence, as a result of the regression analysis in Annex 3. physical violence, is the only category where Table 9: Average Number of Experiences of Violence separated women reported higher levels of for Women, by Population Group GBV. On this note, it is important to consid- er that sexual violence is possibly the most Type of Violence IDP ALP NAP severe form of GBV and, from FGDs, consid- Controlling behavior/ 1.2 1.6* 1.3 ered the most stigmatized as well. This would emotional abuse suggest that women would be less willing to Sexual harassment 1.5 1.9 1.5 admit they were survivors. Divorced or sep- arated women, by contrast, may be less en- Sexual violence 2.2 1.2* 1.7 cumbered by fear of a partner or social stig- Physical violence 2.0 3.2* 2.5 mas, so may be more willing to speak about Source: Population’s Life Experiences in Georgia Survey, sexual violence. 2016. Note: * indicates a p-value < 0.10. ƒƒ Physical violence. Younger, poorer, less ed- ucated women living in inadequate housing in Tbilisi report the most physical violence. significantly lower) as illustrated in Table 9. Separated women report more physical vio- lence than married or single women. Women in early marriages suffer more vio- lence than women who married as adults. In The most frequently victimized women who the survey, adult survey respondents were asked are subject to the measured violence live in if they had been married before the age of 18. poor housing conditions in Tbilisi. The vulner- Through this metric, certain trends for women ability profiles above are computed by looking who married early could be ascertained. Overall, at individuals who reported at least one form of 7.8 percent of the national sample married before violence. However, the analysis also examined the age of 18. The share is significantly higher for which individuals suffered multiple forms of vio- females but not significantly different among the lence. Specifically, by looking at individuals who main sample groups (NAPs, IDPs, ALPs). Com- reported more than one form of violence. These pared to the others, women who got married be- survivors overwhelmingly live in Tbilisi or belong fore the age of 18 were much more likely to suffer to the ALPs group and do not have stable em- emotional abuse, sexual violence, and physical ployment. This shows a similar trend to the earlier violence. This underlies the danger of early mar- analysis, where location seems to be a significant riage and the danger for young brides, especial- risk factor for women. Rural women also report ly. On a separate note, there were no significant significantly more physical violence and con- differences in GBV vulnerability among people trolling behaviors than IDPs or the general popu- with or without children in the home. lation (although the number of sexual violence is 53 Poverty is correlated with increased levels of Figure 15: Top Four Contributing Factors for the economic coercion. To measure economic co- Most Significant Partner Incident, by Gender ercion, respondents were asked if they had ever refused a job because of their partner or had Women Men their money taken away by their partner. From a 1. Alcohol consumption 1. Alcohol consumption bivariate analysis, it appears that ALPs, the poor- est sample group by almost all measures, were 2. Money problems 2. Disobedience the most likely to suffer this type of abuse. IDPs 3. Jealosy (respondent or 3. Difficulties at work in collective centers suffered more economic vi- partner) 4. Disputes over children olence than IDPs who live in private residences 4. Disobedience and who tend to be more affluent. A regression investigated the impact of different factors in a Source: Population’s Life Experiences in Georgia, 2016. Note: Data are available only for respondents who report multivariate sector, where the dependent vari- having experienced a serious incident by a partner or anoth- able is the probability to report at least one form er perpetrator. Due to the small number of responses, data of economic coercion (refusing a job because of in the chart cannot be considered representative. Women the partner and/or had money taken away). De- reporting serious incident by partner n = 111, Men report- tailed results are presented in Annex 3. Among ing serious incident by partner n = 31. the main results is that women living in Tbilisi report a higher level of economic violence. Ed- or at work. It should be noted that relatively few ucation does not appear to be a significant co- respondents chose to answer the questions on variate while not having technical skills increases most serious partner and non-partner incident the probability to report experiences of econom- and data on the most serious incident are not ic coercion. Also, as with survivors of sexual vio- representative of the sample populations.29 In lence, divorced or separated women report a sig- the FGDs, most of the discussions involved vio- nificantly higher level of economic violence than lence within the household or family. Internation- married women or single ones.28 Again, this likely al studies suggest that when women experience reflects their relationship status more than actu- physical violence, it is much more likely to occur al levels of economic violence; in non-egalitari- at the hands of a partner or family member.30 an marriages in Georgia, women are expected While women experience violence by their male to accept that the husband controls the family’s partners in the majority of cases, in the case of wealth. It should be noted that GBV has psycho- non-partner violence, in almost half the cases this logical impacts that can lead to diminished as- was carried out by another woman (for example, pirations, empowerment, and personal endow- a relative or acquaintance). While it is likely that ments. This can diminish an individual’s agency the type of violence and context vary for male and confidence; two factors linked to economic and female perpetrators, the data does not en- wellbeing (Flechtner 2014). able to link the gender of the perpetrator with a specific kind of violence. Women tend to experience physical violence at home. When women were asked where the Individual GBV incidents are often initiated by most serious case of violence at the hands of a drunkenness. Female respondents were asked partner took place, 86 percent said it occurred at what caused the most serious individual GBV in- home. The most serious non-partner incident oc- cident they suffered. Most frequently they said curred inside the home in almost half the cases, that alcohol consumption led to the violence. The while a quarter of the cases occurred at school next most common reasons they gave were mon- ey problems, jealousy, and disobedience. Men 28 Women who had never been in a relationship were excluded from this analysis. 29 In particular, only 5.7 percent of the sample report information about the most serious incident experienced by the current or any previous partner (142 respondents over 2,492 who ever had a partner, of which 111 are female and 31 male) and only 5.9 percent about the most serious incident by other perpetrators (159 over 3,014 respondents of which 76 are female and 83 male). 30 International experience suggests that men are overwhelmingly the perpetrators of physical violence against women and it typically includes a gender motive (Fulu 2013a). 54 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES also said alcohol consumption and disobedience with poor health care, legal, and social infrastruc- by the woman were to blame, but they also said ture, remains unknown.” Also, conflict-affected difficulties at work and disputes over children led people could be more highly traumatized or not to the GBV incident (Figure 15). recognize violence as such relative to the other violence they have experienced or witnessed in Despite experiences with sexual harassment times of conflict. These factors further obfuscate outside the home, generally, women do not how IDPs and ALPs in Georgia report GBV. alter their behavior for fear of violence. When asked whether they had altered their behavior Both women and men concur that violence is for fear of violence, 87 percent indicated they common in families but that this is not talked had taken no action. Of the few women who had about. All focus group discussions highlighted altered their behavior, actions taken included the commonplace nature of violence in the home. avoiding leaving the home alone (17 percent), As one IDP woman explained, “We got beating avoiding opening the door when home alone and cursing in our dowry.” The most common (13 percent), avoiding certain areas (8 percent), form of violence is emotional or psychological or avoiding home for fear of what might happen abuse. For women, emotional abuse most often there (2 percent). However, the data were not rep- takes the form of humiliation, insults, and impor- resentative.31 While respondents living in border tantly, control. Men control their wives’ move- villages noted that in some of them safety and ments outside the home, whom they visit, their security have declined, for the most part, people choice of clothing and hairstyles, and do not let continue to feel safe in public spaces (though it them express their opinions. Focus group partic- should be noted that the reverse is true for mem- ipants observed that emotional or psychological bers of the LGBT community, as is explored in violence is so common that people do not even section 4.8). recognize that it is a form of violence. Typical of women’s views as expressed in the FGDs, one 4.5 UNDERREPORTING OF VIOLENCE woman explained that, “Verbal humiliation and swearing to wife is like saying hello to Megrelian Certain demographics of women are less likely men. Others will agree that Megrelian men often to report GBV and conflict areas are especially openly humiliate women in public,” to which an- troublesome. Palermo and others (2014) carried other participant responded that “Such attitude out a 24-country analysis of DHS data from 2005 is regarded as bearing close to no significance.” to 2011 for over 280,000 women. This seminal re- Sexual violence was less frequently discussed, port finds that “formerly married and never mar- though women and practitioners mentioned that ried status, urban residence, and increasing age this is heavily stigmatized. A practitioner in Tbilisi were characteristics associated with increased said, “Sexual violence is rather prevalent in fami- likelihood of formal reporting.” This means that lies. No one speaks about it but every woman we rural women and married women are less likely had is a victim of it. Sometimes, they don’t even to report experiencing GBV to formal sources. know they are victims, they just think it is their Moreover, the Palermo study shows a weaker, but duty as the wife; this is the function of wife and significant, link between reporting and education she has to do it even against her will. They are levels, with less-educated women reporting less ashamed to speak about it openly.” frequently. They also show an apparent relation- ship between reporting and wealth, but the direc- Overall, the participants noted that violence tion of the correlation is unknown. In reporting in is not talked about given filial duty and stig- conflict zones, they say, “the magnitude of GBV, ma. An older ALP woman from Zugdidi explains, especially in situations of civil conflict or contexts “When a woman goes out and says her husband 31 Data are available only for respondents who report having experienced a serious incident by a partner or another perpetrator. Wom- en reporting serious incident by partner N=111. Women reporting serious incident by other perpetrator N=76. As respondents could mark all the answers that applied, the sum of the shares is not 100 percent. 55 beats her, that woman does not want a family.” to rise in response to the increased amount of This means that if she reports GBV, she risks los- GBV awareness raising initiatives that have taken ing her family or her chance at starting a family. In place in Georgia over the last five years. Second, many FGDs, participants pointed to the concern focus group participants indicate that GBV is very with other members in the community finding common and more prevalent than the survey out about the violence. As one male participant reports, as previously described. Third, younger explained: “There is a serious problem in the generations tend to hold more progressive views family, woman gets married and husband beats on GBV and, over time, more people are likely her to death but parents tell her to endure it be- to report GBV. The fact that younger generations cause it’s a shame to take a woman back home, have more gender egalitarian views could bode relatives, friends will have a topic for gossip.” As well for the future and may be an indication of such, there are indications that experiences of gender norms changing in society. However, violence may have been underreported in this this assumes that only the cohort effect is at play survey.32 here. It is also possible that younger people have more egalitarian views, regardless of the gener- 4.6 TRENDS IN WOMEN’S ation, and that as they grow older, they become more conservative. While the data suggest peo- EXPERIENCES OF VIOLENCE ple may be more comfortable reporting GBV, the Findings on reported violence are not very data cannot confirm it. different from findings in other surveys, al- though not directly comparable. This survey 4.7 TRENDS IN MEN’S EXPERIENCE shows that 9.7 percent of Georgian women expe- OF VIOLENCE rienced at least one instance of sexual or physical violence. About 7.7 percent of women report suf- Men experience physical violence outside the fering physical violence and 4.2 report suffering home from non-partners. Around 14 percent sexual violence. A 2010 UNFPA survey in Geor- of men report physical violence. This includes gia showed that 6.9 percent of women reported street fights, military violence, and other forms of suffering physical violence and 3.9 percent re- violence that are not necessarily gender based. ported suffering sexual violence (Chitashvili and When asked where the most serious non-partner others 2010). The UNFPA survey and this survey event took place, 85 percent of men reported have differing methods for measuring GBV, so that the most serious case occurred in the streets these figures are not comparable, but they high- or other public spaces and that this was by a man light that reporting trends are similar, with report- in 93 percent33 of the cases. In fact, in 54 percent ing possibly increasing. This survey also finds that of the cases, there were multiple perpetrators of 23.8 percent of women report suffering emotion- this violence. al abuse, while the 2010 survey shows that 14.3 percent of women reported suffering emotional Men also frequently report controlling behav- abuse. Again, these data points are not directly ior/emotional abuse. Almost 17 percent of men comparable. reported having experienced this at least once. In the context of relationships, men’s experience Most indications attribute increases in GBV to of emotional abuse seems linked to their gender more openness to report to surveyors rather role. As seen in Section 4.3 on gender norms, than more violence. There are a few reasons both women and men largely expect men to pro- for this. First, key informant interviews show that vide for the family. Being unable to meet these stakeholders, like UN Women, expect GBV levels expectations is a source of anxiety for men. In 32 The same study shows that 41 percent of women in Eastern Europe and Central Asia, of which Georgia is a part, shared their GBV experience with someone, even privately, but only 9 percent reported the event to a formal source (Palermo, Bleck, and Peterman 2014). Understanding that Georgia is not the same as the rest of Eastern Europe and Central Asia, this still allows us to estimate that experiences of GBV are likely to be higher than what the survey shows. 33 Given the few numbers of respondents for the most serious non-partner incident (n = 83), this data is not representative. 56 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Table 10: Percentage of Men Reporting At Least One Form of Violence Respondent Characteristics Types of Violence Emotional Abuse/Controlling Sexual Harassment (%) Sexual Violence (%) Physical Violence (%) Behavior (%) Age 18–24 35 13 7 20 25–34 21 11 4 20 35–44 21 10 3 10 45–54 26 10 2 12 55–64 24 13 2 10 65+ 21* 9 1 13 Education Lower secondary 30 6 3* 17 Upper secondary 22 10 4 13 Vocational 25 12 2 15 University Degree 30 15 2 15 Location Tbilisi 35* 15 1 20* Urban 22 15 5 19 Rural 20* 6* 3 8* Marital Status Single 28 12 4 19* Married-cohabiting 23 11 3 12* Separated 26 4* 4 11 Housing conditions Adequate 25 12 2 16* Marginal 21 7* 5 11 Inadequate 32 14 2 9 Welfare Bottom 40 19* 8* 2 13 Top 60 28* 13* 5 14 Total 24 11 3 14 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: Data are weighted by sampling weights. Data report the share of male respondents who report having had at least one form of each type of violence, by individual characteristics. As mentioned in Box 3, in the violence index that served for the main analysis and regressions in this report, we included only two indicators of controlling behaviors/emotional abuse: (a) ‘Belittled or humiliated you in front of other people?’ and (b) ‘Threatened to hurt you or someone you care about?’ For each characteristic, the table reports the results of the two-tail mean comparison test across each category and the rest of the respondents within the male sample; * indicates a p-value < 0.10. 57 communities such as Zugdidi, where female la- during conflict than women. This is because they bor migration was common, FGDs shared their are either combatants or potential combatants to negative views of the phenomenon. Two male the opposing side. Although this research main- focus group participants remarked: “When men tains a more traditional definition of GBV that fo- send their wives, mothers abroad so they will cuses on violence against women, this expanded send money from there and men are on women’s definition and the dangers presented to young back, it is one of the painful issues”. “Yes, this is vi- men and boys are acknowledged. olence already, you, a grown-up man do not have a job and woman has to go abroad and keep you Finally, men are typically the perpetrators of fed.” In several FGD, discussants remarked that violence. FGD respondents overwhelmingly men humiliate women and undermine the wom- concurred that in cases of physical abuse, men an’s self-esteem, while women humiliate men are almost always the aggressors. A female IDP when men are unable to fulfill their provider role. from Abkhazia asserted, “I’ve never seen an abused man around me, but there are so many While the violence experienced by women abusers from husbands to fathers-in-law.” Other and men is structurally different, men also ex- studies clearly show that men are typically the press experiencing violence in Georgia. Table perpetrators of violence. Romans, Poore, and 10 describes the four types of violence that men Martin (2000) show that males are typically the experience. As indicated earlier, men more fre- perpetrators of domestic violence. And Cincotta quently experience violence by other men. There (2005) argues that a large demographic youth may also be less stigma attached to discussing bulge, especially of young men, increases a such types of violence than for women. This country’s risk of armed conflict. means that contrary to women, occurrences for men may be less underreported. Nevertheless, 4.8 TRENDS IN VIOLENCE AGAINST men’s experience with violence points to a less SEXUAL MINORITIES researched area in Georgia that warrants further exploration. Although legal protections for LGBT people are relatively advanced in Georgia, the so- Men are much more likely to participate in cial stigma thrives. The drivers of stigma are armed conflict. This will be discussed in more numerous and include (a) beliefs in maintaining detail in section 5, but it serves to highlight a few traditional family units and the rise of ‘tradition- points here. Five times as many men than women al values’, (b) the view of non-normative sexual have military experience in Georgia. Moreover, 24 orientation and gender identity as abnormal, (c) percent more men than women have witnessed the visible differences in appearance or behav- or experienced conflict violence. This shows that ior of some LGBT people, and (d) internalized men are much more prone to military violence in homophobia and transphobia within the LGBT Georgia than women. Some researchers argue community. Some quantitative measures of per- that this type of vulnerability to violence from spectives of the general population show that armed conflict is also a form of GBV. Carpenter tolerance toward LGBT people is likely rising. (2006) makes the case that forced conscription But, many focus group discussants did not per- and ‘sex-selective massacres’ are much more like- ceive this. One Tbilisi female said, “Once I asked ly to target men and boys. He suggests that it is my father what he would do if he discovered his because of the victim’s gender that he is targeted. daughter was a lesbian. He said he’d kill her.” The International Peace Research Institute Oslo (PRIO) database (Ormhaug, Meier, and Hernes Members of the LGBT community are vulner- 2009) shows that men are more likely to be killed able to violence.34 Physical, sexual, and psycho- 34 LGBT people were not included in our survey out of both ethical and practical concerns. Questioning LGBT people on sensitive GBV issues can put them at risk within their communities, either because of social stigmas or family concerns. To bring them into the analysis, six FGDs with LGBT people were carried out in safe settings with trusted LGBT-focused nongovernmental organizations. 58 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES logical violence against LGBT people is perva- many LGBT employees report harassment at the sive in Georgia. Participants in every LGBT FGD workplace from bosses or colleagues (UN 2015). reported being a survivor or witness of a physical Taken together, this can exclude LGBT people attack. This was the case in 2012 and 2013, when from attaining promotions or higher-paying jobs. peaceful protests to end LGBT discrimination led Regarding health care, doctors and health staff to violent mobs attacking the crowds, leaving sometimes treat LGBT patients with disdain and many LGBT people and their supporters wound- have been known to break doctor-patient confi- ed. In other cases, violence against LGBT people dentiality or even deny treatment. A male in Ku- culminated in death (Council of Europe 2016). taisi said, “When a gay person visits a doctor and Respondents say violence is committed by many that doctor asks about sexual orientation, the pa- types of people from strangers to family mem- tient immediately feels discriminated against.” A bers. But, almost always, violence is committed woman in Batumi recounted a story, “There were by men. Additionally, these attacks seem to be occasions in which a member of the community on the rise. Civil society organizations show that, disclosed they were HIV positive to a doctor, de- since the mob attacks in 2012 and 2013, there spite not being obliged to, and was denied a sur- has been an increase in official reports of vio- gical procedure as a result. Such incidents occur lence against LGBT people in Georgia (OHCHR frequently.” Finally, stigma is prevalent in schools 2013). and often comes from teachers as well as other students. This leads to LGBT students feeling in- Fairly rigid expectations about how wom- tense psychological pressure to conform. As a en and men should look and behave con- woman in Kutaisi expressed, “I know of one un- tributes to the vulnerability of LGBT people. derage gay-friendly kid who goes to my school. The more visible the perceived difference, the He’s probably in the 10th grade. The teacher in- more at risk LGBT people are. FGD participants structed his class to write an essay on the subject concurred that the most vulnerable to violence of ‘what harm unconventional people cause to a among LGBT people are feminine gay men, mas- traditional country’.” culine lesbians, and transgender individuals. An LGBT woman in Batumi explains, “The most op- This leads to many LGBT people being less pressed groups are those who are blatantly dif- open about their sexual orientation or gen- ferent, therefore transgender women, gay men, der identity and not reporting LGBT violence and butch lesbians are the most mistreated be- to the police. In a 2012 study of 150 LGBT peo- cause of their visible unlikeness from others.” She ple, 73 percent of respondents said they do not continued, “But, lesbianism is considered a less report to the police when violent acts are com- significant issue because of the assumption that mitted against them (OHCHR 2013). A lesbian in women can be dissuaded from being gay.” Batumi said, “If you are identified as a member of the LGBT community, police pay less atten- The impact of stigma on LGBT people’s access tion, or treat you disrespectfully.” FGDs and in- to employment and services is significant. To terviews with civil society organizations show that many in the general population, the antagonism survivors do not report to authorities out of fear they feel toward LGBT people justifies the abus- for public shaming, re-traumatization, or greater es perpetrated against them. Stigma diminishes stigmatization or because sometimes they do not agency, empowerment, and endowments (Klug- even know their rights (KIIs, FGDs, and Council of man et al. 2014) by limiting health, employment, Europe 2016). Civil society organizations also say and education opportunities. Many FGD partici- that ambivalent public officials and inadequate pants report that they were fired or not hired at all enforcement of hate crime laws leave LGBT peo- after employers perceived a different sexual orien- ple unprotected (KIIs, FGDs, and Bureau of De- tation or gender identity. A UN report shows that mocracy, Human Rights, and Labor 2015). 5 GBV, Conflict, and Economic Stressors The legacy of conflict contributes to GBV in Georgia. We found no clear evidence that conflict directly leads to increased GBV among the three population groups stud- ied. Rather, secondary economic stressors related to conflict and displacement, like unemployment, seem to have a consequence. This section shows that IDPs have the most direct experience with conflict and acutely feel these secondary economic stress- ors. Women have an added layer of vulnerability because, to a greater extent than men, they lack financial independence. This section looks at who is the most affected by conflict, how conflict affects livelihoods, how conflict-related economic stressors contribute to violence, and how greater economic independence may support survi- vors of GBV. 5.1 CONFLICT-AFFECTEDNESS these is considered ‘conflict-affected’. By this measure, one-fifth of respondents (20.3 percent) All sample groups from this study have ex- report being conflict-affected. Men were almost perienced some levels of conflict, but IDPs 50 percent more likely than women to be con- have experienced the highest levels. Figure flict-affected. This was driven by mostly men serv- 16 shows over 80 percent of IDPs said they “wit- ing in the military. Table 11 shows the percentage nessed or directly experienced conflict.” By con- of each group to the country’s male, female, and trast, less than 20 percent of the total population total population. can say the same. Of ALPs, 34 percent experi- enced conflict, the second highest proportion. The general population did not directly expe- It is likely that these are individuals who fled the rience the conflict. IDPs and ALPs are 20 times 2008 violence, but returned home. A female IDP more likely to have witnessed conflict than NAPs. now living in Gori recalled witnessing the con- People with military experience are 17 times flict, “Terrible things were happening. I once saw more likely than the national sample to have wit- someone beheaded. I was living in Isani, close nessed conflict. These results emphasize the high to the airport. My niece was five years old at the level of conflict exposure among IDPs, ALPs, and time. I remember planes flying over us. She was former soldiers. They also show that although afraid we would be bombed. This haunts them, most Georgians were alarmed by the armed con- don’t you think?” flict, they were relatively unexposed to its effects. More men than women have direct experi- People who participated in armed conflict are ence of conflict in Georgia. We combined three more exposed to the FFV studied. From the measures of conflict experience to create an survey, we see military experience is associated aggregate metric of ‘total conflict-affectedness’. with a twofold increase in experiencing different These measures include (a) being identified as forms of violence. When controlling behavior/ an IDP or ALP, (b) having been a combatant, or emotional abuse, sexual harassment, sexual vio- (c) having witnessed or experienced conflict in lence, and physical violence are combined, com- either conflict. Those who reported any one of batants were almost 2.5 times as likely to experi- 59 60 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Figure 16: Percentage of Respondents Who Indicate That They or Their Partner Have Directly Witnessed or Experienced Conflict in the 1990s or 2008, by Gender and Population Groupsa 100% 85.4% 87.8% 84.0% 90% 81.6% 78.8% 80% 65.9% 70% 60% 50% 40% 34.1% 30% 16.0% 21.2% 18.4% 20% 14.6% 12.2% 10% 0% NAPs IDPs ALPs Women Men Total a In several figures, data for males and females at the national level is presented but do not disaggregate data by gender for each subpopulation (NAPs, IDPs, and ALPs). The reason is that subpopulation data for each gender tends not to differ significantly from national-level data for each gender. Source: Population’s Life Experiences in Georgia Survey, 2016. Note: Blue is yes, orange is no. Data are weighted by sampling weights. Table 11: Conflict-affected Persons as a Percentage of the National Population, by Gender Men Women Total NAPs with direct experience of conflict 6% 6% 6% IDPs or ALPs with direct experience of conflict 94% 94% 94% Military experience 5% 1% 3% No military experience 95% 99% 97% Witnessed or experienced conflict violence 21% 16% 18% Did not witness or experience conflict violence 79% 84% 82% Source: Population’s Life Experiences in Georgia Survey, 2016. Note: Data are weighted by sampling weights. NAPs = Nonconflict-Affected Persons, IDPs= Internally Displaced Persons, and ALPs= Administrative Line Persons. ence these forms of violence compared to those Perhaps surprisingly, we found no link be- without military experience. Men and women tween being an ALP or IDP and higher levels with military service were also twice as likely to of GBV and the FFV. In fact, for each of the FFV, experience sexual violence or sexual harassment. NAPs reported the highest levels. IDPs, both in The small sample size made it impossible to look collective centers and private residences, report at the experience of only female soldiers. How- lower levels of the FFV than the national popula- ever, male soldiers were almost twice as likely to tion, though as seen in Section 4, there are differ- experience any form of violence. This is notable ences for specific subgroups of conflict-affected because this group is mostly male, suggesting women and depending on the type of violence. that experience in military groups can put men at Table 12 shows that ALPs consistently report the a higher risk of sexual violence. lowest levels of violence. This is true under the 61 Table 12: Percentage of Women Reporting at Least Table 13: Average Number of Violent Events One Form of Violence, by Population Group Experienced by Women, by Population Group Type of Violence IDP ALP NAP Type of Violence IDP ALP NAP Controlling behavior/ Controlling behavior/ 1.2 1.6* 1.3 18 14* 24* emotional abuse emotional abuse Sexual harassment 14 9* 15 Sexual harassment 1.5 1.9 1.5 Sexual violence 4 2* 4 Sexual violence 2.2 1.2* 1.7 Physical violence 6 6 8 Physical violence 2.0* 3.2* 2.5* Source: Population’s Life Experiences in Georgia Survey, Source: Population’s Life Experiences in Georgia Survey, 2016. 2016. Note: Data are weighted by sampling weights. For each indi- Note: Data are weighted by sampling weights. For each in- vidual characteristic, the table reports the results of the two- dividual characteristic, the table reports the results of the tail mean comparison test across each category and the rest two-tail mean comparison test across each category and of the respondents within the female sample; * indicates a the rest of the respondents within the female sample; * indi- p-value<0.10. cates a p-value<0.10. broadest definition of violence—at least one form However, there is a possibility that these rates of violence against both genders—and under the are skewed by a reporting bias. As seen, rural narrowest definition of violence—at least one form Georgians with the least adequate housing con- of physical or sexual violence committed against ditions have the most patriarchal gender views. women only. For younger age groups, these As reported in FGDs, in such communities, wom- trends hold consistent, with ALP youth reporting en feel especially constrained in reporting GBV. lower levels of the FFV and IDP youth reporting This suggests that GBV could be more severely similar levels of the FFV to the national sample of underreported in IDP and ALP communities that youth.35 It should also be noted that among IDPs, are characterized by more conservative gender we saw no statistical difference in the FFV levels norms. There are exceptions to these trends as between those displaced in the 1990s and those IDPs living in collective centers report more sexu- displaced in 2008. al violence and sexual harassment than those liv- ing in private dwellings. In other words, although The most well-off communities report the generally the most affluent report the most of the highest levels of violence. Perhaps more tell- FFV, it is not always the case. This underscores ing, Table 12 also indirectly shows that commu- the complexity within the sample groups. nities with the lowest levels of multidimensional poverty have the highest levels of reported vio- On the other hand, there are some statistically lence, although this does not imply a causal re- significant differences in terms of the average lationship between poorer communities and ex- number of violent events reported among sur- tent of violence. As we saw in the sample group vivors. In particular, ALP women, while character- profiles, ALPs are the poorest in both per capita ized by significantly lower sexual violence, report consumption rates and multidimensional pover- a significantly higher level of controlling behav- ty. Although IDPs have a relatively high per capita ior and physical violence. Table 13 illustrates this. consumption level, they also have a higher mul- IDP women also reported slightly higher levels of tidimensional poverty level and less adequate sexual violence, though this was not statistically housing than the national sample. If one looks at significant. Table 12 through an economics lens, it is clear the most affluent groups (NAPs and IDPs) report Georgians perceive that GBV increased from the most of the FFV. before to after the armed conflicts. Figure 17 shows that a plurality of IDPs and ALPs, by far the 35 Again, youth are defined as those between ages 18 and 26 years. 62 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Figure 17: Comparing Levels of GBV before the Conflict versus after the Conflict, by Sample Groupa 49.5% Men 18.8% 31.7% 41.6% Women 18.5% 39.8% 32.3% ALPs 13.5% 54.3% 35.8% IDPs 25.2% 39.0% 45.8% NAPs 18.5% 35.7% 0% 10% 20% 30% 40% 50% 60% Stayed the same Decreased Increased a In several figures, data for males and females at the national level is presented but do not disaggregate data by gender for each subpopulation (NAPs, IDPs, and ALPs). The reason is that subpopulation data for each gender tends not to differ significantly from national-level data for each gender. Source: Population’s Life Experiences in Georgia Survey, 2016. Note: Data are weighted by sampling weights. Table 14: When the Most Significant Non-partner flict. It is also noteworthy that in terms of the most Event Occurred, by Gender significant event committed by non-partners, for both women and men, this occurred after the Women Men conflict (Table 14). In contrast, the most violent Before the conflict 20% 31% event by a partner occurred in equal proportions 18.3% before and after the conflict. For women, for ex- During the conflict 13% 17% ample, 48 percent of the most significant partner After the conflict 67% 53% events occurred before the conflict, 7 percent Source: Population’s Life Experiences in Georgia Survey, during the conflict, and 45 percent after the con- 2016. flict. While this could be influenced by a recollec- Note: Data are available only for respondents who report to tion bias, 3.3% whereby respondents are 4% more likely to have experienced a serious incident by a non-partner. Due remember more recent events, it is not clear why to the small number of responses, data in the table cannot be considered representative. N total=159, N women=76, the recollection bias would be more marked for and N men=83. non-partners NAP for partners. Again, than IDP ALP this data is not representative given low responses to these most conflicted-affected sample groups, says questions and it is thus not possible to make any GBV increased after the conflicts. NAPs, by con- definite conclusions here. trast, mostly say GBV stayed the same. This differs from the perceptions of GBV levels during the People in conflict zones worried more about conflict, when most report GBV did not increase. random acts of violence than people not in Figure 17 also shows about 8 percent more wom- conflict zones. About one-third of IDPs (35 per- en than men believe GBV worsened after the con- cent) and ALPs (31 percent) worried that strang- 63 ers would attack them during the conflict. Only may have been the relatively short duration of 13 percent of NAPs felt the same. This suggests the conflicts. Because direct impacts of armed that people generally had a greater fear of vio- conflict are not reported to have led to increased lence outside of the home in conflict zones than physical or sexual assault, other factors that may they did in areas not affected by conflict. This was have contributed to them, and GBV more gener- 49.5% 8.8% perhaps because of less law and order in conflict ally, after the conflicts should be looked at. 31.7% zones. Still, most respondents said they did not fear being attacked during the conflict: 85 per- 5.2 CONFLICT’S ECONOMIC LEGACY 41.6% .5% cent of NAPs, 59 percent of IDPs, and 66 percent of ALPs. 39.8% IDPs and others who experienced conflict endure many difficulties, but disruptions to During the conflict, people were not worried 32.3% employment and livelihoods are paramount. about combatants or security forces physical- 54.3% From the questionnaire, information about labor ly or sexually attacking them. Figure 18 shows force status before the conflict and today (2016) 4 percent of ALPs and 3 percent of NAPs wor- 35.8% was collected. For IDP respondents, the conflict 25.2%ried about this.39.0% IDPs, the sample group with the of reference is the one that caused their displace- most conflict exposure, were much more fear- ment (1990s or 2008), while for NAPs and ALPs, ful of soldiers or police, at 18 percent. Overall 45.8% the first set of questions refers to before 2008. Of .5% though, most respondents never feared being IDPs, 85 percent report to have been displaced 35.7% attacked by police or military personnel. This because of the conflict in the 1990s, 12 percent highlights that GBV was not a systematized tool because of the conflict in 2008, and 3 percent % 30% 40% 50% 60% of war in Georgia, like it has been in some oth- because of both.36 To investigate how the con- d the same er conflict-affected Decreased Increased countries. The majority of fo- flict has disrupted the livelihood forms across the cus group respondents from conflict zones reit- different sample groups Figure 19 compares the erated this, saying there were not higher levels distribution of respondents across sectors limit- of GBV in conflict-affected areas during the war ing the analysis to IDPs displaced in 2008. Data though some reported being witness to such vi- show that IDPs endured the greatest disruption olence during the conflict. One factor that may to livelihoods between 2008 and 2016. In Figure have limited increased physical or sexual attacks 19, the closer the two lines on each graph are Figure 18: Respondents Who Indicated That They aligned, the less change occurred to that sam- Worried That Soldiers or Police Might Physically or ple group’s livelihoods. One can see that IDPs Sexually Attack Them or Their Partner during the underwent a major transition away from agricul- Conflict or Displacement, by Population Group ture toward service. The national population also moved slightly from manufacturing toward the 18.3% retail and service sectors, which likely reflects a general population movement to urban centers, but not nearly at the rate of IDPs. Administrative line areas are all rural and over 60 percent of ALPs remained in agriculture before versus after the conflicts. This is because the ALPs who fled 3.3% 4% the conflict in 2008 were largely able to return to their homes and careers, whereas those who remained displaced were forced into new liveli- NAP IDP ALP hood sectors. Source: Population’s Life Experiences in Georgia Survey, Overall, IDPs have also experienced the big- 2016. gest changes in unemployment. Figure 20 Note: Data are weighted by sampling weights. compares the share of respondents who are not 36 In case the respondent was displaced during both periods of conflict, they were asked to refer to the most recent one (2008). 64 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES working but looking for a job before and after employed people after the conflict. Men gener- 2008 across the three different sample groups, ally have higher unemployment than women, but again limiting the analysis to IDPs who were dis- this is because more men than women are active- placed in 2008 to make data comparable. Geor- ly seeking work. gia’s 2016 national census reported an 11.8 per- cent unemployment rate (GEOSTAT 2016). The The IDPs and ALPs who are employed are figure also shows IDPs underwent the greatest in- more likely to have unstable jobs than NAPs. crease in the share of unemployed respondents Figure 21 (chart A) shows that IDPs are between before versus after the conflict in 2008: from 3.4 13 and 20 percentage points less likely to have percent unemployed people to 14.3 percent un- stable work (either fully or partly) than NAPs. On IDPs Figure 19: Changes in Livelihoods for Survey Sample Groups before and after the 2008 Conflict 45 40 IDPs 35 45 30 40 25 35 20 30 15 25 10 20 5 15 0 10 5 forestry forestry and fishing Manufacturing Construction and retail and retail repair of m EducationEducation Service Service Healthcare Security service Energy Energy Transport Transport Arts/culture/ entertainment State sector Other 0 Agriculture, WholesaleWholesale and fishing Manufacturing Construction of m Healthcare Security service Arts/culture/ entertainment State sector Other trade; trade; repair Agriculture, Pre conflict Today Pre conflict Today ALPs 70 60 ALPs 50 70 40 60 30 50 20 40 10 30 0 20 forestry forestry and fishing Manufacturing Construction and retail and retail repair of m Education Education Service Healthcare State sector Transport Transport Security service Arts/culture/ entertainment 10 0 Agriculture, WholesaleWholesale and fishing Manufacturing Construction of m Service Healthcare State sector Security service Arts/culture/ entertainment trade; trade; repair Agriculture, Pre conflict Today Pre conflict Today Whol Agri tr Pre conflict Today 65 NAPs 35 30 25 20 15 10 5 0 Agriculture, forestry and fishing Manufacturing Construction Wholesale and retail trade; repair of m Education Service Healthcare Energy State sector Transport Security service Arts/culture/ entertainment Distribution Other Pre conflict Today Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The blue lines denote the proportion of each group engaged in each of the livelihood sectors before the conflict, and the orange lines denote the proportion of each group engaged in each of the livelihood sectors today. The more the lines diverge, the more disruptive conflict has been to livelihoods. Figure 20: Individuals Not Working but Seeking a Job over the Last Four Weeks, pre and post the 2008 Conflict, by Gender and Population Group 16 14.2% 14.6% 14.2% 14 12.7% 12 10.9% 11.0% 10 8.6% 8.1% 8 6.96% 7.0% 6 5.6% 4 3.1% 2 0 Today Today Today Today Today Today Pre-conflict Pre-conflict Pre-conflict Pre-conflict Pre-conflict Pre-conflict NAPs IDPs ALPs Women Men National Source: Population’s Life Experiences in Georgia Survey, 2016. Note: IDPs are limited to those who experienced the 2008 conflict. 66 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Figure 21: Stability of Employment, by Population Group (A) (A) Do you Do have you have aa stable stable employment? employment? (B) Does (B) your Does partner your have partner aa have stable employment? stable employment? 50% 50% 50% 50% 40% 40% 40% 40% 30% 30% 30% 30% 20% 20% 20% 20% 10% 10% 10% 10% 0% 0% 0% 0% NAPs NAPs IDPs IDPs ALPs ALPs NAPs NAPs IDPs IDPs ALPs ALPs Yes, Yes,fully fully Yes, Yes,partly partly No No Yes, Yes, fully fully Yes, Yes, partly partly No No Source: Population’s Life Experiences in Georgia Survey, 2016. the other hand, ALPs are the most reliant on ag- now than before the conflict. ricultural livelihoods, which are often seasonal or temporary. This partially explains why ALPs have 5.3 ECONOMIC STRESS AND GBV the least stable work situation: 22 percent report not having fully stable work, compared to 39 per- Respondents feel that economic stress trig- cent of IDPs and 41 percent of NAPs. In fact, the gers GBV. Among the people who believe GBV percentage of ALPs with unstable work is inverse- increased because of the conflict, a plurality ly proportionate to the percentage of NAPs with blamed economic stressors for the increase. Of stable work. This was true when asked of both them, 48 percent blamed idleness and unem- the respondents’ employment stability and their ployment, while 22 percent blamed general in- partner’s employment stability (see Figure 21, creases in violence, the next most common factor. chart B).37 Figure 22 shows the responses to this question when broken down by gender and conflict status. The conflicts in Georgia caused economic IDPs possess the largest proportion that feels this stress on the people involved. One IDP wom- way, while ALPs blame unemployment relatively an from the 2008 conflict, living in Tsilkani, elab- less and increased violence relatively more than orated, “After the war, we had to start from zero… the other sample groups. Despite most frequent- we had to rethink and identify once again who ly having unstable employment, rural ALPs seem we were and what we could do to survive. We less concerned with employment than the other were all very stressed and depressed from the groups. very beginning and we still feel this way now.” The survey verifies that most ALPs and IDPs feel Economic stress contributes to sample groups their access to jobs and economic situation be- feeling shame, particularly among IDP men. came worse after the conflict. Conversely, most Table 15 shows 40 percent of IDP men report NAPs (57 percent) felt their economic and jobs feeling shame if they are out of work or are not situations improved or stayed the same. As seen earning enough money. Again, by a large margin, in the survey sample, IDPs have higher per cap- IDPs are the most conflict-affected population in ita consumption than previously thought, but 61 the survey sample. The same table shows that percent still find their economic situation worse men, traditionally the family’s primary breadwin- 37 Key informants said that many Georgians remain unemployed because of a lack of quality jobs adapted to their skill set. This is evi- dent in rural areas where agricultural jobs are available but the pay is low and the jobs are unstable. 67 Figure 22: Response to the Question ‘If GBV Increased from Conflict, What Do You Think Is the Most Important Cause?’ Stress, idleness and lack of employment of perpetrator Overall increase in violent activity Alchol and drug abuse of perpetrator Woman taking over as primary breadwinner 0% 10% 20% 30% 40% 50% 60% Men Women NAPs ALPs IDPs Source: Population’s Life Experiences in Georgia Survey, 2016. Table 15: Percentage of Respondents Who Feel Shame If Poor or Unemployed, by Gender and Population Groups NAP Men IDP Men ALP Men All Women All Men I sometimes feel ashamed because I’m jobless or 30.3% 40.3% 21.8% 21.1% 30.7% don’t make enough money I sometimes drink or stay away from home when I 4.1% 8.4% 6.0% 0.5% 4.4% can’t find work Source: Population’s Life Experiences in Georgia Survey, 2016. ner, are more likely to feel shame from jobless- This link between economic stress and in- ness than women (31 percent to 21 percent). A creased GBV is supported by international male from Telavi reiterated this: “If you tell her to literature. For example, SIDA (2015, 14) argues, get a job, she might doubt your masculinity and “poverty and socio-economic insecurity is (…) strength.” ALPs, again seemingly unconcerned one of the factors contributing to GBV. (…) Inti- about their poverty or job stability, feel the least mate partner violence also tends to increase in shame among male populations. IDP men are contexts of poverty, partly reflecting ideals and also the most likely to channel this shame into expectations linking masculinity to the provider alcohol abuse. Stress, alcoholism, and economic role and subsequent sentiments of ‘failed mas- shame are factors that contribute to GBV (SIDA culinity.” 2015). Encouragingly, the survey shows that IDP and ALP youth (under 26 years) are not more in- Regression analyses confirm that economic clined to drink or experience stress compared to stressors trigger certain forms of GBV. To test other youth. this relationship, other controls were added to the basic regression model presented in Annex 68 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES 3 related to economic stress.38 First, two variables the place they were staying. This means that two that describe the Perception of Unemployment by out of five female GBV survivors returned to an the respondent were added to the model (wheth- abusive situation because they lacked the re- er s/he is ashamed to face the family if s/he is out sources to stay away. Interestingly, a large num- of work, and whether s/he admits sometime to ber of women and men participants across the drink or stay away from home when s/he cannot FGDs framed women’s economic dependency find a job). For male respondents, the probability on men as an issue, and this was true even among to report physical or sexual violence is significant- the most traditional groups, such as men and ly higher for those who admit to being ashamed, women living in rural border communities. As while respondents of both sexes more frequent- one male participant explained, “A woman is less ly report experiences of controlling behavior/ likely to become a victim of violence when she emotional abuse when they feel ashamed. Sec- is employed and [financially] strong.” One ALP ondly, for married or cohabiting respondents, a woman in Zugdidi stated it more optimistically, measure of Partner’s Stress was introduced in the “It is great when the husband works and provides basic model (which identifies whether the part- for the family, yet this is not enough. When you ner spends most of their time looking for a job get your own income as a woman, you feel more and/or if they sometimes drink or stay away from confident and empowered.” home when they cannot find a job) and a vari- able indicating Partner’s job stability. While the Women lack financial independence. Of fe- measure of partner’s stress does not appear to male respondents, 54 percent declared no per- be significant, female respondents whose part- sonal income, compared to 24 percent of male ner does not have a stable job are more likely to respondents. Moreover, women have fewer valu- report episodes of physical violence and sexual able possessions than men. Figure 23 shows harassment. Male respondents whose partner that women own significantly less land, fewer does not have a stable job appear to report more homes, and fewer cars than men. The only item often experiences of sexual harassment (see Ta- they possess more often than men is jewelry, but bles 3.11—3.24 in Annex 3). this is more of a marriage tradition than evidence of wealth. In Georgia, when a man and woman 5.4 WOMEN’S LACK OF FINANCIAL are wed, the jewelry of the man’s family is trans- INDEPENDENCE CONTRIBUTES TO Table 16: Reasons for Returning Home after GBV GBV RISKS Incidents (Women) It is not only that economic stress can trigger I missed my child 16% GBV; a lack of financial independence can also impede a survivor’s ability to seek protection. To keep the family together 4% Table 16 shows when a woman returns home af- My husband apologized 6% ter a GBV incident, it is often for the sake of the I had no money 8% family: 28 percent of the reasons women gave for returning were because they missed their child or I couldn’t stay longer with my friend or relative 12% husband and wanted to keep the family together. My husband had me returned 6% Another 12 percent of the reasons for returning I missed my husband 8% were because of some action by the husband, either he apologized or had her returned some- I had nowhere else to go 26% how. However, 36 percent of the reasons women He sold the apartment where I was living 2% gave to return home could have been avoided if survivors controlled more wealth. These reasons Source: Population’s Life Experiences in Georgia Survey, include not having money and losing access to 2016. Note: Due to the small sample size for this question (N=111), data cannot be considered representative. Multiple re- sponses were possible to this question, and therefore, the 38 Table 34 in Annex 3 total does not add up to 100 percent. 69 Figure 23: Responses to the Question ‘Do You Possess Any of the Following Items?’, by Gender 50% 45% 40% 35% 30% 25% Women 20% Men 15% 10% 5% 0% House Land Car A business Cash savings Jewelry Source: Population’s Life Experiences in Georgia Survey, 2016. Note: Data are weighted by sampling weights. Table 17: Reasons Cited for Not Securing a Job, by ferred to the bride. This works somewhat like a Gender dower, where the bride takes ownership. How- ever, in practice, this wealth generally falls within Women Men the household, of which the man is most often Too many people searching for jobs 77% 75% the head. Generally, neither men nor women in Georgia have any savings. As one female FGD Available jobs are too far away 23% 10% participant expressed, “No matter how long a Lack personal or political connections 20% 7% woman lives in a family, she does not have any Skills not commensurate with jobs 14% 12% rights in the family, nothing belongs to her.” Lack of assets, financial resources 9% 8% It is more difficult for women to find employ- Have given up a job because of 5% 0.5% ment than men. Finding employment is a chal- spouse lenge reported by both sexes in Georgia. The Source: Population’s Life Experiences in Georgia Survey, main difficulty according to Table 17 is the job 2016. Note: Data are weighted by sampling weights. Data refer to market is too crowded. However, for women, the respondents who are currently unemployed (N total=435, N location of the job and the lack of personal or women=131 and N men=304). political connections are additional constraints. With a woman’s defined roles being at home and ment (20 percent) as the most important needs her often being without transportation, she is pre- to reduce joblessness. Men say better access to vented from accepting jobs that are too far away. the job market and more financial support are the Also, because she is often at home, she does not most important factors. Women, by contrast, list have the same networking possibilities as men. skills development as a relatively more important Moreover, a woman is 10 times more likely than a type of support (33 percent for women versus 11 man to give up a job because of her spouse. percent for men). This reflects a lack of capaci- ty among women and suggests job trainings fo- To reduce joblessness, men report needing cusing on women could yield immediate bene- more money and access to the job market, fits. For both genders, the most important type while women report needing more money, of support to start a business is financial support access, and training (Figure 24). Respondents (32 percent), while better access to the job mar- cite better access to job markets (40 percent), fi- ket is more important for males (29 percent) than nancial support (32 percent), and skills develop- females (16 percent). Skills development again 70 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Figure 24: Most Important Type of Support to Get a Job, by Gender 100% 90% 80% Other (social services, application support, networking, others) 70% 60% Better access to job market (information on 50% vacancies, more job opportunities) 40% Financial support or lower loan rates for small 30% business development 20% Skills development 10% 0% Women Men Source: Population’s Life Experiences in Georgia Survey, 2016. Note: Data are weighted by sampling weights. Data refer to respondents who are currently unemployed (N total=435, N women=131 and N men=304). has an opposite trend in starting a business (22 support counts as the most important support to percent of women feel it is important versus 14 start a business by these more conflict-affected percent of men). IDPs and ALPs, by contrast, de- people (40 percent), compared to NAPs (31 per- clare simply requiring more money. Financial cent). 6 Services and GBV Women feel as though they are without recourse in protecting themselves from GBV. Despite Georgia making great strides to combat GBV over the last decade, wom- en at large still believe they are unprotected. They see little help from the police, reli- gious services, the legal system, service providers, or even their families. This section will look at how survivors respond to GBV and their general lack of awareness of GBV services. 6.1 SURVIVOR RESPONSES TO GBV en the situation, or result in violence being di- rected at the person intervening. Focus groups, This section leverages survey data to understand particularly representing male and female ALPs people’s responses to GBV, their awareness of and male IDPs, explained that outsiders can in- services, and their confidence in the institutions terfere with couples experiencing GBV only “if that provide them. The institutions and mecha- someone is getting killed” or “if it happens reg- nisms in place to support survivors are detailed ularly.” Moreover, in those cases, people should in Section 7 and Annex 2. only interfere if they know the couple involved. Most respondents feel that people outside of Figure 26: Where Respondents Went to Escape GBV the family should generally not intervene in cases of GBV, because doing so might wors- A relative’s Don't know Other house 2% Figure 25: Responses 9%to the Question ‘If a Man 10% Mistreats his Partner, Should Others Outside of the A friend or Family Intervene?’ neighbor’s Don't know house A relative’s Other Yes house 8% 2% 9% 48% 10% A friend or No neighbor’s 43% house Yes 8% 48% Parents’ house 80% No 43% Source: Population’s Life Experiences in Georgia Survey, 2016. Note: Data are weighted by sampling weights. Data are Parents’ available only for those house respondents who admit to have left Source: Population’s Life Experiences in Georgia Survey, home at least once after 80% a conflict with a partner N=50. Due 2016. to the small size of the sample, data cannot be considered Note: Data are weighted by sampling weights. representative. 71 72 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Figure 27: Stress, Responses idleness and lack ofto the Question employment ‘Did of perpetrator Figure 28: Reasons for not Reporting an Incidence (women) the Police Learn of the Incident?’ (women) else in violent activity Overall increase Someone Stress, idleness and lack of employment of perpetrator Other, reported Someone it else Did not think anyone 8% Other, 2% it reported Did not could or think anyone 8% Yes, I reported it 2% would do anything could or Alchol and drug abuse of perpetrator 16% it Yes, I reported Overall increase in violent activity would do 9% anything 16% 9% Wanted to keep it private Woman taking over as primary breadwinner 5% it private Wanted to keep Alchol and drug abuse of perpetrator 5% It was a Felt shame It wasmatter family a 0 8% 20 10 Felt shame 30 40 50 family 56% 60matter 8% 56% Woman taking over as primary breadwinner Men Women NAPs ALPs IDPs Too minor to report, No, I did not report it 14% Too minor to report, 82% No, I did not report it 0 10 20 14% 30 40 50 60 82% Men Women NAPs ALPs IDPs Source: Population’s Life Experiences in Georgia Survey, 2016. Note: Data refer only to female respondents who report to have experienced a “most serious incident by partner” N=111. Men: Do you believe more female police authorities would make a 29.6% difference in reporting of GBV by females? Figure 29: Perspectives on the Presence of More Female Police Officers, by Gender 70.4% Women: Men: Do Would you believe youfemale more be more inclined police to approach authorities wouldamake female a 10.7% 29.6% difference in reporting of GBVpolice officer? by females? 70.4% 89.3% Women: Would you be more inclined to approach a female No Yes 10.7% police officer? 89.3% No Yes Source: Population’s Life Experiences in Georgia Survey, 2016. Figure 30: Responses to the Question ‘Are There Female Police Officers in your Community?’, by Gender 89% 62% 70% 38% Yes 89% 30% 11% 70% No 62% 38% IDPs ALPs NAPs Yes 30% 11% 2016. No Source: Population’s Life Experiences in Georgia Survey, Note: Data are weighted by sampling weights. IDPs ALPs NAPs 73 According to them, strangers should not inter- port the incident because they felt ashamed or fere. The survey results show that this is a preva- wanted to keep the matter private. Of the female lent view with 52 percent of respondents saying respondents, 61 percent felt that reporting GBV either people outside the family should not inter- would stigmatize them. This confirms there is a vene or they did not know if people should inter- significant stigmatization involved with reporting vene (Figure 25). A recent femicide in Georgia, as GBV to police in Georgia. One older IDP male in reported by OC Media (2017), demonstrates this. Zugdidi voiced a view shared broadly with men, particularly among the more conservative ALP “the (victim) never appealed for help to and IDP focus groups, toward reporting to police anyone before the day of her murder, when when he said, “Why would anyone appeal to the she asked a neighbor to shelter her (…) police? The institution of snitching is not part of the neighbor refused, and this was when our cultural heritage.” the husband stabbed her in the street in front of their children.” Women are more likely to report GBV to fe- male police officers. Figure 29 indicates that In GBV cases, survivors largely go to friends’ 70 percent of males and almost 90 percent of or family members’ houses (Figure 26). Out females believe that having local female officers of the 50 respondents who answered this ques- would make it easier for GBV survivors to report tion, 44 were women. In all, 98 percent report- incidents. They feel women officers would be ed going to the home of someone they knew, of more likely to understand the situation and as- which 80 percent fled to their parents’ house. No sist if necessary. However, Figure 30 shows most respondent said they went to a shelter. Because communities still do not have female police offi- of the low response rates, these results are not cers: 30 percent of NAP communities, 38 percent representative of the population. of IDP communities, and only 11 percent of ALP communities have female police. It should be Respondents do not usually report GBV to noted, that this distribution indicates a clear ur- the police. If neighbors are not of much help in ban-rural bias. IDPs, who mostly live in cities, have stemming GBV, people feel the police are even the most female officers, while ALPs, who live in less helpful. In all, 82 percent of female respon- rural areas, have the fewest. When disaggregat- dents said they did not report GBV to the police ing the total population by urban-rural location, (Figure 27).39 FGDs show there are several rea- the differences are even more stark: 51 percent sons for this. First, although respondents in ev- of urban areas have female police officers com- ery FGD consider there is a role for police, the pared to only 9 percent of rural areas. large majority feel the police are not properly trained to respond to GBV. This is despite pub- GBV survivors indicate they do not find suf- licized efforts by the police force to sensitize of- ficient support from religious institutions. ficers in gender issues and recruit more women. Overall, 91 percent of Georgians favor the work Second, people believe the police take the side of the church (IRI 2015). In this survey, among of the perpetrator or discriminate against GBV those who think religious insitutions and services survivors, including LGBT people. Third, wom- are helpful, it is because they provide spiritual en tend to believe that if they call the police, the relief after an episode of GBV. In an interview repercussions from their husbands will be much with Eurasianet, Giorgi Zviadadze, rector of the worse. Figure 28 shows that according to the Tbilisi Theological Academy and Seminary, as- survey, 70 percent of women did not report to serted the church helps survivors by preaching the police, generally because they felt it is a pri- the importance of peace and understanding vate family matter or did not think it was serious within the family (Edilashvili-Biermann 2016). enough to merit a police intervention. Thirteen Overall, however, there was general agreement percent of all women also said they did not re- among focus group participants that religious in- 39 GBV here refers to individuals who said they had experienced a “most serious form of GBV.” 74 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES stitutions teach women to “obey” their husbands but most still feel these laws stigmatize women or and, in the best cases, husbands to “be kind” to expose them to more pain. their wives. As one IDP Sukhumi woman put it, “A priest always suggests for a woman to be patient. 6.2 LACK OF AWARENESS OF GBV Always reminds you that you’re a woman, a moth- SERVICES er, you cannot do this or that.” Most FGD partici- pants, discussing the church specifically, believe Respondents generally feel that services re- that it does little to reduce GBV or promote gen- lated to GBV are not available to them, espe- der equality, although ALP participants largely cially in rural areas (Figure 32). As seen in Sec- supported the church’s views on traditional mar- tion 7 and Annex II, the Georgian Government riage. They did, however, feel the church could provides many services to protect women or play a more active role in combating GBV, par- help them recover from GBV. These services in- ticularly in helping men understand that abusing clude legal assistance, medical coverage, men- family members is unjustifiable. IDP focus group tal health counseling, shelter protection, support participants had mixed views on the potential of groups, financial services, and others. However, the church, while FGDs in non-conflict affected only in the case of medical services, are even half areas felt that the church was unlikely to play a of the people aware of them. On average, fewer constructive role in addressing GBV. than 5 percent of respondents are aware of Geor- gia’s domestic violence shelters, and fewer than Women believe the legal system will not pro- 10 percent are aware of women support groups. tect them from GBV. According to Figure 31, Only one in five women say that psychological only two in five women say they are aware of GBV counseling is available in their community. As laws. Moreover, 56 percent of women believe it is mentioned before on legal protections, even if complicated to bring charges against an abuser. women are aware of services, they doubt their This was reinforced by key informants who said efficacy. In the case of domestic violence shel- the courts have a reputation for minimizing the ters, focus group discussants feel there are not severity of GBV cases. Meanwhile, three out of enough, and they are only available for “special four women think GBV laws are too weak to pro- cases,” likely referring to the victim status neces- tect them. The majority of respondents of both sary to access the state shelters. Respondents are genders believe legal mechanisms expose them misinformed on both counts: government-oper- to more problems. Women feel GBV laws are not ated shelters regularly run below capacity, and sufficient for protecting them from GBV by signif- they are generally available to any documented icant margins over men. To sum, fewer than half survivor of domestic violence. Youth were not of the respondents are aware of any GBV laws, more aware of services than older age groups, Figure 31: Perspectives on GBV Legal Protections, by Gender 42% I am aware of GBV laws 39% 56% It is complicated for a woman to charge a man with GBV 38% 74% Laws do not provide enough protection for GBV victims 57% 61% GBV laws expose women to more pain and stigmatization 53% Women Men Source: Population’s Life Experiences in Georgia Survey, 2016. Note: Data are weighted by sampling weights. GBV laws expose women to more pain and stigmatization 53% Women Men 75 Figure 32: Availability of Services for GBV Survivors, by Location, Population Group, and Gender Legal assistance Medical care (clinic or hospital) Mental health counseling Shelter house Women's support group Financial services 0% 10% 20% 30% 40% 50% 60% 70% 80% Rural Urban Men Women ALPs IDPs NAPs Source: Population’s Life Experiences in Georgia Survey, 2016. Note: Data are weighted by sampling weights. but IDP and ALP youth were more than two times All of this leads to women feeling without re- as likely as nonconflict youth to know about GBV course. Women do not believe the police or le- services. This could be because, as we saw in the gal system will help them, services are available sample group profiles, these communities are to them, or the church will assist them. They even more reliant on social assistance than the general think their family will not support them in cases population and are more in contact with service of GBV. One young ALP woman explained, “Do- providing authorities, and would therefore be mestic violence is generally not condemned by more aware of available services. family members, therefore victims are less like- ly to discuss the issue. Domestic violence does 76 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES not induce shock because it is commonplace. It er than do men. Women rate both short-term is considered an aspect of Georgian tradition. and long-term shelters as less useful (both 3.9), Physical abuse is perceived as a small conflict or though this probably reflects a lack of familiarity a quarrel.” When one considers the intense stig- with these services rather than actual usefulness. matization that comes from reporting GBV, it is Table 18 shows the services that received a score no surprise when one IDP woman in Gori said, higher than 4. These are ranked for each gender. “A woman won’t tell anyone even if her husband The table shows that women rank 10 services as skins her alive. She will pretend she has an ide- very useful, while men only rank 4 services. al family.” Both these views were typical of those expressed by women across ALP, IDP and NAP Many people have not heard of GBV aware- focus groups. ness-raising campaigns, but there have been several. These include the Rugby team’s He for Despite believing services are not available, She campaign, USAID’s ‘No Violence Against respondents said services would be useful. In Women!’ campaign, UN Women and the Geor- the survey, people rated availability and impor- gian National Film Center’s ‘Speak Out’ docu- tance of a range of services, not necessarily lim- mentary, and the Dutch and British Embassies’ ited to GBV40 using a mean value measurement ‘Public Awareness Raising and Monitoring of the between 1 and 5, with scores between 4 and 5 State Activities to Combat Femicide,’ among oth- being very useful and scores between 3 and 4 ers. Yet, despite these campaigns, most people being less useful. Overall, survey respondents (52 percent), have not heard of them (Figure 33). rated general medical care as the most useful Men seem particularly ignorant of these cam- (4.6), followed by religious services (4.2) and po- paigns, with 58 percent saying they have not lice services (4.1). According to rankings, men seen them. Of those who have heard of GBV feel the police and religious services are more campaigns, 99 percent saw them on television useful than do women. Women, by contrast, rank or heard them on the radio. Six percent also saw food security and access to clean water high- the campaign on the Internet. Only 5 percent saw Table 18: Perception of the Availability and Usefulness of Services to Protect Against GBV, by Gender Ranking of Importance Services Female Male General medical care 1 1 Food products 2 4 Church, mosque, or other religious services 3 2 Access to clean drinking water sources 4 — Police and security services 5 3 Access to markets 6 — Psychological assistance and counseling 7 — Reproductive health services 8 — Legal assistance 9 — Job, skills, business, and literacy training 10 — Source: Population’s Life Experiences in Georgia Survey, 2016. Note: Data are weighted by sampling weights. 40 Respondents did not assess the service, they only mentioned what they thought was useful. 77 Figure 33: Respondents Who Have Seen a GBV few point to the role of parents in educating their Advertisement or Announcement, by Gender children about violence. Yes No or don’t know Egalitarian gender views are more prevalent among people who have seen awareness-rais- ing campaigns. This relationship was consistent 46% for every gender perspective indicator tested. If 58% the respondent disagreed with the statements in Figure 34, they hold a more egalitarian view on gender issues. It is not clear if these people’s egal- itarian views were formed by GBV campaigns, 54% 42% or if they noticed GBV campaigns because they hold egalitarian views. It could also be a combi- Yes No or don’t know nation of both. What is clear, is that awareness Women Men raising and more egalitarian gender views are correlated though there is no causal relationship Source: Population’s Life Experiences in Georgia Survey, 2016. 46% between the two. Regression analyses indicates 58% Note: Data are weighted by sampling weights. DK denotes that the typical profile of a person aware of GBV ‘Don’t 100% Know’ campaigns is a young, urban, educated, and em- 90% ployed woman living in adequate housing. it through 80% another 54% medium such as print or out- reach organizations. A noteworthy42% point is that People aware of GBV campaigns report more 70% respondents largely advocate for more informa- violence. In the survey, a higher proportion of 60% tion on GBV (from schools, media, nongovern- people reporting violence were aware of GBV 50% Women Men mental organizations, and the government) but campaigns (57 percent) than were not aware (43 40% 30% 34: Relationship between Gender 67.3% Figure 57.1% Perspectives and Awareness 51.0% of 56.5% GBV-related Campaigns 20% 37.3% 34.4% 31.5% 30.5% 10% 100% 0% 90% Saw campaign Did not see Saw campaign Did not see Saw campaign Did not see Saw campaign Did not see 80% campaign campaign campaign campaign 70% A woman should not be able A man should have the A woman's most important Gender equality has 60% to choose her own friends final word duty is to cook and take come far enough 50% even if her partner disapproves on houshold decisions care of the home 40% 30% 67.3% 57.1% 51.0% 56.5% 20% Agree Disagree 31.5% 37.3% 30.5% 34.4% 10% 0% Saw campaign Did not see Saw campaign Did not see Saw campaign Did not see Saw campaign Did not see campaign campaign campaign campaign A woman should not be able A man should have the A woman's most important Gender equality has to choose her own friends final word duty is to cook and take come far enough even if her partner disapproves on houshold decisions care of the home Agree Disagree Source: Population’s Life Experiences in Georgia Survey, 2016. Note: Data are weighted by sampling weights. 78 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES percent). Among those who did not report any tice GBV campaigns, or second, people who see of the FFV in the survey, the split was 50/50. This GBV campaigns are more likely to report abuse. could be explained in two ways. First, people The direction of the relationship, however, is not who experience violence are more likely to no- clear.41 41 Table 51 in Annex 3 investigated which individual characteristics are more correlated with the probability to be aware of GBV cam- paigns and advertisements. Female, younger and more educated respondents, living in urban areas appeared to be more aware of them. Including in the model as an explanatory variable the number of reported violence does not appear to be significant. 7 Policy Implications and Recommendations Georgia has a relatively robust political and institutional framework to combat GBV, but stronger capacity and better implementation are needed. This section looks first at important gender-related policy milestones reached over the past 25 years (Figure 35). It then maps the main stakeholders involved in GBV-related issues in the country. Based on key findings of the survey research and key stakeholder and institutional mapping, a series of recommendations are made to the Government of Georgia and to development partners. 7.1 NATIONAL POLICY AND ual orientation, gender identity and expression, LEGISLATION ON GBV political or other opinions, or other characteris- tics” (President of Georgia 2014). The Georgian government’s policy is to not In Georgia, there are legal protections for LGBT discriminate based on gender. The Constitu- people. Same-sex sexual acts between adults are tion states: “Every human being is free from birth legal, and both straight couples and same-sex and equal before the law regardless of (…) sex.” couples have equal age limits for consensual sex. Moreover, Georgia’s Criminal Code imposes Transgender people can legally change their sex criminal penalty, “for the material infringement of and name on state identification after gender-af- human rights on the basis of sex...” In 2010, Geor- firming surgery. It is prohibited in Georgia to dis- gia adopted the Gender Equality Law which de- criminate against LGBT people in hiring, serving, fines gender equality as “a part of human rights educating, or providing health care (Georgia referring to equal rights and obligations, respon- Labor code 2017). Hate crime legislation specif- sibilities and equal participation of men and ically protects LGBT individuals (ILGA 2017). The women in all spheres of personal and public life.” Law of Georgia on the Elimination of all Forms of This led to corresponding National Action Plans Discrimination, referenced above, includes sex- in 2011, 2014 and 2016. These documents raise ual orientation, gender identity, and gender ex- awareness on gender issues, empower wom- pression. Moreover, there are no laws preventing en, and eliminate gender-based discrimination the establishment or public expression of LGBT (Government of Georgia, 2010, 2011b, 2014 and civil society organizations. Despite these laws, 2016). While the Gender Equality Law defined LGBT people still face substantial discrimination discrimination solely on the basis of sex, the Law in Georgia (see section 4.2). on Elimination of all Forms of Discrimination was adopted in 2014 to “ensure equal rights of ev- Georgia is a signatory to key international ery natural and legal persons under the legisla- conventions on protecting women from vi- tion of Georgia, irrespective of race, skin color, olence, including on women and conflict. In language, sex, age, citizenship, origin, place of 1994, the Georgian Parliament ratified the Unit- birth or residence, property or social status, re- ed Nations (UN) Convention on the ‘Elimination ligion or belief, national, ethnic or social origin, of All Forms of Discrimination against Women’ profession, marital status, health, disability, sex- (CEDAW) without reservations. In 2011, the Par- 79 80 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES liament of Georgia adopted the first Action Plan ternational organizations, and local nongovern- for implementation of UN Security Council Reso- mental organizations (UNFPA 2010). But several lution 1325 on ‘Women, Peace and Security’. Fur- key informants for this study said that the mech- thermore, in 2014, Georgia signed ‘the Council anism’s requirements for referral and the roles of Europe Convention on preventing and com- of stakeholders were not clear. To rectify this, a bating violence against women and domestic vi- more robust version was submitted to Parliament olence’, better known as the Istanbul Convention. for consideration. This new version more pre- The Convention was ratified by Parliament in May cisely identifies the names and roles of national 2017 and took effect on September 1, 2017 (UN stakeholders, both inside and outside the gov- Women 2017). ernment. A major milestone in combating GBV in Geor- Georgia has developed a series of National gia was the 2006 Anti-Domestic Violence Action Plans related to combating domestic Law. Known formally as the law on ‘Prevention violence and women in conflict. Georgia devel- of Domestic Violence, Protection and Assistance oped a 2016 to 2017 National Action Plan relat- of Victims of Domestic Violence’, it imposes re- ed to UN Security Council Resolution 1325. The strictive and protective measures to temporarily resolution—known as the Resolution on Women, restrain abusers. A restrictive order, issued by po- Peace, and Security—urges countries to integrate lice, restricts the abuser’s access to the survivors gender perspectives into their conflict-related for one month. The restrictive order becomes a actions. The action plan directly addresses IDPs protective order, active for six to nine months, if it and others affected by conflict. It requires (a) a is approved by the court. However, in many cas- Gender Equality Council to recommend actions es, the court refuses to rule or denies the request. to prevent GBV during conflict; (b) the Ministry For example, in 2014, 902 restrictive orders were of Justice to develop laws against conflict-related issued by the police, but only 87 protective or- sexual violence; and (c) the Ministry of Defense ders were issued by the court (Office of Public and Ministry of Internal Affairs to train “military Defender 2015). More recently, under guidance units, peacekeeping forces and police units on from UN Women, the Gender Equality Council42 gender issues...especially on prevention, identi- proposed an amendment to exclude courts from fication and response to gender based violence the process, but it has not yet been considered against women and girls” (Government of Geor- by Parliament. Despite this, the anti-domestic vi- gia 2011a). olence law has slowly taken effect (Javakhishvili and Tsuladze 2011). To further protect women Also, in 2016, the Government developed a in relation to marriage, in 2014, forced marriage National Action Plan to combat GBV and pro- was criminalized and marriage before 18 years tect survivors. The plan, called the 2016–2017 now requires court approval (UN Council of Hu- Measures to be Implemented for Combating Vi- man Rights 2016). olence against Women and Domestic Violence and Protection of Victims/Survivors, includes In 2009, the National Referral Mechanism on benchmarks to (a) change attitudes about GBV, Domestic Violence was developed to coor- (b) reduce violence against women, (c) protect dinate national efforts on GBV. Supported by survivors, and (d) allow survivors to provide in- the United Nations Population Fund (UNFPA), the put for future laws on GBV. A major impediment main objective is to create a system for protec- to the implementation of action plans related to tion of victims and provision of timely and effi- women and conflict and domestic violence has cient assistance by identifying health care insti- been a lack of dedicated budgets for their imple- tutions that have initial interactions with survivors mentation (Government of Georgia 2014c and (Government of Georgia 2014c). It emphasizes 2016a). cooperation among government agencies, in- 42 The Gender Equality Council was created in 2012 as part of the National Action Plan related to UN Security Council Resolution 1325. 81 Box 3: Domestic Violence Shelters and Crisis Shelters Georgian women have access to domestic violence shelters but they are underutilized. These shelters are managed by ATIPFUND within Georgia’s Ministry of Labor, Health, and Social Affairs. There are currently four state-run shelters where women can stay for up to three months (renewable). However, the approval process requires a protective or restrictive order, official victim status, or the acknowledgment that the person seeking shelter is the survivor of a domestic violence crime under investigation (UN Women 2014). This burdensome approval process to achieve formal victim status takes a minimum of five days to complete. This means that survivors who cannot wait, do not want to go through the formal process of pressing charges, or feel they will be stigmatized for doing so cannot be admitted to shelters. Ostensibly, these restrictions were put in place to preserve these protective spaces for GBV survivors and prevent the influx of other vulnerable populations like the homeless. A 2014 UN Women study showed that there would not be enough shelters to receive survivors if they all applied or qualified for entry (UN Women 2014). However, the administrative red tape leaves government shelters often running under capacity (according to KIIs with shelter staff). Crisis shelters are also available to GBV survivors. Crisis centers, unlike domestic violence shelters, do not require GBV survivors to attain formal victim status. As a result, crisis centers are available immediately to survivors, without the administrative requirements. For this reason, crisis shelters are consistently at full capacity. Here, survivors can find legal, medical, or psychological support (UNFPA 2010). Currently, there are three crisis centers in Tbilisi and four others throughout the country, including in Gori, Kutaisi, and Zugdidi. Except for one state-run crisis shelter in Tbilisi, these are run by nongovernmental organizations. Acknowledging its importance, ATIPFUND made plans to open more government-operated crisis centers. One has opened already. The crisis shelters are improving their coordination with the domestic violence shelters, so that a survivor can transfer from the first to the second if necessary (according to KIIs). Figure 35: Timeline of Legislation and Policy Developments Related to GBV Ratification of the UN Convention on the First Action Plan to Combat Violence against 1994 2000 Elimination of All Forms of Discrimination Women against Women Adoption of the law on Prevention of Domestic The advisory and coordination body on Domestic Violence, Protection and Assistance of victims 2006 2008 Violence - the State Inter-agency Council is of Domestic Violence established State Fund for the Protection and Assistance of (Statutory) Victims of Human Trafficking also 2009 2009 National Referral Mechanism is developed charged with combating domestic violence and shelters Nationwide 24/7 hotline service for victims of Domestic violence is criminalized and new forms 2010 2012 GBV is launched are incorporated into the Criminal Code of Georgia Reestablishment of Inter-agency Council on Action Plan for implementation of the UN Security 2014 2016- Domestic Violence Council Resolutions on Women, Peace and Security 2017 (UNSCR 1325) Action Plan on the Measures to Be Implemented for Combating Violence against 2106 - Ratification of Council of Europe Convention on 2017 2017 preventing and combating violence against women Women and Domestic Violence and Protection of Victims/Survivors and domestic violence, i.e. Istanbul Convention Source: Stakeholder analysis prepared by ACT, 2016. Background note to the Population’s Life Experiences in Georgia Survey. 82 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES The Government of Georgia provides a range ber of nongovernmental organizations also pro- of services to survivors of domestic violence. vide services to survivors and support awareness The State Fund for Protection and Assistance of raising on GBV. More information on the shelter (Statutory) Victims of Human Trafficking (ATIP- services available in Georgia is provided in Box 3. FUND) was tasked with providing services to domestic violence survivors in 2009 and is the 7.2 STAKEHOLDER MAPPING main institution providing assistance to survivors of domestic violence and human trafficking. The There is a complex network of stakeholders range of services include shelter, medical and working on GBV issues in Georgia. Because no psychological assistance, legal aid, and rehabil- single entity is responsible for all government ac- itation services to the victims/survivors of sexual tivities related to GBV, many separate stakehold- violence (Government of Georgia 2014c). A num- ers focus on different parts of GBV. The closest Figure 36: Main GBV-related Stakeholders in Georgia GOVERNMENT AGENCIES The Interagency Local Self- Council for the Governments Prevention of Domestic Violence Prosecutor’s Ministry of Labor, Ministry of Ministry of Ministry of Ministry of Office Health and Social Justice Internal Affairs Education and Internally Affairs Science Displaced Persons, Accommodation Educational and Refugees of Institutions Georgia (School, College, University) Public The State Fund for Social Service Agency The Court Patrol Police Defender’s Protection and Assistance and District Office of Victims of Human Health Care Police Trafficking Institutions (Hospitals, Clinics) NON GOVERNMENTAL ORGANIZATIONS INTERNATIONAL ORGANIZATIONS Anti-Violence Union “Sapari” Women Consulting Center Georgian Young Network of United Nations Entity for United Nations “Sakhli” Lawyer’s Association Georgia Gender Equality and the Population Fund Empowerment of Women Employment ”Atinati” Association “Amagdari” Women Georgian Swedish International United States Agency Information Association of Georgian Orthodox Church Development Agency for International Center Social Workers Development Kvinna till Kvinna Source: Stakeholder analysis prepared by ACT, 2016. Background note to the Population’s Life Experiences in Georgia Survey. 83 Figure 37: Mapping Stakeholder Support of GBV Survivors Process of acquiring Shelters victim status Medical Institution Social Service ATIPFUND Agency ATIPFUND hotline 116006 Crisis Center GBV 112 Patrol Turns case over to the prosecutor if Prosecutor Survivor hotline Police it falls under the criminal code [NGOs providing various services, Issues short-term restrictive order, including crisis centers] which must be ruled upon by the court Anti-Violence Network Union “Sapari” of Georgia Rules for or against issuing a Georgian Young The Court Women’s Information longer-term protective order Center Lawyer’s Association Women Consulting Atinati Center “Sakhli” Source: Stakeholder analysis prepared by ACT, 2016. Background note to the Population’s Life Experiences in Georgia Survey. entity to a coordinating body is the Interagency ficulties admitting some GBV survivors because Council for the Prevention of Domestic Violence, of the burdensome registration process. These, but they do not provide services, which fall with- and other implementation problems encoun- in the realm of eight different government agen- tered by national stakeholders, are slowly being cies (Annex 2). Figure 36 maps the main gender improved upon. The Ministry of Internal Affairs, stakeholders in Georgia and Figure 37 shows the which includes the national and district-level complex arrangement in which these stakehold- police, are still implementing gender sensitivi- ers provide support to survivors. Here some areas ty trainings to their officers. As first responders, for improvement in the implementation process there has been criticism that many officers still are briefly discussed. Annex 2 reviews the roles of carry national gender biases that suggest GBV is national stakeholders working on GBV and briefly a private matter and not the business of the po- describes each stakeholder’s function. lice (Mushkudiani 2016). The court faces a similar criticism. During KIIs, informants said judges, re- The rapid policy reforms to protect survivors sponsible for issuing protective orders, still disre- from GBV mean that many agencies still lack gard some GBV cases and crimes against LGBT the capacity for proper implementation. As il- people. In July 2016, the Social Service Agency, lustrated in Box 3, the shelter system has had dif- operating under the Ministry of Labor, Health, 84 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES and Social Affairs, became responsible for GBV tial link between awareness-raising campaigns survivors as well, dramatically increasing the and more egalitarian views on gender. However, agency’s workload. This means that few of their less than half of the population has even seen a employees have been properly trained in GBV. GBV campaign. While general awareness cam- paign can play a role in shifting popular percep- 7.3 RECOMMENDATIONS AND tions around GBV, they should focus more on advertising available services for GBV survivors, LOOKING AHEAD particularly shelters. Information campaigns Based on findings of the survey research and the should be tailored to target audiences, including results of the stakeholder analysis and institution- to the vulnerable groups identified in the study— al mapping, the following are recommendations e.g. urban women, those in challenging housing for a development response to challenges of environments, former combatants, and youth/ GBV. young brides. To the Government of Georgia (GoG): To international development partners, includ- ing the World Bank: Focus additional resources and attention on implementation of existing GBV-related legis- Provide capacity building support to insti- lation and policy reforms. The recent national tutions involved in prevention and response action plans provide an enabling policy environ- to GBV in Georgia. Support could include for ment for robust action on GBV prevention and training, expansion of services, mainstreaming response. However, individual agency roles and attention to GBV in line agencies, and analytics/ responsibilities need to be further defined and tracking of GBV trends. For example, the State associated with dedicated budgets and quali- Fund for Protection and Assistance of (Statutory) fied personnel. The mandate of the Interagency Victims of Human Trafficking (ATIPFUND) and the Council for the Prevention of Domestic Violence Social Service Agency are required to provide could be further expanded to ensure oversight social services to survivors of domestic violence. over capacity building for implementation across However, there is limited capacity within the relevant agencies and departments. This would Agencies to implement these new mandates. De- also include a robust monitoring plan to estab- velopment partners could also support the GoG lish implementation milestones and reporting through bringing the experiences of other coun- requirements, and to provide updates to GoG tries and region in successful implementation leadership and the public. of GBV-related legislation and policy, including coordinated cross-agency responses. Partners Consider women’s economic security as a key can also encourage oversight and monitoring by dimension of the GBV challenge and a focus non-governmental organizations to hold GoG to in GBV prevention and response efforts. Study account in implementation of existing legislation findings point to the potential links between con- and policy reforms. flict, economic stress and GBV risk as well as the value of economic security in women accessing Mainstream attention to GBV in large-scale support and GBV services. Law enforcement and development projects, including a focus on social service agencies have traditionally taken those that could advance economic opportu- the lead on GBV prevention and response in nities and employment for vulnerable wom- Georgia; however, a more holistic response, in- en. Study results point to vulnerable populations cluding a focus on employment opportunities that could be targeted through development and economic independence could help in ad- programming. In particular, projects that focus dressing the challenge. on employment-generating activities and skills development to help increase options to women Invest in strategic communication, recogniz- affected by GBV. Gains may be even higher for ing the gaps in knowledge particularly on GBV conflict-affected people, who reported the high- services. Survey research also indicated a poten- est levels of unemployment and unstable work. 85 They also consistently believed these economic by this survey or upcoming surveys could help stressors lead to GBV. Employment focused pro- gauge GBV trends. It would also help understand gramming for IDPs and conflict-affected people the pervasiveness of underreporting and any im- should include attention to potential GBV risks provements in that regard. Future research could and investments to increase economic security. further explore profiles of those most vulnerable to GBV and allow for better targeting of preven- Invest in housing and support for IDP collec- tion efforts, information campaigns and services. tive centers while taking into account GBV This would include a future focus on men’s ex- risk. Study findings point to the potential link be- perience of violence and risk factors to adapt tween housing challenges and increased prev- future services and support. In addition, apply- alence of GBV, including for IDPs living in often ing longitudinal studies over time regarding the overcrowded collective housing. Development same subjects would help track the longer-term partners can consider GBV prevention and re- impacts of conflict and displacement. Improve- sponse strategies in designing housing solutions ments, such as these, to GBV’s knowledge-gen- for these vulnerable groups. erating mechanisms would allow organizations Deepen knowledge and track progress in re- to create benchmarks and better target their in- ducing GBV. Leveraging information provided vestments. 8 Bibliography Bureau of Democracy, Human Rights, and Labor. Chitashvili, M., Javakhishvili, N., Arutiunov, L., 2015. Country Reports on Human Rights Tsuladze, L., and Chachanidze, S. 2010. Practices for 2015: Georgia. United States “National research on domestic violence Department of State. http://www.state.gov- against women in Georgia,” UNFPA, Tbilisi. /j/drl/rls/hrrpt/humanrightsreport/index.ht- m?year=2015&dlid=252849. Cincotta, Richard. 2005. “Youth Bulge, Under- employment Raise Risks of Civil Conflict.” CAP (Center for American Progress). 2011. Global Security Brief No 2. “The Georgia Conflicts: What You Need to Know.” https://cdn.americanprogress.org/ Cornell, S. E. 2001. Small Nations and Great Pow- wp-content/uploads/issues/2011/02/pdf/ ers – A Study of Ethno Political Conflicts in georgia_conflicts.pdf. the Caucasus. London: Routledge. Carpenter, R. Charli. 2006. “Recognizing Gen- Council of Europe. 2016. ECRI Report on Geor- der-based Violence Against Civilian Men gia, Fifth Monitoring Cycle. Strasbourg: Eu- and Boys in Conflict Situations.” Security Di- ropean Commission against Racism and In- alogue 37 (1): 83–103. tolerance, Council of Europe. CEDAW (Convention on the Elimination of All EBRD (European Bank for Reconstruction and Forms of Discrimination Against Women). Development). 2016. “Life in Transition Sur- 2004. “Consideration of Reports Submit- vey 2016.” EBRD. ted by States Parties Under Article 18 of the Convention on the Elimination of All Forms Edilashvili-Biermann, Maia. 2016. “Georgia: of Discrimination Against Women: Geor- Women Taking Action Against Domestic gia.” CEDAW/C/GEO/2-3, CEDAW, New Violence.” Eurasianet.org, August 2. http:// York. www.eurasianet.org/node/79986. ———. 2006. “Concluding Comments of the Com- Ellsberg, M., and L. Heise. 2005. Researching Vio- mittee on the Elimination of Discrimination lence Against Women: A Practical Guide for Against Women: Georgia.” CEDAW/C/ Researchers and Activists. Washington, DC: GEO/CO/3, CEDAW, New York. World Health Organization, PATH. ———. 2012. “Consideration of Reports Submit- Falb, K. L., M. C. McCormick, D. Hemenway, K. ted by States Parties under Article 18 of the Anfinson, and J. G. Silverman. 2013a. “Vio- Convention on the Elimination of All Forms lence Against Refugee Women Along the of Discrimination Against Women Com- Thai–Burma border.” International Journal bined Fourth and Fifth Periodic Reports of of Gynecology and Obstetrics 120 (3): 279– States Parties Georgia.” CEDAW/C/GEO/4- 283. 5, CEDAW, New York. 86 87 Falb, K. L., J. Annan, M. Hossain, M. Topolska, D. ———. 2009, 2010, 2011, 2012, 2013, and 2014. Kpebo, and J. Gupta. 2013b. “Recent Abuse “Integrated Household Surveys.” Nation- from In-laws and Associations with Adverse al Statistics Office of Georgia (GEOSTAT). Experiences During the Crisis Among Ru- http://www.geostat.ge/?action=meurneo- ral Ivorian Women: Extended Families as ba&mpid=1&lang=eng Part of the Ecological Model.” Global Public Health 8 (7): 831–844. ———. 2014. “Preliminary Results of 2014 Gener- al Population Census of Georgia.” General Flechtner, Svenja. 2014. “How Poverty Affects Information. http://census.ge/files/results/ Aspirations; Gap Between Facts and Per- Census%20Release_GEO.pdf. ceptions of Inequality.” Inequality in Focus 2 (4). http://www.worldbank.org/en/topic/ ———. 2017. National Census Database, Nation- isp/publication/inequality-in-focus-janu - al Statistics Office of Georgia. http://www. ary-2014. geostat.ge/?action=page&p_id=1145& lang=eng. FRA (European Union Agency for Fundamental Rights). 2014. Violence Against Women: An Government of Georgia. 2007. “National Action EU-wide Survey: Main results. FRA. Plan for Gender Equality in Georgia 2007– 2009.” www.parliament.ge. Fulu, Emma, Xian Warner, Stephanie Miede- ma, Rachel Jewkes, Tim Roselli, and James ———. 2011a. “2012–2015 National Action Plan for Lang. 2013a. Why Do Some Men Use Vio- Implementation of the UN Security Coun- lence Against Women and How Can We Pre- cil Resolutions 1325, 1820, 1888, 1889 and vent It? Quantitative Findings from the UN 1960 on ‘Women, Peace and Security.’” Multi-Country Study on Men and Violence http://www.peacewomen.org/sites/default/ in Asia and the Pacific. Partners for Peace, files/georgia_napdec_27_2011.pdf. UNDP, UNFPA, UN Women, and UN Volun- teers. http://www.partners4prevention.org/ ———. 2011b. “Action Plan for Establishing Gender node/515. Equality 2011–2013.” https://matsne.gov. ge/en/document/view/1333925. Fulu, E., X. Warner, S. Miedema, and C. Liou. 2013b. Toolkit for Replicating the UN Mul- ———. 2013. “Action Plan on Prevention of Domes- ticountry Study on Men and Violence: Un- tic Violence and Protection of Victims of Do- derstanding Why Some Men Use Violence mestic Violence 2013-2015.” Tbilisi. against Women and How We Can Prevent It. Bangkok: UNDP, UNFPA, UN Women, and ———. 2014a. “Action Plan of the Government of UNV. Georgia on the Protection of Human Rights: 2014–2016.” Tbilisi. GBVIMS (Gender-based Violence Information Management System). 2010. “GBVIMS User ———. 2014b. “National Strategy for the Protection Guide.” GBVIMS. http://www.gbvims.com/ of Human Rights in Georgia: 2014–2020.” gbvims-tools/user-guide/. Tbilisi. GEOSTAT (Georgia’s National Statistics Office). ———. 2014c. “National Review of the Implementa- 2008 and 2013. “Women and Men in Geor- tion of the Beijing Declaration and Platform gia: Statistical Booklets.” Ministry of Eco- for Action.” Tbilisi. nomic Development of Georgia, Depart- ment of Statistics (GEOSTAT). http://www. ———. 2016a. “2016–2017 Measures to be Imple- parliament.ge/files/670_11253_663836_1- mented for Combating Violence against eng.pd.pdf. Women and Domestic Violence and Protec- tion of Victims/Survivors.” Tbilisi. 88 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES ———. 2016b. “National Action Plan of Georgia for Jewkes, R., E. Dartnall, and Y. Sikweyiya. 2012. implementation of the UN Security Council “Ethical and Safety Recommendations for Resolutions on “Women, Peace and Securi- Research on Perpetration of Sexual Vio- ty”” Tbilisi. lence.” Sexual Violence Research Initiative, Medical Research Council. Pretoria, South Heise, Lori L., and Andreas Kotsadam. 2015. Africa. “Cross-National and Multilevel Correlates of Partner Violence: An Analysis of Data Klugman, Jeni, Lucia Hanmer, Sarah Twigg, Jenni- from Population-based Surveys.” The Lancet fer McCleary-Sills, Tazeen Hasan, and Julieth Global Health 3 (6): e332–e340. Andrea Santamaria Bonilla. 2014. “Main Re- port.” World Bank Group, Washington, DC. Hossain, M., C. Zimmerman, L. Kiss, D. Kone, M. http://documents.worldbank.org/curated/ Bakayoko-Topolska, D. K. Manan, H. Leh- en/455941468152971238/Main-report. mann, and C. Watts. 2014. “Men’s and Wom- en’s Experiences of Violence and Traumatic MRA (Ministry of Internally Displaced Persons Events in Rural Cȏte d’Ivoire Before, During from the Occupied Territories, Accommo- and After a Period of Armed Conflict.” Brit- dation, and Refugees). 2011. “Improving ish Medical Journal 4 (2): e003644. Access to Agricultural Land for Forcefully Displaced Persons: Lessons from Georgia.” IASC (Inter-Agency Standing Committee). 2015. World Bank Presentation, November 9. “Guidelines for Integrating Gender-Based Violence Interventions in Humanitarian Ac- ———. 2017. “2013 Registration of IDPs.” Ministry tion: Reducing Risk, Promoting Resilience of Internally Displaced Persons from the and Aiding Recovery.” United Nations Chil- Occupied Territories, Accommodation and dren’s Fund (UNICEF) and United Nations Refugees of Georgia. http://mra.gov.ge/ Population Fund (UNFPA). http://gbvguide- eng/static/3181. lines.org/en/home/ Mushkudiani, L. 2016. “Domestic Violence Cases IDMC (Internal Displacement Monitoring Cen- Addressed by Police: Approach, Attitudes tre).2016. “Country Profile: Georgia.” http:// and Reality.” Master Thesis, Gender Studies www.internal-displacement.org/countries/ Program, TSU. georgia/. OHCHR (Office of the United Nations High Com- ILGA (International Lesbian, Gay, Bisexual, Trans- missioner for Human Rights). 2013. Violation gender, and Intersex Association). 2017. of the Rights of Lesbian, Gay, Bisexual, and “Map.” ILGA. https://rainbow-europe.org/. Transgender People in Georgia. Identoba. http://tbinternet.ohchr.org/Treaties/CCPR/ IRI (International Republican Institute). 2015. Shared%20Documents/GEO/INT_CCPR_ “Public Opinion Survey Residents of Geor- NGO_GEO_15206_E.pdf. gia 2015.” IRS. http://www.iri.org/sites/ default/files/wysiwyg/iri_georgia_pub - Ormhaug, Christin, Patrick Meier, and Helga lic_2015_final_0.pdf. Hernes. 2009. “Armed Conflict Deaths Dis- aggregated by Gender.” PRIO Paper 23, In- Javakhishvili, N., and L. Tsuladze. 2011. “Imple- ternational Peace Research Institute, Oslo. menting Domestic Violence Policy in Geor- gia: Impediments and Their Causes.” 6th Palermo, Tia, Jennifer Bleck, and Amber Peter- ECPR General Conference, University of man. 2014. “Tip of the Iceberg: Reporting Iceland, August 25–27. and Gender-based Violence in Developing Countries.” American Journal of Epidemiol- ogy 179 (5): 602–612. 89 President of Georgia. 2014. “Law of Georgia on ———. 2016. “The First Crisis Center for the Victims of the Elimination of all Forms of Discrimina- Domestic Violence has been opened in Tbili- tion.” Public Defender’s Office of Georgia si.” http://georgia.unwomen.org/en/news/ (Ombudsman). http://www.ombudsman. stories/2016/09/the-frist-crisis-center-for- ge/uploads/other/1/1662.pdf. the-victims-of-domestic-violence-has-been- opened-in-tbilisi. Public Defender’s Office of Georgia. 2015. Vio- lence against Women and Domestic Vio- ———. 2017. “Georgia Ratifies the Istanbul Conven- lence in Georgia. Special Report, Georgian tion, Europe and Central Asia.” UN Women Office of the Public Defender, Georgia. News, May 19. http://eca.unwomen.org/ en/news/stories/2017/05/georgia-rati - Romans, Sarah E., Marion R. Poore, and Judy L. fies-the-istanbul-convention. Martin. 2000. “The Perpetrators of Domestic Violence.” The Medical Journal of Australia ———. n.d. “Georgian National Rugby Team Sup- 173 (9): 484–488. ports He for She at the European Nations Cup Finals.” http://georgia.unwomen.org/ SIDA (Swedish International Development Co- en/news/stories/2016/03/georgian-nation- operation Agency). 2015. Preventing and al-rugby-team-supports-he-for-she-at-the- Responding to Gender-Based Violence: Ex- european-nations-cup-finals. pressions and Strategies. SIDA. UNDP (United Nations Development Programme) Spinelli, Adria. 2014. “Prevalence, Incidence and Georgia. 2013. “Public Perceptions on Gen- Impact of Sexual and Gender-Based Vio- der Equality in Politics and Business, Pro- lence in Areas of Humanitarian Conflict: A moting Gender Equality in Georgia.” UN Systematic Review.” UNC Chapel Hill Under- Joint Program. graduate Honors Theses Collection. https:// cdr.lib.unc.edu/indexablecontent/uuid:1c- UNFPA (United Nations Population Fund). 2010. b4a54c-6db7-4c9d-ac2e-acaad61073cd. “The Collection of Normative Acts on Pre- vention of Domestic Violence, Protection Traynor, Ian, and Luke Harding. 2008. “Surrender and Assistance of Domestic Violence in or Else, Russia tells Georgia.” The Guardian, Georgia.” Tbilisi. August 12. https://www.theguardian.com/ world/2008/aug/12/russia.georgia1. UNHCR (United Nations High Commissioner for Refugees). 2009. “Protection of Internally UN (United Nations). 2015. “Universal Period- Displaced Persons in Georgia: A Gap Anal- ic Review: Georgia.” https://www.upr-info. ysis.” UNHCR and EU. http://www.unhcr. org/sites/default/files/document/georgia/ org/4ad827f59.pdf. session_23_-_november_2015/js5_upr23_ geo_e_main.pdf . UNHCR and ISSA (Institute of Social Studies and Analyses). 2015. “Intentions Survey on Du- UN Council of Human Rights. 2016. “Report of rable Solutions: Voices of Internally Dis- Violence against Women, Reasons of Vio- placed Persons in Georgia.” http://www.ref- lence and Its Results.” Special Reporter’s world.org/pdfid/55e575924.pdf. UNHCR Visit to Georgia. and ISSA, Georgia. UN Women. 2014. “Georgia – National Report.” U.S. Department of State. n.d. “Sexual Harass- Georgia Component of a Multi-Country ment Policy.” https://www.state.gov/s/ocr/ Study on Support Services for Women and c14800.htm. Girls Subjected to Gender-Based Violence, UN Women & Council of Europe, May, Un- published. 90 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Vinck, P., and P. N. Pham. 2013. “Association of en’s Health and Domestic Violence Against Exposure to Intimate-Partner Physical Vio- Women: Initial Results on Prevalence, Health lence and Potentially Traumatic War-Related Outcomes, and Women’s Responses. WHO. Events with Mental Health in Liberia.” Social Science & Medicine 77: 41–49. ———. 2013. Global and Regional Estimates of Vi- olence Against Women: Prevalence and WHO (World Health Organization). 2001. Putting Health Effects of Intimate Partner Violence Women First: Ethical and Safety Recommen- and Non-Partner Sexual Violence. Geneva: dations for Research on Domestic Violence World Health Organization. Against Women. WHO. World Bank. 2013. Inclusion Matters: The Foun- ———. 2002. World Report on Violence and Health. dation for Shared Prosperity (Advance Edi- WHO. http://www.who.int/violence_inju- tion). Washington, DC: World Bank. ry_prevention/violence/world_report/en/ introduction.pdf. ———. 2016. Transitioning from Status to Needs Based Assistance for Georgia IDPs: A Pover- ———. 2005. “Sample Design, Ethical and Safety ty and Social Impact Analysis. World Bank. Considerations, and Response Rates (chap- ter 3).” WHO Multi-Country Study on Wom- ANNEXES Annex 1: Methodology The methodology is a combination of qualitative and quantitative research methods. The re- search implementation lasted 21 months from the first stages of design in November 2015 to the delivery of the final report in September 2017. It followed five phases, as outlined in Figure 1A.1. The conceptual framework and main research questions are presented in Figure 1A.2. The re- search looked at links between GBV and conflict, economic opportunities, and services. It also sought to understand gender norms as they might influence experiences and opportunities in each of these areas. Research questions are asked to each of the study’s three main social groups—NAPs, IDPs, and ALPs—and the men and women within each group. The study also asked these questions to members of the LGBT community but only through FGDs as LGBT people were not included in the survey. This was done to add an extra layer of privacy for LGBT people. A1.1 OVERVIEW OF RESEARCH COMPONENTS This study applies a triangulated research approach. Table 1A.1 shows the study’s basic research components. A1.1.1 Quantitative Survey Sampling Design and Questionnaire The survey targets men and women over 18 years old in all three target groups—NAPs, IDPs, and ALPs. Because of the sensitivity of GBV questions, we surveyed only women or only men in each sam- ple site. This reduced the risk of interviewing the respondent’s partner or friends of the opposite sex, which could lead to interpersonal conflict. Given the specificity of the three target groups, we followed Figure 1A.1: Implementation Phases Phase I Phase II Phase III Phase IV Phase V Research Design Initial FGDs Survey Implementation Second round Data Analysis and Stakeholder and KIIs of FGDs and and report Analysis KIIs writing NOVEMBER 2015 SEPTEMBER 2015 93 94 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Figure 1A.2: Conceptual Framework Conflict ƒƒ Are conflict-affeced people more at risk from GBV? ƒƒ Can displacement induced stress increase risks? Violence Service Livelihoods & economic opportunities ƒƒ What are the service needs of ƒƒ How have conflict-affected people’s livelihoods conflict-affected people? been impacted by the conflict? ƒƒ What are the service needs of the ƒƒ Does household economics impact the general population? experience of violence? ƒƒ What is the level of knowledge and ƒƒ What economic options and coping mechanisms access to services? are available to survivors? Table 1A.1: Basic Research Components Data Collection Stakeholder Analysis Quantitative Survey Focus Group Discussions Key Informant Interviews Method 1. Male NAPs 1. Government officials 2. Female NAPs 2. Academics 1. Male NAPs 3. Male IDPs 1. Institutions 3. Civil society 2. Female NAPs 4. Female IDPs engaged in GBV organization 3. Male IDPs 5. Male ALPs Target Group 2. GBV-related policies members 4. Female IDPs 6. Female ALPs 3. Background note 4. International 5. Male ALPs 7. Young women on LGBT people organization officials 6. Female ALPs 8. Elderly women 5. Thematic and regional 9. LGBT people specialists 10. GBV service practitioners Sample Size — 3,014 interviews 35 FGDs (280 participants) 10 in-depth interviews Random stratified Purposive sampling and Sampling Method Desktop review Purposive Sampling Sampling snowball sampling Source: World Bank team. Note: For the FGDs, LGBT people and GBV service practitioners are represented in italics to signify that these groups were not included in the quantitative survey. ANNEXES 95 three different methods to build samples for these groups: (1) For NAPs, GEOSTAT provided a list of census districts with populations broken down by age, gender, region, and settlement type. We excluded ALP communities that bordered conflict zones and individuals who reported themselves as IDPs from this sample. NAPs live in all areas of Georgia. (2) For IDPs, we used the MRA’s official registry of all IDPs in Georgia. This information was kept in complete confidentiality and only accessed by the local research team to build a sample and identify respondents. IDPs live in all areas of Georgia. (3) For ALPs, we used a government list of the communities where barbed wire fences were erected after the 2008 conflict. The list includes 88 settlements, 86 of which were verified by GEOSTAT. We built the sample and selected respondents from these 86 settlements. All settle- ments border Abkhazia and South Ossetia. The total survey sample size was 3,000 peo- Table 1A.2: Number of Survey Interviews and Size of ple, though we successfully interviewed 3,014 Sample (in brackets) people. To achieve representation, the NAP sam- ple required the most interviews, while the ALP   Women Men Total sample required the fewest. The number of inter- views (shown in Table 1A.2) ensured a 2–3 percent NAP 706 (700) 700 (700) 1,406 (1,400) margin of error, at a 95 percent confidence level IDP 506 (500) 502 (500) 1,008 (1,000) for the total sample. At certain points throughout the report, we examine certain subpopulations— ALP 301 (300) 299 (300) 600 (600) disaggregated by age, education, welfare level, Total 1,513 (1,500) 1,501 (1,500) 3,014 (3,000) housing condition, urban-rural location, conflict experience, and more. Often, these subpopula- Source: World Bank team. tions did not have enough survey responses to achieve these confidence levels. In these cases, we explicitly state they are not representative samples. Sample stratification Individual stratification strategies were applied to each population group based on that popu- lation’s geographic distribution. There were 33 substrata for all target populations. The sample size allocation is proportional to the adult population in each stratum. The resulting numbers were then rounded to multiples of 10, and in every substratum, at least 10 respondents were selected. Table 1A.3 shows how each substratum sample was calculated for the non-conflict-affected national population. The survey questionnaire was modelled after questionnaires that have been previously used and tested.1 The eight modules and themes covered are described in Table 1A.7. The modules were applied to all survey respondents, except for a limited number of questions that are relevant only to people with official IDP status, or minor wording variations for some questions depending on the gen- 1 The following tested survey questionnaires were used to inform the eight modules of the survey: the WHO’s multi-country study on women’s health and domestic violence against women (http://www.who.int/reproductivehealth/publications/violence/24159358X/en/), FRA’s EU-wide Violence Against Women Survey (http://fra.europa.eu/en/publication/2014/violence-against-women-eu-wide-survey-main- results-report), IMAGES (https://www.icrw.org/publications/international-men-and-gender-equality-survey-images/), USAID’s DHS Do- mestic Violence Module (https://dhsprogram.com/pubs/pdf/DHSQMP/domestic_violence_module.pdf.pdf), the World Bank’s Socio-eco- nomic Impact Assessment of the Presence of Syrians under Temporary Protection on Turkish Hosting Communities, The EU and Handicap International’s Livelihoods Assessment questionnaire, the Rural Health Research Center GBV Toolkit, the London School of Hygiene and Tropical Medicine questionnaire on Women’s Health, the Kinyanda War Trauma module, GEOSTAT’s IHS for 2015, and George Washington University’s Database of Questionnaires on Violence Against Women and Girls in Humanitarian Settings. 96 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Table 1A.3: National Sample Calculation Settlement Region Population Sample Stratum Region Type Code Male Female Male Female 1 Urban 11 Tbilisi 334,846 435,920 160 200 2 Urban 15 Adjara 63,920 74,343 40 40 3 Urban 23 Guria 11,321 13,611 10 10 4 Urban 26 Imereti 84,993 102,492 50 50 5 Urban 29 Kakheti 25,245 30,472 10 10 6 Urban 32 Mtskheta-mtianeti 7,642 8,639 10 10 7 Urban 35 Racha Lechkhumi and Kvemo 2,401 3,026 10 10 Svaneti 8 Urban 38 Samegrelo and Zemo Svaneti 37,583 45,841 20 20 9 Urban 41 Samtskhe-javakheti 18,694 22,124 10 10 10 Urban 44 Kvemo Kartli 60,317 69,282 30 30 11 Urban 47 Shida Kartli 34,454 41,348 20 20 12 Rural 11 Tbilisi 10,318 11,693 10 10 13 Rural 15 Adjara 56,292 57,853 30 30 14 Rural 23 Guria 31,609 34,177 20 20 15 Rural 26 Imereti 108,893 111,980 60 50 16 Rural 29 Kakheti 94,213 99,843 50 50 17 Rural 32 Mtskheta-mtianeti 25,977 25,139 10 10 18 Rural 35 Racha Lechkhumi and Kvemo 10,335 10,818 10 10 Svaneti 19 Rural 38 Samegrelo and Zemo Svaneti 66,903 69,596 40 30 20 Rural 41 Samtskhe-javakheti 39,100 41,770 20 20 21 Rural 44 Kvemo Kartli 87,446 92,441 50 40 22 Rural 47 Shida Kartli 47,952 50,054 30 20 Total 1,260,454 1,452,462 700 700 Source: World Bank team. ANNEXES 97 Table 1A.4: IDP Sample Calculation Population Sample Stratum Settlement Type Geographical Division Male Female Male Female 1 1 Urban 1 Tbilisi 25,760 34,266 200 220 2 1 Urban 2 Western Georgia 14,452 18,303 120 110 3 1 Urban 3 Eastern Georgia 5,233 6,472 40 40 4 2 Rural 1 Tbilisi 500 564 10 10 5 2 Rural 2 Western Georgia 8,276 11,553 70 70 6 2 Rural 3 Eastern Georgia 7,208 8,439 60 50 Total 61,429 79,597 500 500 Source: World Bank team. Table 1A.5: ALP Sample Calculation Population Sample Stratum Settlement Type Geographical Division Male Female Male Female 1 1 Urban 1 Tbilisi 25,760 34,266 200 220 2 1 Urban 2 Western Georgia 14,452 18,303 120 110 3 1 Urban 3 Eastern Georgia 5,233 6,472 40 40 4 2 Rural 1 Tbilisi 500 564 10 10 5 2 Rural 2 Western Georgia 8,276 11,553 70 70 6 2 Rural 3 Eastern Georgia 7,208 8,439 60 50 Total 61,429 79,597 500 500 Source: World Bank team. der of the respondent. Respondents self-selected which conflict they had been affected by. Respon- dents who had not been affected by conflict from the general population were asked to compare their current situation to the situation before the most recent conflict in 2008. A1.1.2 FGD Sampling Design The FGD target groups included NAPs, IDPs, ALPs, LGBT people, and GBV service providers like shelter workers. Most focus groups included participants between the ages of 25 and 60, but a select number of focus groups included individuals between the ages of 18 and 25 to gain a better insight into the perspectives of the youth. Of the conflict-affected FGDs, only IDPs and ALPs who experienced conflict in the 1990s or 2008 participated. As such, FGDs excluded IDPs who inherited their IDP status— those born to displaced persons but who were not themselves displaced—and ALPs who moved to, or were born in, ALP communities after the conflict. For example, in Tbilisi FGDs, only IDPs older than 35 years were invited to participate in FGDs. This allowed the research team to exclude persons too young to have lived through displacement. The survey, by contrast, included all IDPs and ALPs regard- less of their personal experience with conflict. LGBT participants were recruited using LGBT-focused nongovernmental organizations. This was done to ensure the safety and comfort of participants. 98 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES FGDs were performed over three rounds. The first round of FGDs informed the survey question- naire. The second round deepened the understanding of survey results. The third round included new demographics that the research team felt were underrepresented in the survey and the previous two rounds of FGDs. These demographics included youth, LGBT people, and GBV service providers. Table 1A.6: Focus Group Composition FGD Composition FGD Location First Round 1 Female IDPs residing in Tbilisi (1900s) (35–55) Tbilisi 2 Male IDPs residing in Mtskheta-mtianeti (2008) (25–50) Tserovani IDP settlement 3 Female ALPs residing in Shida Kartli (25–50) Gori 4 Male ALPs residing in Shida Kartli (25–50) Gori Second Round 5 Female IDPs residing in Tbilisi (1900s) (35–55) Tbilisi 6 Female IDPs in Mtskheta-mtianeti (2008) (25–50) Tsilkani 7 Male IDPs residing in Tbilisi (1900s) (35–55) Tbilisi 8 Male IDPs in Mtskheta-mtianeti (1900s) (35–55) Mtskheta 9 Female IDPs residing in Samegrelo (1900s) (35–55) Zugdidi 10 Male IDPs residing in Samegrelo (1900s) (35–55) Zugdidi 11 Female IDPs residing in Shida Kartli (2008) (18–34) Gori 12 Male IDPs residing in Shida Kartli (2008) (18–34) Gori 13 Female ALPs residing in Shida Kartli (25–50) Gori 14 Female ALPs residing in Shida Kartli (25–50) Gori 15 Male ALPs residing in Shida Kartli (25–50) Gori 16 Male ALPs residing in Shida Kartli (25–50) Gori 17 Female ALPs residing in Samegrelo (35–55) Zugdidi 18 Male ALPs residing in Samegrelo (35–55) Zugdidi 19 Female ALPs residing in Samegrelo (18–34) Zugdidi 20 Male ALPs residing in Samegrelo (18–34) Zugdidi 21 Females in other urban settlements (18–35) Other urban Telavi 22 Males in other urban settlements (18–35) Other urban Telavi 23 Females in rural settlements (25–50) Rural (villages in Kakheti) 24 Males in rural settlements (25–50) Rural (villages in Kakheti) 25 Females in Tbilisi (25–50) Tbilisi 26 Males in Tbilisi (25–50) Tbilisi Third Round 27 Practitioners from the shelters Tbilisi ANNEXES 99 28 Practitioners from the shelters Gori 29 Practitioners with no specific GBV training Tbilisi 30 Female LGBT people (25–50) Tbilisi 31 Male LGBT people (25–50) Tbilisi 32 Female LGBT people (25–50) Kutaisi 33 Male LGBT people (25–50) Kutaisi 34 Female LGBT people (25–50) Batumi 35 Male LGBT people (25–50) Batumi Note: LGBT = Lesbian, Gay, Bisexual, and Transgender. A1.1.3 Forms of Violence Included in the Survey In this report, we look at four forms of violence, henceforth referred to as violence or as FFV. These include controlling behavior/emotional abuse, sexual harassment, sexual violence, and physical violence. For the analysis of the population group’s general experience of violence, this report does not distinguish between partner or non-partner violence, though relevant information pertaining to either partner/non-partner violence is pointed out where this is known and relevant. FFV is typically in- terpersonal violence, or violence perpetrated by one person against another, but it can include more than one perpetrator. Each of the four forms of violence, as used in this survey, are defined in turn, and the specific indicators or questions that were used to measure each form of violence are included in Box 1A.1. A1.1.4 Structure of the Survey and of the Violence Module Given the focus of this survey on exploring the links between GBV, conflict, economic oppor- tunities, and services, the questionnaire modules collected data in these respective areas, in addition to pertinent background information. The modules and themes are detailed in Table 1A.7. Table 1A.7: Survey Modules and Themes Module Themes Covered by the Module 1. Conflict status IDP affected by conflict in 1990s or 2008; ALP affected by conflict in 1990s or 2008; NAP not affected by conflict 2. Basic household data Household members and relationship between them; educational levels; marital status; ethnicity; disability status 3. Welfare Dwelling; land; livestock; durable goods; income; social assistance 4. Employment, skills, and livelihoods Employment status, sector and characteristics (before and after conflict); unemployment; financial autonomy; jobs training 5. Gender norms and attitudes Gender norms; norms and violence; violence and the law 6. Violence Controlling behavior /emotional abuse; sexual harassment; sexual violence; physical violence; most serious partner incident; most serious non-partner incident; health consequences; report- ing of violence 7. Conflict and displacement Pre- and post-conflict situation; perception of GBV and conflict; access to services 8. Services Awareness of GBV campaigns; access to services; female police officers Source: World Bank team. 100 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Box 1A.1: Violence Indicators by Category of Violence Used in This Study Controlling behavior/emotional abuse is threats or acts committed by a person to exert or maintain control over another person. These are psychological in nature and refer to the infliction of mental or emotional pain or injury such as threats or humiliation (adapted from IASC 2015). The questions included in this survey are: 1. **Insulted you or made you feel bad about yourself? 2. Belittled or humiliated you in front of other people? 3. Threatened to hurt you or someone you care about? 4. **Harassed you by repeatedly sending you messages or calling you? Sexual harassment includes unwelcome sexual advances and other verbal or physical conduct of a sexual nature (IASC 2015, from US Department of State, n.d.). The questions included in this survey are: 1. Touched, hugged, or kissed you against your will? 2. Stared or leered at you inappropriately so that it made you feel intimidated? 3. Made sexually suggestive comments or jokes that made you feel offended? 4. Sent or showed you sexually suggestive photos or pictures, e-mails, or texts? 5. Exposed themselves to you indecently? Sexual violence refers to any sexual act committed against the person’s will (IASC 2015, from WHO 2002). Specific forms of sexual violence such as forced pregnancy, sexual slavery, sexual trafficking, or sexual exploitation are not part of this survey. The questions included in this survey are: 1. Physically forced you to have sexual intercourse when you did not want to? 2. Had sexual intercourse with you when you were unable to refuse (for example, too drunk)? 3. Had sexual intercourse with you against your will, because you were afraid of what they might do? 4. Forced you to do something sexual that you found degrading or humiliating? Physical violence refers to a physical assault that is not sexual in nature. This includes hitting, slapping, choking, cutting, shoving, burning, and using weapons against a person (IASC 2015, from GBVIMS 2010.) The questions included in this survey are: 1. Slapped you or thrown something at you that could hurt you? 2. Dragged, pushed, or shoved you? 3. Kicked you or hit you with their fist or with something else that could hurt you? 4. Choked or burnt you on purpose? 5. Threatened to use or actually used a gun, knife, or other weapon against you? Source: World Bank team. Note: Indicators with ** in the controlling behavior/mental abuse section were not included in the violence index that served for the main analysis and regressions in this report. This was because one of the indicators (Insulted you or made you feel bad about yourself?) had results that were atypical, that is, markedly higher than the other indicators and that skewed the reported violence upwards. Conversely, the indicator ‘Harass you by repeatedly sending you messages or calling you?’ was also excluded because it had results that were markedly lower than the other indicators and that consequently added little value to the measure. ANNEXES 101 A1.2 ETHICAL CONSIDERATIONS AND ETHICS REVIEW Answering questions on GBV and conflict experiences can be traumatic. Sharing such experienc- es with strangers can be difficult and potentially cause problems with partners who may prefer that information stays private. As such, this research followed WHO Guidelines (Ellsberg and Heise 2005) for researching violence against women to ensure the safety and well-being of respondents and re- searchers and undertook several activities to sensitize the data collection mechanisms. For example, previously, it was described how only a single gender was interviewed in each survey site and how LGBT FGDs were organized by trusted LGBT nongovernmental organizations. In addition to those techniques, the methodology included a gender sensitivity training and underwent an independent ethics review. Each is described below: ƒƒ A gender sensitivity training was provided to all fieldwork team members. This included re- cruitment specialists, survey interviewers, KII specialists, FGD moderators, and regional fieldwork supervisors. The goal of the training was for participants to acknowledge and break down their own gender biases and understand the safety and ethical issues around conducting GBV research. The training included teaching participants about GBV and available services for GBV survivors, all in the Georgian context. Each participant was provided a referral document to direct possible survivors to support services. The gender sensitivity training was carried out in three training ses- sions for three different groups of participants. The training was informed by several well-known training guidelines (Ellsberg and Heise 2005; Jewkes, Dartnall, and Sikweyiya 2012; WHO 2001). After the fieldwork, researchers were debriefed about their experiences and provided counseling as necessary. The three-day gender sensitivity training was in addition to a two-day training on the study’s methodology. ƒƒ The study’s methodology passed an independent ethics review by an Institutional Review Board (IRB) managed by the Health Research Union in Georgia. The purpose of an IRB is to protect the rights and welfare of research subjects. An ethics review approves, monitors, and re- views behavioral research involving humans. This includes a risk-benefit analysis. The study’s ap- proval by the IRB independently verified that appropriate steps were taken to protect the rights and welfare of participants in this research. A1.3 UNIQUENESS AND LIMITATIONS OF METHODOLOGY This study’s methodology was designed to complement existing GBV research gaps. The survey assesses respondents’ experiences with different forms of violence, including controlling behavior, sexual harassment, sexual violence, and physical violence. Previous studies have rarely included analy- ses of harassment as part of violence assessments,2 but harassment can be widespread and meaning- ful to measure. Our survey also focuses on economic outcomes and experiences from Georgian con- flicts. This approach adds to a limited literature base that examines how conflict correlates to different forms of violence. In current literature, many studies ask survey questions about conflict and violence experiences but most draw on small samples with idiosyncratic populations such as refugees (Saile et al., 2013; Falb et al. 2013a, 2013b; Gupta et al. 2012). Moreover, previous studies often focus only on women and emphasize intimate partner violence (Saile et al. 2013; Falb et al. 2013a; Gupta et al. 2012; Vinck and Pham 2013) rather than assessing multiple forms of violence like this study. However, some studies also look at physical violence outside of the home (Falb et al. 2013b; Hossain et al. 2014). This study represents a contribution to the existing literature by drawing on nationally representative samples, with both male and female respondents, that report conflict and GBV experiences. 2 This is generally true, though some notable studies have included harassment and controlling behavior in their measures of GBV. See FRA (2014). 102 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES This survey draws from other violence-related survey questionnaires but may not be directly comparable. It is similar because it asks about respondents’ experiences with specific types of behav- iors. Their responses are then used to create measures for the FFV measured. It is important to note, however, that the current survey asked these questions of all respondents— men and women—and did not limit the violence modules to ever-partnered women—a common approach in some violence-fo- cused surveys. Therefore, these results may not be directly comparable to other nationally representa- tive surveys. The study also carries with it a few limitations. First, the sensitive nature of some questions could lead to underreporting of certain experiences. Certain gender-inequitable attitudes may be underre- ported because they are socially undesirable. Certain groups, such as men or those living in very rural areas, may have been more likely than other groups to underreport experiences like sexual violence because of the highly stigmatizing nature of these experiences. Similarly, women currently in a rela- tionship may prefer not to disclose experiences that have happened in the context of that relationship to researchers, despite the confidential nature of those exchanges. Second, this research presents a snapshot in time and does not allow the researchers to compare current rates of violence to histori- cal rates. We simply ask about former levels of violence and well-being, but this is still a measure of the respondent’s current thinking, which includes memory biases. In the future, a longitudinal study following subjects over time could track how changes in economic status, exposure to services, and awareness of GBV campaigns are associated with changes in reported rates of violence. Annex 2: Organizations Addressing GBV in Georgia3 A2.1 GOVERNMENT AGENCIES3 In Georgia, eight government agencies address GBV. These include five ministries: the Ministry of Justice; the Ministry of Internal affairs; the Ministry of Education and Science; the Ministry of Labor, Health, and Social Affairs; and the MRA. It also includes the courts and criminal justice system, the Pub- lic Defender’s Office, and the Interagency Council for the Prevention of Domestic Violence. Besides these, local governments also have important roles to play. ƒƒ Ministry of Justice. This ministry is responsible for ‘strengthening law and order’ and ‘developing and improving national legislation’ to make it consistent with international standards. Specifically, it worked to align Georgian legislation with the Istanbul Convention. The ministry also prosecutes criminal offenses of GBV. Recently, the main prosecutor’s office organized GBV training for pros- ecutors and aims to establish a special GBV-focused group of prosecutors this year (2017). The Ministry of Justice Department of Public International Law is responsible for GBV issues. It has no specific gender staff but defines ‘gender equality and prevention of violence against women’ as one of its action fields. ƒƒ Ministry of Internal Affairs. This ministry includes the national and district-level police, who op- erate as GBV first responders. Officers can issue restrictive orders against abusers. The Ministry of Internal Affairs employees are regularly trained in GBV issues like gender sensitivity, prevention mechanisms, and legal rights for victims and abusers. The ministry maintains a website with infor- mation on domestic violence, including a domestic violence hotline, statistical data, and consultan- cy support.4 Despite these efforts, more can be done to prepare the police for dealing with GBV. A recent study says that Georgian police still align with national gender biases. These biases suggest GBV is a family matter and police should not interfere (Mushkudiani 2016). There is also evidence that many police are unable to properly identify and provide support in cases on GBV. ƒƒ Ministry of Education and Science. This ministry is responsible for integrating GBV into educa- tional curriculums at various levels. The ministry does not have a designated specialist on GBV. ƒƒ Ministry of Labor, Health, and Social Affairs of Georgia. This ministry’s main function in relation to GBV is to monitor two agencies that work extensively on GBV. These include the Social Service Agency and ATIPFUND. Healthcare providers also operate under this ministry. • ATIPFUND. Founded in 2006, ATIPFUND became the main governmental service provider for GBV survivors (Javakhishvili and Tsuladze 2011). ATIPFUND runs four shelters for GBV survivors in Tbilisi, Gori, Signagi, and Kutaisi and offers support in approximately 70 locations. According 3 This Annex is based on a Stakeholder analysis prepared by the Analysis and Consulting Team (ACT), 2016 as a background note to the Population’s Life Experiences in Georgia Survey. 4 Ministry of Internal Affairs. “Domestic Violence.” http://police.ge/en/projects/domestic-violence. 103 104 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES to ATIPFUND Georgia (2016), the total number of beneficiaries of the shelter system since 2013 (until November 2016) include 17 minors, 244 adult survivors, and 367 dependents. ATIPFUND established its first crisis shelter last year and plans to open more. It also created a free 24/7 na- tional hotline to report GBV or seek counseling. There have been about 2,500 calls since 2015.5 • Social Service Agency. Founded in 2007, its objective is to provide social services. In 2010, it was tasked with protecting children from a wide range of social problems, including GBV. Starting in July 2016, the agency’s social workers became responsible for other GBV survivors as well, dramatically increasing the agency’s workload. However, according to KIIs and a 2015 study by Public Defender’s Office of Georgia (2015), the agency lacks the human capacity to cover such a broad range of problems. Few of the 240 employees at the agency have been properly trained in GBV or even social work. • Educational and health care institutions. These institutions are required to inform law en- forcement bodies if they see evidence of GBV and inform possible survivors about GBV ser- vices like shelters and the domestic violence hotline (UNFPA 2010). ƒƒ MRA. The MRA is mainly tasked with supporting the country’s IDP population. Currently, the min- istry provides a small stable allowance of GEL 45 per month to registered IDPs, but is considering replacing this program with a ‘means-based’ program. Recently with the support of UN Women, the MRA hired a gender specialist to oversee the integration of gender perspectives in IDP actions. ƒƒ The courts. The courts and the criminal justice system issue protective orders against abusers and define the terms of their validity. ƒƒ Public Defender’s Office. This office houses the Department of Gender Equality, which is tasked with the “examination and response to violation of rights on the basis of gender, including gender identity and sexual orientation.” The department monitors government services for GBV sensitivity and produces independent reports with GBV-related recommendations for other state agencies.6 ƒƒ The Interagency Council for the Prevention of Domestic Violence. Established in 2008 by pres- idential decree, this council is tasked with monitoring the various government efforts to combat GBV. However, it was discontinued after only a few months. In 2014, it was reestablished with a stronger mandate. It now manages the Group to Grant Domestic Violence Victim Status and moni- tors whether government agencies are fulfilling their National Action Plans related to GBV. ƒƒ Local governments. Local governments have the freedom to develop their own social programs. As a result, there are too many local initiatives to list here. That said, some initiatives provide finan- cial support or shelter to GBV survivors. For example, there is a special budget line in Zugdidi’s municipal budget allocated for GBV survivors. This includes GEL 200 per month for lodging and a one-time assistance payment of GEL 500. Unfortunately, these initiatives are not systematic, and in general, local budgets do not take gender issues into consideration (UN Women 2016). A2.2 NONGOVERNMENTAL ORGANIZATIONS At least seven nongovernmental organizations aim to reduce GBV, with another ten that address related challenges. Of the seven, three nongovernmental organizations operate nationwide. These include the Women’s Information Center, the Anti-Violence Network of Georgia, and the Georgian Young Lawyer’s Association. The other four only operate in Tbilisi and other cities. These cities include 5 ATIPFUND. Statistics. http://atipfund.gov.ge/geo. 6 Public defender of Georgia, “About the Department of Gender Equality.” http://www.ombudsman.ge/en/specializirebuli-centrebi/ genderuli-tanasworoba/genderuli-tanasworobis-centris-shesaxeb ANNEXES 105 Sakhli, Sapari, Amagdari, and Atinati. These organizations provide a wide array of survivor services. These include legal aid, crisis shel- ters, social counseling, job market access, rehabilitation, training of state agencies in GBV, the profes- sional development of social workers, and help to survivors for attaining proper documentation to receive state support. Most of these nongovernmental organizations provide support to other vulnera- ble populations as well, including drug addicts, people with disabilities, and LGBT people, among oth- ers. These organizations are funded by international donors. As such, services are typically available to survivors for free (Javakhishvili and Tsuladze 2011). A2.3 INTERNATIONAL ORGANIZATIONS Several international agencies actively support efforts to reduce GBV in Georgia. These include the UN, the World Bank, and several bilateral donors, among others. ƒƒ UN agencies. The three UN agencies—UN Women, UNDP, and UNFPA—work on a joint program to enhance gender equality in Georgia. All three have long-standing gender portfolios in Georgia, providing policy and governance support and coordinating GBV awareness-raising campaigns. In many ways, the UN agencies lead the way on GBV issues among the international organizations in Georgia. ƒƒ The World Bank Group. GBV is a relatively new focus area for the World Bank. GBV-related work focuses support programs on three areas: (a) reduce domestic violence, (b) improve the safety of women in the workplace and public transport systems, and (c) develop health and livelihood approaches for women at risk of conflict displacement. This research is the World Bank’s first GBV engagement in Georgia. ƒƒ Bilateral aid agencies. USAID has allocated over US$1.5 billion to assist Georgia in a wide range of programs. Currently, USAID works with ATIPFUND on the National Referral Mechanism. Among a larger work program, the Swedish International Development Cooperation Agency (SIDA) also helps local organizations fight GBV and gender inequality. SIDA cooperates with UN Women to support domestic violence shelters. As an aside, the Swedish nongovernmental organization, Kvin- na Till Kvinna works in Georgia on women’s rights and sexual or reproductive health, among other efforts. Annex 3: Quantitative Analysis A3.1 MEASURE OF VULNERABILITY TO VIOLENCE We define seven different measures of violence: For Both Men and Women: Physical violence is defined as the number of the statements in this category for which the respon- dents said it ever happened to them. The indicator ranges from 0 to 5. Sexual violence is defined as the number of the statements in this category for which the respondents said it ever happened to them. The indicator ranges from 0 to 4. Sexual harassment is defined as the number of the statements in this category for which the respon- dents said it ever happened to them. The indicator ranges from 0 to 5. Controlling behavior is defined as the number of the statements in this category for which the re- spondents said it ever happened to them. The indicator ranges from 0 to 2. Total violence is defined as the total number of physical violence, sexual violence, sexual harassment, and controlling behaviors reported by the respondents. The indicator ranges from 0 to 14. Sexual abuse is defined as the total number of sexual violence and sexual harassment reported by the respondents. The indicator ranges from 0 to 9. Only for Women GBV is defined as the total number of physical violence and sexual violence. The indicator ranges from 0 to 7. A3.2 BASIC MODEL In this section, we present the results of the analysis whose aim is to identify the main factors that can be associated with a higher level of reported violence. We estimate a basic model, which includes fundamental predictors of number of violence. Subse- quently, we add further control sets to identify the effect of each sets of variables. The explanatory variables included in the basic model are presented in Table 3A.1. From Table 3A.2 to Table 3A.4, we present the result of a linear regression for the total sample, for women only, and men only where the dependent variable is, respectively, the total violence, GBV, and sexual abuse. 106 ANNEXES 107 Table 3A.1: Explanatory Variables in the Basic Model Age In log Gender 1. Female 2. Male Level of education 1. Lower secondary 2. Upper secondary/vocational 3. Higher education/bachelor’s degree Socio-demographic variables 4. Master’s/doctorate degree Marital status 1. Single/widower 2. Married/cohabiting 3. Separated/divorced Household size Location 1. Tbilisi Geographical variables 2. Other urban areas 3. Rural areas 1. National sample 2. IDPs in collective center Conflict status 3. IDPs not in collective center 4. ALPs Before the conflict 1. Completely or partly stable 2. Not stable 3. Unemployed 4. Out of labor force Employment status Currently 1. Completely or partly stable 2. Not stable 3. Unemployed 4. Out of labor force Bottom 40% Housing 1. Adequate Economic status 2. Marginal 3. Inadequate No asset Table 3A.2: Predictors of Number of Total Violence - Basic Model Coefficient/(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center −0.278*** −0.230* −0.319** (0.102) (0.138) (0.145) IDP not in collective center −0.378** −0.499** −0.275 (0.150) (0.222) (0.190) 108 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES ALP −0.004 0.06 0.028 (0.117) (0.135) (0.174) Demographic variables Age (ln) −0.264** −0.449*** −0.044 (0.116) (0.164) (0.166) Education Ref: Lower secondary Upper secondary/vocational −0.454** −0.624** −0.063 (0.208) (0.278) (0.221) Higher vocational/bachelor’s degree −0.565** −0.716** −0.150 (0.236) (0.310) (0.286) Master’s or doctorate degree −0.563** −0.717** −0.126 (0.253) (0.334) (0.278) Gender Ref: Female Male −0.012 (dropped) (dropped) (0.114) Marital status Ref: Single/widower Married/cohabiting −0.105 −0.152 −0.082 (0.091) (0.121) (0.145) Separated/divorced 0.730** 1.300*** −0.323* (0.294) (0.437) (0.166) Household size −0.032 −0.071* 0.004 (0.027) (0.037) (0.041) Location Ref: Tbilisi Other urban −0.535*** −0.728*** −0.225 (0.169) (0.220) (0.236) Rural −0.721*** −0.839*** −0.524** (0.177) (0.240) (0.236) Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable −0.126 0.081 −0.239* (0.123) (0.239) (0.123) ANNEXES 109 Unemployed 0.306* 0.298 0.258 (0.185) (0.287) (0.227) Out of labor force −0.042 −0.039 −0.005 (0.115) (0.158) (0.172) Employment status before the conflict Ref: Employed Unemployed −0.420*** −0.650*** −0.194 (0.155) (0.208) (0.213) Out of labor force 0.013 −0.046 0.127 (0.096) (0.127) (0.156) Economic status Bottom 40% −0.005 0.062 −0.096 (0.116) (0.159) (0.172) Housing Ref: Adequate Marginal 0.072 0.18 −0.037 (0.104) (0.150) (0.136) Inadequate 0.1 0.1 0.178 (0.139) (0.153) (0.245) No asset 0.033 0.012 0.109 (0.087) (0.114) (0.137) Number of observations 2,985 1,504 1,481 R-squared 0.075 0.128 0.055 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Total Violence is defined as the total number of physical violence, sexual violence, sexual harassment and controlling behavior reported by respondents (ranging from 0 to 14). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs, 2) IDPs living in collective centers, 3) IDPs not living in collective centers, and 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Marital status a categorical variable defined as: 1) single or widower 2) Married or cohabiting 3) Separated/divorced. e) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the household by room (excluding kitchen, hallways, stor- age); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indicator their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/other valuables, cash savings. All the regressions are based on the full sample (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. 110 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Table 3A.3: Predictors of Number of GBV - Basic Model Coefficient /(Standard Error) Female Conflict status Ref: National sampling IDP in collective center −0.050 (0.070) IDP not in collective center −0.220** (0.098) ALP 0.040 (0.088) Demographic variables Age (ln) −0.060 (0.084) Education Ref: Lower secondary Upper secondary/vocational −0.358* (0.192) Higher vocational/bachelor’s degree −0.339* (0.197) Master’s or doctorate degree −0.416* (0.214) Marital status Ref: Single/widower Married/cohabiting 0.040 (0.070) Separated/divorced 0.763** (0.307) Household size −0.043* (0.022) Location Ref: Tbilisi Other urban −0.269** (0.111) Rural −0.358*** (0.115) ANNEXES 111 Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable 0.089 (0.145) Unemployed 0.043 (0.131) Out of labor force −0.006 (0.090) Employment status before the conflict Ref: Employed Unemployed −0.263*** (0.098) Out of labor force −0.054 (0.082) Economic status Bottom 40% 0.127 (0.104) Housing Ref: Adequate Marginal 0.109 (0.086) Inadequate −0.024 (0.092) No asset −0.016 (0.069) Number of observations 1,480 R-squared 0.081 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable GBV is defined as the total number of physical violence and sexual violence reported by respondents (ranging from 0 to 9). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Marital status a categorical variable defined as: 1) single or widower 2) Married or cohabiting 3) Separated/divorced. f) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if 112 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indica- tor their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/ other valuables, cash savings. All the regressions are based on the full sample (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. Table 3A.4: Predictors of Number of Sexual Abuse - Basic Model Coefficient /(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center −0.052 −0.040 −0.051 (0.053) (0.076) (0.063) IDP not in collective center −0.170** −0.253** −0.105 (0.080) (0.121) (0.092) ALP −0.015 −0.015 0.006 (0.043) (0.051) (0.061) Demographic variables Age (ln) −0.080 −0.199** 0.065 (0.052) (0.083) (0.069) Education Ref: Lower secondary Upper secondary/vocational −0.146** −0.246* 0.073 (0.015) (0.135) (0.052) Higher vocational/bachelor’s degree −0.186* −0.273** 0.026 (0.023) (0.082) (0.067) Master’s or doctorate degree Marital status −0.113** (dropped) (dropped) Ref: Single/widower Married/cohabiting −0.065** −0.132** −0.025 (0.030) (0.065) (0.068) Separated/divorced 0.119* 0.211 −0.161* (0.063) (0.188) (0.083) Household size 0.002 −0.004 −0.001 (0.011) (0.017) (0.013) ANNEXES 113 Location Ref: Tbilisi Other urban −0.185** −0.325*** 0.003 (0.085) (0.120) (0.100) Rural −0.247*** −0.350*** −0.101 (0.089) (0.130) (0.017) Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable −0.093** −0.086 −0.120** (0.045) (0.089) (0.048) Unemployed 0.174** 0.226 0.102 (0.082) (0.159) (0.071) Out of labor force −0.044 −0.034 −0.078 (0.049) (0.079) (0.054) Employment status before the conflict Ref: Employed Unemployed −0.186*** −0.308*** −0.084 (0.069) (0.106) (0.089) Out of labor force 0.000 0.015 0.000 (0.045) (0.062) (0.067) Economic status Bottom 40% −0.057 −0.045 −0.064 (0.050) (0.073) (0.065) Housing Ref: Adequate Marginal 0.073 0.145* −0.006 (0.049) (0.075) (0.053) Inadequate 0.090 0.128* 0.070 (0.067) (0.075) (0.112) No asset 0.011 −0.036 0.128** (0.040) (0.061) (0.055) Number of observations 2,960 1,497 1,463 R-squared 0.053 0.085 0.043 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Sexual Abuse 114 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES is defined as the total number of sexual violence and sexual harassment reported by respondents (ranging from0 to 7). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respon- dents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Marital status a categorical variable defined as: 1) single or widower 2) Married or cohabiting 3) Separated/divorced. f) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each in- dicator their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/ other valuables, cash savings. All the regressions are based on the full sample (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. Table 3A.5—3A.8 replicate the same regression presented in Tables 3.1–3.4 by type of reported vio- lence (and within each type again for the overall sample and by gender). Distinguishing by type of violence allows us to identify factors that can be correlated with the level of reported violence within each category of violence. Table 3A.5: Predictors of Number of Physical Violence Coefficient /(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center −0.086* −0.086 −0.084 (0.048) (0.054) (0.088) IDP not in collective center −0.128* −0.173** −0.056 (0.071) (0.071) (0.146) ALP 0.070 0.078 0.113 (0.069) (0.080) (0.112) Demographic variables Age (ln) −0.089 −0.094 −0.078 (0.065) (0.078) (0.111) Education Ref: Lower secondary Upper secondary/vocational −0.139** −0.184 ** −0.022 (0.093) (0.06) (0.127) Higher vocational/bachelor’s degree −0.16 *** −0.185 ** −0.041 (0.006) (0.090) (0.164) ANNEXES 115 Master’s or doctorate degree −0.229** −0.296** −0.047 (0.110) (0.132) (0.168) Gender Ref: Female Male 0.114** (dropped) (dropped) (0.054) Marital status Ref: Single/widower Married/cohabiting −0.024 0.024 −0.048 (0.046) (0.058) (0.076) Separated/divorced 0.361** 0.668*** −0.11 (0.154) (0.228) (0.098) Household size −0.019 −0.039** 0.007 (0.015) (0.019) (0.023) Location Ref: Tbilisi Other urban −0.174** −0.216** −0.095 (0.075) (0.089) (0.125) Rural −0.291*** −0.297*** −0.266** (0.078) (0.090) (0.128) Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable −0.037 0.14 −0.128* (0.067) (0.133) (0.065) Unemployed 0.094 0.042 0.117 (0.089) (0.128) (0.123) Out of labor force −0.007 −0.023 0.056 (0.060) (0.074) (0.100) Employment status before the conflict Ref: Employed Unemployed −0.150* −0.229** −0.058 (0.086) (0.088) (0.134) Out of labor force 0.006 −0.065 0.103 (0.054) (0.061) (0.098) 116 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Economic status Bottom 40% 0.089 0.115** 0.047 (0.065) −0.048 (0.105) Housing Ref: Adequate Marginal 0.022 0.047 0.006 (0.051) (0.059) (0.083) Inadequate −0.04 0.001 −0.053 (0.068) (0.082) (0.112) No asset 0.001 0.021 −0.047 (0.045) (0.052) (0.079) Number of observations 2,909 1,473 1,436 R-squared 0.0412 0.0827 0.0402 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Physical Vilence is defined as the total number of physical violence forms (1) slapped you or thrown something at you 2) dragged, pushed or shoved you 3) kicked you, hit you with their fist or with something else 4) chocked or burnt you on purpose 5) threatened to use or actually used a gun, knife orother weapon) reported by respondents.(ranging from 0 to 5). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Marital status a categorical variable defined as: 1) single or widower 2) Married or cohabiting 3) Separated/divorced. f) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the house- hold by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indicator their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/other valuables, cash savings. All the regressions are based on the full sample (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. Table 3A.6: Predictors of Number of Sexual Violence Coefficient /(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center 0.011 0.036 0.001 (0.019) (0.031) (0.017) IDP not in collective center −0.034 −0.048 −0.028* (0.023) (0.040) (0.015) ANNEXES 117 ALP −0.018 −0.036 0.005 (0.018) (0.025) (0.016) Demographic variables Age (ln) −0.081*** −0.095*** −0.043 (0.021) (0.026) (0.028) Education Ref: Lower secondary Upper secondary/vocational −0.115** −0.190* 0.020 (0.039) (0.100) (0.017) Higher vocational/bachelor’s degree −0.121* −0.167* −0.017 (0.067) (0.096) (0.021) Master’s or doctorate degree −0.098*** −0.236** 0.002 (0.006) (0.087) (0.024) Gender Ref: Female Male −0.035** (dropped) (dropped) (0.015) Marital status Ref: Single/widower Married/cohabiting −0.008 0.017 −0.014 (0.022) (0.031) (0.022) Separated/divorced 0.065** 0.107** 0.001 (0.029) (0.049) (0.047) Household size 0.0000 −0.0040 0.0030 (0.005) (0.008) (0.006) Location Ref: Tbilisi Other urban −0.110*** −0.053*** −0.046** (0.025) (0.011) (0.019) Rural −0.122*** −0.061*** 0.025 (0.027) (0.005) (0.020) Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable −0.048*** −0.041 −0.056*** (0.016) (0.029) (0.019) 118 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Unemployed −0.001 0.001 −0.012 (0.026) (0.043) (0.030) Out of labor force 0.001 0.015 −0.017 (0.019) (0.033) (0.018) Employment status before the conflict Ref: Employed Unemployed −0.004 −0.035 0.026 (0.031) (0.032) (0.046) Out of labor force 0.000 0.011 −0.024 (0.022) (0.032) (0.023) Economic status Bottom 40% −0.012 0.015 −0.038 (0.027) (0.043) (0.025) Housing Ref: Adequate Marginal 0.051** 0.063* 0.039* (0.023) (0.038) (0.020) Inadequate −0.015 −0.028 0.019 (0.019) (0.026) (0.026) No asset −0.021 −0.038 0.000 (0.020) (0.032) (0.011) Number of observations 2,842 1,443 1,399 R-squared 0.0398 0.0398 0.0395 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Sexual Violence is defined as the total number of sexual violence forms (1) physically forced you to have sex 2) had sexual intercourse with you when you were unable to refuse 3) had sexual intercourse with you against your will 4) forced you to do something sexual that you found humiliating) reported by respondents.(ranging from 0 to 4). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs leaving in collective centers 3) IDPs not leaving in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Marital status a categorical variable defined as: 1) single or widower 2) Married or cohabiting 3) Separated/divorced. f) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the house- hold by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indicator their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/other valuables, cash savings. All the regressions are based on the full sample (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. ANNEXES 119 Table 3A.7: Predictors of Number of Sexual Harassment Coefficient /(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center −0.06 −0.073 −0.049 (0.045) (0.064) (0.056) IDP not in collective center −0.135* −0.205* −0.078 (0.071) (0.106) (0.086) ALP 0.003 0.02 0.001 (0.038) (0.046) (0.054) Demographic variables Age (ln) −0.080* −0.233*** 0.105* (0.045) (0.076) (0.058) Education Ref: Lower secondary Upper secondary/vocational −0.040 −0.072 0.056 (0.055) (0.078) (0.052) Higher vocational/bachelor’s degree −0.074 −0.120 0.043 (0.063) (0.079) (0.069) Master’s or doctorate degree −0.061 −0.083 0.025 (0.084) (0.118) (0.060) Gender Ref: Female Male −0.079* (dropped) (dropped) (0.048) Marital status Ref: Single/widower Married/cohabiting −0.057 −0.148*** −0.014 (0.035) (0.052) (0.062) Separated/divorced 0.061 0.116 −0.165*** (0.084) (0.118) (0.058) Household size 0.002 0.000 −0.003 (0.008) (0.014) (0.010) Location Ref: Tbilisi 120 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Other urban −0.175** −0.271** −0.041 (0.076) (0.105) (0.092) Rural −0.225*** −0.289** −0.125 *** (0.078) (0.112) (0.030) Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable −0.042 −0.034 −0.065 (0.041) (0.081) (0.042) Unemployed 0.176** 0.228 0.117* (0.072) (0.141) (0.060) Out of labor force −0.045 −0.049 −0.059 (0.041) (0.062) (0.049) Employment status before the conflict Ref: Employed Unemployed −0.184*** −0.275*** −0.107 (0.058) (0.093) (0.074) Out of labor force 0 0.006 0.021 (0.035) (0.048) (0.056) Economic status Bottom 40% −0.044 −0.058 −0.029 (0.040) (0.058) (0.054) Housing Ref: Adequate Marginal 0.023 0.084 −0.043 (0.039) (0.055) (0.051) Inadequate 0.102 0.153** 0.051 (0.063) (0.069) (0.109) No asset 0.031 0.001 0.128*** (0.033) (0.051) (0.047) Number of observations 2,934 1,483 1,451 R-squared 0.0576 0.108 0.0478 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Sexual Harassment is defined as the total number of sexual harassment forms (1) touched, hugged or kissed you against your will 2) stared or leered at you inappropriately 3) made sexually suggestive comments 4) sent or showed you sexually suggestive photos 5) exposed themeselves to you indecently) reported by respondents.(ranging from 0 to 5). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ANNEXES 121 ALPs 2) IDPs leaving in collective centers 3) IDPs not leaving in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Marital status a categorical variable defined as: 1) single or widower 2) Married or cohabiting 3) Separated/divorced. f) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indicator their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/other valu- ables, cash savings. All the regressions are based on the full sample (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. Table 3A.8: Predictors of Number of Controlling Behavior Coefficient /(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center −0.139*** −0.103** −0.181*** (0.029) (0.040) (0.040) IDP not in collective center −0.083* −0.08 −0.105** (0.047) (0.073) (0.050) ALP −0.048 0.011 −0.075 (0.035) (0.047) (0.048) Demographic variables Age (ln) −0.091** −0.154*** −0.026 (0.040) (0.053) (0.059) Education Ref: Lower secondary Upper secondary/vocational −0.175*** −0.189** −0.123 (0.061) (0.077) (0.094) Higher vocational/bachelor’s degree −0.228*** −0.259*** −0.144 (0.076) (0.099) (0.110) Master’s or doctorate degree −0.187** −0.209** −0.11 (0.074) (0.090) (0.115) Gender Ref: Female Male −0.01 (dropped) (dropped) (0.033) 122 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Marital status Ref: Single/widower Married/cohabiting −0.04 −0.01 −0.04 (0.048) (0.047) (0.048) Separated/divorced 0.419*** −0.043 0.419*** (0.122) (0.086) (0.122) Household size −0.017* −0.030** −0.002 (0.009) (0.013) (0.013) Location Ref: Tbilisi Other urban −0.172*** −0.185*** −0.137* (0.049) (0.063) (0.078) Rural −0.190*** −0.201*** −0.163** (0.049) (0.063) (0.077) Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable 0.009 0.027 0.014 (0.043) (0.070) (0.052) Unemployed 0.055 0.069 0.04 (0.056) (0.085) (0.070) Out of labor force 0.011 0.014 0.023 (0.034) (0.046) (0.054) Employment status before the conflict Ref: Employed Unemployed −0.097** −0.138** −0.055 (0.045) (0.061) (0.064) Out of labor force 0.010 0.010 0.020 (0.030) (0.040) (0.050) Economic status Bottom 40% −0.031 −0.001 −0.077 (0.038) (0.052) (0.057) Housing Ref: Adequate Marginal −0.023 −0.006 −0.04 ANNEXES 123 (0.035) (0.052) (0.047) Inadequate 0.045 −0.030 0.154** (0.054) (0.066) (0.076) No asset 0.021 0.028 0.031 (0.029) (0.037) (0.049) Number of observations 2,982 1,502 1,480 R-squared 0.675 0.1184 0.0987 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Controlling Behavior is defined as the total number of sexual violence forms (1) belittled or humiliated in front of other peoplere 2) threatened to hurt you or someone you care about) ported by respondents.(ranging from 0 to 2). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs leaving in collective centers 3) IDPs not leaving in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Marital status a categorical variable defined as: 1) single or widower 2) Married or cohabiting 3) Separated/divorced. f) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the house- hold by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indicator their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/other valuables, cash savings. All the regressions are based on the full sample (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. A3.3 Adding Controls In this section, we present the results we obtained on adding further controls to the basic model. We add these sets of indicators one at a time. In Tables 3A.9 and 3A.10, we report only coefficients and standard errors for the added controls, as results do not change (if not marginally) from the ones presented in the basic model estimations. On the other hand, from Table 3A.11 to Table 3A.17 we present analysis limited to married or cohabiting respondents therefore we report the full regression findings as the basic model has been modified not including the explanatory variable marital status. A3.3.1 Gender Views and Beliefs We estimate a new version of the basic model (Table 3A.9) in which we add an indicators of gender norms and attitudes that can be correlated to the experience of GBV. The indicator Acceptability of vio- lence against women is equal to 1 if the respondent considers acceptable at least one of the following reasons for a man to hit his spouse/partner: ƒƒ She does not complete her household work to his satisfaction. ƒƒ She disobeys him. ƒƒ She refuses to have sexual relations with him. 124 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES ƒƒ She asks him whether he has other girlfriends. ƒƒ He finds out that she has been unfaithful A3.3.2 Controlling for Levels of Stress From the questionnaire, it is possible to get information about the level of stress of the respondents related to the work situation. In particular, we use two variables related to the Perception of Unemploy- ment (for all respondents). The variables are: ƒƒ Ashamed to face family: which is equal to 1 if the respondent admits to be ashamed to face his family because the respondent is out of work or does not make enough money, 0 otherwise ƒƒ Sometimes drink or stay away: which is equal to 1 if the respondents admits to drink sometimes or to stay away from home when he cannot find a job As each one of these variables have a share of missing value higher than 9 percent, we include the category ‘missing’. Therefore, each variable is 0 if no, 1 if yes, and missing if the respondent does not know or rather not say. Table 3A.10 presents the results for this set of added controls. Table 3A.9: Predictors of Violence Number Including Gender Views Variables   Coefficient /(Standard Error) Total Violence GBV Sexual Abuse   Total Female Male Female Total Female Male Acceptability of violence against women 0.257** 0.357** 0.11 0.216** 0.199*** 0.221*** 0.164** (0.109) (0.156) (0.145) (0.095) (0.058) (0.082) (0.077) Number of observations 2968 1499 1469 1475 2944 1492 1452 R-squared 0.079 0.136 0.055 0.088 0.064 0.097 0.059   Coefficient /(Standard Error) Physical Violence Sexual Violence   Total Female Male Total Female Male Acceptability of violence against women 0.014 0.07 -0.067 0.087*** 0.150*** 0.015 (0.047) (0.064) (0.069) (0.029) (0.049) (0.020) Number of observations 2893 1469 1424 2827 1438 1389 R-squared 0.0481 0.0959 0.0449 0.0373 0.0628 0.0413   Coefficient /(Standard Error) Sexual Harassment Controlling Behavior   Total Female Male Total Female Male Acceptability of violence against women 0.116** 0.075 0.152** 0.045 0.066 0.015 (0.047) (0.059) (0.070) (0.033) (0.048) (0.042) Number of observations 2,919 1,479 1,440 2965 1497 1468 R-squared 0.061 0.111 0.051 0.0688 0.1161 0.0488 ANNEXES 125 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variables are Total Violence, GBV, Sexual Abuse, Physical Violence, Sexual Violence, Sexual Harassment, Controlling Behavior. The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Marital status a categorical variable defined as: 1) single or widower 2) Married or cohabiting 3) Separated/divorced. f) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indica- tor their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/ other valuables, cash savings. We add to the basic model the explanatory variable Acceptability of violence against women which is equal to 1 if the respondent considers acceptable at least one of the following reasons for a man to hit his spouse: 1) she does not complete her chores 2) she disobeys him 3) she asks him if he has other girlfriends 4) he finds out that she has been unfaithful. All the regressions are based on the full sample (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. Table 3A.10: Predictors of Violence Number Including Indicators of Economic Stress   Coefficient /(Standard Error) Total Violence GBV Sexual Abuse   Total Female Male Female Total Female Male Ashamed to face family Agree 0.366*** 0.323** 0.426** 0.112 0.076 0.07 0.093 (0.110) (0.137) (0.169) (0.089) (0.047) (0.061) (0.070) Missing -0.321** -0.254 -0.281 -0.054 -0.168* -0.270* -0.017 (0.154) (0.186) (0.213) (0.101) (0.090) (0.137) (0.087) Sometimes I drink or stay away Agree 0.211 0.469 0.239 0.046 0.066 0.228 0.059 (0.262) (0.466) (0.293) (0.247) (0.096) (0.219) (0.095) Missing -0.103 -0.129 -0.028 -0.159 0.018 0.048 -0.02   (0.160) (0.199) (0.225) (0.101) (0.091) (0.133) (0.072) Number of observations 2985 1504 1481 1,480 2960 1497 1463 R-squared 0.089 0.136 0.079 0.085 0.056 0.090 0.049 126 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES   Coefficient /(Standard Error) Physical Violence Sexual Violence   Total Female Male Total Female Male Ashamed to face family Agree 0.128** 0.086 0.169* 0.042* 0.028 0.063** (0.058) (0.071) (0.086) (0.023) (0.038) (0.028) Missing -0.092 -0.014 -0.131 0.014 -0.031 0.05 (0.074) (0.100) (0.103) (0.039) (0.037) (0.060) Sometimes I drink or stay away Agree -0.028 -0.202* 0.029 0.023 0.242 0.002 (0.128) (0.107) (0.142) (0.047) (0.178) (0.038) Missing -0.112 -0.128 -0.065 -0.038 -0.036 -0.014   (0.072) (0.095) (0.114) (0.032) (0.045) (0.038) Number of observations 2909 1473 1463 2842 1443 1399 R-squared 0.057 0.099 0.057 0.028 0.044 0.061   Coefficient /(Standard Error) Sexual Harassment Controlling Behavior   Total Female Male Total Female Male Ashamed to face family Agree 0.035 0.044 0.033 0.165*** 0.167*** 0.170*** (0.036) (0.044) (0.056) (0.039) (0.048) (0.061) Missing -0.178** -0.241* -0.067 -0.059 0.043 -0.135* (0.074) (0.131) (0.054) (0.062) (0.077) (0.078) Sometimes I drink or stay away Agree 0.041 -0.019 0.058 0.164 0.444 0.146 (0.076) (0.062) (0.077) (0.114) (0.338) (0.124) Missing 0.054 0.082 -0.007 -0.016 -0.061 0.05   (0.078) (0.122) (0.052) (0.063) (0.075) (0.089) Number of observations 2,934 1,483 1,451 2982 1502 1480 R-squared 0.0576 0.108 0.0478 0.090 0.131 0.082 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variables are Total Violence, GBV, Sexual Abuse, Physical Violence, Sexual Violence, Sexual Harassment, Controlling Behavior. The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Marital status a categorical variable defined as: 1) single or widower 2) Married or cohabiting 3) Separated/divorced. f) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing ANNEXES 127 in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indica- tors that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indicator their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following as- sets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/other valuables, cash savings. We add to the basic model two explanatory variables: Ashamed to face family: which is equal to 1 if the the respondent admits to be ashamed to face his family because the respondent is out of work or does not make enough money, 0 if not and missing if he doesn’t know or rather not say; Sometimes I drink or stay away: which is equal to 1 if the respondents admit to drink sometimes or to stay away from home when he cannot find a job, and missing if he doesn’t know or rather not say. All the regressions are based on the full sample (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. A3.4 ADDING CONTROLS AVAILABLE ONLY FOR NON-SINGLE RESPONDENTS In this section, we present some estimations obtained including in the basic model further variables which are available from the questionnaire only for subgroups of the sample. A3.4.1 Controlling for Partner’s Issues (Married or Cohabiting Respondents) In the questionnaire, respondents who are married or cohabiting (1,901 observations) are asked two questions about the partner’s level of stress related to possible unemployment (question 4.21): ƒƒ Does the partner (fully or partly) spend most of her/his time out of work or looking for a work? ƒƒ Does the partner (fully or partly) sometimes drink or stay away from home when he/she cannot find work? We combine the two variables creating an indicator of ‘partner issue’ that is equal to 1 if respondents answer yes to at least one of the two previous questions (from Table 3A.11 to Table 3A.17). Table 3A.11: Predictors of Number of Total Violence - Including Indicators of Partner Issues - Only Married or Cohabiting Respondents Coefficient /(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center -0.296*** -0.216* -0.376** (0.097) (0.122) (0.167) IDP not in collective center -0.430*** -0.566*** -0.306 (0.142) (0.184) (0.230) ALP 0.196 0.193* 0.2 (0.123) (0.104) (0.235) 128 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Demographic variables Age (ln) -0.192 -0.243 -0.239 (0.141) (0.208) (0.218) Education Ref: Lower secondary Upper secondary/vocational -0.062 0.218 -0.35 (0.163) (0.171) (0.284) Higher vocational/bachelor’s degree -0.295 0.064 -0.700** (0.192) (0.218) (0.341) Master’s or doctorate degree -0.342* -0.147 -0.537 (0.190) (0.213) (0.331) Gender Ref: Female Male 0.155 (dropped) (dropped) (0.118) Household size -0.023 -0.048 0.007 (0.031) (0.038) (0.050) Location Ref: Tbilisi Other urban -0.628*** -0.651*** -0.553** (0.172) (0.225) (0.251) Rural -0.828*** -0.835*** -0.790*** (0.178) (0.220) (0.275) Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable -0.335*** -0.425** -0.243* (0.106) (0.171) (0.136) Unemployed 0.147 0.14 0.132 (0.185) (0.281) (0.242) Out of labor force -0.112 -0.192 -0.009 (0.124) (0.184) (0.159) Employment status before the conflict Ref: Employed Unemployed -0.259 -0.565** -0.03 ANNEXES 129 (0.167) (0.222) (0.235) Out of labor force 0.031 -0.044 0.109 (0.115) (0.165) (0.150) Economic status Bottom 40% -0.073 -0.119 -0.078 (0.109) (0.133) (0.188) Housing Ref: Adequate Marginal -0.076 0.018 -0.151 (0.097) (0.156) (0.103) Inadequate 0.036 -0.058 0.167 (0.142) (0.163) (0.223) No asset 0.084 0.09 0.1 (0.097) (0.146) (0.138) Partner issue ref: no issue partner has at least one issue 0.165 0.118 0.265 (0.104) (0.131) (0.170) Number of observations 1873 889 984 R-squared 0.087 0.092 0.094 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Total Violence is defined as the total number of physical violence, sexual violence, sexual harassment and controlling behavior reported by respondents (ranging from 0 to 14). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variable defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or doctoral degree. d) Gender: male or female. f) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is de- fined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribu- tion, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water net- works; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indicator their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/other valuables, cash savings. We add to the basic model (which here exclude the variable marital status) the variable Partner’s issues which is equal to 1 if the respondent answers yes to at least one of the following questions: 1) Does your partner (fully or partly) spend most of her/his time out of work or looking for a work? 2) Does the partner (fully or partly) sometimes drink or stay away from home when he/she cannot find work?, and 0 otherwise. All the regressions are based on the sample of married or cohabiting respondents (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. 130 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Table 3A.12: Predictors of Number of GBV Including Indicators of Partner Issues - Only Married or Cohabiting Respondents Coefficient /(Standard Error) Female Conflict status Ref: National sampling IDP in collective center -0.119* (0.064) IDP not in collective center -0.295*** (0.084) ALP 0.061 (0.068) Demographic variables Age (ln) -0.132 (0.118) Education Ref: Lower secondary Upper secondary/vocational 0.174 (0.122) Higher vocational/bachelor’s degree 0.174 (0.118) Master’s or doctorate degree -0.004 (0.137) Household size -0.041 (0.025) Location Ref: Tbilisi Other urban -0.208* (0.125) Rural -0.319*** (0.120) Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable -0.214** (0.084) ANNEXES 131 Unemployed 0.064 (0.164) Out of labor force -0.11 (0.101) Employment status before the conflict Ref: Employed Unemployed -0.279** (0.122) Out of labor force -0.002 (0.091) Economic status Bottom 40% -0.009 (0.091) Marginal -0.009 (0.091) Inadequate 0.048 (0.116) No asset 0.091 (0.089) Partner issue ref: no issue partner has at least one issue 0.02 (0.091) Number of observations 876 R-squared 0.059 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable GBV is defined as the total number of physical violence and sexual violence reported by respondents (ranging from 0 to 9). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indica- tor their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/ other valuables, cash savings. We add to the basic model (which here exclude the variable marital status) the variable Part- ner’s issues which is equal to 1 if the respondent answers yes to at least one of the following questions: 1) Does your partner (fully or partly) spend most of her/his time out of work or looking for a work? 2) Does the partner (fully or partly) sometimes 132 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES drink or stay away from home when he/she cannot find work?, and 0 otherwise. All the regressions are based on the sample of married or cohabiting respondents (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. Table 3A.13: Predictors of Number of Sexual Abuse Including Indicators of Partner Issues - Only Married or Cohabiting Respondents Coefficient /(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center -0.083* -0.029 -0.129* (0.049) (0.068) (0.071) IDP not in collective center -0.233*** -0.310*** -0.183** (0.071) (0.108) (0.090) ALP 0.095* 0.122** 0.055 (0.049) (0.059) (0.079) Demographic variables Age (ln) -0.039 -0.092 0.034 (0.058) (0.065) (0.114) Education Ref: Lower secondary Upper secondary/vocational 0.106** 0.186** 0.034 (0.054) (0.077) (0.073) Higher vocational/bachelor’s degree 0.006 0.096 -0.08 (0.054) (0.072) (0.085) Master’s or doctorate degree 0.007 0.084 -0.054 (0.069) (0.111) (0.085) Gender Ref: Female Male -0.016 (dropped) (dropped) (0.053) Household size 0.005 -0.008 0.018 (0.012) (0.016) (0.017) Location Ref: Tbilisi Other urban -0.247*** -0.320*** -0.138 (0.087) (0.118) (0.115) ANNEXES 133 Rural -0.321*** -0.381*** -0.226* (0.091) (0.123) (0.123) Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable -0.146*** -0.195*** -0.110** (0.045) (0.073) (0.051) Unemployed 0.056 0.081 0.021 (0.076) (0.142) (0.075) Out of labor force -0.044 -0.051 -0.064 (0.052) (0.090) (0.052) Employment status before the conflict Ref: Employed Unemployed -0.106 -0.235** -0.003 (0.067) (0.096) (0.090) Out of labor force 0.01 -0.016 0.046 (0.051) (0.077) (0.060) Economic status Bottom 40% -0.075 -0.092 -0.065 (0.051) (0.062) (0.086) Housing Ref: Adequate Marginal 0.038 0.092 -0.021 (0.052) (0.080) (0.054) Inadequate 0 -0.06 0.076 (0.063) (0.060) (0.103) No asset 0.048 0.027 0.105* (0.043) (0.065) (0.055) Partner issue ref: no issue partner has at least one issue 0.105** 0.081 0.174** (0.047) (0.061) (0.073) Number of observations 1860 884 976 R-squared 0.066 0.082 0.074 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Sexual Abuse is defined as the total number of sexual violence and sexual harassment reported by respondents (ranging from0 to 7). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative 134 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: wheth- er a dwelling has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how indi- viduals scored on each indicator their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/ business, car, jewelry/gold/other valuables, cash savings. We add to the basic model (which here exclude the variable marital status) the variable Partner’s issues which is equal to 1 if the respondent answers yes to at least one of the following questions: 1) Does your partner (fully or partly) spend most of her/his time out of work or looking for a work? 2) Does the partner (fully or partly) sometimes drink or stay away from home when he/she cannot find work?, and 0 otherwise. All the regressions are based on the sample of married or cohabiting respondents (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. Table 3A.14: Predictors of Number of Physical Violence Including Indicators of Partner Issues - Only Married or Cohabiting Respondents Coefficient /(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center -0.078 -0.108** -0.079 (0.049) (0.045) (0.094) IDP not in collective center -0.12 -0.190*** -0.018 (0.084) (0.061) (0.175) ALP 0.128 0.028 0.238 (0.078) (0.064) (0.157) Demographic variables Age (ln) -0.079 -0.125 -0.124 (0.079) (0.105) (0.144) Education Ref: Lower secondary Upper secondary/vocational -0.004 0.116 -0.152 (0.100) (0.105) (0.174) Higher vocational/bachelor’s degree -0.068 0.102 -0.294 (0.103) (0.098) (0.197) Master’s or doctorate degree -0.152 -0.055 -0.276 (0.113) (0.115) (0.206) ANNEXES 135 Gender Ref: Female Male 0.142** (dropped) (dropped) (0.057) Household size -0.021 -0.036 -0.002 (0.018) (0.022) (0.028) Location Ref: Tbilisi Other urban -0.209*** -0.14 -0.281** (0.080) (0.100) (0.129) Rural -0.300*** -0.210** -0.402*** (0.081) (0.090) (0.142) Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable -0.158*** -0.159** -0.137* (0.050) (0.073) (0.071) Unemployed 0.07 0.094 0.063 (0.110) (0.162) (0.152) Out of labor force -0.047 -0.096 0.033 (0.065) (0.087) (0.103) Employment status before the conflict Ref: Employed Unemployed -0.088 -0.252** 0.008 (0.102) (0.114) (0.156) Out of labor force 0.025 -0.03 0.065 (0.057) (0.069) (0.097) Economic status Bottom 40% 0.032 -0.002 0.042 (0.063) (0.074) (0.105) Housing Ref: Adequate Marginal -0.076* -0.052 -0.085 (0.046) (0.059) (0.071) Inadequate -0.017 0.06 -0.085 136 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES (0.084) (0.106) (0.126) No asset 0.03 0.07 -0.019 (0.049) (0.070) (0.081) Partner issue ref: no issue partner has at least one issue 0.033 -0.026 0.086 (0.059) (0.076) (0.092) Number of observations 1831 874 957 R-squared 0.050 0.053 0.061 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Physical Vilence is defined as the total number of physical violence forms (1) slapped you or thrown something at you 2) dragged, pushed or shoved you 3) kicked you, hit you with their fist or with something else 4) chocked or burnt you on purpose 5) threatened to use or actually used a gun, knife orother weapon) reported by respondents.(ranging from 0 to 5). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indica- tor their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/ other valuables, cash savings. We add to the basic model (which here exclude the variable marital status) the variable Part- ner’s issues which is equal to 1 if the respondent answers yes to at least one of the following questions: 1) Does your partner (fully or partly) spend most of her/his time out of work or looking for a work? 2) Does the partner (fully or partly) sometimes drink or stay away from home when he/she cannot find work?, and 0 otherwise. All the regressions are based on the sample of married or cohabiting respondents (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. Table 3A.15: Predictors of Number of Sexual Violence Including Indicators of Partner Issues - Only Married or Cohabiting Respondents Coefficient /(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center -0.016 -0.014 0.003 (0.023) (0.035) (0.021) IDP not in collective center -0.065*** -0.107** -0.031 (0.025) (0.044) (0.020) ALP 0.035** 0.034* 0.03 ANNEXES 137 (0.016) (0.019) (0.022) Demographic variables Age (ln) -0.015 -0.005 -0.035 (0.030) (0.039) (0.049) Education Ref: Lower secondary Upper secondary/vocational 0.025 0.061** -0.009 (0.016) (0.030) (0.020) Higher vocational/bachelor’s degree 0.022 0.076** -0.026 (0.025) (0.036) (0.034) Master’s or doctorate degree 0.008 0.049 -0.02 (0.035) (0.066) (0.027) Gender Ref: Female Male -0.026 (dropped) (dropped) (0.019) Household size 0.003 -0.006 0.011 (0.005) (0.007) (0.007) Location Ref: Tbilisi Other urban -0.022 -0.066 0.034 (0.033) (0.053) (0.021) Rural -0.045 -0.110** 0.033 (0.035) (0.055) (0.023) Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable -0.045** -0.060* -0.037** (0.018) (0.032) (0.019) Unemployed -0.017 -0.035 -0.014 (0.031) (0.059) (0.032) Out of labor force -0.021 -0.018 -0.029** (0.023) (0.039) (0.013) Employment status before the conflict Ref: Employed Unemployed 0.015 -0.028 0.073 138 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES (0.034) (0.036) (0.053) Out of labor force 0.019 0.028 0.009 (0.023) (0.036) (0.022) Economic status Bottom 40% -0.054** -0.049 -0.063* (0.026) (0.037) (0.032) Housing Ref: Adequate Marginal 0.031 0.042 0.019 (0.026) (0.043) (0.019) Inadequate 0.006 -0.013 0.027 (0.020) (0.025) (0.032) No asset 0.014 0.023 0.003 (0.023) (0.036) (0.023) Partner issue ref: no issue partner has at least one issue 0.018 0.049 0.002 (0.026) (0.045) (0.018) Number of observations 1795 861 934 R-squared 0.024 0.038 0.053 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Sexual Violence is defined as the total number of sexual violence forms (1) physically forced you to have sex 2) had sexual intercourse with you when you were unable to refuse 3) had sexual intercourse with you against your will 4) forced you to do something sexual that you found humiliating) reported by respondents.(ranging from 0 to 4). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the house- hold by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indicator their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/other valuables, cash savings. We add to the basic model (which here exclude the variable marital status) the variable Partner’s issues which is equal to 1 if the respondent answers yes to at least one of the following questions: 1) Does your partner (fully or partly) spend most of her/ his time out of work or looking for a work? 2) Does the partner (fully or partly) sometimes drink or stay away from home when he/she cannot find work?, and 0 otherwise. All the regressions are based on the sample of married or cohabiting respondents (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. ANNEXES 139 Table 3A.16: Predictors of Number of Sexual Violence Including Indicators of Partner Issues - Only Married or Cohabiting Respondents Coefficient /(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center -0.066 -0.014 -0.131* (0.042) (0.055) (0.067) IDP not in collective center -0.167*** -0.202** -0.154* (0.058) (0.083) (0.082) ALP 0.064 0.091* 0.026 (0.042) (0.053) (0.066) Demographic variables Age (ln) -0.024 -0.086 0.064 (0.046) (0.056) (0.085) Education Ref: Lower secondary Upper secondary/vocational 0.084* 0.126* 0.045 (0.050) (0.066) (0.072) Higher vocational/bachelor’s degree -0.012 0.025 -0.054 (0.049) (0.062) (0.075) Master’s or doctorate degree 0.001 0.035 -0.034 (0.055) (0.075) (0.079) Gender Ref: Female Male 0.01 (dropped) (dropped) (0.046) Household size 0.002 -0.002 0.007 (0.010) (0.015) (0.012) Location Ref: Tbilisi Other urban -0.225*** -0.255*** -0.172 (0.073) (0.093) (0.108) Rural -0.277*** -0.272*** -0.256** (0.078) (0.099) (0.116) Employment status 140 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable -0.104*** -0.138** -0.073 (0.037) (0.058) (0.045) Unemployed 0.073 0.114 0.039 (0.064) (0.118) (0.069) Out of labor force -0.025 -0.035 -0.033 (0.043) (0.073) (0.051) Employment status before the conflict Ref: Employed Unemployed -0.120** -0.206** -0.069 (0.056) (0.082) (0.074) Out of labor force -0.009 -0.043 0.034 (0.044) (0.067) (0.050) Economic status Bottom 40% -0.022 -0.043 -0.006 (0.040) (0.048) (0.070) Housing Ref: Adequate Marginal 0.007 0.05 -0.04 (0.041) (0.060) (0.053) Inadequate -0.006 -0.048 0.048 (0.057) (0.049) (0.097) No asset 0.035 0.004 0.102** (0.036) (0.053) (0.049) Partner issue ref: no issue partner has at least one issue 0.089** 0.035 0.169** (0.039) (0.041) (0.069) Number of observations 1850 879 971 R-squared 0.069 0.078 0.087 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Sexual Harassment is defined as the total number of sexual harassment forms (1) touched, hugged or kissed you against your will 2) stared or leered at you inappropriately 3) made sexually suggestive comments 4) sent or showed you sexually suggestive photos 5) exposed themeselves to you indecently) reported by respondents.(ranging from 0 to 5). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Household ANNEXES 141 size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indica- tor their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/ other valuables, cash savings. We add to the basic model (which here exclude the variable marital status) the variable Part- ner’s issues which is equal to 1 if the respondent answers yes to at least one of the following questions: 1) Does your partner (fully or partly) spend most of her/his time out of work or looking for a work? 2) Does the partner (fully or partly) sometimes drink or stay away from home when he/she cannot find work?, and 0 otherwise. All the regressions are based on the sample of married or cohabiting respondents (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. Table 3A.17: Predictors of Number of Controlling Behavior Including Indicators of Partner Issues - Only Married or Cohabiting Respondents Coefficient /(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center -0.140*** -0.084* -0.176*** (0.032) (0.045) (0.047) IDP not in collective center -0.084* -0.077 -0.104 (0.050) (0.074) (0.066) ALP -0.013 0.055 -0.071 (0.040) (0.046) (0.062) Demographic variables Age (ln) -0.07 -0.031 -0.138* (0.053) (0.084) (0.074) Education Ref: Lower secondary Upper secondary/vocational -0.168** -0.074 -0.261** (0.083) (0.114) (0.120) Higher vocational/bachelor’s degree -0.245** -0.134 -0.362** (0.103) (0.142) (0.146) Master’s or doctorate degree -0.202** -0.165 -0.236 (0.097) (0.129) (0.144) Gender Ref: Female 142 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Male 0.028 (dropped) (dropped) (0.038) Household size -0.008 -0.006 -0.008 (0.011) (0.017) (0.015) Location Ref: Tbilisi Other urban -0.177*** -0.185** -0.153** (0.056) (0.078) (0.074) Rural -0.219*** -0.253*** -0.180** (0.054) (0.074) (0.074) Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable -0.034 -0.072 0.005 (0.044) (0.075) (0.057) Unemployed 0.038 0.015 0.046 (0.063) (0.095) (0.086) Out of labor force -0.025 -0.051 0.02 (0.039) (0.057) (0.050) Employment status before the conflict Ref: Employed Unemployed -0.079 -0.112 -0.041 (0.055) (0.073) (0.078) Out of labor force 0 0.006 -0.005 (0.040) (0.057) (0.056) Economic status Bottom 40% -0.024 -0.02 -0.048 (0.037) (0.053) (0.059) Housing Ref: Adequate Marginal -0.039 -0.016 -0.055 (0.034) (0.055) (0.037) Inadequate 0.049 -0.057 0.164** (0.055) (0.061) (0.080) No asset 0.013 0 0.025 ANNEXES 143 (0.034) (0.051) (0.051) Partner issue ref: no issue partner has at least one issue 0.024 0.053 0.007 (0.039) (0.059) (0.059) Number of observations 1872 888 984 R-squared 0.051 0.056 0.069 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Controlling Behavior is defined as the total number of sexual violence forms (1) belittled or humiliated in front of other peoplere 2) threatened to hurt you or someone you care about) ported by respondents.(ranging from 0 to 2). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indica- tor their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/ other valuables, cash savings. We add to the basic model (which here exclude the variable marital status) the variable Part- ner’s issues which is equal to 1 if the respondent answers yes to at least one of the following questions: 1) Does your partner (fully or partly) spend most of her/his time out of work or looking for a work? 2) Does the partner (fully or partly) sometimes drink or stay away from home when he/she cannot find work?, and 0 otherwise. All the regressions are based on the sample of married or cohabiting respondents (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. A3.4.2 Controlling for Partner Issues (Married or Cohabiting Respondents with Employed Partner) Moreover, from the questionnaire, we have further information on partners’ issues for respondents who are not only married or cohabiting (question 2.2 codes 4 and 5) but whose partner is currently employed (question 4.6 codes 2, 3, 4, and 5) for a total of 825 observations. In particular, we include in the basic model the variable Partner’s occupational stability which is equal to 1 if the partner has a job which is stable or partly stable and 0 otherwise (from Table 3A.18 to Table 3A.24). 144 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Table 3A.18: Predictors of Number of Total Violence – Including Partner’s Job Stability Coefficient /(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center -0.374*** -0.460*** -0.152 (0.126) (0.139) (0.256) IDP not in collective center -0.572*** -0.758*** -0.113 (0.185) (0.223) (0.321) ALP 0.09 0.094 -0.156 (0.135) (0.136) (0.342) Demographic variables Age (ln) 0.097 0.135 -0.067 (0.252) (0.323) (0.352) Education Ref: Lower secondary Upper secondary/vocational 0.141 0.586** -0.465 (0.315) (0.258) (0.601) Higher vocational/bachelor’s degree 0.111 0.502* -0.435 (0.304) (0.270) (0.576) Master’s or doctorate degree -0.183 0.075 -0.521 (0.294) (0.253) (0.591) Gender Ref: Female Male 0.022 (dropped) (dropped) (0.144) Household size -0.043 -0.028 -0.061 (0.037) (0.046) (0.044) Location Ref: Tbilisi Other urban -0.716*** -0.755*** -0.520* (0.228) (0.279) (0.284) Rural -0.805*** -0.683*** -0.869*** (0.203) (0.244) (0.311) Employment status Employment status currently ANNEXES 145 Ref: Employed - Completely or partly stable Employed - Not stable -0.357*** -0.522*** -0.218 (0.133) (0.195) (0.159) Unemployed 0.016 0.126 0.155 (0.186) (0.241) (0.227) Out of labor force -0.322* -0.308 -0.326 (0.191) (0.248) (0.260) Employment status before the conflict Ref: Employed Unemployed -0.008 -0.476* 0.618* (0.220) (0.252) (0.338) Out of labor force 0.181 0.12 -0.008 (0.188) (0.227) (0.313) Economic status Bottom 40% -0.266** -0.370** -0.102 (0.120) (0.152) (0.162) Housing Ref: Adequate Marginal 0.023 0.057 -0.115 (0.140) (0.184) (0.132) Inadequate 0.258 0.068 0.623 (0.199) (0.206) (0.397) No asset 0.089 0.041 0.228 (0.140) (0.177) (0.192) Partner issue ref: no issue partner has at least one issue 0.106 0.255* -0.349 (0.138) (0.149) (0.273) Number of observations 820 510 310 R-squared 0.106 0.114 0.214 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Total Violence is defined as the total number of physical violence, sexual violence, sexual harassment and controlling behavior reported by respondents (ranging from 0 to 14).. The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) 146 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water net- works; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indicator their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/other valuables, cash savings. We add to the basic model (which here exclude the variable marital status) the variable Partner’s job stability which is equal to 1 if the partner has a job which is stable or partly stable and 0 otherwise. All the regressions are based on the sample of married or cohabiting respondents (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. Table 3A.19: Predictors of Number of Sexual Abuse - Including Partner’s Job Stability Female Coefficient /(Standard Error) Female Conflict status Ref: National sampling IDP in collective center -0.254*** (0.090) IDP not in collective center -0.395*** (0.097) ALP 0.081 (0.089) Demographic variables Age (ln) 0.046 (0.173) Education Ref: Lower secondary Upper secondary/vocational 0.383** (0.162) Higher vocational/bachelor’s degree 0.394** (0.158) Master’s or doctorate degree 0.122 (0.152) Household size -0.045 (0.033) Location Ref: Tbilisi Other urban -0.283* (0.157) ANNEXES 147 Rural -0.292** (0.134) Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable -0.225*** (0.085) Unemployed 0.112 (0.156) Out of labor force -0.154 (0.119) Employment status before the conflict Ref: Employed Unemployed -0.22 (0.147) Out of labor force 0.093 (0.120) Economic status Bottom 40% -0.198* Housing Ref: Adequate Marginal -0.011 (0.114) Inadequate 0.126 (0.125) No asset 0.172 (0.123) Partner issue ref: no issue partner has at least one issue 0.208** (0.092) Number of observations 502 R-squared 0.105 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable GBV is defined as the total number of physical violence and sexual violence reported by respondents (ranging from 0 to 9). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without 148 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variable defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or doctoral degree. d) Gender: male or female e) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indica- tor their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/ other valuables, cash savings. We add to the basic model (which here exclude the variable marital status) the variable Part- ner’s job stability which is equal to 1 if the partner has a job which is stable or partly stable and 0 otherwise. All the regressions are based on the sample of married or cohabiting respondents (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. Table 3A.20: Predictors of Number of Sexual Abuse - Including Partner’s Job Stability Coefficient /(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center -0.053 -0.1 0.1 (0.068) (0.084) (0.116) IDP not in collective center -0.208** -0.346*** 0.032 (0.101) (0.123) (0.163) ALP 0.135* 0.051 0.181 (0.073) (0.060) (0.168) Demographic variables Age (ln) 0.04 0.001 0.103 (0.086) (0.105) (0.161) Education Ref: Lower secondary Upper secondary/vocational 0.127 0.172 0.057 (0.106) (0.153) (0.115) Higher vocational/bachelor’s degree 0.05 0.09 0.039 (0.081) (0.132) (0.109) Master’s or doctorate degree -0.001 0.032 -0.045 (0.092) (0.142) (0.101) Gender Ref: Female ANNEXES 149 Male -0.064 (dropped) (dropped) (0.066) Household size -0.003 -0.012 0.002 (0.016) (0.020) (0.022) Location Ref: Tbilisi Other urban -0.273** -0.333** -0.099 (0.113) (0.138) (0.143) Rural -0.314*** -0.310** -0.228 (0.101) (0.122) (0.148) Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable -0.107** -0.178** -0.113** (0.048) (0.083) (0.052) Unemployed 0.121 0.107 0.219 (0.120) (0.180) (0.166) Out of labor force -0.086 -0.118 -0.027 (0.085) (0.120) (0.127) Employment status before the conflict Ref: Employed Unemployed 0.005 -0.181 0.282 (0.113) (0.115) (0.216) Out of labor force 0.078 0.087 -0.104 (0.082) (0.101) (0.119) Economic status Bottom 40% -0.100* -0.151** 0.031 (0.054) (0.071) (0.098) Housing Ref: Adequate Marginal 0.06 0.1 -0.034 (0.073) (0.096) (0.059) Inadequate 0.026 -0.022 0.148 (0.071) (0.083) (0.105) No asset 0.056 0.021 0.205*** 150 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES (0.065) (0.080) (0.071) Partner issue ref: no issue partner has at least one issue 0.110** 0.133** -0.061 (0.052) (0.066) (0.089) Number of observations 813 506 307 R-squared 0.080 0.093 0.198 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Sexual Abuse is defined as the total number of sexual violence and sexual harassment reported by respondents (ranging from0 to 7). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: wheth- er a dwelling has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how indi- viduals scored on each indicator their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/ business, car, jewelry/gold/other valuables, cash savings. We add to the basic model (which here exclude the variable marital status) the variable Partner’s job stability which is equal to 1 if the partner has a job which is stable or partly stable and 0 oth- erwise. All the regressions are based on the sample of married or cohabiting respondents (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. Table 3A.21: Predictors of Number of Physical Violence - Including Partner’s Job Stability Coefficient /(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center -0.152** -0.202*** -0.070 (0.068) (0.066) (0.163) IDP not in collective center -0.195** -0.262*** 0.039 (0.091) (0.065) (0.235) ALP 0.009 0.032 -0.126 (0.065) (0.080) (0.141) Demographic variables Age (ln) 0.007 0.029 -0.049 (0.122) (0.152) (0.207) Education ANNEXES 151 Ref: Lower secondary Upper secondary/vocational 0.122 0.306** -0.136 (0.148) (0.134) (0.301) Higher vocational/bachelor’s degree 0.175 0.317** -0.045 (0.148) (0.133) (0.329) Master’s or doctorate degree -0.027 0.084 -0.173 (0.144) (0.139) (0.292) Gender Ref: Female Male 0.086 (dropped) (dropped) (0.074) Household size -0.036 -0.032 -0.036 (0.023) (0.030) (0.028) Location Ref: Tbilisi Other urban -0.202* -0.186 -0.211 (0.110) (0.126) (0.197) Rural -0.244** -0.166 -0.368* (0.098) (0.111) (0.208) Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable -0.168** -0.189** -0.158 (0.065) (0.076) (0.110) Unemployed -0.040 0.097 -0.086 (0.112) (0.159) (0.143) Out of labor force -0.158* -0.125 -0.208 (0.089) (0.106) (0.148) Employment status before the conflict Ref: Employed Unemployed -0.028 -0.190 0.186 (0.114) (0.129) (0.266) Out of labor force 0.076 0.046 0.069 (0.086) (0.092) (0.194) Economic status Bottom 40% -0.103 -0.127 -0.069 152 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES (0.077) (0.095) (0.102) Housing Ref: Adequate Marginal -0.005 -0.044 0.037 (0.062) (0.068) (0.111) Inadequate 0.162 0.130 0.217 (0.123) (0.116) (0.266) No asset 0.076 0.123 -0.025 (0.067) (0.085) (0.138) Partner issue ref: no issue partner has at least one issue 0.059 0.196*** -0.175 (0.074) (0.067) (0.179) Number of observations 805 501 304 R-squared 0.082 0.094 0.121 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Physical Vilence is defined as the total number of physical violence forms (1) slapped you or thrown something at you 2) dragged, pushed or shoved you 3) kicked you, hit you with their fist or with something else 4) chocked or burnt you on purpose 5) threatened to use or actually used a gun, knife orother weapon) reported by respondents.(ranging from 0 to 5). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indica- tor their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/ other valuables, cash savings. We add to the basic model (which here exclude the variable marital status) the variable Part- ner’s job stability which is equal to 1 if the partner has a job which is stable or partly stable and 0 otherwise. All the regressions are based on the sample of married or cohabiting respondents (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. ANNEXES 153 Table 3A.22: Predictors of Number of Sexual Violence - Including Partner’s Job Stability Coefficient /(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center -0.019 -0.056 0.092 (0.042) (0.045) (0.083) IDP not in collective center -0.094*** -0.136*** -0.042 (0.034) (0.048) (0.030) ALP 0.068** 0.051* 0.053 (0.027) (0.026) (0.050) Demographic variables Age (ln) 0.018 0.017 -0.024 (0.045) (0.053) (0.063) Education Ref: Lower secondary Upper secondary/vocational 0.035 0.08 -0.047 (0.040) (0.052) (0.067) Higher vocational/bachelor’s degree 0.03 0.078 -0.035 (0.038) (0.048) (0.079) Master’s or doctorate degree 0.01 0.036 -0.029 (0.038) (0.043) (0.077) Gender Ref: Female Male -0.02 (dropped) (dropped) (0.027) Household size -0.002 -0.013 0.011 (0.006) (0.009) (0.007) Location Ref: Tbilisi Other urban -0.045 -0.093 0.06 (0.051) (0.067) (0.045) Rural -0.089** -0.126** 0.007 (0.042) (0.057) (0.026) Employment status Employment status currently 154 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Ref: Employed - Completely or partly stable Employed - Not stable -0.02 -0.042 -0.012 (0.021) (0.032) (0.029) Unemployed 0.01 0.011 0.03 (0.054) (0.072) (0.073) Out of labor force -0.024 -0.033 -0.001 (0.024) (0.035) (0.016) Employment status before the conflict Ref: Employed Unemployed 0.043 -0.034 0.175 (0.064) (0.043) (0.139) Out of labor force 0.035 0.044 -0.046 (0.034) (0.040) (0.038) Economic status Bottom 40% -0.067** -0.072* -0.045 (0.030) (0.039) (0.030) Housing Ref: Adequate Marginal 0.01 0.031 -0.028 (0.039) (0.054) (0.023) Inadequate 0.016 -0.005 0.084 (0.035) (0.030) (0.069) No asset 0.044 0.055 0.046 (0.037) (0.051) (0.030) Partner issue ref: no issue partner has at least one issue 0.009 0.015 -0.037 (0.032) (0.047) (0.041) Number of observations 792 495 297 R-squared 0.048 0.068 0.175 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Sexual Violence is defined as the total number of sexual violence forms (1) physically forced you to have sex 2) had sexual intercourse with you when you were unable to refuse 3) had sexual intercourse with you against your will 4) forced you to do something sexual that you found humiliating) reported by respondents.(ranging from 0 to 4). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed ANNEXES 155 - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the house- hold by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indicator their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/other valuables, cash savings. We add to the basic model (which here exclude the variable marital status) the variable Partner’s job stability which is equal to 1 if the partner has a job which is stable or partly stable and 0 otherwise. All the regressions are based on the sample of married or cohabiting respondents (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. Table 3A.23: Predictors of Number of Sexual Harassment - Including Partner’s Job Stability Coefficient /(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center -0.033 -0.043 0.013 (0.053) (0.064) (0.095) IDP not in collective center -0.111 -0.210** 0.074 (0.088) (0.105) (0.160) ALP 0.072 0.003 0.132 (0.059) (0.055) (0.135) Demographic variables Age (ln) 0.023 -0.014 0.124 (0.064) (0.077) (0.139) Education Ref: Lower secondary Upper secondary/vocational 0.09 0.093 0.106 (0.093) (0.138) (0.114) Higher vocational/bachelor’s degree 0.018 0.014 0.073 (0.074) (0.124) (0.088) Master’s or doctorate degree -0.014 -0.002 -0.015 (0.081) (0.135) (0.077) Gender Ref: Female Male -0.045 (dropped) (dropped) (0.057) Household size -0.001 0.001 -0.008 (0.014) (0.018) (0.019) 156 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Location Ref: Tbilisi Other urban -0.227** -0.240** -0.158 (0.090) (0.106) (0.143) Rural -0.224** -0.184* -0.238 (0.089) (0.105) (0.154) Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable -0.088** -0.135* -0.099** (0.040) (0.070) (0.047) Unemployed 0.112 0.099 0.189 (0.089) (0.131) (0.150) Out of labor force -0.063 -0.084 -0.028 (0.073) (0.102) (0.126) Employment status before the conflict Ref: Employed Unemployed -0.038 -0.148* 0.105 (0.083) (0.087) (0.161) Out of labor force 0.044 0.045 -0.06 (0.069) (0.086) (0.114) Economic status Bottom 40% -0.035 -0.079 0.075 (0.048) (0.063) (0.099) Housing Ref: Adequate Marginal 0.052 0.069 -0.005 (0.057) (0.075) (0.057) Inadequate 0.011 -0.018 0.073 (0.054) (0.068) (0.102) No asset 0.014 -0.033 0.160** (0.053) (0.061) (0.067) Partner issue ref: no issue partner has at least one issue 0.102*** 0.117** -0.022 (0.039) (0.049) (0.076) Number of observations 809 503 306 R-squared 0.068 0.074 0.153 ANNEXES 157 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Sexual Harassment is defined as the total number of sexual harassment forms (1) touched, hugged or kissed you against your will 2) stared or leered at you inappropriately 3) made sexually suggestive comments 4) sent or showed you sexually suggestive photos 5) exposed themeselves to you indecently) reported by respondents.(ranging from 0 to 5). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indica- tor their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/ other valuables, cash savings. We add to the basic model (which here exclude the variable marital status) the variable Part- ner’s job stability which is equal to 1 if the partner has a job which is stable or partly stable and 0 otherwise. All the regressions are based on the sample of married or cohabiting respondents (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. Table 3A.24: Predictors of Number of Controlling Behavior - Including Partner’s Job Stability Coefficient /(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center -0.178*** -0.172*** -0.184** (0.042) (0.044) (0.086) IDP not in collective center -0.175*** -0.158* -0.185*** (0.064) (0.089) (0.062) ALP -0.038 0.031 -0.198 (0.058) (0.063) (0.147) Demographic variables Age (ln) 0.047 0.098 -0.12 (0.103) (0.129) (0.135) Education Ref: Lower secondary Upper secondary/vocational -0.103 0.115 -0.384 (0.140) (0.112) (0.257) Higher vocational/bachelor’s degree -0.12 0.088 -0.428 (0.156) (0.132) (0.299) Master’s or doctorate degree -0.149 -0.031 -0.303 158 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES (0.151) (0.116) (0.314) Gender Ref: Female Male -0.006 (dropped) (dropped) (0.055) Household size -0.007 0.012 -0.028 (0.018) (0.025) (0.018) Location Ref: Tbilisi Other urban -0.240*** -0.231*** -0.210** (0.070) (0.087) (0.095) Rural -0.254*** -0.217** -0.273*** (0.071) (0.090) (0.100) Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable -0.08 -0.155 0.053 (0.069) (0.098) (0.100) Unemployed -0.021 0.013 0.022 (0.081) (0.111) (0.102) Out of labor force -0.088 -0.075 -0.091 (0.065) (0.082) (0.080) Employment status before the conflict Ref: Employed Unemployed -0.011 -0.159 0.151 (0.101) (0.101) (0.170) Out of labor force 0.035 -0.005 0.029 (0.072) (0.087) (0.150) Economic status Bottom 40% -0.06 -0.089 -0.065 (0.058) (0.083) (0.081) Housing Ref: Adequate Marginal -0.028 0.009 -0.118 (0.050) (0.064) (0.074) Inadequate 0.07 -0.039 0.260* ANNEXES 159 (0.077) (0.084) (0.133) No asset -0.038 -0.09 0.047 (0.045) (0.055) (0.071) Partner issue ref: no issue partner has at least one issue -0.054 -0.053 -0.114 (0.066) (0.079) (0.106) Number of observations 820 510 310 R-squared 0.074 0.084 0.190 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Controlling Behavior is defined as the total number of sexual violence forms (1) belittled or humiliated in front of other peoplere 2) threatened to hurt you or someone you care about) ported by respondents.(ranging from 0 to 2). The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: whether the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indica- tor their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/ other valuables, cash savings. We add to the basic model (which here exclude the variable marital status) the variable Part- ner’s job stability which is equal to 1 if the partner has a job which is stable or partly stable and 0 otherwise. All the regressions are based on the sample of married or cohabiting respondents (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. A3.5 FACTORS CORRELATED TO TRADITIONAL VIEWS In this section, we want to identify what characteristics are correlated with the prevalence of norms and attitudes toward GBV. We use again the indicators described in Section 1 Acceptability of violence (Table 3A.25). Table 3A.25: Factors Related to Norms and Beliefs toward Gender Coefficient/(Standard Error) Acceptability of Violence Conflict status Ref: National sampling IDP in collective center −0.017 (0.031) IDP not in collective center −0.055* (0.031) 160 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES ALP −0.141*** (0.046) Demographic variables Age (ln) 0.097*** (0.035) Education Ref: Lower secondary Upper secondary/vocational −0.067 (0.048) Higher vocational/bachelor’s degree −0.174*** (0.052) Master’s or doctorate degree −0.192*** (0.054) Gender Ref: Female Male 0.058* (0.031) Location Ref: Tbilisi Other urban −0.01 (0.039) Rural 0.077** (0.039) Economic status Bottom 40% −0.025 (0.031) Number of observations 2,967 R-squared 0.0515 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Acceptability of which is equal to 1 if the respondent considers acceptable at least one of the following reasons for a man to hit his spouse: 1) she does not complete her chores 2) she disobeys him 3) she asks him if he has other girlfriends 4) he finds out that she has been unfaithful. The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise. All the regressions are based on the full sample. Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. ANNEXES 161 A3.6 PREDICTORS OF THE PROBABILITY OF BEING A VICTIM OF THE FFV OR MORE FORMS OF VIOLENCE Among the 708 respondents who report to have had at least one form of violence, 129 declare to have been victim of more than one form of violence. To verify if there are some factors that can predict the probability to be a more forms of victim of vi- olence, we estimate a linear regression where we investigate which variables of the basic model are significantly correlated with the probability to have reported more than one form of FFV. The analysis is limited to respondents who report at least one form of violence (Table 3A.26). Table 3A.26: Factors Related to the Probability to Have Reported the FFV or More Forms of Violence Coefficient /(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center −0.05 −0.019 −0.085 (0.039) (0.052) (0.109) IDP not in collective center −0.081* −0.081 −0.072 (0.046) (0.075) (0.157) ALP 0.227*** 0.246*** 0.256 (0.068) (0.088) (0.232) Demographic variables Age (ln) 0.019 0.04 0.006 (0.055) (0.076) (0.059) Education Ref: Lower secondary Upper secondary/vocational 0.004 −0.039 0.205 (0.072) (0.099) (0.191) Higher vocational/bachelor’s degree −0.049 −0.095 0.171 (0.086) (0.105) (0.121) Master’s or doctorate degree −0.09 −0.127 0.102 (0.086) (0.125) (0.102) Gender Ref: Female Male −0.05*** (dropped) (dropped) (0.008) Marital Status Ref: Single/widower 162 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Married/cohabiting −0.004 −0.032 0.061 (0.049) (0.070) (0.050) Separated/divorced −0.032 −0.039 −0.057 −0.004 (0.090) (0.098) Household size −0.015 −0.003 −0.01 (0.013) (0.016) (0.014) Location Ref: Tbilisi Other urban −0.083** −0.068 −0.029 (0.037) (0.058) (0.047) Rural −0.109*** −0.188*** −0.022 (0.039) (0.060) (0.049) Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable 0.067 0.348** 0.144** (0.079) (0.156) (0.072) Unemployed 0.138* 0.216* 0.093* (0.077) (0.118) (0.055) Out of labor force 0.004 0.066 −0.032 (0.046) (0.063) (0.049) Employment status before the conflict Ref: Employed Unemployed −0.159*** −0.135 −0.211*** (0.060) (0.207) (0.073) Out of labor force 0.035 0.034 0.047 (0.039) (0.054) (0.052) Economic status Bottom 40% −0.026 −0.007 −0.037 (0.038) (0.056) (0.050) Housing Ref: Adequate Marginal 0.082* 0.09 0.067 (0.049) (0.056) (0.045) Inadequate −0.02 −0.095 0.095 (0.063) (0.057) (0.075) ANNEXES 163 No asset 0.082* −0.033 0.082** (0.031) (0.047) (0.040) Number of observations 698 347 351 R-squared 0.0665 0.0802 0.1308 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Probability of systematic violence which is equal to 1 if the respondent reported to have been a victim of four or more forms of violence. The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Marital status a categorical variable defined as: 1) single or widower 2) Married or cohabiting 3) Separated/ divorced. e) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: wheth- er the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indicator their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jew- elry/gold/other valuables, cash savings. All the regressions are based on the full sample (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. A3.7 AWARENESS OF AVAILABILITY OF SERVICES To verify what factors can help prevent GBV, we use two questions of the questionnaire: ƒƒ Question 8.1: Have you heard of any campaign or activities in your community or workplace about preventing GBV? ƒƒ Question 8.3: Have you seen or heard an advertisement or public service announcement about GBV? ƒƒ We built a measure of awareness that is equal to 1 if the respondent reports to have heard at least of a campaign/activity or of an advertisement once or more than once (Table 3A.27). Table 3A.27: Awareness of Services Coefficient /(Standard Error) Total Female Male Conflict status Ref: National sampling IDP in collective center 0.062 0.081 0.047 (0.044) (0.082) (0.072) IDP not in collective center −0.001 −0.049 0.08 (0.108) (0.090) (0.118) 164 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES ALP 0.197** 0.204* 0.196* (0.094) (0.108) (0.108) Demographic variables Age (ln) −0.135*** −0.111*** −0.121*** (0.012) (0.011) (0.013) Education Ref: Lower secondary Upper secondary/vocational 0.151*** 0.124** 0.166** (0.022) (0.048) (0.064) Higher vocational/bachelor’s degree 0.266*** 0.177*** 0.348*** (0.031) (0.033) (0.036) Master’s or doctorate degree 0.248*** 0.214*** 0.276*** (0.015) (0.034) (0.045) Gender Ref: Female Male −0.165*** (dropped) (dropped) (0.022) Marital Status Ref: Single/widower Married/cohabiting −0.012 0.042 −0.067* (0.033) (0.045) (0.042) Separated/divorced −0.028 0.013 −0.064 (0.029) (0.072) (0.077) Household size 0.000 0.008 −0.006 (0.008) (0.009) (0.012) Location Ref: Tbilisi Other urban −0.018 −0.019 −0.017 (0.031) (0.039) (0.067) Rural −0.179*** −0.161*** −0.198*** (0.028) (0.029) (0.027) Employment status Employment status currently Ref: Employed - Completely or partly stable Employed - Not stable 0.02 0.018 0.023 (0.047) (0.059) (0.038) ANNEXES 165 Unemployed 0.058* 0.011 0.085** (0.039) (0.059) (0.044) Out of labor force −0.015 −0.018 −0.008 (0.028) (0.034) (0.039) Employment status before the conflict Ref: Employed Unemployed −0.090** −0.054 −0.110** (0.038) (0.074) (0.048) Out of labor force −0.085*** −0.099*** −0.081** (0.019) (0.022) (0.033) Economic status Bottom 40% 0.037* 0.036 0.04 (0.025) (0.042) (0.042) Housing Ref: Adequate Marginal −0.067*** −0.105*** −0.03 (0.028) (0.026) (0.034) Inadequate −0.068* −0.114** −0.001 (0.039) (0.047) (0.053) Economic empowerment −0.001 No official technical training −0.081*** −0.105*** −0.057* (0.018) (0.025) (0.030) No technical skills −0.031 −0.014 −0.055* (0.024) (0.032) (0.031) Number of observations 2,798 1,401 1,397 R-squared 0.1367 0.1309 0.1225 Source: Population’s Life Experiences in Georgia Survey, 2016. Note: The table reports the linear regression results of the OLS estimation equation. The dependent variable Awareness which is equal to 1 if the respondents reported to have heard of a campaign and/or of an advertisment related to GBV. The explanatory variables include: a) Conflict Status: which identifies the four target groups: 1) nationally representative respondents without IDPs and ALPs 2) IDPs living in collective centers 3) IDPs not living in collective centers 4) ALPs. b) Age: age of the respondents in log c) Education of the respondent, a categorical variables defined as 1) Lower secondary education 2) Upper secondary education 3) Higher vocational/bachelor degree 4) Master or docatoral degree. d) Gender: male or female e) Marital status a categorical variable defined as: 1) single or widower 2) Married or cohabiting 3) Separated/ divorced. e) Household size: ranging from 1 to 10 f) Location: categorical variable defined as 1) residing in Tbilisi 2) residing in other urban areas 3) residing in rural areas g) Current Employment Status: which is defined in the four categories: 1) Employed - Stable or partially stable 2) Employed - not stable 3) unemployed 4) out of the labor force h) Employment Status before the conflict: which is defined in the three categories: 1) Employed 2) Unemployed 3) Out of the labor force i) Bottom 40%: which is equal to 1 if the respondent belong to the lower 40% of the welfare distribution, 0 otherwise j) Housing: A housing index was created using three indicators that capture living conditions in the dwelling: (a) Overcrowding: whether a dwelling has more than two members of the household by room (excluding kitchen, hallways, storage); (b) Bathroom: wheth- er the dwelling has a bathroom, even if shared; this indicates access to sewage and water networks; and (c) Heating: whether the dwelling has a heating system, which means the dwelling relies on burning fuel for heat. Based on how individuals scored on each indicator their housing was considered 1) adequate 2) marginal 3) inadequate k) No asset: which is equal to 1 if 166 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES the respondent doesn’t own any of the following assets: land/plot, apartment/house/summer house, company/business, car, jewelry/gold/other valuables, cash savings. All the regressions are based on the full sample (column Total), on women only (column Female) and on men only (column Male). Standard errors, clustered, are reported in parentheses. Significance: p-value<0.01 ***; p-value<0.05 **; p-value<0.1 *. Annex 4: Socio-economic Characteristics of the Sample Groups Table 4A.1: Socioeconomic Characteristics of the Sample Groups Nonconflict-Affected Person Internally Displaced Person Administrative Line Person ƒƒ Any non-IDP or non-ALP ƒƒ A person forced to flee his or her home ƒƒ Person residing in areas bordering ƒƒ 93.8 percent of the national population but who remains within his or her current or former conflict zones ƒƒ From all regions of Georgia country’s borders ƒƒ 1.1 percent of national population ƒƒ 5.1 percent of national population ƒƒ Defined as areas where barbed wire ƒƒ Displaced from Abkhazia and South fences were installed in 2008 by the Ossetia during conflicts in the 1990s government along former conflict and 2008; now living in all parts of zones Georgia Age 18-24 18-24 18-24 12% 13% 11% 65+ 65+ 65+ 25-34 18% 16% 25-34 25-34 22% 16% 55-64 19% 55-64 20% 16% 16% 55-64 17% 35-44 35-44 35-44 16% 45-54 45-54 18% 17% 17% 18% 45-54 18% Education Primary Primary Master Primary Master 1% Master 1% Bachelor 4% 1% 16% 17% 5% Bachelor Vocational 11% Bachelor 20% 14% Vocational Secondary Secondary Secondary Vocational 25% 47% 46% 70% 22% Rural-urban Rural 25% Rural Rural 100% 45% Urban 55% Urban 75% 167 Welfare Top 60 Bottom 40 16% 26% Urban 55% Urban 75% 168 GENDER BASED VIOLENCE IN GEORGIA: LINKS AMONG CONFLICT, ECONOMIC OPPORTUNITIES AND SERVICES Welfare Top 60 Bottom 40 16% 26% Bottom 40 Top 60 40% 60% Top 60 74% Bottom 40 84% Housing Marginal Marginal 30% Marginal 31% 31% Adequate 33% Adequate Adequate 58% 62% Inadequate Inadequate 7% 12% Inadequate 36% Women Men ƒƒ A nationally representative sample of all ƒƒ A nationally representative sample of all women over 18 years in Georgia men over 18 years in Georgia ƒƒ 54 percent of national population ƒƒ 46 percent of national population Includes NAPs, IDPs, and ALPs Includes NAPs, IDPs, and ALPs Age 18-24 65+ 18-24 65+ 11% 25-34 15% 14% 21% 55-64 18% 16% 25-34 20% 55-64 16% 35-44 45-54 35-44 45-54 16% 18% 18% 18% Education Primary Primary Master 2% Master 0% 16% 17% Bachelor Bachelor Secondary Secondary 10% 11% 45% 50% Vocational Vocational 22% 27% Rural-urban Rural Rural 42% 48% Urban Urban 52% 58% Welfare Bottom 40 Bottom 40 Rural Rural 42% 48% Urban Urban 52% 58% ANNEXES 169 Welfare Bottom 40 Bottom 40 39% Top 60 41% 59% Top 60 61% Housing Marginal Marginal 31% 32% Adequate Adequate 61% 61% Inadequate Inadequate 8% 7% Source: Population’s Life Experiences in Georgia Survey, 2016. Note: Chart data weighted by sampling weights.