This publication was produced for review by the United States Agency for International Development. It was prepared by NORC at the 
University of Chicago, with Jocelyn Kelly (Harvard Humanitarian Initiative), Sarah Khan (Yale University), and Maarten Voors (Wageningen 
University) as Principle Investigators. The authors’ views expressed in this publication do not necessarily reflect the views of the United 
States Agency for International Development or the United States Government.
COUNTERING GENDER BASED 
VIOLENCE IN EASTERN DEMOCRATIC 
REPUBLIC OF CONGO
BASELINE REPORT
TASKING 059
Contract No. GS-10F-0033M/AID-0AA-M-13-00013
MAY 2019
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | i 
COUNTERING GENDER BASED 
VIOLENCE IN EASTERN DEMOCRATIC 
REPUBLIC OF CONGO
BASELINE REPORT
TASKING 059
MAY 2019
Prepared under Contract No.: GS-10F-0033M/AID-0AA-M-13-00013
SUBMITTED TO:
Brandy Witthoft
SUBMITTED BY:
NORC at the University of Chicago
Attention: Renee Hendley, Program Manager
Bethesda, MD 20814
Tel: 301634-9489
E-mail: Hendley-Renee@norc.org
DISCLAIMER 
The authors’ views expressed in this publication do not necessarily reflect the views of the United 
States Agency for International Development or the United States Government.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | ii
CONTENTS
ACRONYMS......................................................................................................................................................... VII
EXECUTIVE SUMMARY....................................................................................................................................... 1
1. INTRODUCTION............................................................................................................................................. 4
2. BACKGROUND................................................................................................................................................ 5
3. PLANNED IMPACT EVALUATION DESIGN............................................................................................ 7
3.1 LEVEL OF IMPLEMENTATION AND COMMUNITY SELECTION........................................... 7
3.2 DESCRIPTION OF CBTH INTERVENTION ................................................................................... 7
3.3 HYPOTHESIS ..........................................................................................................................................10
4. SAMPLE ..............................................................................................................................................................11
4.1 SELECTION OF STUDY AREA .........................................................................................................11
4.2 TREATMENT ASSIGNMENT .............................................................................................................11
4.3 RESPONDENT SELECTION...............................................................................................................12
5. BASELINE DATA COLLECTION STRATEGY ........................................................................................13
5.1 DATA COLLECTION..........................................................................................................................13
6. KEY FINDINGS ................................................................................................................................................14
6.1 RESPONDENT CHARACTERISTICS...............................................................................................14
6.1.1 RESPONDENT GENDER AND AGE....................................................................................14
6.1.2 RESPONDENT EDUCATION AND LITERACY................................................................15
6.1.3 INCOME, ASSETS AND EMPLOYMENT .............................................................................16
6.1.4 ECONOMIC SHOCKS ..............................................................................................................19
6.2 CONFLICT AND DISPLACEMENT ................................................................................................. 20
6.3 SECURITY, DISPUTES & GROUP RELATIONS ............................................................................24
6.4 MENTAL HEALTH, TRAUMA AND RESILIENCE........................................................................28
6.4.1 DEPRESSION, ANXIETY, PTSD .............................................................................................28
6.4.2 SELF-EFFICACY AND RESILIENCE .......................................................................................30
6.5 PREVALENCE OF GBV ........................................................................................................................33
6.5.1 INTIMATE PARTNER VIOLENCE..........................................................................................33
6.5.2 NON-PARTNER SEXUAL VIOLENCE .................................................................................35
6.5.3 SEXUAL VIOLENCE BY AN ARMED GROUP...................................................................35
6.5.4 PERPETRATION OF IPV AND NON-PARTNER SEXUAL VIOLENCE......................35
6.6 ATTITUDES TOWARDS GENDER EQUALITY...........................................................................35
6.7 ATTITUDES AND NORMS AROUND GBV.................................................................................37
6.8 AWARENESS OF GBV-RELATED SERVICES ACCESS ...............................................................39
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | iii
7. RISK FACTORS FOR GBV VICTIMIZATION AND PERPETRATION.............................................42
7.1 IPV VICTIMIZATION AMONG WOMEN......................................................................................43
7.2 MEN’S PERPETRATION OF IPV........................................................................................................43
7.3 MEN’S PERPETRATION OF NON-PARTNER SEXUAL VIOLENCE......................................44
8. CONCLUSIONS..............................................................................................................................................47
REFERENCES.........................................................................................................................................................49
APPENDIX A: INTERVENTION TIMELINE..................................................................................................51
APPENDIX B: MEMO ON POPULATION-LEVEL ESTIMATES OF THE BASELINE REPORT .......52
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | iv
TABLES 
Table 1: Intervention Areas.............................................................................................................................11
Table 2: Gender Breakdown of Household Survey Sample.....................................................................14
Table 3: Level of education..............................................................................................................................15
Table 4: Literacy, by gender.............................................................................................................................15
Table 5: Household’s primary source of income........................................................................................16
Table 6: Household and individual income...................................................................................................17
Table 7: Individual primary occupation .........................................................................................................18
Table 8: Proportion of respondents reporting witnessing types of conflict events, by gender......23
Table 9: Proportion of respondents reporting personally experiencing types of conflict 
events, by gender...............................................................................................................................24
Table 10: Depression, Anxiety and PTSD Checklist Items in Survey ......................................................28
Table 11: Self-Efficacy and Resilience Survey Items......................................................................................31
Table 12: Index of attitudes towards gender equality in different domains...........................................36
Table 13: Index of attitudes and norms towards GBV................................................................................38
Table 14: Barriers to accessing GBV-related services.................................................................................39
Table 15: Barriers to access to GBV-related service if perpetrated by non-family member.............41
Table 16: Barriers to access to GBV-related service if perpetrated by family member .....................41
Table 17: Village-Level Correlation Between GBV Measures ...................................................................42
Table 18: Correlates of IPV ...............................................................................................................................45
FIGURES
Figure 1: Experimental Sample .........................................................................................................................12
Figure 2: Map indicating location of baseline villages in Eastern DRC....................................................13
Figure 3: Age distribution, by gender..............................................................................................................14
Figure 4: Proportion of respondents who are literate, by zone and gender........................................16
Figure 5: Mean wealth score by zone and gender.......................................................................................18
Figure 6: Proportion of villages having experienced various types of economic shocks in the 
past 5 years..........................................................................................................................................20
Figure 7: Proportion of respondents who are currently displaced, by health zone and gender......21
Figure 8: Proportion of respondents who witnessed at least 1 conflict-related event, by 
health zone and gender.....................................................................................................................22
Figure 9: Proportion of respondents who personally experienced at least 1 conflict-related 
event, by health zone and gender...................................................................................................22
Figure 10: Mean perceived safety in community, by health zone and gender.........................................25
Figure 11: Proportion of households facing exposure in last 12 months, by event type and 
gender...................................................................................................................................................26
Figure 12: Respondent rating of relations with others, by health zone....................................................27
Figure 13: Respondent rating of relations with others, by gender............................................................27
Figure 14: Proportion of respondents with scores above cutoff for depression and anxiety 
symptom checklist, by gender and health zone ..........................................................................30
Figure 15: Proportion of respondents with scores above cutoff for PTSD symptom checklist, 
by gender and health zone...............................................................................................................30
Figure 16: Mean self-efficacy score, by health zone and gender................................................................. 32
Figure 17: Mean individual resilience score, by health zone and gender..................................................32
Figure 18: Mean community resilience score, by health zone and gender ..............................................33
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | v 
Figure 19: Proportion of ever-partnered women having ever experienced IPV (physical or 
sexual), by health zone......................................................................................................................34
Figure 20: Proportion of ever-partnered women having experienced IPV (physical or sexual) 
during the last 12 months, by health zone ...................................................................................34
Figure 21: Gender equitable attitude index, by health zone and gender..................................................37
Figure 22: Index of acceptability of GBV, by health zone and gender.......................................................38
Figure 23: Number of indicated barriers to access GBV-related services if perpetrated by a 
family member, by health zone and gender.................................................................................40
Figure 24: Number of indicated barriers to access GBV-related services if perpetrated by non￾family member, by health zone and gender.................................................................................40
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | vi
IMA INDICATORS
IMA Indicator 1: G1.1 Prevalence rate of GBV among women, men, and children in target 
geographic areas........................................................................................................................................................33
IMA Indicator 2: IR 1.2.1 Percentage of target population reporting increased agreement that 
males and females should have equal access to social, economic and political opportunities...............35
IMA Indicator 3: IR 1.2.4 Percentage of target population that views GBV as less acceptable 
after participating in or being exposed to USG programming.......................................................................37
IMA Indicator 4: IR 2.3.2 Percentage of target population reporting disagreement with identified 
barriers to accessing GBV-related community services..................................................................................39
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | vii
ACRONYMS
CBTH Community-Based Trauma Healing
DRC Democratic Republic of Congo
GBV Gender-based violence
HZ Health zone
IE Impact evaluation
IPV Intimate partner violence
SFCG Search for Common Ground
GBV Gender-based violence
SV Sexual violence
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 1 
EXECUTIVE SUMMARY
This report presents key results from a baseline survey implemented in selected villages in Eastern 
Democratic Republic of Congo (DRC) during fall 2018. The baseline is part of a randomized Impact 
Evaluation (IE) of a program that aims to change norms around Gender Based Violence. The impact 
evaluation assesses the impact of selected aspects of the Tushinde Ujeuri program to reduce gender￾based violence, coordinated by IMA World Health and implemented by several partners including 
Panzi Hospital, Heal Africa, ABA Rule of Law Initiative and Search for Common Ground. 
The field experiment tests the effectiveness of Community-Based Trauma Healing (CBTH) as a core 
approach to change norms and behavior around gender-based violence (GBV), and particularly 
intimate partner violence (IPV) as well as sexual violence (SV) more broadly. The evaluation 
randomly selects 80 villages from a list of 160 villages across 3 health zones in Eastern DRC to 
receive the CBTH program. The IE assesses impacts on perceived norms, and individual attitudes 
and behaviors (self-reported and observed). Moreover, the IE aims to assess whether and how these 
interventions affect the uptake of services provided to survivors of GBV through the health area￾level service delivery of the Tushinde Ujeuri program. 
This baseline survey was collected in all 160 villages that are part of the Impact Evaluation across 
Walikale, Katana, and Nyangezi Health Zones, and in an additional 32 avenues in the semi-urban
Karisimbi Health Zone1. The data reported below presents results across all 192 villages. We also 
report on four core indicators relevant to IMA’s monitoring and evaluation plan, namely (i) 
prevalence of GBV, (ii) gender equality, (iii) attitudes towards GBV and (iv) attitudes towards 
barriers to accessing GBV services.
The results from the baseline survey reflect the harsh realities that decades of conflict have brought 
to this region. Political instability and other hardships have wrought changes at both the village and 
individual level. Villages reported extremely high levels of social and economic shocks in the past five 
years, including drought, flooding, crop failure and disease outbreak. Witnessing or experiencing 
war-related abuses was extremely common. More than half of respondents reported having 
witnessed or experienced at least one conflict-related event. Respondents also reported high 
exposure to sexualized violence related to conflict. Over 22% of all respondents report having 
witnessed rape or sexual abuse of a woman; 9% of all respondents report having witnessed rape or 
sexual abuse of a man and 10% report having personally experienced sexual abuse by an armed 
group. 
A core indicator of GBV relates to Intimate Partner Violence (a first component of IMA indicator 1).
Over half of all ever-partnered women surveyed reported having experienced IPV at some point in 
their lives (53.4%), a number that is well above the regional average for Sub-Saharan Africa. Over a 
third of women (38.9%) reported experiencing IPV in the past 12 months. The survey also examined 
women and men’s experiences with sexual violence committed by a non-partner. One-fifth of 
women and one in twelve men reported being ever raped by a non-partner. Again, these numbers 
are much higher than the regional victimization rate for women in Sub-Saharan Africa, where in the 
past year, 14.6% of women and 6.4% of men reported experiencing sexual violence. 
1 These 32 additional avenues in Karisimbi Health Zone were included in the baseline data collection as a 
precaution in case accessibility or safety issues prevented the implementation of the CBTH intervention in 
Walikale. More detail is provided in the Sample section below. 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 2 
Almost one-third of men and 13% of women report perpetrating physical IPV against a partner in 
their lifetime. Roughly one in seven men have reported ever forcing someone other than their 
partner to have sex against their will, and 4% of men reported doing so in the past year. Just over six 
percent of women report ever forcing someone to have sex against their will, and 3% reported 
doing so in the past year.
A second core indicator of GBV relates to non-partner sexual violence (a second component of IMA 
indicator 1). One-fifth of women (20.5%) and one in twelve men (8.23%) reported being ever raped 
by a non-partner. Again, these numbers are much higher than the regional victimization rate for 
women in Sub-Saharan Africa, which is 14%. In the past year, 14.6% of women and 6.4% of men 
reported experiencing sexual violence.
The survey assessed individuals’ own attitudes toward the acceptability of GBV, as well as individuals’ 
beliefs about the attitudes of others in their community (IMA indicator 2 and 3). In general, there 
exists a sizable gender gap in attitudes; women hold less accepting attitudes towards GBV while men 
have more permissive attitudes. The most common barriers to accessing GBV services differed 
slightly for family versus non-family perpetrators of GBV. However, in general, the most commonly 
reported problems were not having money for transport or for services fees, and not knowing 
where to access services. 
Individuals participating in the research had high levels of depression and anxiety as well as PTSD – 
an outcome that is expected given the high levels of trauma reported elsewhere in the survey. 
Across all health zones, a significantly larger proportion of women report anxiety and depression 
and PTSD than men.
There are stark gendered differences in many of the outcomes assessed for this project, including 
but not limited to: access to education, access to income-generating activities and resilience and 
mental health outcomes. There is also substantial variation among health zones in a number of 
outcomes, including literacy, employment, mental health outcomes and perceptions of security. This 
highlights the importance of disaggregating these data to gain a more nuanced picture of how 
outcomes may vary across populations.
There are strong and significant links between the GBV outcomes and exposure to war-related 
violence. Women who have witnessed or experienced war-related events are more likely to be 
victims of IPV. Men who witnessed war-related events are more likely to report perpetration of IPV 
and SV. These results suggest a strong link between exposure to political conflict and perpetration of 
violence within the home. These findings are particularly striking given the extremely high-levels of 
war-exposure reported in the survey.
Finally, there were strong links between certain measures of poor social cohesion and men’s 
perpetration of different forms of IPV. Men who report worse relations with neighboring villages are 
also more likely to report perpetrating SV and IPV.
Taken together, these results suggest that violence within communities – including IPV and other 
forms of GBV – are highly associated with individuals’ experiences with the conflict, and are also 
associated with measures of social cohesion. Addressing the underlying trauma of conflict and trying 
to rebuild trust and social cohesion may provide an avenue to addressing some of the underlying 
mechanisms that propagate violent behaviors in highly war-affected areas.
We summarize the key findings from the baseline survey below:
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 3 
• Exposure to witnessing or experiencing war-related abuses was extremely common. More than 
half of respondents reported having witnessed or experienced at least one conflict-related event.
• Individuals participating in the research had high levels of depression and anxiety as well as 
PTSD. Across all health zones, a significantly larger proportion of women report anxiety and 
depression and PTSD than men.
• Over half of all ever-partnered women surveyed reported having experienced IPV at some point 
in their lives. One-fifth of women and one in twelve men reported being ever raped by a non￾partner. These numbers are higher than the regional victimization rate for women in Sub￾Saharan Africa. 
• Men who witnessed war-related events were more likely to report perpetration of IPV and SV.
• War-related experiences are highly and significantly associated with higher odds of women 
reporting IPV victimization.
• Men who reported more mistrust and resentment of neighboring communities also had higher 
odds of perpetrating IPV and sexual violence against non-partners.
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 4 
1. INTRODUCTION
USAID's Counter Gender-Based Violence (C-GBV) program known as Tushinde Ujeuri in Swahili) in 
the Democratic Republic of the Congo (DRC) seeks to:
1. promote positive gender norms, power-relations and attitudes to address a core driver of
GBV;
2. respond to survivors' medical, emotional, and legal needs; and
3. reduce stigma to enhance reintegration of survivors.
Tushinde Ujeuri builds on the previous 5-year program “Overcoming Sexual and Gender Based 
Violence in Eastern DRC” (known as Ushindi in Swahili), which has implemented a holistic approach 
to supporting survivors of Sexual and Gender Based Violence (SGBV) through the integrated 
delivery of psychosocial, medical, legal and economic services to survivors and their communities. 
While there is an overwhelmingly positive perception of the holistic service delivery approach for 
survivors of GBV among implementing partners, and suggestive evidence from an independent study 
to confirm this (Amisi 2016), far less is known about the effectiveness of preventative strategies. 
During a scoping trip conducted by the research team in 2017, implementing partners identified two 
key factors driving both the incidence of GBV and perpetuating stigma among survivors: (i) 
“traditional culture and norms'', and (ii) exposure to, and fallout from, conflict-related traumatic 
events.2 
Partners also pointed to the importance of sensitization efforts in the form of trainings, workshops 
and "Social and Behavior Change Communication" (SBCC) campaigns conducted as part of Ushindi in 
creating awareness about specific issues relating to GBV, and the program services. It is an open 
question as to whether these efforts lead to enduring changes in attitudes, norms and behaviors so 
as to prevent future incidence of GBV, and which types of efforts are most effective in doing so. This 
mirrors an overall dearth of empirical evidence from interventions oriented towards primary 
prevention of GBV (Ellsberg 2015).
The proposed impact evaluation (IE) seeks to narrow this evidentiary gap by evaluating a trauma 
healing strategy for GBV prevention in a post-conflict context. The program, Community Based 
Trauma Healing (CBTH) involves explicitly addressing post-conflict trauma resulting from, among 
other things, systematic rape in the past, indigenous conflict, internal displacement and inheritance 
conflicts through community detraumatisme sessions that bring together potential victims, 
perpetrators, traditional leaders and faith-based leaders.
Principal Investigators for this evaluation are Jocelyn Kelly, Sarah Khan, and Maarten Voors. The 
research protocol for this study was approved by the NORC Institutional Review Board 
(IRB00000967), Project Number 7554.059.01, IRB Protocol Number 18.08.13 on 09/10/2018.
2 See Kelly, Khan and Voors. "Scoping Trip Report". December 7, 2017.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 5 
2. BACKGROUND
GBV is one of the most pervasive and harmful forms of 
violence globally. It encompasses any form of violence that 
is perpetrated on a person as a result of their gender 
identity. Men, women, boys and girls can all be victims as 
well as perpetrators of GBV. However, global evidence 
suggests that the great majority of GBV victims are women 
and that men have higher rates of GBV perpetration than 
women. 
Thirty-five percent of women globally report being victims 
of physical or sexual abuse during their lifetime and one￾third of women who have been in a relationship have 
experienced physical or sexual violence from a partner. 
Globally, one out of every five women will become a victim 
of rape or attempted rape during her lifetime, and intimate 
partner violence (IPV) is the most common form of 
violence against women. Sequelae of sexual abuse include 
injury and death as well as poorer mental and reproductive 
health outcomes, including unintended pregnancy; physical 
harm; increased vulnerability to sexually transmitted 
diseases, and HIV.
GBV has many forms; a complex interplay of risk factors that contribute to this form of violence. 
While the types and causes of GBV may vary from one context to another, a number of multi￾country research efforts have begun to uncover some of the mechanisms underlying GBV. Unequal 
power relations between men and women, harmful gender norms and beliefs, experiences with 
violence in childhood, and drug and alcohol abuse are some of the factors that are consistently 
associated with GBV.
There is also a link between political instability and increased rates of GBV. Research from Uganda, 
Thailand, Cote d’Ivoire and Liberia has found that women who are subject to higher levels of 
conflict-related abuses also report higher levels of IPV victimization during and after conflict. 
Exposure to political violence and to human rights abuses has been linked to higher rates of IPV 
perpetration among men in conflict and post-conflict settings. Thus, personal experience with 
conflict-related abuse appears to enhance vulnerability to IPV. 
Relatively little systematic information is available about the factors associated with men’s 
perpetration of sexual violence, particularly in war-affected areas. Rates of perpetration vary greatly 
from context to context – reinforcing the finding that social norms and gendered inequalities, as well 
as other factors, can affect levels of violence across the globe. Men’s reports of perpetrating physical 
IPV vary from 42% in South Africa to 22% in India. Reports of non-partner rape vary from 9% in 
Chile to 37% in South Africa.
The United Nations Multi-Country Study on Men and Violence in Asia and the Pacific is one of the 
most comprehensive efforts to examine the prevalence and risk factors of GBV perpetration among 
men (Fulu et al, 2013). This study found that a number of factors are linked to sexual violence 
perpetration, including: having more sexual partners, seeking transactional sex, using physical 
violence against female partners, men’s own victimization and low socioeconomic status.
United States Government 
Definition of Gender-based Violence
Violence that is directed at an individual based 
on his or her biological sex, gender identity, 
or perceived adherence to socially defined 
norms of masculinity and femininity. It includes 
physical, sexual, and psychological abuse; 
threats; coercion; arbitrary deprivation of 
liberty; and economic deprivation, whether 
occurring in public or private life. GBV takes 
on many forms and can occur throughout the 
life cycle. Types of gender-based violence can 
include female infanticide; child sexual abuse; 
sex trafficking and forced labor; sexual 
coercion and abuse; neglect; domestic 
violence; elder abuse; and harmful traditional 
practices such as early and forced marriage, 
“honor” killings, and female genital 
mutilation/cutting. 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 6 
While the UN study included two post-conflict sites, this was not a focus of their study. Indeed, very 
little is known about why men perpetrate rape in active conflict settings. Yet, data suggests that war 
may increase the rates of sexual victimization among women. It is therefore particularly important to 
understand how experiences related to political conflict may be associated with men’s perpetration 
of sexual violence. 
The data presented here suggest links between political instability and GBV, although practitioners 
and policy makers are still working to fully understand these links and how to address them. The 
Tushinde Ujeuri project attempts to directly address the harmful gender attitudes, norms and 
practices that may underlie perpetration of IPV and other forms of GBV. This evaluation aims to 
strengthen this effort by more fully examining how political conflict and war-abuses may be linked to 
the perpetration of multiple forms of GBV, and particularly IPV. 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 7 
3. PLANNED IMPACT EVALUATION DESIGN
The proposed IE assesses the impact of Community Based Trauma Healing (CBTH) as a strategy to 
address GBV in a post-conflict context. During the scoping trip, partners repeatedly mentioned 
conflict-related factors as contributing both to the incidence and prevalence of GBV, and the 
stigmatization of survivors of GBV. This activity has not been previously implemented as part of 
Ushindi, although it has been implemented by Catholic Relief Services in other parts of the DRC. The 
Search for Common Ground (SFCG) Technical Strategy Document for CBTH is used as a reference 
to highlight the main components of this activity. 
3.1 LEVEL OF IMPLEMENTATION AND COMMUNITY SELECTION
CBTH activities will be implemented at the village level in three Health Zones of Eastern DRC: 
Nyangezi, Katana and Walikale. Within these three Health Zones, CBTH will be implemented in 2 of 
4 randomly selected villages within each of 40 study Health Areas, creating a study sample of 160 
villages with 80 villages receiving the CBTH treatment. 
The choice of Health Zones was based on IMA World Health's (IMA) objective to “Identify CBTH 
beneficiary geographic entities according to geographical accessibility and security context: for this 
project, the implementing partner will target at least 3 to 5 CBTH beneficiary communities per 
health area. The choice will also be guided by the communities most affected by the post-conflict 
trauma (massive rape in the past, high scale / prevalence of sexual and gender-based violence, 
indigenous conflict and internally displaced persons, inheritance conflicts, rejection of marginalized 
people, mistrust between local communities and the security forces or institutions supposed to 
protect them etc.)''.
3.2 DESCRIPTION OF CBTH INTERVENTION
This programmatic description draws on a design developed by implementing partner Search for 
Common Ground (SFCG) to outline the objectives and the related activities that will constitute the 
CBTH intervention:
OBJECTIVE 1: TO STRENGTHEN THE COMMUNITY CAPACITY TO SUPPORT TRAUMA 
HEALING PROCESS
Result 1.1: Community trauma healing facilitators have increased capacity to support the 
community-based trauma healing process. The approach around this result is grounded in the idea 
that communities will be supported to select and train trusted interlocutors to become CBTH 
facilitators (called healing companions) to then steward healing activities in communities. The 
following activities will be undertaken in each CBTH community:
• Activity 1.1: Participatory, community-led identification of facilitators - Working with 
NORC as per the IMA agreed plan, SFCG will implement this action in 80 villages of the selected 
health zones of NYANGEZI, KATANA in South Kivu and WALIKALE in North Kivu. To cover 
the 80 villages, SFCG will retain 80 local facilitators here called THC (Trauma Healing 
Champions). To build trust, the THCs will be nominated by the community. SFCG will set a list 
of criteria including but not limited to:
− Be a member of the community
− Deserve trust from the community members
− Facilitation skills
− Gender balance (they will work in pair of man/women)
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 8 
− Sensitivity of local identity with representation where needed
− Knowing the conflict issues and GBV issues that affect the community welfare
− Be a leader for the community (doesn't mean with official responsibility) but an accepted 
person that draw persons around its vision
− Effort to include victims of: past rape, indigenous conflict, internal displacement and 
inheritance conflicts, traditional leaders and faith-based leaders, taking into account all 
sensitivities
− Be able to read and write to use the facilitation guide and fill out to data collection 
sheets
SFCG will convene a public forum in each health area where these criteria will be explained 
before the community proceeds with selection of their THC. Prior to this selection forum, 
SFCG field assistants will sensitize different people in the community to explain the approach 
and the purpose of the public forum so that everyone is prepared and all groups have a chance 
to participate and to tolerate others' participation.
• Activity 1.2: Training manual designed covering trauma healing, conflict resolution, 
and community facilitation techniques - Following the selection of the THCs, SFCG will 
assess their knowledge and skills to design/update the training manual. On this basis, SFCG’s 
training department will design the training manual with support from an in-house or external 
trauma healing expert (either from SFCG Burundi or from COPA that also supported CBTH in 
the SEEC program). The training manual will cover knowledge around CTH, GBV, Conflicts and 
violence and interrelations between all these and finally facilitation technics adapted to 
community level. SFCG will involve NORC in the design/adaptation of the training manual to 
ensure that learning since the start of the project is captured. 
• Activity 1.3: Training of THCs on trauma healing, GBV, conflict resolution and 
community participatory facilitation techniques - To train the 80 THCs, SFCG will 
conduct 7 experiential training workshops gathering together 10-12 persons per training 
workshop. Each training will last 5 days, including 3 days of practical learning and 2 days of 
application of the learning, facilitating sessions and being assessed by the other participants and 
the trainers. Participants will re-try their session until the other participants and the trainer are 
fully satisfied and the participant has adequate self-confidence to lead a trauma healing session. 
Over the course of the project, 4 months following the training and in the 12th month of the 
project, SFCG will conduct 3 refresher trainings (1 Katana for 36 THC, 1 in Nyangezi for 20 
THC and 1 in Walikale for 24THC) with the aim of refreshing the skills of the THC, 
understanding common issues and updating the facilitation materials where required.
• Activity 1.4: CBTH facilitated sessions - The 80 THCs will be deployed within the 80 
villages targeted in the 40 health areas. One trauma healing (TH) session will be conducted in 
each village per month leading to a total of 960 CBTH sessions with the goal of reaching a target 
population of 19,200-24,000 persons, including 50% women and 40% youth below 35 years. The 
THC will work in peer groups, each one moving to the village of his peer to co-facilitate with 
him/her.
Each TH session will bring together 20-25 persons. At the beginning groups could be based on 
identity, for example sessions for women themselves, without men; depending on progress 
sessions could be opened to person of different sex, or different generation (mixed groups). 
Each session will be conducted during three days. Participants will select the times that work for 
them to attend the meeting 2-3 hours each day.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 9 
Using the facilitation guide developed by SFCG, the THC will facilitate the session to enable 
participants themselves to identify incidents that affect their personality and causes them trauma. 
They will also help participants understand the signs of trauma, think about healing and solutions, 
and understand psychosocial services and other forms of support that exist in the community or 
near the community. Gender and conflict issues and all other trauma-related factors will be 
discussed. The facilitator will foster open dialogues using the skills they acquired so all 
participants feel comfortable to share throughout the session.
These sessions will also be as engaging as possible. For instance, trauma healing through stories
drawn from myths, legends or the Bible (according to the faith of the participants) will be read 
during the group discussions to inspire, nourish hope, and provide relief. There is evidence, 
including from an experiment attempted in many areas of the DRC, that in areas with strong
oral tradition and for participants with a low level of education, stories have the capacity to 
captivate participants who identify themselves to them, to provoke sharing of personal 
experiences and to nurture personal motivation to surpass oneself.
Participants requiring additional support will be referred as appropriate. For example, mediation 
and dialogue will be facilitated for victims of crimes, burdened by the fact that the perpetrator is 
known but has not faced justice, and who would like to talk to the perpetrator as part of their
healing. If a judicial or medical response is required, the person will be referred to appropriate 
resources as part of the consortium’s holistic response. If a societal norm is the cause of the 
trauma, mediation by traditional leader could be conducted, etc. 
OBJECTIVE 2: INFORM AND MOBILIZE COMMUNITIES AROUND THE TRAUMA HEALING 
SENSITIVITY AND SURVIVORS' REINTEGRATION
Result 2: The population of the targeted villages have increased knowledge about trauma healing 
and survivors feel their communities are more positively engaged for their effective reintegration
• Activity 2.1: Trauma healing and conflict resolution community celebrations - SFCG 
will organize in the target areas, at times identified with the communities, occasional cultural 
public events. These events could take the form of festivals, such as traditional celebrations 
including dances, drums, music, shared meals, etc. SFCG will discuss the content with the 
community to make sure this is a community proposed (led) initiative. GBV and conflict issues 
will be subtly integrated into the event while emphasizing enjoyment and community cohesion: 
edutainment. Once a quarter between the 5th and the 17th month, SFCG will support such an 
activity within each health area, bringing together 500 persons. A total of 160 community events 
will be organized with a goal to reach 80,000 persons with balanced gender and youth 
participation. 
Problems underlined during sessions will be weaved in these celebrations in combination with 
fun and enjoyment for a large public to raise community awareness, value the victims, and 
mobilize community engagement to overcome trauma healing issues and establish secure space 
for survivors.
SFCG will also encourage voluntary and spontaneous testimonies of former victims as part of 
the trauma healing approach. Some trauma victims who have been accompanied through a 
community-facilitated process could testify about the value of community support and the 
resulting changes to their lives. They will be given room to highlight the contribution of family 
members, neighbors, colleagues, and friends who helped them or are helping them overcome 
the tough traumatic situation. 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 10
These testimonies will not be compelled but instead will be welcomed when individuals express 
interest in sharing their experiences with community members. For the sake of the impact 
evaluation, only individuals from the treatment villages will be admitted to these events.
• Activity 2.2: Trauma healing through art using de-traumatization participatory 
theatre - SFCG has been supporting communities to use participatory theater for more than 
ten years in the most remote parts of the DRC. Participatory theater is an opportunity for the 
audience to learn, understand, and contribute to the solution on serious matters, all while having 
fun. Learning from the use of participatory theater for conflict transformation, SFCG is 
developing the "theater of de-traumatization" approach that will be utilized to build compassion 
and restore a sense of joy for the victims of conflicts and the community.
SFCG will train the community facilitators on organizing participatory theater shows within 
communities, with the support and monitoring of Project Assistants. Facilitators will form six
production teams comprised of 5-6 members each. The six production teams will prepare and 
deliver participatory theater productions within each of the 80 selected villages on a quarterly 
basis, bringing together 100 persons per production each time. Between the 4th month and the 
16th month of the project, a total of 320 participatory de-traumatization theater productions will 
be delivered within the targeted villages of the 40 health areas, bringing together a total of 
32,000 persons. Of the participants, 60% are expected to be females and 30% to be children. 
Finally, to be able to address secondary trauma, it will be important for facilitators themselves to 
be connected to available counseling services within the Tushinde Ujeuri program. They should 
also be trained identifying and referring severely affected trauma healing session participants to 
these services. For example, Cilliers et al (2016) found that a community reconciliation 
intervention in Sierra Leone improved social trust in treatment communities but worsened
individual psychological wellbeing. In this regard, the holistic service delivery approach of the 
Tushinde Ujeuri program, which provides community healing interventions in the presence of 
psychological services to deal with and address potential negative psychological consequences is 
especially promising. 
3.3 HYPOTHESIS 
The impact assessment tests the hypothesis that CBTH activities will:
1. Improve attitudes and promote positive norms around GBV.
2. Decrease GBV incidence.
3. Increase the uptake of psychological, legal and health services provided under Tushinde Ujeuri.
This baseline report serves to present descriptive statistics, examine associations between key 
variables and undertake regression modeling to better understand the relationship between conflict￾related variables and key outcomes for this project, including different forms of GBV.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 11
4. SAMPLE
4.1 SELECTION OF STUDY AREA
The intervention areas for the Tushinde Ujeuri program include health areas (aires de santé) across 
five health zones in the North Kivu and South Kivu provinces of Eastern DRC. The first four 
columns of Table 1 below provide additional information about these five health zones. 
The study sample for the IE (two rightmost columns of Table 1) comprises a total of 40 health areas 
within three largely rural Health Zones (Nyangezi, Katana and Walikale) out of the five in which the 
Tushinde program is taking place. Given the largely rural settings of USAID programs in Congo, the 
USAID Mission has greater interest in learning what works in rural settings, which led to the 
exclusion of the semi-urban Karisimibi health zone. Nonetheless, baseline data was collected in 
Karisimbi and we will report baseline results from across four Health Zones.3 In addition, and in 
consultation with the implementing partner, the Bunyakiri health zone was excluded from the survey 
sample due to accessibility concerns (an escort for the field teams would be required to gain access). 
Table 1: Intervention Areas
PROVINCE HEALTH 
ZONE CHARACTERISTICS
PROGRAM 
HEALTH 
AREAS 
BASELINE SURVEY IMPACT EVALUATION
HEALTH 
AREAS 
TOTAL 
VILLAGE 
CLUSTERS*
HEALTH 
AREAS 
TOTAL 
VILLAGE 
CLUSTERS*
North Kivu Karisimbi Semi-Urban 16 8 32 0 0 
North Kivu Walikale Rural & Access Issues 14 12 48 12 48 
South Kivu Katana Rural and Accessible 18 18 72 18 72 
South Kivu Nyangezi Rural & Accessible 12 10 40 10 40 
South Kivu Bunyakiri Rural & Access Issues 27 0 0 0 0 
Total 87 48 192 40 160
* For this study (and by extension for the baseline survey), 4 village clusters were selected in each selected health area.
4.2 TREATMENT ASSIGNMENT 
The random assignment procedure proceeded in two stages. First, within each of the three selected 
health zones, IMA and its implementation partners created a list of 40 health areas that met three 
criteria: they are receiving Tushinde Ujeuri interventions, they are relatively accessible, and they are 
not anticipated to present major security risks that might undermine or interrupt program and 
evaluation implementation. Within this subset of health areas, 160 eligible villages were identified by 
IMA and its implementation partners on the basis of their accessibility, relative security, population 
size and household count, and a limited presence of other projects that might contaminate any 
impact of the intervention. In each health area, they identified on average four villages. The baseline 
survey was conducted in these villages during fall 2018, before they were randomly assigned to the 
treatment or control group. Within each health area, approximately two villages were randomly 
 
3 Baseline data was collected in Karisimbi both in order to provide better coverage for the IMA indicators and 
to serve as a backup for the IE in case the security and accessibility conditions change in one of the other three 
health zones selected for the IE.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 12
assigned to receive the CBTH program, while the other two were assigned to not receive the 
intervention. 
The Stata Package "randomize" (Kennedy and Mann, 2017) was used to assign these 160 villages to 
either receive the CBTH intervention (treatment) or no CBTH intervention (control). The health 
areas were randomized in blocks, meaning 50% of villages in each of the 40 health areas across 3 
health zones (Katana, Nyangezi, Walikale) were assigned to receive the CBTH intervention. Figure 1 
shows the overall distribution of villages by health zone and treatment status.
Figure 1: Experimental Sample 
4.3 RESPONDENT SELECTION
Within each of the 192 villages selected for the baseline survey, field teams used a "random-walk" 
methodology to randomly select households. Enumerators interviewed every 3rd household on the 
right in rural areas and every 5th household on the right in semi-urban areas, starting from a pre￾determined village landmark (market, administrative building, well, tree, etc.) Prior to the start of 
daily data collection, the team supervisor was in charge of dispersing enumerators around the village
so that they could cover all the area and that the sample would not be confined to a small portion of 
the observation area.
In each selected household, enumerators interviewed a random adult respondent (age 18 or over) 
selected from within the household (using a kish-grid method programmed on the tablet). As per the 
programming, and to ensure the highest level of responsiveness on sensitive topics, female 
enumerators only interviewed female respondents and male enumerators only interviewed male 
respondents.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 13
5. BASELINE DATA COLLECTION STRATEGY
5.1 DATA COLLECTION
The baseline data collection, which occurred in fall 2018, comprised of both household and 
community-level surveys. The household survey was administered to a target of 20 households per 
village across all 192 villages. Of these, the 160 villages in the Katana, Nyangezi and Walikale health 
zones were part of the IE sample, while the remaining 32 villages in Karisimbi were not. Within each 
household, one randomly chosen adult household member was interviewed. The final sample 
comprises 4,223 respondents (of which 3,678 are in our experimental sample, split equally across 
treatment and control locations). Figure 2 displays a map with the location of the villages included in 
the baseline survey. The community survey was conducted with village chiefs in each of the 192 
villages and aimed to capture information about each community and its characteristics, as well as 
the perspective of the community leaders. Note that this document reports on data from all 
communities and households surveyed, including those that are not part of the IE sample.
Figure 2: Map indicating location of baseline villages in Eastern DRC
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 14
6. KEY FINDINGS
6.1 RESPONDENT CHARACTERISTICS
This section presents relevant background characteristics of the individual respondents and 
households interviewed in the baseline household survey. Background characteristics may be 
predictive of psychosocial wellbeing and GBV-related outcomes. In addition, the intervention may 
also affect beneficiaries with different background characteristics differently. 
6.1.1 RESPONDENT GENDER AND AGE
The respondents of the baseline household survey were almost equally divided into females and 
males. Of the total 4,223 respondents surveyed, 50.1% were female and 49.9% were male (Table 2). 
Table 2: Gender Breakdown of Household Survey Sample
PERCENTAGE N 
Male 49.9 2,109 
Female 50.1 2,114 
Total 100 4,223 
The average age of respondents is 35.9 years, with the youngest respondent being 18 and the oldest 
97. Female respondents in the sample are younger, with an average age of 32.8 years, while male 
respondents have an average age of 39.0 years. Figure 3 shows the distribution of age by gender in 
the sample. Note the spikes in the age distribution at the 5 and 10-year marks (more pronounced 
for female respondents) which are typical of survey data, as respondents tend to round up self￾reported age. 
Figure 3: Age distribution, by gender
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 15
6.1.2 RESPONDENT EDUCATION AND LITERACY
Nearly half of respondents (47%) did not complete any form of formal education and only 16% of 
respondents completed secondary school or higher. On average, male respondents had higher 
education levels, as only 35% of males did not complete any formal education and 22% completed 
secondary school or higher. Of the female respondents, 59% did not complete any level of formal 
education and only 10% attained secondary education or higher (Table 3). 
Table 3: Level of education
LEVEL OF EDUCATION MEN
(%)
WOMEN
(%)
TOTAL
(%)
None 265 
(12.6) 
717 
(33.92) 
982 
(23.3) 
Primary, not completed 477 
(22.6) 
534 
(25.3) 
1011 
(23.9) 
Primary, completed 207 
(9.8) 
142 
(6.7) 
349 
(8.3) 
Secondary, not completed 697 
(33.1) 
503 
(23.8) 
1,200 
(28.4) 
Secondary, completed 316 
(15.0) 
170 
(8.0) 
486 
(11.5) 
College, not completed 60 
(2.8) 
22 
(1.0) 
82 
(1.9) 
College, completed 63 
(3.0) 
21 
(1.0) 
84 
(2.0) 
Higher than college 23 
(1.1) 
4 
(0.2) 
27 
(0.6) 
Missing 1 
(0.1) 
1 
(0.1) 
2 
(0.1) 
Total 2,109
(100)
2,114
(100)
4,223
(100)
In addition to education, literacy was also measured in the survey sample by asking whether the 
respondent is able to write a simple message, e.g. “It is hot today”, in any language. By this measure, 
just over half of all respondents were literate, but substantially fewer women (42%) than men (68%) 
(Table 4). 
Table 4: Literacy, by gender
PERCENTAGE N 
Male 67.8 2,109 
Female 41.8 2,114 
Total 54.8 4,223 
Importantly, there was substantial geographical variation in educational attainment and literacy in the 
sample. In Karisimbi (mostly urban sites surveyed), 77% of respondents were literate, while in 
Walikale, Nyangezi and Katana only 56%, 45% and 49% of the respondents were literate, 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 16
respectively. The gender gap in literacy was also less pronounced in Karisimbi relative to the other 
health zones (Figure 4), although it still exists. 
Figure 4: Proportion of respondents who are literate, by zone and gender
Note: Proportion is a value between 0 and 1. For example, 0.85 implies 85% of men in our sample in Karisimbi are 
literate. 
6.1.3 INCOME, ASSETS AND EMPLOYMENT
HOUSEHOLD
Table 5 reports the primary sources of income for households in the sample. A majority of 
respondents’ households primarily live from agriculture, farming or livestock rearing. Trading was
the second most commonly reported primary source of income. 
Table 5: Household’s primary source of income
MEN COUNT (%) WOMEN COUNT (%) TOTAL COUNT (%)
Agriculture/Farming/Livestock 1,076 
(51.0) 
1,169 
(55.3) 
2,245 
(53.16) 
Trading 301 
(14.3) 
345 
(16.3) 
646 
(15.3) 
Private firm employment 160 
(7.6) 
138 
(6.5) 
298 
(7.1) 
Civil Service 120 
(5.7) 
99 
(4.7) 
219 
(5.2) 
Fishing 117 
(5.6) 
66 
(3.1) 
183 
(4.3) 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 17
MEN COUNT (%) WOMEN COUNT (%) TOTAL COUNT (%)
Any Mining 29 
(1.4) 
19 
(0.9) 
48 
(1.1) 
Military employment 15 
(0.7) 
24 
(1.14) 
39 
(0.9) 
Other 275 
(13.0) 
228 
(10.8) 
503 
(11.9) 
Missing 16 
(0.8) 
26 
(1.2) 
42 
(1.0) 
Total 2109
(100)
2114
(100)
4223
(100)
Table 6 below reports on the monetary income earned by the household as a whole and by the 
individual respondents in the past month. The mean individual income of survey respondents was 
77,331 CDF (Congolese franc). Of all male respondents in the sample, 33% were currently 
personally engaged in a job or activity from which cash income was earned. For the female 
respondents, this was 15.5%.
Table 6: Household and individual income
N MEAN SD MIN MAX
Wealth score based on assets and dwelling status 4,223 0.29 0.15 0 0.87 
Household income flow (CDF in last month) 4,223 68,769.9 98,162.2 0 600,000 
Individual Respondent Income (in CDF/month) 1,022 77,331.1 108,236.8 0 1,600,000 
Notes: outliers, values above 99th percentile, are removed.
Figure 5 below shows the mean wealth score disaggregated by zone and gender. We see that 
Walikale has the lowest mean wealth score compared to the other health zones. In all zones, 
women earned less than men but the difference was most pronounced in Walikale and least 
pronounced in Karisimbi.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 18
Figure 5: Mean wealth score by zone and gender
Table 7 below describes the occupation of survey respondents. Overall, we see that the majority of 
respondents described farming and tending livestock as their primary occupation (43.6%), with 
trading as the second most common occupation (11.8%). However, a significant percentage of 
respondents (14.9%) reported being unemployed. Women were more likely than men to report 
working in agricultural activities and in trading. Men more commonly reported being unemployed 
than women.
Table 7: Individual primary occupation
PRIMARY OCCUPATION
MEN
COUNT
(%)
WOMEN
COUNT
(%)
TOTAL
COUNT
(%)
Agriculture/farming/livestock 796 
(37.7) 
1,043 
(49.34) 
1,893 
(43.6) 
Trading 210 
(10.0) 
290 
(13.7) 
500 
(11.8) 
Unemployed 349 
(16.6) 
279 
(13.2) 
628 
(14.9) 
Other unskilled work 149 
(7.1) 
32 
(1.5) 
181 
(4.3) 
Student 124 
(5.9) 
79 
(3.7) 
203 
(4.8) 
Fishing 113 
(5.4) 
14 
(0.7) 
127 
(3.0) 
Private firm employee 96 
(4.6) 
55 
(2.6) 
151 
(3.6) 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 19
PRIMARY OCCUPATION
MEN
COUNT
(%)
WOMEN
COUNT
(%)
TOTAL
COUNT
(%)
Teacher 72 
(3.4) 
25 
(1.2) 
97 
(2.3) 
Civil Service 61 
(2.9) 
10 
(0.5) 
71 
(1.7) 
Mining 42 
(2.0) 
5 
(0.2) 
47 
(1.1) 
Other skilled Work 41 
(1.9) 
7 
(0.3) 
48 
(1.1) 
Religious leader 25 
(1.2) 
1 
(0.1) 
26 
(0.6) 
Military 14 
(0.7) 
1 
(0.1) 
15 
(0.4) 
Retired 7 
(0.3) 
1 
(0.1) 
8 
(0.2) 
Housewife / househusband 1 
(0.1) 
224 
(10.6) 
225 
(5.3) 
Missing 9 
(0.4) 
48 
(2.3) 
57 
(1.4) 
Total 2,109
(100)
2,114
(100)
4,223
(100)
6.1.4 ECONOMIC SHOCKS
The survey measures community-level shocks that may have affected villages in the last 5 years 
through surveys with community leaders. The surveys reveal overall high levels of exposure, with a 
majority of villages having experienced drought, flooding, crop failure and disease outbreak. Figure 6 
below shows the proportion of villages that experienced economic shocks in the past five years. 
Seventy-eight percent of villages reported disease outbreak, making this the most common form of 
shock, followed closely by crop failure or low crop yield (76%) and then drought (68%.) The very 
high reported levels of community shocks emphasize how common it is for village to experience 
extreme stressors in the survey area. 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 20
Figure 6: Proportion of villages having experienced various types of economic shocks 
in the past 5 years
Note: Proportion is a value between 0 and 1. For example, 0.68 implies 68% of villages in our sample experienced a 
drought. 
6.2 CONFLICT AND DISPLACEMENT
The survey measures levels of past and current displacement as well as exposure to conflict and 
conflict-related events in the population using individual questions asked to respondents, as well as 
through the community-level survey conducted with the village chief in each of the villages. 
A majority of respondents in the household survey (52.5%) report that they have been displaced 
sometime over the course of their lives as a result of armed conflict; 12.4% of respondents report 
that they are currently displaced. These levels are not significantly different for men and women. 
There is however geographical variation in current displacement levels, with most of the 
displacement concentrated in Walikale (Figure 7). 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 21
Figure 7: Proportion of respondents who are currently displaced, by health zone and 
gender 
Note: Proportion is a value between 0 and 1. For example 0.05 implies 5% of males in our sample in Karisimbi are 
currently displaced.
Respondents were asked whether they have witnessed as well as personally experienced a number 
of different conflict-related events at any point in their lives. Exposure to these events either by 
witnessing or personally experiencing them may be a source of present trauma for respondents, 
which has direct relevance for the CBTH programming. 
We constructed binary summary measures to assess whether respondents witnessed and 
experienced any conflict-related events to assess the levels of exposure among the population. The 
distribution of these measures by gender and health zone is visualized below in Figures 8 and 9. 
Overall, the rates of exposure are high, as more than half the respondents reported having 
witnessed and experienced at least one conflict-related event. Exposure seems to vary by region, 
with the highest levels of exposure in Walikale, followed by Nyangezi. While there are differences in 
exposure by gender in Karisimbi and Nyangezi, with women being less likely to report exposure to 
conflict events, this pattern does not hold in the other two health zones. 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 22
Figure 8: Proportion of respondents who witnessed at least 1 conflict-related event, by 
health zone and gender 
Note: Proportion is a value between 0 and 1. For example, 0.62 implies 62% of males in our sample in Karisimbi have 
witnessed at least 1 conflict-related event.
Figure 9: Proportion of respondents who personally experienced at least 1 conflict￾related event, by health zone and gender
Note: Proportion is a value between 0 and 1. For example 0.54 implies 54% of males in our sample in Karisimbi have 
personally experienced at least 1 conflict-related event.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 23
This section also reports disaggregated exposure by type of conflict event witnessed or experienced 
and by gender in the Tables 8 and 9 below. This is important since the type of traumatic event 
experienced may have implications for later outcomes of resilience and post-traumatic stress.4 
There is some evidence of gendered experiences of conflict events in these comparisons. Table 8 
shows the proportion of respondents witnessing types of conflict events by gender. Men are more 
likely to have witnessed harassment, the assassination of a non-family member, and the sexual abuse 
of other men. On the other hand, women are more likely to have witnessed the assassination of a 
family member.
Table 8: Proportion of respondents reporting witnessing types of conflict events, by gender
(1)
MALE
MEAN/SE
(2)
FEMALE
MEAN/SE
T-TEST
DIFFERENCE
(1)-(2)
Pillaging/Property Destruction 0.429 
[0.011] 
0.421 
[0.011] 0.008 
Harassment/Torture 0.367 
[0.010] 
0.329 
[0.010] 0.037** 
Assassination of family member 0.207 
[0.009] 
0.242 
[0.009] -0.035*** 
Assassination of other person 0.340 
[0.010] 
0.270 
[0.010] 0.070*** 
Rape/Sexual abuse of woman 0.228 
[0.009] 
0.213 
[0.009] 0.015 
Rape/Sexual abuse of man 0.098 
[0.006] 
0.082 
0.006] 0.016* 
N 2,109 2,114
Note: The value displayed for t-tests are the differences in the means across the groups. 
***, **, and * indicate significance at the 1, 5, and 10 percent critical level.
Table 9 shows the proportion of respondents experiencing types of conflict events by gender. Men 
are more likely to have personally experienced being forced to pay cash or goods, being physically 
beaten and being forced into labor by armed group. Women are more likely to have had their house 
destroyed, to have been sexually abused and to have been forced to kill another person.
In terms of absolute levels of exposure to conflict-related sexual violence, more than 22% of all 
respondents report having witnessed rape or sexual abuse of a woman, 9% of all respondents report 
having witnessed rape or sexual abuse of a man, and 10% report having personally experienced 
sexual abuse by an armed group.
 
4 In a study of gender differences in resilience among US veterans Portnoy et al (2018) find that women exhibit 
lower levels of resilience, however these differences are insignificant once accounting for the fact that women 
are more likely to experience sexual abuse, inter-personal violence and military sexual trauma. 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 24
Table 9: Proportion of respondents reporting personally experiencing types of conflict events, by 
gender
(1)
MALE
MEAN/SE
(2)
FEMALE
MEAN/SE
T-TEST
DIFFERENCE
(1)-(2)
Forced to Flee 0.355 
[0.010] 
0.368 
[0.010] -0.013 
Abducted by Armed Group
0.163 
[0.008] 
0.167 
[0.008] -0.004 
House Destroyed 0.237 
 [0.009] 
0.269 
[0.010] -0.033** 
Looting or Theft 0.370 
[0.011] 
0.348 
[0.010] 0.022 
Forced to pay cash/goods 0.328 
[0.010] 
0.228 
[0.009] 0.100*** 
Physically beaten 0.241 
[0.009] 
0.176 
[0.008] 0.065*** 
Sexually abused 0.086 
[0.006] 
0.115 
[0.007] -0.029*** 
Forced labor 0.289 
[0.010] 
0.157 
[0.008] 0.132*** 
Forced to beat other 0.080 
[0.006] 
0.074 
[0.006] 0.006 
Forced to kill other 0.032 
[0.004] 
0.055 
[0.005] 
-0.024*** 
N 2,109 2,114
Note: The value displayed for t-tests are the differences in the means across the groups. 
***, **, and * indicate significance at the 1, 5, and 10 percent critical level.
6.3 SECURITY, DISPUTES & GROUP RELATIONS
In addition to explicitly conflict related events, respondents are also asked about disputes in their 
community, with other individuals, and their perceptions of security and inter-group relations. 
When asked to think about the security situation in their village or community, around 36% of 
respondents reported feeling very unsafe or unsafe, and this did not appear to differ significantly by 
gender. A 5-point measure of perceived safety was constructed, where 1 is very unsafe and 5 is very 
safe, and the average perception of safety by region and gender is visualized below in Figure 10. 
Respondents in Karisimbi appear to have the lowest evaluation of safety. In Walikale and Nyangezi, 
women seem to have lower perceptions of safety than do men, but this is reversed in the case of 
Karisimbi and there are no significant gender differences among respondents in Katana. 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 25
Figure 10: Mean perceived safety in community, by health zone and gender
Note: question is coded on a 5-point scale of perceived safety where 1 is very unsafe and 5 is very safe
Twenty percent of respondents report having been personally involved in a conflict/dispute with 
another individual from the same or another community in the past year; among the 850 individuals 
reporting involvement in a conflict(s), 80% (N= 673) reported that at least one of these disputes 
turned violent. However, it is not possible to conclusively interpret whether this indicates that a 
large majority of inter-personal disputes turn violent, or if individuals are more inclined to say “yes” 
to having experienced a dispute if it led to violence. 
The survey also asked respondents to report whether their household has fallen victim to non￾conflict related insecurity in the form of theft, physical assault, witchcraft, forced bribe payment, and 
other criminal acts. Figure 11 presents the distribution of respondents reporting exposure to such 
events during the past 12 months. Both men and women report theft as the most common form of
non-conflict related insecurity, followed by physical assault, witchcraft, having to pay a bribe and 
other forms of criminality.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 26
Figure 11: Proportion of households facing exposure in last 12 months, by event type 
and gender
Note: Proportion is a value between 0 and 1. For example in the figure 0.41 implies 41% of respondents in our sample 
experienced theft during previous year.
GROUP RELATIONS
To assess social cohesion, the survey asked respondents about their own perception of the quality 
of relations with others in and outside their village. Respondents are asked to rate relations with 
members of their household, friends and neighbors, others in the village, neighboring villages, other 
religious groups and ethnic groups other than their own on a 5-point scale.
▪ 1 – Very Good
▪ 2 – Good 
▪ 3 – Neither good nor bad
▪ 4 – Bad 
▪ 5 – Very bad
There was low variation across these measures; i.e. on average there is no significant difference in a 
respondent’s rating of their relations with their household members and with other religious groups 
for example. Very few respondents rate relations with any group as bad or very bad, or even neutral 
(Figures 12 and 13), and little variation is seen across health zones, or between men and women. 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 27
Figure 12: Respondent rating of relations with others, by health zone
Note: Respondents rate relations on a scale of 1 (Very good) to 5 (very bad)
Figure 13: Respondent rating of relations with others, by gender
Note: Respondents rate relations on a scale of 1 (Very good) to 5 (very bad) 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 28
6.4 MENTAL HEALTH, TRAUMA AND RESILIENCE
The CBTH intervention is expected to help participants cope with trauma and have positive impacts 
on individual well-being and community resilience. The survey assessed baseline levels of these 
outcomes in the population using a number of standard psychometric scales that have been validated 
in a number of contexts.
6.4.1 DEPRESSION, ANXIETY, PTSD
To assess depression, anxiety and PTSD, this analysis draws on the methodology outlined by Bass et 
al (2013) from a mental health project in eastern DRC and uses the Hopkins Symptom Checklist (25 
items) and Harvard Trauma Questionnaire (16 items). 
Respondents report the frequency of symptoms in the past month on a 4-point scale:
▪ 0 – Not at all 
▪ 1 – A little
▪ 2 – Moderately 
▪ 3 – A lot 
We generated an average score of the 25 items and 16 items respectively (with each item weighted 
equally) for each individual respondent, which ranges from 0 to 3, with higher scores indicating 
worse symptoms. Following Bass et al (2013) we used a cutoff to determine prevalence of 
depression, anxiety, and PTSD in the population: “An average HSCL-25 score of 1.75 or higher and 
an average PTSD Checklist score of 1.75 or higher were considered to be predictive of clinically 
significant depression or anxiety and PTSD, respectively, on the basis of data from other conflict￾affected populations”
Table 10: Depression, Anxiety and PTSD Checklist Items in Survey
HOPKINS SYMPTOM CHECKLIST 25 (HSCL-25) HARVARD TRAUMA 
QUESTIONNAIRE
DEPRESSION ANXIETY PTSD
1. Feeling low in energy, slowed 
down 
1. Suddenly scared for no 
reason 
1. Recurrent thoughts or memories of the 
most hurtful or terrifying event
2. Blaming yourself for things 2. Feeling fearful 2. Feeling as though the hurtful or terrifying 
event is happening again 
3. Crying easily 3. Faintness, dizziness or 
weakness 3. Recurrent nightmares about the event
4. Loss of sexual interest or 
pleasure 
4. Nervousness or shakiness 
inside 
4. Feeling detached or withdrawn from 
others 
5. Poor appetite 5. Heart pounding or racing 5. Unable feel emotions
6. Difficulty falling asleep, staying 
asleep 6. Trembling 6. Feeling jumpy, easily startled 
7. Feeling hopeless about the 
future 7. Feeling tense or nervous 7. Difficulty concentrating 
8. Feeling sad or blue 8. Headaches 8. Feeling on guard
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 29
HOPKINS SYMPTOM CHECKLIST 25 (HSCL-25) HARVARD TRAUMA 
QUESTIONNAIRE
DEPRESSION ANXIETY PTSD
9. Feeling lonely 9. Spells of terror or panic 9. Feeling irritable or having outbursts of 
anger
10. Thoughts of ending your life 10. Feeling restless, can’t sit 
still 
10. Avoiding activities that remind of the 
traumatic or hurtful event
11. Feeling of being trapped or 
caught 
11. Inability to remember parts of the most 
traumatic or hurtful events
12. Worrying too much about 
things 12. Feeling as if you don’t have a future
13. Feeling no interest in 
things/less interest in daily 
activities
13. Avoiding thoughts of feelings associated 
with the traumatic or hurtful events
14. Feeling everything is an effort 
14. Sudden emotional or physical reaction 
when reminded of most hurtful/traumatic 
events
15. Feelings of worthlessness
15. Feeling no interest in things/less interest in 
daily activities
16. Difficulty falling asleep, staying asleep
Figure 14 shows the proportion of the sample population that is above the 1.75 score cutoff on the 
depression and anxiety scales. Across all health zones, a significantly larger proportion of women is 
above the cutoffs for both scales. Notably, the proportion of male and female respondents reporting 
high scores on depression, anxiety and mental health is substantially greater in Walikale (more than 
30% of all respondents) compared to other health zones. Figure 15 shows the proportion of the 
sample population that is above the score cutoff for PTSD, which shows very similar patterns to the 
depression and anxiety scores. Walikale has a notably higher burden of PTSD than other health 
zones, with women in Walikale showing the highest rates of PTSD in the sample. In all health zones, 
women have worse PTSD outcomes than men.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 30
Figure 14: Proportion of respondents with scores above cutoff for depression and 
anxiety symptom checklist, by gender and health zone
Note: Proportion is a value between 0 and 1. For example in the figure 0.09 implies 9% of females in our sample in 
Karisimbi have scores above cutoff for depression and anxiety symptom checklist.
Figure 15: Proportion of respondents with scores above cutoff for PTSD symptom 
checklist, by gender and health zone
Note: Proportion is a value between 0 and 1. For example in the figure 0.06 implies 6% of females in our sample in 
Karisimbi have scores above cutoff for PTSD.
6.4.2 SELF-EFFICACY AND RESILIENCE
This survey used a number of psychometric scales and other measures to triangulate the 
measurement of what is often looked at as a hard-to-define concept: resilience. This survey looked 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 31
at in-group and out-group attitudes, measures of self-efficacy, measures of individual resilience and 
measures of community resilience to build a comprehensive picture of this multifaceted concept. 
Self-efficacy, while related to resilience, represents a distinct measure related to a person’s ability to 
accomplish undertakings – this characteristic can help determine how one would approach tasks and 
challenges and overcome difficulties (Bandura, 1982). It also relates to belief in one’s ability to 
control one’s own social environment (American Psychological Association, no date). On a distinct 
but related note, resilience represents the process of adapting to one’s circumstances in the face of 
trauma, threat, and adversity. It can be manifested at the individual and community level. The 
approach used here was to build on resilience scales used in a previous USAID project in the 
Philippines (NORC, 2018) while also adding additional measures. 
We constructed scales for self-efficacy, individual resilience, and community resilience from three 
different sets of questions shown in Table 11. Respondents were asked to respond to each of the 
questions on a 5-point scale:
▪ 0 – Strongly Disagree
▪ 1 – Somewhat Disagree
▪ 2 – Neither Agree nor Disagree
▪ 3 – Somewhat Agree
▪ 4 – Strongly Agree
We generated an average score for each scale respectively (with each item weighted equally) for 
each individual respondent, which ranges from 0 to 4, with higher scores indicating greater self￾efficacy and higher resilience.
Table 11: Self-Efficacy and Resilience Survey Items
SELF-EFFICACY SCALE INDIVIDUAL RESILIENCE 
SCALE 
COMMUNITY RESILIENCE
SCALE
1. I can manage to solve difficult 
problems if I try hard enough
1. I feel that I belong in my 
community
1. People in my community feel like 
they belong to the community.
2. I can deal with unexpected 
events.
2. I know where to go if I need 
help
2. People in my community have hope 
for the future.
3. I can handle unexpected 
situations.
3. I cooperate with people 
around me
3. People in my community help each 
other.
4. I can remain calm when facing 
difficulties
4. I have people in my life who I 
can respect
4. My community treats people fairly 
no matter what their background is.
5. If I am in trouble, I can usually 
think of a solution
5. Spiritual beliefs are a source of 
strength for me 5. My community has effective leaders. 
6. Even when things are tough, I 
am able to succeed
6. If I am hungry, I can usually get 
enough food to eat
6. People in my community discuss 
issues so they can improve the 
community.
7. I think it is important to help 
out in my community
7. People in my community work 
together to improve the 
community.
8. My community keeps people 
informed about important news
9. People in my community trust the 
leaders
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 32
Figures 16, 17, and 18 show the distribution of scores on these measures by gender and health zone. 
For self-efficacy, individual resilience and community resilience, there is a pattern of gender 
differences with women consistently scoring lower than men on these measures in Katana which is 
reversed in Walikale with women consistently scoring higher than men. 
Figure 16: Mean self-efficacy score, by health zone and gender
Note: Figure reports mean score on 6 item self-efficacy scale
Figure 17: Mean individual resilience score, by health zone and gender
Note: Figure reports mean score on 7 item individual resilience scale
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 33
Figure 18: Mean community resilience score, by health zone and gender
Note: Figure reports mean score on 9 item community resilience scale
6.5 PREVALENCE OF GBV 
IMA Indicator 1: G1.1 Prevalence rate of GBV among women, men, and children in 
target geographic areas 
Preventing and responding to GBV is a core goal of the Tushinde Ujeuri project. The baseline survey, 
therefore, assessed two key types of GBV perpetration and victimization: sexual violence by a non￾partner as well as IPV. The war trauma section also asked about experiences of being a victim or 
being forced to perpetrate sexual violence by an armed group. All respondents (both women and 
men) were asked whether they personally had experienced various forms of sexual violence or IPV 
in the past 12 months (past-year violence) and at any point in their lives (ever violence). 
Note that the statistics below in 6.5.1 through 6.5.3 are broken down by perpetrator, namely sexual 
violence by 1) an intimate partner and 2) any non-partner, including by armed groups which is
analyzed separately as it is particularly relevant for conflict experience. This was done because the 
first type is limited to women who have had an intimate partner (the survey question is asked only 
to those who were ever married or partnered); while the second type of sexual violence is not 
conditional on partner status. As such, pooling the two estimates together would not be meaningful.
6.5.1 INTIMATE PARTNER VIOLENCE
Intimate partner violence among ever-partnered women was measured through the conflict tactics 
scale (CTS) and included acts of physical or sexual violence perpetrated by a partner. Ever-partnered 
women were asked to report the frequency (Never, Sometimes, or Often) with which their current 
or most recent partner engaged in a set of actions. For respondents reporting that a behavior took 
place Sometimes or Often, the survey asked a follow-up as to whether the partner had perpetrated 
the action in the last 12 months. 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 34
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
Over half of all ever-partnered women surveyed reported having experienced IPV (either physical or
sexual) at some point in their lives (53.4%). This was well above the regional average for Sub-Saharan 
Africa, which is 39.9% according to Demographic and Health Survey (DHS) data. Over a third of
women (38.9%) reported experiencing IPV (either physical or sexual) in the past 12 months.
Figures 19 and 20 report the proportions of ever-partnered female respondents reporting 
experiencing IPV at some point in their lives and in the last 12 months, by health zone. The only 
significant regional differences are between Nyangezi (lowest prevalence rates) and Katana (highest 
prevalence rates). 
Figure 19: Proportion of ever-partnered women having ever experienced IPV (physical 
or sexual), by health zone
Note: Proportion is a value between 0 and 1. For example in the figure 0.54 implies 54% of ever-married females in our 
sample in Karisimbi have experienced IPV in their lifetime
Figure 20: Proportion of ever-partnered women having experienced IPV (physical or 
sexual) during the last 12 months, by health zone 
Note: Proportion is a value between 0 and 1. For example in the figure 0.39 implies 39% of ever-married females in our 
sample in Karisimbi have experienced IPV in the last 12 months.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 35
6.5.2 NON-PARTNER SEXUAL VIOLENCE
The survey also examined women and men’s experiences with sexual violence committed by a non￾partner. Those people who answered yes to the following question were coded as having 
experienced sexual violence perpetrated by a non-partner, “At any time in your life, as a child or as 
an adult, has anyone ever forced you in any way to have sexual intercourse or perform any other 
sexual acts when you did not want to?”
One-fifth of women (20.5%) and one in twelve men (8.23%) reported being ever raped by a non￾partner. Again, these numbers are much higher than the regional victimization rate for women in 
Sub-Saharan Africa, which is 14%. In the past year, 14.6% of women and 6.4% of men reported
experiencing sexual violence. 
6.5.3 SEXUAL VIOLENCE BY AN ARMED GROUP
While war-related sexual violence was covered in a separate section, it is also worth revisiting here. 
More than one in ten women (11.4%) and one in twelve men (8.61%) stated they had experienced 
sexual violence at the hands of armed groups. This statistic speaks to the heavy burden of armed 
group violence on the population surveyed for this project. 
6.5.4 PERPETRATION OF IPV AND NON-PARTNER SEXUAL VIOLENCE
Almost one-third of men (28.5%) and 12.5% of women report perpetrating physical IPV against a 
partner in their lifetime. Roughly one in seven men (14.6%) have reported ever forcing someone 
other than their partner to have sex against their will, and 4% of men reported doing so in the past 
year. Just over six percent of women (6.7%) report ever forcing someone to have sex against their 
will, and 2.8% reported doing so in the past year. 
6.6 ATTITUDES TOWARDS GENDER EQUALITY
IMA Indicator 2: IR 1.2.1 Percentage of target population reporting increased 
agreement that males and females should have equal access to social, economic and 
political opportunities 
Note here that we cannot report on change statements due to the program, as is implied by the 
indicator, because this report is a baseline. Rather here we report on the level of agreement in the 
sample. In the endline survey, and by comparing across villages with and without the program, we 
will capture any changes.
The survey assessed respondents’ attitudes towards gender roles and gender equality in various 
domains by asking them to state the extent to which they agree with the following statements:
1. Changing diapers, giving a bath, and feeding kids is the mother’s responsibility.
2. A woman’s role is taking care of her home and family.
3. The husband should make the final decision about all major household purchases.
4. A man should have the final word about decisions in his home.
5. A woman should obey her husband in all things.
6. When jobs are scarce, men should have more right to a job than women.
7. Women should have the same rights as men to own and inherit land.
8. Men make better political leaders than women, and should be elected rather than women.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 36
The first 5 items correspond to the Domestic Chores and Daily Life Items in the Gender Equitable 
Men (GEM) Scale used to “measure attitudes toward gender norms in intimate relationships or 
differing social expectations for men and women”. The last 3 items measure attitudes towards 
equality of economic and social opportunity and are adapted from survey items in the Afro￾barometer. 
Respondents were asked to rate each statement on a 5-point scale:
▪ 1 – Strongly Disagree
▪ 2 – Somewhat Disagree
▪ 3 – Neither Agree nor Disagree
▪ 4 – Somewhat Agree
▪ 5 – Strongly Agree
We generated an index by averaging responses to the 8 statements (with each item weighted 
equally) for each individual respondent, thus ranging from 0 to 40. A higher score denotes less
equitable attitudes (Statement 7 is recoded so that higher values indicate disagreement). A 
respondent would score 0 if they strongly disagree with all inequitable statements and would score 
40 if they strongly agree with each one. The second row rescales this to range 0 to 1 to allow for 
comparisons across tables. 
Table 12: Index of attitudes towards gender equality in different domains
MEAN SD MIN MAX
Gender equitable attitudes (out of 40) 19.7 6.19 0 40 
Gender equitable attitudes (rescaled to 0-1) 0.49 0.15 0 1 
Gender equitable attitudes (men only) 21.1 6.61 0 40 
Gender equitable attitudes (women only) 18.3 5.38 0 39 
Note: Minimum value of 0 represents most favorable towards gender equality, maximum value of 40 (or 1) represents least 
favorable towards gender equality
Figure 21, showing variation by health zone and gender, reveals that overall, respondents in Katana 
and Nyangezi seem to hold slightly less favorable views on gender equality. In all regions except for 
Karisimbi (where this is no significant difference), there is a gender gap in attitudes towards equality 
where women seem to hold more favorable views towards gender equality. 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 37
Figure 21: Gender equitable attitude index, by health zone and gender
Note: Min value of 0 represents most favorable towards gender equality, max value of 40 towards least favorable
6.7 ATTITUDES AND NORMS AROUND GBV
IMA Indicator 3: IR 1.2.4 Percentage of target population that views GBV as less acceptable after 
participating in or being exposed to USG programming. 
The following section assesses the acceptability of GBV among the sample through questions about 
individually held attitudes, as well as questions pertaining to respondents’ beliefs about the attitudes 
and behaviors of others. Note that, as this stage, we cannot report on any change in the perception 
of GBV due to the program (as is implied by the indicator). Rather, we report here on the 
perception levels in the sample at baseline. Once the endline survey is completed, we will capture 
any changes in perception by comparing across villages with and without the program.
On individual attitudes, respondents were asked about the extent to which they agree with the 
following 6 Violence Domain Items of the Gender Equitable Men (GEM) Scale: 
1. There are times when a woman deserves to be beaten.
2. A woman should tolerate violence to keep her family together.
3. It is alright for a man to beat his wife if she is unfaithful.
4. A man can hit his wife if she won’t have sex with him.
5. If someone insults a man, he should defend his reputation with force if he has to.
6. A man using violence against his wife is a private matter that shouldn't be discussed outside
the couple.
We generated an index by averaging the responses to the 6 statements (with each item weighted 
equally) for each individual respondent, which ranges from 0 to 30. A higher score denotes greater 
acceptance of GBV. A respondent would score 0 if they strongly disagree with all statements 
indicating acceptability of GBV and would score 30 if they strongly agree with each one.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 38
Figure 22: Index of acceptability of GBV, by health zone and gender
Note: Min value of 0 represents least accepting of GBV, max value of 30 represents most accepting
As Figure 22 shows, there do not seem to be substantial differences across regions in men’s 
attitudes towards acceptance of GBV. However, in Karisimbi there is no gender gap in accepting 
attitudes and women report generally more accepting attitudes than in the other health zones. In the 
other three health zones, there is a sizable gender gap in attitudes whereby women hold less 
accepting attitudes towards GBV.
In addition, the survey asked respondents to report their beliefs about the attitudes of others in 
their community. Specifically the survey asked each respondent to think of 5 men and 5 women in 
their community and ask them how many of these men and women would agree with the following 
statements: 
• There are times when a woman deserves to be beaten, and
• A man using violence against his wife is a private matter that shouldn't be discussed outside the
couple.
We averaged the responses to construct a measure of each respondent’s perception of other men 
and women’s beliefs. Overall it appears that respondents believe men to be more accepting of GBV 
than women – on average respondents think that 2.2 men but only 1.75 women would agree with 
these statements. At the same time, female respondents seem to perceive both other women and 
other men to be more tolerant of GBV than do male respondents (Table 13).
Table 13: Index of attitudes and norms towards GBV
N MEAN SD MIN MAX
GBV Acceptability (out of 30) 4,223 17.2 5.49 0 30 
GBV Acceptability (rescaled to 0 - 1) 4,223 0.57 0.18 0 1 
GBV Acceptability (men only) 2,109 18.7 5.44 0 30 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 39
N MEAN SD MIN MAX
GBV Acceptability (women only) 2,114 15.6 5.10 0 30 
Norms (men) 4,223 2.03 1.09 0 5 
Norms for men (men only) 2,109 1.78 1.03 0 5 
Norms for men (women only) 2,114 2.29 1.09 0 5 
Norms (women) 4,223 1.75 1.21 0 5 
Norms for women (men only) 2,109 1.78 1.03 0 5 
Norms for women (women only) 2,114 2.29 1.09 0 5 
Note: For acceptability Min value of 0 represents being most acceptable of GBV (strongly agree to all statements), max 
value of 30 (or rescaled to 1.) 
Norms indicate how many out of 5 men/women in a respondent’s surrounding are accepting 
towards GBV. 
6.8 AWARENESS OF GBV-RELATED SERVICES ACCESS 
IMA Indicator 4: IR 2.3.2 Percentage of target population reporting disagreement with 
identified barriers to accessing GBV-related community services 
A key objective of the Tushinde Ujeuri program is to raise awareness and uptake of GBV services.
The baseline study used a 6-point scale to assess possible barriers to access to GBV-related services.
The six barriers were: “Doesn't want family to know”; “No money for transport”; “Doesn't know
where to get help”; “Doesn't want others to know”; “No money for fees”; and “Doesn't think
services will help”. Respondents reported more barriers to seeking services for GBV if the
perpetrator was a family member (2.53) compared to a non-family member (2.36). Interestingly, men
reported more barriers to service seeking than women for both family and non-family perpetrators
(Table 14). Figures 23 and 24 show number of indicated barriers to GBV-services by health zones
and gender. 
Table 14: Barriers to accessing GBV-related services
N MEAN SD MIN MAX
Barriers to GBV service access no family (out of 6) 4,223 2.36 1.93 0 6 
Barriers to GBV service access no family (women only) 2,114 2.26 1.93 0 6 
Barriers to GBV service access no family (men only) 2,109 2.46 1.92 0 6 
Barriers to GBV service access family (out of 6) 4,223 2.53 1.93 0 6 
Barriers to GBV service access family (women only) 2,114 2.39 1.92 0 6 
Barriers to GBV service access family (men only) 2,109 2.67 1.92 0 6 
Note: Indicates how many out of 6 possible barriers to access to GBV-related services are identified by respondents.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 40
Figure 23: Number of indicated barriers to access GBV-related services if perpetrated 
by a family member, by health zone and gender
Figure 24: Number of indicated barriers to access GBV-related services if perpetrated 
by non-family member, by health zone and gender
The most common reasons for not seeking services for GBV perpetrated by a non-family member
were: no money for transport, followed by doesn't think services will help and doesn't know where to get
help (Table 15). For family GBV perpetrators, the most common reason for not seeking services was
no money for transport, followed by no money for fees and doesn’t know where to get help (Table
16). 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 41
Table 15: Barriers to access to GBV-related service if perpetrated by non-family member
MEAN MIN MAX
Doesn't want family to know 0.40 0 1 
No money for transport 0.48 0 1 
Doesn't know where to get help 0.45 0 1 
Doesn't want others to know 0.28 0 1 
No money for fees 0.43 0 1 
Doesn't think services will help 0.47 0 1 
Table 16: Barriers to access to GBV-related service if perpetrated by family member
N MEAN SD MIN MAX
Doesn't want family to know 3776 0.38 0.49 0 1 
No money for transport 3776 0.54 0.50 0 1 
Doesn't know where to get help 3776 0.49 0.50 0 1 
Doesn't want others to know 3776 0.27 0.44 0 1 
No money for fees 3776 0.49 0.50 0 1 
Doesn't think services will help 3776 0.48 0.50 0 1 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 42
7. RISK FACTORS FOR GBV VICTIMIZATION AND
PERPETRATION
This section aims to better understand the factors, such as demographics, war exposure, norms etc., 
associated with the perpetration and victimization of both non-partner sexual violence and IPV. The 
following analysis focuses on three main GBV variables (i) IPV victimization among women (ii) Men’s 
perpetration of IPV and (iii) Men’s perpetration of non-partner sexual violence. 
First, we examine the relationship between the levels of these three GBV outcomes at the village 
level to see whether villages where IPV victimization of women is high also have higher levels of 
perpetration by men, both in terms of IPV and non-partner sexual violence. Indeed, the village-level 
correlations between these measures are high (see Table 17)—in villages where women reported 
higher levels of IPV, men perpetration of IPV and non-partner sexual violence is also higher. The 
village-level correlation between women experiencing IPV and men’s perpetration of IPV is 0.38 and 
between women experiencing IPV and men’s perpetration of non-partner sexual violence is 0.45. 
Both of these correlations are highly statistically significant. It is not surprising that in villages where 
men’s perpetration of IPV is high, men’s perpetration of non-partner sexual violence is also high—
the correlation is 0.55 and is highly statistically significant. These results suggest that there could be 
community-level factors that affect the experiences and behaviors of individuals within the same 
communities in a similar way. These factors could be observed ones such as the demographic 
composition of the population (age distribution, ethnic composition, etc.) or unobserved ones such
as norms regarding and/or tolerance for GBV.
Table 17: Village-Level Correlation Between GBV Measures
IPV VICTIMIZATION 
AMONG WOMEN
MEN’S PERPETRATION 
OF IPV
Men’s perpetration of IPV 0.380*** 
Men’s perpetration of non-partner sexual violence 0.451*** 0.550*** 
Note: *** p-value ≤ 0.01
Next, we explore the correlation between the three GBV outcomes with known contributing 
factors related to individual experiences regarding conflicts, perceptions of social relations, 
demographic characteristics and attitudes, and partner behavior. Given that there are several 
potential contributing factors, we use multivariate regression analysis. 
Specifically, we use a linear probability model with village fixed effect that allows for examining the 
contributing factors by comparing individuals within the same village. This is especially important 
because of the likely village-level effects on the outcome measures described above. In addition, we 
include the following contributing factors in the regression model: war exposure (witnessing or 
experiencing conflict related events), within and between village social relations (ranging from very 
good to very bad), whether a respondent experienced a personal conflict with someone else, 
respondent age, marital status, if primary occupation is farming, religion, whether the respondent 
had abusive parents, whether they have gender inequitable attitudes, and (for female respondents) 
whether husbands use alcohol or drugs.
Finally, we run three regression models, one with each of the three outcome measures as dependent 
variables. We include the same contributing factors, as listed above, for each of the three 
regressions. We present the regression results for women that experienced any form of IPV, IPV 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 43
perpetration by men, and non-partner sexual violence perpetration by men in Columns 1, 2, and 3, 
respectively of Table 18 and describe them below.
7.1 IPV VICTIMIZATION AMONG WOMEN 
The regression model for examining the correlations between IPV victimization among women with 
various contributing factors includes responses from and information on 1,560 female respondents 
from the baseline survey. Of the women in our sample, over 53% report having experienced physical 
or sexual IPV in their lives (Column 1, Table 18). 
Women exposed to war-related experiences were significantly more likely to be victims of IPV 
compared to women who had fewer war-related experiences. Specifically, women who have 
witnessed or experienced conflict events are about 6% more likely to report IPV victimization.
However, both of these correlations are weakly statistically significant, at the 10% level. These 
results highlight how personal experiences with conflict may put women at greater risk of violence 
within the home. 
Some background characteristics, such as age, marital status, and experience with violence in 
childhood are also important contributing factors. Younger women are 2% more likely and women 
in polygamous marriages are 12% more likely to report they have experienced any form of IPV, and 
both correlations are statistically significant at the 5% level. Women who had an abusive father (who 
they witnessed beating their mother or those women who were beaten themselves when young) are 
13% more likely to report IPV victimization and this relationship is highly statistically significant, at 
the 1% level. 
Not surprisingly, women with husbands with alcohol and drug abuse problems are more likely to 
report IPV victimization. The correlations between alcohol and drug use by husbands and IPV 
victimization are quite large and highly statistically significant at the 1% level. Women who report 
their husband drinks are 23% more likely to report having experienced IPV, and those with husbands 
that use drugs are 20% more likely to report having experienced IPV. 
Other contributing factors included in the regression model were not statistically significant. 
The factors we found to be statistically significant align with the existing literature on IPV 
victimization. Globally, risk factors for GBV include younger age, education level, wealth, experience 
of or witnessing of violence previously, and alcohol or substance abuse (Abramsky et al., 2011; 
Hindin et al., 2008; Jewkes, 2013, Linos et al., 2013; WHO, 2013; WHO, 2014). For IPV specifically, 
having more children under five in the household and attitudes condoning domestic violence by 
victim and perpetrator are also risk factors (Jewkes, 2002; WHO, 2013).
7.2 MEN’S PERPETRATION OF IPV 
The regression model for examining the correlations between men’s perpetration of IPV with 
various contributing factors includes responses from and information on 2,030 male respondents 
from the baseline survey. Of the men in our sample, about 29% report to have “hit, slapped, kicked, 
or done anything else to hurt their current or previous wife/husband or girlfriend/boyfriend 
physically” (Column 2, Table 18). 
Similar to the results for women, men who witnessed war-related events are about 9% more likely 
to report perpetration of violence against women, compared to men that did not, suggesting a 
strong link between exposure to political conflict and perpetration of violence within the home. In 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 44
addition, social relations with neighboring villages also seem to be an important factor contributing 
to perpetration of IPV. Men that report worse relations with other (nearby) villages are about 6% 
more likely to report IPV perpetration. Both of these correlations are highly statistically significant at 
the 1% level. Men who had a personal conflict with someone in the year preceding the survey were 
also more likely, by about 6%, to report IPV perpetration and this correlation is statistically 
significant at the 5% level. 
Background characteristics, such as age, marital status, and experience with violence in childhood are 
also important contributing factors. Older men are slightly more likely (by 1%) to report IPV. 
Compared to single men, married men (by about 7%) and those in polygamous marriages (by about 
16%) are more likely to report they hurt their partner(s). Also, men who had an abusive father are 
12% more likely to report IPV perpetration. These results are in line with the finding above that 
women in polygamous marriages are more likely to report experiencing any form of IPV. 
Also, men who hold gender inequitable norms, such as “I strongly agree there are times when a 
woman deserves to be beaten” are 12% more likely to report they hurt their partner and this is 
highly statistically significant at the 1% level. We did not observe similar results for women; 
therefore, women’s attitude did not have any correlations with them reporting IPV victimization. 
This suggests that for perpetrating men, women’s views regarding equitable gender norm is not so 
important. 
Finally, Muslim men are 13% more likely to report IPV perpetration compared to Christian men and 
this is weakly statistically significant at the 10% level. Other contributing factors included in the 
regression model were not statistically significant. 
7.3 MEN’S PERPETRATION OF NON-PARTNER SEXUAL VIOLENCE 
The regression model for examining the correlations between men’s perpetration of non-partner 
sexual violence uses the same sample of the 2,030 male respondents as in the model above. Of the 
men in our sample, 10% answered “yes” to the question “Have you forced someone OTHER THAN 
your current or previous wife/husband or girlfriend/boyfriend to have sex with you when she did not 
want to?” (Column 3, Table 18). These respondents were coded as having perpetrated non-partner 
sexual violence.
Unlike the previous two sets of results, war-related experiences were not a contributing factor for 
men to perpetrate non-partner sexual violence. However, men who report worse relations with 
other (nearby) villages are about 8% more likely to report perpetrating sexual violence. Men who 
had a personal conflict with someone in the year preceding the survey were also about 6% more 
likely to report perpetrating sexual violence. Both of these correlations are highly statistically 
significant. 
Other factors contributing to men reporting perpetrating sexual violence are similar to those that 
we found to have contributed to men’s reports of IPV perpetration: Compared to single men, 
married men are about 5% more likely to report they perpetrated sexual violence, while Muslim 
men are about 13% more likely compared to Christian men, both statistically significant at the 5% 
level. Also, men who had an abusive father are about 7% more likely and men who hold gender 
inequitable norms, such as “I strongly agree there are times when a woman deserves to be beaten” 
are about 9% more likely to report sexual violence, both results being highly statistically significant at 
the 1% level.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 45
Table 18: Correlates of IPV
(1)
EXPERIENCED 
ANY FORM OF 
IPV (EVER 
MARRIED 
FEMALE)
(2)
IPV 
PERPETRATION 
(MEN)
(3)
SEXUAL 
VIOLENCE 
PERPETRATION 
(MEN)
Witnessed Conflict Events (1=yes) 0.056* 
(0.032) 
0.088*** 
(0.025) 
0.029 
(0.020) 
Experienced Conflict Events (1=yes) 0.062* 
(0.033) 
-0.035 
(0.025) 
-0.004 
(0.020) 
Inter-village relations are bad (1= very good – 5 = 
very bad)
0.036 
(0.027) 
0.004 
(0.021) 
0.009 
(0.017) 
Intra-village relations are bad (1= very good – 5 = 
very bad)
-0.015 
(0.027) 
0.057*** 
(0.022) 
0.075*** 
(0.017) 
Own village is worse off than neighboring villages 
(1= much better – 5 = much worse)
0.007 
(0.011) 
0.005 
(0.010) 
-0.012 
(0.008) 
Personal Conflict (last year) (1=yes) 0.034 
(0.032) 
0.057** 
(0.024) 
0.058*** 
(0.019) 
Age (years) -0.002** 
(0.001) 
0.001* 
(0.001) 
0.000 
(0.001) 
Married/cohabiting (1=yes) 0.074*** 
(0.027) 
0.053** 
(0.022) 
Married polygamous (1=yes) 0.117** 
(0.046) 
0.155*** 
(0.046) 
0.029 
(0.036) 
Divorced or widowed (1=yes) -0.006 
(0.036) 
0.065 
(0.063) 
0.079 
(0.049) 
Main occupation if farming (1=yes) 0.016 
(0.032) 
-0.036 
(0.023) 
0.011 
(0.018) 
Protestant (1=yes) 0.049 
(0.030) 
-0.030 
(0.022) 
0.012 
(0.018) 
Muslim (1=yes) 0.046 
(0.120) 
0.130* 
(0.074) 
0.134** 
(0.059) 
Other (1=yes) 0.054 
(0.045) 
0.034 
(0.034) 
0.032 
(0.027) 
Abusive father when young (1=yes) 0.133*** 
(0.026) 
0.123*** 
(0.024) 
0.068*** 
(0.019) 
Gender inequitable attitude (1=yes) -0.028 
(0.036) 
0.122*** 
(0.021) 
0.085*** 
(0.017) 
Husband uses alcohol (1=yes) 0.232*** 
(0.027) 
Husband uses drugs (1=yes) 0.199*** 
(0.054) 
Respondent gender FEMALE MALE MALE
Village Fixed Effects YES YES YES 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 46
(1)
EXPERIENCED 
ANY FORM OF 
IPV (EVER 
MARRIED 
FEMALE)
(2)
IPV 
PERPETRATION 
(MEN)
(3)
SEXUAL 
VIOLENCE 
PERPETRATION 
(MEN)
Observations 1,560 2,030 2,030 
Mean of Dependent Variable 0.534 0.285 0.106 
Notes: i) OLS regressions; ii) significance levels indicated by * p < 0.10, ** p <0.05, *** p < 0.01; ii) standard errors in 
parenthesis; iii) specifications include village fixed effects; iv) Column 1, female respondents only; v) Column 2 and 3, male respondents only. 
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 47
8. CONCLUSIONS
A foundational goal of this research was to understand the baseline levels of different forms of GBV 
that affect communities in the project areas. The survey asked a number of questions to assess GBV 
perpetration and GBV victimization. For victimization, the survey assessed lifetime IPV, past-year IPV 
for women, as well as lifetime SV and past-year SV for women and men. For perpetration, the survey 
assessed perpetration of IPV and SV among women and men. 
Over half of all ever-partnered women surveyed reported having experienced IPV at some point in 
their lives, a number that is well above the regional average for Sub-Saharan Africa. Over a third of 
women (38.9%) reported experiencing IPV in the past 12 months. The survey also examined women 
and men’s experiences with sexual violence committed by a non-partner. One-fifth of women and 
one in twelve men reported being ever raped by a non-partner. Again, these numbers are much 
higher than the regional victimization rate for women in Sub-Saharan Africa. In the past year, 14.6% 
of women and 6.4% of men reported experiencing sexual violence. 
Almost one-third of men and 13% of women report perpetrating physical IPV against a partner in 
their lifetime. Roughly one in seven men has reported ever forcing someone other than their partner 
to have sex against their will, and 4% of men reported doing so in the past year. Just over six 
percent of women report ever forcing someone to have sex against their will, and 3% reported 
doing so in the past year.
Another core goal of this research was to better understand the norms and beliefs related to GBV 
and GBV help seeking in this area. The survey assessed individuals’ own attitudes toward the 
acceptability of GBV, as well as individuals’ beliefs about the attitudes of others in their community. 
In general, there exists a sizable gender gap in attitudes; women hold less accepting attitudes 
towards GBV while men have more permissive attitudes. Overall it appears that respondents believe 
men to be more accepting of GBV than women. At the same time, female respondents seem to 
perceive both other women and other men to be more tolerant of GBV than do male respondents.
Related to GBV help-seeking, overall men reported more barriers to accessing GBV services on 
average than women. The most common barriers to accessing services differed slightly for family 
versus non-family perpetrators of GBV. However, in general, not having money for transport or for 
services fees, and not knowing where to access services were common problems. 
There are stark gendered differences in many of the outcomes assessed for this project, including 
but not limited to: access to education, access to income-generating activities and resilience and 
mental health outcomes. There is also substantial variation among health zones in a number of 
outcomes, including literacy, employment, mental health outcomes and perceptions of security. 
While all health zones included in this survey have been highly affected by conflict, Walikale stood 
out as a site with higher levels of displacement, lower levels of literacy, and higher levels of 
depression and anxiety than other health zones. There were significant geographic and gendered 
differences among many outcomes, suggesting the importance of disaggregating this data to gain a 
more nuanced picture of how outcomes may vary across populations.
There are strong and significant links between the project outcomes – IPV and SV – and exposure to 
war-related violence. Women that witnessed or experienced conflict are 6% more likely to report 
IPV victimization. These results highlight how personal experiences with conflict may put women at 
greater risk of violence within the home. Men who witnessed war related events were more likely 
to report perpetration of IPV and SV, suggesting a strong link between exposure to political conflict 
and perpetration of violence within the home.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 48
In addition, there were strong links between certain measures of poor social cohesion and men’s 
perpetration of different forms of IPV. Men who report worse relations with other (nearby) villages 
are more likely to report perpetrating SV and IPV. 
Taken together, these results suggest that violence within communities – including IPV and other 
forms of GBV – are highly associated with individuals’ experiences with the conflict, and are also 
associated with measures of social cohesion. Addressing the underlying trauma of conflict and trying 
to rebuild trust and social cohesion may provide an avenue to addressing some of the underlying 
mechanisms that propagate violent behaviors in highly war-affected areas.
We summarize the key findings from the baseline survey below:
• Exposure to witnessing or experiencing war-related abuses was extremely common. More than 
half of respondents reported having witnessed or experienced at least one conflict-related event.
• Individuals participating in the research had high levels of depression and anxiety as well as 
PTSD. Across all health zones, a significantly larger proportion of women report anxiety and 
depression and PTSD than men.
• Over half of all ever-partnered women surveyed reported having experienced IPV at some point 
in their lives. One-fifth of women and one in twelve men reported being ever raped by a non￾partner. These numbers are higher than the regional victimization rate for women in Sub￾Saharan Africa. 
• Men who witnessed war-related events were more likely to report perpetration of IPV and SV.
• War-related experiences are highly and significantly associated with higher odds of women 
reporting IPV victimization.
• Men who reported more mistrust and resentment of neighboring communities also had higher 
odds of perpetrating IPV and sexual violence against non-partners.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 49
REFERENCES 
Abramsky, T., Watts, C. H., Garcia-Moreno, C., Devries, K., Kiss, L., Ellsberg, M., Jansen, H., & 
Heise, L. (2011). What factors are associated with recent intimate partner violence? Findings from 
the WHO multi-country study on women’s health and domestic violence. BMC Public Health, 
11(109).
Amisi, Christine, Rosalie Biaba, Aline Cikara, Ragnhild Nordas, Gudrun Ostby, John Quattrochi and 
Siri Aas Rustad. 2016. Social Inclusion of Survivors of Sexual Violence: How Do Support Programs 
Work? Micro-level evidence from eastern Democratic Republic of the Congo (DRC)." PRIO Paper
Bandura, A. (1982). Self-efficacy mechanism in human agency. American Psychologist, 37(2), 122-147.
Bass, J.K., Annan, J., Murray, S.M., Kaysen, D., Griffiths, S., Cetinoglu, T., Wachter, K., Murray, L.K., & 
Bolton, P.A. (2013). Controlled trial of psychotherapy for Congolese survivors of sexual violence. 
New England Journal of Medicine, 368(23), 2182–2191.
Carey, M.P. & Forsyth, A.D. (n.d.) Teaching Tip Sheet: Self-Efficacy. Retrieved from 
https://www.apa.org/pi/aids/resources/education/self-efficacy. 
Cilliers, J., Dube, O., & Siddiqi, B. (2016). Reconciling after civil conflict increases social capital but 
decreases individual well-being. Science, 352(6287): 787-794.
Ellsberg, M., Arango, D.J., Morton, M., Gennari, F., Kiplesund, S., Contreras, M., & Watts, C. (2015). 
Prevention of violence against women and girls: what does the evidence say? The Lancet, 385(9977), 
1555-1566.
Fulu, E., Warner, X., & Moussavi, S. (2013). Why do some men use violence against women and 
how can we prevent it?: Findings from the UN multi-country study on men and violence in Asia and 
the Pacific. Phnom Penh: UN Women Cambodia.
Garcia-Moreno, C., Pallitto, C., Devries, K., Stockl, H., Watts, C., & Abrahams, N. (2013). Global and 
regional estimates of violence against women: prevalence and health effects of intimate partner 
violence and non-partner sexual violence. Geneva, Switzerland: World Health Organization, 
Department of Reproductive Health and Research.
Gracia, E. (2014). Intimate partner violence against women and victim-blaming attitudes among 
Europeans. Bulletin of the World Health Organization, 92(5), 380-381.
Hindin, M.J., Kishor, S., & Ansara, D.L. (2008). Intimate partner violence among couples in 10 DHS 
countries: Predictors and health outcomes. DHS Analytical Studies, 18. 
Jewkes, R. (2002). Intimate partner violence: causes and prevention. The Lancet, 359(9315): 1423-
1429. 
Jewkes, R. (2013). Intimate partner violence: the end of routine screening. The Lancet, 382(9888): 
190-191. 
Kelly, J., Khan, S. & Voors, M. (2018). Impact Evaluation Design: Changing Norms around Gender 
Based Violence in Eastern Democratic Republic of Congo. Bethesda, MD: NORC at the University 
of Chicago.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 50
Kennedy, C., & Mann, C.B. (2017). RANDOMIZE: Stata module to create random assignments for 
experimental trials, including blocking, balance checking, and automated rerandomization. Statistical 
Software Components S458028, Boston College Department of Economics.
Linos, N., Slopen, N., Subramanian, S.V., Berkman, L., & Kawachi, I. (2013). Influence of community 
social norms on spousal violence: A population-based multilevel study of Nigerian women. American 
Journal of Public Health, 103(1): 148-155.
NORC at the University of Chicago. (2018). Philippines Trafficking in Persons: Knowledge, 
Awareness, and Victim Identification Survey Final Report. Bethesda, MD: NORC at the University of 
Chicago. 
Nordas, R., Ostby, G., Rustad, S.A., Amisi, C., Biaba, R., Quattrochi, J., Allden, S., & Cikara, A. 
(2016). Social inclusion of survivors of sexual violence: How do support programs work? Micro-level 
evidence from eastern Democratic Republic of the Congo (DRC). PRIO Paper. Oslo: PRIO.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 51
APPENDIX A: INTERVENTION TIMELINE
We include the expected timeline of the impact evaluation below. A key element of the impact 
evaluation is to include the collection of data, as part of the data collection for the impact analysis, 
which will contribute to the monitoring and evaluation needs of the implementing partners. 
• September 2017-2018. Develop research design with partners
• September 2018. Finalize research design (see Kelly, Khan and Voors (2018) Impact Evaluation 
Design: Changing Norms around Gender Based Violence in Eastern Democratic Republic of 
Congo) 
• September/October 2018: Baseline data collection
• October 2018: Assign Treatment (blocked cluster randomization, using replicable procedure in 
STATA/R)
• January 2019: Program CBTH activities commence by SFCG and partners 
• Continuous: Monitoring
• Spring 2020: Endline data collection
• Spring 2020: Analyses and impact report and results presentations
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 52
APPENDIX B: MEMO ON POPULATION-LEVEL ESTIMATES OF 
THE BASELINE REPORT
July 9, 2019
PURPOSE AND NOTES ON METHODOLOGY
This memo reports population-level estimates for certain key indicators on which IMA relies for its
M&E activities. These IMA indicators were originally presented and discussed in Sections 6.5 to 6.8 
of the baseline report for the sample households included in the impact evaluation study. For the 
baseline study, we randomly selected 20 households each from 192 villages, irrespective of village 
population size, from 48 Health Areas (HAs) in 4 Health Zones5 (HZs). In this memo, we 
reproduced the IMA indicators by weighting the village-level estimates appropriately to reflect that 
the responses to baseline survey represent different proportions of population in different villages. 
As such, the IMA indicators reported in this memo are the population averages for the 
192 villages that were included in the study. Note that we can “weight up” the sample 
averages reported in the baseline study to population averages of the 192 villages because the study 
sample is a random sample of the population in the 192 villages included in the study.6
However, note that this approach does not provide population averages for either the North Kivu 
and South Kivu Provinces as a whole or the four HZs as a whole or the 48 HAs as a whole in which 
the 192 villages surveyed are located. This is because the selection of villages within the HAs were 
not random—the criteria for the villages to be included in the sampling frame were accessibility, 
relative security, population size and household count, and a limited presence of other projects that 
might contaminate any impact of the intervention. This has two implications:
(1) We do not have the statistical ability to weight the sample estimates to all villages (192 plus 
the ones that are not selected) in the 48 HAs, and
(2) The population-level estimates presented in this memo for the 192 villages may not 
represent the population-level estimates for all villages (192 plus the ones that are not 
selected) in the 48 HAs had we been able to calculate them. This is specifically likely because 
of the non-randomness in the village selection—the villages that were not selected for the 
study were less accessible and less secure and thus are likely to have worse outcomes on 
GBV.
Nonetheless, absent the appropriate sampling strategy, the estimates presented in this memo are the 
best available population-level estimates for the 192 villages in the 48 HAs in the 4 HZs in South 
Kivu and North Kivu Provinces. We recommend that IMA exercises caution in using these 
estimates. Specifically, given the possibilities in #2 above, the estimates presented here are likely to 
be lower-bound estimates compared to the estimates that would include all villages (192 plus the 
ones that are not selected). 
5 Namely Karisimbi, Walikale, Katana, and Nyangezi.
6 The weight is calculated as the inverse probability of being selected for the survey in the village. This 
probability is calculated by dividing the sample size in each village by the population size of that village. 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 53
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
TABLES AND CHARTS FOR POPULATION-LEVEL ESTIMATES 
IMA Indicator 1: G1.1 Prevalence rate of GBV among women, men, and children in 
target geographic areas
Figure 1. Proportion of ever-partnered women having ever experienced IPV 
(physical or sexual), by health zone, at the population-level7
Figure 2. Proportion of ever-partnered women having experienced IPV (physical 
or sexual) during the last 12 months, by health zone, at the population-level8
7 Population-level version of Figure 19 in the baseline report. 8 Population-level version of Figure 20 in the baseline report.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 54
IMA Indicator 2: IR 1.2.1 Percentage of target population reporting increased 
agreement that males and females should have equal access to social, economic and 
political opportunities
Table 1. Index of attitudes towards gender equality in different domains (population-level)9
MEAN SE MIN MAX 
Gender equitable attitudes (out of 40) 20.1 0.13 0 40
Gender equitable attitudes (rescaled to 0-1) 0.50 0.00 0 1 
Gender equitable attitudes (men only) 21.4 0.19 0 40
Gender equitable attitudes (women only) 18.7 0.15 0 39
Note: Minimum value of 0 represents most favorable towards gender equality, maximum value of 40 (or 1) represents least 
favorable towards gender equality
Figure 3. Gender equitable attitude index, by health zone and gender, at the 
population-level10
Note: Min value of 0 represents most favorable towards gender equality, max value of 40 towards least favorable
9 Population-level version of Table 12 in the baseline report. 10 Population-level version of Figure 21 in the baseline report.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 55
IMA Indicator 3: IR 1.2.4 Percentage of target population that views GBV as less 
acceptable after participating in or being exposed to USG programming. 
Figure 4. Index of acceptability of GBV, by health zone and gender11
Note: Min value of 0 represents least accepting of GBV, max value of 30 represents most accepting
Table 2. Index of attitudes and norms towards GBV (population-level)12
N MEAN SE MIN MAX
GBV acceptability (out of 30) 4223 17.4 0.11 0 30 
GBV Acceptability (rescaled to 0 - 1) 4223 0.6 0.00 0 1 
GBV acceptability (men only) 2109 18.9 0.16 0 30 
GBV acceptability (women only) 2114 15.8 0.15 0 30 
Norms (men) 4223 2.0 0.02 0 5 
Norms for men (men only) 2109 1.8 0.03 0 5 
Norms for men (women only) 2114 2.3 0.03 0 5 
Norms (women) 4223 1.7 0.02 0 5 
Norms for women (men only) 2109 1.8 0.03 0 5 
Norms for women (women only) 2114 2.3 0.03 0 5 
For Acceptability, Min value of 0 represent being most acceptable of GBV (strongly agree to all 
statements), max value of 30 being least acceptable of GBV (strongly disagree to all statements). 
Norms indicate how many out of 5 men/women in a respondents surrounding are acceptable 
towards GBV.
 
11 Population-level version of Figure 22 in the baseline report. 12 Population-level version of Table 13 in the baseline report.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 56
IMA Indicator 4: IR 2.3.2 Percentage of target population reporting disagreement with 
identified barriers to accessing GBV-related community services
Table 3. Barriers to accessing GBV-related services (population-level)13
N MEAN SE MIN MAX
Barriers to GBV service access no family (out of 
6) 4223 2.43 0.04 0 6 
Barriers to GBV service access no family 
(women only) 2114 2.38 0.06 0 6 
Barriers to GBV service access no family (men 
only) 2109 2.48 0.06 0 6 
Barriers to GBV service access family (out of 6) 4223 2.60 0.04 0 6 
Barriers to GBV service access no family 
(women only) 2114 2.51 0.06 0 6 
Barriers to GBV service access no family (men 
only) 2109 2.70 0.06 0 6 
Indicates how many out of 6 possible barriers to access to GBV-related services are identified by respondents
Figure 5. Number of indicated barriers to access GBV-related services if perpetrated by 
family member, by health zone and gender, at the population-level14
 
13 Population-level version of Table 14 in the baseline report. 14 Population-level version of Figure 23 in the baseline report.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 57
Figure 6. Number of indicated barriers to access GBV-related services if perpetrated by 
non-family member, by health zone and gender, at the population-level15
Table 4: Barriers to access to GBV-related service if perpetrated by non-family member
N MEAN SE MIN MAX
Doesn't want family to know 3734 0.39 0.01 0 1 
No money for transport 3734 0.50 0.01 0 1 
Doesn't know where to get help 3734 0.46 0.01 0 1 
Doesn't want others to know 3734 0.27 0.01 0 1 
No money for fees 3734 0.45 0.01 0 1 
Doesn't think services will help 3734 0.48 0.01 0 1 
Table 5: Barriers to access to GBV-related service if perpetrated by family member
N MEAN SE MIN MAX
Doesn't want family to know 3776 0.38 0.01 0 1
No money for transport 3776 0.57 0.01 0 1
Doesn't know where to get help 3776 0.51 0.01 0 1
Doesn't want others to know 3776 0.28 0.01 0 1
No money for fees 3776 0.50 0.01 0 1
Doesn't think services will help 3776 0.49 0.01 0 1
 
15 Population-level version of Figure 24 in the baseline report.
Contract No. GS-10F-0033M/AID-0AA-M-13-00013 
USAID.GOV Countering Gender Based Violence in Eastern Democratic Republic of Congo | 58
U.S. Agency for International Development
1300 Pennsylvania Avenue, NW
Washington, DC 20523