Refugees and host population report similar levels of human security in terms of attachment to home, community, and a positive grasp of the future. Though some human security indicators differ between these populations, overall they reflect a relatively stable population at the time of this study, where refugees not only feel safe but also have access to land, livelihoods, clean water, shelter, and wish to remain in their newly adopted villages for many years to come. Refugees appeared to be able to attain stability and security in Eastern Cameroon—in fact, though the proportion of refugees that owned land was lower than the proportion of host population that owned land, refugees and Cameroonians who owned land owned in similar amounts, suggesting equal opportunity to grow livelihoods between the two groups. Similarly, refugees had equal levels of assets as many host population households. Refugees and Cameroonians were both similarly attached to their villages, desiring to stay in the same village well into the future.
However, despite the unusual and encouraging stability and egalitarian quality of these human security findings, these human security indicators missed an epidemic of sexual violence that endangered the lives and health of refugee and host population women. ROC analysis shows that human security indicators measured in this study did not uncover either lifetime or six-month sexual violence. These data suggest that current, gender-blind means of describing human security are missing serious threats to the safety of one half of the population. If human security is ‘concerned with the protection of people from critical and life-threatening dangers’ , this presents a major gap.
Why is it crucial that measures of human security be sensitive to gender? A robust body of evidence has explicitly linked measures of gender inequity and violence against women, including intimate partner and domestic violence, to higher levels of conflict [8, 12, 26–29]. Much of this work has been done by Caprioli, whose analyses have shown that increased domestic gender equity has a pacifying effect on state behavior at an international level , and that states with higher levels of gender equality are less likely to resort to violence first in resolving international disputes . Caprioli went on in 2005 to link lower levels of gender equity to higher rates of internal conflict , and more violent conflict . Melander reproduced these findings, demonstrating that societies with higher female representation in parliament and higher female to male education attainment ratios have lower levels of internal conflict . There are several theories as to why this may be true. Some suggest that a tendency toward peaceful behavior is associated with a tolerance of the rights of others , while others cite that states with a tendency toward inequity and oppression are inherently likely to be more violent both internally and internationally [8, 11]. Violence against women not only has implications for the human security of half of the world’s population, it has consequences for the stability of the state.
Human security indicators represent current feelings of respondents, and thus are not likely to be sensitive to distant events. Lifetime sexual violence could have occurred any time over the respondent’s lifetime, and in either CAR or Cameroon, making it difficult to interpret the meaning of high human security indicators in the context of a high prevalence of lifetime sexual violence. However, human security indicators were also not associated with any meaningful association with sexual violence during the past six months, suggesting an insensitivity of this model to a major threat to the security and health of women and girls. Further, lifetime and six month prevalence for sexual violence reported by both women in the CAR and Cameroon were similar, and both proportions were substantial. Sexual violence represents one extreme manifestation of gender inequity—this level of violence is likely to be associated with other manifestations, including limited access to education and access to justice. Low levels of literacy, low rates of completion of primary and secondary school education, poor access to justice for survivors of sexual violence, and high rates of early marriage were outlined in a previous publication focused on sexual violence findings of this study . Thus, sexual violence against women represents a proxy indicator—it suggests that gender inequity has real human security consequences in this population, consequences that are not captured by this quantitative model, and consequences that extend beyond sexual violence itself.
These data also confirm what was described in the report of the Human Security Research Group published in 2012 . This report outlined data suggesting that rape as a weapon of war, while important in many contexts, does not constitute the majority of rape faced by women in conflict. The findings of this report support this assertion [Tables 8 and 9], as the majority of women who experienced sexual violence reported that the perpetrator was their intimate partner (64.0%, 95% CI 54.3-72.5) and the second most common perpetrator was a friend or member of the community (20.0%, 95% CI 14.2-27.5). The group identified as ‘combatants’ - soldiers, rebels, and coupers de route— who, in most conflict contexts, would be the focus of sexual violence related insecurity—constituted a comparatively less commonly reported perpetrator (17.1%, 95% CI 10.7-26.1). When disaggregated data is examined [Tables 8 and 9], data show that while refugees report a higher proportion of sexual violence at the hands of combatants (39%, 95% CI 25.6-54.2) than Cameroonian women, refugees still report that a majority of the perpetrators of sexual violence are known to them (intimate partners: 52%, 95% CI 35.4-68.1: friends/community members: 15.6%, 95% CI 8.3-27.2). Based on these findings, we estimate 22.3% (95% CI 17.1-28.5) of all women in Djohong District, Cameroon have endured sexual violence by their intimate partner during their lifetimes. Several authors have reported high rates of sexual violence in times of conflict [31–34], and as suggested by the Human Security Research Group, much of that occurs at the hands of intimate partners and community members. While a focus on sexual violence perpetrated by armed forces remains important, the much more common epidemic of violence perpetrated by intimate partners and community members against women must also be addressed, particularly given that this violence presents risks not only for the affected women but for the development and maintenance of their communities and the world as a whole.
Several measures have been developed to understand gender equity at the country level, including UN Women’s Indicators and Statistics Database, GenderStats, and the World Economic Forum’s Gender Gap Project, and many studies analyze other measures including female to male education attainment ratio, female representation in parliament, life expectancy, literacy, and participation in democracy. However as Hudson points out, these do not include measures of physical violence against women . These are also aggregated sources at the country level, and not likely to aid in a nuanced understanding of community level human security, as is the goal of the Leaning and Arie model.
Based on the findings of this study, we suggest that, at a minimum, measures of human security should include prevalence of sexual violence at the hands of all types of perpetrators, including both combatants and non-combatants. Additionally, measures of gender inequality that have been previously linked to conflict in other research (female participation in parliament and female to male higher education attainment ratio) should be included in human security frameworks.
All interviews were conducted privately in their homes in the respondent’s preferred language. Nuanced understanding of questions at the household level could be distorted in the course of translation and risked being asked differently in Fulfilde across data collectors. Though all efforts were made to ensure the respondents privacy at the time of the interview, it is possible that women felt reluctant to report sexual violence given strong social stigma associated with sexual violence. Whenever possible, respondents were interviewed by women, however, several members of the surveying team were male, due to limitations in availability of experienced multilingual staff. Although the few male data collectors had prior experience in medical and psychological intake of sexual violence survivors, it is possible that women felt less comfortable reporting their experiences with sexual violence to them regardless. We proceeded with male data collectors based on the acceptance in this community of male sexual violence counselors in a pre-existing international NGO program on sexual violence. Although female data collectors are recommended according to WHO guidelines , several studies have successfully engaged male data collectors in similar research [19, 33]. As all data reported here refer to the respondent’s last episode of sexual violence, it is likely that these statistics under-represent the true burden of sexual violence in this community. Results are specific to refugee and host population female heads of household over the age of 18 only, thus results cannot be generalized to all women in the region and are likely biased towards the experiences of older women. Given these female heads of household tend to be first wives in polygamous households, this study represents the experiences of this group of women primarily. Arguably any form of sexual violence could affect human security and data was not collected on men or children.
Several of the data collectors were NGO sexual violence program staff and thus known to be affiliated with these programs. Although interviewers were carefully trained to emphasize that no aid or compensation would be given for participation in the survey, it is possible respondents may have altered their responses. Respondents may either have reported themselves to be refugees in the interests of obtaining services, or Cameroonians in order to avoid stigma. Given the anonymous nature of this survey and initial disclosure regarding aid and compensation, we do not feel it is likely that this issue significantly impacted responses.
Human security indicators were developed by the research team based on the Leaning-Arie model, used during a 2009 iteration of a population based survey in the region and refined based on the results. Outcomes are self-reported, thus assets, income, and land holdings were not verified objectively.