When discussing the measurement of sexual- and gender-based violence (SGBV) prevention in humanitarian settings, reactions often polarize around two main arguments: that measurement is impossible (“How can we measure what hasn’t happened?”), or that prevalence alone signals success (“Do you have data showing decreased perpetration?”). Both perspectives overlook the complexity of SGBV prevention – and the valuable insights gained from measuring it through alternative approaches.

In this post, Zuleyka Piniella, ICRC’s Global Coordinator for the Prevention of Sexual Violence Programme, and Jessica Lenz, Senior Protection Advisor for InterAction, explore new ways to capture the nuanced impacts of prevention efforts, moving beyond prevalence data to understand what truly drives change.

Researchers agree: measuring what didn’t happen is extremely challenging. Yet often, it’s not prevention work itself, but the design of programs that obstructs evaluation. Without specific, well-defined risk assessments, prevention programs become impossible to evaluate. Effective prevention of sexual violence, in turn, requires a precise understanding of the factors fuelling it.

Prevalence studies – a tool to measure the scope, scale and patterns of a population or community that is affected by a specific condition, event, or behaviour at a given point in time or over a specified period – gained prominence from the 1950s onwards, and for very good reasons. In the field of public health, they quickly became a key tool to understand how infectious diseases spread and to assess the impact of prevention measures, such as vaccination programs, sanitation improvements, or awareness campaigns. From the 1990s, their use expanded beyond purely medical conditions, to include social issues like sexual- and gender-based violence.

While prevalence studies provide valuable data, they have limitations when applied to sexual violence and other violations of international humanitarian law (IHL) in armed conflict. Severe barriers – like underreporting, legal narrow definitions and sampling restrictions – create incomplete data. For example, torture of a sexual nature has been extensively documented and associated to detention in relation to conflict; however, it can be rendered invisible under the wider label of “torture”. Certain populations, such as men, boys, and LGBTIQ+ individuals, often go unaccounted for. Regions controlled by specific groups remain inaccessible (26% of the armed groups globally fully control their territory), and fear of retaliation may prevent survivors from disclosing information.

Conflict dynamics further skews data, as sexual violence can surge around certain moments (for example around ceasefires in Liberia in 2003 or in Ethiopia in 2022) or due to war tactics (such as using sexual violence to force displacement of entire communities expanding control towards new territories). Conversely, it can decrease in certain phases, but these variations are not necessarily linked to prevention efforts, nor do they represent a change in the motivation to perpetrate.

Figure 1. Comparison of incidents vs examinations in Ethiopia in the period September 2020-July 2023. The graphic shows the increase of incidents in the months preceding and following the ceasefire. Prevalence studies usually take a very strict approach to timelines and can easily misguide the interpretation of perpetration rates. Source: Physicians for Human Rights (2023) Broken promises. Conflict-related sexual violence before and after the Cessation of Hostilities Agreement in Tigray, Ethiopia.

Global underreporting and the extreme low capacity to identify cases of sexual violence even in stable settings – only 2.17% of cases are recorded annually in Europe – further complicates data accuracy. ICRC’s Stigma Impact Model shows the multi-layered effects of victim-blaming beliefs, attitudes, policies, and practices. Consultation with victims/survivors of sexual violence supported by the ICRC has made clear that stigma is the main barrier for disclosing and reporting. Efforts to reduce stigmatisation may improve reporting without reflecting a drop in violence, evident in countries like Sweden, which has higher reporting rates than the surrounding countries due to major legislation reforms, intensive media campaigns, and better victim support.

Given these factors, prevalence studies, while insightful, are not reliable for assessing prevention outcomes. Relying on prevalence as the main success metric risks misguiding prevention efforts.

If not prevalence, then what?

Some prevention efforts target long-term social norm changes at community, institutional, or societal levels. However, humanitarian teams often face significant barriers, such as short program timelines, limited population access, and ethical issues with collecting SGBV data during crises. These factors complicate the collection, analysis, and attribution of SGBV prevention outcomes to specific program activities, especially without control groups to verify causality.

Recognizing these challenges, humanitarian actors identified the need for alternative approaches to measure SGBV prevention safely, ethically, and effectively. In 2019, InterAction led an interagency initiative to map SGBV risk patterns, interventions, and tools used to track risk changes and prevention outcomes in crisis contexts.

The study found that many prevention programs relied on generalized SGBV risk analyses rather than context-specific analysis, leading to theories of change that didn’t fully address local nuances. Consequently, projects often focused on broad risks or solely on risk mitigation (i.e., safe programming, for example making sure lights and locks are on latrines), rather than specific prevention strategies suited to the actual context. Evaluability issues were also prevalent, with common problems including:

• Vague program designs lacking theories of change or clarity in targeting.
• Unclear articulation of intended outcomes for SGBV prevention.
• Missing baseline data and limited monitoring to address data gaps.
• Insufficiently disaggregated data reflecting unique vulnerabilities in context-specific SGBV risks.

These findings informed the development of the GBV Prevention Evaluation Framework (GBV PEF), launched in 2021. This interagency-tested framework aids humanitarian actors in analyzing, designing, implementing, and measuring GBV prevention outcomes. At its core is the “risk equation” tool endorsed by ICRC Professional Standards and the IASC Principles, supporting the Centrality of Protection in SGBV prevention.

Applying the risk equation helps avoid assumptions about how SGBV manifests within a community by focusing on critical aspects of the SGBV context: the specific threats (e.g., types of perpetrators and their behaviours), community vulnerabilities to the specific threat, and existing capacities that can reduce SGBV risk in the specific context. Contextual use of the risk equation allows for disaggregated analysis based on gender and sexuality and can help identify any harmful traditional practices or norms that contribute to risk, countering any preconceived notions about SGBV’s presence within the community.

For example, InterAction’s action-based research initiative in Somalia that looks at the links between conflict-induced food insecurity and protection, used the risk equation, alongside additional analysis tools, with affected communities to better identify and analyze risks within the community. A conflict-related sexual violence risk emerged that identified female herders experiencing sexual violence by militia. The use of the risk equation allowed humanitarians to better understand how the risk was manifesting and who was most vulnerable. The example below is a simplified depiction of the analysis.

The use of the risk equation has demonstrated that continuous application of this tool also enables responsive analysis as crisis dynamics evolve, affecting SGBV threats, vulnerabilities, and mitigation capacities.

How does this improve understanding and measurement of SGBV prevention?

By using the risk equation, we shift focus from prevalence or incidence data towards the components of SGBV risk: threat, vulnerability, and capacity to counteract SGBV risks. This component-based approach lets us track changes in behaviours of the threat and behaviours of those vulnerable to the threat to understand if threats are decreasing and if community capacities to counter them are strengthening. Analyzing all three dimensions provides a clearer picture of the SGBV risks and what is necessary to prevent or mitigate them.

The risk equation also serves as a tool for designing context-specific theories of change (ToC). These theories outline pathways by which behaviour of threat actors can change, vulnerabilities to threats can be reduced, and existing protective capacities can be enhanced. Validated by communities, these pathways clarify the interdisciplinary actions required to effectively reduce SGBV risk.

Unlike broader organizational theories of change, the ToC within the GBV PEF is designed to be highly context-specific, aligning directly with the risk equation’s SGBV protection analysis. With clearly defined outcomes, this tool enables measurement of meaningful results at the outcome level rather than merely tracking activities or outputs. It highlights the importance of examining all three risk components – threat, vulnerability, and capacity – to achieve effective SGBV prevention outcomes.

Using the risk equation to break down the component parts of SGBV allowed humanitarians and the community to identify specific pathways to bring about change. It also helped to explore how different actors and disciplines are relevant in helping to solve the problem to change the behaviour of the threat actor, decrease vulnerabilities to the threat, and enhance the capacities. The graph below is a simplified version of the context-specific theory of change that illustrates the pathways along the lines of reducing the threat, reducing vulnerability, and enhancing capacities–all contributing to intermediate results that aim to achieve SGBV outcomes.

Armed with this improved analysis and targeted ToC, humanitarians are better equipped to utilize outcome-level measurement tools that capture changes in risk patterns, paving the way for more impactful SGBV prevention outcomes.

Breaking down how behaviour changes

By using context-specific ToC, measuring SGBV prevention need not focus solely on what hasn’t happened; it should, and must, capture the concrete changes that signal progress. In frontline SGBV prevention in conflict settings, these are often behavioural. Such changes are observable, recordable, and measurable.

When ToC are further broken down through a behavioural change model, this perspective provides an inventory of enabling beliefs and practices that define the prevention scope. Tracking how these elements shift in response to interventions is a key method for measuring prevention.

But this approach is only part of the picture. For successful prevention of conflict-related sexual violence, as with other IHL violations, understanding and reinforcing sources of restraint is as essential as addressing enabling factors.

“There is always one.”

ICRC staff often hear this phrase when supporting survivors of sexual violence, referring to the one armed actor who refrained from violence, urged others to stop, or helped victims escape. ICRC’s study, The Roots of Restraint in War, emphasizes that “focusing on restraint as well as violence” expands our understanding of the influences on behaviour.

Mapping beliefs and practices underlying restraint and incorporating them into prevention programming enables a dual focus: reducing enabling factors and strengthening preventive ones. SGBV prevention efforts traditionally rely on prohibitive norms and sanctions but do not always leverage the impact of peer influence and positive reinforcement. The Roots of Restraint in War highlights the power of informal socialization: “…even adoption by half of the group or less can sensitize their comrades to the need to spare civilians.”

To measure SGBV prevention effectively, interventions should track the reduction of enabling factors and the reinforcement of sources of restraint (for example, individual values and religious beliefs that preserve humanity, code of conducts that prohibit sexual violence, or social practices that protect children). ICRC’s Prevention of Sexual Violence Programme has been exploring this approach, combining InterAction’s GBV PEF tools with behavioural change theory. By adapting UNICEF’s Behavioural Drivers Model, ICRC staff have developed a framework to measure progress in reducing enabling factors and reinforcing restraint among armed actors and communities.

This approach uses, for example, participatory risk analysis with armed actors and consultations with survivors to identify and track drivers of the behaviour change that the intervention needs to contribute to. In its second year, the framework shows promising results, such as:

• reductions in cognitive biases that normalize SGBV perpetration by 14% for weapon bearers (annual results 2023) and by 18% in communities in South Sudan and Ethiopia (up to September 2024).
• 24% knowledge gain about the causes and consequences of sexual violence linked to conflict among service providers.
• 21% increases in self-efficacy among members of armed groups.

Though still in its early stages, this measurement and evaluation approach differs from prevalence studies in a key way: it reveals what works – and what doesn’t – in the frontline prevention of sexual violence linked to armed conflict.

Prevalence studies can be one important tool in the toolbox to understand and inform SGBV prevention policies. But it is not the only and final tool. True SGBV prevention work requires nuanced methods adapted to the complexity of the environments they seek to improve. Continuous, context-specific analysis of SGBV risk patterns and Theories of Change that incorporate community perspectives gives humanitarians a detailed view of enabling behaviours and sources of restraint that prevention efforts can target.

See also:

• Kristina Lindvall, Paula Iwanowska, Vanessa Murphy, Applying IHL with a gender perspective in the planning and conduct of military operations, October 24, 2024
• Maria Carolina Aissa de Figueredo, Forced to report: mandatory reporting of sexual violence in armed conflict, July 4, 2024
• Peixuan Xie, What are the transformative potentials of sexual and reproductive health and rights in humanitarian assistance: a feminist inquiry, April 18, 2024