What are reliable sources and studies measuring gender differences in sexual violence perpetration?

1. What “reliable” means in this literature — large samples, validated instruments, transparent bias adjustments

Reliable studies combine representative sampling or pooled, multi‑country meta‑analysis, validated perpetration questionnaires, and explicit handling of disclosure bias and sample composition. The Lancet global analysis used age‑standardisation and disclosure‑adjustment methods across locations and explicitly cautioned that data sparsity and measurement challenges likely underestimate true prevalence and bias female/male comparisons because far more data exist for females ^[1]. Measurement reviews emphasize that perpetration research requires purpose‑built self‑report instruments covering a broad range of tactics; measures that omit coercive behaviours systematically undercount perpetration ^[2].

2. Best large‑scale prevalence sources to consult

For childhood sexual violence, the Lancet’s 1990–2023 global analysis is a top source: it reports age‑standardised global prevalence and documents methodology for disclosure adjustment and age windows critical to prevention ^[1]. For adult and regionally focused prevalence and policy framing, WHO/UN and UN Women publications and the WEF report aggregate global estimates and measurement guidance and highlight persistent gendered patterns and data gaps ^{[3] [4] [1]}.

3. Measurement and instrument reviews researchers cite most

Systematic reviews of perpetration measures and psychometrics are essential. Ngai Lam Mou’s 2025 systematic review focuses on self‑reported measures used to quantify men’s sexual violence perpetration and documents that broader tactic lists produce higher detection rates; it argues for measurement advancement to evaluate prevention and screen perpetrators ^[2]. Meta‑analytic work on attitudes and perception (Trottier et al.) shows how rape myths and gender‑role expectations correlate with interpretation and reporting — relevant because measurement instruments and respondent framing change observed gender patterns ^[5].

4. Population and subgroup gaps: males, LGBTQ+ people, and underrepresented locations

Available sources repeatedly flag gaps. The Lancet notes most locations have more female than male data and that sparse male, boy, transgender, and gender‑diverse data limit interpretation of sex/gender differences ^[1]. A scoping review of LGBTQ+ adult sexual‑violence research likewise documents knowledge gaps about perpetration by and against LGBTQ+ people and calls for more inclusive data ^[6]. UN and UN Women guidance stresses that a lack of consistent data on perpetrator sex and victim gender identity impedes understanding and accountability ^[3].

5. Mode and question wording matter — why estimates vary across studies

Perpetration rates change with question framing, tactic breadth, and survey mode. Studies of adolescents and specific risk groups show large variance: a South African adolescent study found 59% of sexually active adolescents reported sexually violent behaviors though only 8% labeled it “rape,” illustrating how label versus behaviour questions diverge ^[7]. Measurement reviews show rewording victim‑oriented items for perpetrators can miss relevant tactics unless instruments are purpose‑designed ^{[2] [7]}.

6. Conflict, institutions, and context produce different perpetrator gender patterns

Conflict‑related sexual violence analyses and UN reporting show that perpetration contexts (armed actors, detention) create distinct patterns: verified CRSV cases overwhelmingly affect women and girls, but men and boys are also targeted in specific contexts — these are documented separately from population surveys and require different measurement approaches ^{[8] [9]}. Likewise, national surveys and sector studies (workplace, gangs, campuses) reveal context‑specific prevalence and perpetrator profiles ^{[10] [11] [12]}.

7. How to use this evidence responsibly in research or policy

Use multi‑source triangulation: combine high‑quality pooled prevalence estimates (Lancet), measurement reviews (Ngai Lam Mou), and UN/WHO guidance on data collection to avoid overinterpreting sex differences where data are sparse ^{[1] [2] [3]}. Explicitly report instrument wording, sample composition, disclosure adjustment, and where subgroup data are missing. Available sources do not mention a single, universally accepted metric that resolves all biases; instead, they urge iterative measurement improvement and transparency ^{[2] [1]}.

Limitations: major global and regional sources acknowledge systematic gaps in male, transgender, and nonbinary data and warn that comparative claims often reflect measurement artifacts and compositional bias in available datasets ^{[1] [3]}.