Are there reliable academic studies comparing prevalence (victimization-based) versus conviction-based rates of child sexual abuse across races?
Executive summary
There is a substantial body of empirical work comparing victimization-based (self-report or population surveys) and case-based (reports, substantiations, convictions) measures of child sexual abuse (CSA), and that work finds disparities across racial and ethnic groups — but the literature is fragmented: many studies examine reporting/substantiation disparities rather than direct side‑by‑side academic comparisons of victimization prevalence versus conviction rates by race. Key, recent population‑level CPS analyses document higher CPS victimization rates for American Indian/Alaska Native and Black children (e.g., 14.3 and 12.1 per 1,000) while survey and retrospective prevalence studies show different patterns and larger absolute prevalence numbers that vary by sample [1] [2] [3] [4].
1. What the literature actually compares: victimization surveys vs. administrative outcomes
Academic work separates two measurement worlds: retrospective or prospective prevalence (self‑report surveys and community samples) versus administrative case counts (reports to CPS, substantiations, prosecutions, and convictions). Studies using CPS or NCANDS data map incidence and substantiation rates by race and state and routinely report non‑White overrepresentation in reports and substantiations (e.g., higher rates for American Indian/Alaska Native and Black children) [1] [4] [2]. Parallel survey research and ACEs‑style retrospective studies produce higher overall prevalence estimates and often different race patterns (for example, community samples finding high CSA prevalence among Black women) [3]. These two approaches are not identical measures and academic papers explicitly note that disparities can reflect both true differences in victimization and system biases [4] [5].
2. Studies that directly examine racial disparities in substantiation, prosecution or conviction
Multiple peer‑reviewed analyses focus on substantiation and case outcomes by race. State‑level and national NCANDS analyses show consistent racial/ethnic and gender disparities in substantiated maltreatment and in how cases are processed; for example, Latine girls were more likely to have sexual abuse substantiated in one large state‑level study, and Black children were often overrepresented in reports and substantiations [6] [7] [4] [8]. Researchers at Johns Hopkins used NCANDS data to test how child and accused‑person characteristics drive substantiation disparities [9] [10]. Regarding prosecutions, some older reviews note mixed findings on whether victim race affects prosecution decisions, but models generally emphasize that type of abuse and system referral patterns matter more than victim race alone [11].
3. Evidence from victimization (prevalence) surveys and retrospective studies
Prevalence research — often retrospective adult reports or convenience samples — yields substantially higher rates of CSA than administrative counts and sometimes different racial patterns. For example, Utah BRFSS and other ACEs modules report higher childhood sexual abuse prevalence among Black, American Indian/Alaska Native, and some other groups compared with state‑level averages [12]. Academic reviews and targeted studies of college or community samples have documented significant ethnic differences in reported CSA prevalence and in characteristics of abuse and disclosure [13] [14] [12] [3]. These surveys capture unreported abuse that never entered CPS or criminal systems, so they are essential for a victimization‑based picture [12].
4. Gaps, methodological reasons for discordant findings, and interpretive cautions
Authors repeatedly warn that differences between victimization prevalence and conviction/substantiation rates can reflect multiple causes: differential exposure/risk, socioeconomic confounds (poverty), differential reporting and disclosure, reporter/source bias, variability in state CPS policies, and prosecutorial discretion [4] [5] [15] [6]. Many studies note limitations: convenience samples for prevalence studies; limited racial categories or aggregation that obscures subgroup patterns; and that conviction data is sparse and influenced by legal‑system practices rather than true incidence [14] [10] [11].
5. Do “reliable academic studies” directly compare prevalence vs conviction rates by race?
Available sources document both sides (prevalence surveys and administrative/conviction analyses) and several papers discuss disparities between reporting/substantiation and population prevalence — but “side‑by‑side” academic studies that explicitly match representative victimization prevalence estimates to conviction rates by race in the same population and timeframe are rarely presented in the sources you supplied. Many high‑quality papers analyze disparities within administrative datasets or within survey data and explicitly call for integrated work to reconcile measurement differences [6] [4] [9]. If you seek a single study that neatly juxtaposes population prevalence and conviction outcomes by race across the same sample and period, available sources do not mention such a definitive, unified comparison.
6. Practical next steps and what to look for in future research
Look for studies that (a) use representative population surveys linked or comparable to NCANDS/CPS years, (b) disaggregate race/ethnicity beyond White/Black/Latinx, and (c) model poverty, reporter source, and state policy differences. Recent whole‑population NCANDS/state analyses and systematic reviews flag these needs and authors are calling for integrated designs to separate true prevalence differences from system bias [6] [4] [8]. For immediate citations, consider the state‑level NCANDS work (Fix et al., 2025) and NCANDS/BRFSS comparative literature summarized above [6] [12] [9].