Sexual violence risk factors help assess patterns and guide prevention. A risk factor alone does not show that someone will cause harm. This guide explains what the evidence can support. It also shows where a claim goes beyond the data.
What are sexual violence risk factors, and what are they not?
A risk factor is a trait or condition linked to a higher chance of harm in a group. It may or may not directly cause violence. The CDC sorts factors into individual, relationship, community, and societal levels (Centers for Disease Control and Prevention, 2024). Examples include hostility toward women, violent peers, and a community that tolerates violence. These links are not proof about any one person.
Can risk factors predict what one person will do?
Risk factors can inform risk estimates. But one factor does not settle what someone will do. Nor does a list that has not been tested for that purpose. A formal risk tool needs to be tested for the group, outcome, time span, and setting where it will be used. It cannot promise certainty or assign guilt based on group identity. The absence of a factor is not a guarantee. The presence of one is not proof. Actual conduct and the setting matter too. This is a guide to reading research, not a way to assess someone's personal risk. It is not advice to ignore credible threats.
Does a research label describe a whole group?
Hostile masculinity names a set of measured beliefs. These include sexual dominance, hostility toward women, beliefs that frame sex as conflict, rape myths, and acceptance of violence. A review covered 95 articles published from 1990 to 2020, using US male samples. It found links between these components and sexual aggression toward women. Definitions varied, and wider application needs more evidence (Ray & Parkhill, 2023). The label does not describe all men or male identity. It is not a diagnosis.
What does research on impersonal sex show?
Impersonal sex is a research measure of greater engagement in sexual activities with more casual sexual partners (Davis et al., 2018), and definitions can differ across studies (Zheng et al., 2026). It is not a synonym for sex drive or sexual assault.
A 2026 review included 39 eligible articles. It pooled 28 studies and 61 effects from US male samples published between 1995 and May 2023. The review found a correlation of .19 between impersonal sex and sexual aggression (Zheng et al., 2026). The results differed a lot across studies. That calls for care when applying the pooled figure to a new setting. The number is not 19 percent or a person's chance of causing harm. It does not measure libido or mean that casual sex equals assault. A link alone does not prove cause.
Who is being counted in the survey figures?
The 2016/2017 US adult survey asked victims about their lifetime experiences. Its weighted estimates show:
- 94.0% of female rape victims reported male perpetrators only.
- 76.8% of male rape victims reported male perpetrators only.
- 69.6% of male victims made to penetrate reported female perpetrators only.
These are victim reports (Basile et al., 2022). These are not shares of all offenders, nor percentages of all men or women who offend. Cases with unknown or partly unknown perpetrator sex are not shown. Some estimates were too unstable to report; a blank is not zero. Rape and being made to penetrate are separate survey categories. These US figures should not be assumed to hold worldwide.
The 2023/24 survey used a new sample drawn from addresses. It also changed questions and moved to web responses with a phone option. The methods report advises against comparing those results with prior years (Smith et al., 2025). The older numbers are historical, not the latest results. A report's date is not always the year of its data. The 2022 report, for example, uses 2016/17 data.
What is a practical framework for checking a risk claim?
Use five questions to trace a claim back to its evidence.
- Population: who took part, and does that group match the claim?
- Outcome: what was measured, and has a related act been swapped in?
- Design: can this design show cause, or only a link?
- Denominator: a share of whom, counted how, and who was left out?
- Limits: what do the authors say the finding cannot show, and does that caveat carry forward?
One hypothetical example shows how this works. A headline says, "All offenders are group X." The study behind it sampled victims of one form of harm in one country. The denominator is victims, not offenders. The unit is wrong, so the claim fails. The research may be sound. The summary is not.
A second hypothetical: "A factor linked to harm means a person with it will offend." That is not a probability. To estimate personal risk, a tool would need testing for that group and setting. Its outcome and time span would need to match the claim. A list on a website does not establish those things.
Here is a template you can use to rewrite an overstated claim: "In [studied group], [measured factor] was linked to [specified outcome]." The brackets mark placeholders, not real data. Each one forces you to name the population and the outcome.
Keep three rules in view. A missing or suppressed estimate is not zero. The absence of a factor is not a guarantee. And proof that a program worked needs outcome evidence, not just a risk list.
| Type of claim | What you need | What not to infer |
|---|---|---|
| A factor is linked to harm | The group studied, the outcome, and the limits | That any one person with the factor will offend |
| A figure describes who causes harm | The unit, denominator, and missing cases | That it counts all offenders or applies elsewhere |
| A program works | Evidence that tests the program's effects | That listing risks proves the program helps |
What should prevention and support actually focus on?
Prevention can work with people, relationships, communities, and policy. A risk framework suggests where to act. It does not prove that a given program works. To judge that, look at what the program changed and how its effects were tested.
If this material touches something personal, support matters. Seeking it is a reasonable step, not a last resort. A trusted clinician, a local service, or a qualified professional can help. The right route depends on your situation. Reading about harm is not a substitute for care.
Next in this series: Why women live longer. Browse the full Insights library.
Key takeaways
- A risk factor can inform prevention. Its presence alone does not prove what someone will do.
- Hostile masculinity is a measured construct, not a label for men or for masculinity in general.
- A correlation is not a percentage or a probability. Substantial differences across studies call for caution.
- These survey categories come from victim reports. They do not count all offenders worldwide.
- Check any risk claim on population, outcome, design, denominator, and stated limits.
Frequently asked questions
References
Davis, K. C., Neilson, E. C., Wegner, R., & Danube, C. L. (2018). The intersection of men's sexual violence perpetration and sexual risk behavior: A literature review. Aggression and Violent Behavior, 40, 83-90.
Basile, K. C., Smith, S. G., Kresnow, M., Khatiwada, S., & Leemis, R. W. (2022). The National Intimate Partner and Sexual Violence Survey: 2016/2017 report on sexual violence. Centers for Disease Control and Prevention.
Centers for Disease Control and Prevention. (2024). Risk and protective factors. Sexual Violence Prevention.
Smith, S. G., Chen, J., Basile, K. C., Siordia, C., Zhu, S., Kudon, H. Z., Chen, B., & Friar, N. W. (2025). The National Intimate Partner and Sexual Violence Survey (NISVS): 2023/2024 methodology report. Centers for Disease Control and Prevention.
Ray, T. N., & Parkhill, M. R. (2023). Components of hostile masculinity and their associations with male-perpetrated sexual aggression toward women: A systematic review. Trauma, Violence, & Abuse, 24(2), 355-368.
Zheng, W., Parkhill, M. R., Mongene, N. G., & Dimock, C. D. (2026). Impersonal sex and sexual aggression perpetrated by men against women: A systematic review and multilevel meta-analysis. Trauma, Violence, & Abuse. Advance online publication.
RAINN. (2026, May 31). Consent 101: Respect, boundaries, & building trust.
This article is for informational and educational purposes only and does not constitute financial, legal, tax, medical, or professional advice. Individual results vary.