Sexual violence surveys should measure institutional response, not just prevalence

Updated Sep 07, 2026

A campus safety survey can examine both experiences of harm and confidence in an institution's response. Steele and colleagues' second OUR SPACE survey connects these questions at one UK university, while also showing why participation and study design matter.

What the study found

The June 2024 survey invited approximately 26,000 students. Its analysis included 1,097 completed responses, about 4% of the student population, and used survey weights. Women reported higher levels of sexual assault and harassment than men; perpetrators were commonly fellow students and incidents frequently occurred on university premises.

Respondents reporting sexual violence generally viewed institutional response less favourably. The sexual-assault association was statistically significant among women but not men; the harassment association was significant in both groups. This is evidence of associations, not proof of what caused those perceptions.

The low participation rate leaves potential self-selection bias. The authors also discuss pandemic-related differences from the earlier survey. Repetition alone cannot attribute changes to a particular policy.

Questions for local survey design

Our recommendation is to distinguish a research survey from a route for reporting an incident or requesting help. Tell participants which purpose applies, who can see responses and how to access support. Survey designers should involve the institution's specialist staff and students in deciding which questions are appropriate.

For institutional review, keep separate records of prevalence estimates, perceived safety, confidence in support and perceptions of handling. Avoid treating a single average as a verdict on all four. Small groups require particular care: suppressing an identifying figure must not make their concerns disappear from the review process.

Open questions may help students describe barriers, but their inclusion needs a specific handling plan. Ordinary course-feedback analysis should not be assumed to provide the right process for sensitive disclosures. Our governance checklist is a starting point for responsibilities, not a substitute for specialist procedures.

FAQ

Can these figures be used as a sector benchmark?

No. One institution's responding sample should not stand for every university or student group.

Does this assess whether a new policy worked?

No. The design does not isolate a policy's effect. A local evaluation needs to account for participation, changing context and other explanations.

Reference

Steele, B., Tan, V., Giraldi, D. M., Degli Esposti, M., & Humphreys, D. K. (2026). Experiences and perceptions of sexual violence in a UK university: results from the second iteration of the OUR SPACE survey. Studies in Higher Education.

Correction, 7 September 2026: added participation and subgroup limitations, qualified claims about repeated surveys and revised the guidance on sensitive evidence after reviewing accessible primary sections. This is not a full-paper verification.