How do Crime Survey for England and Wales sexual assault estimates compare with police-recorded crime over the last 20 years?

1. The core contrast: survey prevalence versus police counts

The fundamental difference is methodological: the CSEW measures victims’ self-reported experiences in the population and therefore captures incidents that never come to police attention, producing estimates in the hundreds of thousands per year (for example, the CSEW estimated around 898,000 people experienced sexual assault in YE Mar 2025) while police-recorded sexual offences number in the low hundreds of thousands (about 209,079 offences recorded in YE Mar 2025), meaning survey estimates are larger because many incidents are not reported to police ^{[1] [2]}.

2. Divergent trends over 20 years: steady survey signals, volatile police figures

Over the long run the CSEW is used as the preferred measure of trend for sexual assault because it is less affected by administrative changes; it shows sexual assault as an exception to the general long-term fall in other crime types and provides prevalence percentages rather than raw counts (for example, 1.9% experienced sexual assault in YE Mar 2025) ^{[3] [1]}. By contrast police-recorded sexual offences have risen markedly in recent years, but analysts warn that much of that rise reflects improvements in recording practices, legislative change (new online sexual offences), and greater reporting, not purely an increase in victimisation ^{[1] [2] [4]}.

3. Why police numbers moved up: recording, legislation and reporting

Multiple ONS releases and parliamentary analysis attribute a substantial share of the rise in recorded sexual offences to non-crime drivers: police forces have improved recording accuracy, victims have become more likely to report, and new offences—such as those introduced by the Online Safety Act—expanded what is counted; the Home Office-supplied police data therefore trace demand on policing and criminal justice rather than the full prevalence in the community ^{[5] [1] [2] [4]}.

4. Strengths and limits: when to trust the CSEW and when to use police figures

The CSEW’s strength is that it captures unreported crimes and yields a stable long-term trend for sexual assault that is less susceptible to administrative artefacts, which is why ONS recommends it for prevalence and trend interpretation; however the survey has methodological caveats—questionnaire changes, temporary data gaps (for example an error affected some months in 2022–23) and design changes that can affect comparability across years—so researchers triangulate both sources and academic studies highlight that different surveys can yield different magnitudes, sometimes substantially ^{[3] [1] [6]}.

5. What the comparison implies for policy and public debate

The consistent gap—CSEW prevalence far exceeding police counts—means criminal justice statistics understate victimisation and that rising recorded offences signal both improved visibility of sexual crime and increased strain on police and courts; policymakers and advocates therefore draw on CSEW for scale and on police data to understand system demand and operational pressures, while cautioning that short-term movements in recorded crime may reflect recording and legislative shifts rather than sudden changes in offending ^{[1] [2] [7]}.

6. Alternative interpretations and unresolved questions

Some researchers note survey estimates themselves vary by instrument and sampling (the UK’s other national surveys can produce larger or smaller estimates), so disagreement about absolute levels persists and underlines the need for methodological transparency and parallel data sources; ONS and independent reviews have recommended ongoing improvements to both survey design and police recording to reduce ambiguity in trend interpretation ^{[6] [8]}.