False Rape Reports Are Rare, but Reports Show Wide Documentation Gaps in “Unfounded” Cases, New Study Finds
FOR IMMEDIATE RELEASE
24-year analysis of 5,638 Cleveland rape reports finds just 2.7% were false. Another 2.4% of cases labeled “unfounded” lacked the evidence to support that classification.
CLEVELAND, Ohio — August 31, 2026 — A new study published in the Journal of Criminal Psychology finds that false rape reports are far less common than police departments’ use of the “unfounded” (or non-crime) classification would suggest and that a substantial share of cases closed as unfounded were never adequately documented to justify that decision. The study, conducted by Rachel Lovell, PhD, Associate Professor of Criminology and Director of the Criminology Research Center in the Levin College of Public Affairs and Education at Cleveland State University, and Lacey Caporale, PhD, Research Associate at the Criminology Research Center, analyzed nearly 25 years of rape reports from the Cleveland Division of Police (CDP) to produce one of the largest and most generalizable estimates of false and baseless reporting rates to date.
Using a dataset of 5,638 rape reports linked to sexual assault kits from CDP spanning 1991 to 2015, Lovell and Caporale identified 386 reports that had been classified by officers as “unfounded.” A random, representative sample of 122 of those narratives was then analyzed in depth using qualitative content analysis to determine whether each case was genuinely false (clear admission or strong evidence the victim lied), baseless (meaning the account simply didn’t meet the legal definition of rape), or neither.
The findings challenge a persistent assumption in the criminal justice system: that when police close a rape case as “unfounded,” it means the victim lied. Extrapolated across the full dataset, only 2.7% of all rape reports were determined to be false (of which 1.1% were false due to victim recantation and another 1.6% were determined false based on third-party statements and physical evidence (e.g., a nurse or doctor noting “no signs of trauma”)). Just 1.7% were baseless. Meanwhile, 2.4% of cases had been labeled unfounded despite insufficient documentation in the report to justify the designation, a rate nearly as high as the false-report rate itself.
“For decades, the myth that women frequently lie about being raped has shaped how these cases are investigated and perceived,” said Lovell. “What we found is that when you actually read these reports and separate false claims from cases that were closed without enough evidence or documentation, the true rate of false reporting is very low, consistent with, and even more conservative than, prior research. What’s far more common is a report that’s too brief or too vague to justify closing a case as a non-crime in the first place. That’s not a victim problem. That’s a documentation and training problem.”
The study also draws on a high-profile local case to illustrate the real-world stakes of premature or poorly evidenced unfounding decisions. In 2008, a woman escaped the Cleveland home of Anthony Sowell after he raped and strangled her; she fled bleeding and sought help from police, who arrested Sowell but released him two days later, and the case was closed as unfounded. Less than a year later, another woman reported being raped at the same address. When police eventually returned, they discovered the remains of 11 women in and around Sowell’s house. The study’s authors note that the 2008 case shares striking similarities with reports in their sample — brief, credibility-focused narratives that discounted the victim’s account without sufficient investigation.
“Reading through these narratives case by case, what stood out was how often the entire justification for closing a case as unfounded amounted to a sentence or two, often surrounding perceived victim credibility flaws,” said Caporale. “We also found that more than half of the reports we classified as false involved minors, which speaks to how vulnerable young victims already are in this process. These point directly to where better training, oversight, and documentation are needed most.”
The research builds on the authors’ prior machine-learning analysis of the same dataset, which found that unfounded reports were consistently shorter and more negatively worded than other rape reports over the full 24-year period. While that earlier work could detect broad linguistic patterns, it could not explain why individual cases were closed as non-crimes. This new study fills this gap through detailed human coding of the underlying narratives.
The authors argue that conflating “false” and “baseless” under the single “unfounded” label and, more troublingly, closing cases without adequate evidence at all, inflates public perceptions that rape victims commonly fabricate allegations, while also removing many legitimate cases from official crime statistics, since unfounded reports are excluded from federal crime data. The study calls for additional officer training on trauma-informed report writing, stricter adherence to federal unfounding guidelines, and more detailed documentation of the evidentiary basis for closing a rape case as a non-crime.