Evidence and opinion, by Jonathan Egelston · October 26, 2026
Is there one reliable percentage for all false accusations?
The sources discussed here do not establish a single rate for every kind of accusation. A statistic about police reports, documented exonerations, or one institution’s cases answers a different question. Before quoting a number, identify who was counted, what “false” meant, and the period studied.
One study with a specific scope
Ferguson and Malouff’s 2016 meta-analysis combined seven studies of confirmed false sexual-assault reports to police. It estimated 5.2%, with a 95% confidence interval of 3.0%–8.9%, and reported variation across studies. Those findings do not measure all accusations, all offenses, or the percentage of men who experience a false accusation.
A confirmed-false category also does not turn every other report into a proven-true case. Uncertainty has to remain visible. This page summarizes the published abstract; it does not present a new analysis of the underlying cases.
Ask what is in the denominator
Imagine a fictional review of 100 reports. If five are classified as confirmed false, that is five out of the reviewed reports. It is not five out of every person in the community. If the remaining reports include unresolved cases, the number alone does not explain their outcomes.
This fictional example illustrates how a percentage works. It is not an additional research finding.
My opinion: the harm matters without inflating a number
I believe men who have been wrongfully accused deserve serious support. That position does not depend on claiming that most reports are false. A smaller documented group can still experience profound injustice, and people reporting genuine harm also deserve to be taken seriously.
I want the discussion to be strong enough to withstand scrutiny. That means acknowledging limits, correcting errors, and refusing to use one statistic as a shortcut to judging an individual.
A checklist before reposting a statistic
- Find the original study rather than a screenshot of a claim.
- Record the publication year and the population studied.
- Check the definition of the outcome being counted.
- Keep confidence intervals and stated limitations with the estimate.
- Distinguish a finding from an opinion about its importance.
Frequently asked questions
Does a disputed accusation automatically count as false?
No. Disagreement is not itself verification. A classification requires evidence and a stated method.
Can this research settle an individual dispute?
No. A particular dispute requires examination of its own information and circumstances.
Read my personal statement and related articles.
