Frequency of FRA

How often do false allegations occur?

On the 9th of June 2010, during PMs Questions, the then Prime Minister, David Cameron, stated that home office research showed that between 8-10% of allegations made to police were false.

In this section below, we examine whether this is likely to be true.

False allegations (as opposed to true ones)

After someone makes a false allegation to the police, amongst many other outcomes, there are two things that may happen.  A) the complainant admits that no crime had occurred (i.e. they lied) and B) where the police proved it hadn’t happened.

A) and B) typically result in the police recording these complaints as ‘no-crime’ events.  They are not that uncommon.  In a detailed Home Office (2007) research study from Feist et al., 676 complaints of rape were analysed.  100 of these were ultimately recorded as ‘no crimes’ by the police.  In other words, the police found evidence that no crime had occurred.  In 42 of these, the ‘victim’ “admitted that no offence had taken place”.  In another 8, the police found evidence or information that no offence took place.  This is 50 out of 676, or 7.7%.  Hang onto that.

Now, take A) & B) (the proven liars) and scale up these numbers to England and Wales – i.e. the ones where the victim admitted no offence or the police proved it false (42+10).  Apply the 7.7% figure to the 18,800 rape allegations made to police in 2011/12 in England and Wales. That translates to about 1,450 really false allegations per year.  1,450 men’s lives ruined by the crime of attempting to pervert the course of justice.  And provably ruined – either because of a retraction (yes, we’re aware that there are double-retractions), or because of contradictory evidence.

There is of course another group of men who would claim that they have been are falsely accused – C) all of those who are not charged (74.8%), and D) all of those who are found not guilty at a trial (6%).  Are these all liars – rapists who are getting away with it?  Or are are they all innocent men, roughly treated by the police and the court of public opinion?

The answer of course, is somewhere between the two.  So the proportion of allegations that are false must lie between 7.7% and and 91.6%, this figure being 100% minus the 8.4% of men accused, who were found guilty.

And before you think we aren’t aware, then of course there are some rapists mixing in amongst the 7.6% of provably false allegations or retractions.  And of course there are some men who are wrongfully convicted of rape.  This is all very muddy water.


Sexual offences


The total number of people who are falsely accused is uncertain

The research in the section below attempts to put some of the studies that look at this in one place. The range of estimates from these papers is enormous – from about 2% to 90%.

The major problem has been in weaknesses and lack of consistency in their study design.  We intend to produce a later article with a series of recommendations as to how this issue should be addressed in future.

When is a lie not a lie?

In the interests of fairness, we should note a degree of caution that should be applied to what is meant by an “admission of falsehood”.  Put it another way, ‘when is a lie, not a lie?’.  In Para 140, p39, of the Investigation of Rape in London by the R Hon Dame Elish Angiolini QC (2015), she talks about the reasons why people might retract an allegation and admit that it was false.

‘fear of violence, or intimidation, which may itself constitute a criminal offence. Other factors which may also be relevant include, a desire to give an existing relationship another chance; the impact  upon children; embarrassment; a fear of going to court; family/community pressure which may stem from the immediate and extended family, the wider community and from cultural traditions; insecure immigration status; and mental health issues/learning disabilities’.

So if a retraction is admitted to be a lie, it might not be.  Alice through the looking glass again.  A world of double- and triple-negatives.

But as there is no particularly good evidence of the proportion of these no-crimes that fall into the categories above, then for the time being, let’s run with the quoted figure of admissions or situations where the police have proved the allegation didn’t happen.

Further research about false allegations

The evidence above is particularly applicable to the UK in 2016, because it’s from the UK, and it’s recent.  There has however been over 40 years of research here and elsewhere into the problem of people who make up lies about sexual offences.  If we look at this, can we make any generalisations about whether these lies are rare, or whether they are common?

First, let’s dispose of one heavily quoted number.  Susan Brownmiller is one root source for a discredited assertion that 2% of allegations made to police are false.  She is a US feminist author and journalist.  Importantly, she does not deny that false allegations of rape occur.  Even this feminist agrees that some people make up false allegations.  In the 1975 edition of her book, Against our Will, she attributed the quoted 2% to a speech by Lawrence H. Cooke, Appellate Division Justice, Before the Association of the Bar of the City of New York, Jan. 16, 1974.  This, she asserted, was the proportion of allegations made to the police that are false.  It was a figure that isn’t backed up by data or research.  That hasn’t stopped it from being heavily quoted, unfortunately.

Research into the number of false allegations of rape has actually been going on since the 1960s.  Most of it is of limited quality, as the authors have no consistent definition of what they mean by a false allegation of rape and as you can see from the data, the range of estimates of falsehood is enormous.  Is it when someone admits to the police that they made it up?  Or is it then the police reckon the complainant isn’t telling the truth?  There’s a world of difference here.  In 2006, Philip Rumney from the University of the West of England in Bristol made a very good attempt at pulling the research together.  Surprisingly, he didn’t add all the numbers together and try to reach some sort of average for the proportion of all allegations made to the police that might be false.  We have done that for him, here.

What is the best estimate of the number of false allegations?

To work this out, we’ve used a tried and tested statistical method, called a ‘random effects’ meta-analysis, and we’ve applied it to Rumney’s data.  Meta-analysis is an attempt to summarise all of the research and come up with a number that might be a reasonable estimate of the proportion of all reports to police that are false.  Our number is 17% – see the figure below, which summarises the studies.  There’s a margin for error to this number, so it could easily range from 10% to 22%.  Whilst it is our best estimate, it is inadvisable to quote this figure without understanding that these studies have used many different types of methodologies over a long period of time, have been done in different countries, with different jurisdictions.  The proportion of false allegations in these studies of course varies – significantly.

So please understand that the numbers below are also unpublished and have not been subject to peer review in an academic journal.  We would argue however, that 17% is definitely a more reasoned number than 2%.


Adding some more recent evidence to this number, there is another estimate, which comes from the MoJ itself in 2010.  It is quoted on p12 of their publication on false allegations of rape.  It is again, the figure of 8-10%, derived from the analysis by Feist (2004/7), a researcher for the Home Office.  We use 8% as our ‘low-estimate’ and 17% as the ‘high-estimate’.  Of course, the true proportion of false allegations in the UK in 2016 could be outside these limits.




The estimate of the number of rapes that occur to persons in the UK is 97,000 per year (range 77,000-116,000).  Reference 2011/12 CSEW Table 2.2 – Estimated numbers of victims of sexual offences in the last 12 months among adults aged 16 to 59, average of 2009/10, 2010/11 and 2011/12 CSEW.

Research data links as of Oct 2015:

MoJ review of anonymity in rape allegations (2010)

MoJ Report on Sexual Offences in the UK (2013)

Report into FRA for DPP by Allison Levitt QC (2013)

Rumney review of evidence into FRA (2006)

Crimes detected in England and Wales (2012)

Provenance of assumptions

The links above have provided all of the evidence for the numbers quoted in the figure above.  The three figures marked with a dagger are estimates.

(1) The number of males in the UK aged 16-64 inclusive comes from a zip file, published by the office of national statistics (ONS).  UK Population Estimates 2011-12.

(2) Crime Survey for England & Wales 2010-12 (central estimate quoted refers to women 16-59 years).

The first estimate (1) comes from a meta-analysis that has performed on the papers included in Rumney’s (2006)review of the evidence on the proportions of reports of rape that might be false.  These reports are culled mostly from the UK and the US.  They were collated over a number of different state jurisdictions and at different times over a 50 year interval.  The studies are very probably ‘heterogeneous’.  They should therefore be considered as if there is no common underlying rate of false allegation that they are all sampling and a random effects meta-analysis is the most appropriate method to use.  The resulting number of 17.0 percent may not be appropriate to the UK in today’s conditions, but it remains a best-estimate.

The second estimate of the number of trials of PCJ/WPT (2) of is derived from Allison Levitt QC’s report into cases passed to the CPS for consideration of Perversion of the Course of Justice (PCJ) or Wasting Police Time (WPT).  Figures collected over a 17 month period in 2011-12 were scaled to a 12-month period to make them comparable to the data available for several 12-month intervals used in other analyses (e.g. the CSEW analysis) in 2011-12.  No other assumptions have been applied.

This number of trials must have been taken from the larger number in (1) of the people derived from Rumney’s analysis who might have been presumed to have been falsely accused of rape.

The third estimate (3) is derived from ONS data about the number of crimes detected in England and Wales in 2012.  On p20, this shows that the sanction detection rate for Perversion of the course of justice was 54% in bother 2010/11 and 2011/12.

Estimates (1), (2) and (3) together provide a best estimate for the proportion of all attempts to pervert the course of justice through making a false allegation of rape that result in a successful prosecution.  This is 0.53%.

0.53% is 22 times less than the proportion of true rape allegations to the police that result in a successful prosecution (11.9%) .  This is the disparity between the credence the police / CPS give to true allegations of rape as opposed to the credence given to false allegations of rape.  Assuming that the same problems apply in Scotland and Northern Ireland, then across the UK as a whole it must be concluded that men who are being framed for a false allegation of rape, are being appallingly treated by the state compared to people who are raped.