Sexual Scripts, Regret, and How Experience Becomes an Allegation

There is a perfectly legitimate question to ask about modern rape statistics, particularly following Labour’s announcement today that it intends to create specialist rape courts across England and Wales.

One statistic that will inevitably be cited in support of the argument for reform is the apparently tiny proportion of recorded rape allegations which result in a charge and ultimately reach the courts.

The figure is often reported as being around 2.7 per cent.

The reality is slightly more complicated. The latest Home Office figures show that, of rape offences recorded in England and Wales in the year ending March 2026, 3.4 per cent had been assigned a charge or summons outcome at the time the statistics were published. For the previous year, the figure was initially 2.8 per cent, but subsequently increased to 5.2 per cent as more investigations were completed, so as we see it changes from year to year.

The Home Office expressly warns that the apparent charge rate rises as cases are closed and outcomes are assigned. [See Home Office, Crime outcomes in England and Wales 2025 to 2026, section 2.4, Rape offences.]

But there is another statistic which is perhaps even more important when considering what these numbers actually mean.

The impression created by a figure of only around 3.7 per cent is that rape is somehow uniquely unlikely to result in a prosecution. Yet the proportion of recorded crime resulting in a charge or summons is also remarkably low across the entire criminal justice system as a whole.

In the year ending June 2025, the Ministry of Justice recorded 6.59 million crimes in England and Wales, with 492,256 resulting in a charge or summons. That equated to approximately 9 per cent of recorded crime in England and awakes reaching court.[See Ministry of Justice, Criminal Justice Statistics quarterly: June 2025, Overview and Figure 1.]

The figures vary considerably between offences. The latest Home Office statistics show that, in the year ending March 2026, 7.3 per cent of victim based offences resulted in a charge or summons. For robbery the figure was 10.5 per cent and for theft 9.1 per cent. Sexual offences were considerably lower, at 4.7 per cent, while rape was lower still. [See Home Office, Crime outcomes in England and Wales 2025 to 2026, sections 2.1, 2.3 and 2.4.]

It should also be noted that reaching court does still not guarantee a conviction.

Why does that matter?

Because different crimes produce very different evidential circumstances.

A stolen vehicle may have CCTV, registration details, forensic evidence, a digital trail and an identifiable suspect. A shoplifting offence may be captured on cameras. An assault may have independent witnesses, photographs or medical evidence.

A rape allegation can present at any time and have an entirely different evidential problem.

The central question may be what happened between two people behind a closed door, sometimes with no independent witnesses and no forensic evidence capable of determining whether the sexual activity was consensual.

That does not make the allegation less serious.

It makes the evidential problem different.

And it means that simply comparing the proportion of rape allegations resulting in prosecution with the proportion for other crimes does not, by itself, tell us whether the criminal justice system is treating rape allegations properly or improperly.

This is precisely why the numbers require careful examination.

If approximately 3 per cent of recorded rape offences initially result in a charge or summons, while approximately 9 per cent of recorded crime overall does.

But it is not, by itself, proof that 97 per cent of rape allegations are being ignored, nor is it proof that 97 per cent represent genuine rapes which the justice system has failed to prosecute.

Indeed, the Home Office data show that rape investigations are particularly likely to remain open for longer and to receive evidential difficulty outcomes. In the year ending March 2026, 51.7 per cent of rape offences were assigned an evidential difficulties outcome. [See Home Office, Crime outcomes in England and Wales 2025 to 2026, section 2.4.]

This brings us to the central question of this essay.

What exactly are these statistics measuring?

The controversy begins when we look at the numbers which purport to tell us how much rape actually occurs.

That is where the methodology matters.

One of the most striking episodes in the history of rape research is the work of Eugene J. Kanin, published in Archives of Sexual Behavior in 1994. In False Rape Allegations, Kanin examined 109 forcible rape complaints made to one small Midwestern American police department over a nine year period. Forty five were classified as false, producing the extraordinary figure of 41 per cent. [See the Abstract and Incidence of False Allegations section, pp. 81–92.]

That figure has subsequently become something of a statistical football. It is sometimes invoked as evidence that false rape allegations are extraordinarily common. It is also routinely dismissed or marginalised in modern discussions because of the study’s small geographical base, its reliance on a single police department and the difficulty of generalising from 109 cases.

The criticisms are legitimate but in themselves do not disprove his work.

Others say that Kanin’s 41 per cent cannot be extrapolated to an entire country and is not establish by other studies that put false claims be 2 to 10 per cent, as we often hear.

And that is the broader point.

The same methodological caution which should be applied to Kanin’s 41 per cent should also be applied to every other figure in this debate.

The much more familiar modern figure comes from research such as Lisak and colleagues’ 2010 study, False Allegations of Sexual Assault: An Analysis of Ten Years of Reported Cases, published in Violence Against Women.

The researchers examined only 136 sexual assault reports made to a major Northeastern American university over a ten year period. Eight cases, or 5.9 per cent, were classified as false. The authors then considered their findings alongside previous research and concluded that the existing literature suggested a prevalence of between 2 and 10 per cent. [See Abstract and pp. 1318–1334, particularly the Results and Discussion sections.]

That is an enormous difference in the underlying figures found across different studies.

But it is important to be precise about what it demonstrates.

It does not demonstrate that the false allegation rate is 2 per cent, 5.9 per cent or 10 per cent.

It demonstrates that different studies, using different populations, definitions and methodologies, have produced substantially different results. This is a major issue.

Ferguson and Malouff subsequently conducted a meta analysis of seven studies in Assessing Police Classifications of Sexual Assault Reports: A Meta-Analysis of False Reporting Rates, published in Archives of Sexual Behavior in 2016. They calculated an overall confirmed false reporting rate of 5.2 per cent. [See Abstract and Results section, pp. 1185–1193.]

Five per cent sounds reassuringly precise.

It is not.

And that is where we need to be extremely careful about what the statistics are actually telling us.

The first thing to understand is that these studies are generally measuring something very specific: reports made to the police which researchers or investigators subsequently classified as confirmed false.

That is not the same question as asking how many sexual encounters are misremembered, misunderstood, regretted, retrospectively reinterpreted or incorrectly described by the complainant or go on and result in acquittal against the accused.

Nor is a report being closed without prosecution equivalent to a false allegation.

This is one of the great statistical problems in the subject.

A rape allegation can disappear from the criminal justice system for many reasons. The complainant may withdraw. The police may decide there is insufficient evidence. The prosecutor may consider that the evidential test cannot be met. The suspect may never be identified. There may be contradictory accounts. There may be no forensic evidence. There may be evidential problems concerning consent. Or the allegation may genuinely and obviously be false.

Those are not interchangeable categories.

Research into attrition illustrates precisely this problem. In A Gap or a Chasm? Attrition in Reported Rape Cases, Liz Kelly, Jo Lovett and Linda Regan examined 2,643 reported rape cases in England and Wales and distinguished between cases classified by police as false, cases withdrawn by complainants and cases lost at later stages of the criminal justice process. [See A Gap or a Chasm?, particularly Chapter 3 and the discussion of false allegations and attrition.]

That does not mean that every allegation which is not prosecuted is true also.

It means that we need to distinguish between an allegation which is false, an allegation which cannot be proved and an allegation which is simply withdrawn or otherwise lost from the criminal justice process.

But there is another problem which is discussed rather less often.

The thing being measured has itself changed.

For centuries, the traditional conception of rape was heavily associated with force, resistance and the overpowering of the complainant’s will.

The historical importance of force, fear and absence of consent can be seen in the development of English rape law and in the discussions surrounding the Sexual Offences (Amendment) Act 1976. [See the parliamentary debate on the Sexual Offences (Amendment) Bill, 13 February 1976, particularly the discussion of the traditional common law definition and the Heilbron Committee’s recommendation that absence of consent should be the essence of the offence.]

In England and Wales, however, the law progressively moved away from that model. The Sexual Offences (Amendment) Act 1976 amended the law so that rape was defined by sexual intercourse without consent, together with the required mental element of knowledge or recklessness as to consent. [See section 1 of the Sexual Offences (Amendment) Act 1976.]

The Sexual Offences Act 2003 subsequently created the modern statutory framework, including section 74.

Section 74 provides that a person consents if they agree by choice and have the freedom and capacity to make that choice. The prosecution must also consider the defendant’s reasonable belief in consent in accordance with the statutory framework. [See Crown Prosecution Service, Rape and Sexual Offences Prosecution Guidance, sections headed Consent, Capacity and Reasonable Belief in Consent.]

This is an important development in the law and this has a statistical consequence.

When the legal and social concept of rape changes, the population being counted changes with it.

That makes simplistic historical comparisons extremely dangerous.

If rape was once principally understood through force and resistance and is now determined through questions of choice, freedom, capacity and reasonable belief, we cannot simply take the number of cases recorded under one conceptual framework and compare it with the number recorded under another as though the underlying phenomenon had remained static.

The same problem becomes particularly acute with alcohol.

Alcohol is unquestionably associated with a substantial proportion of sexual assaults. A UK study by Lily Monk and Alyson Jones, Alcohol Consumption as a Risk Factor for Sexual Assault: A Retrospective Analysis, examined 286 cases involving complainants aged 12 to 25 who attended the Lancashire SAFE Centre between May 2011 and April 2012. It found that 70.6 per cent had consumed alcohol during the 24 hours before the reported assault. [See Abstract and Results section, Journal of Forensic and Legal Medicine, volume 23, pp. 55–61.]

But ‘had been drinking’ and ‘was incapable of consenting’ are obviously not synonymous.

The modern legal position recognises precisely this distinction. The CPS guidance states that prosecutors should consider both capacity and freedom to choose, and specifically explains that alcohol may affect capacity but that the question remains whether the complainant had the capacity and freedom to make a choice. [See CPS, Rape and Sexual Offences Prosecution Guidance, section on Consent, Capacity and Intoxication.]

That distinction is enormously important.

Someone can be drunk and still capable of consenting.

Someone can also be so intoxicated that they lack the capacity to consent.

Between those two points lies an enormous grey area.

And science itself demonstrates the difficulty.

Recent research by Ellen Laughlin, Molly Pettitt, Veronica M. Lamarche and Laurie James-Hawkins, Just One Shot? The Contextual Effects of Matched and Unmatched Intoxication on Perceptions of Consent in Ambiguous Alcohol fueled Sexual Encounters, examined how the amount of alcohol consumed by each participant affected perceptions of consent, coercion, sexual assault and responsibility. Across four studies involving 535 participants, the researchers presented participants with different hypothetical sexual encounters involving alcohol. [See Abstract and Results section, Journal of Interpersonal Violence, volume 38, pp. 11445–11474.]

The study found that participants assessed otherwise similar scenarios differently depending upon whether the alcohol consumption of the two people was matched or substantially different. [See Results and Discussion sections.]

Other recent research demonstrates another problem: the wording of the question itself.

In A Qualitatively-Informed Approach to Examining the Construct Validity of Alcohol-Involved Sexual Violence, researchers used cognitive interviews to examine how people understood different descriptions of alcohol related sexual experiences. Participants were asked to distinguish between concepts such as being ‘too drunk to consent’ and someone ‘taking advantage’ of them while drunk. [See Abstract, Results and Discussion sections.]

The researchers found substantial differences in how participants interpreted those phrases, including differences between impairment and incapacitation and differences in whether participants understood the wording as describing a victim’s state or another person’s conduct.

This matters enormously when we start generating population statistics from questionnaires.

If two people can read essentially the same description of a sexual encounter and categorise it differently, then the wording of the questionnaire becomes part of the measurement.

That does not make the research worthless.

It means we need to understand what the research is actually measuring.

There is another layer which is even more uncomfortable.

People do not interpret sexual experiences in a vacuum.

They interpret them through relationships, friendship groups, cultural expectations, social media, political movements, prevailing theories about sexual consent and the language available to them.

The academic literature on sexual scripts has been examining this for decades.

More recent research by Kathryn Rittenhour and Michael Sauder, Identifying the Impact of Sexual Scripts on Consent Negotiations, examined how dominant and alternative sexual scripts influence sexual communication and consent negotiations. The researchers conducted 51 in depth interviews across heterosexual, queer, non monogamous and kink communities. [See Abstract and Findings section, Journal of Sex Research, volume 61, issue 3, pp. 454–465.]

The researchers found that traditional sexual scripts remained common and that adherence to those scripts could affect communication around consent. They also found that alternative scripts could encourage greater sexual agency and communication.

This cuts both ways.

A woman can genuinely fail to recognise that what happened to her was coercive or non consensual because she has internalised a script which tells her that sexual pressure is simply part of heterosexual relationships.

That is an important insight.

But the converse deserves consideration too.

A person can also subsequently interpret an ambiguous or regretted sexual experience through a social framework which encourages them to understand themselves as having been victimised.

That possibility is not merely some internet conspiracy theory.

Academic researchers have explicitly investigated it.

Samuel Demarchi, Frédéric Tomas and Laurent Fanton published False Rape Allegation and Regret: A Theoretical Model Based on Cognitive Dissonance in Archives of Sexual Behavior in 2021. The paper proposes a theoretical model explaining how regret following consensual intercourse might, in certain circumstances, contribute to a false rape allegation through cognitive dissonance. The authors develop thirteen hypothetical scenarios illustrating their model. [See Abstract, sections on Regret and Cognitive Dissonance, and the thirteen scenarios, pp. 2067–2083.]

Their argument is not that this explains rape allegations generally.

It is that regret, shame, social norms and psychological conflict may, in some circumstances, contribute to a subsequent reinterpretation of consensual sex as rape.

A 2020 case study by Demarchi, Tomas, Franchi and Fanton, Cognitive Dissonance and False Rape Allegations: A Case Study, explored the possibility through an individual case involving consensual intercourse which was subsequently alleged to have been rape following regret. [See Abstract and Case Report section.]

Importantly, a case study cannot establish how frequently such a process occurs.

Nor can a theoretical model establish a prevalence rate.

That qualification matters enormously.

We should not turn an interesting psychological mechanism into another simplistic statistic.

There is, however, a fascinating point here.

If an individual is repeatedly told that a particular type of sexual experience is evidence of victimisation, that social interpretation may potentially affect how the person understands their own experience and may create a victim.

Equally, a person may be pressured by friends or social groups in the opposite direction and told that something which felt wrong to them was merely normal sex.

The literature on sexual scripts demonstrates that people do not make sense of sexual experiences independently of their social environment. [See Rittenhour and Sauder, Identifying the Impact of Sexual Scripts on Consent Negotiations, Abstract and Findings.]

That means the process by which an experience becomes a reported ‘rape’ is not necessarily a simple process of remembering a fixed historical fact.

Sometimes it may be.

Sometimes it may be a straightforward account of a violent crime.

Sometimes it may involve genuine uncertainty.

Sometimes it may involve impaired memory.

Sometimes it may involve regret.

Sometimes it may involve coercion that the person did not initially understand as such.

And, in a proportion of cases identified by research, it may involve deliberate fabrication.

The difficulty is that statistics can flatten all of those very different human experiences into one number.

This is why the language matters so much.

There is an enormous difference between saying:

‘X million people reported experiences which the survey classified as rape or sexual assault’

and saying:

‘X million rapes occurred.’

The first is a statement about what respondents reported and how researchers classified their answers.

The second is a statement about objective criminal events.

Those are not the same thing.

The criminal courts know this distinction instinctively.

A criminal trial does not ask whether the complainant’s account fits a statistical category. It asks whether the prosecution has proved the elements of the offence beyond reasonable doubt.

That is an extraordinarily high threshold, deliberately so.

A person can have experienced something deeply unpleasant without the conduct constituting rape in law.

A person can also have experienced rape without there being sufficient evidence to prove it in court.

And a person can make an allegation which is ultimately demonstrated to be false.

These propositions can all be true simultaneously.

The problem arises when one of them is allowed to swallow the others.

For years, we were told that we underestimated sexual violence, particularly where the complainant knew the alleged perpetrator, had been drinking, had not physically resisted or continued contact afterwards. Campaigns have tried to correct these alleged misconceptions.

They claim the correction was necessary.

But if it was a correction it’s can be argued that this is an overcorrection.

It is possible to take seriously the crime of rape without assuming that every allegation represents an objectively established rape.

It is possible to recognise underreporting without treating every unreported experience as a known criminal offence.

It is possible to recognise the reality of sexual coercion without eliminating the concept of individual agency.

And it is possible to recognise the vulnerability of intoxicated people without pretending that drunkenness creates a simple binary between consent and rape.

The fundamental problem is that we are attempting to put numbers on something which contains extraordinarily complicated questions of memory, language, psychology, social expectation, legal definition and individual behaviour.

That does not mean we abandon statistics.

It means we interrogate them.

Kanin’s 41 per cent should not be treated as gospel.

The modern 2 to 10 per cent figure should not be treated as gospel either.

Nor should a survey estimate of millions of ‘rapes’ be presented as though millions of criminal offences had been independently established.

The honest approach is much less satisfying.

It requires us to say that rape is real, that some victims never report it, that some allegations are false, that some experiences are difficult to classify, that definitions have changed, that alcohol complicates both capacity and interpretation, that social expectations influence how sexual experiences are understood, and that the difference between an allegation and a proven offence matters.

Most importantly, it requires us to resist the temptation to turn a morally important subject into a numbers game.

Because once we start treating every statistical estimate as an established fact, we are no longer doing science.

We are doing advocacy with numbers.

And that is a dangerous place for any criminal justice system to be where both the risk and actual miscarriages of justice has never been so real.

Edward Graham