As number of sex crimes rises but conviction rate falls, 'lie detector' could help police weed out bogus claims

By Jason Bennetto, The Independent, UK, 31 October 2000

A "lie-detector" test to help uncover false allegations of rape is being developed by the police and criminologists.

Between 10 and 41 per cent of allegations of rape are made up by the "victim", according to previous research. In the new test, a claimant's statement is analysed and points are given from a list of set clues - people who have made up a rape allegation get a low score, while genuine complaints get a high score.

Using this technique in two studies, police officers and researchers had a success rate of between 72 and 100 per cent in identifying genuine rape victims from liars. All the cases were real-life examples selected because of the existence of strong evidence, such as closed-circuit television footage, to prove the defendants' guilt or innocence.

The two pieces of research indicate that police officers who rely on their detective skills and intuition when examining a statement by an alleged rape victim are no better than a member of the public at identifying a genuine complainant from a false one.

The researchers found that the three main motivations for reporting false rape are

• to provide an alibi where the accuser believes she may be pregnant or has had illicit consensual sex,
• revenge against a partner, and
• a desperate attempt to seek sympathy or attention.

The National Crime Faculty, which develops new investigative techniques, is studying the rape testing system, known as statement validity analysis (SVA). This form of analysis is already used by a handful of officers in Britain. Criminologists hope it will help detect malicious complaints - which can ruin the reputations of the people falsely accused - to allow police to concentrate more on genuine cases.

The small number of false allegations that make it to court have harmed the campaign by women's groups, the Home Office and police to increase the conviction rate for rape, which fell from 24 per cent in 1985 to 9 per cent in 1997.

In the year ending March 2000, the number of rapes recorded by police in England and Wales increased by 9.5 per cent to 7,809 offences involving a female victim, and by 19 per cent in male rapes to 600.

The two studies adapted SVA, which was first developed to test allegations of child abuse, to help assess rape cases. Points were given for different aspects of the complainant's statement including how it was structured, what details were included and excluded and the type of language and description used. The researchers say that genuine rape victims include unexpected details and structure their statements in a different way from false claimants.

In one of the studies, Andrew Parker, a Metropolitan Police officer who works at the Institute of Psychiatry in London, and Jennifer Brown, professor in the psychology department at the University of Surrey, tested 43 statements of alleged rape victims from London, 38 of which were made by women.

The testing system successfully predicted all the eventual outcomes of their cases -

• 24 were believed to be "credible",
• 16 "non-credible", and
• three undecided.

Of the non-credible cases, the police later found that

• six involved serial allegers suffering from psychiatric disturbances,
• four involved complainants who admitted making malicious allegations, and
• two were made by delusional alcoholics.

Their report, "Detection of Deception", published in the Legal and Criminological Psychology journal, concluded:

• "SVA is not intended to replace the detective, but it can be used as a significant aid in the assessment of complex and emotional victim and witness testimony."

In the second study, Detective Sergeant Chris Few, working at the Centre for Applied Psychology at Leicester University, tested the system on 32 police officers. He trained half in the new analysis technique, which uses a check list of 15 criteria or clues when studying a statement. They examined four statements, of which four were genuine and four were false. The officers using the system correctly predicted the validity of the statements in 72 per cent of the cases, while those using their own judgement were accurate in just over half.

• DS Few said:
  "Virtually all the training police officers are given to assess whether someone is telling the truth, focuses on verbal and non-verbal behaviour such as body language. Most research suggests this does not work.
  Police officers are no better than anyone else at telling whether someone is lying or not about a rape allegation. Most have about the same chances of success as tossing a coin."
  He added:
  "The tests need refining further but they should be able to help investigators better assess the credibility of witnesses in serious sexual offences."