Can a reasoning test predict who will make a good detective?

By Alex Fradera
Although criminal investigation has been transformed through technological developments in DNA, phone tracking, and online data, the way a detective works through a crime has remained much the same. The first suspect is often the true perpetrator, but not always, and snowballing biases continue to lead to miscarriages of justice. Proficient detectives need the ability to generate and evaluate different explanations and keep an open mind. New research in the Journal of Police and Criminal Psychology investigates whether it’s possible to use established tests of reasoning ability to identify who has the skills necessary for thinking this way.
University of Gothenburg psychologists Ivar Fahsing and Karl Ask asked 166 recruits (60 women, aged around 23) from the Norwegian Police University College to complete two tests of reasoning ability. One involved deductive reasoning – the ability to apply rules to reach a correct conclusion, in this case, combining shapes to form a new target shape. The other involved inductive reasoning – viewing different images, and using these to figure out the rule that is governing them all. These tests already feature in Norwegian police recruitment but it’s not clear whether they are useful for predicting detective skills, which are more dependent on abductive reasoning, which is much harder to test.
Whereas deduction seeks linear, definitive conclusions, and induction tries to identify a category based on the available clues (“all birds, none fly” therefore the category must be “flightless birds”), abduction amplifies the known information to generate imaginative possibilities (“an open wine bottle and smashed glasses could suggest a thwarted seduction… or a failed reconciliation with his wife?”). It requires a leap of logic, a creative act, hard to measure but known to differentiate stronger investigators, whether criminal or scientific. The ability to make many such leaps helps avoid premature foreclosure on the possibilities of a case.
Fahsing and Ask were hopeful that measures of deductive and inductive reasoning would help identify better detectives because investigation isn’t rooted solely in abduction, and because different forms of reasoning ability are known to correlate (so a candidate skilful at deduction and induction ought to be skilled at abduction and therefore detective thinking too).
To assess detection nous, the researchers asked their participants to review two case vignettes, each describing a woman missing under ambiguous circumstances, and some versions also mentioned the arrest of a suspect. The participants’ task was to outline all the possible explanations that could account for the facts. Both vignettes had been reviewed by an expert panel of detectives, who generated a “gold standard” of 9 viable hypotheses for the first and 11 for the second.
The participants tended to neglect the non-criminal hypotheses, such as the possibility the missing woman had suffered an accident. More importantly, scores on the reasoning ability tests didn’t predict their detective performance. Nor did they predict immunity to a commonly observed detecting error: becoming more narrow-minded (generating fewer hypotheses) after reading that a suspect had been arrested.
The current data suggest that commonly used cognitive assessments are unlikely to help identify candidates with the best aptitude for becoming a detective. It seems police forces will need to think more inventively about the actual skills that contribute to good detection if they are to succeed in separating the Maigrets from the Clouseaus.
— In Search of Indicators of Detective Aptitude: Police Recruits’ Logical Reasoning and Ability to Generate Investigative Hypotheses
Image: Peter Sellers In ‘The Pink Panther Strikes Again’ is disguised as a mountain climber, while hiding in a trash can, in a scene from the film ‘The Pink Panther Strikes Again’, 1976. (Photo by United Artists/Getty Images)
Alex Fradera (@alexfradera) is Staff Writer at BPS Research Digest