Psychological research suggests more deep flaws in latest search strategy for Madeleine McCann?
Raj Persaud and Adrian Furnham
Release of the new images coincides with the broadcast of a new televised reconstruction of events on 3rd May 2007. Investigators now claim the timeline and “accepted version of events” surrounding Madeleine’s disappearance have significantly changed.
The police’s new strategy dominates the headlines, particularly in the UK media.
The same pattern of saturation coverage followed the release in May 2013 of a new ‘age-progressed’ picture of Madeleine McCann. However, we previously reported how scientific research on such ‘aged-progressed’ images suggests this approach might in fact harm recognition chances.
But the press were convinced a new breakthrough was imminent then, as it is again, now.
The study we reported on; ‘Age-progressed images may harm recognition of missing children by increasing the number of plausible targets’, by Steve Charman and Rolando Carol, from Florida International University, found that age-progressed images were not just simply decreasing the likelihood of recognizing anyone, they seemed to be systematically leading people away from recognizing the target (and toward mistakenly ‘recognizing’ non-targets).
So what does the latest psychological research reveal about the new strategy? Is it similarly flawed?
Professor Elizabeth Loftus, from the University of California, has recently pointed out that public appreciation of deep psychological pitfalls involved in eyewitness reports, only really came about as a result of progress in forensic DNA testing.
Her paper entitled ’25 Years of Eyewitness Science… Finally Pays Off’, points out DNA technology helped exonerate many wrongfully convicted individuals in the mid 1990s; today over 300 innocent people owe their freedom to such DNA testing. Elizabeth Loftus demonstrates in her paper, published in the academic journal ‘Perspectives on Psychological Science’, that faulty eyewitness testimony was a major factor in wrongful convictions, present in roughly three quarters of those cases.
Psychologists Charlie Frowd, Derek Carson, Hayley Ness, Dawn McQuiston-Surrett, Jan Richardson, Hayden Baldwin and Peter Hancock, recently conducted a study seemingly relevant to the current Madeleine McCann search strategy. They tested how new high-tech ‘composite’ pictures of a person the police want to interview, actually performed in correctly identifying targets.
In the experiment, recently published in the academic journal, ‘Legal and Criminological Psychology’, ‘witnesses’ inspected a photograph of a celebrity for 1 minute, then just 2 days later constructed a ‘composite’ image from one of these ‘photo-fit type’ systems, using procedures closely matching real police work; for example, the use of a specialised ‘cognitive’ interview (used to enhance a witness’s memory of a suspect) and computer operators/artists appropriately trained and experienced.
The experiment introduced just a two day gap between witnessing the relevant face, and producing the composite picture, while in the Madeleine McCann case, several years have now elapsed.
The study entitled ‘Contemporary composite techniques: The impact of a forensically-relevant target delay’, tested how good the composite picture generated by witnesses actually was, by asking independent observers to correctly identify and name the ‘composite’ images. Several different computerised systems, many used in Forensic contexts, were included – E-FIT, PROﬁt, FACES and EvoFIT.
The study from the Universities of Stirling and Abertay in the UK, and Arizona State University West, USA, found correct identification of the celebrity was surprisingly low (3% overall), with old-fashioned sketches named best, with an accuracy rate of 8%.
The authors conclude that after a 2 days delay to construction, few composite pictures would be correctly identified in police work. Their experiment provides further evidence that the computerized systems (so beloved by the media because it seems advanced computerised technology must be giving the police a new edge) tested inferior to manually-generated sketches.
The authors of the study suggest that possible reasons why more old-fashioned manual sketching might in fact be superior to more modern computerised composites, may be something to do with the psychology of memory for faces. It may be that witnesses sometimes remember a face as a set of particular features, yet on other occasions perceive a countenance as a whole. More ‘holistic’ memory encoding of faces appears best for face recognition and composites constructed via a sketch artist.
Time delays may mean memory for faces is more of a whole impression than an analysis of individual features. The issuing of a computerised image could be a flawed approach.
But another problem is not just the time delay in the Madeleine McCann case but the very way the media has reported the crime could have damaged the possibility of correct identification.
This is based on a part of psychology which finds that expectations, such as the pre-existing beliefs of guilt or innocence of a suspect, dramatically influences subsequent identification decisions. Yet some sections of the press have already gone as far as suggesting the image of the man wanted for questioning by the police, might be the perpetrator of the crime.
Psychologists Steve Charman, Amy Hyman Gregory, and Marianna Carlucci from Florida International University examined how pre-existing beliefs of guilt influence similarity ratings between a suspect and a facial composite image, similar to the kind just released in the Madeleine McCann investigation.
Beliefs in a suspect’s guilt inflated similarity ratings, and beliefs in a defendant’s guilt predicted similarity ratings. The study entitled ‘Exploring the Diagnostic Utility of Facial Composites: Beliefs of Guilt Can Bias Perceived Similarity Between Composite and Suspect’ highlights a deep problem in using facial composites as evidence against a suspect, because of the malleability of similarity judgements.
The study, published in the ‘Journal of Experimental Psychology’ was partly inspired by an infamous case in the USA where Vishnu Persad (no relation to one of the authors of this article) was eventually found innocent of a crime for which he was convicted. He turned out not to be the person represented in the witness-generated facial composite image. Yet an FBI veteran expert witness, plus an established prosecutor (and, presumably, the jury in the original trial) seemed to have perceived extremely high levels of similarity between Vishnu Persad, and the composite image of the actual perpetrator.
Vishnu Persad spent two years in prison before all charges were dropped at a later retrial. “. . . completely and uncannily consistent” was how Prosecutor Moira Lasch, described the similarity of the composite image to Persad.
To explain the psychology of this effect the authors of the study ask readers to consider the following sequence of letters—hijkl—and the following sequence of numbers – 5432l. Although the last character in each of the two sequences is objectively identical (“l”), most people’s expectations lead them to read it differently—as the lower-case letter “L” in the former and the number “one” in the latter. The ambiguous character “l” is perceived in accordance with the expectation that it would represent either a letter or a number.
Expectations influence what we see.
Psychologists Hans Crombag, Willem Wagenaar and Peter Van Koppen investigated Dutch residents’ memories of an El Al Boeing 747 crashing into an 11-story Amsterdam apartment building. Media coverage never actually depicted the collision itself, but included the aftermath – ensuing fire and rescue operations.
Yet 10 months afterwards, the study entitled ‘Crashing memories and the problem of ‘source monitoring” found that over 60% of respondents reported having actually seen the crash on TV. The investigation published in the journal ‘Applied Cognitive Psychology’ found many gave vividly detailed descriptions of a cataclysm they could not have actually witnessed, even via television. Women were more prone to these false memories than men.
The authors, from the University of Limburg, Leiden University and The Netherlands Institute for the Study of Criminality and Law Enforcement, attributed these false memories to the effects of widespread media descriptions of the crash, leading participants to visualize it. Later, the public probably falsely attribute the source of these images to media coverage of the incident, rather than their own imaginations.
Those launching this new search strategy for Madeleine McCann need to be aware of these various psychological pitfalls. Otherwise false arrests and wrongful convictions could be the result, or another dead-end.