September 30, 2012

The taint of a false confession

Ripple effects bias parties, contaminate "independent" evidence 

Michael Crowe, age 14, falsely confessing to murdering his sister
With the recent tidal wave of scholarly research into false confessions, informed forensic psychologists are by now tuned in to the phenomenon. We know, for example, that they played a role in one out of four DNA exoneration cases. We are aware of their compelling nature, and can cite examples such as the Central Park Jogger case in which they produced profound miscarriages of justice.

But let's take it one step further. What if, once police elicit a false confession from a suspect, it contaminates everything and everyone in touches -- from the prosecutor, the judge, and even the suspect's own attorney all the way to the fingerprint identification and even, perhaps, the DNA match?

That is the troubling thesis raised by Saul Kassin, a pioneer in the psychological study of false confessions, in an article in the current issue of the American Psychologist.
"Corroboration inflation"

Research shows us that such a contaminating effect is plausible. For example:
  • Fingerprint experts who were told the suspect had confessed were more likely to change their opinion and make an incorrect match, as compared with experts who were told the suspect was already in custody at the time of the crime. (1)
  • Polygraph examiners were significantly more likely to opine that an inconclusive chart showed deception when they were told the suspect had confessed. (2)
Bizarre case of multiple false confessions and prosecutions
Such findings may extend to other forensic science that requires subjective judgments, Kassin argues, including comparative analyses of ballistics, hair and fiber, shoeprints, tire tracks, handwriting and even DNA. Although CSI-style TV shows portray such evidence as infallible, a 2009 study by the National Academy of Sciences found widespread errors and bias in the collection and analysis of evidence.

That's not to mention egregious cases of intentional fraud in forensic laboratories that pop up with alarming regularity, such as a case in Boston, Massachusetts currently garnering headlines. There, a lab worker with allegedly bogus credentials as a chemist intentionally fabricated positive drug test results. Over a 9-year period, Annie Dookhan tested an estimated 60,000 drug samples confiscated from about 34,000 criminal defendants. Dookhan reportedly admitted writing reports listing samples as positive for illicit drugs even though she had never tested them; sometimes, "if a sample tested negative, she would take known cocaine from another sample and add it to the negative sample to make it test positive for cocaine," according to the Huffington Post's account. Dookhan has been arrested and the lab is temporarily shuttered.

Kassin points to an archival study conducted by he and two colleagues which found that, in DNA exoneration cases, false confessions were often accompanied by other errors, including improper forensic science, mistaken eyewitness identifications and/or the testimony of dishonest informants. Importantly, the confession preceded the other case errors in two-thirds of cases, suggesting it may have had a corrupting influence.

Such findings suggest that the legal system's longstanding assumption that independent sources of evidence provide confirmation of a suspect’s guilt may be wrong. Rather, Kassin writes, "confessions can spawn other incriminating evidence, creating an illusion of corroboration":
Amanda Knox, wrongly convicted in Italy
"Supported by 100-plus years of basic psychology and the research reviewed herein, confession-induced corroboration inflation challenges a core premise in law. Both pretrial corroboration requirements and a harmless error analysis on appeal rest on the assumption that the corroborating evidence on record is nonredundant and independent of the confession. It now appears that this assumption is often incorrect, that the other evidence may be tainted by confession, and that the appearances of corroboration at pretrial and the sufficiency of evidence on appeal may be more illusory than real."
"Hollywood productions"

Especially pernicious is the frequent situation in which police -- either intentionally or inadvertently -- feed an innocent suspect information that only a guilty party should know. Taking on the aura of a carefully scripted movie production with the confession as the central plot device, the confession is carefully drawn out of the suspect over hours and even days until in its final version it includes vivid details and plausible motivations.

Such an account proves virtually impossible for a judge or a jury to discount. The scripted confession thus becomes the be-all, end-all of the case, contaminating the minds of all who are exposed to it:
  • POLICE close the investigation, deem the case solved, and overlook exculpatory information, even when (as Richard Leo and his colleagues have shown) the confession is internally inconsistent or contradicted by independent evidence.
  • PROSECUTORS stubbornly cling to false confession cases, refusing to admit the possibility of their falsity even when DNA testing unequivocally excludes the confessor. (The New York Times Magazine has more on this phenomenon, describing -- in an article titled "The prosecution's case against DNA" -- the improbable arguments manufactured by prosecutors to explain away negative DNA findings.)
  • Perhaps most dangerously, even DEFENSE ATTORNEYS succumb to the allure. Individuals who falsely confess are much more likely to be pressured into accepting a guilty plea, which bars future appeals. In an archival study conducted by Kassin and a colleague of 273 DNA exoneration cases, those based on false confessions were three times as likely to involve bad lawyering.
Matias Reyes, the actual rapist
in the Central Park Five jogger
wrongful conviction case

"Taken together," Kassin concludes, "research suggests that judges, juries, and others are doomed to believe the false confessions of innocent people not only because the phenomenon strongly violates common sense but because of corroboration inflation -- a tendency for confessions to produce an illusion of support from other evidence."

All of this suggests that it is essential for courts to allow the testimony of forensic experts who can explain the mechanisms of false confessions, including both what types of police practices are more likely to generate them, and what types of individual vulnerabilities make a person especially prone to cave in under such pressure.

More broadly, this line of analysis suggests the need for changes in police practices, for example an end to the routine practice of lying to suspects about incriminating evidence, and greater government oversight and regulation of police interrogations. Moreover, safeguards on the analysis of supposedly independent evidence, such as evidence technicians being blind to a suspect's confession status, must be implemented in order to ensure that corroborating evidence truly is independent.

The article is: "Why confessions trump innocence." Members of the American Psychological Association may download it for free as part of their member benefits; others may request a copy from the author (HERE).

Related blog posts:

For a complete list of my many other posts on the topic of confessions and interrogations, click HERE.


(1)   Dror, I. E., and Charlton, D. (2006). Why experts make errors. Journal of Forensic Identification, 56, 600–616.
(2)   Elaad, E., Ginton, A., and Ben-Shakhar, G. (1994). The effects of prior expectations and outcome knowledge on polygraph examiners' decisions. Journal of Behavioral Decision Making, 7, 279–292.

Hat tip: Tim Derning

No comments: