How Innocence Puts Innocents At Risk

False confessions in criminal cases are very hard to reverse in the minds of police, jurors and judges, even when substantial evidence is available to show that the confession is false.

Saul Kassin of the John Jay College of Justice has a fascinating paper in American Psychologist [abstract free, $ for full article] on false confessions, in which he presents some startling findings on why innocent people confess.

Police often bluff to suspects that they have airtight evidence and the confession is just a formality to confirm what they can already prove. Kassin and his colleagues tried this same technique in experiments in which subjects who were typing were told that they had hit the wrong key and crashed a computer’s hard drive. When told that all keystrokes were recorded for later review, innocent subjects became more likely to confess to having hit the wrong key.

Apparently, innocent experimental subjects assumed there was no harm in falsely confessing as they would be exonerated by the airtight evidence later. In an actual criminal case, this same assumption by an innocent suspect can lead to a wrongful conviction when the police are only bluffing about the existence of the airtight evidence.

Author: Keith Humphreys

Keith Humphreys is the Esther Ting Memorial Professor of Psychiatry at Stanford University and an Honorary Professor of Psychiatry at Kings College Lonon. His research, teaching and writing have focused on addictive disorders, self-help organizations (e.g., breast cancer support groups, Alcoholics Anonymous), evaluation research methods, and public policy related to health care, mental illness, veterans, drugs, crime and correctional systems. Professor Humphreys' over 300 scholarly articles, monographs and books have been cited over ten thousand times by scientific colleagues. He is a regular contributor to Washington Post and has also written for the New York Times, Wall Street Journal, Washington Monthly, San Francisco Chronicle, The Guardian (UK), The Telegraph (UK), Times Higher Education (UK), Crossbow (UK) and other media outlets.

26 thoughts on “How Innocence Puts Innocents At Risk”

  1. Cops know this. That’s why good cop practice is to withhold information from the suspect that only a truly guilty person would know. If the confession does not contain this information, it isn’t reliable. Unfortunately, not all cops or prosecutors are good. With plea-bargaining, the judge and jury are “just a formality to confirm what” the cops and prosecutors are determined to do.

    1. Another one of the fascinating things about Kassin’s article is that he documents that even when cops have used deception to get a confession that may well be false, it dramatically changes their subsequent investigatory behavior. They look for evidence consistent with the confession and ignore leads that are inconsistent with it, including for example the fact that the subject doesn’t know about held back evidence that the true criminal would know. When the case goes to trial, juries do the same thing, *even when the judge says the confession was false, disregard it* juries tend to believe that no one who was really innocent would confess, so they base their verdicts on that assumption even when very strong disconfirmatory evidence is available.

  2. Are there any legal sanctions against cops and prosecutors who use deception to obtain confessions?

    Early in the first volume of Les Miserables, the Bishop of D___, the “just man” whose example transforms Jean Valjean, the book’s hero, into a just man, is told of a case in which a woman was tricked by the attorney for the crown into denouncing her lover when there was a lack of evidence of the crime. Most people admire the prosecutor for his resourcefulness. The bishop asks where the couple is to be tried, and is told “at the assizes.” The bishop then asks, “And where will the attorney for the crown be tried?”

    So where are the authorities tried in cases like these in the USA?

    1. Mark downthread is correct. Deception is generally kosher. But there is an exception to this. Deceptions that amount to threats are forbidden. You can’t tell a suspect that he will get beaten up if he doesn’t confess, even if you have no plan to beat him up. Closer to the line is falsely telling the suspect that the victim is dying, and the doctors need to know precisely what the suspect did to the victim in order to save the victim’s life. This is tantamount to a threat: “Don’t confess, and you will be guilty of murder rather than assault.” I think that the New York courts are currently considering this.

      To answer the rest of your question, prosecutors enjoy something akin to absolute civil immunity, at least within the scope of their duty. (Cops don’t.) Criminal immunity is kind of irrelevant–who is going to prosecute a prosecutor for doing something that prosecutors do all the time? I remember one fellow in New Jersey who might have been prosecuted for his prosecutorial misdeeds (they amounted to extortion), but he committed suicide when he was found out.

      The usual remedy of a defendant for prosecutorial misconduct is a retrial. It’s kind of a weak remedy, but it is the only common one. The usual grounds are suppression of exculpatory evidence.

  3. I took a class on precisely this topic in law school. While there were many compelling anecdotes of false confessions, nothing emerged during the term that suggested that false confessions are a systemic problem. And yes, I of course believe that any false confession is a problem, but as one of my grad school professors oft said, extreme cases make for bad policies.

    And I do think that there’s a slight difference between misremembering whether you hit one key amongst thousands of keystrokes and misremembering whether you murdered someone on a given night. But even in such a case, the corpus delicti rule does not allow conviction based on nothing more than the uncorroborated statement of a defendant.

      1. i did, and cannot tell what I should be looking for.

        Four of the first 5 listings are official sites (including the chamber of commerce and a newspaper) for Wenatchee, WA. The fourth listing is the Wikipedia entry, where nothing relevant jumps out at me. Then follow links to TripAdvisor, Craigslist more government sites, a local college…

        How about some additional information, ES?

    1. In the studies of false confessions that have been done, the percentage of miscarriages of justice caused by false confessions was found to be between 14% and 25% (p. 907 of the paper above). While there are obviously plenty of other causes, this is beyond “anecdotal” and not just limited to “extreme cases”; in fact, the evidence is pretty strong that false confessions contribute significantly to miscarriages of justice.

  4. Keith, I think this is a fascinating and appropriate topic for my favorite law blog, Concurring Opinions. You might want to write to one of their regulars; Howard Wasserman and Frank Pasquale are names that come to mind.

  5. I think that this particular research is of questionable value because a) by definition you only have innocent subjects and b) in research situations there is (more of) a relationship of trust between research and subject and c) the definition of “wrong key” doesn’t include mens rea. In fact, I think there’s an argument to be made that increased confessions in the supposed presence of a complete record here results from recognizing that the “wrong key” can be identified retrospectively even if the subject had no idea about it at the time.

    In some ways, of course, this is analogous to the “good cop/bad cop” routine where one member of the interrogation team develops a relationship of trust with the suspect and uses that to extract an admission of guilt, whether accurate or otherwise.

    Although deception seems not entirely out of line (and would be a nightmare to create bright-line rules against), the implicit threat of additional punishment if the suspect doesn’t cooperate by confessing does seem problematic. (In cases where a suspect/defendant can’t make bail, the mere prospect of going to a jury trial serves as a threat of additional punishment.) Similarly the issue of inducements for accomplices or jailhouse informants — technically, offering anything of value is unlawful, but from my spotty reading anything that’s within prosecutorial or administrative discretion isn’t considered of value for deciding whether testimony has been bought.

    1. Paul wrote “I think that this particular research is of questionable value because a) by definition you only have innocent subjects and b) in research situations there is (more of) a relationship of trust between research and subject and c) the definition of “wrong key” doesn’t include mens rea.”

      This is just one study mentioned in a review paper with many other studies which vary method substantially, as well as analysis of real criminal cases in which people falsely confessed. I wasn’t trying to review all the studies in this area, that’s what Kassin’s paper is for.

  6. Keith, aren’t you positing that people are more rational, and their attachment to “facts” more solid, than is in fact the case? You’re assuming that test subjects told of impeccable keystroke data confess because they think these data will clear them. But couldn’t they confess because they convince themselves that they must in fact have crashed the hard drive if the keystroke monitor says they did? Similarly: couldn’t suspects of weak will, after many hours of interrogation in an enclosed room with nothing but cops for company–I’m taking for granted that they’ve been tricked into waiving their right to counsel–could be influenced into thinking that if there’s airtight physical evidence against them they must have committed the crime and then somehow failed to remember it?

    1. Andrew, are you thinking of the Asch conformity experiments here?

      Related: Why, even if you’re innocent, you still want an attorney. Because (aside from the obvious benefits of having somehow along who knows how to deal with what is a frightening situation for most people) an attorney is what Asch calls a “true partner” and dramatically reduces the pressure to conform.

      1. As a philosopher who works on the epistemology of testimony, the usual interpretation of the Asch conformity experiments bugs me. It seems to me that, if there’s a big group of people in the room and everyone disagrees with you about a simple observational matter, it’s rational to suppose that they’ve got it right. Unless you have some sort of special expertise or other explanation as to why everyone else might be systematically deceived, the chance that you’re making a mistake is greater than the chance that everyone else is.

        That particular quibble doesn’t have anything to do with Andrew’s point, though.

      2. Katja: As Kassin discusses, in real criminal cases, people who are in fact innocent are the ones least likely to ask for an attorney, because they know they are innocent, why would they need one? But of course, they DO need one.

        1. Yes. I remember a discussion I had with my godfather (who is a criminal prosecutor in Germany). He told me that if I were ever to be arrested for a crime (the assumption was that I would be innocent), I should immediately get a lawyer and NOT talk to the police or the prosecution. His argument, as I recall it, was along the following lines: “You are innocent. Yet, you were arrested. That means that something went very, very wrong with the investigation, and you need competent advice by an expert ASAP.”

        2. So out of curiosity, what is typically the recommended approach to a typical person finding himself in an interrogation room?

          I’m certainly not a hardened criminal, but it’s certainly possible that I could find myself busted for something or other, whether guilty or innocent. I’m comfortably middle-class, and could easily afford to hire an attorney if the need was urgent, but I obviously don’t “have a lawyer.”

          So in extremis, should someone like me ask for a public defender? Use my one call to call my wife and have her hit our contacts to find an attorney? Yellow pages? What?

          1. A competent lawyer would likely advise an arrestee to say nothing to police; any information which may be helpful can be relayed through counsel at a later time. The call to your wife to arrange for a lawyer with experience in criminal defense is the prudent thing to do. Assume that the phone conversation is being recorded. Do not speak of what circumstances led to the arrest. If you know what the charged offense or the amount of your bail is, tell her. (That will facilitate the process of finding counsel.)

    2. You would have to be an extraordinary combination of naive and cynical to believe that

      1) there is “airtight” evidence that exonerates you, but
      2) the police do not recognize this fact and instead consider the evidence to be proof of your guilt, yet
      3) once you have confessed and then repudiated your confession this evidence will be turned over to your defense team in a form that a jury will agree exonerates you.

      Occam’s razor would suggest instead that “we have airtight evidence” is, like “we have a keylogger” is actually a signalling mechanism that the people in authority intend to blame you for the bad event regardless, and thus you might do better by confessing than by continuing to oppose (and presumably anger) them. If you don’t trust the authorities, the announcement of airtight evidence (or of a keylogger) is an assertion of control over the story that will ultimately be released to the public or a jury.

      By the way, is there research on the level of punishment meted out to otherwise-equivalent suspects depending on whether they cop a plea or hold out for trial?

      1. I don’t know what the research says, but the prosecutors have arranged things so the gap is enormous. If a prosecutor wants you to go to jail for 2-3 years, s/he can probably threaten you with a good enough chance of a 20-year sentence so that you will plea out, even if you’re innocent and have a good shot of persuading a jury of this.

        It’s hard to find equivalence among suspects for a number of reasons. First, the prosecutor will typically charge plea-bargainers more lightly than folk who go to trial: manslaughter rather than murder, for instance. Second, it’s hard to put a metric on the heinousness of a crime. Third, judges have some discretion over sentencing these days, although charging belongs to the prosecutor. A fourth confounding factor is strength of evidence. Prosecutors hate to lose at trial, and will give a better bargain to defendants with better cases. A fifth confounding factor is snitching–people who don’t snitch upon demand can get pounded: plea or not. (Snitching is not a factor in all cases, of course.)

        I’m just a lawyer; I’m sure a criminologist can handle your question better. But the nub of the matter is that except for wealthy defendants, the prosecutor has almost all the cards.

    3. Andrew’s explanation seems right to me. If I were convinced that I hadn’t hit the wrong key and crashed the system, but the experimenters told me I had and they had airtight proof, I’d confess. Because I just don’t type that well, and I hit wrong keys all the time.

      I can imagine that someone who had been in an interrogation room for fifteen hours without sleep might be similarly confused. I can imagine that I’d be similarly confused, if I had been interrogated for hours and hours and hours, and the police told me they had airtight proof that I had done the crime I didn’t think I had done. After fifteen hours (or whatever) I’d start to doubt my memories. After all, everyone has the experience of “remembering” something that turns out not to be true.

    4. Hi Andrew: Kassin discusses this phenomenon in his paper as well, and it is certainly real. The book Remembering Satan includes vivid examples of how this happened in a child sexual abuse case.

      That said, these are non-competing explanations — there is no logical contradiction between some false confessions being false in the eyes of the confessor, and others not being.

Comments are closed.