Jan 20 2015

False Confessions

When a suspect confesses to a crime, that is a powerfully persuasive piece of evidence. People assume (reasonably) that no one would confess to a crime they didn’t commit, extrapolating from the conviction that they would never do so themselves.

Of course there are legal reasons why someone accused of a crime would confess to guilt, in order to plea to a lesser charge or sentence and avoid the risk of a much more serious sentence. It is an unfortunate reality that innocent people can be wrongly convicted, and therefore an innocent person might confess in order to avoid the worst consequences of such a wrongful conviction.

Would, however, someone confess not to secure a plea bargain but simply because they are being aggressively interrogated? New research suggests that it is possible to form false memories of having committed crimes that one never committed. This is not surprising given all the previous memory research showing  how easy it is to form false memories.

In this study college students were interviewed and asked to recall two events, one of which was a true event from their teen years, and another was a false event invented by the researchers. The false events were either of the subject committing a crime, or of an emotional event, such as being attacked by a dog. The researchers had first contacted the caregivers of each of the subjects and gave them a series of questions about the subject, so the researchers were armed with details about their teen years. The false stories had true details within them to give them more plausibility.

When asked to recall each of the two events (one true, one false) 21/30 (71%) of the subjects were able to “recall” the fake crime, including detailed descriptions and reenactments, and reported high confidence in the memory. They therefore internalized the fake crime as a false memory. A similar number, 76%, of the subjects came to believe the false emotional story.

The students were able to report more details for the true events, and had a higher confidence, but the recall and confidence in the false stories were also high.

Of course, this is one smallish study. Psychological studies like this are always contrived to some extent. This is more of a proof of concept than an indication of what will happen in real life situations. It is in line with a great deal of research, however, showing how easy it is to form false memories.

The implications of this research, if it holds up, are obvious. In any situation where someone is being interrogated or simply questioned it is possible for the interview process to form false memories, even of a crime the subject never committed.

This is already conventional wisdom in medicine. We are taught not to ask leading questions or suggest symptoms to patients. Also, if the patient has given their medical history to others previously this may have contaminated their history with suggested details. Since we often have the experience of multiple people taking the history from the same patient, and recording that history, we see first hand how histories can evolve with the retelling.

The key insight here is that memories are narratives, not just objectively stored facts. The narrative drives the details, more than the other way around. Memories are malleable, and will shift in order to support the narrative. This can happen even in the most benign setting, such as telling your history to a concerned health care professional who only wants to help you.

Imagine the process of a hostile interrogation. The interrogators are trying to trip you up, confuse you, and force mistakes. They make you tell your story over and over, and try to force your story into a narrative in which you are guilty. The assumption is that this procedure will only work if you are truly guilty. This is a naive assumption, however, and goes against decades of memory and witness research.

A similar problem exists when interrogating children who are potential victims. Young children may not fully distinguish reality from fantasy, and may also have an incredible desire to give the “right answer” and please adults. Any leading questions therefore invalidate the outcome.

I previously discussed the Reid technique of interrogation, which is popular in the US. This technique provides the suspect with two narratives, one worse than the other but both involving guilt. The goal is to make them confess to the lesser scenario, which is still a confession of guilt. The assumption (almost certainly wrong) is that an innocent person cannot be made to confess. In light of this and other research, the Reid technique should be questioned, if not abandoned.

An alternate technique is the PEACE technique which is not confrontational. It is based on the premise that lying requires a much greater cognitive load than telling the truth. Interrogators are not looking for a confession, they simply circle back through the suspect’s story over and over to see if the details hold up.

The increased cognitive load of lying seems to be a reasonable premise and a valid basis for interrogation techniques. The assumption that the guilty will act anxious or guilty, and the innocent cannot be made to confess are not valid assumptions and should be abandoned as premises for interrogation.

Given that the stakes for individuals can be extremely high, this is an area that deserves greater research. It also seems reasonable, given the current state of the research, that policy should err on the side of protecting the innocent from wrong conviction.

In general the research on human biases, memory, and the tendency to favor certain narratives indicates that we are always one step away from an outright witch hunt. The witch hunt is the psychological pathway of least resistance, and we need checks and procedures in place to guard against them.

This latest research adds further to this conclusion.

33 responses so far