A Collection of Strange Beliefs, Amusing Deceptions, and Dangerous Delusions

From Abracadabra to Zombies


Skeptimedia is a commentary on mass media treatment of issues concerning science, the paranormal, and the supernatural.

»Skeptimedia archives

Skeptimedia replaces  Mass Media Funk and Mass Media Bunk. Those blogs are now archived.

Evaluating Evidence

5 Feb 2009. Of all the courses I taught in my 32-year career as a philosophy teacher, my favorite was the philosophy of law. I designed the course and called it Law, Justice, and Punishment (LJP). The course was for community college students, some with career goals in law or law enforcement. Most of my students, however, seemed to sign up because of the course title and the content it suggested. LJP students were my best students, too, which partly explains why it was my favorite class. What I really enjoyed about LJP, though, was the way I could teach critical thinking skills in a concrete way, as well as explore many interesting philosophical issues with my students.

LJP afforded me the opportunity to use news-of-the-day items to illustrate particular legal issues and general critical thinking problems in the law. There is scarcely a topic or an issue that we took up in the course that didn't involve some aspect of critical thinking. Some are obvious, such as the problem of  confirmation bias during an investigation. Detectives who make decisions about guilt or innocence too early in an investigation often err because they close themselves off to alternative hypotheses. Like most of us, detectives don't realize how easy it is to find support for one's hypothesis when one systematically excludes alternatives from consideration.

We also discussed problems with perception and memory, including problems brought out by the exercise of staging a crime for the students and having them write down what they saw. We reviewed the role of suggestion in affecting eyewitness testimony. Students were surprised that eyewitness testimony is the least reliable but most influential form of evidence. They were not surprised that different people witnessing the same event would differ in their memories, but they were impressed with how widely eyewitness accounts differ. Students were not surprised to know that coaching a witness during a line-up, even in very subtle ways, can influence an eyewitness's testimony. Students immediately understood the problem with placing a single picture in front of an eyewitness and asking "is this the guy?"

Most of my students assumed that a confession is the most reliable form of evidence. Some, however, had had more than a passing relationship with the legal system and understood the role of coercion in extorting confessions. In the general critical thinking course we list argumentum ad baculum or the appeal to fear as one of the fallacies to learn. In LJP, we studied case law involving various ways prosecutors can frighten a person into confessing to a crime he did not commit. Torture isn't necessary when you are gifted with the tools of plea bargaining and vast resources.

News stories with a similar theme appeared with increasing frequency over the years I taught LJP: someone who had been convicted on the basis of eyewitness testimony or a confession had been exculpated (relieved of guilt) by DNA evidence. By the time I retired in 2007, the Innocence Project had brought about the release of many prisoners who had been wrongly convicted of serious crimes like rape and murder. Mistakes in eyewitness testimony account for 75% of these wrongful convictions and "25% of wrongful convictions later overturned through new tests on old DNA evidence" involved false confessions.* (The count is now over 200. Does the data represent the tip of the iceberg or are these cases outliers?)

I regret that over the course of fifteen years that I taught LJP, I didn't think of the problem that confessions might play on eyewitness testimony. An article by Lisa E. Hasel and Saul M. Kassin demonstrates what now appears obvious to me: when an eyewitness is told of a confession, that information taints the evidence. The article is called "On the Presumption of Evidentiary Independence: Can Confessions Corrupt Eyewitness Identifications?" Knowledge of a confession alters eyewitnesses' identification decisions. This seems obvious. If I have identified a suspect and then am told that the suspect confessed, my confidence in my identification should increase. What's not so obvious is that if eyewitnesses have looked at a photo lineup and could not identify anyone as the perpetrator of the crime witnessed, when told that a particular suspect in the lineup has confessed, 50% of the eyewitnesses change their mind and now identify the confessor as the perpetrator.

Hasel and Kassim found in their study of 206 subjects (undergraduates) that 84% of the students who witnessed a staged crime identified the perpetrator when shown a photo lineup where the perpetrator was not included. This is good news for criminals who aren't caught, but not very encouraging for innocent suspects.

Even experts can be tainted by knowing of confessions:

Recent empirical research also suggests that confessions can corrupt other evidence. Dror and Charlton (2006: Why experts make errors, Journal of Forensic Identification; Volume 56, 2006.) re-presented fingerprint experts with pairs of prints from a prior case. Instructed that the suspect had confessed (which suggested a match) or was in custody when the crime was committed (which suggested an exclusion), the experts changed 17% of their previously correct decisions.*

Hasel and Kassim conclude:

There are two ways for confessions to corrupt other evidence. Most noticeably, they may enable police to procure corroboration through additional confessions, eyewitness reports, or other incriminating evidence. Perhaps less apparent is that confessions may also suppress exculpatory evidence, leading individuals who had provided alibis to doubt their own recollections and forensic experts to interpret physical evidence differently than they would otherwise. At this point, systematic research is needed to determine if the effects on eyewitnesses would extend to other types of evidence. In this regard, it is noteworthy that whereas physical evidence is immutable (once collected and preserved, it can always be retested), an eyewitness's identification decision cannot later be revisited without contamination. Once informed of a confession, an eyewitness is forever tainted.

So, for those still teaching critical thinking, I highly recommend you read this article and find some way to incorporate it into your lesson plan. A little knowledge can be a dangerous thing. It can taint how we evaluate evidence that we think we are evaluating in an unbiased and independent way.

new further reading

Are Eyewitnesses in the Zimmerman Trial Reliable? External input makes eyewitness testimony unreliable. Eyewitnesses are generally unaware that their memory has been altered by post-event information By Natalie Wolchover Scientific American


(thanks to Tim Boetcher)

 Skeptimedia archives rarrow.gif (1048 bytes)


Media Mythmakers
How Journalists, Activists, and Advertisers Mislead Us

Don't Get Taken!
The book "they" don't want you to know about!

The Skeptic's Shop
No shirts, no mugs, no tinfoil hats.



alt med (CAM)
critical thinking (CT)
cryptozoology (CZ)
Frauds, Hoaxes, Conspiracies (FH)
junk science (JS)
paranormal (P)
science (SC)
skepticism (SK)
supernatural/religion (SU)
UFOs & ETs (U)
other (O)

Recent Posts

Evaluating Evidence (CT)

PBS Infomercial for Daniel Amen's Clinics (JS)

Framing the Medicine Wars (JS, CAM)

Statistics and Medical Studies (SC)

Sharon Begley: Skeptics Think They're Intellectually Superior (CT)

Prescribing Placebos (CT) (SC)

Religious Fascism (SU)

The Skeptics' Circle (CT)

Honoring Carl Sagan (FH)

Ancient Wisdom (CAM) (CT)

The FBI on Trial (JS)

The Paralyzing Precautionary Principle (JS)



This page was designed by Cristian Popa.