Pseudoscience in the media
Last week, I saw a slew of media headlines about the HaLeigh Cummings case and how certain methods used by investigators indicated that HaLeigh’s step-mother was lying or being deceptive. What I didn’t see was anyone holding up a hand and saying, “Wait a minute. Are these methods really scientifically accepted as deception detectors?” This might be a good time to point out that they’re not.
Desperation doesn’t confer reliability
A missing child is a serious matter, and it is not surprising that investigators want to know if HaLeigh’s step-mother, Misty Croslin-Cummings, is telling the truth about how HaLeigh disappeared. But at least three of the methods used to question her and to determine her truthfulness — a polygraph test, hypnosis, and a layered voice analysis — have either been disconfirmed in terms of their accuracy as deception detectors or have been seriously challenged in terms of their accuracy, reliability, and/or validity.
Some techniques have more basis in research and science than others but still don’t enable us to conclude deception or truth-telling with accuracy and reliability. For example, nonverbal communication (or “body language”) is an integral part of communication with others. Our interpretation of others’ body language often determines our impression of them to a greater degree than their words. But although there has been research on body language, none of the research, to my knowledge, demonstrates that body language can reliably indicate when a suspect is telling the truth or is being deceptive. Claims like “the body doesn’t lie!” are just, well, not scientific. Of course the body can lie. If it couldn’t, there wouldn’t be courses and consultants teaching people how to change their body language and people wouldn’t be able to learn how to beat a polygraph. Then, too, the interpretation and meaning of any particular nonverbal communication may vary as a function of the gender of the observer or cultural factors.
Body language may be communicating something, but as far as I know, there is no body language that has been scientifically validated as a “deception detector” that we could apply to any one individual with confidence. Criminal accusations are not a parlor game, and any claims that we “know” that HaLeigh’s step-mother or anyone else is definitely being deceptive because they look in a particular direction or sit or move in a particular way would not be claims supported by science. If anyone thinks that they have the research otherwise, please post it.
Methods such as polygraph testing and voice stress analysis or layered voice analysis also do not cut the scientific mustard when it comes to deception detection. Available critiques of research suggest that if these methods are detecting anything, it is more likely to be the individual’s arousal level or stress level, and there many be many innocent explanations for elevated readings. That said, it’s important to note that some of these methods have been described as being no more accurate than flipping a coin (cf, this study on layered voice analysis or some controversial research over voice risk analysis). If some methods do have any value, it may be because those being interrogated are fearful that they will be proven to be lying and therefore confess. But if that’s the case, then why not just set up an empty black box with tons of wires and lie to the suspect about the box and hope they confess?
Why does the media report so much pseudoscience?
So why do we see so much pseudoscience in media reports or “experts” giving their results and interpretations based on methods that do not meet rigorous scientific standards? Perhaps part of the reason is that proponents of specific methods authoritatively present their methods as if they were valid and scientifically accepted methods when they are not — and no one challenges them about the basis for using that method. I heard one reporter trying to ask questions about one particular method. It seemed to me that he was trying to get to the issue, but every question he posed was met with a firm answer from the investigator assuring that the method was a valuable tool. The one question that the reporter didn’t ask, however, was, “Is this method generally accepted as reliable and valid in the scientific community, and do the courts accept this method?”
Most people know that polygraphs are inadmissible in court. Criminal defense attorney Scott Greenfield informs me that body language, voice stress analysis, or layered voice analysis are also inadmissible in criminal proceedings as deception-detecting methods and that there is no method to detect deception that is currently admissible in criminal proceedings.
So if a layered voice analysis or any other such technique would not be admitted in a criminal proceeding against HaLeigh’s step-mother, why is she being tried in the court of public opinion based on that technique? Should the media report “news” or “findings” based on what may be or is often pseudoscience or junk science?
In the meantime, my email query to Scott seems to have inspired him to write his own blog entry on handwriting analysis and junk science. You can read his commentary here. After having been involved in one case where the prosecution’s main witness was a so-called “handwriting expert” who was anything but, you can bet I will be leery of any claims that handwriting can be used to conclusively determine deception.
Update: I just found this article by Matt Clarke on how law enforcement expanded its use of some techniques after 9/11 and the problems with that. Clarke writes, in part:
The big lie about lie detectors is that they can detect lies. At most they measure stress. Studies have generally shown that polygraphs can successfully measure stress levels; however, they do not differentiate stress due to the testing itself from deceptive or untruthful responses. Dr. Ken Alder, author of The Lie Detectors: The History of an American Obsession (Free Press 2007), calls the practice of measuring physiological responses to detect lies “a farce.”
He also reviews some of the research on voice stress analysis.
THE POLYGRAPH AND LIE DETECTION by THE NATIONAL RESEARCH COUNCIL OF THE NATIONAL ACADEMIES (2003) [http://www.nap.edu/openbook.php?isbn=0309084369] offers a very detailed examination of the subject—both the polygraph and other methodologies—including a review of the scientific literature. The Committee found that these methods, for the polygraph after a century of research, generally lacked any scientific validity to support their claims of detecting a subject’s deception. Indeed, the minority of research which does tend to establish scientific validity is typically produced by those with a vested interest in making such a claim. (E.g., The Department of Defense Polygraph Institute). There certainly is no empirical evidence, except anecdotally, to support the 90% accuracy rate asserted by some polygraphers. The Committee observed that the primary utility of the polygraph rests with its “mystique…that is, a culturally shared belief that the polygraph device is nearly infallible.” Ibid. at 18. Comparing the polygraph to the “[r]itualized lie detection techniques” of other societies and cultures, e.g., the medieval practice of immersion in water. As such, in the context of a criminal investigation the polygraph’s primary purpose is not the search for truth but to compel a confession. In such a situation, where subjective belief trumps objective efficacy, the witchdoctor is in and it’s best not to look behind the curtain.
As for the subject of the media’s scientific illiteracy and naivety: A deeply disturbing trend—not only because of the lives ruined or potential jurors tainted, as in this particular example—but because one hopes (vainly in most cases) that the professional media would show a greater level of skepticism and also because such reporting only tends to propagate the belief in such pseudoscience in the general public. However, such reporting should not be a surprise. The optimist in me would say this is the result of a lazy and ignorant media. The cynic in me would answer that such rubbish is sexy and sells and these are the only incentives of the News-Entertainment Complex.