When my totally satisfactory son was a child, we would watch Sesame Street together. At that time, Sesame Street had the hots for "Some" "All" and "None". Every show had something about inclusive and exclusive adjectives. Unfortunately, some people did not watch SS enough, as evidenced by the current confusion about whether cops are any good at lie detection.
Look at these two syllogisms:
1. Most people are only at chance in detecting lies.
Most cops are people.
Therefore, most cops are only at chance in detecting lies.
Is this correct, or is number 2 better?
2. Truth wizards can detect lies most of the time.
Some truth wizards are cops.
Therefore, some cops can detect deception most of the time.
Both syllogisms are correct (at least correct enough for a post!). Most cops, like most people are at chance in their lie detection accuracy. But some cops are expert lie detectors. The conclusion about most cops comes from research on unselected groups of law enforcement personnel; the conclusion about some cops comes from research with highly selected individuals and groups of law enforcement personnel.
In the 1990's, Paul Ekman, Mark Frank and I described two highly accurate law enforcement groups (Secret Service agents and highly motivated federal investigators). We also reported only chance accuracy scores for several other groups of police professionals. More recently, in a study done in England, Mark Frank found differences in accuracy among different types of law enforcement personnel. Homicide detectives were very accurate while patrol officers were not. So some kinds of cops can catch liars sometimes.
In 2000, I started interviewing so-called "Truth Wizards," people who obtained scores of 80% or better on at least two of three standardized lie detection measures. These were not easy tests since most people achieved only chance scores (near 50%) on them. All three tests show "high stakes" lies in which the liars and truth tellers were highly motivated, either for personal reasons or by some significant positive and/or negative reinforcement (substantial reward, serious punishment).
Paul Ekman has argued for many years that high stakes testing materials are necessary to determine lie detection accuracy, otherwise the lies and truths will not have the relevant emotional and cognitive behaviors necessary to uncover the truth. Remember the last time a police officer interviewed you? Did your heart race? Was your mind numb? Was your mouth dry? And this can happen even if interview is about a broken tail light or a missing tag. Imagine if the interview were about a murder. Police are used to interviewing people in situations in which they are both emotionally and cognitively aroused.
My colleagues and I recently examined this hypothesis in a review examining all studies of lie detection accuracy in police groups. We found that police officers tested with high stakes lies were significantly more accurate than those tested with low stakes lies. No police officers tested with low stakes lies were above chance in accuracy.
So to determine whether police officers are accurate in lie detection be sure to specify whether it is some, all, or some