Social Engineering Blogs

What’s the latest research news in detecting lies with behavioral anomalies?
You asked and Humintell has answered

Presupposing that clues to deception do exist (which science has proven), an important question is whether individuals in general can be trained to improve their ability to detect lies via training tools.

Drs. David Matsumoto and Hyisung Hwang along with colleagues Dr. Mark Frank and Lisa Skinner researched that question but took a closer look at Law Enforcement Officers (LEO’s) who’s job description includes deducing truths from lies.

They examined the question whether training in both verbal and nonverbal indicators of truth telling and lying would have positive effects on Law Enforcement Officer’s (LEO’s) ability to evaluate truth from lies.

Their results, “The Positive Effects of Detecting Lies from Training to Recognize Behavioral Anomalies“, were published in the Journal of Police and Criminal Psychology.

All trainees were from the Federal Bureau of Investigation’s (FBI) National Academy (NA) located in Quantico, VA. The FBI NA is a professional course of study for U.S. and international law enforcement leaders that serves to improve the administration of justice in police departments and agencies in the U.S. and elsewhere.

The researchers noted that even despite the fact that two major sources of indicators exist (Statement Analysis & Nonverbal Behavior), training programs typically focus only on one or the other. In reality both co-occur simultaneously, especially in face-to-face interactions but even when interviewees are alone writing statements.

Therefore, LEO’s should be trained in both techniques to ensure better accuracy at predicting truths and lies.

Procedures used:

Participants were tested in their ability to accurately detect lies via a pre and post training using truthful and deceptive videos of mock crimes and opinions. It is important to note that the participants were ONLY tested in scientifically validated methods of deducing truths from lies such as Statement Analysis and Nonverbal Behaviors (SA & NVB). Over a 10 week period for a total of 22 instructional hours, they were then provided with college level training on validated verbal and nonverbal indicators of truth telling and lying (SA & NVB).

Statement Analysis focuses on not only a person’s verbal recounting of an event (pauses, errors in speech etc) but also on their written recollection of an event as well (Undeutsch, 1989). Nonverbal Behavior (NVB) is when there are conflicting thoughts and feelings that occur when a person is lying. These often leak out despite attempts to control them (i.e. micro and subtle expressions).

An important note is that all trainees completed all tests individually and independently of each other, not in groups with any discussion.

The researchers used a variety of training tools: From lecture/discussions, web-based exercises, video review exercises and role playing. Exercises specific to nonverbal behavior were facial expression of emotion training (web-based) and verbal cues such as pitch, pauses, and speech errors (videos of interviews).

Results indicated:

Interesting results were that there was a marginally significant truth bias that existed at pre-test; training. Accuracy improved for videos of mock crimes but not for opinions. A substantial proportion of the variance in accuracy scores was accounted for by improvement from pre- to post. Therefore, this study purports that training in nonverbal and verbal indicators has positive effects on law enforcement officer’s ability to delineate truths from lies.

The training effect size for Crime videos, in fact, was substantially large, with a 25% increase in accuracy rates. These latter findings speak to the effectiveness of the training to the most relevant types of lies LEOs deal with. The findings did support the existence of a modest truth bias in the trainees at pre-test, replicating previous similar findings (Feeley et al.1995; Levine et al.1999; Masip et al. 2009b).

Limitations of the Study:

1. This study was not conducted without limitations, perhaps the largest of which was the lack of a control group. Given that the training occurred within classes offered at the FBI NA, it was logistically impossible to include a wait-list or placebo control or comparison group.

2. The small number of videos used in the pre- and post-tests and their brevity, and the fact that the trainees were mere observers of the interaction and did not have the freedom to question the interviewees themselves.

Other studies in the literature have suggested that just increasing the amount of active processing of verbal and nonverbal behaviors can lead to improvements in accuracy judgments regardless of having been trained to identify valid indicators of truth and lies (Bond et al. 2004; Levine, et al. 1999).

To test whether this may have accounted for increases in the accuracy rates observed in this study, they examined the amount of time trainees spent completing the pre- and post-tasks, considering the time spent as a proxy for active processing. Across the entire sample the time spent by the trainees did not differ between the pre- and post-tests. These data suggested that the improvements in accuracy rates did not occur because of increases in active processing of the stimuli.

Summation:

Identifying valid behavioral anomalies that indicate truth telling and lying–both verbal and non-verbal – can be an incredibly useful aid for any investigator.

Recognizing behavioral anomalies in Verbal and NVB can not only aid investigators in detecting lies more accurately; they can also be used as aids during interviews and interrogations to help the investigator gain insights about the personality, motivation, and internal conflicts of their interviewees, and to identify meaningful content areas of the interview that deserve further exploration and discovery.

Using behavioral anomalies to evaluate truthfulness and detect lies in investigative interviewing is not a silver bullet that will solve every case. As always interviews and interrogations need to be augmented by other sources of evidence such as witness statements and forensics.