Should Police Use ACID? Training and Credibility Assessment Using Transcripts Versus Recordings

Abstract

There is a need for an applicable system of investigative interviewing and credibility assessment that extends science to practice. Experiment 1: Officers assessed the credibility of colleagues in a face-to-face interaction. Fourteen of sixteen officers were wrong in determining whether their colleague was responding honestly to them or lying to them. Experiment 2 compared untrained officers to officers trained in Assessment Criteria Indicative of Deception (ACID) in their ability to assess the credibility of statements regarding real and mock crimes by analyzing either verbatim transcripts or audio recordings. Officers who were trained in ACID performed significantly better after training than before, and significantly better than a group of officers who were never trained in ACID (89% correct versus 53% correct). No differences emerged due to analyzing transcripts versus recordings. The practical aspects of these results are discussed, including success rate, ease of application, and ease of training. The theoretical discussion includes differential recall enhancement and impression management. Finally, limitations and suggestions for future research are presented.

KEYWORDS:

• investigative interviewing
• credibility assessment
• deception
• police
• training
• Assessment Criteria Indicative of Deception (ACID)
• differential recall enhancement

Notes

1 This use of a lie script as a basis for responding, rather than thinking about the memory for the target event, was initially described as superficial encoding (Porter & Yuille, 1996).

2 The RI is a permutation of the Cognitive Interview with a structure and content designed to detect deception (K. Colwell et al., 2002, 2013; Fisher, Geiselman, & Amador, 1989; Geiselman & Fisher, 1986).

3 TTR is a ratio of lexical diversity that is related to control of information. It is calculated as the number of unique words in an utterance divided by the total number of words in that utterance (K. Colwell et al., 2002).

4 This is a mixed design, with a within-participants examination of training and modality and a between-participants examination of training and modality.

5 Also, this training was twice as long as the training in the L. Colwell et. al. (2012) study, and this increased time is likely the reason for improved performance by this sample relative to the 2012 study.

6 The trainees summed this up with, “You mean, we have been told before to look for micro-expressions and fidgeting and all sorts of stuff. That didn’t work. Now, you are telling us that all we have to do is just learn to see who is adding more relevant information … that they are telling the truth?!”

7 This was a powerful motivational tool, by the way. The officers saw that their colleagues had no idea who was honest and who was deceptive. Consequently, there was no way to deny the need for additional training or to attempt to rely upon the superiority of previous knowledge, experience, or other sources of information.

8 It is possible that practice accounted for some of the improvement, because the officers received feedback about whether they were correct in their judgments between Experiment 1 and the beginning of Experiment 2. They did not receive feedback regarding their performance in Experiment 2 until all of the data were collected.

9 Police officers in Britain have performed better than chance at detecting deception using real-life events (Mann et al., 2004), while police in the United States have always been found at or near chance, and have demonstrated mistaken beliefs about deception (L. H. Colwell et al., 2006).

10 Similarly, honest respondents are less likely to attempt to appear calm or helpful (K. Colwell et al., 2006; Hines et al., 2010).