The linguistic correlates of conversational deception: Comparing natural language processing technologies

Abstract

The words people use and the way they use them can reveal a great deal about their mental states when they attempt to deceive. The challenge for researchers is how to reliably distinguish the linguistic features that characterize these hidden states. In this study, we use a natural language processing tool called Coh-Metrix to evaluate deceptive and truthful conversations that occur within a context of computer-mediated communication. Coh-Metrix is unique in that it tracks linguistic features based on cognitive and social factors that are hypothesized to influence deception. The results from CohMetrix are compared to linguistic features reported in previous independent research, which used a natural language processing tool called Linguistic Inquiry and Word Count. The comparison reveals converging and contrasting alignment for several linguistic features and establishes new insights on deceptive language and its use in conversation.

Differences in Word Usage by Truth Tellers and Liars in Written Statements and an Investigative Interview After a Mock Crime

“…Bond and Lee () demonstrated that truthful statements of incarcerated prisoners included more sensory and perceptual processes when they told the truth compared to when they lied. Other studies have reported similar results (Duran, Hall, McCarthy, & McNamara, ; Hancock, Curry, Goorha, & Woodworth, ; Schelleman‐Offermans & Merckelbach, ; Toma & Hancock, ; Vrij, Mann, Kristen, & Fisher, ). To be sure, findings have not always been consistent; two studies reported negative findings on words related to complexity and pronoun use (Hancock et al ., ; Toma & Hancock, ).…”

“…One source of indicators is verbal and comes from analyses of words used to provide statements and answer questions. Lies contain fewer words and omissions of information; are less plausible, structured and logical; are more internally discrepant and ambivalent; contain repeated details and lack contextual embedding; and include more descriptions of what did not occur (DePaulo et al ., ; Duran, Hall, McCarthy, & McNamara, ; Newman, Pennebaker, Berry, & Richards, ; Porter, Birt, Yuille, & Lehman, ; Porter & ten Brinke, ; Vrij, ).…”

Expect the Unexpected? Variations in Question Type Elicit Cues to Deception in Joint Interviewer Contexts

“…However, this does not preclude any possible differences between truth tellers' and liars' recollections. Script generation from imagined events is likely to be different, because mental representations of false events are likely to be less coherent than those of truthful events (Duran, Hall, McCarthy, & McNamara, ). We examined the frequency of temporal connectives in responses to a reverse order question, and a distinction was made between normal (‘and then’) and reverse order (‘before that’) connectives.…”

Discriminative Models for Predicting Deception Strategies

“…The use of linguistic cues to indicate potentially deceptive communications has also been studied in a variety of modalities and contexts (e.g., [24,13,7]). While text-based deception in computer mediated communication is a specific area of study [29,30,3], there has been less of a concentration on the various types of deception strategies that individuals utilize in their duplicitous communications, especially as people more often use informal, succinct messages as a result of the widespread adoption of social media.…”