Detecting concealed language knowledge via response times

Abstract

In the present study, we introduce a response-time-based test that can be used to detect concealed language knowledge, for various potential applications (e.g., espionage, border control, counter-terrorism). In this test, the examinees are asked to respond to repeatedly presented items, including a real word in the language tested (suspected to be known by the examinee) and several pseudowords. A person who understands the tested language recognizes the real word and tends to have slower responses to it as compared to the pseudowords, and, thereby, can be distinguished from those who do not understand the language. This was demonstrated in a series of experiments including diverse participants tested for their native language (German, Hungarian, Polish, Russian; n = 312), for second language (English, German; n = 66), and several control groups (n = 192).

Prolonged response time concealed information test decreases probe-control differences but increases classification accuracy.

“…A recently conducted series of experiments introduced an RT-CIT variation that serves to detect concealed language knowledge (Lukács et al, in press): For example, a suspect, speaking in English, may claim to not understand anything in Polish. In that case, Polish words could serve as probes in the RT-CIT, while graphemically similar pseudowords would serve as controls.…”

Motor Reaction Times as an Embedded Measure of Performance Validity: a Study with a Sample of Austrian Early Retirement Claimants

“…The common denominator of most studies is the assumption that reaction and response times are under lower conscious control and, therefore, may be harder to manipulate, and such manipulations might be more difficult to coach (Lupu et al, 2018;Vagnini et al, 2008). It also relates to studies in other fields of deception detection where manipulations in response behavior led to elevated reaction times (Lukács et al, 2021;Suchotzki et al, 2017).…”

Response Time Concealed Information Test Using Fillers in Cybercrime and Concealed Identity Scenarios

“…As compared to the Verbal fillers, it does only require selecting three relevance-or meaningfulness-referring target fillers. However, in our experience (with implementing various fillers in various languages; e.g., Lukács, Kawai, et al, 2021), this latter task is not too difficult, partly because only three fillers are needed, and partly because, in contrast to irrelevance-referring words, there are typically several obviously suitable choices for relevance-referring ones. Hence, the Mixed version does not substantially complicate practical implementation as compared to the purely Nonverbal version.…”

“…We collected data from 593 valid individual tests (all on "guilty" participants, i.e., ones recognizing the relevant probe) using RT-CIT with fillers, from seven different previous experiments (Lukács, 2021a; Exp. 1, 2, and 4 from Lukács, Kawai, et al, 2021; Exp. 1 from Lukács, Kleinberg, et al, 2017;Exp. 2 from Lukács & Ansorge, 2019; and one yet unpublished experiment).…”