Familiarity-related fillers improve the validity of reaction time-based memory detection.

Abstract

The reaction time (RT)-based Concealed Information Test (CIT) allows for the detection of concealed knowledge (e.g., one's true identity) when the questions are presented randomly (multiple-probe protocol), but its performance is much weaker when questions are presented in blocks (e.g., first question about surname, then about birthday; single-probe protocol). The latter test protocol, however, is the preferred and sometimes even the only feasible interviewing method in real-life. In a first, preregistered, experiment (n = 363), we show that the validity of the single-probe protocol version can be substantially improved by including familiarity-related fillers: stimuli related to either familiarity (e.g., the word "FAMILIAR,") or unfamiliarity (e.g., the word "UNFAMILIAR"). We replicated these findings in a second, preregistered, experiment (n = 237), where we further found that the use of familiarityrelated fillers even improved the classic multiple-probe protocol. We recommend the use of familiarity-related filler trials for the RT-based CIT.

Deleterious effects of probe‐related versus irrelevant targets on the “CIT effect” in the P300‐ and RT‐based three‐stimulus protocol for detection of concealed information

“…Contrary to expectations based on Farwell () and our understanding of Lukács et al (), using crime‐familiar targets from the mock crime scene versus crime‐unfamiliar irrelevant target items (defined solely with unique response requirements as in the original 3SP) leads to smaller CIT effects for RT and P300 variables. Of course, Lukács et al did not use familiar targets but actually added familiarity‐related fillers to the probes, irrelevant, and targets in the usual P300‐based CIT.…”

“…As in the case of Lukács et al (), more targets and more familiar targets clearly improved the CIT effect. But, there were two important differences between those studies and the present one: In the studies of both Suchotzki et al () and Lukács et al (), the information to be detected in the CIT was autobiographical and likely based on semantic memory, whereas in the present P300 study, the information to be detected more likely involved episodic memory for incidentally acquired information during a mock crime. Indeed, besides the probe information, the familiar target information in the Suchotzki et al study was also autobiographical, whereas here the target stimuli were also incidentally acquired.…”

“…(Farwell, 2012). Although empirical support for advocating the use of these crime-familiar targets has not, to our knowledge, been published in the P300 CIT literature, a recent report by Lukács, Kleinberg, & Verschuere (2017), using a modified reaction time (RT)-based CIT, did suggest a possible rationale for us, although certainly Lukács et al were not suggesting using familiar targets.…”

“…Within the usual three stimulus types in a 3SP, Lukács et al () additionally interspersed familiarity‐related filler trials. These contained unfamiliar stimulus items such as unfamiliar, unknown, other, theirs, them , and foreign , in addition (and as opposed) to familiar items, such as familiar, recognized , and mine .…”

“…For the other control group (crime-unfamiliar or U), the target item was another unseen irrelevant item in the same category as the stolen probe item, as in the original 3SPs such as Farwell and Donchin (1991); Rosenfeld et al (1988). Based on Farwell (2012) and Lukács et al (2017), we would expect that, with meaningful targets possibly serving a role similar to that of fillers in Lukács et al, the F group would show a larger CIT effect (probe-irrelevant P300 difference) than the U group. Alternatively, it could be that meaningful targets divert participant attention from the probe stimuli, which would support the expectation that, contrary to Farwell (2012), the U group would show the superior CIT effect.…”