The importance of detailed context reinstatement for the production of identifiable composite faces from memory

Abstract

Memory is facilitated by reflecting upon, or revisiting, the environment in which information was encoded. We investigated these "context reinstatement" (CR) techniques to improve the effectiveness of facial compositesvisual likenesses of a perpetrator's face constructed by eyewitnesses. Participant-constructors viewed a face and, after a one-day-delay, revisited (Physical CR) or recalled the environmental context (Mental/Detailed CR) before recalling the face and constructing an EvoFIT or a PRO-fit composite. Detailed CR increased correct naming of ensuing composites, but only when participant-constructors suitably encoded the environment. Detailed CR was also effective when combined with another interviewing technique (Holistic-Cognitive Interview), with focus on a target's character; it was no more effective prompting constructors to engage in greater environmental recall. Analyses indicate that the Detailed CR advantage was mediated by an increase in face recall. Results are applicable by forensic practitioners to aid eyewitness memory, thereby potentially increasing suspect identification and subsequent arrest rates.

Fixing the stimulus-as-a-fixed-effect fallacy in forensically valid face-composite research.

“…Comparisons between the six difference conditions were evaluated using a by-participant-only analysis. Fodarella et al's (2021) analysis committed the stimulus-as-thefixed-effect fallacy and so we are unable to generalize these results beyond the composites employed. It is possible to demonstrate the size of the problem by carrying out a simulation on data of the same format but with no difference between the conditions (i.e., a null effect).…”

“…Across the field, as Table 1 shows, there are times when byparticipant only analyses are employed (e.g., Fodarella et al, 2021) and times when by-items only analyses are employed (e.g., C. D. Frowd, Nelson, et al, 2012). There were five papers out of the 33 that employed both by-item and by-participant analyses and interpreted the combination of these (e.g., C. D. .…”

“…Fodarella et al (2021) investigated the effects of context reinstatement and composite construction method on the naming of face composites. This article was chosen for illustrative purposes; however, it is the case that this article originally had additional analyses that were removed at the request of a reviewer (Hancock, personal communication) and so the fallacy may only be a feature of the published version and not the original article.…”

“…Across the field, as Table 1 shows, there are times when by-participant only analyses are employed (e.g., Fodarella et al, 2021) and times when by-items only analyses are employed (e.g., C. D. Frowd, Nelson, et al, 2012).…”

“…From these randomly generated 16 participants' data, a between-participant independent samples t test was conducted, and it was recorded whether the p value was less than .05. This by-participants analysis models the type of analysis reported in Fodarella et al (2021). A second by-items analysis was carried out where the means for the performance of each of the 20 items (over the 16 participants) were analyzed using an independent samples t test assessing whether the p values was less than .05.…”