Narrative recall in an investigative interview: insight into witness metacognition

Abstract

Compared to other interview procedures, the Cognitive Interview produces a larger amount of information without compromising accuracy, and uses techniques that support memory retrieval and socio-communication. Metacognition plays a key role in regulating recall performance but it is unclear how metacognition regulates narrative recall in response to these techniques. Importantly, the grainsize of information elucidates the metacognitive mechanisms regulating recall, yet it is unknown how Cognitive Interview techniques affect narrative grainsize. This thesis examined how these techniques impact narrative performance (quantity, grainsize, and accuracy) and, by applying Koriat and Goldsmith’s (1996) framework of metacognition to narrative recall, elucidated the regulatory role of metacognition in the efficacy of the Cognitive Interview. Experiment 1 tested if the mental-reinstatement-of-context instruction improves monitoring performance, and if the naivety instruction (i.e., the interviewer states their naivety about the witnessed event) encourages the decision to produce more informative testimony. Both instructions produced a greater quantity of information but only the naivety instruction elicited finer-grained accounts. Results suggest that a statement of naivety promotes the decision to give a more informative report, and the mental-reinstatement-of-context instruction reduces the monitoring sensitivity to errors. Experiment 2 examined the mechanism that may lead a witness to respond to the naivety instruction. Specifically, it was assumed that the witness’ decision to report is influenced by their belief in the statement of naivety. When the interviewer made a naivety statement, participants rated their belief in the interviewer’s naivety higher and produced more informative reports. Results suggest belief is a necessary state for the efficacy of the naivety instruction. Additionally, Experiment 2 examined if the report-detail instruction also encourages a witness’ decision to produce more informative testimony and, importantly, if this moderates the efficacy of the naivety instruction. Participants produced more informative accounts, and interactions on quantity and grainsize precision, indicate that the report-detail instruction moderates the impact of the naivety statement. Experiment 3 applied Ackerman and Goldsmith’s (2008) dual-criterion model to narrative recall, to examine how the report-detail (informativeness incentive) and do-not-guess (accuracy incentive) instructions impact witness knowledge state. Linguistic qualifiers (e.g., “I think”) were also examined for how they communicate recall uncertainty. The study tested if: (a) the report-detail instruction manifests unsatisficing knowledge in more informative, less accurate reports communicated with greater uncertainty (i.e., more linguistic qualifiers); and (b) the do-not-guess instruction manifests conservative satisficing knowledge in less informative, more accurate reports communicated with less uncertainty. The report-detail instruction produced more information (in quantity and finer grainsize) without compromising accuracy or recall uncertainty, suggesting satisficing knowledge is used to give detailed accounts. The do-not-guess instruction produced more correct information, suggesting that the instruction enhances monitoring performance. Across all studies, accuracy was uncompromised when instructions produced more informative reports, suggesting the primary goal in narrative reporting is informativeness and not accuracy. This thesis makes theoretical contributions in applying metacognition theory to narrative recall, and elucidating how component Cognitive Interview techniques impact report informativeness (quantity and grainsize). Findings are useful to practitioners with understanding how different techniques produce informative and accurate testimony.