The Self‐Administered Witness Interview Tool (SAW‐IT): Enhancing witness recall of workplace incidents

Abstract

Given the often crucial role of witness evidence in Occupational Health and Safety investigation, statements should be obtained as soon as possible after an incident using best practice methods. The present research systematically tested the efficacy of a novel Self-Administered Witness Interview Tool (SAW-IT); an adapted version of the Self-Administered Interview (SAI ©) designed to elicit comprehensive information from witnesses to industrial events. The present study also examined whether completing the SAW-IT mitigated the effect of schematic processing on witness recall. Results indicate that the SAW-IT elicited significantly more correct details, as well as more precise information than a traditional incident report form. Neither the traditional report from, nor the SAW-IT mitigated against biasing effects of contextual information about a worker's safety history, confirming that witnesses should be shielded from extraneous post-event information prior to reporting. Importantly, these results demonstrate that the SAW-IT can enhance the quality of witness reports.

A meta‐analytic review of the Self‐Administered Interview©: Quantity and accuracy of details reported on initial and subsequent retrieval attempts

“…The following inclusion criteria were applied: (a) Participants were randomly allocated to complete the (original, unmodified) SAI© or to a comparison condition. Modified versions of the SAI© such as the SAW‐IT© (MacLean, Gabbert, & Hope, 2019) and ad‐hoc structured recall questionnaires (e.g., Roos af Hjelmsäter, Strömwall, & Granhag, 2012) were not eligible; (b) Participants observed a staged event (filmed or live) with known “ground truth,” allowing for scoring of correct and incorrect details; (c) The SAI© was completed in the same experimental session as the witnessed event (i.e., within 1 hr). This mimics the way the SAI© was designed to be used in practice, and in which it has been recommended for use by policing bodies (College of Policing, 2019); (d) The study reported at least one of three dependent measures (number of correct details, number of incorrect details, accuracy of reported details); and (v) the results were reported in English.…”

Cognitive bias in workplace investigation: Problems, perspectives and proposed solutions

“…We can see the effectual possibilities of irrelevant contextual information in workplace investigation with research that has explored biasing contextual information on witnesses' (not investigators) understanding of event cause. Irrelevant contextual information that indicated an unsafe worker that was shared with witnesses after they experienced a workplace event and before they reported it, biased the content of their testimonies to emphasize worker fault, as well as, their allocation cause to be more worker focused [82].…”

“…The majority of workplace investigation techniques and tools have not been empirically vetted with the exception of a few. This vetting has revealed the utility of some tools to support the investigation while not adding bias (e.g., Cognitive Interview [139]; Self-Administered Witness Interview Tool [82]) whereas limited evidence suggests others should be used with caution (e.g., SCAT [72]).…”

Examining the efficacy of a self‐administered report form in missing person investigations

“…Specifically, we sought to address whether retrieval support in the form of a mental reinstatement of context elicited more person descriptors in comparison to a simple free-recall instruction. Based on research showing that both the original and bespoke versions of the SAI recall tool elicit more information than a free-recall comparison (Gabbert et al, 2009;MacLean, Gabbert, & Hope, 2019), it was predicted that participants who complete a SAI-MISSING form will report more person descriptors than those who complete a self-administered control form (that contains the same questions but an absence of cognitive retrieval instructions). In addition, we explored which types of information were most likely to be reported or withheld when reporting a missing person, and why.…”