Eliciting best evidence from autistic interviewees

Study 1 tested the efficacy of a 'Witness-Aimed First Account' (WAFA) interview with the primary aim of supporting autistic witnesses to improve the quality of their accounts within a legally appropriate, non-leading framework. The study adopted a 2 (Group: Autistic vs Typically Developing) x 2 (Interview: WAFA vs. Control ABE interview) x 2 (Video: Scrambled vs. Unscrambled) mixed design, where Video was within participants. Participants were interviewed about their memory for two specially developed videos depicting criminal events. One video was 'scrambled' whereby clip segments were re-ordered, thus removing the narrative structure of the video; the other video was watched intact. Participants were interviewed using the same interview (WAFA vs. Control) for both videos. In WAFA interviews, rather than having a free flow verbalization of the entire event – which is difficult and inevitably results in underperformance by autistic adults – participants self-segmented their free narrative recollection right from the beginning. Once complete, they then revisited each of the self-directed free narrative topics in turn, in the order that they were recalled. 

Study 2 tested how individual questions could be adapted to elicit more detailed yet relevant accounts from autistic witnesses. Based on an a-priori sample size calculation (assuming α =.05), in order to achieve Cohen's (1992) recommended power of .80 to detect a medium-to-large effect size, 32 ASD and 31 typically developing (TD) participants were recruited. Participants were asked to recall specific instances from their past that were relevant to CJS interviews. Questions were split into three different blocks comparing the use of: (1) open questions; (2) cueing with general autobiographical information before asking for specific details; (3) visual-verbal prompting involving a visual pie-diagram with prompts to recall when, who, what and where. Half of the participants also received preparatory information which included the question topics. The study used a mixed factorial design: 2 (Autistic, TD) x 2 (Preparation, No Preparation) x 3 (Support: Open, Semantic Support, Verbal Labels), where Support was within-subjects.