The effects of different paraphrasing styles on the quality of reports from young child witnesses

Abstract

Young children's descriptions of sexual abuse are often sparse thus creating the need for techniques that elicit lengthier accounts. ‘Paraphrasing’, or repeating information children have just disclosed, is a technique sometimes used by forensic interviewers to clarify or elicit information (e.g. if a child stated ‘He touched me’, an interviewer could respond ‘He touched you?’). However, the effects of paraphrasing have yet to be scientifically assessed. The impact of different paraphrasing styles on young children's reports was investigated. Overall, paraphrasing per se did not improve the length, richness, or accuracy of reports when compared to open-ended prompts such as ‘tell me more’, but some styles of paraphrasing were more beneficial than others. The results provide clear recommendations for investigative interviewers about how to use paraphrasing appropriately, and which practices can compromise the quality of children's reports.

Keywords:

• child eyewitness
• child witness
• children's memory
• interviewing
• investigative

Notes

1. Five interviews from each paraphrasing condition, in each study, were independently coded for interviewer aggressiveness (on a scale of 1–5) and challenging tone (on a scale of 1–5). In general, interviewers were considered to be very low in aggression (Experiment 1: Ms = 1.04, 1.044, SDs = 0.04, 0.05, for the yes/no-paraphrasing and expansion-paraphrasing condition, respectively; Experiment 2: Ms = 1.25, 1.30, SDs = 0.10, 0.20, for the expansion-paraphrasing and paraphrasing-only condition, respectively) and challenging tone (Experiment 1: Ms = 1.03, 1.04, SDs = 0.04, 0.05, for the yes/no-paraphrasing and expansion-paraphrasing condition, respectively; Experiment 2: Ms = 1.23, 1.18, SDs = 0.20, 0.12, for the expansion-paraphrasing and paraphrasing-only condition, respectively). Thus, there were no differences across condition, ts < 0.50, all NS.

2. The finding that children provided fewer inaccurate details after yes/no paraphrases than expansion paraphrases could also be explained by a ‘yes bias’. Some researchers have noted that children's tendency to reply with a yes bias may increase children's accuracy to yes/no questions if the correct response is ‘yes’ (Peterson, Dowden, & Tobin, 1999; Steffensen, 1978). Given that in the present study children's responses were for the most part accurately paraphrased, the correct response to yes/no paraphrases would indeed be ‘yes’. Thus, a yes bias may help explain why children's responses contained a greater number of inaccurate details after expansion paraphrases than yes/no paraphrases. Since children in the yes/no paraphrasing condition were likely to give a simple ‘yes’ response (which would most often be coded as accurate) while children in the expansion paraphrase condition gave additional details (which could be coded as accurate or inaccurate), children in the expansion paraphrase condition provided both more accurate and inaccurate details.