| Summary: | The aim of the present study was to experimentally test the Undeutsch Hypothesis, which holds that children's statements based on self-experienced events are qualitatively and quantitatively different from statements based on coaching. Specifically, this study tested the validity of Criteria-Based Content Analysis (CBCA, a system for assessing the credibility of eyewitness reports) for discriminating between credible and non credible eyewitness reports by children. As well, two other tests of the quantitative and qualitative differences between credible and noncredible eyewitness reports were included. One hundred and forty-two children (74 Grade4, 68 Grade 2) were tested in three conditions: (1) Live Event, in which children were actively involved in a staged event (the event was complex and included many features considered relevant to credibility), (2) Heavily Coached, in which children did not experience the event but were told in detail about it (including details which, if reported, would be assigned significance by CBCA),and (3) Lightly Coached, in which children did not experience the event but were provided with a brief account of it, with the expectation that they would fill in details to make their reports believable. Children were asked to recall the event in individual interviews. Transcribed interviews were evaluated using CBCA. Results of the study provided mixed support for the Undeutsch Hypothesis. For Grade 4 children, CBCA significantly discriminated between the Live Event and Lightly Coached conditions, but not between the Live Event and Heavily Coached conditions. Thus, although CBCA accurately distinguished credible from lightly coached reports by this older group of children, reports of the heavily coached children fooled CBCA evaluation. For Grade 2 children, CBCA did not discriminate between the three conditions. This result raised questions about the applicability of CBCA to the reports of younger children. Results of the other two tests of quantitative and qualitative characteristics indicated that these systems did not aid in discriminating between credible and noncredible reports. The implications of these findings for the empirical validation of CBCA and for the use of this system in making credibility decisions in the forensic context are discussed. At this point in time, the assessment of CBCA is still taking place. Until further testing is completed, CBCA should be viewed as one approach to credibility assessment that has clinical support but limited empirical validation.
|
|---|
