The effect of anxiety on eyewitness testimony
Anxiety at the time of interview in witnesses to a crime has received scant attention from researchersin the eyewitnesst estimony field. In this thesis, the effects of state and trait anxiety on memory accuracy, suggestibility and confidence of accuracy were investigated. In addition, with respect to suggestibility, the effects of anxiety at the time of encoding misleading information was evaluated and compared to the effects of anxiety at the time of its possibler etrieval as a suggestibler esponse. Five laboratory-based studies were conducted. The first three used a standard suggestibility paradigm and anxious mood was induced by an experimental manipulation. The fourth study was a source identification experiment and the fifth was a line-up recognition study. The fourth and fifth studies both used questionnaire measures of state anxiety. Questionnaire measures of trait anxiety were used in all experiments. Neither state nor trait anxiety was found to affect memory accuracy, measured by cued-recall in Studies I to 3, nor did anxiety affect correct attributions to source in Study 4. In Study 5 there was a modest improvement in identification accuracy in the high state-anxious group, but only when the target was present in the line-up. Misleading post-event information (MPI) did not lead to suggestibility in Study 5, but it did affect accuracy. There was a strong effect of NTI in all other studies with misinformed groups more suggestible than controls. However, anxiety moderated the effect of MPI, with high state anxiety (whether experimentally induced or measured by questionnaire) being associated with reductions in suggestibility. In contrast, trait anxiety was associated with higher levels of suggestibility, but the effects of state anxiety were stronger. In Studies I-3 it was found that anxiety at either the encoding or the retrieval of WI generally resulted in lower levels of suggestibility. Anxiety affected confidence of memory accuracy, with more accurate confidence judgements generally observed in low anxious participants. In contrast, anxiety did not affect confidence in responsesto questionst esting suggestibility. Overall, the findings indicate that elevated state anxiety at interview does not adversely affect eyewitness performance and can reduce the negative effects of WI. Results are discussed in the light of both theories of cognition and emotion, for example cognitive biases associated with anxiety, and eyewitness testimony research,. in particular theories of suggestibility.