An investigation of the role of cognitive style as a mediator of eyewitness memorial performance
Six separate experiments were conducted to investigate the role of Field Dependency (FDI) in determining the susceptibility of eyewitnesses to context reinstatement (CR); their performance in free, cued, and multi-choice recall, and facial identification accuracy; and finally the confidence expressed before and after performing these tasks, and the confidence accuracy relationship (C/AR). Questions were also addressed in relation to the measurement scale status of confidence ratings and the validity and utility of the Calibration / Resolution techniques for analysing the C/AR. Experiments I & II focussed on FDI and CR susceptibility, together with accuracy in free recall, cued recall, and recognition. Experiments III & IV focused on FDI, CR, and recognition. Experiment IV also utilised the Calibration / Resolution technique. Experiment V used a computer presented format to focus on FDI and recognition through simultaneous and sequential line-ups presented upright and inverted. Experiment VI included multi-choice recall testing along with free and cued recall and focussed on FDI, CR, and both correct and incorrect information produced. Experiments I, II, & VI indicated that Field Dependent (FD) participants benefited significantly from CR whilst Field Independent (FI) participants did not. FIs consistently outperformed FDs in cued recall. Experiment VI indicated that for FDs CR in free recall increased correct information and decreased erroneous information, leading to a significant improvement in the 'quality' of the information produced. In cued recall, however, correct information produced by FIs was greater than that produced by FDs and erroneous information lower, leading to a significant difference between FDs and FIs in the 'quality' of cued recall produced. Experiments III and IV indicated a significant superiority in facial recognition for FDs as compared to FIs across time delays of one week and three months in filled line-ups but not in blank. The explicit encoding format and upright and inverted lineups used in experiment V indicated a role for both attentiveness to others and configural processing in the superiority of FDs seen in experiments HI & IV. Results across experiments I, II, and VI indicated that confidence was significantly higher following a free recall test than when assessed at other points during recall testing. In experiment III, IV, & V confidence ratings were higher following exposure to the line-up than confidence ratings given before exposure. Results across all six experiments in relation to the C/AR show a marked level of inconsistency, however the use of the calibration / resolution techniques in experiment IV indicated a level of over-confidence on the part of FIs and underconfidence on the part of FDs that is in keeping with the personality aspects of FDI. In experiment VI the same techniques in relation to cued recall indicate a degree of overconfidence for 'easy' questions and under-confidence for 'hard' questions and point to a new approach to the investigation of the role of 'item difficulty' in determining the C/AR. Overall, the thesis argues that cognitive style is an important factor in predicting when context reinstatement will be beneficial, who will perform better in different recollection and recognition tasks, and how confidence will relate to accuracy.