The Effect of Instructions on the Accuracy and Confidence of Eyewitnesses - Assignment Example

Name: College: Course: Lecturer: Date: EYE WITNESS (FACE RECOGNITION) EXPERIMENTS: THE EFFECT OF INSTRUCTIONS ON THE ACCURACY AND CONFIDENCE OF EYE WITNESSES Abstract This report examines the accuracy and confidence level of eyewitnesses in perpetrator identification. The experiment was tailored to determine the impact of instruction on ability of the eyewitness to accurately and confidently identify the perpetrator. Instruction was the single factor for manipulation, whereas the sets of photographs (faces) became the factor for manipulation within the subjects. It was found that low expectancy instruction reduces the confidence and accuracy, while high expectancy instruction raises the confidence and accuracy of the eyewitness. Therefore, instruction actually affects the confidence and accuracy of eyewitness in perpetrator identification, which in turn compromises the reliability of their testimony in a court of law. Introduction Every year, thousands of crimes are committed with some having witnesses at the scene. These witnesses may be called upon to testify in a court of law so as to prove the guilt or innocence of the perpetrators in question. Numerous factors determine the reliability of such the witnesses. Studies show that some of the factors include system variables, confounders, mental capacity and the environment. Other than these, external factors such as instruction by police or pre- line up instructions by the administrator; also do affect the reliability of eyewitness in perpetrator identification (Wells, 603- 613). Given all these factors, it can be seen that not all eyewitness testimony is correct, as about 250 people have been found not guilty (www, innocentproject.org) following DNA tests. Prior to appearing before a court of law, the witness must identify the suspect from a line- up mounted by the police, to ascertain that they have the right person. Before the lineup, the police or the line- up administrator instructs the witness on what to do. Some of the instructions given include telling the witness that the suspect is not in the lineup, or by not telling them about whether or not the suspect is in the lineup. The instructions given influence the accuracy of the witnesses, leading them to make wrong decisions (Wells 603- 613). The police can avoid this by, warning the witness that the probability of the suspect being the in lineup is 50%, maintaining an environment where a detective cannot influence judgement and getting a statement of the witness certainty of their identification (Wells, in Stambor 26). If detectives followed the above steps, this will help the witnesses make their decisions without undue influence from them. This increases the probability that if they selected someone, they did it because they were right, not due to the instructions given. There are many different variations to instructions, yet little is done to find out how they affect the witnesses’ decision, and thereby their accuracy and confidence levels during the line- up. We therefore set to find out how the instructions given affect the accuracy and confidence levels of the witnesses, in recognition of faces, hence the experiment. Methodology Participants To do this experiment, participants were psychology students taking Psychology II. The experiment involved both the students who were the supposed eyewitnesses and the instructor. The participants therein role-played as the eyewitnesses a versus the police or line-up administrator respectively. Materials The study was a single factor within subject design. The factor being manipulated was the instructions given. The participants looked at fifty photographs in two batches of 25 each to constitute the between-group degree of freedom for the first trial. In the second trial, the group divided into two, and then looked at another fifty that consisted of some of the previous photographs. This was to deduce their level of accuracy and confidence after being given the instructions that one group of fifty contained 70% of the previous photographs while the next group were informed that their group of fifty contained 30% of the previous photographs. Procedure The study was a single factor within-subjects design, where all of the participants completed both conditions. This method is referred to as the test recognition memory as elaborated by psychologists (Van Zandt 582- 600). In a face recognition experiment, participants view photographs and the order of manipulation is counterbalanced. The factor manipulated in the experiment was the instructions. The participants were expected to complete two blocks of trials where they were shown twenty-five photographs in each. Afterwards, they viewed 50 photographs out of which only 50% were from the previous (old batch) they had seen while the rest were new. High expectancy (70%) is expected to result in higher confidence and accuracy, and the reverse applies for the low expectancy (30%) as shown in Figure 1 below. Figure 1 We used the hypothesis that the frequency of positive decisions would be higher in the high expectation level than in the low one, and that after every positive identification, confidence levels would rise much higher in that group as compared to the confidence levels of the low expectations group. Results The results of this experiment revealed that the subjects enjoyed a degree of freedom (df=37) at 95% confidence level. Figure 2: Two-Tailed Normal Distribution curve to which the confidence level = 95% 0.96 p=0.05 0.96 The t-test, p-values varied from (p = -0.02278) to (p =0.02815) for low expectancy and high expectancy respectively. On the other hand, with a two-tailed ANOVA, the degree of freedom significantly varied whereas ANOVAt-between = 5.805 and ANOVAt-within = 0.214. This translated to the expected value of F-ratio ≠ 1. Consequently, the experiment resulted in an F-score of 0.832 falling within the expected distribution of possible F-scores: (F-score = 0.832) hence (Fcrit, < 0.832 >1). Illustrations of the experiment and results Table 1Results of the two-trial (unbiased) High and Low Expectancy face recognition experiment Figure 3 Mean frequency for positive recognitions for the two trials Table 2 Analyzed results of the single-factor face recognition experiment Discussion Figure 1 illustrates the two-trial single factor face recognition experiment. In this case the experiment was designed to be unbiased hence the two groups completed both Trial 1 and Trial 2 simultaneously. Since the experiment was manipulated by the instruction, it was of importance to subject both groups to similar condition so as to achieve homogeneity and control the error of biasness. In terms of sensitivity and accuracy, this experiment was drawn at a 95% confidence level at p = 0.05. Given that experiment was two tailed, the results would only be significant at values where p is less than 0.05, hence the shaded proportions, p1), the established effect herein is real. Meaning, instruction actually influenced the observed variation of confidence and accuracy among the low and high expectant eyewitnesses. The materiality of this experimental outcome is therefore in accord with other knowledge of psychology articulating the influence of instruction on eyewitness decision-making ability. In fact, one can anticipate that in a contemporary society whereas an eyewitness is not instructed on the possibility of this probable outcome, the witness’ positive recall would be more accurate. The jury deems such precision as more authentic since the identification process and police has not solicited judgement through instruction. (Steblay 283- 297) To confirm positive identifications, scientists carry out DNA tests. Surprisingly, most results where one instructs eyewitness have turned out to be negative. Meaning that the perpetrator identified by the ‘instructed eyewitness’ is actually not guilty. The reverse is true for ‘uninstructed eyewitnesses’ as the DNA test is consistent with every positive identification; certainly not always. (Vickers and Lee 169- 194) We concluded that indeed, instruction affects the confidence and accuracy of eyewitness. It would thence be modest and advisable not to instruct eyewitnesses since studies have established the ‘instruction’ idea to create biasness in positive judgement. Works cited Steblay, N. M. Social influence in eyewitness recall: A meta-analytic review of lineup instruction effects. Law and Human Behavior, 21(3), 1997: 283-297. Van Zandt, T.L. ROC curves and confidence judgments in recognition memory. Journal of Experimental Psychology: Learning, Memory and Cognition, 26(3), 200: 582-600. Vickers, D., & Lee, M. Dynamic models of simple judgments: I. Properties of a self-regulating accumulator module. Nonlinear Dynamics, Psychology, and Life Sciences, 2(3), 1998 :169-194 Wells, Gary. Eyewitness identification procedures: recommendations for the lineup and photo spreads. 22Law. and Human Behavior. 1998:603- 613. Innocent Project. Eye witness misidentification. Wells, Gary in Stambor, Zak. How reliable is eyewitness testimony? Psychologists are helping police and juries rethink the role of eyewitness identifications and testimony. The Monitor, Vol 39 No 4. April 2006: 26 Read More