Scaling the perception of crime severity using Thurstone's method of paired comparisons : a thesis presented in partial fulfilment of the requirements for the degree of Master of Arts in Psychology at Massey University

Abstract

The present research assessed Thurstone's (1927a,b) method of paired comparisons as a technique for scaling crimes according to their seriousness as perceived by a New Zealand population. In the first of two experiments, 10 crimes, ranging from murder to possession and use of cannabis, were judged for their seriousness by 78 male and female subjects made up of University students and New Zealand Army personnel. Subjects were given a questionnaire in which each of 10 single-word crime descriptors was paired with every other crime. For each of the possible 45 crime pairs, subjects selected the most serious in the pair. Due to numerous boundary probabilities, Edwards (1957) Case V Incomplete Data Scaling Method was employed to construct crime seriousness scales. The method of paired comparisons produced similar results in the ranking of crime seriousness to a previous New Zealand study (Davis, 1992) that employed magnitude estimation scaling. A high level of relative consensus was found between different community groups based upon occupation and sex. This relative consensus extended to crime severity evaluations obtained from the sample employed and the New Zealand Judiciary and Legislature. A second study was carried out to examine whether the degree of relative consensus could be manipulated by varying the seriousness of the crimes. Subjects (24 males and 27 females) were given a forced-choice, computer-generated questionnaire that presented high and low serious versions of each of the 10 crimes employed in Experiment 1 in a one or two sentence description. Crime seriousness was manipulated by varying the quantity of economic or physical harm inflicted upon the victim. In general, the results showed that the degree of crime seriousness altered the position of any given crime on the scale in a very systematic way. Nonetheless, the relative consensus found was much the same as for the first experiment, suggesting that subjects do not envision a specific crime scenario when evaluating a crime's seriousness.