| Abstract: | The classical studies by Thurstone (1927b) and Coombs (1967) on the seriousness of crimes and offences are replicated here for a sample of male and female students. Eighty subjects used both forced choice and graded dominance ratios to assess all pairs of 10 crimes/offences. The direct ratio estimates correlate very highly with pair-comparison probabilities, thus making it possible to interpret the former as direct estimates of dominance probabilities. These data can be scaled with excellent fit in Thurstone's sense. The resulting scales are quite similar to those obtained by Thurstone and Coombs. An analysis of the individual data shows, however, that 69 per cent of the average data are aggregated over a variety of distinct subgroups of individuals who partition the crimes/offences into subsets that are incomparable in their seriousness. Thurstone's and Coombs' seriousness scales must therefore be considered a population lawfulness. |
