Different Measures of Violent Attitudes May Not Be So Different After All

Previous research has produced mixed evidence regarding the distinctiveness of measures of violent attitudes. In two studies, we examined this issue further with different approaches. In the first study, 97 male undergraduate students were randomly assigned to a message to reduce violent attitudes or to a control condition and then responded to three self-report measures of violent attitudes (Evaluation of Violence Questionnaire, Violence Scale of the Revised Measures of Criminal Attitudes and Associates, and the Criminal Attitudes to Violence Scale). Though the effects of the experimental manipulation (Cohen’s d = -0.33, −0.03, −0.21, respectively) ranged from negligible to small, their 95% confidence intervals were largely overlapping. Further, the three measures were highly intercorrelated (r > .70). In the second study, 164 male undergraduate students were presented with a uniform set of violent behaviors paired with various response scales drawn from different violent attitude measures (e.g. bad vs. good, not understandable vs. understandable, unacceptable vs. acceptable). The different response scales were highly intercorrelated, with most correlations greater than .70. Taken together, the results of the current studies generally did not support the hypothesis that different measures of violent attitudes are practically or meaningfully distinct from one another.