The Study Of Reliability And Validity Of The Aggression Questionnaire(Revised Edition)

Objective:To explore reliability and validity of the revision aggression questionnaire and to supply scientific foundation for expanding the applied range.Methods:LiYuHe,MenXianZhang in 2005 revised questionnaire based on the various types of criminal conduct determined by a survey on 1201 prisoners;Use of an assessment questionnaire conducted on 220 prisoners by police investigation.Confirmatory factor analysis to verify whether the amended version of the questionnaire fits this targeted surveyed group;Finally,empirical testing,in aim to demonstration study.Results:1.The revised questionnaires in this study received entries Identification Index range from 0.18 to 0.43,with a total score of the correlation between 0.335～0.699,with a magin error of 0.05 of that correlation.2.Reliability:Guttman Rtt Spilt-half reliability is 0.850,Cronbach'sαof homogeneity reliability is 0.875.After 2 weeks,to re-test and obtain the test-retest reliability is 0.881.3.Construct validity:Though the confirmatory factor analysis,the results suggest that the four-factor model is appropriate.4.Demonstration study:The police officer grades with the AQ total score correlation coefficient is 0.640(p＜0.01),connecting assumes medium. Evaluates high groups AQ to score points compared to lowly groups high. Masculine AQ scores points is higher than the female.The violence violates AQ scores points compared to the non- violence violates high.Different professional group AQ scores points has obvious difference.The years of schooling are not big to the aggressive influence. Conclusion:1.Revised edition AQ has a better reliability and validity,can achieve the surveying the request.2.Revised edition AQ may the quite different sex,the crime type,the professional crowd attack the behavior the difference.3.Revised edition AQ to be allowed the quite common crowd high attack behavior difference.