人格测验的变异性及效应量分析

An analysis of variances of personality testing and effect sizes

目的探讨重复测量中不同人格分量表得分的特点及其变化规律。方法323名某部官兵在相隔42 d两次参加了中国士兵人格问卷（CSPQ）测验。结果相关结果表明在6个分量表中,除防御性分量表外,测验分数的变异性与相应的人格分数都有显著的相关（r=0.11-0.37）。效应量统计结果表明,求异性、敏感特质、分离特质分量表分数的变异性随分数的增加而增加。除防御性分量表外,高分组与中等组之间差别有实际意义（Cohen d〉0.20）。结论人格测验分数的变异性不是完全随机的,是具有一定稳定性或规律性的。

Objective To find out the characteristics of variance of personality tests in repeated testing. Methods 323 soldiers attended the CSPQ twice for an interruption of 42 days. Results The correlation analysis found out that variances of all but one subtests （Defense） are significantly correlated with respective standardized scores （0.11 - 0.37, P 〈 0.05 ）. Analysis of effect size showed that the variances of the subtests of Difference Seeking, Sensitive, Dissociate increased with the growth of respective standardized scores. And the differences between high-score group and middle-score group of all but one subtests （Defense） have real effects （ Cohen d 〉 0.20）. Conclusion The variances of personality scores are not completely random.