16PF中文版的信度、结构效度问题以及测试方式对信度的影响

The Reliability and Construct Validity Problems of Chinese Version of 16PF and the Effect of Testing Methods on Reliability

目的:揭示16PF中文版的信度和结构效度问题以及测试实施方式对问卷信度的影响。方法:以上海的一所大学的最大时间跨度16年的11个年级的53750名新生的16PF测验历史数据(包括有效数据52958份)为样本对16PF问卷中文版进行内部一致性信度和结构效度分析。结果:①无论是纸笔测试总体还是网络测试总体,16种因素的信度系数接近或者大于0.60的只有5个(乐群性、兴奋性、敢为性、忧虑性和紧张性),接近或大于0.70只有2个(兴奋性和敢为性),不大于0.40的有5个(聪颖性、怀疑性、幻想性、世故性和实验性);②各因素的信度系数在纸笔测试的9年间波动很小且具有随机性,在包括网络测试在内的16年间未表现出逐渐变小的趋势;③各因素在两种测试条件下的信度系数很接近,但整体而言,网络测验条件下的信度略占优势;④一些因素的总分之间有中等或中等以上的相关;⑤来自5种因素的6道题目之间的相关均具有统计学意义(均P<0.01),并且来自同一种人格因素的两道题目之间的相关并无优势。结论:①由于问卷设计本身的原因,16PF中文版存在明显的信度和结构效度问题;②合理设计的网络测试有助于提高问卷的信度。

Objective:To reveal the reliability and construct validity problems of the Chinese version of 16 PF and the effect of testing methods on reliability.Methods:The internal consistency reliability and construct validity of the questionnaire were analyzed based on 53750 freshmen’s history data of 16 PF test(including 52958 valid data)of 11 grades with a maximum time span of 16 years from a university in Shanghai.Results:①Whether to the paper-pencil testing data totality or to the network testing data totality,only the Cronbach’sαof the 5 factors as Warmth(A),Liveliness(F),Social Boldness(H),Apprehension(O)and Tension(Q4)were close to or greater than 0.60,only 2 of them for Liveliness(F)and Social Boldness(H)were close to or greater than 0.70,and the Cronbach’sαof the 5 factors as Reasoning(B),Vigilance(L),Abstractedness(M),Privateness(N)and Openness to Change(Q1)were not more than 0.40.②The Cronbach’sαof each factor fluctuated very little and randomly during the 9 years of paper-pencil condition,and there was no tendency to decrease gradually for the Cronbach’sαof each factor in the whole 16-year time span which including the network condition.③The Cronbach’sαof each factor under the two conditions were very close,but overall,the reliability under the network condition was slightly superior.④There were moderate or even higher correlation between the total scores of some factors.⑤The correlations between the 6 items from 5 factors were statistically significant(P<0.01 for all of them),and there was no advantage for the correlation between the 2 items from the same factor.Conclusion:①Due to the questionnaire designing itself,there were obvious reliability and structural validity problems in the Chinese version of 16 PF.②Reasonably designed network testing can help to improve the reliability of the questionnaire.