对过去20年(1994~2013年)间国内有关大五人格测验的研究文献进行信度概化分析。结果表明:(1)检索到的文献中约68.15%存在"信度引入"现象;(2)未加权估计中,A和O的均值最低,N和C的均值最高,国内所得结果均略低于国外(O除外),而后者的变异性略大(E除外);采用α系数效果量方法,在随机效应模型中,N的估计值最高,O和A的估计值最低;(3)回归分析显示,分数均值、量表来源和南北地域差异是N维度信度的预测变量;量表来源、文章专业类型、测验版本和测验记分对E维度信度具有预测作用;样本量、文章专业类型和量表来源是O维度信度的预测变量;量表来源、文章专业类型、项目数和样本类型对A维度信度具有预测作用;量表来源、项目数、文章专业类型和测验记分是C维度信度的预测变量。
In this paper,we attempted to analysis the reliability generalization on studies of the Big-Five factor personality tests during the last 20 years( 1994 ~ 2013) in China. The results showed that:( 1) Among all studies,68. 15% caused reliability induction phenomenon.( 2) In the unweighted estimating,the lowest reliability coefficient was found for the Agreeableness and Openness,whereas the highest was found for the Neuroticism and Conscientiousness. What’s more,the reliabilities from Chinese researchers were less than Western counterparts,except the Openness. And the latter had higher variance,except the Extraversion. With the meta-analysis of coefficient alpha,the coefficient alpha of Neuroticism was the highest,whereas Agreeableness and Openness were the lowest,in the random effects model.( 3) The result of multivariate regression analysis indicated that sample characteristics such as the mean of scores,regional differences between the south and the north,inventory origin had predictive functions on the reliability for Neuroticism; inventory origin,paper types,testing version,and testing pointing were the main predictor variable for the reliability for Extraversion; sample size,paper types,and inventory origin had predictive functions on the reliability for Openness; paper origin,paper types,the number of the items,and sample types had predictive functions on the reliability for Agreeableness; inventory origin,the number of the items,paper types,and testing pointing were the main predictor variable for the reliability for Conscientiousness.