结构化面试中评分量表对评分准确性的影响

采用2(组内变量:量尺大小(25分和9分))×2(组间变量:评分方法(相对和绝对))的混合实验设计探讨评分量表对115名大学生新手评委评分准确性的影响。对于评分准确性,采用Cronbach1955年提出的四个指标,Elevation(EL)、Differential elevation(DE)、Stereotype accuracy(SA)、Differential Accuracy(DA)。结果发现,评分方法只在SA上主效应显著,量尺大小在只在DA上主效应边缘显著,评分方法和量尺大小在DE、SA和DA三个指标上均有交互作用。总体上看,在结构化面试评分中,对于评分准确性,相对评分量表优于绝对评分量表,小量尺量表优于大量尺量表。

Effects of Rating Scale on Rating Accuracy in Structured Interview

The researchers conducted an empirical research on rating scale.A 2(within group:scale size(25 vs.9))× 2(between group:rating method(absolute vs.comparative))multivariate analysis was adopted to study the influence of scale size and rating method on novice interviewers' rating accuracy.119 undergraduates served as novice raters were randomly assigned to watched several interview video clips recorded from civil servants selection tests,at the same time,making ratings for each applicant individually.The classical Cronbach's(1955)four accuracy components were used as rating accuracy indexes.The results show:(a)raters who used comparative rating method had higher stereotype accuracy(SA)than those who used absolute method;(b)comparing to large-scale,small-scale rating had higher differential accuracy(DA);(c)when small-scale rating form was used,comparative rating showed better differential evaluation(DE)than absolute rating;(d)when large-scale rating form was used,comparative rating showed better differential accuracy(DA)than absolute rating.In conclusion,small-scale and comparative rating scales have higher rating accuracy.Hence,small-scale and comparative rating scales should be extensively employed in the future selection tests.