摘要
以大学生为研究对象,实施无领导小组讨论测评,探讨培训组与非培训组评分者的信效度差异。通过相关分析、验证性因素分析等方法检验评分者的内部一致性、评分者间信度、组间与组内差异、结构效度及效标关联效度。结果显示,培训效应对评分者影响显著,经培训与未经培训的评分者组间差异显著,培训组评分结果的效标关联效度优于非培训组。对评分者进行系统的培训能有效促进客观公正地实施测评,提高测评结果准确性。
With university students as subjects, this paper carried out a Leaderless Group Discussion Test and probed into differences of raters' validity in Training group and non - training group in LGD Test. Through correlation analysis and confirmatory factor analysis, it tested internal consistency, inter - rater reliability, differences between groups, group differences, construct validity and criterion - related validity. The results illustrate that training of raters has significant effects on scores of LGD tests. There are obvious differences between groups among trained and untrained raters; the criterion - related validity of Training group rating is better than that of non - training group. The results show that raters' systematic training will ensure more objective and impartial implementation in LGD Test and improve the rating accuracy.
出处
《广州大学学报(社会科学版)》
CSSCI
2013年第2期60-65,共6页
Journal of Guangzhou University:Social Science Edition
基金
广东省哲学社会科学"十一五"计划2010年课题(GD10CGL08)