大学生学习适应量表的信效度检验及改进:基于概化理论的视角

Psychometric Properties and Improvement of the College Learning Adjustment Scale:on the Perspective of Generalizability Theory

为了探讨大学生学习适应量表(CLAS)的测量学性能,对700名大学生进行问卷调查,并采用多元概化理论对所得数据进行分析.结果显示:CLAS各因子间的相关系数大都高于0.6;总量表的概化系数和可靠性指数均高于0.86;学习动机、学习能力和环境因素3个因子在理论贡献和实际贡献上相差3%.因此,CLAS的整体信效度较好,能有效地区分出大学生的学习适应水平,可用于常模和标准参照测验,但各因子不宜;因子结构和分值比例不尽完善;学习动机因子条目数量过多,可删1个;环境因素因子应增加1个条目;而学习能力因子则需改善条目质量.

To explore the psychometric properties of the college learning adjustment scale（CLAS）, 700 un- dergraduates were surveyed with CLAS. We analyzed the data by using multivariate generalizability theory. It turned out that most of the correlation coefficients of the factors were higher than 0.6, the generaliz- ability coefficient and reliability index of the whole scale were higher than 0.86, and there was a difference of 3% among the three factors including learning motivation, learning ability and environment factors in both theoretical and practical contribution. Therefore, the overall reliability and validity of CLAS were bet- ter. It was efficient to distinguish the learning adaptability level of college students and could he used for the norm-referenced test and the criterion-referenced test, but it was not applicable for a single factor. However, the factor structure and score proportion of each factor in the scale were not perfect. To improve CLAS, we suggest to delete one item from the learning motivation, add an item to the environment factor and improve the quality of learning ability items.