运动成绩观测中广义信度的理论与方法

The Theory and Method of Generalizability in the Observation of Motor Performance

介绍了体育测量中的一种新的广义信度理论以及它在实践中的应用,并对古典测量理论与广义信度理论作了比较,重点研究了一组受试者在操作一系列测试中被一组观测者评价所得资料,并对算出的各种广义信度系数进行检验。文中使用的实例是由3名教师对10名学生在4次测试中的运动成绩作出评分。我们讨论了上述三种因子进行充分交叉的观测设计中能应用的各种广义信度系数的公式。最后还简要地讨论了把测试嵌套入受试者之内两种观测设计特例。

The purpose of this paper is to introduce a new generalizability theory of sports measurement and its application to practice. In our study, a comparison between the classical measurement theory and the generalizability theory was made. We put emphasis on the study of the data obtained from the evaluation of a group of observers when they observed a group of subjects taking part in a series of tests and then after calculation, we examined the various coefficients of generalizability. The examples used in our study were the motor performance marks given respectively by 3 teachers when they tested the motor performance of 10 students four times. We discussed various formulas of generalizability coefficients which could be used in crossed observational designs of the three kinds of factors. Finally, we discussed briefly two special cases of observational designs in which tests were set in subjects.