多元概化理论在教育教学能力测试中的应用

The Application of the Multivariate Generalizability Theory to Teaching Capability Assessment

使用多元概化理论对教育教学能力测试实测数据进行了分析。结果表明:1.教育教学能力测试将说课、答辩和面试成绩进行合成是比较合理的;2.三个测评任务中说课和答辩的评分质量较好,面试评分的质量较差;3.总体测试结果较适合于相对决策,做绝对决策时要谨慎处理;4.影响教育教学能力测试质量的主要原因是评分者宽严程度不一;5.增加评分员可以提高测试的精度,但增幅递减,当评分员人数为5时能够较好满足测试要求。

This study analyzes the data of teaching capability assessment based on the Multivariate Generalizability Theory. Both D study and G study were carried out in the analysis. The results showed that ： 1 ） It is reasonable to combine the there part composite universe scores into a composite universe score in teaching capability assessment; 2 ） The quality of both the summarizing lessons and replication is high, while the interview is not good enough ; 3 ） The result of Teaching Capability Assessment is only suitable for the relative decision not for the absolute decision; 4） The main factor that affects the results of teaching capability assessment are whether the raters are strict or not; 5 ） By increasing the number of raters, the accuracy of the teaching capability assessment can be improved, When the number of raters is five, the results will meet the testing requirements well.