以學生評鑑教師教學量表決定教師的開課或去留可行嗎？混合IRT分析取向

Abstract

本研究主要探討以學生評鑑教師教學量表決定大學教師開課或去留的可行性。研究對象為東部某大學大學部學生，總樣本數為6,111 人。有別於過往學生評鑑教師教學的實證研究皆建構在古典測驗理論，本研究為了更嚴謹地回應學生評鑑教師教學的評鑑結果，因此使用近代測驗理論進行分析，並進一步考量學生的潛在異質差異對學生評鑑教師教學的影響。研究結果顯示，在未考慮評鑑教師的學生潛在異質差異時，教師可輕易通過學校所訂定在學生評鑑教師教學量表的效標門檻。但進一步考慮學生潛在異質性發現，不同潛在類別的學生評鑑教師教學的方式差異頗大。針對上述結果，本研究對大學教師及學生評鑑教師教學提出相關的建議。

This study examines the effects of variability in student ratings regarding instruction on decision-making for faculty teaching evaluation. A total of 6,111 undergraduate students from 173 classes in a university on the east coast of Taiwan were included in the research sample. This study is different from previous studies regarding student ratings for instruction that are constructed in classical test theory. We use item response theory to analyze the heterogeneity of students, to rigorously examine the effects on student ratings regarding instruction. The results show that teachers may easily exceed the teaching criterion score set by the university when not considering the heterogeneity of the student ratings. However, the different latent types of the variability of student ratings may be important for interpreting the results of different student rating scores. The recommendations for university teaching and student ratings regarding instruction are created based on the results from this study.