學生評鑑教師教學題目安排順序不同對學生評鑑教師的影響：MI與MMI分析取向

The Effects of Question Order Difference on Student Ratings of Instruction: MI and MMI Analysis Approaches

本研究主要探討學生評鑑教師教學量表的題目在安排順序不同時，對學生評鑑教師教學的影響。研究對象為東部某大學的大學部182班學生，各班級人數介於11和92人之間，總樣本數為6，568 人。每班學生隨機分為A 組和B 組，A 組學生填寫學生評鑑教師教學題目安排在總量表之前的問卷，簡稱A組；B組則剛好相反。經測量恆等性檢定發現，題目安排順序不同對學生評鑑教師教學的分數在學生層次保持恆等，在班級層次因為聚合後產生的共線性而出現無法解釋的現象，因此後續分成兩部分加以深入探究。首先，在控制母體組間變異數的情況下，發現因素構念及學生填答分數仍具有恆等性；其次，在跨層級的分析架構下，其結果依舊保持恆等。本研究依照研究發現，對學生評鑑教師教學提出相關建議。

The purpose of this study is to investigate the effects of question order difference on student ratings of instruction. The sample consisted of 182 undergraduate classes from a university in the east coast of Taiwan， class size ranging between 11 and 92， and the total sample were 6，568 students. There were two forms (Form A and Form B) of the Student Ratings of Instruction (SRI) Instrument. Form A， the student ratings of instruction items were placed before the student self-evaluation items. On Form B， the question order was opposite to the order of Form A. The measurement invariance test revealed that subjects in a different order on theSRI at the student level scores remain identical， but at the class level the result was influenced by a high degree of polymerization of linear without clear explanations. Therefore， the follow-up analysis was discussed in two parts to explore them in depth. First， in the case of the model in the control group variance， SRI construct factors and scores were found to be identically equal. Second， in cross-level analysis framework， the results also remain identical. Based on the results of this study， some suggestions were provided for the practice and the future study as well.