Please use this identifier to cite or link to this item: https://ah.lib.nccu.edu.tw/handle/140.119/98197
題名: 脈絡效果的階層線性模型分析:以學校組織創新氣氛與教師創意表現為例
其他題名: Hierarchical Linear Modeling of Contextual Effects: An Example of Organizational Climate of Creativity at Schools and Teacher`s Creative Performance
作者: 邱皓政 ; 溫福星
關鍵詞: 脈絡效果 ; 脈絡變數 ; 迴歸同質假設 ; 組織創新氣氛 ; 階層線性模式
Assumption of homogeneity of regression ; Contextual effects ;Contextual variables ; Hierarchical linear modeling ; Organizational climate of creativity
日期: Mar-2007
上傳時間: 22-Jun-2016
摘要: 社會科學研究所使用的量化資料多涉及階層性或叢集性的結構,階層資料的一個重要特性,是低階層次的解釋變數可以透過組內聚合程序,產生相同測量內容的「脈絡變數」(contextual variable)。透過脈絡變數與個體解釋變數相互間統計控制,得出解釋變數對於依變項的影響,稱為脈絡效果。本文的目的,主要在陳述脈絡變數的處理對於階層資料結構分析的重要性,並以688位來自22個學校的中學教師的組織創新氣氛調查資料為例,以傳統的分析取向(ANOVA、多元迴歸分析)與「階層線性模式」(Hierarchical Linear Modeling, HLM),進行脈絡效果的分析。研究數據指出,利用HLM的脈絡效果模型,證明了脈絡變數與個體解釋變數的相互影響,影響了各變數對於依變項(教師創意表現)的解釋,以致得出不同的研究結論,同時也發現脈絡變數會因為組間差異的存在,造成參數估計的變動。此外,本研究也發現組內迴歸不同質現象會影響變數的解釋,但可利用HLM分析來加以處理,是HLM優於其他分析技術之處。本文的結論指出,脈絡效果的檢驗,應被視為HLM分析的必要程序。其他有關HLM方法學與技術層面的考量,也影響數據分析與解釋的正確性,進一步的研究可再就脈絡效果在實際研究中的檢驗策略,以及HLM的原理與假設問題深入探究。
The data collected in social research frequently involve multilevel or cluster structure. One of the features of multilevel data is that the predictive variables in the individual level could be aggregated into the group level, served as a contextual variable. In turn, contextual effects, defined as the net effect of a group analytic variable after having controlled for the effect of the same variable on the individual level, may confound with other predictive variables in the model. The purpose of present paper is to clarify the significance of the contextual effects in the social research. A dataset based on a nation-wise survey on organizational climate of creativity, consisted of 688 high-school teachers at 22 high-schools, was taken for examining the contextual effects. Traditional analytic strategies such as ANOVA and multiple regression along with hierarchical linear modeling were applied. Results indicated that contextual variables on the group level have significantly confounding effects on the parameter estimates of predictive variables in the models, resulting in the different conclusions. Violation of the assumption of homogeneity of regression homogerenality may also affect the analysis of contextual effects. The present study proved that the HLM approach could be used to explore the contextual effects. However, further study on the basic technical issue of HLM for applying to the examination of contextual effects is expected. It is noted that the methodological issues of HLM have to be emphasized while the contextual effects been in interest.
關聯: 教育與心理研究, 30(1),1-35
Journal of Education & Psychology
資料類型: article
Appears in Collections:期刊論文

Files in This Item:
File Description SizeFormat
index.html107 BHTML2View/Open
index.html151 BHTML2View/Open
Show full item record

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.