本研究主要在探討個人與主管適配之前因及後果變項。本研究以主管與部屬之關係做為前因變項,將其又分為共事前既定存在的華人關係以及共事後才引發的主管部屬交換關係,另外將組織承諾、主管承諾、離職意圖視為結果變項。其次,本研究另一研究目的為將個人與組織適配和個人與群體適配視為干擾變素,觀察兩者是否會影響原本個人與主管適配對於結果變項之關係。 本研究問卷發放對象同時為主管和員工,有效問卷的計算為對偶方式,總計發放問卷419份,最後回收有效樣本為360份,其中主管188份,員工360份。統計方法則採用敘述性統計、信效度及相關分析、迴歸分析等基本統計分析方法,此外,為了忠實呈現干擾情況,本研究採用Edwards(1991)所提出的Polynomial Regression分析方法,此種分析方法將可針對干擾效果描繪出更為精密的3D變化,以提供更為全面的分析結果。 研究結果顯示,主管部屬交換關係越好,個人與主管適配程度越高,並進而造成組織承諾較高、主管承諾較高及離職意圖較低的結果;然而華人關係卻無顯著預測效果。再者,個人與組織適配和個人與群體適配的確存在干擾效果,說明不同的適配是會相互作用的,其中較為特別的是針對主管承諾和離職意圖方面,個人與主管適配皆吸收了大部份的解釋能力,可以推論在華人社會裡,個人與主管之間的相處情況的確在企業中扮演很重要的角色。 整體而言,本研究彌補過去文獻關於個人與主管適配上的不足,並企圖以新的迴歸分析方法觀察不同適配之間的互動,來探討個人與主管適配在華人社會的重要性,因此對於學術上和實務上皆有所貢獻。
This study mainly examined the antecedences and outcomes of person- supervisor fit. Person- supervisor fit can be explained by leader-member relationship, including the relationship before and after co-working and can explain organization commitment, supervisor commitment, and intention to quit. Also, this study examined the moderating effects of person-organization fit and person-group fit to observe the interaction effect of different fits. This study has distributed 419 questionnaires, and there are 360 effective ones surveyed by 188 supervisors and 360 employees. Then, we adopt descriptive statistics, reliability, correlation, and regression analysis. In order to realize the moderating effects more precisely, we also adopt Edwards’(1991) Polynomial regression method to get 3D graphics. The result indicates that the better the relationship after co-working, the higher the person-supervisor fit which predicts higher organization commitment and supervisor commitment, and lower intention to quit. However, the relationship before co-working has no significant effects. Furthermore, person-organization fit and person-group fit really have moderating effects, and person-supervisor fit almost predicts whole supervisor commitment and intention to quit, so we can conclude supervisors indeed play important roles in Chinese context. Totally speaking, this study supplements the lack of related researches and tries using new statistics method to demonstrate the interaction effects of different fits, so this study can contribute on academic and practical fields.