本論文根據無望感的相關理論，以三種方式探討貝氏無望感量表內的正負向期待為一個概念的兩面，或是表徵兩種不同的概念。首先，本論文進行探索性與驗證性因素分析，以考驗一與二因素模式的模式適合度。再者，MacLeod與Byrne（1996）認為兩類未來思考對憂鬱症狀的影響是相互獨立運作，但可能有過於簡化的限制，故本論文提出四種的可能模式並加以檢驗。最後，根據Trew（2011）所提出的整合性模式，提出行為抑制與激發系統導致憂鬱症狀之兩種競爭模式，檢驗無望感或正負向未來思考在此模式中所扮演的中介角色，以及兩個系統之機制間有互動的可能性。主要的研究結果如下：（1）探索性與驗證性因素結果均支持貝氏無望感量表的二因素結構，並以正負向未來思考加以命名；（2）支持模式二的假設，負向未來思考為正向未來思考與憂鬱症狀間的部分中介變項，但正向未來思考並非是負向未來思考與憂鬱症狀間的部分中介變項；（3）競爭模式二具備較佳的模式適合度，支持Trew（2011）認為憂鬱症時須同時注意BAS與BIS各自不同影響途徑的觀點，亦彰顯正向未來思考的保護因子角色；（4）更重要的是，支持無望感量表中正負向期待內容應被視為兩種不同且各自存在的概念。最後並提出本論文研究限制與對憂鬱症的臨床理論與實務上之建議。 / This study investigated the relation between the positive and negative expectations assessed in Beck Hopelessness Scale (BHS) by three ways. First, exploratory and confirmatory factor analysis were used to test the goodness of fit of one-factor and two-factor models. Besides, MacLeod and Byrne (1996) stated that two kinds of future thinkings influenced the depressive symptoms independently, but this statement had some limitations. Therefore, this study proposed four models to test the hypotheses. Based on Trew ‘s (2011) integrated model, two competing models illustrating the relations among BAS, BIS and depression were proposed to examine the mediation effect of hopelessness or future thinkings. The main results were: (1) the two-factor model of BHS was supported in exploratory and confirmatory factor analyses; (2) the second model was supported that negative future thinking was the partial mediator between depression and positive future thinking; (3) the competing model 2 had the better goodness of fit, supporting that BAS and BIS had important but different pathways to influence the development of depression, and positive future thinking played the protective role in this process; (4) Most importantly, the perspective that the positive and negative expectations assessed in BHS should be treated as two different kinds of constructs respectively was supported. Finally, the limitations of this study and the suggestions for the theories and clinical treatment of depression were discussed.