本研究分爲兩大部份，研究一是以問卷方式對「生活事件知覺」、「歸因」、「適應方式」、「自尊」及「身心適應」間，作一整體性的探討。以301人大學生爲研究樣本，結果發現：1.以逐步延歸分析的結果，解釋「身心適應」的最主要變項為「自尊」與「歸因」；解釋「自尊」的最主要變項爲「適應方式」；解釋「歸因」的最主要變項為「事件影響程度」與「事件重要性」；對於「適應方式」，各變項之解釋力均低。2.在控制度方面，身心適應較差者，評量自己實際的控制度較低，且理想與實際控制程度間的差距較大，評量他人與自己實際控制程度間的差距亦較大。3.在歸因的原始分數上，這應較差者，在各項原因的評量上皆較高；以歸因的調整後分數來看，則無論正、負事件，適應較差者，外因歸向傾向較高，而內在歸因傾向較低。承續研究一的結果，對其中的重要變項再加以釐清；研究二從時間序列上，來探討事件成敗與適應好壞對歸因、期望改變的影響歷程。結果發現：1.在「生活事件歸因問卷」的原始分數上，遥應較差組比適應較好組的外在歸因為高；以調整分數來看，則適應越差，其「外在——不穩定歸因」傾向越强。2.實驗情境中的歸因，皆傾向於作「內在——不穩定歸因」；在作業二上，成功情境的內在歸因較失敗情境的為高；適應較好與較差兩組間則皆無顯著差異。3.期望改變主要受成功或失敗的立即影響，而不受症狀嚴重指標（GSI） 分數的高低與整個成敗序列的影響。4.在其他有關的變項上，亦受到成敗的影響。如在成功情境，自評的工作表現較好，設定的滿意標準較高，認爲作業較需技巧性，覺得自己的控制程度較高，並認為他人的工作表現會較好，而自己表現與滿意標準間的差距較小，評量自己與他人的工作表現間的差距亦較小。

The purpose of this research is to understand the integrated variables of mental health, including perception of life events, personal attributions, coping styles and self-esteem, then to know how success or failure of events influence attributions and expectancies of two groups with different mental health status from time serial approach. 301 college students are used as subjects, we find "self-esteem" and "attributions" explain major variance of mental health (by "Globle Severity Index"); "Coping styles" is the main explaining variable of "self-esteem", and "attribution" is mainly explained by "influentiality of life event" and "importance of life event". Subjects with poor mental health status (having higher GSI score) evaluate themselves at lower actual self-control level, show more discrepancies between actual and ideal self-control, and between other's and self-control. The also rate most of attribution items higher than better mental health subjects. With corrected attributional scores, the poor mental health students tend to have more external attribution and less internal attribution, regardless toward positive or negative events. When subjects are offered two seguential tasks (success and failure tasks in random order), all subjects tend to make internal-unstatable attribution in the experiemental situation, and at the second task, subjects make more internal attribution in the success situation than the failure situation. No group differences are found. Expectancy changes are mainly influenced by success or failure trials immediately, and not affected by GSI scores or proportion/order of success and failure.