The aim of this research was to test a series of theoretical models based on Beck (1967) cognitive diathesis-stress and Kwon and Oei (1994) linear mediational models as well as earlier research findings to determine the best-fitting model to explain the aetiological processes of suicide attempts in Taiwanese people. The participants were patients diagnosed with Major Depressive Disorders (MDD) recruited from three hospitals in Taiwan. They were used for data analyses in both cross-sectional (main) study and longitudinal (follow-up) study. In addition, a sample of students recruited from three universities in Taiwan was used for data analyses in the generalized study to examine the generalization of the results from clinical depressed patients to nonclinical university students. In the main study, by the application of structural equation modeling (SEM) techniques, four initial models were compared using the MDD patients (N = 162). The SEM analyses showed that two interactional models failed to provide an adequate fit to the given data, suggesting that the hypothesis of interaction between dysfunctional attitudes and negative life events in predicting the psychopathology of Taiwanese MDD patients was not supported. The SEM analyses supported two mediational models in terms of goodness-of-fit. Because the two mediational models were very similar, they were combined to form a combined mediational model. The SEM analyses indicated that the combined model provided an adequate fit to the given data. After modifying the model to improve its goodness-of-fit, the final modified combined mediational model was selected as the most appropriate in representing the data of Taiwanese MDD patients. The final model revealed that dysfunctional attitudes mediated the relationship between negative life events and depressive hopelessness, which in turn increased depression, which then precipitated suicidal ideation, which finally resulted in suicide attempts. In addition, it was found that negative life events exerted direct influences on depressive hopelessness and suicide attempts; sex and age exerted direct influences on negative life events. However, social support buffered the impact of negative life stress on dysfunctional attitudes and compliance with medications prevented the development of depression. In the follow-up study, the final modified combined mediational model was validated and reexamined with two-wave panel data gathered from the same population of Taiwanese MDD patients who participated in assessments twice, separated by a six-month interval (N = 142). The SEM analyses showed that the model provided an adequate fit to the two-wave panel data, suggesting that the model can be applied for predicting suicide attempts over six months in Taiwanese MDD patients. In the generalized study, the findings obtained from the MDD patients were replicated in a sample of Taiwanese university students (N = 324). Results revealed that the final modified combined mediational model failed to fit the given data. The result suggests that the most appropriate model for Taiwanese MDD patients can not be generalized to Taiwanese students. Some cautions and limitations should be noted. First, the models obtained from clinical and nonclinical people in Taiwan should not be directly generalized to people outside Taiwan. Further research using clinical and nonclinical samples from other countries to cross-validate the models was suggested. In addition, the researcher’s interventions during the follow-up period may disturb the relationship between predictor variables and subsequent suicide attempts. However, the problems appear to be unavoidable because of the research ethics of protecting participants from suicidal risk. / http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1339675 / Thesis (Ph.D.) - University of Adelaide, School of Psychology, 2008