The Relationships Among Paradigm Beliefs, Attributional Styles and Depression Scores

Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)



First Advisor

William Lyddon

Advisor Department



The purpose of this study was to examine the degree of relationships among paradigm beliefs, attributional styles, and depression scores using a college student sample. Two-hundred forty adult participants enrolled in undergraduate psychology courses at a southeastern university completed the Social Paradigm Belief Inventory, the Attributional Style Questionnaire, and the Beck Depression Inventory. Data analyses involved techniques of multiple linear regression. Results revealed no statistically significant predictors of depression by absolute, dialectical, and relativistic paradigm beliefs or by internal, global, and stable attributional style dimensions. The paradigm beliefs and attributional style dimensions accounted for 3% of the variance of depression scores. Results further indicated only two statistically significant paths among the variables in question. The stable and global attributional dimensions for negative events were significantly predicted by absolute paradigm beliefs. Results revealed no other statistically significant direct or indirect paths among attributional styles, paradigm beliefs, and depression scores. Results suggest that the proposed model may not be a viable theoretical scheme for understanding the variables under question. Paradigm beliefs and/or attributional styles may only be predictive of future risk of depression rather than current depression. The influence of paradigm beliefs and/or attributional styles on depression may be most relevant when a negative event happens directly to an individual, or when he or she is cognitively or emotionally challenged in some way, at which time a particular set of causal ascriptions is employed to interpret the event. If the proposed model is indeed a viable theoretical scheme for understanding the variables under question, some factors may have limited the degree to which relationships among the constructs could have been detected. Examples of such factors include low sample reliabilities of the ASQ dimensions, lack of full comprehension of the ASQ instructions by many participants, and the relative absence of depression in the present sample. Such factors are explored to help explain the lack of statistical significance in the overall findings. Suggestions for future research are also provided.