Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)




Although there is considerable evidence to support the contributions of life event stress to subsequent psychological and physical symptoms, the correlations reported between obtained stress scores and illness measurements have been low. This study was undertaken to improve prediction of physical and psychological complaints subsequent to life event stress by more comprehensive measurement of life event characteristics and by assessment of locus of control, social supports, and somatization as moderating variables in the life stress-illness process. An overview of psychosomatic medicine was given and the relevant life event research was reviewed. Fifty undergraduate students completed the experimental version of the Life Experiences Survey (LES), Rotter's Internal-External Control of Reinforcement Scale (I-E Scale), the Minnesota Multiphasic Personality Inventory (MMPI), and the Interview Schedule for Social Interaction (ISSI). Subjects then completed psychological and somatic complaint checklists (modified SCL-90-R, modified Wahler Physical Symptoms Inventory) every four weeks for a three month period. Pearson product-moment correlations were calculated, providing an intercorrelation matrix for all independent and dependent variables, and multiple regression analyses were used to determine the best predictive models. Results did not support the first hypothesis that inclusion of the four idiographic subscales of the experimental LES (desirability, change, anticipation, control) would improve prediction of psychological and/or physiological complaints over the counting of event occurrences. Statistically significant improvement in prediction was found when the measures of social support (AVAT, ADAT%, AVSI, and ADSI scales of the ISSI), locus of control (I-E Scale), and somatization (HS and HY scales of the MMPI) were added to LES scores in an heirarchical regression analysis. However, a more parsimonious and powerful predictive model was derived using a simultaneous stepwise regression analysis: This model consisted of the HS and I-E scales and the change and desirability subscales of the LES. Social support measures were not correlated with either criterion variable nor with any of the predictor variables. These results were discussed in relation to previous studies concerning life event stress, social supports, somatization, and locus of control in relation to health. The clinical implications of this study were also discussed.