A social relations model test of Sullivan's anxiety hypothesis.

Interpersonal relationships and mutual influence are important aspects of both personality and behavior. However, empirical tests of mutual influence in anxiety have not occurred because of difficulties in design and assessment. In this report, we present a study of two training groups of graduate students and a study of an outpatient psychotherapy group. In both studies relationship-specific variance was significant and accounted for a substantial proportion of the systematic variance. In the training groups, there were also significant individual differences in experienced anxiety. These studies support the importance of relationships in anxiety but not Sullivan's hypothesis of the exclusive interpersonal nature of anxiety (Sullivan, 1964). The results address Endler and Magnusson's (1976a, 1976b) interactional approach to anxiety by assessing dynamic interaction rather than mechanistic interaction. In addition, these studies extend the use of the Social Relations Model to a new area, anxiety, and demonstrate its use in separating relationship-specific adjustments in anxiety from individual differences in anxiety.