Abstract

The present studies tested a linear mediation model of social phobia derived from Beck, Emery, and Greenberg's(1985) cognitive theory of anxiety disorder. The linear mediation model assumed that dysfunctional beliefs as a cognitive vulnerability variable might predict socially phobic symptoms mainly by mediating of maladaptive automatic thoughts as a cognitive product of the beliefs activated by threatening social situations. In Study 1, the linear mediation model and two alternative models were tested with cross-sectional data from 542 undergraduates. The results of structural equation analyses showed that the linear mediation model fitted the data very well, while the simple effect model didn't and in the full model direct path of dysfunctional beliefs to socially phobic symptoms wasn't significant. Study 2 attempted to examine whether the results of Study 2 might be cross-validated in short-term longitudinal data from 289 undergraduates. The results of structural equation analyses showed that the linear mediation model fitted the data, while the simple effect model didn't and the full model as a saturated model was not adequate in explaining obtained data. Hence, the short-term longitudinal data of Study 2 as well as the cross-sectional data of Study 1 supported the linear mediation model. In conclusion, the findings of the present studies suggest that the linear mediation model apply to the explanation of socially phobic symptoms well.