Abstract

The present studies tested the validity of a linear mediation model in explaining somatization. The linear mediation model is a comprehensive model that includes the personality variables, such as negative affectivity and alexithymia, and cognitive processes, such as somato-sensory amplification and somatic attribution. The assumption was that negative affectivity and alexithymia as vulnerability variables might predict somatic symptoms mainly through the mediation of cognitive variables such as somato-sensory amplification and somatic attribution. The linear mediation model and two alternative models were tested in 124 psychiatric patients and 200 undergraduates. The results of structural equation analyses showed that the linear mediation model fitted the data very well in both groups, while the simple effect model didn't. The full model didn't fitted the data of psychiatric patients, but fitted that of undergraduates. In conclusion, the linear mediation model of somatization was supported in two groups, but the full model also may be applied to undergraduates. Finally, these results with those of sex difference suggest that follow up studies are needed.