In this paper the results of a meta-analytic review about the effectiveness of psychological treatments for social phobia are presented. The literature search enabled to identify a total of 48 studies that met the selection criteria. The average effect size (as weighed by the inverse of the variance) was d+= 0.724 in the posttest and d+= 0.902 in the follow-up. These results permit us to affirm, in global terms, that these treatments of social phobia are clearly effective. Nevertheless, the homogeneity test did not reach statistical significance, by which we can assume that the effectiveness of exposure techniques, cognitive restructuring techniques, and social skills training can be considered homogeneous. This fact questions the psychological principles that underlie the effectiveness of these treatments. An explanation based on feared social stimuli exposure is proposed.