An additional unexpected finding which was observed in both men and women was that social concerns was not a significant predictor of social anxiety symptoms as it is described in the model of AS. Moreover, social concerns did not predict any of the symptoms of anxiety or depression. Several reasons may explain this finding. Firstly, the internal consistency of this subscale was considerably low. Based on Nunnally’s (1978) requirement of reliability of an instrument to be no lower that .3, the social concerns subscale did meet the cut- off; however, the low reliability is questionable as it has also been observed in previous studies. If the anxious and depressive symptoms were differentially related to some of the items in the ASI social concerns subscale, the poor performance of the ASI social items might have masked any observable differential relationships between social concerns and the symptoms of anxiety and depression.