Social phobia and problem drinking often co-occur leading to greater disability and poorer response to treatment. There are broad models of the comorbidity of these disorders. Morris, Stewart & Ham (2005) suggest a broad two step model – that drinking begins as a acceptable way of controlling social anxiety and that as drinking continues dependence develops. Continued drinking may maintain or worsen social anxiety. However more specific mechanisms are unclear. Rapee & Abbott (2007) provide a model of the cognitive mechanisms underlying social phobia. In social situations attentional resources are diverted from the threat, performance appraisal is negatively biased, the probability and consequences of negative evaluation are inflated and thus state anxiety increases. This study examines alcohol expectancy effects on mechanisms in a sample of 87 undergraduate drinkers (67% screened positive on the AUDIT). We expected that alcohol expectancies would lead to reduced anxiety. Participants were randomly allocated to consume a drink they were told contained 2 standard drinks of alcohol but did not (n=50) or a control condition (n=40). After 20 minutes participants gave a 3 minute speech into a video camera that they were told would be rated by a panel of experts. Then self report ratings of intoxication, anxiety, attentional focus, appraisal of performance and the probability and cost of negative evaluation were taken. Only three in the placebo alcohol condition did not believe they had consumed alcohol and were excluded from following analyses. Contrary to hypotheses greater levels of perceived intoxication were related to higher levels of anxiety. Preliminary analyses indicate that as perceived intoxication increased, the focus of attention was more inappropriate and the probability (but not cost) of negative evaluation increased. Thus the unexpected effect on anxiety was probably due to these processes working together. Further analyses and their immpications will be presented.