A number of limitations indicate caution in interpreting this study’s results. First, all self-report data are subject to unintentional (e.g., through recall error) or intentional misreporting. If the ability or willingness to report AUD symptoms were positively correlated with recall or reporting of other mental disorders, this could upwardly bias estimates of their associations. Despite evidence that social desirability bias is inversely related to reporting of both substance abuse (Cox et al., 1994; Davis et al., 2010; Johnson and Fendrich, 2005; Welte and Russell, 1993) and other psychopathology (Kuentzel et al., 2008; Logan et al., 2008; McEwan et al., 2009), we are unaware of any evidence that this form of bias contributes to exaggerated associations between the two. However, several recent studies have reported slightly increased OR between psychological and somatic conditions/symptoms resulting from social desirability bias (Baumeister et al., 2010; Horn et al., 2010; Ladwig et al., 2010). Fortunately, social desirability bias and its associations with reporting alcohol use and problems do not appear to vary by sex (Welte and Russell, 1993; Kolarcik et al., 2009); accordingly, there is
