Despite these findings, several shortcomings regarding personality as a moderator of the pharmacological effects of alcohol (at least regarding response to stress) should be noted. Laboratory-based findings have been inconsistent (e.g., Sher, & Walitzer, 1986; see Sher & Wood, 2005) and there is a lack of laboratory studies that used structured and empirically verified measures of personality (e.g., Five-Factor measures of personality; Costa & McCrae, 1992). Thus, integrating these initial findings into contemporary models of personality is somewhat difficult. Recent research using within-person, process-oriented methodology (e.g., daily-diary data) and five-factor measures of personality (e.g., NEO–PI–R; Costa & McCrae, 1992) has also produced inconsistent findings regarding personality as a moderator of the alcohol-stress relationship. For example, Armeli et al. (2003) hypothesized that the stress-response dampening effects of alcohol should be stronger for participants high in neuroticism, extraversion, and self-consciousness. Though there was some support for these hypotheses involving extraversion and self-consciousness, results were inconsistent across the five moods assessed in the study (i.e., sad, nervous, relaxed, happy, and angry). Thus, though pharmacological vulnerability models of personality are conceptually important, empirical support for these models has been somewhat inconsistent.
