Psychometric properties of the Important People Instrument with college student drinkers.

ABSTRACT

Alcohol use in college students is associated with a number of negative consequences, and specific characteristics of college students' social networks have been linked to hazardous or harmful alcohol use. College represents a time when social networks are changing significantly, both in terms of composition and relative influence of network members. However, the reliability and validity of one of the most common measures for assessing alcohol-specific social support, the Important People Instrument (IPI; P. R. Clifford & R. Longabaugh, 1991), have not been established in college student samples. The aim of the current study was to examine the psychometric properties of the IPI administered in computerized and paper-and-pencil formats within a non-treatment-seeking sample of college drinkers (N = 197). Test-retest reliability estimates for the overall sample indicated that all indices had acceptable reliability. While the models tested within a confirmatory factor analysis (CFA) framework exhibited mediocre fit, a three-factor model appeared to offer the best fit overall. Consistent with previous findings with treatment-seeking samples, negative consequences of drinking were positively associated with network drinking behavior, but not network support for drinking or network general support. The IPI can provide reliable and valid information about network drinking behavior, network support for drinking, and general support among college students, although there is room for improvement in the measure. Future research may improve the measurement of alcohol-specific social support by measuring additional domains of the construct or by taking more contextualized approaches.