Influencing College Student Drinking Intentions With Social Norms and Self-Schema Matched Messages: Differences Between Low and High Self-Monitors

Abstract

College students were exposed to either a self-schema matched message (emphasizing how binge drinking is inconsistent with personal values) or a social norms message (highlighting the true normative drinking behavior of peers). As predicted, low self-monitors intended to drink significantly less alcohol if they received the self-schema matched message versus the social norms message, and high self-monitors intended to drink less if they received the social norms message versus a self-schema message. While previous research supports both techniques for marketing responsible college student drinking, the current results suggest that each method may be especially effective for certain audiences.

KEYWORDS:

• alcohol
• messages
• self-monitoring
• self-schema
• social norms

Notes

Binge drinking is defined as consuming at least four drinks on one occasion for females and at least five drinks on one occasion for males (Wechsler, Dowdall, Davenport, & Rimm, 1995).

Self-schema should not be confused with the function the person's attitudes serve (see Petty, Wheeler, & Bizer, 2000). A self-schema is more complete depiction of the individual's self.

The initial sample included 188 students. Data from two participants were excluded because the participants’ responses to a few questionnaire items did not follow instructions, which was an indication that the participants may have not followed instructions appropriately throughout the study.

The reliability of the self-schemas as an individual difference measure has been demonstrated with large numbers of participants (i.e., over 7,000) being able to easily select a schema with which they identify, and construct validity has been established from comparisons of participants’ responses to these descriptions and their responses to an alternate measure of self-schema (Brannon & Brock, 1994; Brock et al., 1990).

The messages have been pretested and have been used in published research (i.e., Pilling & Brannon, 2007) conducted with college students from the same population that was sampled from for the current study.

Participant characteristics may be associated with differences in baseline drinking behavior, and drinking intentions may vary by baseline drinking behavior. Therefore, including baseline drinking behavior as a covariate allowed us to identify any effects of the independent variables on drinking intention controlling for differences in pre-intervention drinking behavior. Including baseline drinking behavior as a covariate also would control for any differences in baseline drinking behavior associated with self-monitoring. We conducted an ANOVA to examine whether baseline drinking behavior differed between low and high self-monitors and found a statistically nonsignificant difference, F(1, 184) = 1.63, p = .20.

An additional analysis showed that among all participants in the self-schema matched message condition, drinking intention did not differ significantly by participant schema type. That is, tailoring to schema type is not more effective for one type than for any other.

It is impressive that the current analyses found significant differences based upon self-monitoring even though they included some persons who were near the midpoint of the scale (and therefore not strongly high or low self-monitors). Therefore, as is customary in research examining the effects of individual difference variables, the data were also analyzed using more extreme groups (by eliminating the third of participants who scored near the middle of the self-monitoring scale). This ensured that our participants clearly fell into one or the other category of self-monitoring. These results were particularly impressive given that the statistical power of the analyses was reduced due to the elimination of the participants for whom self-monitoring was not as relevant.