NDARC Technical Report No. 1 (1989)
It is widely accepted that alcohol use is influenced by peer and parents' use and attitudes towards alcohol, but the mode of the influence is unclear. Is it direct? That is, do our parents' and peers' behaviour and attitudes serve as a model for our own? Or is it indirect? That is, do our parents' and peers' behaviour and attitudes influence the norms we acquire about drinking which, in tum, influence our own alcohol use?
In this Technical Report, Keats and his colleagues have used the method of path analysis to address this issue. Path analysis tests hypotheses about the direction of causal influence between variables by examining the pattern of correlations between them when the effect of other variables has been statistically controlled for. For example, if our parents' use of alcohol indirectly affects our own use via our norms, then any correlation between our parents' use of alcohol and our use should disappear when our own norms are controlled for. If, however, our parents' use of alcohol directly affects our use, then controlling for our norms will leave the correlation between our parents' and our use of alcohol unaffected. If our parents' use of alcohol affects our use of alcohol both directly and indirectly then we can estimate the size of both effects by examining the degree to which controlling for our nonns reduces the correlation between our own and our parents' use of alcohol.
Keats and his colleagues have studied the influence of parental and peer alcohol use and attitudes on adult alcohol use in both an alcohol-dependent and a community sample. In the process of carrying out this work, they have developed and validated the Australian Inventory of Alcohol Usage (AIAU). This is a specifically Australian questionnaire for assessing frequency and quantity of an individual's alcohol use, and the extent of his or her dependence on alcohol and alcohol-related problems.
The first question concerned the influence of demographic variables such as age and occupational status upon alcohol use (there were too few women in the sample to make an analysis of the effects of sex worthwhile), The results showed predictable relationships between age and occupation in the control sample. Younger persons consumed more alcohol than older persons, and the unskilled workers consumed more than the semi-skilled and skilled.
The second question Keats and his colleagues addressed was: Do social processes influence alcohol use among adults in the same way as they do among adolescents? To answer this question, they compared the results of path analyses in a corrununity sample of adults with those previously obtained in community samples of adolescents. The results suggested that adult alcohol use was more strongly influenced directly by parental and peer use and attitudes than was adolescent alcohol use. Adult alcohol use was equally influenced by preferences and norms whereas adolescents' alcohol use was more strongly related to their preferences than to their norms.
The third question Keats and his colleagues posed was: What role do these social processes play in alcohol dependence? This question was addressed by comparing the results of path analyses in the alcohol-dependent and control samples. The result indicated that in alcohol-dependent subjects, parental and peer attitudes did not have an indirect effect on alcohol use via internalized values. Rather, there was a direct effect of peer use on the alcohol consumption of the alcohol-dependent sample. In the community control sample, by contrast, the effects of parental and peer use and attitudes on the person's alcohol use were indirect.
Their findings have implications for future investigations of social processes in alcohol use and dependence. They have provided a promising instrument for assessing alcohol usage and dependence which can be applied in Australian samples. The study also reveals the potential usefulness of path analytic methods in disentangling the contribution of social variables to alcohol consumption and dependency. As theoretical ideas mature, and as valid and reliable measures of social variables are developed, more sophisticated methods of exploring patterns of causal relationships between variables (such as COSAN and LISREL) can be used.
It is perhaps a little premature to derive strong practical implications from the fmdings of the study. As the authors note, the current findings are based upon cross-sectional our confidence in the relationships observed will be increased if the same sorts of relationships are borne out in the prospective studies that the authors recommend should be undertaken.