Drinking behaviors of large groups : studies disentangling population drinking in Sweden

Abstract: Background: The starting point for this thesis was the unexpected development of per capita alcohol consumption and youth drinking in Sweden. There has been an observed decline for both: in the adult population since 2004 and among youth since 2000. Alcohol has at the same time become cheaper and more available. Given these circumstances, the expectations from previous research would be that consumption increased. The recent development in Sweden provided an opportunity to examine how changes in population drinking are dispersed and furthermore how these overall changes transfer to changes in various sub-groups of the population. Objective: The overall aim of this thesis is to disentangle changes in overall population means to get a better understanding of what happens within a population when consumption changes and also, to some extent, what influences those changes. More specifically, the individual studies focus on; (a) Are temporal changes in the general population’s alcohol consumption collective? (b) Is the theory of collectivity of drinking cultures applicable to changes in alcohol consumption also in a population of youth? (c) Is there a transmission of drinking from the adult population to the youth population? (d) Are there differences between generations in alcohol consumption and can these differences be explained by long-term effects of alcohol policy? Method: Studies I & II focused on the dispersion of alcohol consumption within populations and how this changes when there are changes in the overall mean consumption. Means were compared across different sub-groups and across time using both parametric and non-parametric tests. Study III used aggregate time-series analysis to examine the association between changes in per capita alcohol consumption and drinking among youth. Study IV used a fixed-effect regression model and a post-estimation Wald test to examine differences in drinking between cohorts that grew up during periods with different alcohol policies. Results: Studies I & II showed collective displacements of consumption in all sub-groups when the overall mean changed. In absolute terms, the changes were most pronounced among the heaviest drinkers while the relative decrease was inversely related to the initial consumption level. The results from study II further showed that the decline in consumption among the 10 % that drank the most accounted for 37.5 % of the overall decline in youth drinking in Sweden, while the decline among the bottom half that drank the least only accounted for a little more than 14 % of the overall decline. The increasing rate of non-drinkers among youth thus had a marginal effect on the overall level of alcohol consumption among youth. The results from study III showed that there was a positive association between changes in per capita alcohol consumption and changes in youth drinking. This association has however become weaker since 1995 and thus a change in per capita alcohol consumption is not a viable explanation for the decline in youth drinking. Study IV showed that there were significant differences in drinking between cohorts that grew up during periods with different alcohol policies. The cohort that grew up during a more restrictive period had a significantly lower consumption level than the reference cohorts. However, all cohorts changed their consumption in the same way, albeit from different starting points. Discussion and conclusion: Even though the consumption trends during the past decade were unexpected, based on assumptions drawn from previous research, a main conclusion of this thesis is that changes in drinking are collective, for all consumption segments from light to heavy drinkers and for most population sub-groups. The results from all four studies in this thesis lend support to the theory of collectivity of drinking cultures. With the use of different statistical methods and the availability of high quality data, the results corroborate the theory by providing empirical findings of collective displacements of consumption across time. Furthermore, the findings also complement and expand the theory, by showing that adolescents should be incorporated in the collectivity and that there is a collectivity of drinking that transcends generations. A change in per capita alcohol consumption is, however, not a viable explanation for the observed drop in youth drinking.