|
Article Excerpt
There is a large body of research in different cultures on the distribution of drinking problems across different sociodemographic categories, although most of the studies come from outside the Nordic countries. Overall, it is not always made clear why studying this is important. In our view, the sociodemographic variables reflect important social differentiations in the population, in terms of life stage and life chances. They are also related to visible or easily understood conditions that make a difference in social life. As variables relating to social realities, they also provide a way to talk about differential risk and risk groups which can be of practical significance in policy and prevention planning.
The default hypothesis in most analyses is that there will be significant differences by sociodemographic variables--in amount and pattern of drinking, and in rates of alcohol-related problems. Over 30 years ago, Mark Keller made fun of the default hypothesis in formulating Keller's Law: "The investigation of any trait in alcoholics will show that they have either more or less of it." Keller based his law on a literature review of (primarily biochemical) studies addressing the differences between alcoholics and non-alcoholics. To mention just a few examples, alcoholics were found to be more allergic, less bold, more color blind, less religiously active, less hypnotizable, less married, less fecund, less body-hairy (men), more rigid--and more thirsty (Keller, 1972).
However, many studies, in their ambition to predict problems by sociodemographic variables, ignore the maybe too self-evident intermediating role of drinking habits (see also e.g. Makela, 1999b; Room, 2000). Modeling drinking consequences by sociodemographic background without controlling for drinking itself may result in modeling predictors of drinking rather than the propensity to develop problems. To identify groups with a high prevalence of drinking problems has, as mentioned above, undoubtedly its own value. However, a distinction is too seldom made between a high prevalence of problems on the one hand, and a high risk of problems given a particular level of drinking, on the other. An exception are risk curves analysis predicting risk of harm from a particular level of drinking--typically for one sociodemographic characteristics at a time, thus not examining the independent contribution of each (e.g. Midanik et al., 1996; Ramstedt, 2002).
In this article, we will explore the relation of sociodemographic categories with rates of alcohol-related problems in a national sample of Swedish adults. We are also interested in pushing the analysis a step further, addressing the question: for a given level and pattern of drinking, how much difference do different sociodemographic variables make in predicting whether or not an alcohol-related problem will result? To some extent, this issue has been previously addressed, often with the default hypothesis. But there is also a counterhypothesis in the literature, usually attributed to Lee Robins (Robins et al., 1962; Cahalan & Room, 1974). This "two-step hypothesis" expects that sociodemographic variables will indeed predict levels and patterns of drinking, but that they will not strongly predict who gets into problems for a given level and pattern of drinking. Here, it is hypothesized, sociodemographic factors will give way to other predictors such as personal characteristics.
Prior research
Sociodemographic background as a significant predictor
There are studies suggesting that affiliation to a specific population subgroup per se, irrespective of drinking, implies an enhanced risk for alcohol-related harm. Examples of such risk groups found in the literature are youth, sometimes men and sometimes women, divorced or single people, people from low social strata, certain ethnicities, or people resident in regions with "dryer" drinking subcultures. Roughly speaking, there might be three main explanations to these differences: a) variations in actual patterns of drinking, b) biological differences, c) differences in social response. Considering social response, many kinds of alcohol-related harm, and social harm in particular, involve a strong element of social construction (Room, 1996). Thus, some of the differences found by sociodemographic background, controlling for drinking, may be attributed to lower social tolerance and a higher degree of stigmatization and social control directed to certain population groups.
Considering variations by age, the fact that the same amount of alcohol seems to be associated with more harm in young people is sometimes explained by their lower physical tolerance (e.g. Hurst et al., 1994). In addition to a different physiological reaction to alcohol, young people might be more targeted in terms of social control. Moreover, there might be an actual age difference in subjectively experienced problems, e.g. impaired control, from the same drinking (Harford et al., 1991) and thus in the propensity to report these consequences. On the other hand, different age-related norms around how to behave when drinking may imply that people above a certain age are more likely to be stigmatized given the same drinking behavior. This can increase the level of experienced social problems but, at the same time, might lead to a lower willingness to admit and report these experiences. Results from survey data show a positive association between young adult age and harm from alcohol, but often not a very strong one (Cahalan & Room, 1974; Clark & Hilton, 1991).
Another variable included in most analyses of drinking behavior and assumed to have a high explanatory value is gender. Indeed, men tend to report more negative consequences from drinking in all cultures where drinking has been studied (Wilsnack et al., 2000; Fillmore et al., 1997). But, when gender differences in drinking habits are taken into account, the pattern is not quite the same. In fact, there are both biological (different metabolism and distribution of alcohol in the body) and social (different norms and expectations) explanations of why the same level of drinking might actually lead to more or different kinds of harm among women (Graham et al., 1998). In an analysis of U.S. data by Harford et al. (1991), for an equivalent level of daily average consumption women were at higher risk of experiencing physical health problems than were men. For other problem areas studied (belligerence, social/family problems, work/financial problems, and legal problems), however, gender turned out not to be important in the relationship between drinking habits and harm. Similarly, in an analysis by Caetano et al. (1997), little gender effect was seen in the risk curves for DSM-IV (Diagnostic and Statistical Manual of Mental Disorder-fourth edition) dependence. Also analyses of risk curves by Midanik et al. (1996) show a rather similar degree of risk for harm (drunken driving, dependence, and job problems) for men and women at the same level of drinking, except for the highest consumption level, where the risk is higher for women than for men. In contrast, according to Canadian data, modeling drinking consequences within six life areas (friendships/social life, health, happiness, home life/marriage, work/studies, and financial position) by mean consumption, frequency of 5+ drinks, gender, age and educational level, men were 21 percent more likely to report two or more consequences (Room et al., 1995). That women's drinking is more controlled and/or women's bodies respond differently to alcohol is supported by a Canadian experiment conducted by Fillmore & Weafer (2004), showing that men displayed greater impairment of inhibitory control compared to women when drinking the same amount of alcohol. Women, instead, reported more sedation effects. Analyzing data from a Dutch population sample, Bongers et al. (1998) found that women were as likely as men to report at least one alcohol-related problem for the same level of drinking--except for female light drinkers, who reported less problems. No gender difference in the severity of problems could be found, but men tended to accumulate more problems than women did.
The issue of social stigma has received particular attention in a number of studies of the association between socio-economic status (typically indicated by education, income, or occupational status) and harm from alcohol. For instance, in a historical analysis of the association between social class and alcoholism, Park (1983) came to the conclusion that there is a tendency to exaggerate the rate of alcoholism among the lower classes. Similarly, Cahalan & Room (1974) found that both drinking problems and abstinence were associated with lower social position (among American men) while the ratio of problems to drinking was lower at the top of the social scale. They suggested that poor people were more exposed to social consequences from a given pattern of drinking, quoting an old English music-hall song "It's the poor wot gets the blyme" [blame]. On the other hand, Harford et al. (1991) found a higher ratio of consequences to intoxication among the highly educated, studying a general population sample of both American men and women. In Sweden, Leifman (1998) found no significant association between socio-economic background and frequency of binge drinking when controlling for volume, while a register-based analysis of alcohol-related mortality clearly showed higher figures in low social strata. Thus, there seems to be stronger evidence that more harm related to the same level of drinking can be observed in the lower strata rather than the other way around. However, it must be remembered that the relationship is not static. Studies from Nordic countries have shown that, when the total consumption in a given culture moved in a certain direction, a disproportionate development could be observed in different socio-economic groups (Makela, 1999a; Romelsjo & Lund berg, 1996; Sulkunen, 1989). In addition, the direction of the relationship is certainly not fixed, as discussed later.
It has been shown in a number of studies that marital status is connected to risky drinking habits; individuals without a spouse, and in particular those who once had a spouse, are at a higher risk of heavy drinking and alcohol-related harm. According to some results, the mechanisms do not seem to work the same way for men and women. For instance, analyzing data from multinational longitudinal studies, Fillmore et al. (1997) found that having never been married was associated with more frequent binge drinking, but only among young men. On the other hand, separation or divorce increased young women's consumption per occasion more that it did men's. Considering our research question--what is the difference in alcohol-related harm when controlling for actual drinking--the relationship has not been studied much for marital status.
Location of residence is another variable often included in models predicting variations in alcohol-related harm. Two main indicators are typically used, one referring to the geography and the other to population density. Understanding geographical differences in drinking habits and related harm requires looking beyond the region variable, taking into account also subcultures typical of specific areas, associations with other demographic circumstances such as ethnicity or religion (Room, 1983), and country-specific regional sub-cultural variations. For Sweden, for instance, the difference between South and North reflects both the distance from other European countries--making alcohol more available in the South--as well as different norms around drinking, what is called a "continental" drinking style being more prevalent in the southern regions. In the U.S. literature, a substantial emphasis has been put on a distinction between "wetter" and "dryer" regions, where dryer regions were those with a history of stronger temperance movements, more abstainers, and less frequent heavy drinking. In most studies, more problems from the same level of drinking are observed in "dryer" regions. In terms of population density or urbanization, dry regions most often apply to the countryside in contrast to cities (Room, 1983). When Cahalan & Room (1974) compared drinking practices in wet and dry U.S. regions, they observed that heavy drinking led to more trouble in dryer than in wetter areas, particularly problems with wife and job. In the wetter geographic regions, only problems associated with continuous heavy intake...
|
