Public Opinion on Alcohol Policy: A Review of U.S. Research

Article excerpt

Public opinion on alcohol policy has received little research attention until recently, although a coherent research agenda in this area holds promise of a better understanding of the public health policymaking process. This paper reviews and critiques the major analyses of policy opinion in the United States, discusses national support for specific policy options, and addresses trends in policy opinion. The dynamics of policy opinion are complex, allowing for few generalizations. Environmentally based alcohol policies receive mixed levels of public support, and, overall, support decreased somewhat throughout the early 1990s. Recommendations for continued monitoring and improved assessment of policy opinions are made.

KEY WORDS: Alcohol, policy, United States, public opinion, research.

Although public sentiments concerning alcohol issues have long been a subject of historical and social research, public opinion in relation to specific alcohol policies has only recently emerged as a research area in its own right. In the late 1970s, Goodstadt et al. stated, "[apparently, no sophisticated studies of public attitudes towards [policies regarding] the pricing of alcoholic beverages have been completed" (1978: p. 1630). Approximately ten years later, while Crawford (1987) reviewed 266 studies of a variety of attitudes towards alcohol use and misuse, another review turned up "no published studies of public opinion on alcohol policy in the scientific literature" (Wagenaar & Streff, 1990: p. 190). (Wagenaar has subsequently added to the literature, as discussed later.) Leedham (1987: p. 935) observed that "public opinion data on alcohol and tobacco policy are rarely collected on a regular basis by policymakers. Only where the data feed into decision making in other policy areas, such as taxation policy, is the information collected as part of a data base." Leedham adds that tobacco policy has received more attention than alcohol policy. Thus it would appear that attention given to policy opinions per se has lagged behind that given to generic attitudes towards alcohol. Although the body of literature in this area has been growing, especially subsequent to the U.S. policy development project discussed in this special issue (Greenfield et al., 2004), the literature remains fragmented. Our intention here is to critically review major analyses of public opinion on alcohol policies in an effort to set a benchmark for the field and make suggestions for future research.

The emergence of a coherent agenda for alcohol policy opinion research has the potential to improve both theoretical understanding of the role played by public opinion in public policy making (Greenfield, 1994) and the evaluation of which public health policy strategies hold the greatest promise of support (Room et al., 1995). While public opinion in and of itself is seldom the central determinant in policy agenda setting, it is in principle an important part of the policymaking mix. A considerable literature in political science addresses the role of public opinion in policymaking (Arnold, 1990; Gozenbach, 1996; Kingdon, 1995, p. 1994; Stimson et al., 1994). Leedham's succinct description is that public opinion "forms part of the context within which strategy is determined and provides an aspect of legitimacy for a chosen course of action," playing a validating role in the policymaking process (Leedham, 1987, p. 935). From a public health perspective, policy opinion research also has the potential to reveal gaps in public understanding of health issues and opportunities for enhanced educational efforts aimed at helping the public understand the rationales for particular policies.

To some extent, the growth of interest in alcohol policy opinion research reflects the shift, begun in the 1970s and established by the mid-1980s, from an alcoholism focus in research literature to an alcohol problems focus generative of a diverse array of public policy proposals directed toward strengthening environmental controls (Edwards et al., 1994; Moore & Gerstein, 1981). Several sources of policy opinion data exist from the earlier period, including federally sponsored research (Harris, 1972, 1973a, 1973b, 1974) and studies of alcohol abuse in a social-problems context (Cahalan et al., 1974; Cameron, 1981; Goodstadt et al., 1978; McKenzie & Giesbrecht, 1981). Starting in 1989 as part of an emerging trend of systematic attempts to address policy opinion in a social epidemiological context, Canadian and American national alcohol surveys added items on specific alcohol control policies. Work reviewed comes from five major data sources on alcohol-related public attitudes and policy opinions: (a) the inclusion of 11 policy opinion questions (see Table 1) on the Canadian National Alcohol and Other Drugs Survey (NADS) in 1989 (Eliany et al., 1992), subsequently used extensively in U.S. national and Canadian provincial and national surveys; (b) the Impact of Alcoholic Beverage Warning Labels Survey conducted by the Alcohol Research Group (ARG) in Berkeley, California, providing both a U.S national cross-sectional data series from 1989 to 1994 and an Ontario, Canada, no-intervention reference series covering the NADS 11 policy opinion item set (Greenfield, 1997a; Hilton & Kaskutas, 1991; Kaskutas, 1993a); (c) ARG's quinquennial U.S. National Alcohol Survey (NAS) cross-sectional series with data collection on policy opinions in 1995/96, 2000/01, and 2004/05 (now being fielded); (d) the annual Ontario Alcohol and Other Drug Opinion Surveys (OADOS; now called CAMH Monitor), sponsored by the Addiction Research Foundation (ARF) from 1992 onward (Anglin, 1995; Bondy, 1994; Ferris et al., 1994; Paglia, 1995), and (e) a 1997 telephone survey of 7,021 "mainland" U.S. adults introduced as a study of "views on teenage drinking problems," funded by a University of Minnesota grant from the Robert Wood Johnson Foundation (A. C. Wagenaar, P.I.) (Wagenaar et al., 2000; Latimer et al., 2001, 2003). Private opinion polling companies have, in addition, periodically included questions about alcohol issues in national surveys (Gallup, 1972, 1985, 1987, 1989; Harris, 1983). The extent of this latter data resource is not assessed here, but private polling data are acknowledged to have utility if used carefully in policy opinion research (Gozenbach, 1996; Hilton & Kaskutas, 1991; Leedham, 1987; Wagenaar & Streff, 1990).

In this review we focus on public opinion on alcohol control policies, drawing on published and some unpublished analyses of the data sets outlined above and other research on public support for specific policy preferences. Consideration is given to county- or state-level data (Fischer, 1989; Jones & Greenfield, 1991; Jones-Webb et al., 1993; Moloney & Emanuels, 1995; Wagenaar & Streff, 1990; Wagenaar et al., 1988; Schmid et al., 1990) as well as national-level data. While a sizeable body of non-U.S. data and research reports exists (e.g., Ahlström & Österberg, 1992; Anglin, 1995; Anglin et al., 2001; Bondy, 1994; Cameron, 1981; Casswell et al., 1989; Eliany et al., 1992; Ferris et al., 1994; Giesbrecht & Greenfield, 1991, 1999; Giesbrecht & Kavanaugh, 1999; Giesbrecht et al., 2001; Goodstadt et al., 1978; Paglia, 1995; Room et al., 1995), we restrict our attention to public opinion research in the United States (Giesbrecht & Greenfield, 1999; Latimer et al., 2001, 2003; Room et al., 1995). In addition, this review focuses on analytic rather than purely descriptive reporting on policy opinion. We begin by reviewing norms around alcohol use and then consider dimensions of public support for specific policy alternatives and critical issues in refining and improving policy opinion research.

Norms, attitudes, and drinking behaviors

Attitudes toward drinking and drunkenness, while known to predict behavior only imperfectly, reflect stances regarding the social location (status and function) and context of alcohol use that underlie the formulation of specific policy preferences. A 1964-to-1984 trend study examined such attitudes by inquiring about acceptable consumption levels in four social contexts: drinking with people from work, with neighbors, with close friends, or among people from church (Hilton, 1991a). In this 20-year period (during most of which time per capita consumption was increasing), three of these four settings became "wetter" (more permissive); only "among people from church," originally a very dry setting, showed no change. These results were interpreted to suggest that norms about drinking were becoming more liberal. In a study of situational drinking norms (defined as subjective levels of acceptable consumption in various situations), Greenfield and Room (1997) used NAS U.S. adult data from 1979, 1984, and 1990 to examine trends in acceptance of "drinking at all" and of "drunkenness." The authors found considerable stability in the normative structure of American drinking during the 1980s, but they noted some changes as well, finding "a diminished acceptability of any drinking in dryer but not in wetter situations, and of heavier drinking in public situations or in situations which will require travel afterwards" (p. 45). These reductions were seen as paralleling the downturn of alcohol consumption in the U.S. from the beginning of the 1980s through to 1990 (Midanik & Clark, 1994), which has not continued beyond 1995 (Greenfield et al., 2000a) and in fact has begun to trend upward (Kerr et al., 2004). Thus, based on studies in the 1980s and early 1990s (the period from 1995 to 2000 has not yet been studied), norms around alcohol use seem to be associated with levels of consumption in the population, as do sociocultural variables.

Attitudes toward drinking and drunkenness have also been looked at in terms of reasons for drinking (or abstaining from drinking) and by regional and gender differences. Using NAS data, Greenfield (1993) found between 1984 and 1990 some weakening over time among abstainers in "goes against religion" as a reason for not drinking, while moderate drinkers' reasons for limiting their intake saw little change, with a majority invoking health reasons. The same trends (downward for abstainers but not for drinkers) continued to 1995 (Greenfield & Klingemann, 1998). In the 1984 NAS data, Hilton (1991b) found that residents of wetter regions (defined in terms of percentage of abstainers and rates of heavy drinkers) reported more favorable opinions toward alcohol than those in drier regions. For men, this difference seemed attributable to regional differences in the proportion of abstainers. For women, differences seemed focused in the South; while there were equally heavy drinking rates and problem levels compared with other areas, Southern women were less likely to report favorable attitudes toward drinking and drunkenness than their counterparts elsewhere. In a similar vein, the Hilton (1991a) study cited above found that Catholics were considerably wetter than conservative Protestants, and men wetter than women, suggesting that the composition of social reference groups can also exert some effect on norms around alcohol use (see also Clark, 1985; Greenfield et al., 2000b).

Caetano and Clark (1999) examined trends in situational norms and alcohol-related attitudes among whites, blacks, and Hispanics from 1984 to 1995, extending the period considered by Greenfield and Room (1997)-1979 to 1990and investigating ethnic influences. No overall trends in norms and attitudes were seen, but variations between 1984 and 1995 were ethnicity- and gender-specific. Increased conservatism in attitudes toward drinking was seen among black respondents. Demographic predictors of attitudes were also ethnic-specific, again indirectly reinforcing the need to take ethnic group differences into account in policy opinion studies (Jones-Webb et al., 1993).

The overall picture of attitudes toward alcohol is complex. Changes over time in consumption and situational norms beg further research on the nature and direction of the relationships between these, changes in aggregate consumption, and changes in access and availability (see Giesbrecht & Greenfield, 2003; Holder, 1993; Room, 1987, 1991). A focus on the relative continuity of norms and attitudes may obscure changes in tenor-for example, the gradual replacement of moralistic attitudes with health-based ones (Beauchamp, 1988; Greenfield, 1993). Observed regional and gender differences suggest that studies of policy opinion should continue to include such variables and add measures of social context and locally prevailing conditions.

Variation in support by policy issues

Opinions on specific alcohol policies tap both attitudes toward alcohol and drinking, on one hand, and social and political values, on the other. Distinctions between attentive and inattentive populations (Arnold, 1990) and between specific policy opinions and global policy preferences (Stimson et al., 1994) are important to keep in mind in this regard. Global preferences measured by broadly phrased opinion items may not predict judgments about specific policy alternatives made by an attentive and informed public (Page, 1994; Yankelovich, 1991). Like changes in economic and social problem levels (Hacker, 1997), alterations in the gross national mood (Kingdon, 1995) can affect public support for alcohol control policies by shaping the political backdrop against which alternatives are presented and can interact with media coverage (Gozenbach, 1996; Lemmens et al., 1999). In the next section, we restrict ourselves to reporting on levels of public support and to some cautious analyses of underlying dynamics.

Support for an outright prohibition on the manufacture and sale of alcoholic beverages is the only control policy for which public opinion has been tracked for decades (Gallup, 1987; Hilton & Kaskutas, 1991; Room, 1991). Between 1936 and 1989, support for prohibition has fluctuated between 17% and 38%. As late as 1989, prohibition was favored by approximately one-quarter of the U.S. population-mostly by those who themselves abstain from use of alcohol (Hilton & Kaskutas, 1991). This group constitutes a core of support for stringent alcohol control policies (Greenfield, 1997a).

In a seminal paper on policy opinion, Wagenaar and Streff (1990) reviewed a series of national private opinion polls done between 1986 and 1989. These polls used a variety of questions related to excise taxes on alcohol and/or tobacco. They found levels of support for increased taxation ranging from a high of 81% to a low of 65%. Such high levels of support for increased taxation stand in contrast to less-thanmajority support for similar items found in other studies. Notably, question wording in polls finding stronger support for tax increases frequently linked policy alternatives with specific and widely supported policy outcomes (e.g., reducing the federal deficit).* For example, Wagenaar et al. (2000) found in a 1997 U.S. national survey that 82% of the adult respondents supported increased alcohol taxes when the item stated that funds were to be used for treatment and prevention of alcohol-related problems. The importance of making a cognitive link between a policy proposal and its intended outcome, discussed in the political science literature (Arnold, 1990) with regard to legislative decision-making, is also relevant to designing and interpreting results of public opinion surveys.

Wagenaar and Streff (1990) also reported on a 1988 telephone survey of the State of Michigan that found "overwhelming public support" (82%) for increased alcohol excise taxes, with support for other measures ranging downward to a low of 31% favoring social host liability. Other highly to moderately supported policy approaches were prohibiting concurrent sales of beverage alcohol and vehicle fuels (74%), administrative driver's license suspension on first driving while intoxicated (DWI) offenses (67%), limiting alcohol outlets (63%), and reducing blood alcohol concentration (BAC) to 0.05 (55%). Of the eight policies assessed, only limiting hours of sale (40%), commercial liability (35%), and social host liability were favored by a minority. Question wording may have affected observed levels of support. For example, the excise tax item read: Increasing efforts to reduce drunk driving will cost money. In order to raise the money, would you favor or oppose an increase in the tax on each bottle of beer, wine or liquor sold to pay for programs to reduce drunk driving? Heavily stressing the problem of drunk driving and the need to generate money that would solve the problem is likely to inflate endorsement in comparison with responses to less "linked" items. During political campaigns, the valence of such links is prone to political manipulation, as in the case of the California tax initiative where earmarking of the "nickel a drink" tax for (among other things) emergency rooms was portrayed in industry advertising as going into the pockets of "greedy" ER doctors.

In another study of policy opinion based at the state level in Minnesota, Schmid et al. (1990) found greatest support for educational programs and increased penalties for selling to minors and for minors caught drinking. Restricting ads on billboards, magazines, and newspapers had low overall support, as did eliminating sales of wine coolers and beer at gas stations and convenience stores. Extremely low support was found for eliminating happy hours or for increased alcohol sales taxes. While one must exercise caution in generalizing from state-level data, Schmid et al. asked a question having potential bearing on generic support for national as well as state policy: Who do you think is most responsible for today's alcohol-related health problems? Individuals, stores, manufacturers, or others? They noted that "the fact that 82% of the respondents indicated that the individual was responsible for alcohol-related problems and the fact that 78% felt the individual was the most important focus of intervention illustrate the dilemma of public health advocates-implementing policies most effective at the population level versus the individual level. Efforts to introduce policies . . . should recognize this dilemma and the need for additional public education and consensus building on the value of environmentally based intervention" (Schmid et al., 1990: p. 442).

A similar conclusion was reached by Wagenaar and Streff (1990), who argued that more research on attitudes toward alcohol and alcohol policy is needed to inform educational programs aimed at reducing social costs of alcohol use. However, education and consensus building around the need for and utility of environmentally based alcohol control policies faces a significant obstacle in the form of the cognitive dynamics underlying the prevention paradox: "Mostly people act for substantial and immediate rewards . . . [t]heir health next year is not likely to be much better if they accept [the advice of public health experts] or if they reject it. Much more powerful as motivators for health education are the social rewards of enhanced self-esteem and social approval" (Rose, 1985: p. 38). Public opinions on alcohol policy are inextricably tied up with such cognitions, which help explain responses to specific policy opinion items. Implicit or explicit policy objectives, personal salience of specific policies and problems, and the social and jurisdictional context in which a policy is intended to take effect are all plausible contributors to an individual's opinions (Giesbrecht & Greenfield, 1999; Room et al., 1995).

Typical questions/topics addressed in surveys

Table 1, which presents policy opinion item-wording drawn from selected research reports, highlights such dynamics. Considerable variation can be noted in the specificity of item framing. Variation in response categories can also be significant; Leedham notes that "provision of an 'indifferent' category makes a significant difference to the proportion expressing disapproval of the policy" (1987: p. 938), as respondents have a tendency to favor the status quo. Item examples in the drinking age and host/server intervention categories illustrate the potential for substantially differing policy alternatives to fall under the same rubric (for example, underage purchasing versus sales enforcement (Mosher, 1995), social versus commercial host liability). Table 1 illustrates a part of the range of items that can be adapted for use in future research on policy opinion.

A preliminary conceptual schema

Categorizing the range of control policy options in an effort to understand the dynamics of public opinion has been an issue in the analysis of national data sets. In analyzing the Canadian NADS data and U.S. warning-labels survey data series on policy opinions, Giesbrecht and Greenfield (1991, 1999) divided 11 policy items into a tripartite characterization of policy types (see Table 2). Highest support tended to be for those policies termed "interventions" (including increasing treatment, prevention, and server intervention programs), most of which were favored at above 80%. In addition, one policy from the message-oriented set termed "promotion controls/counter-promotions," beverage warning labels, was supported at around 90%. Other policies in this group tended to have modest majority support: TV alcohol advertising bans and government counter-advertising were supported by about 57%, while support for banning alcohol sponsorship of sports and cultural events was somewhat lower, at about 40%. "Access controls" are viewed with considerable specificity depending on the measure: decreasing hours of sale or raising the drinking age have low support (of course, the drinking age had already been raised in a number of states in the 1980s); in contrast, however, nearly two-thirds of Americans don't think alcohol should be available in corner stores. The last of the four items in this set, increased alcohol taxation, is supported by approximately half the population, although there appears to be more fluctuation over time in support for this measure than for others (this item is neutral or "unlinked" to desirable outcomes).

The pragmatic conceptual schema developed by Giesbrecht and Greenfield raises a number of questions. The range of policies included in this 11-item set is not exhaustive, and there is variation in policies within each category both in policy content and in level of support, suggesting complex underlying dimensions of opinion. The schema was not designed to assess the extent to which specific policies are perceived by respondents as personally intrusive and/or politically restrictive. The schema, originally developed in 1991 (Giesbrecht & Greenfield, 1991), was pragmatically adopted by Room et al. (1995). Other investigators, facing similar problems, have wondered whether there is an empirical structure to sets of policy opinions that might make them suitable for scaling. Schmid (1990) developed an additive scale measuring levels of support for control policies that had a moderately high Cronbach's alpha of .80, indicative of adequate internal reliability. This scale was used as the dependent variable for regression analyses predicting overall alcohol policy support from demographics and personal characteristics. In another attempt, Wagenaar & Streff (1990) used cluster analysis to try to organize policies into content groups. They found clusters for server liability, retail availability, and alcohol-impaired-driving countermeasure policies. An analysis undertaken by Kaskutas (1992) used Guttman scaling, Stouffer's H technique, and factor analysis to identify dimensions of policy opinions. Results were mixed, in part because of the limited number of policy indicators available (13). Factor analysis and Stouffer's H technique showed the most promise for future analyses of policy opinion items. Development of true multidimensional alcohol policy scales must await more complete analyses of larger sets of policy opinion, including development of attitude items better covering domains of alcohol policy, the dynamics of policy opinion expression, and better item response scaling. Broad-based studies that include public opinion items on alcohol policies, unless this topic is the sole aim, are unlikely to be able to afford to include a wide range of types of policies, multiple item sets, or item formats needed for a full-dimensional analysis. However, one recent study, promising in this regard, includes scale development as a methodological aim.

Single policy items on specific policies, such as those in Table 1, may be criticized as having limited and unknown reliability. Recently, using a large 1997 U.S. national telephone survey conducted by Wagenaar, Latimer et al. (2003) reported a factor analysis and scale development process based on numerous items designed to assess policy "issues." By including multiple items on a given alcohol policy, they formed scales including the following policy measures: regulate alcohol marketing, regulate alcohol consumption in public places, regulate alcohol distribution, regulate sellers and servers, regulate youth access, increase alcohol taxes, and penalize youth. Cronbach's alphas ranging from .59 (three items on taxes) to .88 (five items on marketing) were obtained (Latimer et al., 2001). It is noteworthy that alcohol taxation items included all identify link ... to pay for alcohol treatment, ... to lower other taxes, and ... to pay for any government purpose (a fourth item, "tax drinkers to pay for alcohol costs," was not included in the scale). Thus it is not surprising that alcohol taxes so measured have higher support than the non-linked (neutral) question in the NAS, WL and NADS surveys (see Table 1). The development of psychometrically sound "policy issue" scales is a noteworthy development, although the number of policy items required will limit use to specific kinds of study.

Trends in levels of policy support

Pertinent trend data on alcohol control policy opinion come from the warning-label data series, which spans the five years from 1989 to 1994 (Greenfield, 1997b; Greenfield et al., 1993, 1999). Using identical questions and methods during the five-year period, the researchers found considerable stability in levels of public support for alcohol control policies (see Table 2). Linear trends in four of the 11 policies show downward trends in support; only one policy, beverage warning labels, shows a modest increase in support, from 87% in 1989 to 92% in 1994. Most change with respect to warning labels occurred quickly, in the initial pre- to postimplementation period of the federal container warning mandate (Kaskutas, 1993a). In addition, support for increased alcohol taxes, while not showing a consistent change over time, has fluctuated significantly, declining, then rising again to a slight majority of support in 1993 (Greenfield & Kaskutas, 1994), only to descend to a low of 43% the following year (Greenfield et al., 1995). These fluctuations might be linked to the rise and decline in the public salience of and support for national health care reform in the early 1990s (Hacker, 1997). Finally, there was a consistent trend for all remaining policy opinions, including those that did not change significantly, to show declining levels of support of a few percentage points. Generally, the changes over time that were observed were modest in comparison both with the a differing preferences for various policies and with cross-country differences (e.g., U.S. vs. Canada) (Giesbrecht & Greenfield, 1999).

This can be seen graphically in Figure 1, which includes, in addition to results from the warning-labels series in Table 2, data from the most recent NAS survey also conducted by telephone in 2000. It should be kept in mind that the gap between 1994 and the final 2000 epoch is in fact longer than the range displayed from 1989 to 1994. The continued decline in support for the selected policies is evident. Excepting the warning-label policy, which despite all expectations prior to this federal law's adoption has proved increasingly popular, support for all policies lose ground. Nonetheless, the relative ordering of the policies in the public's view remain the same. Support for prevention and not serving to inebriated customers remains high, and price and availability measures considerably lower. The success story is the warning-label policy. Around the time of its adoption, 87.3% endorsed this policy; by 2000, national support is assessed at 93.9%, or 6.6% higherremarkable in that one might have anticipated a ceiling effect given such high levels of support at the outset. Clearly, substantial numbers of drinkers support the policy. Conversely, support for increasing alcohol taxes, reaching a high of 52.5% in 1993, has slid by 2000 to its lowest level of support at 36.1%, more than 15% lower. It is apparent that unprimed by an active "conceptual link," desirable rationales, or earmarks they can approve of, the public is not intrinsically inclined to see taxing alcohol as a particularly good measure-with the exception of a core of abstainers and very abstemious drinkers.

Correlates of public opinion

Correlates of policy opinions have also been investigated. Schmid et al. (1990) found that age, marital status, being a current smoker, and especially gender and alcohol consumption predicted overall support for alcohol control policies. Women, older persons, nonsmokers, and those who were married or never drank were most strongly and globally supportive of stricter control policies. In another study, Room et al. (1995) used logistic regression and policy-specific models to predict each of 11 surveyed policies in turn, comparing data from the U.S. and Ontario. Women were uniformly more supportive of control policies, including those aimed at conditions of sale, promotions/counterpromotions, and handling problematic behavior, prompting the authors to speculate that "women's tendency to favor restrictive state policies relates to their greater role in trying to control the drinking of relatives, and particularly of husbands and sons" (p. 45)-a finding replicating earlier research (Room et al., 1991). Compared with light drinkers, abstainers were more supportive of all policies, except those greater efforts to prevent drunken customers from being served. Conversely, frequent heavy drinkers were uniformly less disposed to these alcohol policies than other drinkers. Patterns of age showed variability across policies, but older individuals generally favored alcohol control policies more than did younger ones, except for mandated beverage warning labels and increased education and treatment programs. A number of the above conclusions have been replicated by the more elaborate measures in the 1997 University of Minnesota study. Latimer et al. (2001) found that women, infrequent drinkers, and those with greater knowledge or concern for youth showed greatest policy support across the five policy scales. Older adults supported restricting alcohol in public places, while younger adults interestingly tended to have greater support for rationalized tax measures.

Data on the impact of ethnicity on policy opinion are complex. Greenfield and Kaskutas (1993), in a five-year trend study, found minority groups less supportive of restricting service to intoxicated patrons and more supportive of other policies, especially of banning sales at corner stores. Discrimination experienced by members of ethnic minorities, such as having more often been denied service or living in low-income areas with high outlet densities, were hypothesized to underlie such differences. In a county-level survey on policy opinion, Jones-Webb and colleagues (1993) also found ethnic differences in levels of support on a minority of policy items. They note, however, that ethnicity effects are not consistently found in the literature. Mixed results in this regard, they suggest, may be a result of omission of critical consumption, social class, and cultural context variables from surveys and polls. Thus it is difficult to make a single characterization of correlates of policy opinions. Sheer variation in predictors of policy opinion indicates the need for improved measurement. Yet in the recent Minnesota national study using policy issue scales, Latimer and colleagues (2001) conclude that "the findings . . . [were] largely consistent with previous research that has not found ethnicity to predict alcohol policy attitudes" (p. 555). Although a number of other demographic variables were controlled in these analyses, the variables suggested by Jones-Webb and colleagues were not part of the regressions. It appears true that, substantively, the achievement of public health aims will depend upon better efforts to inform the public about the need for alcohol control regulations. Authors agree that education and consensus-building for alcohol policies should be done with sensitivity to the demographic and ethnic composition of targeted populations.

Need for theoretical models

Development of theoretical models of policy opinion dynamics has been limited because alcohol policy opinion data have been collected at a variety of jurisdictional levels and with a range of question items. In one analysis, Greenfield and Kaskutas (1993), stimulated by the variation in predictors among policies, hypothesized that attitudes would be influenced by congruent beliefs, relevant risk perceptions, and personal salience factors. They found overall support for this model. For example, beliefs that drinking can cause health problems, birth defects, and driving impairment are held more by those favoring a beverage label bearing such warning messages. These three hazard beliefs were also weakly related to some other control policies, especially increased alcohol taxes. The authors argued, following Room et al. (1995), that "the findings on association between certain policies and relevant beliefs about alcohol hazards suggest there is [public policy] value in reinforcing people's understandings of the known hazards of drinking" (Greenfield & Kaskutas, 1993: p. 20). In another analysis, Kaskutas (1993b) explored support for alcohol control policies as a function of perceptions about the effectiveness of beverage warning labels and taxation policies. Results for each policy showed that beliefs about policy efficacy are important: "Those who believed the label had affected their [own] drinking were six times more likely to support the label policy, and this belief was the strongest predictor of support for the policy. [In the case of] tax policy . . . the strongest predictors of support were lighter drinking style and belief that higher prices affect heavy drinkers" (p. 433).

This finding about the importance of effectiveness beliefs is consistent with evidence from a descriptive study of statelevel substance abuse policy opinion (Moloney & Emanuels, 1995) and theories of legislative decision-making (Arnold, 1990), as well as Hacker's ( 1997) analysis of the dynamics of public opinion on national health care reform. At the same time, Hilton and Kaskutas (1991) and Kaskutas (1993a) raised the possibility that other policies may be supported most strongly by those who feel their own behavior will be less affected by the policy. This suggestion stands in contrast to Hacker's analysis of public support for national health care reform, which grew in magnitude as more people came to feel they would be more affected by such a policy. The difference may owe to perceived personal benefits of policies. It is not unreasonable to speculate that the dynamics of public support for alcohol control policies differ significantly from those for related health policies because of ambivalence regarding alcohol and locus of control: "Control them, not me" (see, for example, Anglin et al., 2002). The "logic models" that respondents bring to bear on policy questions deserve careful analysis, both through more testing of cognitive models via survey research and through qualitative studies, including protocol analysis and the use of focus groups (Kaskutas, 2004).

Summary and conclusions

Based on widely differing levels of support for individual policy alternatives, the environmental alcohol policy approach supported by public health advocates and expert panels since the 1980s (Edwards et al., 1994; Moore & Gerstein, 1981; Mosher & Jernigan, 1989; U.S. Surgeon General, 1989; Babor et al., 2003) has mixed levels of public support. Overall slippage in support for environmentally (Room, 1999) and, to some extent, even individually (e.g., increasing treatment access) oriented policies suggests that the advice of Schmid et al. (1990) to redouble efforts in public education and consensus building around public health interventions remains apt to this day. As noted by Room et al. (1995), a short-term decline in support for measures that have clear potential for affecting rates of alcohol-related problems is significant: "Efforts to increase public support for policies designed to reduce rates of alcohol-related problems may, indeed, be a prerequisite to maintaining and strengthening the present structure of alcohol controls in North American provinces and states" (p. 46). Routine assessment to monitor and address these trends is vital. Even more to the point, this research should seek innovative ways to address questions about the relative influence of public opinion on alcohol policy development.

Continued policy opinion research may bear other fruit as well. On the broadest level, inclusion of political values and behavior (e.g., conservatism, voting, involvement in social movements), social-psychological variables (e.g., effectiveness beliefs, locus of responsibility), and sociocultural context variables (e.g., ethnicity, socioeconomic status) would expand dimensions of analysis and add explanatory scope to social epidemiological approaches. While this review has focused on control policy opinion, future research would benefit by including other critical health policy issues as well (e.g., alcohol treatment parity, support for national health care programs). Tracked over time, results of such research might illuminate the history and sociology of sentiments toward alcohol and the dynamics by which such sentiments translate into action in the form of social movements or policy enactment. Policy opinion research will likely remain mportant to local- and state-level public health advocates in designing and gaining support for promising alcohol policies. Standardization of policy opinion measures across jurisdictional levels and policy issues, with attention to the role played by item wording and format, would offer both researchers and community groups better tools for policy analysis. It needs to be recognized that since alcohol policies themselves-and policy contents-differ across countries and between country and local levels, one size may not fit all. Different questions tailored to specific issues will need to be developed, although the contributions of Latimer et al. (2001, 2003) and Wagenaar et al. (2000) are noteworthy in this regard. Finally, qualitative research on the policymaking process, on the nature and dynamics of community members' reactions to proposed and implemented policies, and on the social and political history of policy proposals in context should also inform public health efforts with respect to alcohol problems.

[Sidebar]

AUTHORS' NOTE: This research was funded by a grant from the Robert Wood Johnson Foundation to the Alcohol Research Group, Public Health Institute (Grant No. 030290). This work was presented in part at the annual meeting of the American Public Health Association, Washington, DC, November 15-19, 1998. Opinions represent those of the authors rather than the sponsoring institutions.

Note

* Similarly, in an earlier study of 800 Ontario residents, Goodstadt et al. (1978) found that "most drinkers (68%) expressed a willingness to pay more for alcoholic beverages if the long-term result would be a decrease in the prevalence of alcoholism" (p. 1632; italics added).

[Reference]

References

Ahlstrôm, S. & Ôsterberg, E. (1992). Changes in climate of opinion concerning alcohol policy in Finland in the 1980s. Contemporary Drug Problems, 19,431-457.

Anglin, L. (1995). The Ontario Experience of Alcohol and Tobacco: New Focus on Accessibility, Violence and Mandatory Treatment-a Report of the Ontario Alcohol and Other Drug Opinion Survey, 1995. Toronto, Ontario: Addiction Research Foundation.

Anglin, L., Kavanagh, L. & Giesbrecht, N. (2001). Alcohol-related policy measures in Ontario: Who supports what and to what degree? Canadian Journal of Public Health, 92(1), 24-28.

_____ (2002). Public opinion analysis suggesting demographic characteristics of person tending to favour internal versus external control of drinking behaviour. Journal of Substance Use, 1, 214-220.

Arnold, R. D. (1990). The Logic of Congressional Action. New Haven: Yale University Press.

Babor, T. F., Caetano, R., Casswell, S., Edwards, G., Giesbrecht, N. A., Graham, K., Grube, J., Gruenewald, P., Hill, L., Holder, H. D., Homel, R., Ôsterberg, E., Rehm, J., Room, R. and Rossow, I. (2003). Alcohol: No Ordinary Commodity. Research and Public Policy. New York: Oxford University Press.

Beauchamp, D. E. (1988). The Health of the Republic: Epidemics, Medicine, and Moralism as Challenges to Democracy. Philadelphia: Temple University Press.

Bondy, S. (1994). Attitudes and Experiences with Treatment of Alcohol and Tobacco Problems: A report of the Ontario Alcohol and Other Drug Opinion Survey, 1993. Toronto, Ontario: Addiction Research Foundation.

Caetano, R. & Clark, C. L. (1999). Trends in situation norms and attitudes toward drinking among whites, blacks, and Hispanics: 1984-1995. Drug and Alcohol Dependence, 54, 45-56.

Cahalan, D., Roizen, R. & Room, R. (1974). Findings of a Statewide California Survey on Attitudes Related to Control of Drinking Problems. Berkeley: Social Research Group.

Cameron, T. (1981). Traffic accidents and drunken driving: The problem as seen in public opinion polls. Abstracts and Reviews in Alcohol and Driving, 2, 3-5.

Casswell, S., Gilmore, L., Maguire, V. & Ransom, R. (1989). Changes in public support for alcohol policies following a community-based campaign. British Journal of Addiction, 84, 515-522.

Clark, W. B. (1985). Alcohol use in various settings. In E. Single & T. Storm, (eds.), Public Drinking and Public Policy (pp. 49-70). Toronto: Addiction Research Foundation of Ontario.

Crawford, A. (1987). Attitudes about alcoholism: A general review. Drug and Alcohol Dependence, 19, 279-311.

Edwards, G., Anderson, P., Babor, T. F., Casswell, S., Ferrence, R., Giesbrecht, N., Godfrey, C., Holder, H., Lemmens, P., Makela, K., Midanik, L. T., Norstrom, T., Ôsterberg, E., Romelsjo, A., Room, R., Simpura, J. & Skog, O.-J. (1994). Alcohol Policy and the Public Good. Oxford: Oxford University Press.

Eliany, M., Giesbrecht, N., Nelson, M., Wellman, B. & Wortley, S. (1992). Alcohol and drug use by Canadians: A national alcohol and other drugs survey (1989) technical report. Ottawa: Health and Welfare Canada.

Ferris, J., Templeton, L. & Wong, S. (1994). Alcohol, Tobacco and Marijuana: Use, norms, problems and policy attitudes among Ontario adults-a report of the Ontario Alcohol and Other Drug Opinion Survey, 1992. Toronto, Ontario: Addiction Research Foundation.

Fischer, P. J. (1989). Estimating the prevalence of alcohol, drug and mental health problems in the contemporary homeless population: A review of the literature. Contemporary Drug Problems, 16, 333-389.

Gallup, G. (1972). The Galhip Poll: Public Opinion 1935-1971. New York: Random House.

_____ (1985). Gallup Report. Princeton: The Gallup Poll.

Gallup Inc. (1987). Alcohol use and abuse in America. Gallup Report No. 265.

Gallup Poll News Service. (1989). Alcohol consumption at its lowest level in thirty years. Gallup Poll News Service.

Giesbrecht, N. & Greenfield. T. K. (1991). Public opinion on alcohol policy issues: A comparison of American and Canadian surveys. Paper presented at the Alcohol Policy section sessions, 36th International Institute on the Prevention and Treatment of Alcoholism, Stockholm, Sweden, 2-7 June 1991.

_____ (1999). Public opinions on alcohol policy issues: A comparison of American and Canadian surveys. Addiction, 94, 521-531.

Giesbrecht, N., Ialomiteanu, A., Room, R. & Anglin, L. (2001). Trends in public opinion on alcohol policy measures: Ontario 1989-1998. Journal of Studies on Alcohol, 62(2), 142-149.

Giesbrecht, N. A. and Greenfield, T. K. (2003). Preventing alcohol-related problems in the US through policy: Media campaigns, regulatory approaches and environmental interventions. Journal of Primary Prevention, 24, 63-104.

Giesbrecht, N. & Kavanagh, L. (1999). Public opinion and alcohol policy: Comparisons of two Canadian general population surveys. Drug and Alcohol Review, 18, 7-19.

Goodstadt, M., Smart, R. G. & Gillies, M. (1978). Public attitudes toward increasing the price of alcoholic beverages. Studies on Alcohol, 39, 1630-1632.

Gozenbach, W. J. (1996). The Media, the President, and Public Opinion: A Longitudinal Analysis of the Drug Issue, 1984-1991. Mahwah, NJ: Lawrence Erlbaum Associates.

Greenfield, T. K. (1993). Reasons for abstaining or limiting drinking: U.S. national trends between 1984 and 1990. Paper presented at the annual meeting of the Research Society on Alcoholism, San Antonio, Texas.

_____ (1994). The next agenda for prevention research: Studying alcohol policy formation. Skrifter fra Center for Rusmiddelforskning (Journal of the Center for Alcohol and Drug Research), 1, 35-55.

_____ (1997a). Alcoholism, alcohol problems, and social policies. In W. N. Shelton & R. B. Edwards (eds.), Values, Ethics, and Alcoholism. Advances in Bioethics (Vol. 3, pp. 221-256). Greenwich, Connecticut: JAI.

_____ (1997b). Warning labels: Evidence of harm-reduction from long-term American survey. In M. Plant, E. Single & T. Stockwell,(eds.), Alcohol: Minimizing the Harm (pp. 105-125). London: Free Association Books.

Greenfield, T. K., Giesbrecht, N. A., Johnson, S. P., Kaskutas, L. A., Anglin, L., Kavanagh, L., Room, R. & Mackenzie, B. (1999). U.S. Federal Alcohol Control Policy Development: A Manual. Berkeley, CA: Alcohol Research Group.

Greenfield, T. K., Giesbrecht, N.A., Kaskutas, L.A., Johnson, S. P., Kavanagh, L. & Anglin, L. (2004). A study of the alcohol policy development process in the United States: Theory, goals, and methods. Contemporary Drug Problems, 31, 591-626 (this issue).

Greenfield, T. K., Graves, K. L. & Kaskutas, L. A. (1993). Alcohol warning labels for prevention: National survey findings. Alcohol, Health & Research World, 17, 67-75.

Greenfield, T. K. & Kaskutas, L. A. (1993). Accounting for public opinion on alcohol control policies: General and individual predictors in surveys of U.S. adults. Paper presented at the annual meeting of the American Public Health Association, San Francisco, CA.

_____(1994). Changes in public opinion about alcohol control and intervention policies in the U.S.: A five-year trend study. Paper presented at the annual meeting of the American Public Health Association, Washington, DC.

Greenfield, T. K., Kaskutas, L. A. & Giesbrecht, N. (1995). Review of research on public opinion about alcohol policies. Berkeley: Alcohol Research Group.

Greenfield, T. K. & Klingemann, S. D. (1998). Reasons for abstaining or limiting drinking: U.S. national trends between 1984 and 1995. Paper presented at the annual meeting of the Research Society on Alcoholism, Hilton Head Island, SC, June 20-25, 1998.

Greenfield, T. K., Midanik, L. T. & Rogers, J. D. (2000a). A ten-year national trend study of alcohol consumption 1984-1995: Is the period of declining drinking over? American Journal of Public Health, 90, 47-52.

Greenfield, T. K., Rogers, J. D. & Trocki, K. (200Ob). Changes in contexts of drinking in the 1984 and 1995 U.S. national alcohol surveys: What light does context shed on drinking patterns? Paper presented at the 26th annual Alcohol Epidemiology Symposium, Oslo, June 5-9, 2000.

Greenfield, T. K. & Room, R. (1997). Situational norms for drinking and drunkenness: Trends in the U.S. adult population, 1979-1990. Addiction, 92, 33-47.

Hacker, J. S. (1997). The Road to Nowhere: The Genesis of President Clinton's Plan for Health security. Princeton, NJ: Princeton University Press.

Harris Associates (1972). Public Awareness of a NlAAA Advertising Campaign and Public Attitudes Toward Drinking and Alcohol Abuse: Phase One, Fall 1972. Rockville, MD: National Institute on Alcohol Abuse and Alcoholism.

_____ (1973a). Public Awareness of a NIAAA Advertising Campaign and Public Attitudes Toward Drinking and Alcohol Abuse: Phase Two, Spring 1973. Rockville, MD: National Institute on Alcohol Abuse and Alcoholism.

_____ (1973b). Public Awareness of a NIAAA Advertising Campaign and Public Attitudes Toward Drinking and Alcohol Abuse: Phase Three, Autumn 1973. Rockville, MD: National Institute on Alcohol Abuse and Alcoholism.

_____ (1974). Public Awareness of a NIAAA Advertising Campaign and Public Attitudes Toward Drinking and Alcohol Abuse: Phase Four, Winter 1974. Rockville, MD: National Institute on Alcohol Abuse and Alcoholism.

_____ (1983). People of the U.S.A. New York: Louis Harris and Associates.

Hilton, M. E. (199Ia). The presence of alcohol in four social situations: Survey results from 1964 and 1984. In W. B. Clark & M. E. Hilton (eds.), Alcohol in America. Albany. State University of New York Press.

_____ (199Ib). Regional diversity in U.S. drinking practices. In W. B. Clark & M. E. Hilton (eds.), Alcohol in America: Drinking Practices and Problems. Albany: State University of New York Press.

Hilton, M. E. & Kaskutas, L. A. (1991). Public support for warning labels on alcoholic beverage containers. British Journal of Addiction, 86, 1323-1333.

Holder, H. D. (1993). Changes in access to and availability of alcohol in the United States: Research and policy implications. Addiction. 88, 67S-74S.

Jones, R. J. & Greenfield, T. K. (1991). Alcohol and Other Drug Problems in Marin County, California: Results from a General Population Survey. San Rafaël: Marin Institute for the Prevention of Alcohol and Other Drug Problems.

Jones-Webb, R. J., Greenfield, T. K. & Graves, K. (1993). The relationship between ethnicity, social class, alcohol use, and public opinion regarding alcohol control policies. Contemporary Drug Problems, 20,719-738.

Kaskutas, L. A. (1992). Scaling Attitudes Toward Alcohol Control Policies: A Blending of Methods. Berkeley, CA: Alcohol Research Group.

_____ (1993a). Changes in public attitudes toward alcohol control policies since the warning label mandate of 1988. Journal of Public Policy and Marketing, 12, 30-37.

_____ (1993b). Differential perceptions of alcohol policy effectiveness. Journal of Public Health Policy, 14, 415-436.

_____ (2004). Division in the ranks: An analysis of alliances and dissension in the alcohol industry and advocacy communities. Contemporary Drug Problems, 31, 655-672 (this issue).

Kerr, W. C., Greenfield, T. K., Bond, J., Ye, Y. and Rehm, J. (2004). Age, period and cohort influences on beer, wine and spirits consumption trends in the US national surveys. Addiction, 99, 111 1-1120.

Kingdon, J. W. (1995). Agendas, Alternatives, and Public Policies. New York: HarperCollins.

Latimer, W. W., Harwood, E. M., Newcomb, M. D. and Wagenaar, A. C. (2001). Sociodemographic and individual predictors of alcohol policy attitudes: Results from a US probability sample. Alcoholism: Clinical & Experimental Research, 25, 549-556. _____ (2003). Measuring public opinion on alcohol policy: A factor analytic study of a US probability sample. Addictive Behaviors, 28, 301-313.

Leedham, W. (1987). Data note-10. Alcohol, tobacco and public opinion. British Journal of Addiction, 82, 935-940.

Lemmens, P. H., Vaeth, P. A. C. & Greenfield, T. K. (1999). Coverage of beverage alcohol issues in the print media in the US, 1985-1991. American Journal of Public Health, 89, 1555-1560.

McKenzie, D. & Giesbrecht, N. (1981). Changing perceptions of the consequences of alcohol consumption in Ontario, 1950-1981. Contemporary Drug Problems, 10, 215-242.

Midanik, L. T. & Clark, W. B. (1994). The demographic distribution of U.S. drinking patterns in 1990: Descriptions and trends from 1984. American Journal of Public Health, 84, 1218-1222.

Moloney, T. W. & Emanuels, K. (1995). Living Well: Californians Consider Public Policies That Foster Better Health. Sacramento, CA: California Center for Health Improvement.

Moore, M. & Gerstein, D., eds. (1981). Alcohol and Public Policy: Beyond the Shadow of Prohibition. Washington, DC: National Academy Press.

Mosher, J. F. (1995). The merchants not the customers: Resisting the alcohol and tobacco industries' strategy to blame young people for illegal alcohol and tobacco sales. Journal of Public Health Policy, 16,412-432.

Mosher, J. F. & Jernigan, D. F. (1989). New directions in alcohol policy. Annual Review of Public Health, 10, 245-280.

Page, B. I. (1994). Democratic responsiveness? Untangling the links between public opinion and policy. Political Science and Politics, 27, 25-29.

Paglia, A. (1995). Alcohol, Tobacco, and Drugs: Dependence, problems, and consequences of use: A report of the J 994 Ontario Alcohol and Other Drug Opinion Survey. Toronto, Ontario: Addiction Research Foundation.

Room, R. (1987). Alcohol monopolies in the U.S.: Challenges and opportunities. Journal of Public Health Policy, 8, 509-530.

_____ (1991). Cultural changes in drinking and trends in alcohol problems indicators: Recent U.S. experience. In W. B. Clark & M. E. Hilton (eds.), Alcohol in America: Drinking Practices and Problems (pp. 149-162). Albany, NY: State University of New York Press.

_____ (1999). The idea of alcohol policy. Nordic Studies on Alcohol and Drugs, 16 (English Supplement), 7-20.

Room, R., Graves, K., Giesbrecht, N. & Greenfield, T. K. (1995). Trends in public opinion about alcohol policy initiatives in Ontario and the U.S.: 1989-91. Drug and Alcohol Review,14, 35-47.

Room, R., Greenfield, T. & Weisner, C. (1991). People who might have liked you to drink less: Changing responses to drinking by U.S. family members and friends, 1979-1990. Contemporary Drug Problems, 18,573-595.

Rose, G. (1985). Sick individuals and sick populations. International Journal of Epidemiology, 14, 31-38.

Schmid, T. L., Jeffery, R. W., Forster, J. L., Rooney, B., Klepp, K.-I. & McBride, C. (1990). Public support for policy initiatives regulating alcohol use in Minnesota: A multi-community survey. Journal of Studies on Alcohol, 51, 438-442.

Stimson, J. A., MacKuen, M. B. & Erikson, R. S. (1994). Opinion and policy: A global view. Journal of Political Science & Politics, 21, 29-35.

U.S. Surgeon General (1989). Surgeon General's workshop on drunk driving: Proceedings. Rockville, MD: U.S. Department of Health and Human Services, Public Health Service.

Wagenaar, A. C., Harwood, H. J., Toomey, T. L., Denk, C. E. and Zander, K. M. (2000). Public opinion on alcohol policies in the United States: Results from a national survey. Journal of Public Health Policy, 21, 303-327.

Wagenaar, A., Molnar, L., Streff, F. & Schultz, R. (1988). Michigan Omnibus State Safety Survey. Ann Arbor, MI: University of Michigan Transportation Research Institute.

Wagenaar, A. & Streff, F. M. (1990). Public opinion on alcohol policies. Journal of Public Health Policy, 11, 189-205.

Yankelovich, D. (1991). Coming to Public Judgment: Making Democracy Work in a Complex World. Syracuse, NY: Syracuse University Press.