Australian Drug and Alcohol Review 1988; 7:273-284

The availability of alcohol : prior research and future directions

Eric Single Addiction Research Foundation, Toronto, Canada, and Department of

Preventive Medicine, University of Toronto

Abstract: The relationship between the availability of alcohol and the amount of alcohol consumed is described. An historical overview is presented, followed by a review of the empirical evidence of the relationship. Neglected aspects of availability, including the impact of the physical environment within alcohol outlets, the behaviour of aleohol servers, and the structure of the alcohol industry, are considered.

Keywords: alcohol drinking; alcoholism - prevention and control.

Introduction The subject of my presentation is the relation-

ship between availability and the consumption of alcohol. I should state from the outset that my remarks will be limited in two major respects. Firstly, the subject of pricing and taxation will be taken up in another part of these meetings, so I will use the term availability to refer solely to physical access to alcoholic beverages. Secondly, much of the literature concerning the availability of alcohol focuses on its impact on alcohol-related problems. I will focus instead simply on the impact of availability on consumption, which of course represents a vital, but not the only, connection between availability and alcohol-re- lated problems. Thus, I will not attempt to deal with the important issue of the relationship between consumption levels and alcohol-related problems, except insofar as it relates to the policy implications of the impact of availability on consumption.

Furthermore, I should note that I will em- phasize what I consider to be relatively neglected aspects of the connection between availability and consumption, which may have important implica- tions regarding future directions for research and intervention programming. I could have instead focused exclusively on the empirical evidence concerning the relationship between physical ac- cess and consumption, as there are numerous studies which could be subject to review. Al- though I intend to review this rich literature briefly, it strikes me as unnecessary to consider it in great detail as several excellent recent reviews of its exist.

Therefore I will first present an historical overview of the concern with the connection

between availability and consumption, followed by a brief review of the empirical evidence concerning this relationship. I will then discuss three aspects of availability which have been relatively neglected by alcohol policy research in the past yet represent potentially important new directions for research in the future. I refer specifically to the impact of the physical environ- ment within alcohol outlets, the impact of the behaviour of persons who serve alcoholic be- verages and the structure of the alcohol industry as it relates to availability and consumption patterns.

The impact of availability on consumption: an historical overview

Over the past two decades there has been increasing interest in the use of controls over the availability of alcohol as a means to prevent alcoholism and other alcohol-related health and social problems. In the recent past, a number of international, ~ ' national' and regional bodies 8 have advocated restrictions on alcohol availability as a preventive measure. Interest in the regulation of alcohol availability is hardly a new or recent phenomenon. Many authors have documented historical and even ancient proscriptions regarding the distribution of alcohol, 9 dating as far back as the Code of Hamurabi circa 2000 B.C. 'a

The renewed attention placed on alcohol con- trol measures in the post-World War II era is due to a number of factors. First, there was a marked increase in alcohol consumption and in various indices of alcohol-related problems throughout the industrialized world up to the mid-1970s." At the same time, treatment systems appeared to be limited in effectiveness ~2 despite their high cost,

274

thus providing further impetus to the search for preventive measures.

Another significant influence has been the declining popularity of the ~disease concept" of alcoholism among public health professionals. In the post-war era, controls over the availability of alcohol were generally viewed as unpopular ana- chronisms dating from the earlier era, strongly associated with the failure of Prohibition. *' Moral- istic conceptions of drinking which had been the basis of Prohibition, were largely replaced by the ~disease concept" of alcoholism. According to this viewpoint, normal social drinkers are fun- damentally different from alcoholics, who are seen as suffering from a special disease. As noted in the report of the WHO-affiliated International Study of Alcohol Control Experiences:"

... the locus of alcohol problems tended to be redefined from 'the bottle' to 'the man': if alcohol problems were a matter of specific defective individuals, then there was no need to control the drinking of the majority who were not defective. If, indeed, alcohol problems were specific to these individuals, an ap- propriate means of managing alcohol problems would be to provide treatment for them, and to try to persuade them - - but not others - - to abstain. In this way, the expansion of treat- ment system may be seen as a kind of cultural alibi for the normalization of drinking and relaxation of controls. (pp. 64-65)

Because the unique predisposing factors which cause certain social drinkers to become alcoholics remained unknown, the conceptualization of al- coholism as a disease led to an emphasis on treatment rather than prevention. After all, if the cause of a disease is unknown, how can it be prevented?

More recently, the disease concept of alcohol- ism has been called to question for various reasons, including epidemiological research which shows that the risk of various adverse health effects increases significantly even for those who consume well below levels associated with alco- holism. 's For example, Pequinot and his colleagues '6 found that even a relatively moderate intake of alcohol (between 60 and 80 grams of ethanol per day) carries twice the risk of death from liver cirrhosis as the next lowest level of consumption (between 40 and 60 grams). 2 To put these numbers into perspective, 60 grams of ethanol is equal to 4.4 %tandard drinks", a standard drink being 12 oz. of 5% beer, or 5 oz. of 12% table wine or a 1.5 oz. measure of 40%

spirits. This level of consumption is far below that of clinical alcoholics." Thus, even in moderation, drinking is associated with the development of adverse consequences.

With the increases in alcohol-related problems in the 1960s and a declining enthusiasm for the disease concept of alcoholism on the part of health professionals, interest began to focus on the potential of alcohol control measures. As Room has explicated, '7 three hypotheses emerged con- cerning the general effects of controls over availability. The first is that controls over availability have no effects on alcohol problems; the second is that they have perverse effects; and the third is that they have positive effects.

The first argument, claiming no impact, was based mainly on a priori assumptions that alcohol- related problems stem mainly from alcoholics who are unaffected by controls over access. Empirical evidence severely challenges both of these assumptions. First, it cannot be assumed that all, or in some instances even most, of alcohol problems can be attributed to alcoholics. Second, as we shall discuss shortly, alcoholics are, if anything, even more affected by controls over availability than social drinkers.

The second perspective, claiming that controls over availability have negative effects and actually exacerbate alcohol problems, became known as the %ociocultural" perspective." It was argued that alcohol problems stem in large part from the lack of adequate norms governing its use, i.e. rules which would enable drinkers to distinguish

roperly between moderate and problematic drin- ng. It was therefore argued that preventive

efforts should focus on providing normative rules regarding appropriate drinking behaviour. Thus, according to this perspective, parents should introduce drinking to teenagers at home and restrictions on access to alcohol should be min- imized to avoid the ~forbidden fruit ~ attraction of alcohol.

The sociocultural perspective has been dis- credited on both theoretical and empirical grounds. Room 2° noted that cultural ambivalence is not a property of the individual's mind, but rather a collective variable; therefore, alcohol analyses should focus on conflicting norms and values in society rather than treat individual ambivalence as an explanatory variable. Makela 2. critically evaluated the empirical evidence for the sociocuhural perspective and found no support. Variations in rates of alcoholism for cultural subgroups could largely be explained by varia- tions in overall rates of drinking, and efforts to

275

substitute moderate drinking practices for un- desirable drinking patterns have consistently failed without restrictions over access?

The third line of argument is that controls over availability are negatively related to rates of alcohol problems and can therefore be used as a mechanism for the prevention of alcohol-related problems. The position is supported by con- siderable empirical evidence concerning the im- pact of physical availability on consumption levels and alcohol-related problems, to be summarized shortly. Additional support for this position derives from econometric studies on alcohol consumption. Contrary to what some persons might have expected given the addictive properties of alcohol and given the high proportion of consumption accounted for by heavy drinkers, alcoholic beverages have been found to be gen- erally responsive to incremental changes in price. 2~

The responsiveness of alcohol consumption to controls over availability and price manipulation, combined with the finding that mean levels of alcohol consumption tend to be closely related to various indices of alcohol-related problems, poin- ted to the possible use of controls over availability as a tool for the prevention of alcohol problems. In 1975 an international group of alcohol resear- chers critically reviewed the evidence regarding the relationship between legal controls on alcohol availability, mean levels of consumption and indices of alcohol-related problems." To quote from their report:

... our main argument is well substantiated: changes in the overall consumption of alco- holic beverages have a bearing on the health of the people in any society. Alcohol control measures can be used to limit consumption: thus, control of alcohol availability becomes a public health issue (pp. 12-13)

In the decade that has passed since it first appeared, this influential report has been a central focus of a continuing policy debate. The notion that controls over availability could be used to prevent alcoholism still faces considerable opposi- tion from many organizations and individuals involved with the treatment of alcoholism. As Augustus Hewlett, former President of the Alco- holism and Drugs Programs Association of North America, has stated:

Regardless of our good intentions, commun- ications of support for the wide range of stricter controls ... over availability will be perceived by the public as a denial of the disease concept of alcoholism (quoted in Room, 17 p. 306).

The controversy surrounding the %ingle distribution" or ~Ledermann" approach

An historical overview of the issues concerning alcohol availability would not be complete with- out mentioning the controversial work of the French social scientist Sully Ledermann. 2' Indeed, due to his influence, the position that controls over alcohol availability can be used to prevent alcohol problems has frequently been incorrectly identified as the ~Ledermann" or %ingle distribu- tion" or ~distribution of consumption" theory. The controversial nature of contention that con- trois over availability can be used to prevent alcohol problems is reflected by the variety of terms used to describe it. The 1975 international group" referred to it as the ~public health approach". This term incorrectly implies a false consensus among public health professions. At least one critic provocatively labelled it as ~neoProhibitionist "2~ a term which is certainly inaccurate. The position neither advocates that alcohol be prohibited nor is it based on the same reasoning as that of Temperance interests which spawned Prohibition in many countries. Al- though Temperance groups and public health interests may both advocate greater restraints on alcohol availability, the former is based on moral- istic conceptions regarding drinking while the latter is based on epidemiological evidence regard- ing the consequences of alcohol use. Perhaps the best way to describe the position favouring controls over access to alcohol is to simply label it as the %vailability" hypothesis.

Therefore, before considering the empirical evidence concerning the relationship between alcohol availability and consumption, I will brie- fly review this controversy in order to place the'Ledermann" or %ingle distribution" theory into proper perspective. It will be argued that although the ~Ledermann" approach helped to bring prevention issues to the forefront of policy debates, it is fundamentally flawed. Perhaps most importantly, the ~Ledermann" or %ingle dis- tribution" approach is unnecessary to the ar- gument that controls over availability are an effective mechanism for the prevention of alcohol- related problems.

There is little dispute that the mean level of alcohol consumption in a society is closely related to the number of persons consuming at levels deemed to be of high risk. There has never been documented a society with low mean consump- tion which did not also have a relatively low number of ~heavy" drinkers, and by the same

276

token, societies with high levels of mean consum- ption tend to have high numbers of persons deemed to be ~heavy" drinkers. Ledermann argued that a single function, a lognormal curve, provides an accurate estimate for the distribution of alcohol consumption in any society.

Given the addictive properties of alcohol, some might have expected alcohol distribution to be bimodal, with one peak level among moderate or %ocial" drinkers and a second peak level among those who are physically dependent on alcohol. Contrary to this expectation, Ledermann found a unimodal distribution which is continuous and highly skewed.

Perhaps the most controversial aspect of Leder- mann's work is the contention that the dispersion of this distribution is relatively invariate and can be estimated from the mean. This implies that once the mean consumption among drinkers is known, one can predict with reasonable accuracy the number of persons consuming at any level, including the number who might be deemed to be ~high risk" or %lcoholic'. Bruun and his colleagues" critically examined and extended Ledermann's analysis of the distribution of alco- hol consumption and their work largely suppor- ted these conclusions.

The contention that the Ledermann lognormal distribution represents an accurate description of alcohol consumption in any society has been challenged by many authors. 2 .... Much of the criticism concerns assumptions underlying the development of the Ledermann distribution func- tion, such as the assumption that the upper limit of alcohol consumption per person is 365 litres of pure ethanol per year. More importantly, Skog" has demonstrated several significant deviations from the Ledermann model.

One of the most perplexing issues with the Ledermann distribution is the assumption of homogeneity, whereby the model is only meant to apply to populations in which there are no major social groups which deviate substantially from the overall mean level of consumption. This assump- tion would seem to eliminate the model's ap- plicability to any known society because all societies consist of an approximately equal num- ber of males and females, and male drinking levels are invariably a good deal higher than those of females. For example, survey data indicate that Canadian males drink at least 50% more per capita than females (Canada Health Survey, 1983). The Ledermann distribution would yield one estimate of the number of persons consuming 15 cl. or more per day, given the total population of

Canada, but a very different and considerably higher estimate if the Ledermann formula were applied separately for males and females. In Ledermann's original formulation, a second parameter was added to the distribution to adjust for heterogeneous populations, but this very important aspect of the distribution function was later ignored or omitted in Ledermann's sub- sequent work and in the work of others (the one exception being Skog). ~° Since no society meets the assumption of homogeneity, it would have been preferable if the single parameter Ledermann function had been ignored instead.

The controversy surrounding the Ledermann hyypothesis continues. Skog, in particular has carried on this line of work despite the fact that he views the Ledermann theory as ~untenable" in its original form. '~ By modifying Ledermann's work to include a number of factors which combine in a multiplicative rather than an additive fashion, as well as considering the drinker's social network, Skog has transformed a theoretically crude em- pirical generality into a potentially viable theory concerning the etiology of alcoholism.

The purpose of this discussion is not to summarize the Ledermann controversy fully, as this has been done elsewhere. '°' ~ Rather, the intent is to point out that the policy debate surrounding the feasibility of controlling alcohol availability as a means of preventing alcohol-related problems has focused on a relatively narrow aspect of the theoretical connection between availability and problems which is not at all a necessary, or integral, part of the argument favouring controls. The Ledermann distribution is only one of several possible explanations for the observation that mean levels of consumption in a society are related to the number of heavy drinkers, and this relationship is by no means the only way in which alcohol availability and alcohol-related problems are theoretically connected? 2

The empirical evidence regarding the relationship between access to alcohol and consumption rates

To this point, it has been argued that contrary to what some authors have asserted, *~ the position favouring controls over alcohol availability does not derive from nor depend on any particular hypothesis concerning the distribution of alcohol consumption. The cornerstone of this position is the empirical relationship of control measures to alcohol consumption and to alcohol-related problems. A brief summary of research on these issues is presented below. More thorough reviews appear elsewhere. ''''7'33k~

277

Type of control systems There is enormous variability in the general

approach to alcohol control in different societies. Prohibition is not an historical anachronism; it is common among Moslem countries today and there are even regions within Western countries such as Canada where the sale of alcohol is proscribed by law. On the other hand, many countries have an essentially laissez-faire ap- proach. This is typical of the wine-producing countries of the Mediterranean. In Nordic coun- tries such as Finland, both production and dis- tribution are in the hands of a state monopoly. In Canada and several states in the U.S. the distribu- tion of alcohol is regulated by a state monopoly but production is privately owned. In other states in the U.S. and most Western European coun- tries, both production and distribution are in the hands of privately owned business under licence by the government.

The evidence concerning the impact of the nature of the control system is mixed. On one hand, there is little doubt that regardless of other consequences, total prohibition does succeed in reducing alcohol consumption and alcohol-related problems. Those countries which enacted pro- hibition laws in the early part of this century, and areas subjected to de facto prohibition due to war, experienced substantial reductions in health and social problems associated with alcohol use."

The evidence regarding the impact of less dramatic differences in alcohol control systems is not conclusive. Monopoly systems do not neces- sarily have lower rates of alcohol consumption than private systems. By the same token there are laissez-faire systems with relatively low consump- tion, the Netherlands being a recent example?' In the U.S., there is little difference between those states which have a monopoly system and those with a private licence system in terms of alcohol sales or rates of liver cirrhosis? ~

Restrictions on distribution: hours of operation, days of sale and age limitations

Comparisons between societies with different types of control regimes are necessarily incon- clusive, regardless of the findings, due to the confounding influence of alcohol culture. Changes in the nature and extent of controls over time within particular societies indicate that the liberalization of controls which has occurred in the post-World War II era throughout the indus- trialized world has been accompanied by an increase in alcohol consumption and increases in indicators of alcohol-related problems. ~' As noted

earlier, it is not always clear whether increased consumption is an effect, a cause or a concomitant of liberalization of the alcohol laws. It is also not clear whether the apparent positive relationship between consumption and controls would hold true if greater restrictions were put in place.

Nonetheless, increased access to alcohol has been generally found to be related to increased consumption. When local prohibition was ended in selected towns in rural Finland, there was a dramatic increase in alcohol consumption and alcohol-related problems. '8 Controls on hours of operation of alcohol outlets have been found in a number of settings to be related to consumption patterns and alcohol problems. Popham ~9 found arrests for public drunkenness to be correlated with hours of tavern operation in Toronto. I hardly need remind this audience that here in Australia, tavern hours were changed in an attempt to reduce the problems associated with a 6 p.m. closing time. Workers had been drinking large quantities so rapidly during the brief time after work when drink was available that the time immediately prior to closing was known as the ~six o'clock swill". When the closing time was extended to 10 p.m., there was a corresponding change in the temporal pattern of traffic accidents, but overall consumption was not clearly affected. ~

The findings regarding the impact of age restrictions are unequivocal. The lower the drink- ing age, the lower the age at which adolescents first use alcohol, the higher the consumption of alcohol and the higher the indices of alcohol- related problems among teenagers. ~' A case in point was the situation in Ontario, where the drinking age was lowered from 21 to 18 in 1971, but subsequently raised to 19 in 1978 as a result of increased drinking problems among teenagers, particularly a sharp rise in teenage drinking and driving accidents. 42

Density of outlets There have been a number of recent reviews of

the research literature concerning the relationship of the number of outlets to alcohol consumption and to alcohol problem indicators, ~'''43 all of which arrive at the same basic conclusion. As stated by Macdonald and Whitehead (1983): 43

The weight of evidence, especially when one takes into account the quality of the studies, is on the side of the availability of outlets accounting for some of the variance in the extent of alcohol consumption. The availability of outlets for off-premise consumption appears to be more sensitive than the availability of on- premise consumption, but the impact of the

278

latter cannot be treated as trivial without further study (p. 482).

In an excellent review of this literature Ian Smith" summarized the research with the follow- ing points:

• While giving little or no attention to possible confounding variables a number of 'early' studies gave insignificant results.

• The introduction of on-premise drinking, when off-premise drinking was already permitted, did not result in any significant effects. A similar comment is also applied to the liberalization of on-premise drinking.

• The introduction of extension of liquor licences for off-premises sales by grocery stores or supermarkets possibly results in increased con- sumption of wine and more convictions for drunkenness.

• A number of North American studies have provided evidence of a positive relationship between the number of liquor stores and con- sumption of spirits, wine and alcohol-related problems. However, the above findings cannot be generalized to packaged beer as liquor stores in North America sell primarily wine and spirits.

• The number of licences for on-premise drinking appear to be positively related to a variety of indices for the use and abuse of alcohol. However, on the basis of only two studies the relationship should be regarded as tentative.

• Factor analytic studies suggest that on-premise availability and off-premise availability bear little relationship to each other. However, as can be seen above, each appears to be positively related to various indices for the use and abuse of alcohol. This may explain why the results of studies in which variables for the total number outlets have been used have also been positive.

• Virtually all the studies were conducted in North America.

• The problem of determining the direction of the apparent causation was commented on by a number of researchers. Does increased availability cause the increased use and abuse of alcohol, or does the demand for more alcohol lead to greater availability? (p. 88)

The available evidence from Ontario is limited to correlational analyses, based on county data, but the results serve to illustrate these findings. Rush and his colleagues" found both on-premise and off-premise density of outlets to be related to consumption and problem indicators. Unfortun- ately, correlations convey relatively little intuitive understanding. Such statistics are useful for deter- mining whether an independent variable makes a difference, but they convey little sense of how much of a difference it makes. Therefore, the raw data presented in Rush et al." have been re- analyzed in simple tabular form and are presented in Table 1.

Table 1: Alcohol consumption and selected problem statistics by urbanization and outlet density

Urbanization: % living in towns

Rural: ~50% Mixed: 50%-75%

above below above below mean mean mean mean £ of £ of £ of £ of

outlets outlets outlets outlets

of >2000 Urban: >75%

above below mean mean £ of £ of

outlets outlets

Consumption' 15.15 10.26 13.31 10.70 13.87 11.28 Cirrhosis mortality 2 .22 .15 .22 .15 .20 .16 Traffic mortality 3 2.04 2.01 1.94 1.92 2.13 1.90 Cirrhosis morbidity 4 .79 .73 1.07 .75 .98 .74 (Number of counties 5) (9) (7) (4) (15) (2) (12)

Source: reconstructed from data presented in Rush et al." per capita consumption in litres of absolute alcohol

2 rate pe 1000 adult deaths rate per 1000 adult deaths

4 rate per 1000 hospital separations s The consumption, morbidity and mortality rates are county averages, not weighted by population size. They do

not represent individually based data. Thus, for example, the first row presents the mean county rates of alcohol consumption for each type of county, not the average individual consumption in each type of county.

279

It should be noted that Table 1 does not simply present the selected consumption and problem statistics by outlet density. This is because the term ~outlet density" can be misleading. Many persons would think that urban areas such as Toronto have a high density of outlets, and that rural areas, such as Northern Ontario, have a low density. If outlet density is operationalized as the number of outlets per unit of population, however, one finds just the opposite. Metro Toronto in fact had the lowest density of outlets per unit of population (.52 per 1000 adults), while remote sections of the province tended to have the highest density of outlets. One could instead attempt to compute a measure of outlet density based on the number of outlets per unit of territory in a county, but this would also be less than adequate because many counties contain enormous tracts of uninhabited land and would be artificially underscored on outlet density. Ideally, one would prefer a measure which considers the mean distance between residences and the nearest outlets. Unfortunately, this is impractical.

In view of these difficulties, Table 1 presents a three-variable table, showing differences in alco- hol consumption and alcohol problems between counties with above average number of outlets per unit of population and below average number of outlets per unit of population, with the degree of urbanization controlled for. Thus, from the first two columns of the table, it can be seen that among rural counties (those with less than half of the population living in communities of 2,000 or more), consumption rates are much higher and problem rates are somewhat higher in those counties which have a greater than average density of outlets. The same findings hold true for mixed and urban counties. Therefore, these Ontario data corroborate the general pattern found in other jurisdictions - - the number of outlets is positively related to rates of alcohol consumption and alcohol-related problems.

Differential impact on ~social ~ versus ~heavy ~ drinkers

All of the forementioned research on the impact of controls over availability would provide only weak support for the use of alcohol controls as a preventive policy if it were found that these controls only affected moderate or infrequent drinkers and had relatively little effect on high volume or ~heavy" drinkers. However, there has gradually accumulated more and more evidence that controls over alcohol availability have an even greater impact on high risk or ~heavy ~ drinkers than on moderate drinkers.

Room" has summarized this evidence, which stems mainly from econometric studies on the impact of price changes and from a rich literature on the impact of strikes by alcohol workers. As Room notes:

Though the strikes often attracted extraordin- ary popular attention, and the population as a whole often substantially shifted its choices of beverage and drinking locale, the amount of consumption of moderate drinkers, par- ticularly those of middle class, was often little affected. On the other hand, the strikes often had dramatic effects in reducing alcohol-re- lated problem rates associated with poor habitual heavy drinkers. '7

Room concludes that it is no longer tenable to claim that there are no studies which link de- creased availability with decreased consumption by heavy consumers of alcohol.

Neglected areas of research on availability and potential future directions

Thus far research on alcohol availability has concentrated on demonstrating the straightfor- ward empirical relationship between the existence or non-existence of controls and consumption or problem rates. As the weight of evidence grows that controls do impact on alcohol consumption, it would be expected that future research will focus on other aspects of availability as they affect alcohol consumption and problems. I would expect that greater attention will be paid to the following aspects of availability: the impact of situational factors within alcohol outlets, par- ticularly the physical environment; the impact of serving staff and training programmes aimed at promoting responsible beverage service; and the structure of the alcohol industry as it affects the availability and consumption of alcohol.

The impact of situational factors within alcohol outlets

Most of the studies on alcohol controls cited above consider availability in terms of number and types of outlets, age restrictions, and/or the existence or non-existence of particular control measures, with little or no consideration of potentially important differences within alcohol outlets. However, there is a small but growing literature based on tavern studies which has identified five major sets of situational determin- ants of heavy or problem-related drinking: tem- poral variations, characteristics of the drinking group, the length of time spent on one drinking occasion, the characteristics of the physical en- vironment and price differentials. 's

280

Perhaps the most obvious correlates of heavy drinking are temporal variables. There is almost universally greater drinking during the evening hours and weekends than on weekday mornings and afternoons, and on special holiday or festival events. Alcohol is often more available at such times. Indeed, the temporal relationships are so commonly assumed that researchers rarely measure or report temporal variations in drinking rates.

A situational factor which has received a great deal of attention from researchers has been the influence of the drinking group. A number of studies have shown that the drinking rate of drinking confederates is a strong influence on consumption and that the influence of low-rate drinking companions is less than that of high-rate drinking companions. '~-'9 It would appear that the drinking norms described by Kettil Bruun in 1959 among subjects in a small group drinking ex- periment continue to have wide applicability. Bruun observed that it was permissable for a group member to drink more than other members of the group but not less. ~

A second way in which drinking in groups may influence patrons to drink more is via its effect on duration of stay. In a study of Edmonton beer parlors, Sommer 5~ found isolated drinkers con- sume less than persons drinking in groups, but he attributed this to the shorter duration of stay in the drinking establishments. The finding that group drinkers tend to consume more than isolated drinkers has been replicated in Canada 52," and in the U.S. s'

In addition to the size of the drinking group, a further influence on length of stay is the physical environment. By providing television, games, and other activities, the tavern can increase the dura- tion of patron visits and thus increase volume of consumption. Even the tempo of music can influence drinking. Bach and Schaefer ss claim that the slower the tempo of country and western music played in bars in Western Montana in the U.S., the faster the patrons consumed their drinks. Ratcliffe and his colleagues ~ found that in Alberta beverage rooms where dancing and games were provided, the patrons who participated in these activities stayed longer and drank more than non-participants. Sommer ~ contrasts the Edmon- ton beer parlors with English pubs and attributes the short duration of stay by isolated drinkers in the former situation to the lack of anything to do other than sit and watch other people drink.

Apart from its impact on duration of stay, the physical environment can influence drinking be-

haviour in other ways. Cloyd ~' details how the layout of "market-place" bars is designed to maximize patrons' discretionary powers to initiate and control encounters with other patrons. Cavan ~8 describes bars as "open regions" where patrons have the right to initiate conversation with others and an obligation to accept overtures from others. As Radcliffe and his colleagues point out, 83 the degree of such "openness" is manipulable: the closed nature of lounges tends to foster a genteel atmosphere and greater privacy, as compared to the more open nature of taverns and pubs. Contrary to expectations, however, Radcliffe et al. found no support for the assumption that the lounge environment fosters moderate drinking. Thus, the physical layout of a drinking establish- ment affects the nature of the drinking group, the atmosphere within the setting and the length of drinking occasions, but the net impact on con- sumption levels is not what one might have expected.

A further situational determinant of heavy drinking which has received surprisingly little attention in tavern studies is the price of alcoholic beverages. There is, however, at least one study on the impact of price differentials within alcohol outlets as they affect consumption. Babor and his colleagues" found that "happy hour" price reduc- tions significantly increased alcohol consumption by both moderate and heavy drinkers.

The lack of attention paid to price in tavern studies is perhaps not surprising given the essen- tially social nature of public drinking. After all, if the purpose of going to taverns and pubs was simply to drink, one would find very few people patronizing them because it is generally much less expensive to drink at home. Nonetheless, price and other situational features of the drinking environment within outlets may well prove to be important influences on drinking patterns, despite the fact that they have been relatively neglected in the past.

The civil liability of alcohol providers Less than fifteen minutes after finishing his last beer and leaving the Arlington House Hotel, Clayton Sharpe failed to negotiate a curve in the road. His car plunged over a steep embank- ment, rolled over repeatedly, and came to rest 124 meters away. Sharpe suffered only minor injuries, but his passenger, 16 year old Andreas Schmidt, was critically injured. Although he survived, Schmidt was rendered a quadriplegic.

Schmidt and his family sued not only Sharpe, but also the Arlington House Hotel for

281

$13,000,000. After a lengthy trial, judgement was awarded against both defendants for over $1,390,000 ...

As discouraging as it may to Canadian tavern owners and their insurers, the Schmidt case cannot be written off as a legal aberration. The clear trend in the law is toward the expansion of liability for all those who sell or supply alcohol to others. ~°

The Arlington House Hotel was sued in the Schmidt case as a provider of alcohol for breach- ing an obligation to control the conduct of its intoxicated patrons. There is a distinct trend in Canada, the U.S., Australia and other countries, toward the expansion of civil liability of providers of alcohol. This liability may be based on statute, as is the case in three Canadian provinces and those U.S. states with "dram shop" laws, it can be based on common law, or it may be based on both. In addition to their liability as providers of alcohol, tavern owners and social hosts are further liable for the conduct of intoxicated guests as occupiers under statutory occupier liability acts in virtually all jurisdictions.

The factors responsible for the recent emer- gence and expansion of civil liability of the intoxicated are: an expansion of the duties of affirmative action to control the conduct of others in a wide variety of circumstances; the develop- ment of mandatory breath and blood testing, making it easier to establish intoxication as a cause of injury in court cases; the narrowing of traditional defenses such as the voluntary assump- tion of risk such that now courts will not accept a claim that a drunk driver or his or her passengers voluntarily accepted liability for their actions unless there is an explicit agreement to do so; increases in the size of damage of awards, due to changes in court's method of determining awards and advances in medical technology which enable many seriously injured accident victims to survive; and finally, changing public and judicial attitudes which are much less accepting of drunk driving. 6.

The hospitality industry has responded to the increased threat of civil liability in a generally positive and constructive manner, developing programmes aimed at preventing problems which may give rise to liability law suits. Practices and policies which reduce liability risks include ensur- ing the safety of premises, controlling access, promoting food consumption with alcohol, offer- ing nonalcoholic or low-alcohol alternatives, set- ting drinking limits for patrons, and training staff in responsible serving practices.

The available evidence from tavern studies indicates that serving practices have a marked impact on drinking behaviour. "Waiters and bartenders grant regular status to patrons via friendship behaviour, protective functions, special privileges and other preferential treatment. Their influence is so strong that it may be claimed that they select regular patrons and determine drinking rates". 4s Therefore, the widespread training of servers offers a good deal of potential for the prevention of alcohol problems arising from acute incidents of intoxication, particularly in licensed establishments.

At least one U.S. state, Oregon, and two Canadian provinces, Ontario and British Colum- bia, have developed a policy of mandatory server training for all persons employed in the hosp- itality industry. The development and evaluation of server training programmes in both mandatory and voluntary situations represents a promising area for future research on alcohol availability.

The structure of the alcohol industry Finally, there is one further aspect of

availability which has been neglected in the past and which may be given greater attention in the future. I refer to studies on the structure of alcohol production and distribution. Two recent World Health Organization projects ~'63 have ad- dressed this topic, and the W H O has repeatedly expressed concern over the excessive marketing and promotion of alcoholic beverages in the Third World on the part of private producers and monopolies from developed countries.

These projects spawned a number of case studies on particular aspects of alcohol distribu- tion. Thus, for example, Casswell and Smythe ~ documented the excessive promotion of alcohol in five South Pacific societies by multinational cor- porations. An analysis of the intercorporate ow- nership and interlocking directorates of the alco- hol industry in Canada ~s noted several findings which bear upon alcohol availability. It was found that ownership is highly complex. The alcohol industry has been generally transformed from a large set of small firms owned by entrepreneurs to a small set of large international corporations with extensive and complex intercorporate connec- tions. The distilling and brewing industries in particular have become highly oligopolistic.

From a public health perspective, this has had both positive and negative effects on alcohol availability. On one hand, the industry structure facilitates large advertising layouts, the develop- ment of new products and the promotion of

282

consumption in other ways as well. On the other hand, oligopolistic pricing and high barriers of entry in the market have restricted competition and thus avoided certain excesses associated with fierce competition.

Summary and conclusions In sum, the relationship between alcohol

availability and consumption has long been a contentious issue in the alcohol policy field. Although much of the debate on this issue in the recent past has focused on the ~Ledermann ~ or %ingle distribution" theory, the cornerstone of the position favouring strict controls over alcohol availability is not any claim regarding a particular distribution of consumption, but rather the em- pirical relationship of availability to alcohol con- sumption and problems. In general, a large body of research has found that relaxation of controls over availability such as the expansiion of the type and number of outlets, days and hours of opera- tion, and lower age restrictions, tend to be positively related to levels of alcohol consump- ilion, not only among moderate drinkers but also among ~heavy ~ drinkers as well. Future research on availability should focus on relatively neglected areas of enquiry such as the impact of the physical environment within outlets, the behaviour of servers and the structure of the alcohol industry.

Acknowledgements The author gratefully acknowledges the assis-

tance of Diane McKenzie, Jennifer Evans and Paulette Waiters in the preparation of this article. The views expressed in this paper are those of the author and do not necessarily reflect those of the Addiction Research Foundation.

Correspondence and requests for reprints to: Eric Single, Addiction Research Foundation, 33 Russell St., Toronto, Ontario Canada M5S 251.

References 1. World Health Organization. WHO Official Re-

cords No. 226, 48 Resolution WHA 28.81 .Geneva: WHO, 1975.

2. World Health Organization. Alcohol-related problems: The need to develop further the WHO initiative. Report to the Executive Board. WHO Document EB 63/23. Geneva: WHO, 1978.

3. World Health Organization. Thirty-second World Health Assembly, 7-25 May, 1979. Resolution WHA 32.40. Geneva: WHO, 1979.

4. World Health Organization. Regional Office for Europe. Public health aspects of alcohol and drug dependence. Report of a WHO Conference. Copenhagen: WHO Regional Office for Europe, 1979. (EURO Reports and Studies; 8).

5. World Health Organization. Prevention of alcohol- related problems: an international review of preven- tive measures, policies, and programmes. Geneva: WHO, 1979. (WHO document MNH/79.16).

6. World Health Organization. Report of a WHO Expert Committee on Problems Related to Alcohol Consumption. Geneva: WHO, 1980.

7. Moore M, Gerstein D. Alcohol and public policy: beyond the shadow of prohibition. Washington: National Academy Press, 1981.

8. Addiction Research Foundation. A strategy for the prevention of alcoholism: background information and recommendations for the parliamentary debate on control measures. Reprinted in: The Journal 1978; June.

9. Robinson S. Factors influencing alcohol consump- tion. In Edwards G, Grand M, eds. Alcoholism - - new knowledge and new responses. London: Croom Helm, 1977.

10. Popham RE. The social history of the tavern. In: Israel Yet al., eds. Research advances in alcohol and drug problems. Vol. 4. New York: Plenum Press, 1978.

11. Bruun K, Edwards G, Lumio M, Makela K, Pan L, Popham RE, et al. Alcohol control policies in public health perspective. (The Finnish Foundation for Alcohol Studies; Vol. 25). Forssa: Aurasen Kirjapaio, 1975.

12. Edwards G, Orford J, Egert S, Gutherie S, Hawker A, Hensman C et al. Alcoholism: a controlled trial of ~treatmenC and ~advice ~. J Stud Alcohol 1977; 38: 1004-1031.

13. Davies P, Walsh D. Alcohol problems and alcohol control in Europe. Beckenham, Kent: Croom Helm, 1983.

14. Makela K, Room R, Single E, Sulkunen P, Walsh B et al. Alcohol, society and the State. 1: a comparative study of alcohol control. Toronto: Addiction Research Foundation, 1981.

15. Schmidt W, Popham RE. Heavy alcohol consump- tion and physical health problems: a review of the epidemiological evidence. Drug Alcohol Depend 1975-76; !: 27-50.

16. Pequignot G, Chabert C, Eydoux H, Courcoul MA. Augmenation du risque de cirrhose en func- tion de la ration d'alcool. Rev Alcsme 1974; 20: 191.

17. Room R. Alcohol control and public health. Annu Rev Public Health 1985; 5:293-317.

18. deLint J, Schmidt W. The distribution of alcohol consumptio problem in Ontario. Q J Stud Alcohol 1968; 29: 968-973.

19. Chafetz M. The prevention of alcoholism. Int J Psychiatry 1970; 9: 329-348.

20. Room R. Ambivalence as a sociological explanation: the case of cultural explanations of alcohol problems. Am Soc Rev 1976; 41: 1047-1065.

21. Makela K. Consumption level and cultural drinking patterns as determinants of alcohol problems. J Drug Iss 1975; 5: 344-357.

22. Single E. Estimating the number of alcoholics in Ontario: a replication and extension of an earlier study. J Stud Alcohol 1979; 40:11, pp. I046-1052.

283

23. Osterberg E. The price of alcoholic beverages as an instrument of control policy. Reports of the Social Research Institute on Alcohol Studies, Helsinki No. 83, 1975.

24. Ledermann S. Alcool, alcoolisme, alcoolisation, Vol. 1. Donnees scientifiques de caractere physiologique, econiomique et social. (Institute National d'Etudes Demographique, Travaus et Documents, Cah. No. 29) Paris: Presses Univer- sitaires de France, 1956.

25. Pittman DJ. Primary prevention of alcohol abuse and alcoholism: a critical analysis of the control of consumption policy. Paper presented at the 26th International Institute on the Prevention and Treat- ment of Alcoholism, Cardiff, Wales, 1980.

26. Duffy JC. Estimating the proportion of heavy drinkers. In: The Ledermann curve. London: Alco- hol Education Centre, 1977.

27. Parker D, Harman M. The distribution of consum- ption model of prevention of alcohol problems: a critical assessment. J Stud Alcohol 1978; 39: 377- 399.

28. Pittman DJ, Strickland D. A critical evaluation of the control of consumption policy. Paper presented at conference on ~Control Issues in Alcohol Abuse Prevention: Local, State and National Designs for the 1980s ~, Charleston, September 1981.

29. Skog O-J. Social interaction and the distribution of alcohol consumption. J Drug Issues 1980; 10(1).

30. Skog O-J. The distribution of alcohol consumption 1: a critical discussion of the Ledermann model. Oslo: Statens Institutt for Alkoholforskning, 1982. (Report; 64).

31. Skog O-J. The collectivity of drinking cultures: a theory of the distribution of alcohol consumption. Oslo: National Institute for Alcohol Research, 1983. (SIFA Mimeograph No 69).

32. Single E. International perspectives on alcohol as a public health issue. J Public Health Policy 1984; 5(2): 238-256.

33. Popham RE, Schmidt, deLint J. The effects of legal restraints on drinking. In: Kissin B, Begleiter H. The biology of alcoholism. Vol. 4. New York: Plenum, 1976.

34. Smith DI. Effectiveness of restrictions on availability as a means of reducing the use and abuse of alcohol. Aust Alcohol Drug Rev 1983; 2(2): 84- 90.

35. Farrell S. Review of national policy measures to prevent alcohol-related problems. Geneva: World Health Organization, 1985.

36. Moskowitz J. The primary prevention of alcohol problems: a critical review of the research literature. J Stud Alcohol 1988; (in press).

37. deLint J. Recent trends in Dutch drinking: a problem case for availability explanations. Contemp Drug Problems 1981; 10: 179-192.

38. Kuusi P. Alcohol sales experiment in rural Finland. Helsinki: Finnish Foundation for Alcohol Studies, 1957.

39. Popham RE. Working papers on the tavern III: Notes on the contemporary tavern. Toronto: Ad- diction Research Foundation, 1982. (Substudy No. 219).

40. Raymond A. Ten o'clock closing - - the effect of the changes in hotel bar closing time on road accidents in the metropolitan area of Victoria. Aust Road Research 1969; 3: 3-17.

41. Douglass R, Filkins F. The effect of lower legal drinking ages on youth crash involvement. Ann Arbor: University of Michigan, Highway Safety Research Institute, 1974.

42. Single E, Morgan P, deLint J, eds. Alcohol, society and the State 2: the social history of control policy in seven countries. Toronto: Addiction Research Foundation, 1981.

43. Macdonald S, Whitehead P. Availability of outlets and consumption of alcoholic beverages. J Drug Issues 1983; 13: 477-486.

44. Rush BL, Gliksman L, Brook R. Alcohol availability, alcohol consumption and alcohol-re- lated damage. J Stud Alcohol 1986; 47: 1-10.

45. Single E. Studies of public drinking: an overview. In: Single E, Storm T, eds. Public drinking and public policy. Toronto: Addiction Research Foun- dation, 1985.

46. Caudill BD, Marlatt GA. Modeling influences in social drinking: an experimental analogue. J Con- sult Clin Psychol 1975; 43: 405-415.

47. DeRicco DA, Garlington WK. The effect of model- ing and disclosure of experimenter's intent on drinking rate of college students. Addict Behav 1977; 2: 135-139.

48. DeRicco DA. Effects of peer majority on drinking rate. Addict Behav 1978; 3: 29-34.

49. DeRicco DR, Niemann JE. In vivo effects of peer modeling on drinking rate. J Appl Behav Anal 1980; 13: 149-152.

50. Bruun K. Drinking behaviour in small groups. Helsinki: Finnish Foundation for Alcohol Studies, 1959.

51. Sommer R. The isolated drinker in the Edmonton beer parlor. Q J Stud Alcohol 1965; 26: 95-110.

52. Storm T, Cutler R. Observations of drinking in natural settings: Vancouver beer parlors and cock- tail lounges. J Stud Alcohol 1981; 42: 972-997.

53. Radcliffe WD, Nutter R, Hewitt D, Flanders P, Caverhill K, Gruber G. Drinking behaviours in lounges and taverns. Research Report. Edmonton: Alberta Alcoholism and Drug Abuse Commission, 1980.

54. Foy DW, Simon SJ. Alcoholic drinking topography as a function of solitary versus social context. Addict Behav 1978; 3: 39-41.

55. Bach PJ, Schaefer JM. The tempo of country music and the rate of drinking in bars. J Stud Alcohol 1979; 40: 1058- 1959.

56. Sommer R. Designed for drinking. In: Personal space: the behavioural basis of design. Englewood Cliffs, N.J.: Prentice Hall, 1969.

284

57. Cloyd JW. The market-place bar: the interaction between sex, situation and strategies in pairing ritual of homo ludens. Urban Life 1976; 5: 293-312.

58. Cavan S. Liquor license: An ethnography of a bar. Chicago: Aldine, 1966.

59. Babor T, Mendetson J, Greenberg I, Kuehnle J. Experimental analysis of the happy hour: effects of purchase price on alcohol consumption. J Psycho- pharmacol 1978; 58: 35-41.

60. Solomon R, Boake B, Gleason ML. One for the road: a tavern owner's liability as a provider of alcohol. In: Single E, Storm T, eds. Public drinking and public policy. Toronto: Addiction Research Foundation, 1985.

61. Solomon R, Single E. The civil liabilities of providers of alcohol. Background paper prepared for the Ontario Advisory Committee on Liquor Regulation. Toronto: Liquor Licence Board of Ontario, 1986.

62. World Health Organization. Public health aspects of alcohol availability. Report on working meeting held in Geneva, 1982.

63. World Health Organization. Potential contribution of state monopoly systems to the control of alcohol- related problems. Report on working meeting held in Skarpo, Sweden. Geneva: WHO, 1987. (WHO Document No. MNH/PAD/87.15).

64. Casswell S, Smythe M. Alcohol trade in the South Pacific: a pilot study in Papua New Guinea, Vanuatu, Fiji, Western Samoa and Tonga. Paper presented at meeting of the WHO Project on Public Health Aspects of Alcohol Availability, Geneva, October 1982.

65. Single E. The structure of the alcohol industry as it relates to transportation issues. Accident Anal Prey 1987; 19: 419-431.