Estimating the Effects of Advertising: Application to a Social Problem

ABSTRACT - "Per capita" consumption of alcohol has increased dramatically over the last twenty years; for instance, between 1954 and 1974, the total intake of raw alcohol has more than tripled. To many, this social problem is cause for concern. This paper investigates those variables which may have had a significant impact on this increase. It was found that non-controllable variables (e.g., income, unemployment, and urban/rural mix) were generally more significantly related to higher per capita consumption than were marketing/controllable variables (e.g., advertising, distribution, price).

Citation:

James G. Barnes and Jacques C. Bourgeois (1979) ,"Estimating the Effects of Advertising: Application to a Social Problem", in NA - Advances in Consumer Research Volume 06, eds. William L. Wilkie, Ann Abor, MI : Association for Consumer Research, Pages: 550-556.

Advances in Consumer Research Volume 6, 1979      Pages 550-556

ESTIMATING THE EFFECTS OF ADVERTISING: APPLICATION TO A SOCIAL PROBLEM

James G. Barnes, Memorial University of Newfoundland

Jacques C. Bourgeois, Carleton University

ABSTRACT -

"Per capita" consumption of alcohol has increased dramatically over the last twenty years; for instance, between 1954 and 1974, the total intake of raw alcohol has more than tripled. To many, this social problem is cause for concern. This paper investigates those variables which may have had a significant impact on this increase. It was found that non-controllable variables (e.g., income, unemployment, and urban/rural mix) were generally more significantly related to higher per capita consumption than were marketing/controllable variables (e.g., advertising, distribution, price).

THE PROBLEM

Recent years have seen a marked increase in the per capita consumption of beverage alcohol products in Canada. In the year ending March 31, 1974, for example, the average Canadian aged 15 and above consumed 25.24 Imperial gallons of beer, 1.67 gallons of wine, and 2.11 gallons of spirits. Since 1954, the annual per capita intake of raw alcohol (from beer, wine and spirits) has increased by more than 57 per cent (see Table 1).

TABLE 1

PER CAPITA CONSUMPTION OF BEVERAGE ALCOHOL PRODUCTS IN CANADA, BY PERSONS AGED 15 AND ABOVE: 1954, 1964 AND 1974

Such statistics have been considered cause for alarm by many groups. The social costs associated with the rapid increase in consumption of beverage alcohol products have been widely debated. Lalonde (1975) estimated that seven per cent of the drinking population consume 40 per cent of all alcohol sold. This translates into 15 gallons of raw alcohol consumed each year by each person in the heavy drinking population. It is the alcoholic and the occasional heavy drinker who contribute most to the onerous social costs associated with the consumption of beverage alcohol products.

The Commission of Inquiry into the Non-Medical Use of Drugs concluded in their Final Report (1973) that

alcohol is, and is likely to remain, Canada's most serious non-medical drug use problem ... From almost any point of view the effects of the excessive use of alcohol are more harmful than those of any other form of non-medical drug use. (p. 37)

The problems associated with excessive use of alcoholic beverages have been discussed under three major headings: physical health problems, mental health problems, and social problems.

It has been clearly demonstrated that excessive use of alcohol contributes to diseases such as cirrhosis of the liver, hepatitis, cancer of the upper respiratory and digestive tracts, tuberculosis, and malnutrition. Heavy use of alcohol has similarly been associated with a variety of psychiatric and neurological disorders. In Canada in 1972, 17 per cent of first admissions and 16 per cent of readmissions to psychiatric wards and hospitals were for alcoholism and alcoholic psychosis.

The social problems associated with excessive alcohol use are numerous. The Commission of Inquiry into the Non-Medical Use of Drugs (1973) concluded that, of all drugs used medically and non-medically, alcohol has the strongest and most consistent relationship to crime. In their Final Report (1973) this Commission cited a recent Philadelphia study which indicated that alcohol was present in either the offender or the victim in 64 per cent of homicide cases over a five-year period. Similarly, the contribution of alcohol use to motor vehicle accidents has been well documented. About half of all drivers killed in automobile accidents have been found to have positive blood alcohol levels.

In industry, excessive alcohol use contributes to high rates of absenteeism and to numerous on-the-job accidents. Equally serious, but considerably more difficult to quantify, are the problems which are imposed by problem drinkers on their family and friends. The impact of the alcoholic parent on the lives of his or her children may be significant.

Some attempts have been made to quantify the social costs associated with excessive use of alcohol. The costs of lost production, health and medical care, motor vehicle accidents, alcohol programs and research, criminal justice and social welfare systems have been estimated at over $25 billion per year in the United States and $2.5 billion per year in Canada (Health and Welfare Canada, 1975). Since these estimates were made in 1971, it is likely that the total identifiable costs of alcohol-related problems today are closer to $40 billion in the United States and $4 billion in Canada.

ADVERTISING'S CONTRIBUTION TO INCREASED CONSUMPTION

The question of the effect of advertising on aggregate demand has been debated at length. In the tobacco and alcohol industries in particular, critics have been advocating for some time that advertising should be reduced or eliminated. The assumption obviously is that such a reduction in advertising would lead to a corresponding reduction in aggregate demand or per capita consumption. But studies of the impact of advertising on total sales of cigarettes have tended to indicate that advertising's effect on industry sales can not be demonstrated. Schmalensee (1972) concluded that "advertising seems unlikely to be a very powerful force in the cigarette industry". In Canada, per capita consumption of cigarettes has continued to increase, despite the fact that manufacturers voluntarily withdrew all cigarette advertising from the broadcast media in the late 1960's.

The lack of proof concerning the effect of advertising on industry sales has not stopped the critics of advertising from proposing that the advertising of all alcoholic beverages be banned. The association of advertising with increasing consumption of alcoholic beverages has not been made only by religious groups or social activists. Some leading politicians have been outspoken in their criticism, not only of the sheer volume of advertising, but also of the fact that much advertising, especially for beer, presents the product in the context of a glamorous life style, suggesting that use of the product is essential to having a good time (Lalonde, 1974).

The criticism of beverage alcohol advertising is very often directed toward the observation that advertising leads young people to start drinking. The very rapid increase in consumption of alcoholic beverages among young people in recent years is often cited as an indication of the influence of advertising (Ontario Youth Secretariate, 1976). A more plausible explanation of the increase might be found in the fact that all ten Canadian provinces have reduced the legal drinking age from 21 to 18 or 19 within the past five years, although two provinces have recently raised it again from 18 to 19.

Advertising of alcoholic beverages is regulated in Canada by both the federal and provincial governments. At present, advertising of spirits is not permitted on radio or television and all beer advertising in these media is subject to the approval of the Canadian Radio-Television and Telecommunications Commission. Media other than broadcast are regulated by provincial governments and the stringency of such regulations is quite variable. For example, three provinces in effect prohibit all forms of advertising for beverage alcohol products, while one province has no regulations dealing with such advertising.

VARIABLES INFLUENCING PER CAPITA CONSUMPTION

The social problems associated with the excessive use of alcohol are also associated primarily with the alcoholic and the heavy drinker. A number of authors have explored the possibility that the distribution of alcohol consumption is positively skewed with a curve approximating the lognormal distribution (Miller and Agnew, 1974). If such a distribution applies, then the number of heavy drinkers and alcoholics in a population may be determined as a function of the mean (per capita) level of consumption of alcohol. This observation has led certain authors to suggest that the best way to reduce the number of heavy drinkers, and the problems associated with excessive use of alcohol, is to reduce the mean level of consumption in the population (Schmidt and de Lint, 1971). Before a program can be launched to reduce per capita alcohol consumption, some information must exist concerning those factors which influence the mean level of consumption.

Governments in many jurisdictions are currently being pressured to take some action to reverse the trend toward increased per capita consumption of alcoholic beverages. But, before governments can take any action, it would be appropriate for them to attempt to determine which factors are most important in influencing per capita consumption. The factors which have been associated with per capita consumption of beverage alcohol products may be divided into two categories: (1) the controllable or 'marketing' variables such as price, advertising, and availability, which may be influenced by governments in the short run; and (2) the uncontrollable or socio-economic variables such as per capita. income, urbanization rate, and ethnic mix of population, which are not subject to change in the short run.

Availability

Various aspects of distribution or availability have been linked to levels of per capita consumption. Schmidt (1972) suggested that recent increases in consumption are a result of reductions in the legal drinking age. Popham, et al. (1971) have concluded that per capita consumption of alcoholic beverages is unaffected by the hours of sale in liquor stores and by the number of outlets at which liquor is sold. Simon (1966b) concluded that per capita consumption is not related to the fact that alcoholic beverages are sold by state monopoly rather than by private enterprise. Smart (1974) and Goldberg and Gorn (1975) have observed different purchase patterns by customers in self-serve liquor stores as compared with customers in conventional counter-service stores.

Price

Much has been written about the impact on the level of per capita consumption of alcoholic beverages. Seeley (1960) has shown that death rates from cirrhosis of the liver have increased as the price of alcohol (expressed as a percentage of personal disposable income) declined. Since the level of government taxation significantly influences the price of alcoholic beverages, a number of attempts have been make to measure the price elasticity of demand for alcoholic beverages (Simon, 1966a; Johnson, 1973). The results of these attempts have been largely inconclusive, as they tend to vary with the country studied, size of sample, and the objectives of the study. Several authors (Popham, et al., 1971; Johnson, 1973) have concluded that the tax structure can be used as a valuable tool to reduce per capita consumption of alcoholic beverages. The price of alcoholic beverages relative to personal disposable income has been declining for at least forty years and the per capita consumption of these products and the levels of alcoholism in the population continue to rise.

Advertising

Despite the ongoing criticism of beverage alcohol advertising and the demands for increased regulation, few studies have been devoted to examining the question of advertising's effect. Globetti (1973) discussed the potential of alcohol education programs as measures to counteract advertising for alcoholic beverages. Other authors have examined the effect of cigarette and proprietary drug advertising on per capita consumption of cigarettes and on drug abuse. Hulbert (1974) concluded that television and the advertising of proprietary drugs on television do not appear to be important influences on usage of illicit drugs.

Socio-economic Variables

A series of socio-economic variables has also been associated with per capita consumption of beverage alcohol products. Wechsler, et al. (1970) examined the drinking behavior of various religious and ethnic subgroups and found that alcohol consumption was significantly related to religious-ethnic subgroup membership. Bales (1946) observed that factors affecting problem drinking are often culturally based. Cisin and Cahalan (1971) concluded that whether or not a person drinks is primarily related to socio-cultural rather than psychological variables. They concluded that socio-economic status and urbanization are important correlates of heavy drinking. In a recent study, Johnson and Oksanen (1974) incorporated ethnic, religious, educational and occupational variables into a study of per capita consumption and concluded that "these variables, along with traditional economic ones, are significant determinants of the demand for alcoholic beverages" (p. 12). In one of the most detailed studies to date, Cahalan, et al. (1969) associated per capita consumption of alcoholic beverages with education levels, degree of urbanization, and religion, among other variables. He found that the proportion of heavy drinkers was found to be lowest in the income group below $4,000. per year, and highest in the $10,000. to $14,999. range. Heavy drinkers tended to be more common among urban business and semiprofessional people; among men who had completed high school and not graduated from college; and in large cities. Catholics were above average in the proportion of drinkers considered 'heavy drinkers', but the highest proportion of heavy drinkers was among Irish Protestants.

THE DATA

Data were obtained primarily from Statistics Canada reports and from the Annual Reports and price lists of the various provincial liquor boards and commissions, with the assistance of the Canadian Association of Provincial Liquor Commissioners. The data were gathered from all ten provinces of Canada for the period 1966 to 1973.

Dependent Variables

Measures were obtained of per capita consumption of beer, wine, and spirits by province for each of the eight years from 1966 to 1973. A measure of per capita consumption of actual beverage alcohol was also obtained. (1) per capita consumption of beer: the measure of per capita consumption of beer represents consumption by residents of each province who are aged 15 or older. It was felt that this was the age group which is most likely to be involved in the consumption of beverage alcohol products. The total provincial consumption of beer and ale for each year was simply divided by the population aged 15 and older to obtain the relevant figure.

(2) per capita consumption of wine: this figure was obtained in exactly the same way as was the beer consumption figure discussed above. The total provincial consumption of wine was divided by the population aged 15 and older to obtain the per capita consumption rate.

(3) per capita consumption of spirits: this figure was also obtained by dividing the total provincial consumption of spirits for each year by the provincial population aged 15 and older.

(4) per capita consumption of beverage alcohol: this is a measure of per capita raw alcohol consumption and is included in the analysis in order to provide a measure of actual alcohol consumption. This is necessitated by the fact that the three major types of beverage alcohol products differ in actual alcohol content and an analysis based simply on gallonage of each product type consumed could be misleading. For example, a person who consumes 15 gallons of beer is obviously consuming less alcohol than is a person who consumes 15 gallons of spirits, simply because beer is much lower in alcohol content than is spirits.

This measure of actual beverage alcohol consumption is obtained by converting actual volume of beer, wine, and spirits consumed into beverage alcohol volume through application of the following standard conversion factors:

beer: 5% alcohol by volume

wine: 16% alcohol by volume

spirits: 40% alcohol by volume

Thus a person who consumes 20 gallons of beer in a year is actually consuming one gallon of beverage alcohol. Similarly, 10 gallons of wine is equivalent to 1.6 gallons of beverage alcohol, and 15 gallons of spirits is equivalent to 6 gallons of beverage alcohol.

Independent Variables

Measures were obtained on a total of nine different variables which were hypothesized to be related to per capita consumption of beverage alcohol products.

(1) price of beer: this variable represents simply the price set by the provincial liquor boards at which a case of 24 bottles of beer may be sold at retail.

(2) price of wine: this variable is an index of wine prices. A total of twenty popular brands of imported and domestic wines which are sold in all ten provinces were selected and the prices at which these brands were sold in each province for each year were obtained. This variable represents the average price per bottle of these wines.

(3) price of spirits: this variable was calculated in the same manner as was the wine price discussed above. Twenty popular brands of spirits were selected and their prices in each province calculated. The variable represents the average price per bottle of these spirits.

(4) percentage rural population: this variable represents the percentage of the populations of each province which is classified as rural dwellers by Statistics Canada.

(5) per capita income: this variable represents the per capita income of the residents of each province aged 15 and older and is obtained by dividing the gross provincial income by the total population aged 15 and older.

(6) unemployment rate: this variable represents the average annual unemployment rate for each province as published annually by Statistics Canada.

(7) number of liquor stores per 100,000 population: this variable simply represents the number of liquor stores operated in each province by the provincial liquor commission, divided by the population aged 15 and older.

(8) licensed establishments per 100,000 population: this variable represents the number of establishments licensed by the provincial liquor commission to sell beverage alcohol products in the province, divided by the population aged 15 and older.

(9) advertising regulation index: this variable represents the extent to which each province regulates the advertising and promotion of beverage alcohol products.

A province which permits all forms of advertising and promotion and has no regulations or legislation pertaining to the promotion of beverage alcohol products would score 1 on this variable. Conversely, a province which prohibits all forms of advertising and promotion for such products would score 9 on this regulation index. Scores were assigned to each of the ten provinces for each of the eight years covered by this study. The scores were assigned using a 'panel of experts' approach which involved a detailed study of the regulations and legislation of each provincial government and provincial liquor commission for each of the eight years pertaining to any form of advertising or promotion of beverage alcohol products. Detailed descriptions of the regulatory picture in each province, with changes in regulations and laws being noted as they occurred during the eight-year period, were prepared and a group of university professors who were familiar with marketing and its regulation were asked to rate each province each year on a scale from 1 to 9, depending on the stringency of the provincial regulations. These ratings were then averaged and became the scores for each province on this variable.

Scores on each of the independent variables were obtained for each of the ten provinces for each of the eight years from 1966 to 1973 with a small number of exceptions. Annual reports and price lists were unavailable from four provincial liquor commissions and consequently the 'price of wine', 'price of spirits', 'licensed establishments per 100,000 population', and 'liquor stores per 100,000 population' variables could not be calculated for these provinces for certain of the years in the eight-year period. The analysis of these variables was carried out using complete data only.

The data on the four dependent and nine independent variables from the ten provinces for each of the eight years from 1966 to 1973 were then analyzed using a step-wise multiple regression analysis. In all, four different regression analyses were performed, with each of the four per capita consumption measures discussed above as dependent variables.

ANALYSIS

Per Capita Consumption - Beer

A total of only four independent variables were significantly (at the .05 level) related to per capita consumption of beer. The results of this first regression analysis are presented in Table 2.

TABLE 2

SUMMARY OF REGRESSION ANALYSIS FOR PER CAPITA CONSUMPTION OF BEER

These results indicate that the per capita consumption of beer in Canada's ten provinces is influenced most significantly by the distribution of population in urban centres, by per capita income levels in the provinces, by the level of unemployment, and by the price being charged for competing products, namely spirits. No other variable was found to have a significant effect on per capita consumption of beer.

The regression equation presented in Table 2 indicates that the more urbanized the population of a province, the higher will be the consumption rate of beer among the residents of that province. Those provinces which have a large percentage of their residents considered rural dwellers generally have lower per capita consumption rates for beer. Secondly, more beer is consumed by the residents of those provinces where per capita incomes are highest. Residents of low income provinces generally have lower consumption rates for beer. Thirdly, unemployment rates appear to influence the level of per capita consumption of beer in a negative direction. Those provinces where unemployment is highest generally have lower per capita consumption rates for beer.

Finally, a substitution effect of sorts seems to operate between beer and spirits in that there is a direct link between the price of spirits and per capita consumption of beer. These data would suggest that as the price of spirits is increased, consumers will consume larger quantities of beer.

The remaining five independent variables were not found to be significantly related to per capita consumption rates for beer. That is, consumption of beer appears to be affected neither by the price being charged for beer in the provinces, nor by the number of liquor stores and licensed establishments, nor by the extent to which the provinces regulate the advertising of beverage alcohol products.

Per Capita Consumption - Wine

Four independent variables also entered the regression equation with per capita consumption of wine as the dependent variable. These four variables explained 56 per cent of the variance in the data. The results of this regression analysis are presented in Table 3.

These results suggest that per capita consumption of wine in Canada is influenced significantly by income levels in the various provinces, by the rural/urban distribution of population, by the number of liquor stores operated by the provincial liquor authorities (per 100,000 population), and by the price being charged by the provincial authority for beer. None of the other five independent variables was found to be significantly related to per capita consumption of wine.

TABLE 3

SUMMARY OF REGRESSION ANALYSIS FOR PER CAPITA CONSUMPTION OF WINE

The regression equation presented in Table 3 indicates again that the higher the per capita income in a province, the higher will be the volume of wine consumed by the residents of that province.

Secondly, as was the case in the previous analysis (Table 2), the extent to which the population of a province is urbanized significantly affects the level of per capita consumption of wine. The more urbanized the population of a province, the higher will be the per capita consumption rate for wine in that province.

Thirdly, the accessibility of retail outlets for wine would appear also to significantly influence per capita consumption rates. Those provinces which have larger numbers of government-operated liquor stores per 100,000 population also exhibit higher rates of per capita consumption of wine. Such a relationship does not exist between consumption rates for wine and the numbers of licensed establishments in a province as, presumably, most wine is sold through liquor stores for home consumption.

Finally, a substitution effect is again evident in the link between the price of beer and consumption rates for wine. As the price of beer increases, so also does the per capita consumption of wine as consumers would appear to be substituting wine for beer as the price of the latter is increased.

Per capita consumption rates for wine appear to be significantly influenced neither by unemployment rates, nor by the prices being charged for wine and spirits, nor by the extent to which the provinces regulate the advertising of beverage alcohol products.

Per Capita Consumption - Spirits

Only two independent variables were found to be significantly (at the .05 level) related to per capita consumption of spirits. These two variables explained 44 per cent of the variance in the data. The results of this analysis are presented in Table 4.

TABLE 4

SUMMARY OF REGRESSION ANALYSIS FOR PER CAPITA CONSUMPTION OF SPIRITS

These results indicate that per capita consumption rates for spirits are most significantly affected by per capita income levels. Again, those provinces where personal incomes are highest demonstrated the highest per capita consumption rates for spirits, as they had for beer and wine.

Of the four regression analyses performed, this analysis with per capita consumption of spirits as the dependent variable was the only one where the relationship between the regulation of advertising and the dependent variable was significant. It is suggested in this analysis that the more stringent the regulation of advertising of beverage alcohol products in a province, the lower will be the per capita consumption rate for spirits in that province.

The other seven independent variables failed to produce a significant relationship with the dependent variable. Per capita consumption of spirits appears to be significantly affected neither by the prices being charged for beer, wine and spirits, nor by the number of liquor stores and licensed establishments serving the public, nor by unemployment and population urbanization variables.

Per Capita Consumption - Alcohol

An attempt was made to determine those factors which influenced overall alcohol consumption among Canadians. Per capita alcohol consumption figures were obtained by converting consumption rates for beer, wine and spirits into equivalent volumes of beverage alcohol, as discussed earlier.

A regression analysis revealed that four variables were significantly related to per capita alcohol consumption. These four variables explained 68 per cent of the variance in the data. The results of this final analysis are presented in Table 5.

TABLE 5

SUMMARY OF REGRESSION ANALYSIS FOR PER CAPITA CONSUMPTION OF TOTAL ALCOHOL

These results indicate that overall per capita consumption of beverage alcohol among Canadians is influenced most significantly by the urban/rural distribution of the population in specific provinces, by personal income levels, by the prices charged by provincial liquor authorities for spirits, and by levels of unemployment in the provinces.

This final regression equation presented in Table 5 indicates again that the more urbanized the population of a province, the higher will be the overall rate of per capita consumption of beverage alcohol in that province. Similarly, those provinces which have the highest levels of personal income also exhibit the highest rates of beverage alcohol consumption.

An interesting relationship between the price of spirits and overall consumption of beverage alcohol products may be observed in Table 5. These results suggest that as the price charged by provincial liquor authorities for spirits is increased, a corresponding increase will occur in per capita consumption of beverage alcohol. While this relationship is statistically significant, it is not the relationship which would have been hypothesized to exist. Normally, one would expect consumption to decline, or at least increase at a slower rate, as prices increase. These results, obtained with regard to the effect of the price of spirits on per capita consumption of alcohol, would suggest that other variables are operating through the price variable and which are highly correlated with both the price and consumption variables.

Finally, overall per capita alcohol consumption was found to be significantly affected by unemployment rates. Those provinces which exhibit the highest rates of unemployment among their residents generally exhibit the lowest rates of per capita consumption of beverage alcohol.

SUMMARY AND CONCLUSIONS

Those variables which have been found to be most significantly related to per capita consumption rates of beverage alcohol products are generally 'societal' variables rather than 'marketing' variables. Consequently, significant relationships were found between per capita consumption of beer, wine, spirits, and beverage alcohol and such variables as unemployment levels, per capita income and urban density of population. All of these variables are not controllable in the short and medium run by provincial liquor authorities or by manufacturers of beverage alcohol products. Generally, those variables which are controllable such as price, number of liquor stores and licensed establishments, and stringency of advertising regulation were found to not significantly affect per capita consumption rates.

Such results would suggest that attempts on the part of governments to reduce, or at least slow the increase in, per capita consumption rates will have little effect if such attempts take the form of price increases, reduction in the number of licenses issued or liquor stores opened, or increased regulation of advertising of beverage alcohol products. Consumer behavior with respect to beverage alcohol products would appear to be not significantly influenced by traditional 'marketing' variables such as price, distribution, and promotion. While such variables may be influential in affecting consumer choice among brands, this study would suggest that they have little effect on overall rates of consumption of beverage alcohol products. Governments should not expect consumption to decrease dramatically if prices are raised or advertising is restricted. If reduced, or more moderate, consumption of beverage alcohol products is desired, then possibly increased consumer education is a preferred alternative.

The results presented in this paper should be interpreted with caution. This study was intended as a preliminary examination of an area of study which was felt to have promise as a fertile area for future research. Certainly, not all variables which might be hypothesized to influence per capita rates of beverage alcohol consumption have been included in this study. Also, those variables which have been included, both dependent and independent, may suffer from problems of definition.

Despite the shortcomings of this preliminary study, it is possible to draw some interesting observations from it. It is possible also to regard this study as a 'hypothesis generating' exercise, important in the questions which it raises, rather than in those which it answers. From the point of view of the authors, this study was designed to explore primarily the question of whether increased governmental regulation of advertising might be expected to contribute to a reduction in per capita consumption of beverage alcohol products. Certainly, the preliminary results presented herein do not permit a conclusive answer to this question, although they would suggest that regulation of advertising does not have a significant effect on per capita consumption, at least of beer and wine.

Additional research in this area is planned to explore more deeply the question of whether increased government regulation is necessarily the most valid approach to the solution of certain perceived social problems.

REFERENCES

Robert Freed Bales, "Cultural Differences in Rates of Alcoholism", Quarterly Journal of Studies on Alcohol, 6, 1946, 480-499.

Don Cahalan, Ira H. Cisin and Helen M. Crossley, American Drinking Practices: A National Study of Drinking Behavior and Attitudes (New Brunswick, NJ: Rutgers Center of Alcohol Studies, 1969).

Canada, Final Report of the Commission of Inquiry Into the Non-Medical Use of Drugs (Ottawa: Information Canada, 1973).

Ira H. Cisin, and Don Cahalan, "Some Correlates of American Drinking Practices", in Nancy K. Mello and Jack H. Mendelson (eds.), Recent Advances in Studies of Alcoholism: An Interdisciplinary Symposium (Rockville, Md.: National Institute of Mental Health, 1971).

Gerald Globetti, "Approaches to the Control of Alcoholic Beverages in the United States", Journal of Drug Issues, 3 (No. 3, 1973), 260-266.

Marvin E. Goldberg, and Gerald J. Gorn, "The Influence of In-Store Marketing on the Sale of Liquor", Working Paper, Faculty of Management, McGill University, 1975.

Health and Welfare Canada, "Alcohol Problems in Canada: A Summary of Current Knowledge", No. ERD-75-22, Ottawa: 1975.

James Hulbert, "Applying Buyer Behavior Analysis to Social Problems: The Case of Drug Use", in Ronald C. Curhan (ed.), 1974 Combined Proceedings (Chicago: American Marketing Association, 1974).

J. A. Johnson, "Canadian Policies in Regard to the Taxation of Alcoholic Beverages", Working Paper 73-24, Department of Economics, McMaster University, 1973.

James A. Johnson, and Ernest H. Oksanen, "Socio-Economic Determinants of the Consumption of Alcoholic Beverages", Working Paper 74-02, Department of Economics, McMaster University, 1974.

Marc Lalonde, "On Marketing the Malt", Addictions, 21 (Winter, 1974), 32-40.

Marc Lalonde, A New Perspective on the Health of Canadians: A Working Document (Ottawa: Information Canada, 1975).

Gary H. Miller, and Nell Agnew, "The Ledermann Model of Alcohol Consumption", Quarterly Journal of Studies in Alcohol, 35, 1974, 877-898.

Ontario Youth Secretariat, "Youth and Alcohol", Toronto: 1976.

Robert E. Popham, Wolfgang Schmidt and Jan de Lint, "The Prevention of Alcoholism: Epidemiological Studies of the Effects of Government Control Measures", Toronto: Addiction Research Foundation, 1971.

Richard Schmalensee, The Economics of Advertising (Amsterdam: North-Holland Publishing Company, 1972).

Wolfgang Schmidt, "A Note of the Effect of Lowering the Legal Drinking Age on the Consumption of Alcoholic Beverages", Toronto: Addiction Research Foundation, 1972.

Wolfgang Schmidt, and Jan de Lint, "Alcohol Use and Alcoholism", Addictions, 18 (No. 2, 1971), 1-15.

John R. Seeley, "Death by Liver Cirrhosis and the Price of Beverage Alcohol", The Canadian Medical Association Journal, 83(December 24, 1960), 1361-1366.

Julian L. Simon, "The Price Elasticity of Liquor in the U.S. and a Simple Method of Determination", Econometrica, 34 (January, 1966a), 193-205.

Julian L. Simon, "The Economic Effects of State Monopoly of Packaged-Liquor Retailing", Journal of Political Economy, 74(April, 1966b), 188-194.

Reginald G. Smart, "The Effect of Self-Service Stores on the Purchase of Alcoholic Beverages", Toronto: Addiction Research Foundation, 1974.

Henry Wechsler, Harold W. Demone, Jr., Denise Thum and Elizabeth H. Kasey, "Religious-Ethnic Differences in Alcohol Consumption", Journal of Health and Social Behavior, 11 (March, 1970), 21-29.

----------------------------------------