This research stemmed from two related but conflicting standpoints. The first is a social concern resulting from the fact that smoking is a health hazard. The second is the fact that tobacco growers and producers are putting every effort possible to promote their product despite the fact that smoking is ruled a major cause of lung cancer. The objective of the research was then to define, empirically, any relationship that may exist between tobacco consumption and a number of demographic and socioeconomic factors. To this end, the authors used a micro-data of the 1973 Consumption Expenditure Interview Survey involving 10,105 consumer units and about 2,500 items conducted by the Bureau of the Census for the Bureau of Labor Statistics. The data pertaining to the 5,034 smoking consumer units identified were subjected to multiple regression analysis using tobacco consumption expenditure as the dependent variable and about 60 other economic and demographic attributes or possible explanatory variables as independent variables. However, except for the dependent variable, income, and family size, the rest of the variables were treated as dummy variables for specific groups of interest as was necessary. As usually expected with cross-sectional data, R square was low. Nonetheless, it was observed that several socioeconomic factors affect significantly consumption of tobacco products. They include level of education attained, family size, race, marital status, geographic region, location of residence, occupation, and industry of employment. To mention just a few of these relationships, other things being equal: (1) Heads of a consumer unit or spouses with college education, blacks, and married couples spend less on tobacco consumption; (2) Tobacco consumption and family size are positively correlated; and (3) Consumer units living in central cities of large metropolitan areas tend to spend relatively more on tobacco consumption than their counterparts living in other areas. Chi-square analysis was used to determine tobacco consumption patterns of various demographic groups. The results are consistent with the ones arrived at using multiple regression analysis. It must be pointed out, however, that these findings represent merely symptomatic phenomena of smoking behavior of the population. Literature on health and marketing has tried to identify the real causes and effects of smoking. However, their findings are still inconclusive. That is, in fact, the reason why the government and consumer advocates and tobacco industry cannot agree even in their honest debate. It is hoped that the results of statistical analysis using a large-scale micro-data base such as this will shed more light on smoking behavior when subjected to more advanced techniques of analysis so that policymakers can utilize the findings in their public or private decisions. To this end, the authors plan to employ discriminant analysis by taking smoking and nonsmoking consumer units as the subfiles of interest. They also plan to analyze the Consumption Expenditure Interview Data for I984 so that a comparative analysis can be made between the two years.
Read full abstract