Exploring Generalized and Personalized Beliefs Among Smokers and Non-Smokers: a First Look

Kjell Gronhaug, The Norwegian School of Economics and Business Administration
Norman Kangun, The Norwegian School of Economics and Business Administration
ABSTRACT - In this paper we indent to look at awareness of public statements about the health hazards associated with smoking. We also want to determine what general beliefs, both health and non-health related, are held by the public, and to what extent these beliefs are personally held. The findings based on a Norwegian sample, showed great variations in generalized and personalized beliefs among smokers and non-smokers.
[ to cite ]:
Kjell Gronhaug and Norman Kangun (1979) ,"Exploring Generalized and Personalized Beliefs Among Smokers and Non-Smokers: a First Look", in NA - Advances in Consumer Research Volume 06, eds. William L. Wilkie, Ann Abor, MI : Association for Consumer Research, Pages: 184-190.

Advances in Consumer Research Volume 6, 1979      Pages 184-190


Kjell Gronhaug, The Norwegian School of Economics and Business Administration

Norman Kangun, The Norwegian School of Economics and Business Administration


In this paper we indent to look at awareness of public statements about the health hazards associated with smoking. We also want to determine what general beliefs, both health and non-health related, are held by the public, and to what extent these beliefs are personally held. The findings based on a Norwegian sample, showed great variations in generalized and personalized beliefs among smokers and non-smokers.


Smoking has been under attack almost continuously since tobacco was introduced to the civilized world over 400 years ago. As medical research tightened the statistical connection between incidence of smoking and various diseases (i.e., lung and throat cancer and various respiratory ailments), demands have been made in the last 15 years by various health interests groups, that governmental action should be taken to reduce smoking. To reduce the incidence of smoking, governmental bodies in both the United States and Europe have initiated a wide range of actions. Foremost among these actions has been taxation and regulation, which have been directed at influencing the conduct of producers and intermediaries. In addition, government sponsored anti-smoking information campaigns have been directed toward present and prospective smokers in the hope of discouraging smoking behavior.

This paper focuses in a limited manner on the effectiveness of the anti-smoking information program of the Norwegian government. More specifically, we seek to determine as a result of that government's campaign the extent to which Norwegian consumers are aware of public statements about the health hazards associated with smoking. We also want to determine what general beliefs, both health and non-health related, Norwegians hold with respect to the consequences of smoking, and to what extent, these beliefs are personally held. Before examining these questions, a brief discussion of anti-smoking information activities by most public authorities will be undertaken to be followed by an attempt to characterize the government model of smoking behavior which underlies their influence attempts. Furthermore, the kind of information public officials need to assess, the effects of their anti-smoking campaigns is presented.


The most common information activities utilized by governmental officials to discourage smoking have been (1) advertising bans, (2) health warnings and (3) educational campaigns. Each has been used widely. (Kangun and Gr°nhaug, 1978.) Bans on advertising place restrictions on the conduct of sellers. The rationale for such bans is principally that the absence of positive information about smoking may reduce the propensity to smoke, particularly among young people. To date, it appears that such bans have had limited influence on cigarette consumption (Hamilton, 1977; McGuiness and Cowling, 1975). Health warnings represent a form of negative point-of-purchase information. Their use seems to be based on the assumption that continually reminding people of the health hazards associated with smoking will discourage them from smoking. The underlying idea behind the use of educational campaigns appears to be that since many individuals lack information about the consequences of smoking, that if such information is provided intensively and continually, this may affect their smoking behavior. Actual assessment of the effectiveness of such programs has been extremely difficult.

Every program designed to affect behavior makes some assumption about how and why people behave as they do. It appears that governmental activity to influence smoking behavior, i.e., to convert smokers to non-smokers, and/ or to prevent non-smokers, particularly young people, from becoming smokers, is based on the following paradigm (Kuhn, 1962; March and Olsen, 1976):

1) People are all alike. In this case, the underlying assumption is that everyone holds the same motives and values for smoking and will react in the same way to a given stimulus (message).

2) People lack appropriate information (i.e., health related information). By giving people this information, they will be properly informed, and thus, behave as the new information directs them. [This is very similar to the "hierarchy-of-effect-models" published extensively in the marketing literature. See, among others, Lavidge and Steiner, (1961).]

A corollary to point 2 is as follows:

2a) The implicit assumption which may be deduced from most anti-smoking information programs seems to be that health related beliefs are the predominant factors for either stopping smoking or not starting to smoke. Consequently lack of information about the negative health consequences associated with smoking may be interpreted as the major reason why many people are still smoking.

3) Even though social aspects inherent in smoking have been stressed in some anti-smoking campaigns, the major emphasis has been on communicating about the health consequences of smoking, thus neglecting the social aspects of smoking behavior. [This partly may be attributed to the resource situation which naturally affects the activities performed and the themes to be spotlighted. We believe that the chosen activities and themes reflect the models or paradigms held by the government officials (Sheth & Mammana, 1976).]


Despite our comments in the previous section about the effectiveness of various informational programs, there is little doubt that such programs have had some effect over the last decade in reducing the incidence of smoking among adults (Fishbein, 1977). It is, however, difficult to isolate the precise effect of such activity since reduction of smoking may be attributed to other factors as well, i.e., increasing concern with physical fitness. During this same time span, ironically, the incidence of smoking among youngsters, particularly teenage girls, appears to have increased in the past decade despite the anti-smoking activities of government (Kangun and Gr°nhaug, 1978).

Perhaps, the greatest limitation to assessing the impact of anti-smoking campaigns stems from lack of understanding of the smoking decision process. As noted by Fishbein (1977), there exists little systematic theory of smoking behavior despite the vastness of the literature. More importantly, we know very little about the levels of understanding among the population, regarding both health and non-health related beliefs they hold with respect to smoking. Fishbein suggests three levels of information corresponding to three levels of beliefs, that may be germane in understanding the smoking decision process. They are as follows:

- level 1 (awareness): A person may believe that "The Surgeon General or some other public agency has determined that cigarette smoking is dangerous to health".

- level 2 (general acceptance): A person may believe that "Cigarette smoking is dangerous to health".

- level 3 (personalized acceptance): A person may believe that "My cigarette smoking is dangerous to my health".

In this review, Fishbein concludes that relatively little is known in the United States about the public's beliefs about smoking, particularly their personal beliefs, and that there is sufficient evidence to conclude that the American public is not well informed at the present time. A review of literature and empirical data from other European countries seems to show that the situation is the same in those countries (NU 1975). In sum, we know little about the accuracy of the public's awareness of governmental statements about the health consequences of smoking. Nor do we know much about the extent to which people have accepted the health consequences of smoking on a personalized level. In the remaining sections of this paper, we shall try to shed some light on these questions based on a limited and exploratory Norwegian study.


To get a better understanding of people's beliefs with respect to smoking, an exploratory study was undertaken. Lack of secondary data forced us to conduct primary research. The data were gathered by means of structured personal interviews conducted in March, 1978.

Based on national data (Statistisk Sentralbyra, 1976) the universe was defined to include all individuals 15 years and older. Because of financial and technical restrictions, however, this universe was restricted to residents in Bergen, the second largest city in Norway. In the absence of a register of the actual population, the following sampling procedure was applied. Based on a voting register of households in Bergen, 40 addresses were selected at random. Interviewers were instructed to contact five households in a specific order around each address, and interview the first household member 15 years or older that they came across., The age screening (if in doubt) was conducted by posing the following question: "How old are you?" If the respondent was less than 15 years old, the interviewer asked: "May I talk to one of your parents or another member of the household at least 15 years old?" Interviews were conducted with 136 persons. The absence of 64 interviews from the planned 200 (40 . 5 - 64 = 136) can be attributed to either no household around the selected address, appropriate respondent not at home during interviewer call and subsequent recall, or respondent unwilling to participate in the study. Five questionnaires also had to be deleted because of inadequate information. Thus, 131 respondents formed the basis of our sample.


The variable and their measurements to be reported on were smoking behavior, motives for smoking or not smoking, awareness of public statements of the health hazards associated with smoking, general beliefs, and personal beliefs.

Smoking behavior included actual smoking/non-smoking, quantity smoked, past smoking history, and future smoking intentions. Questions [In order to make comparisons, these measurements are somewhat similar to what has been used in previous research (cf. National Council on Smoking and Health 1977).] used to map this behavior are shown in Figure 1.



Previous research has uncovered several motives that may be relevant to either smoking or non-smoking behavior. Based on the research of Leventhal & Avis (1976), and Jaccard (1975), and related studies by NU (1975), National Council (1977), and Socialdepartementet (1969), the following general motives have been identified as being information in the smoking decision process - health, habit, social reward, pleasure and stimulation. These motives were utilized in this study. For each of the more specific components of these general (see Figure 2), respondents were asked to rate their perceived importance (5 = extremely important, . . . ., 1 = not at all important).



To determine the accuracy and awareness of who said what about smoking and its consequences, respondents were shown the following statements:

1. Smoking may be dangerous to your health

2. If one stops smoking, one will usually gain weight

3. The incidence of heart attacks increases if one smokes

4. Physical fitness is inconsistent with smoking.

5. Smoking may cause lung and throat cancer

6. Smoking may cause bronchitis

7. Smoking can lead to the development of ulcers.

After being shown each statement, the respondents then were asked to identify the agency making the statement. They were given a list of four agencies which are shown here:

(1) The Norwegian Tobacco Council

(2 ) The Consumer Ombudsman

(3) The Norwegian Medical Association

(4) National Council on Smoking and Health

To ascertain generalized beliefs (level 2), the respondents were given a series of statements and then asked to indicate the extent to which they agreed with these statements. Some of the statements used were as follows: It is pleasant to smoke; It is relaxing to smoke; and Smoking may cause bronchitis. With respect to identifying personal beliefs (level 3), the respondents were con- fronted with the statements similar in content to statements above, but phrased in a personal way: I find pleasure in smoking a cigarette; I smoke to relax; and If I smoke, I may get bronchitis. For each of these statements, respondents were asked the extent to which they agreed or disagreed (5 = strongly agree, . . . ., 1 = strongly disagree).

Finally some socioeconomic and demographic data was collected. Respondents were classified according to sex, age, education, and household income.


The sample consisted of 131 respondents - 70 (i.e., 53%) were smokers and 61 (i.e., 47%) were non-smokers. The proportion of smokers in this sample is somewhat higher than that found in recent national surveys (cf. National Council on Smoking and Health, 1977). With respect to age, the highest percentage of smokers were found in the 25-35 age bracket. Men were found to be more frequent smokers than women, and the incidence of smoking was found to be inversely related to the level of education attainment. All of these results are similar to findings in recent Norwegian and U.S. surveys.

When looking at the smoker sub-sample, 55 of 70 (79%) claimed that they had tried to stop smoking, and 50 of these respondents (71%) maintained that they had made serious attempts. Approximately half of those who really tried to stop smoking, claimed that they succeeded in doing so, but for the majority, this period of abstinence lasted less than three months. Of the non-smokers sub-sample, 45 of 61 (74%) had previously tried smoking. The time period for which they had smoked was denoted by a U-shaped curve indicating a significant number of both less than two-week smokers, and more than two-year smokers. These findings indicate that some individuals are capable of trying smoking without becoming permanent adopters of that behavior while some adopters are able to change their behavior.

When smokers were asked about their smoking behavior in the next 12 months, 21 (30%) claimed that they still would be smoking, 15 (21%) said they would not be smoking, and the remaining (49%) claimed they "did not know".

Compared to American findings reported by Fishbein (1977), these results reveal dramatic changes. At least in Norway, these results indicate that the militant anti-smoking policy of the government (Bjartveit 1977) has resulted in smokers giving serious consideration toward quitting. Furthermore, many smokers indicated they would reduce their level of consumption in the future. No one from the non-smoking group believed they would be smoking in the next 12 months. No significant differences were found across sex, age, educational, or income groups with regard to anticipated changes in smoking behavior.


Distributions of the various motives for smoking or not smoking are shown in Table 1. The table reveals that reasons for smoking or not smoking do vary in importance across individuals within the same group. Variations in perceived importance are similarly observed with respect to health related motives.

When looking at the various groups of generalized motives, high scores (i.e., high perceived importance) are found - as expected for the health related motives as well as for "pleasure" and "stimulation/well-being".

When looking at the importance attached to certain motives by non-smokers and smokers, some dramatic differences emerge: Regarding health related motives, nonsmokers' scores for all six motives in this category are higher than smokers scores. If we assume independence between the various health related motives, the probability of getting such a result is P(B=6)=.0156, n=6, p = 1/2. When examining the proportion of respondents in each group that assigned extreme values, significant differences occur for "taking care of health" (p <.01);"avoiding bronchitis"(p <.001, "avoiding heart problems" (p <.05), and "being in good physical condition" (p<.05). [The differences among proportions have been tested by using t-test (cf. Blalock 197, p. 228-232). The tests are two-tailed, because of the absence of explicit hypotheses.]



For habit-related motives, non-smokers do place more emphasis on "fixed habits" than do smokers (p <.02). However, smokers attach more importance having something in their mouths (p <.10).

With respect to the remaining generalized motives, the following findings emerge. Non-smokers stress social reward more than smokers.

No significant differences are found for either the pleasure or stimulation/well-being motive categories. With respect to the socio-economic characteristics in this study, no significant differences were found.


Table 2 reports on the levels of awareness among respondents in associating statements about the health hazard of smoking with the appropriate public and/or private bodies. The Consumer Ombudsman, The Norwegian Medical Association and The National Council of Smoking and Health have all declared that smoking may be dangerous to health, contribute to heart problems, cause lung and throat cancer, ulcers, bronchitis and lead to inferior physical fitness. Furthermore, the National Council of Smoking and Health has tried to inform people how to avoid putting on weight after giving up smoking. Aside from required warnings on packages, the tobacco industry has not given the public any information about the health hazards associated with smoking.

The results presented clearly indicate that a substantial portion of the public is not aware of the pronouncements by public agencies about smoking. For example, even on the statement relating smoking to lung and throat cancer, substantial proportions of both smokers and non-smokers are not aware of agency communications.

When comparing the awareness among non-smokers and smokers with respect to these statements, an interesting finding emerges. It appears that smokers are equally as aware of these statements as non-smokers. This result stands in contradiction to "dissonance theory" which seems to suggest that such statements should be avoided by smokers since they're contrary to smoking behavior. A number of explanations can be offered to explain this finding. First, smokers may have a higher tolerance for dissonance than non-smokers, and thus not avoid pronouncements about the health hazards associated with smoking. Second, smokers may either overlook the consequences of smoking, or tend to give health-related consequences less weight in the decision to smoke. And, finally, smokers may accept the consequences, but either distort the consequences through reinterpretation or distort the information.





General beliefs

Table 3 reports on general beliefs (level 2 information) related to various aspects of smoking among the respondents. Some interesting variations in the degree of agreement/disagreement occur between smokers and nonsmokers. With respect to general beliefs about the health consequences of smoking, a greater proportion of non-smokers seem to believe that smoking can lead to impaired health than do smokers. For example, significant differences between these two groups are found for statements 5 (p <.01) and 6 (p <.01). Ironically, non-smokers are more likely to express the belief that smoking cannot cause ulcers than smokers - the opposite of the expected direction.

When looking at the statements related to pleasure and stimulation/well-being, significantly more smokers express positive beliefs than do non-smokers. All the differences are significant at the .001-level. For example, significantly more smokers (46%) than non-smokers (16%) believe smoking to be pleasant. Or, almost 4/5's of the smokers agree that "smoking is relaxing" whereas less than one-third of the non-smokers agree with that statement. Interestingly, the greatest differences between these two groups lie in the pleasures-stimulation categories, yet most of the anti-smoking information campaigns are directed toward influencing health-related beliefs. Yet among smokers, these beliefs may be less important in the decision to continue to smoke than beliefs in the pleasures and stimulation areas.

When relating general beliefs to the various socioeconomic characteristics, no differences were found for the general beliefs listed in Table 3.

Personalized beliefs

Table 4 focuses on personalized beliefs related to various aspects of smoking. The interesting thing about Table 4, particularly as compared to Table 3, is that significant differences emerge between the two groups on almost all the personalized statements. Smokers generally tend to exhibit less agreement than non-smokers with respect to the consequences to their personal health of smoking. The greatest differences between these two groups in this general category are found in statements 3 (p <.02), 5 (p <.01) and 6 (p <.05). Smokers also tend to agree more extensively with importance of habit in maintenance of their smoking behavior than do non-smokers. For example, significantly more smokers agree with the statement about automatically lighting a cigarette than do non-smokers (p <.001).

The smoker and non-smoker sub-samples differ in their views on smoking and social rewards. Non-smokers were more likely to perceive the social rewards of smoking negatively than smokers, i.e., for statements 9 and 10, the differences between the two groups were p <.001 and p <.05 respectively. For personalized beliefs regarding pleasures, smokers, as one might expect, found smoking pleasant in various ways. All differences in these statements were significant at .001 level. Finally, smokers exhibited significantly more agreement about the personal well-being derived from smoking, i.e., relaxation, calming effect, energy, etc. than did non-smokers. The differences here were also significant at the .001 level.

In evaluating personalized beliefs, smokers who had tried to stop smoking were compared with those who had not tried to stop. The former exhibited more pleasure in smoking (p <.02) and were more distressed to be without cigarettes (p <.05) than the latter. They also believed that they were more accepted among friends when smoking (p <.05), found cigarettes more helpful when "blue" (p <.001), and more effective in calming them down (p <.05).



When examining the differences in personalized beliefs among smokers by daily consumption, heavy smokers (i.e., those smoking 20 or more cigarettes a day) felt it to be more unpleasant to be without cigarettes than did light smokers (p <.01). No significant differences were found when relating socio-economic characteristics to the various statements of personalized beliefs. Differences in generalized and personalized beliefs

Fishbein (1977) assumed the negative consequences of smoking to health would be less accepted by smokers at the personalized than at the generalized level. Table 5 is based on the extreme values extracted from Tables 3 and 4. As Fishbein predicted, lower values were found among smokers for all the six health-related statements at the personalized level vis-a-vis the generalized level (p <.02). For the non-smokers, a higher acceptance of the health-related statements was observed at the personalized than at the generalized level. However, the results relating smoking to ulcers was interesting. At the generalized level only 6% of the non-smokers disagreed with the statement "smoking cannot cause ulcers", while at the personalized level 45% disagreed with this statement (p <.001). This result may indicate that negative effects attributed to smoking by non-smokers partly based on relevant information, may also be generalized to other events where they lack information.



For the statements related to habitual factors and social rewards, the changes are much more dramatic for non-smokers than for the smokers. [The differences of differences (cf. Blalock 1972, p. 230-232) are all significant at the .001 level.] As shown in Table 5, non-smokers, to a lesser extent, accept smoking as habitual and social rewards from smoking at the personalized level. All the differences were significant at the .001 level. For smokers, however, no systematic changes were observed between generalized and personalized beliefs.

Turning to the statements about pleasures and stimulation/well-being, remarkable decreases in the degree of acceptance at the personalized level were found among smokers. Non-smokers, however, on these same dimensions personally found smoking less pleasurable, felt less dependent on tobacco (p <.001), found smoking less relaxing (p <.001), less helpful in overcoming depression (p <.001), and felt it did not have a calming effect (p <.001) than they generally believed. In contrast, smokers found smoking more pleasurable (p <.001) personally than they generally believed. For the other items modest, not significant, changes were observed indicating a high degree of correspondence between personalized and generalized beliefs.


The findings that emerge from this exploratory study are as follows:

1. Many smokers and non-smokers exhibit uncertainty about the risks to health associated with smoking.

2. Respondents in both groups expressed doubt on both a general and personal level about public statements regarding the health hazards associated with smoking.

3. The smoker and non-smoker sub-samples differed significantly on many of the motives, generalized beliefs and personalized beliefs associated with smoking behavior.

4. The socio-economic characteristics of our sample population were found to be less useful in revealing significant differences in smoking behavior.

Several implications may be drawn from the reported findings of this exploratory study. For the public policy makers intent on curbing the incidence of smoking, changing smokers to non-smokers will require programs with broader focus. These programs should encompass factors other than health related beliefs in seeking to influence smoking behavior. Our data show that pleasure and social consequences associated with smoking, must also be taken into account and dealt with when designing anti-smoking communication strategies.

For researchers, the present study, being exploratory in nature, does pose questions which need to be investigated more thoroughly. As noted at the outset of this paper, we need to know more about people's personal beliefs about smoking and non-smoking. Since these beliefs are most material to the smoking decision, lack of information about them is critical and raises serious questions about the success of government initiatives to discourage smoking. Emphasis also should be placed on longitudinal studies in order to map changes in perceived beliefs among relevant populations. Emphasis, of course, should be placed on tracing factors which account for the differences between smokers and non-smokers. Large scale studies of specific segments of smokers and non-smokers would also be of extreme importance in allowing us to gain a better understanding of this complex phenomenon.

A note of caution is appropriate in closing. Behavioral change, of any kind, is neither easy nor automatic. Knowing more about smoking may not result in changing the behavior of current and prospective smokers. Thus, we may know why people act as they do and be powerless to change them using morally acceptable procedures. Knowing less, however, almost certainly guarantees that most anti-smoking information programs will fail. The serious consequences from smoking, both at the individual and societal level, deserve the best of our efforts, not the least.


Bjartveit, "The Norwegian Tobacco Act," Health Education Journal, Vol. 36, No. 1, 1977 (reprint).

H. B. Blalock, Social Statistics (New York: McGraw-Hill Book Company, 1972).

Fishbein, Consumer Beliefs and Behavior with Respect to Cigarette Smoking: A Critical Analysis of the Public Literature (Staff Report of the Federal Commission, May 1977).

J. L. Hamilton, The Cigarette Advertising Ban as a Public Health Policy, Working Paper No. 5, Wayne Economic Papers (Detroit: Wayne State University, 1977).

N. Kangun and K. Gr°nhaug, Government Anti-Smoking Information Campaigns: Gaps in the Strategy of Deterrence (Bergen: Paper presented at Corporate Social Concerns and Public Policy Conference, April 2-4, 1978).

T. S. Kuehn, The Structure of Scientific Revolutions (Chicago: The University of Chicago Press, 1962).

R. C. Lavidge and G. A. Steiner, "A Model of Predictive Measurement of Advertising," Journal of Marketing, 25 (October, 1961), 59-62.

H. Leventhal and Nancy Avis, "Pleasure, Addiction and Habit: Factors in Verbal Rapport on Factors in Smoking Behavior," Journal of Abnormal Psychology, 85(1976), 478-488.

J. McGuiness and K. Cowling, "Advertising and the Aggregate Demand for Cigarettes," European Economic Review, 6(1975), 311-328.

J. G. March and J. P. Olsen, Ambiguity and Choice in Organizations (Bergen: Universitetsforlaget, 1976)

National Council on Smoking and Health, Evaluation of Anti-Smoking Programs and Activities (Oslo: 1977)

N. U., Nordiska Tobaksarbetsgruppens Betankande (Helsingfors: 1975).

J. N. Sheth and N. J. Mammana, "Recent Failures in Consumer Protection," California Management Review, Vol. XVI, Spring 1974, 64-72.

Sosialdepartementet, Om Pavirkning av R°ykevaner, St. meld. nr. 62 (1968-69).

Statistisk Sentralbyra, R°ykevaneunders°kelse 1974, Rapport fra underavdelingen for intervjuunders°kelser (Oslo: Statistisk Sentralbyra, 1976)