Journal of Applied Psychology 1982, Vol. 67, No. 5, 616-622 Copyright 1982 by the American Psychological Association, Inc. 002I-9010/82/6705-0616S00.75 ', ', and Nonsmokers' Beliefs About Cigarette Smoking Barbara Loken Department of Marketing, University of Minnesota Cognitive variables relating to both the health- and non-health-related consequences of smoking among 178 heavy, light, and nonsmoking college women were examined. Using Ajzen and Fishbein's (1980) theory of reasoned action, salient behavioral beliefs about the positive and negative outcomes of smoking, evaluations of these outcomes, salient normative beliefs about the prescriptions of important referents, and motivations to comply with these referents were selected and contrasted among the three groups of women., light, and nonsmokers were found to differ significantly in their belief structures, particularly with regard to the non-health-related factors presumed to affect smoking decisions. and nonsmokers showed the most substantial differences with regard to most of the examined components, and light smokers did not consistently represent a middle ground. smokers concurred with heavy smokers as to the negative, but not the positive, consequences of smoking, and tended to feel less antismoking pressure from certain referents than either heavy or nonsmokers. Researchers generally agree that many factors underlie a person's decision to smoke or not to smoke, For example, health, habitual, social, personality, and other psychological motives have been postulated to influence smoking behaviors (see, e.g., Cartwright & Martin, 1958; Levitt, 1971; Levitt & Edwards, 1969; Matarazzo & Saslow, 1960; Mausner & Platt, 1971). Most research has been directed toward studying only a small subset of these factors at one time. Among these influences, perhaps the most frequently studied variables pertain to the health-related consequences of smoking. Furthermore, as the health dangers of smoking have become increasingly known, health interest organizations have attempted to heighten people's awareness of these dangers. Informational antismoking campaigns have often focused on the health-related dangers of smoking, assuming that knowledge of health hazards of smoking will inhibit nonsmokers from starting to smoke, will convert smokers into nonsmokers, and/or will change I would like to thank Jean Chung for her assistance during various phases of the research project. Requests for reprints should be sent to Barbara Loken, Department of Marketing, School of Management, University of Minnesota, Minneapolis, Minnesota 55455. heavy smokers into lighter smokers. Such an approach also assumes that the target audience is not aware of the health hazards of smoking, an assumption that is not always well substantiated (see, e.g., National Institute of Education, Note 1), and that a link exists between a person's beliefs about health hazards of smoking and his or her subsequent smoking behavior. However, a person's beliefs about the health consequences of smoking need not be the major determinants of smoking behavior (see, e.g., Jaccard, 1975; Fishbein, Note 2). Other negative consequences, such as bad odors or unnecessary expense, and the many positive consequences, such as relaxation, pleasant taste, keeping weight down, or relieving nervous tension (see, e.g., Ikard & Tomkins, 1973), may also influence smoking decisions. One's overall affect toward smoking may be some combination of these positive and negative consequences. For some individuals, such as heavy smokers, the positive consequences may tend to compensate for the negative consequences of smoking. For other individuals, such as nonsmokers, the negative consequences may outweigh the positive consequences. Furthermore, normative prescriptions of important others 616
BELIEFS ABOUT CIGARETTE SMOKING 617 such as friends and family may influence smoking decisions. The purpose of the present paper is to examine college women's beliefs about both the health- and non-health-related consequences of smoking, using a systematic framework for analysis. In particular, using components identified from Fishbein and Ajzen's theory of reasoned action (Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975), which incorporates both attitudinal and normative influences, three groups of women are contrasted: heavy, light, and never-smokers. According to the theory, heavy, light, and nonsmokers will vary in their beliefs about smoking when and if they vary in their intentions to smoke. Intentions to smoke are argued to be ultimately a function of two sets of cognitive variables: (a) salient beliefs about the consequences of smoking (i.e., behavioral beliefs) and the evaluations of these consequences (i.e., outcome evaluations), mediated by one's attitude toward smoking; and/or (b) salient beliefs about the normative prescriptions of referent others (i.e., normative beliefs) and motivations to comply with those referents (i.e., motivations to comply), mediated by a more general measure of subjective norms. Therefore, if heavy, light, and nonsmokers differ in the strength of their smoking intentions, further analyses of their beliefs, evaluations, normative beliefs, and motivations to comply may provide some insights as to why some women cigarette smokers tend to smoke less than others. smoking may be regarded as one step toward quitting or, alternatively, as uniquely different from heavy and nonsmoking. College women, among whom the rate of smoking has been increasing and only recently is leveling off, were the population investigated. Method Elicitation Questionnaire Before examining the issues discussed above, a set of "model salient beliefs" about smoking cigarettes were identified with a standardized elicitation procedure (see Ajzen & Fishbein, 1980). Thirty college women, participating as partial fulfillment of a course in introductory psychology, were asked to complete five open-ended questions about the advantages and disadvantages of their smoking cigarettes, anything else they associate with smoking cigarettes, and the persons or groups of persons who would approve or disapprove of their smoking cigarettes. The most frequently mentioned positive and negative responses, along with one other behavioral consequence ("smoking helps me to concentrate") were used to form 16 behavioral beliefs and 11 normative beliefs. Main Questionnaire Subjects. Respondents were college women (independent of the sample obtained for the elicitation analysis) participating as partial fulfillment of a course in introductory psychology. Procedure. In a laboratory setting, subjects (in groups of about 30) were requested to complete a brief questionnaire about cigarette smoking. Prior to completing the questionnaire, subjects were provided detailed instructions on how to use rating scales. Measures. The main questionnaire of study included measures of 16 behavioral beliefs, 16 outcome evaluations, 11 normative beliefs, 11 motivations to comply, and behavioral intention, all measured on 7-point, bipolar scales from +3 to -3 (for model validation of these components, see Chung, 1980). Each behavioral belief (e.g., "My smoking cigarettes keeps [would keep] my weight down") and behavioral intentions ("I intend to smoke cigarettes") were measured on likely-unlikely scales. Each outcome evaluation (e.g., "Keeping my weight down") was measured on a good-bad affective scale. Normative beliefs were measured on a shouldshould not scale (e.g., "My friends think I should/should not smoke cigarettes"). Motivations to comply had the following format: "In general, I want to do/i don't want to do what my friends think I should do." Finally, respondents were asked to indicate their smoking status as either "a current smoker," "an occasional smoker," "someone who never smokes," or "an ex-smoker." Ex-smokers were subsequently eliminated from analyses of results, given their possible uniquenesses, leaving a sample size of 178: 28 current (or heavy) smokers, 27 occasional (light) smokers, and 123 never (non-) smokers. Results Occasional smokers were midway between (M = -.44) and significantly different (p <.001) from heavy smokers (M = 2.11) and never-smokers (M = -2.93) in their mean intentions to smoke. Further analyses were performed to examine and compare the behavioral beliefs, outcome evaluations, normative beliefs, and motivations to comply of heavy, light, and never-smokers. At the outset, four three-group discriminant analyses were performed to determine overall differences between heavy, light, and nonsmokers in their (a) beliefs about the outcomes of smoking cigarettes, (b) evaluations
618 BARBARA LOKEN of those outcomes, (c) normative beliefs, and (d) motivations to comply with important referents. Results indicated a significant difference between the three groups for each set of dependent variables (p <.01, for the 16 behavioral beliefs, 16 outcome evaluations, 11 normative beliefs, and 11 motivations to comply). To determine differences between groups on the individual components, oneway analyses of variance were performed, followed by Newman-Keuls contrasts of each pair of means. Behavioral Beliefs and Outcome Evaluations With respect to belief strength, results (see Table 1) show that the three groups differed significantly on all of the eight positive and six of the eight negative outcomes of smoking. Interestingly, the only nonsignificant differences pertained to health-related outcomes. In these cases, all three groups tended to already strongly agree that their smoking would be harmful to their health and would increase their chances of getting cancer. Thus, it is the non-health-related outcomes that are informative in differentiating the three groups. smokers were significantly more likely than nonsmokers to believe that their smoking would lead to all eight positive outcomes, and both heavy and light smokers were less likely than nonsmokers to believe that their smoking would lead to four of the eight negative outcomes: bad breath, bad odors on clothing, breathing problems, and offending other people. smokers also differed from nonsmokers on four positive outcomes. smokers more strongly believed that their smoking would relieve nervous tension, would be relaxing, would help them interact more easily with others, and would be a pleasant taste experience. Unlike heavy smokers, light smokers did not differ from nonsmokers in their belief that smoking would keep their weight down, would give them something to do with their hands, would help them to concentrate, and would lead to peer acceptance. Finally, heavy and light smokers significantly differed in Table 1 Results of Analyses of Variance and Newman-Keuls Tests of Behavioral Beliefs M belief strength Outcome Positive 1. Keeps my weight down 2. Relieves nervous tension 3. Helps me interact easily 4. Leads to peer acceptance 5. Something to do with hands 6. Pleasant taste experience 7. Is relaxing 8. Helps me to concentrate 75 a, b 1.50 n.43 a. b -1.25 a 1.89.J,. -96 a, b.68 a 32 a. b -.81 b.67 C -89 b, c -1.74.15 b - 93 b, c.74 C -1.19 b -1.41. -.98*0-1.84,, c -2.08 a -.07. -2.34 a, c -1.2[ a, c -1.80. 16.01*** 26.35*** 25.00*** 3.78* 12.54*** 56.39*** 36.18*** 21.34*** 1. 2. 3. 4. 5. 6. 7. 8. Causes bad breath Is harmful to my health Is expensive Increases chances of cancer Offensive to others Causes breathing problems Increases dependency on cigarettes Leaves a bad odor on clothing Negative 1.57. 2.29 2.25 2.36 1.00 a 1.57 a 2.39 1.14 a 1.48 C 2.33 2.11 2.37.70 C 1.44 C 1.56 1.78 C 2.31., c 2.63 2.64 2.64 2.12 a, c 2.17 a, c 2.24 2.43 a, c 6.01** 1.57 3.68* 1.25 16.08*** 5.04** 3.15* 12.68*** * p <.05. ** p <.01. *** p <.001.
BELIEFS ABOUT CIGARETTE SMOKING 619 beliefs about five positive outcomes of smoking. smokers were more likely than light smokers to believe that their smoking keeps their weight down, gives them something to do with their hands, helps them interact more easily in social situations, is a pleasant taste experience, and helps them to concentrate. It is noteworthy that heavy and light smokers differed in terms of positive outcomes of smoking, but, for the most part, were in agreement about the negative outcomes of smoking. With respect to outcome evaluations (see Table 2), the three groups showed significant overall effects of six negative outcomes. Individual contrasts of means showed the differences occurred primarily between heavy and nonsmokers, in that heavy smokers were less likely to negatively evaluate potential health problems (general harm to health, cancer, breathing problems), offending others, bad odors on clothing, and increased dependency on cigarettes. Thus, interestingly, heavy smokers on the average perceived "increases the chances of getting cancer" as less negative than nonsmokers. Such findings suggest that although heavy smokers may already be aware of (i.e., believe) the general health hazards of smoking, they may need to take the harmful effects of smoking more seriously (i.e., evaluate them more negatively). In this regard, perhaps providing more specific, concrete, or descriptive statements about health hazards would be useful to increase the perception of seriousness of the health consequences (see, also, Loken, Note 3). Finally, light smokers were more likely than heavy smokers to negatively evaluate increased dependency on cigarettes. Interestingly, like nonsmokers, light smokers perceived "increased dependency on cigarettes" as very unfavorable, and significantly more unfavorable than did heavy smokers. Normative Beliefs and Motivations to Comply Results of normative belief measures (see Table 3) indicate that nonsmokers were more Table 2 Results of Analyses of Variance and Newman-Keuls Tests of Outcome Evaluations M outcome evaluation Outcome Positive 1. Keeps ray weight down 2. Relieves nervous tension 3. Helps me interact easily 4. Leads to peer acceptance 5. Something to do with hands 6. Pleasant taste experience 7. Is relaxing 8. Helps me to concentrate 2.71 2.43 2.11 1.61.64 2.18 2.54 2.29 2.52 2.52 2.15 1.81.30 2.15 2.63 2.00 2.50 2.41 2.08 1.50.65 1.96 2.46 2.22.58.15.03.58.90.59.35.56 1. Causes bad breath 2. Is harmful to my health 3. Is expensive 4. Increases chances of cancer 5. Offensive to others 6. Causes breathing problems 7. Increases dependency on cigarettes 8. Leaves a bad odor on clothing Negative -2.39-2.21 a -.82-2.18 a -2.1 l a -2.39 a -2.04 a, b -1.86 a -2.63-2.52 -.96-2.37-2.48-2.52-2.63 b -2.22-2.52-2.70 a -.78-2.80 a -2.64 a -2.75 a -2.90 a -2.54 a.44 2.98*.25 4.43** 4.84** 3.61* 20.03*** 8.09*** * /)<.05. ** p <.01. *** p <.001.
620 BARBARA LOKEN Table 3 Results of Analyses of Variance and Newman-Keuls Tests of Normative Beliefs M normative belief Influence Smoking Most smokers Friends who are smokers Cigarette manufacturers Nonsmoking Most nonsmokers Friends who are nonsmokers Doctors Other Friends Mother Father Other family members Boyfriend -.39 -.32 a 2.25-2.61 b -2.18-2.61 -l.lla -2.65-2.36-1.64 a -1.86 -.26.00 C - 2.15-2.11 b.c -2.00 C -2.59-1.15 C -2.37-2.48-1.74 C -1.81 -.92 -.97«2.37-2.67 C -2.59 C -2.89-2.11 a,c -2.61-2.56-2.44 a, c -2.39 2.70 5.60**.39, 5.15** 5.07** 3.19* 11.47***.85.52 7.62*** 3.95* * p <.05. ** p <.01. *** p <.001. likely than either heavy or light smokers to perceive strong nonsmoking influences from family members (except parents) and from both smoking and nonsmoking friends. Interestingly, light smokers were less likely than either heavy or nonsmokers to perceive strong nonsmoking pressure from nonsmokers in general. Perhaps light smokers perceive less pressure because they do not see themselves as highly dependent on cigarettes (i.e., they do not include themselves in most generalizations about smokers), or because the occasions on which they smoke (e.g., at parties) are less likely to induce pressure from others. With regard to motivations to comply (see Table 4), both smokers and nonsmokers said they were in general motivated to comply with nonsmoking, but not smoking, influences. However, nonsmokers were less motivated than heavy smokers to comply with cigarette manufacturers and were more motivated than heavy smokers to comply with doctors and with nonsmokers in general. Discussion, light, and nonsmokers were found to significantly differ with respect to their underlying belief structures. Nonsmokers were the least likely to believe that smoking would lead to favorable outcomes and most likely to believe that smoking would lead to unfavorable outcomes. Furthermore, nonsmokers more negatively evaluated the negative consequences of smoking and perceived more nonsmoking normative influence than heavy smokers. smokers were between heavy and nonsmokers in their strength of beliefs about the positive outcomes of smoking, but tended to concur with heavy smokers in their beliefs about the negative consequences of smoking. In regard to both outcome evaluations and motivations to comply, light smokers again fell between heavy and nonsmokers, but they tended not to be significantly different from either heavy or nonsmokers. Finally, light smokers tended to report less antismoking pressure from family and friends than did nonsmokers and less antismoking pressure from nonsmokers in general than either heavy or nonsmokers. These results contrasted smoking subgroups of women, and are limited to the extent that they do not indicate differences between men and women. Nevertheless, the population studied is highly important, and of increasing concern to antismoking organizations. Furthermore, these results have several important implications for research
BELIEFS ABOUT CIGARETTE SMOKING 621 Table 4 Results of Analyses of Variance and Newman-Keuls Tests of Motivation to Comply M motivation to comply Influence Smoking Most smokers Friends who are smokers Cigarette manufacturers Nonsmoking Most nonsmokers Friends who are nonsmokers Doctors Other Friends Mother Father Other family members Boyfriend -1.11 -.36-1.71 a --61 a.14 1.04 a.46 1.18 1.14.79 1.43 -.74 -.22-2.22.04.11 1.74.78 1.59 1.52 1.30 1.11 -.93 -.45-2.59 a.51..79 2.04 a.59 1.58 1.51 1.25 1.30 *p<.05. **/><.01. ***/><.001..38.26 7.99*** 5.10** 3.58* 6.44**.33 1.09.79 1.40.41 on smoking decisions. First, the present findings corroborate those of other research (e.g., Chung, 1980; Jaccard, 1975; Mausner & Platt, 1971) that has found that the health hazards of smoking are not the only, or even the more important, determinants of smoking. Furthermore, the differences that existed between heavy and light smokers and between smokers and nonsmokers were generally not related to the health hazards of smoking. Thus, although the health hazards of smoking are of primary importance to many researchers and antismoking institutions, a campaign that focuses exclusively on the proposition that "smoking is dangerous to your health" may not necessarily be effective in reducing smoking intentions. Many smokers are already aware of the most common health hazards of smoking; they smoke cigarettes in spite of this awareness. Rather, heavy and light smokers appear to perceive that, for them at least, the benefits of smoking outweigh the dangers of smoking. However, it may also be important to note that less commonly known health hazards of smoking, which may not be salient to this population and, thus, may not be analyzed in the present study, may yield certain differences between groups (see, e.g., Loken, Note 3). Second, light or occasional smoking was not more likely than heavy smoking to be consistently or systematically associated with antismoking tendencies. smokers fell midway between heavy and nonsmokers with respect to some, but certainly not all, of the factors analyzed. In many respects, the smoking determinants of light smokers were unique and interesting in their own right. Thus, for example, light smokers concurred with heavy smokers as to the negative, but not the positive, consequences of smoking. Moreover, although light smokers did not always feel the antismoking pressure that heavy smokers did, they tended to have lower smoking intentions than did heavy smokers. Finally, the finding that light smokers were significantly less favorable than heavy smokers concerning the benefits of smoking suggests that beliefs about positive, rather than negative, consequences of smoking differentiate the two groups. Thus, if only the negative consequences of smoking had been examined, no potential influences on lighter smoking would have been identified. In conclusion, Fishbein and Ajzen's (1975; Ajzen & Fishbein, 1980) theory of reasoned action provided a useful guide for examining contrasts among heavy, light, and nonsmokers. This approach allows a systematic investigation of general classes of factors (e.g.,
622 BARBARA LOKEN behavioral beliefs, outcome evaluations, normative beliefs, and motivations to comply) that are presumed to affect smoking intentions, as well as the more specific content and nature of these variables. Future research on smoking could perhaps be directed toward obtaining evidence that the links between the cognitive components investigated and actual smoking behavior are, in fact, causally linked according to model specifications. Reference Notes 1. National Institute of Education. Teenage smoking: Immediate and long term patterns (Contract No. 400-79-0010). Washington, D.C.: U.S. Department of Health, Education, and Welfare, November 1979. 2. Fishbein, M. Consumers' beliefs and behavior with respect to cigarette smoking: A critical analysis of the public literature (Staff report). Washington, D.C.: Federal Trade Commission, May 1977. 3. Loken, B. Factors influencing ad ratings and memory for health warnings. Paper presented at the meeting of the American Psychological Association, Los Angeles, August 1981. References Ajzen, I., & Fishbein, M., Understanding attitudes and predicting social behavior. Englewood Cliffs, N.J.: Prentice-Hall, 1980. Cartwright, A., & Martin F. M. Some popular beliefs concerning the causes of cancer. British Medical Journal, 1958, 2, 592-594. Chung, J. Factors influencing young women's smoking intentions and behavior. Unpublished master's thesis, University of Illinois at Urbana-Champaign, 1980. Fishbein, M., & Ajzen, I. Belief, attitude, intention and behavior: An introduction to theory and research. Reading, Mass.: Addison-Wesley, 1975. Ikard, F. F., & Tomkins, S. The experience of affect as a determinant of smoking behavior: A series of validity studies. Journal of Abnormal Psychology, 1973,81, 172-181. Jaccard, J. A theoretical analysis of selected factors important to health education strategies. Health Education Monographs, 1975, 3, 152-167. Levitt, E. E. Reasons for smoking and not smoking given by school children. Journal of School Health, 1971, 41, 101-105. Levitt, E. E., & Edwards J. A. A multivariate study of correlative factors in youthful cigarette smoking. Developmental Psychology, 1969, 2, 5-11. Matarazzo, J. D., & Saslow, G. Psychological and related characteristics of smokers and nonsmokers. Psychological Bulletin, 1960, 57, 493-513. Mausner, B., & Platt, E. S. Smoking: A behavioral analysis. New York: Pergamon Press, 1971. Received March 22, 1982 Revision received April 8, 1982