A COMPARATIVE ANALYSIS OF TEENAGERS WHO SMOKE DIFFERENT CIGARETTE BRANDS.

Enomoto, Carl E.

ABSTRACT

This study analyzed and compared the survey responses of teenagers who smoke different cigarette brands. It was found that teen Marlboro and Camel smokers perceived themselves as having more stress in their lives as compared with teen Newport smokers. On average, Marlboro smokers were depressed or sad more often than were Camel or Newport smokers. Camel smokers were generally less interested in a peer stop-smoking program than were Newport or Marlboro smokers. Finally, these teen smokers, regardless of cigarette brand, had similar opinions about quitting. Given the differences across brands, more flexible approaches may be needed to address teenage smoking.

According to the American Heart Association (1998), there are approximately 4.4 million teenagers aged 12 to 17 who smoke. Three thousand teenagers begin to smoke every day. Furthermore, "the Department of Health and Human Services estimates that 90 percent of smokers begin tobacco use before age 20; 50 percent of smokers begin tobacco use by age 14; and 25 percent begin their smoking addiction by age 12" (American Heart Association, 1998).

Much has been written on the factors that influence teenage smoking. Peer influence, parental influence, advertising, and number of persons in the household who smoke are just a few of the factors that have been analyzed. The results, as well as conclusions about the best way to reduce teen smoking, have been varied. With few exceptions, these studies have dichotomized teenagers into smokers and nonsmokers. Herein may lie the problem-teenage smokers are not necessarily a homogeneous group. While it is true that differences in smoking behavior between males and females, and white and nonwhite individuals, have been investigated, few studies have examined teenagers who smoke different brands of cigarettes. if, for example, Marlboro smokers have attitudes that are different from Camel or Newport smokers, then research that groups them together will produce results that are misleading. Furthermore, the implication is that stop-smoking programs may need to be designed for specific types of teen smokers.

REVIEW OF THE LITERATURE

Why would there be differences in the attitudes and beliefs of teenagers who smoke different brands? It is not uncommon for firms to target a particular market for their product, the result being that consumers of that product are somewhat similar. As an example, people who subscribe to Reader's Digest are likely to have attitudes and interests that differ from those who subscribe to Forbes Magazine. In the case of cigarettes, it has been argued that Joe Camel advertisements targeted very young smokers. The cowboy images used by Marlboro are believed to be aimed at individuals who see themselves as rugged and strong. Thus, different types of teens may be drawn to the various brands of cigarettes because of the images associated with those brands (or some other reason, such as what their friends smoke), and their attitudes and concerns may vary accordingly. The purpose of the present study was to examine the attitudes and concerns of teen smokers who were classified by cigarette brand.

Krosnick and Judd (1982) explored the theory that peer influence on adolescent smoking behavior increased with age, while the influence of parents decreased. Using LISREL, they concluded that the influence of peers during adolescence did increase, but that the influence of parents did not change.

Stanton, Currie, Oei, and Silva (1996) found that peer influence declined during late adolescence. Logistic regression revealed that having a close friend who smoked influenced the smoking behavior of 15year-olds but not 18-year-olds.

Conrad, Flay, and Hill (1992) presented a comprehensive review of studies that analyzed the predictors of smoking behavior in children. The important explanatory variables included socioeconomic status, peer pressure, knowledge, attitudes, and self-esteem.

Chassin, Presson, Sherman, and Edwards (1990) examined the percentages of adolescents who became young adult smokers. These researchers classified adolescent smokers into four groups: those who never smoked, those who tried it, ex-smokers, and regular smokers. The percentages for the four groups were calculated in 1980, 1981, 1982, and 1983. A chi-square test for homogeneity indicated that there was no significant variation in the relative responses across time for the four groups.

Glynn, Greenwald, Mills, and Manley (1993) concluded that the most important factors influencing adolescent smoking include the following: having parents who smoke, having friends who smoke, doing poorly in school, and advertising. They also outlined several possible solutions to the teenage smoking problem, such as involving the family of the teen smoker, physicians, and counselors in efforts to discourage tobacco use, along with implementing school programs.

Chassin, Presson, Sherman, Corty, and Olshavsky (1984) conducted a longitudinal study to predict the start of adolescent smoking. Using multiple regression, they found that intention to smoke, attitudes, beliefs, academic values, relationships with friends and parents, and the smoking behavior of friends, siblings, and parents were important predictors.

Hirschnian, Leventhal and Glynn (1984) investigated second through tenth graders to determine what influenced their progression from trying one cigarette to becoming a regular smoker. Important variables included risk-taking behavior, changing schools, meeting expectations of parents, and having friends who smoked.

Wang, Fitzhugh, Westerfield, and Eddy (1995) used logistic regression to determine the factors associated with adolescent (ages 14 to 18) smoking behavior. Same-gender effects were prominent. Specifically, the strongest predictor of smoking for adolescent males was smoking by best male friend. For adolescent females, the strongest predictor was smoking by best female friend.

Byrne, Byrne, and Reinhart (1995) used principal components analysis to examine the relation between stress and smoking in adolescence. Stress from school, stress from family, and uncertainty about the future were related to smoking.

Najem, Batuman, Smith, and Feuerman (1997) investigated smoking among inner-city teenagers. Using logistic regression, they found that African-American adolescents were less likely to smoke as compared with whites and Hispanics. Boys and girls were equally likely to smoke. Relative to nonsmokers, smokers knew less about the dangers associated with smoking.

Nagel, Mayton, and Walner (1995) analyzed differences in values between adolescents who used tobacco and those who did not. They found that those who did not use tobacco were more concerned about world peace and having a sense of accomplishment. In addition, nonusers were less interested in material comforts, pleasure, and having an exciting life.

Cavin and Pierce (1996), in one of the few studies to look at choice of cigarette brand, included adult and adolescent smokers. Using logistic regression, they found that women, low-income individuals, whites, and rural smokers were more likely to smoke generic brands. For adolescents, Marlboro was the most popular brand, followed by Camel. There was very little interest in generic cigarettes among adolescent boys and girls.

The above are just a few of the studies that have examined the predictors of adolescent smoking behavior. Almost every conceivable explanatory variable has, at one time or another, been investigated. Notably, the one area that has not been fully explored is choice of cigarette brand. Gavin and Pierce (1996) looked at brand selection and found that various groups have different preferences. However, it is not known whether there are differences in concerns, attitudes, and opinions (e.g., regarding stop-smoking programs) among those who select particular brands.

METHOD

The data used in this study came from a Gallup poll taken in 1992 (provided by the Inter-University Consortium for Political and Social Research, ICPSR 6252; Robert Bezilla, principal investigator). The nationwide survey used a telephone sampling procedure (random digit dialing) targeting persons aged 12 to 17 living in the United States (Bezilla, 1994). The respondents were asked about their smoking behavior, age, attitudes, brand of cigarettes purchased, reasons for smoking, amount of peer and family pressure, and other demographic and smoking-related topics. The sample contained 1,125 adolescents (51.2% males, 48.8% females). Further, 83% were white and 12.7% were black. When asked if they had smoked a cigarette in the past 30 days, 22.2% responded that they had and 77.8% that they had not. Of those who had smoked in the past 30 days, 135 bought Marlboro cigarettes the last time they purchased cigarettes for themselves, 38 bought Camel cigarettes, and 20 bought Newport cigarettes. The remaining brands wer e bought by 4 or fewer adolescents, with the exception of generic cigarettes (6 adolescents). This study concentrated on the responses of teens who purchased the three major brands: Marlboro, Camel, and Newport.

Survey responses were tabulated for the Marlboro, Camel, and Newport smokers. A pairwise comparison (chi-square test) was then performed to see if teen smokers of two of the three brands had the same relative response rates. Pearson's chi-square statistic was employed (it is commonly used to test for homogeneity of groups). [1]

RESULTS

Response categories for the first two items were (1) strongly agree, (2) agree, (3) disagree, (4) strongly disagree, (5) do not know, and (6) refused to answer. Categories 5 and 6 were infrequent and thus dropped from the analysis.

The first survey item that was analyzed was, "You have a lot of stress in your life." The response rates of the 135 Marlboro smokers, 38 Camel smokers, and 20 Newport smokers are shown in Figure 1.

A comparison of the response rates of Marlboro and Newport smokers yielded a chi-square statistic of 10.80 with a p value of .013. [2] Thus, at a 2% significance level, the two groups of smokers differed in regard to perceived stress. Many Marlboro smokers (43.7%) agreed with the statement that they have a lot of stress in their lives, while 40.0% of the Newport smokers disagreed. With responses coded 1 for strongly agree, 2 for agree, 3 for disagree, and 4 for strongly disagree, the mean score for Marlboro smokers was 2.24, while the mean score for Newport smokers was 2.85, indicating that the teen Marlboro smokers felt they experienced more stress.

A comparison of the response rates of Camel and Newport smokers yielded a chi-square statistic of 10.05 with a p value of .018. [3] Thus, at a 2% level of significance, the Camel and Newport smokers differed in terms of perceived stress. The mean score for Camel smokers was 2.18, while the mean score for Newport smokers was, as previously noted, 2.85, indicating that Newport smokers felt less stressed than did Camel smokers.

Finally, a comparison of Marlboro and Camel smokers resulted in a chi-square of 5.24 with a p value of .16. Thus, there was no significant difference in the responses of these two groups at the 10% probability level.

The second survey item that was analyzed was, "You often get depressed or feel sad." The response rates are shown in Figure 2.

When comparing the response rates of Marlboro and Newport smokers, a chi-square statistic of 14.12 with a p value of .003 was obtained. [4] This indicates that their responses were significantly different at the 1% level. Even though large percentages of both Marlboro and Newport smokers agreed that they are often depressed or sad, the mean score for Marlboro smokers was 2.38, while that of Newport smokers was 2.85. On average, Marlboro smokers were depressed or sad more often than were Newport smokers, which is consistent with the previous finding for perceived stress.

A comparison of the response rates of Camel and Newport smokers resulted in a chi-square statistic of 4.38 with a p value of .223. [5] At a 10% significance level, the responses of the two groups of smokers were indistinguishable.

Comparing the responses of Marlboro and Camel smokers resulted in a chi-square of 12.86 with a p value of .005. [6] At a 1% significance level, the responses of these groups were different. Many of the Marlboro smokers (46.7%) agreed with the statement that they often get depressed or feel sad, while 57.9% of the Camel smokers disagreed. The mean response of Marlboro smokers was, as previously noted, 2.38, while that of Camel smokers was 2.84.

Next, responses to the following question were analyzed: "How much interest would you have in participating in a stop-smoking program that was designed for people your own age?" Response categories were (1) great deal of interest, (2) some interest, (3) little interest, and (4) no interest at all. The response rates of the teen smokers by cigarette brand are presented in Figure 3.

A comparison of the response rates of Marlboro and Newport smokers yielded a chi-square statistic of 5.019 with a p value of .170. [7] Thus, at a 5% significance level, the two groups did not differ. With responses coded 1 for great interest, 2 for some interest, 3 for little interest, and 4 for no interest, the mean score for Marlboro smokers was 2.81, while the mean score for Newport smokers was 2.70. This indicates that, on average, Marlboro smokers were slightly (but not significantly) less interested in a peer stop-smoking program.

Comparing the responses of Camel and Newport smokers resulted in a chi-square statistic of 9.70 with a p value of .021. [8] Thus, their responses were significantly different at the 3% level. The mean score for Camel smokers was 3.31, which exceeded the mean score for Newport smokers (2.70). It appears that the Camel smokers were far less interested in a peer stop-smoking program than were Newport smokers.

The comparison of Marlboro and Camel smokers yielded a chi-square of 9.12 with a p value of .028. Their responses were significantly different at a 3% significance level, with the average Camel smoker less interested in a peer stop-smoking program as compared with the average Marlboro smoker.

Finally, responses to the following question were examined: "All things considered, would you like to give up smoking?" Response categories were yes, no, and do not know. Response rates for the three groups of teen smokers are reported in Figure 4 (only yes/no responses are presented).

A comparison of the response rates of Marlboro and Newport smokers resulted in a chi-square of 2.60 with a p value of .107. At a 5% level of significance, their responses were not different. There were similar results for the remaining two comparisons. That is, the response rates of Camel and Newport smokers and Camel and Marlboro smokers were not significantly different.

DISCUSSION

This study examined teenagers who smoke the top three brands of cigarettes: Marlboro, Camel, and Newport. There were several interesting findings: (1) Teen Marlboro and Newport smokers differed significantly in regard to the amount of stress they believed they had in their lives. Newport smokers, on average, felt they experienced less stress. (2) Camel and Newport smokers also differed regarding stress. Again Newport smokers, on average, perceived less stress. (3) Marlboro and Camel smokers had similar beliefs concerning stress levels. (4) Teen Marlboro smokers, on average, were more often depressed or sad as compared with teen Newport and Camel smokers. (5) Camel and Newport smokers were found to be similar in terms of how often they felt depressed or sad. (6) Camel smokers, compared with Newport and Marlboro smokers, were generally less interested in a peer stop-smoking program. (7) Marlboro and Newport smokers had similar attitudes concerning such a program. Fifty percent of Newport smokers and 40.2% of Marlboro smokers either had some or a great deal of interest in a peer stop-smoking program. (8) All three groups of teen smokers had similar opinions about quitting. Seventy-one percent of the Marlboro smokers, 71.9% of the Camel smokers, and 52.6% of the Newport smokers reported that they would like to quit.

What are the reasons for these differences? One possibility is that it is likely that teen smokers know others who smoke Marlboro cigarettes (the leading brand for teens and adults), and perhaps that is what got them started in the first place. Those who smoke Marlboro cigarettes may be more inclined to follow what others are doing, desiring to be part of the crowd. Yet, there is also the image of the rugged, strong individual that is associated with the Marlboro brand, which some of these smokers may be trying to emulate. Perhaps those who share these conflicting traits experience greater depression and stress. The Newport smoker, on the other hand, may tend to be less influenced by advertising and what the crowd does. Newport does seem to be the least advertised brand. One question in the survey asked teen smokers and nonsmokers which cigarette brand they had seen advertised most. The 1,125 survey respondents listed Camel first, followed by Marlboro, with Newport last. Thus, Newport smokers maybe a bit mor e independent and self-assured, traits linked to a happier, more stress-free life.

Camel smokers were least interested in a peer stop-smoking program. It has been argued that the Joe Camel advertising campaign attempted to attract a younger audience, although this was not borne out in this study. (The average age of the Marlboro smokers was 15.95 years; Camel smokers, 15.97 years; and Newport smokers, 16 years.) If, however, those smoking Camels find the youthful image portrayed in Camel ads appealing (compared with the Marlboro cowboy or the image of the "sophisticated" Newport smoker), then Camel smokers may have a younger mind-set. This may be related to a belief that they are more immune from the dangers of smoking and thus, in the present study, showed little interest in a peer stop-smoking program. Perhaps they also believe that they have more time before they must stop smoking to reverse damage caused by cigarettes.

It is interesting to note that Cavin and Pierce (1996) found little interest in generic cigarettes among adolescent boys and girls. This may confirm that, for many teens, image associated with cigarettes is important. This attraction to different images through brand identification may enable the separation of teens into different attitudinal categories.

The findings are somewhat encouraging in that a large percentage of the teen smokers indicated that they would like to quit. If, however, teen smokers are not a homogeneous group with similar preferences and attitudes, then stop-smoking programs will have to be designed to meet the needs of this diverse population. If some smokers experience more depression and stress in their lives compared with others, then there may be more serious problems that need attention (e.g., through counseling) before cigarette smoking can be addressed.

This study was funded, in part, by a grant from the College of Business Administration and Economics at New Mexico State University; however, the author is solely responsible for its content.

Reprint requests to Carl E. Enomoto, Department of Economics, New Mexico State University, Las Cruces, New Mexico, 88003. Electronic mail may be sent via Internet to cenomoto@nmsu.edu.

ENDNOTES

(1.) The chi-square statistic used for the test of homogeneity is calculated as follows: [[chi].sup.2] = [sigma][([f.sub.o] - [f.sub.e]).sup.2]/[f.sub.e] where [f.sub.e] is the expected frequency and [f.sub.o] is the observed frequency.

(2.) One of the conditions that should be met when using the chi-square test is that the expected frequency of each response equal 5 or more. If this is not the case, then the chi-square test may not be a valid one. A common way of overcoming this problem is to combine two or more responses. In comparing the Marlboro and Newport smokers, at least one of the expected frequencies was less than 5. To see if the conclusion would be altered, the strongly agree and agree responses were combined. This resulted in a chi-square of 10.28 with a p value of .006. Thus, the conclusion that the responses of the two groups were different remained valid.

(3.) In comparing Camel and Newport smokers, at least one of the expected frequencies was less than 5. To see if the conclusion would be altered, the disagree and strongly disagree responses were combined. This resulted in a chi-square of 4.737 with a p value of .094. Thus, the conclusion that the responses of the two groups were different remained valid at the 10% probability level.

(4.) In comparing Marlboro and Newport smokers, at least one of the expected frequencies was less than 5. To see if the conclusion would be altered, the agree and strongly agree responses were combined. This resulted in a chi-square of 13.998 with a p value of .001. Thus, the conclusion that the responses of the two groups were different remained valid.

(5.) In comparing Camel and Newport smokers, at least one of the expected frequencies was less than 5. To see if the conclusion would be altered, the agree and strongly agree responses were combined. This resulted in a chi-square of 4.182 with a p value of .124. Thus, the conclusion that the responses of the two groups were the same remained valid.

(6.) In comparing Marlboro and Camel smokers, at least one of the expected frequencies was less than 5. To see if the conclusion would be altered, the agree and strongly agree responses were combined. This resulted in a chi-square of 12.847 with a p value of .002. Thus, the conclusion that the responses of the two groups were different remained valid.

(7.) In comparing Marlboro and Newport smokers, at least one of the expected frequencies was less than 5. To see if the conclusion would be altered, the some interest and great interest responses were combined. This resulted in a chi-square of 4.730 and a p value of .094. Thus, the conclusion that the responses of the two groups were not different remained valid at the 5% probability level.

(8.) In comparing Camel and Newport smokers, at least one of the expected frequencies was less than 5. To see if the conclusion would be altered, the some interest and great interest responses were combined. This resulted in a chi-square of 9.259 with a p value of .010. Thus, the conclusion that the responses of the two groups were different remained valid.

REFERENCES

American Heart Association. (1998). Statistics available from the American Heart Association: http://www.americanheart.org/Heart__and__Stroke__A__Z__Guide/tobta.ht ml.

Bauman, K. E., & Chenoweth, R. L. (1984). The relationship between the consequences adolescents expect from smoking and their behavior: A factor analysis with panel data. Journal of Applied Social Psychology, 14(1), 28--41.

Bezila, R. (1994). Teenage attitudes and behavior concerning tobacco (June--July 1992; United States) [Computer file]. Princeton, NJ: The George H. Gallup International Institute [producer]; Ann Arbor, MI: Inter-University Consortium for Political and Social Research [distributor].

Byrne, D. G., Byrne, A. E., & Reinhart, M. I. (1995). Personality, stress and the decision to commence cigarette smoking in adolescence. Journal of Psychosomatic Research, 39(1), 53--62.

Cavin, S. W., & Pierce, J. P. (1996). Low-cost cigarettes and smoking behavior in California, 1990--1993. American Journal of Preventive Medicine, 12(1), 17--21.

Chassin, L., Presson, C. C., Sherman, S. J., Corty, E., & Olshavsky, R. W. (1984). Predicting the onset of cigarette smoking in adolescents: A longitudinal study. Journal of Applied Social Psychology, 14(3), 224--243.

Chassin, L., Presson, C. C., Sherman, S. J.,. & Edwards, D. A. (1990). The natural history of cigarette smoking: Predicting young-adult smoking outcomes from adolescent smoking patterns. Health Psychology, 9(6), 701--716.

Conrad, K. M., Flay, B. R., & Hill, D. (1992). Why children start smoking cigarettes: Predictors of onset. British Journal of Addiction, 87, 1711--1724.

Cutler, T., & Nye, D. (1997). Subjects and accomplices: Regulation and the ethics of cigarette advertising. International Journal of Health Services, 27(2), 329--346.

Eiser, J. R., & Van Der Pligt, J. (1984). Attitudinal and social factors in adolescent smoking: In search of peer group influence. Journal of Applied Social Psychology, 14, 348--363.

Glynn, T. J., Greenwald, P., Mills, S., & Manley, M. W. (1993). Youth tobacco use in the United States--Problem, progress, goals, and potential solutions. Preventive Medicine, 22, 568--575.

Hirschman, R. S., Leventhal, H., & Glynn, K. (1984). The development of smoking behavior: Conceptualization and supportive cross-sectional survey data. Journal of Applied Social Psychology, 14(3), 184--206.

Kaplan, M. S., & Weiler, R. E. (1997). Social patterns of smoking behavior: Trends and practice implications. Health and Social Work, 22(1), 47--52.

Kauffman, S. E., Silver, P., & Poulin, J. (1997). Gender differences in attitudes toward alcohol, tobacco, and other drugs. Social Work, 42(3), 231--241.

Krosnick, J. A., & Judd, C. M. (1982). Transitions in social influence at adolescence: Who induces cigarette smoking? Developmental Psychology, 18(3), 359--368.

McCaul, K. D., Glasgow, R., O'Neill, H. K., Freeborn, V., & Rump, B. S. (1982). Predicting adolescent smoking. Journal of School Health, 52(6), 342--346.

Nagel, L., Mayton, D. M., & Walner, T. (1995). Value differences across tobacco use levels. Health Values, 19(6), 39-44.

Nagel, L., McDougall, D., & Granby, C. (1996). Students' self-reported substance use by grade level and gender. Journal of Drug Education, 26(1), 49-56.

Najem, G. R., Batuman, F., Smith, A. M., & Feuerman, M. (1997). Patterns of smoking among inner-city teenagers: Smoking has a pediatric age of onset. Journal of Adolescent Health, 20, 226-231.

Rogers, R. G., Nam, C. B., & Hummer, R. A. (1995). Demographic and socioeconomic links to cigarette smoking. Social Biology, 42(1-2), 1-21.

Stanton, W. R., Currie, G. D., Oei, T. P. S., & Silva, P. A. (1996). A developmental approach to influences on adolescents' smoking and quitting. Journal of Applied Developmental Psychology, 17, 307-319.

Wang, M. Q., Fitzhugh, E. C., Eddy, J. M., Fu, G., & Turner, L. (1997). Social influences on adolescents' smoking progress: A longitudinal analysis. American Journal of Health Behavior, 21(2), 111-117.

Wang, M. Q., Fitzhugh, E. C., Westerfield, R. C., & Eddy, J. (1995). Family and peer influences on smoking behavior among American adolescents: An age trend. Journal of Adolescent Health, 16, 200-203.