Addictive Behaviors, Vol. 18, pp. 107- 116,1993 Printed in the USA. All rights reserved.
0306-4603/93 $6.00 + .OO Copyright 0 1993 Pergamon Press Ltd.
PSYCHOLOGICAL PROFILE OF COLLEGE STUDENTS WHO USE SMOKELESS TOBACCO JOHN P. FOREYT Department of Medicine, Baylor College of Medicine
ANDREW S. JACKSON Department of Health and Human Performance, University of Houston
WILLIAM G. SQUIRES, JR. Department of Biology, Texas Lutheran College
G. HARLEY HARTUNG Department of Medicine, Baylor College of Medicine
TINKER D. MURRAY Department of Physical Education, Southwest Texas State University
ANTONIO M. GOTTO, JR. Collegeof Medicine
Department of Medicine, Baylor
Abstract - The purpose of this study was to develop a psychological profile of smokeless tobacco users. We surveyed 1991 college students regarding their use of tobacco products. Twenty-one percent of the white males used smokeless tobacco compared to only 10.4% of blacks, 5.4% of Hispanics, and 5.4% of others, primarily Asians. Although the prevalence of smokeless tobacco use reported by Asians was relatively low, their rate of reported smoking was over twice as high as any other ethnic group, 43.6%. The most important reasons for beginning to use smokeless tobacco were to “see if I would enjoy it,” “most friends used it,” and “try something new.” The personality profile of the smokeless tobacco users differed from that of smokers and non-users of tobacco. College students with the highest probability of being a smokeless tobacco user were white males who scored higher in extraversion and neuroticism but lower on state anxiety than non-users.
Despite the continuing decline in the percentage of individuals who smoke cigarettes in the United States, production and use of smokeless tobacco products have at least until recently continued to grow, especially in adolescent males (Boyd & Glover, 1989; Glover, Schroeder, Henningfield, et al., 1988; U.S. Department of Health and Human Services, 1988). Total production of snuff increased 56% in the United States from 3 1.3 to 48.7 million pounds between 1970 and 1985 (Advisory Committee to the Surgeon General, 1986). Chewing tobacco has shown a similar rise, increasing 36% in the same time period, from 63.9 to 86.9 million pounds. Estimates of smokeless tobacco users in the United States range from 6 to 22 million, with the Surgeon General’s Report on Smokeless Tobacco placing the figure at approximately 12 million (Advisory Committee to the Surgeon General, 1986). There has been an especially sharp increase in the use of both snuff and chewing tobacco among young people. Lisnerski, McClary, Brown, Martin, and Jones (199 1) reported that 36% of male rural first graders in North Carolina had tried smokeless tobacco, increasing to Preparation of this paper was supported by a grant from manufacturers of smokeless tobacco. Requests for reprints should be sent to John P. Foreyt, 6535 Fannin, MS. F-700, Houston, TX 77030. 107
108
JOHN P. FOREYT et al.
70% by seventh grade. In Bogalusa, Louisiana, tobacco use was assessed in 8- to 17-yearold children and adolescents in 1976- 1977 and again in 198 l- 1982 (Hunter, Croft, Burke, Parker, Webber, & Berenson, 1986). Over this 5-year period, cigarette smoking decreased. Among white males, however, use of snuff tripled at all ages with the highest use, 32%, occurring in 12- and 13-year-olds (Hunter et al., 1986). Although more boys chewed tobacco, the rate of increase was greater for snuff. There was an eightfold increase reported in the use of snuff in 14-year-old boys (Hunter et al., 1986). Although personality variables related to smoking have been studied extensively (e.g., Cherry & Keirnan, 1978; Eysenck, 1980; Eysenck, Tarrant, Woolf, & England, 1960; McCrae, Costa, & Bosse, 1978; Parkes, 1984; Smith, 1970; Spielberger, 1986; Spielberger & Jacobs, 1982; Stanaway & Watson, 198 l), the psychological profile of smokeless tobacco users has not been as clearly defined (e.g., Botvin, Baker, Tortu, Dusenbury, & Gessula, 1989; Chassin, Presson, Sherman, & Margolis, 1988; Chassin, Presson, Sherman, McLaughlin, & Gioia, 1985; Edmundson, Glover, Alston, & Holbert, 1987; Edmundson, Glover, Holbert, Alston, & Schroeder, 1988; Glover, Holbert, Alston, White, & Edmundson, 1990; Jacobs, Neufeld, Sayers, Spielberger, & Weinberg, 1988). Edmundson, et al. (1987) and Jacobs et al. (1988) reported significant personality differences between users and non-users and suggest that a psychological profile of the smokeless tobacco user may exist. These preliminary data support the need for research to identify the factors involved in the initiation of smokeless tobacco usage (Jacobs et al., 1988). The primary purpose of this study was to determine if a multivariate personality profile could be developed that differentiated the smokeless tobacco user from the nonuser. This involved three interrelated steps: (a) To examine the gender and ethnic characteristics of smokeless tobacco users; (b) to identify the major reasons that motivate students to begin using smokeless tobacco; and (c) to develop a psychological profile of students most likely to be smokeless tobacco users.
METHOD
Subjects A total of 199 1 male and female undergraduate students at four different colleges and universities in Texas were asked to complete a self-report anonymous questionnaire. Slightly over 50% of the respondents attended a large university located in a major city. About 37% of the respondents came from a large state university located in a city of 25,000. The remaining respondents were students from two small liberal arts colleges located in rural towns. Students enrolled in courses required for graduation were asked to volunteer for the study. Those who volunteered were given the questionnaire during their regular class, asked to take it with them, complete it, and return it during the next class. They were free to either fill it out or not, because the questionnaire was anonymous and instructors did not check to see who returned it.
Questionnaire The survey, developed by the authors, was entitled “Lifestyle Questionnaire.” Students were told that it was designed to acquire general information about the lifestyles of college students. They were instructed to answer each question as truthfully as possible and were assured anonymity. The instrument was designed to elicit student demographic data and their habits, including use of tobacco products. The Eysenck Personality Inventory (EPI) (Eysenck & Eysenck, 1986) and the Spielberger State-Trait
Psychologicalprofile of smokeless tobacco users
109
Anxiety Inventory (STAI) (Spielberger, 1983) were also given to assess personality characteristics. The questionnaire took approximately 30 min to complete. Data analysis From their questionnaire responses, students were classifed into one of several groups: (a) non-users were students who were not current users and who had never smoked or used smokeless tobacco; (b) current cigarette smokers: (c)former cigarette smokers; (d) current smokeless tobacco users; (e) former smokeless tobacco users; and (f) current cigarette and smokeless tobacco users. Chi-square analysis was used to determine if tobacco use was related to gender and ethnicity. Multivariate analysis of variance (MANOVA) was used to evaluate group differences on the personality profiles from the EPI and the STAI scales. First, the personality profiles of the smokeless group were compared. Next the personality profiles of the combined smokeless group was compared with the smokers and the nonusers. Stepwise multiple logistic regression analysis was employed to develop a multivariate model for estimating the probability that a student was a regular user of smokeless tobacco. The logistic analysis was used to accomplish two research objectives: (a) To identify the psychological dimensions that separated nontobacco users from smokeless tobacco users; and (b) to define a psychological model that could be used to estimate the probability of being a smokeless tobacco user. The probability was estimated from the logistic model by: p=e -bx+a/(l + e-bx+a) where - bx + a is the regression equation defined from the logistic analysis. RESULTS Prevalence of smokeless tobacco use Of the questionnaires given out, 94% were returned, but some questionnaires could not be scored properly and were not be used. The most common reason was missing data on the personality inventories that were part of the questionnaires. There were a total of 1637 usable questionnaires (82%). Due to the large differences in sample sizes, institutional differences were not examined. This opportunistic sample represented a large, heterogeneous group of college students. The mean age of the students was 20.9 years (SD = 4.1). The gender composition of the sample was females, 54%, and males, 46%. A total of 63.7% of the students (60.1% of the males and 66.7% of the females) reported that they were not current or past users of any form of tobacco. Percentages of male and female tobacco users are presented in Figure 1. Smokeless tobacco use was related to gender (chi-square = 133.49; df = 3, p < 0.001). Less than 1.O% of the females (N = 8) reported that they had ever used any form of smokeless tobacco. In contrast, over 17% of the male students reported that they either were users of, or are presently using, smokeless tobacco. The prevalence of females who only smoked was almost 10% higher than that found for males; however, 6.1% of the males reported that they both smoked and used smokeless tobacco. Because our primary focus was to investigate factors related to smokeless tobacco use and the number of females who reported use of the products was less than 1.O%, our remaining analyses used only the male students’ data. Figure 2 provides the percentages of college males who reported some form of tobacco usage grouped by ethnicity. Chi-square analysis showed that tobacco use was related to ethnicity (chi-square = 32.96; df = 9;p < 0.0 1). A total of 109 white students
JOHN P. FOREYT et al.
110
40 .
q
(3 Females
I. Percentages of male and female tobacco users. (2 1% of total) used smokeless tobacco, compared
to only 5 blacks (10.4% of total), 4 Hispanics (5.4% of total), and 3 “others” (5.4% of total). Of the smokeless group, 23.1% were former users, 28.3% used both cigarettes and smokeless tobacco, 22.2% currently used chewing tobacco, 17.2% currently used both chewing tobacco and snuff, and 9.1% currently used snuff. Percentages of white, black, and Hispanic males who reported that they only smoked were similar, ranging from 19.2 to 2 1.O%,but the prevalence of smoking in the remaining ethnic group, primarily Asians, was nearly double, 43.6%, suggesting a high prevalence of smoking among Asian students.
Reasons for beginning use of smokeless tobacco Subjects were asked to “rank the following reasons for beginning to use smokeless tobacco in order of their importance to you. Put a ‘ 1’ next to the most important reason,
White
Black
Hispanic
Other
ETHNIC GROUP Fig. 2. Percentages of college males who reported tobacco usage grouped by ethnicity.
Psychological profile of smokeless tobacco users
Ill
Table I. Reasons smokeless tobacco users reported they started using smokeless tobacco listed by the rank of the sum of their first three choices Percentages of first three choices Reason
1st
2nd
3rd
Sum
See if 1 would enjoy it Most friends used it Try something new Thought it was satisfying because other people used tobacco Made me feel more relaxed in social situations Parents seemed to enjoy it Media advertisements Parents disapproved Other siblings enjoyed it Did not want to refuse friends
33.5 22.5 14.1 5.8 5.8 5.2 5.8 4.2 2.6 2.6
13.6 15.2 14.7 5.8 6.8 7.3 6.3 8.4 5.2 4.2
12.6 8.9 16.2 7.9 6.8 6.3 5.2 3.7 6.3 5.7
59.7 46.6 45.0 19.5 19.4 18.8 17.3 16.3 14.1 12.5
Rank : 3 4 2 s’ 9 IO
a ‘2’ at the second most important reason, etc.” Table 1 lists the reasons users reported that they started using smokeless tobacco listed by the rank of the sum of their first three choices. The most important reasons reported for starting to use smokeless tobacco were: (a) See if I would enjoy it (sum = 59.7%); (b) Most friends used it (sum = 46.6%); and (c) Try something new (sum = 45.0%). These first three reasons were clearly the most important, being selected by at least 45% of all users. The remaining seven reasons were selected by less than 20% of the users. Personality factors The personality profile of the smokeless tobacco group was compared first. MANOVA revealed that the differences in the personality profiles of the smokeless groups were within chance variation (Wilkes lambda = p > 0.05). Because there were no personality profile differences among the smokeless users, all were combined into one smokeless group. Similar analyses were done with the current and former smokers groups, and because there were no differences, they were also combined into one smokers group. MANOVA simple contrasts showed that the combined smokeless group was significantly different from both the non-users (Wilkes lambda = 0.948; p -C 0.001) and smokers (Wilkes lambda = 0.995; p < 0.001). Group means for the personality tests contrasted by the three groups and the univariate analysis of variance are provided in Table 2. In general, smokeless tobacco users tended to be more extraverted and had less state and trait anxiety then either the non-users or the smokers.
Table 2. Means, standard deviations, and univariate F ratios for the personality dimensions for tobacco use groups Reported tobacco use group None
Extraversion Neuroticism Lie score State anxiety Trait anxiety
*p < 0.01. **p < 0.05.
Smokeless
Smoke
Mean
SD
Mean
SD
Mean
SD
F ratio
13.0 10.3 3.2 39.9 41.3
3.9 4.8 I.6 10.0 8.8
12.8 II.8 3.0 40.8 42.3
3.2 4.8 1.6 10.4 8.9
14.4 11.3 2.9 37.3 39.1
3.1 6.4
6.75* 5.34* 2.82 3.82** 3.90**
i.2 a:5
JOHN P. FOREYT et al.
112
Table 3. Stepwise logistic regression equation for estimating the probability that a male college student is a regular smokeless tobacco user Variable
Logistic equation
Step entered
Improved chi-square
I
0.1415 -0.0612 0.0850 - I .93 I I
1 2 3
14.191* 10.325* 6.873*
-3.56* 3.90* - 2.60*
Extraversion State anxiety Neuroticism Constant
*p < 0.01.
Personality model for predicting smokeless tobacco users Stepwise logistic regression analysis was used to develop a personality profile model for estimating the likelihood that a male college student was a user of smokeless tobacco. Data from the non-user and smokeless groups were used for this analysis and results are provided in Table 3. The stepwise solution revealed that three of the five personality variables - extraversion, state anxiety and neuroticism - significantly enhanced the estimate of smokeless tobacco usage in college students. The logistic equation is furnished in Table 3. All coefficients were significantly greater than zero. The logistic equation provides a probability estimate for smokeless tobacco usage for any combination of the three personality variables. Compared to non-users, the male students most likely to be smokeless tobacco users tended to score high on the extraversion and neuroticism scales but low on the state anxiety scale. Figure 3 provides some examples of various combinations and the estimated probabilities. DISCUSSION
Among the nearly 2000 college students surveyed, 17% of the males reported that they had used smokeless tobacco while less than 1% of the females (all college athletes) reported that they had used the products. These results correspond to prevalence data SCALE SCORE EX NE SA 20 19 19 20 15 29 16 19 29 16 15 29
13 15 40 13 10 50 1010 50 6 6 60 16 15 39 A 0
20
40
60
80
PROBABILITY OF USING SMOKELESS TOBACCO Fig. 3. Probability of a male college student being a smokeless tobacco user for selected scores on the Extraversion (Ex), Neuroticism (Ne), and State Anxiety (SA) psychological scales.
Psychological profile of smokeless tobacco users
113
reported by researchers in other parts of the United States. The Surgeon General (Advisory Committee to the Surgeon General, 1986) estimated that 16% of the males between 12 and 25 years ofage have used smokeless tobacco during the past year, with 33 to 50% of them using the products at least once a week. Glover, Edmundson, Alston, Holbert, and Schroeder ( 1987) reported a prevalence of 19% (SE = IfI2%) and Glover, Laflin, Flannery, and Albritton (1989) reported a prevalence of 22%, both among male college students. Although our prevalence data were slightly lower, they were within chance variation. Our results agree with published data that show that the use of smokeless tobacco is primarily a behavior of white males with a prevalence of about 17 to 22%. Although our sample was opportunistically selected, it was representative of four schools in Texas. The similarity of the tobacco use of our sample and that reported in the literature suggest that our sample is representative of the college population. Based on personality differences found between male college students who used smokeless tobacco and their non-user counterparts, Edmundson et al. ( 1987, p. 68 1) portrayed the smokeless tobacco user as follows: Smokeless tobacco users were found to be more reserved and less socially outgoing than non-users. They also tend to be self-reliant, less sentimental, and see themselves as realistic. Users were also found to be more conforming when compared to nonusers, suggesting that they are susceptible to group influence but may subscribe to subcultural or small-group norms as opposed to the norms of the general culture. Overall, smokeless tobacco users appeared less personally optimistic and socially participating than non-users.
Because of the multivariate nature of personality, this psychological profile of smokeless tobacco users should be viewed with caution. Using univariate statistical models, Edmundson et al. ( 1987) reported that users and non-users were significantly different on 3 of 16 personality factors. These significant differences were used to develop their psychological profile of smokeless users. The statistical procedures used did not account for the possible intercorrelations among their 16 dependent variables and the authors failed to show that the 3 significant factors were independent. Multivariate statistical models are better used to accomplish this. Edmundson et al.‘s (1987) study involved only 55 users and, with such a small subject to variable ratio, a multivariate model will lack statistical power. The results of our study demonstrate that the psychological profile of smokeless tobacco users is different from college students who do not use tobacco of any kind or who smoke. The logistic analysis provides evidence that the personality dimensions of extraversion, state anxiety, and neuroticism separate the college students who use and do not use smokeless tobacco. The college student most likely to be a smokeless tobacco user was a white male who scored higher on extraversion and neuroticism but lower on state anxiety, compared to non-users. As shown in Figure 3, the probability of being a smokeless tobacco user was 80% if a student’s personality profile was 20 on extraversion, 19 on neuroticism, and 19 on state anxiety. Stated differently, 80% of the students in our sample with the 20- 19- 19 profile were users of smokeless tobacco. ‘As a student’s profile deviated from this extreme, the probability ofbeing a user lessened (see Figure 3). According to researchers (Eysenck & Eysenck, 1986; Spielberger, 1983) individuals with this personality profile are likely to be relatively outgoing and uninhibited, usually with low levels of worry or tension. However, when feeling provoked, they tend to be
114
JOHN P. FOREYT
et al.
somewhat impulsive and may overreact, oftentimes by losing their temper. They frequently have difficulty cooling down after such experiences. These individuals are typically described as undersocialized. The stereotype that comes to mind by the above profile is the Southern good-old-boy who tends to be easygoing, gregarious, likes to attend parties, has lots of friends, but is somewhat unpredictable and aggressive when drinking. This picture is an obvious exaggeration with many users of smokeless tobacco not fitting this profile. However, our results suggest that as these psychological characteristics became more pronounced, the probability of becoming a user was enhanced at a linear rate. This psychological profile of an outgoing, uninhibited, gregarious individual is consistent with the college students’ self-report reasons for starting to use smokeless tobacco. Of the 10 major reasons, three (“See if I would enjoy it,” “Most friends used it,” “ Try something new”) were selected by nearly half of the students as one of their first three choices. The fourth highest reason (“Thought it was satisfying because other people use tobacco”) was selected by less than 20% of the students. These three primary reasons are consistent with the multivariate profile of a smokeless user. These reasons for starting use of smokeless tobacco describe an adventurous person who is being influenced by his peer group. This psychological profile tends to agree with part of the Edmundson et al. (1987) profile that users are more conforming and susceptible to group influence, but does not agree with their portrait of a user being “more reserved and less socially outgoing than non-users.” This difference may be due to their failure to demonstrate that their statistically significant differences were independent. The multivariate model provides a personality profile of college students most likely to use or formerly use smokeless tobacco of some type during their school years. Edmundson et al. ( 1987) compared the psychological differences between snuff dippers and tobacco chewers by weekly dosage. Usage was not found to be related to personality variables. Likewise, Jacobs et al. ( 1988) reported no differences between current users, occasional users, and ex-users. In contrast, Glover et al. ( 1990) found that current users differed from ex-users on one factor of the 16-PF. Current users scored a mean of 10.98 compared to a mean of 9.31 for the ex-users on the trusting-skeptical dimension. A limitation of the Glover et al. ( 1990) study is that a multivariate test of group differences was not reported, which increases the probability that this small difference was Type I statistical error. These conflicting results indicate that the personality difference between users and former users needs further examination with sample sizes large enough to support the use of multivariate analyses. Edmundson et al. ( 1988) reported that the type of smokeless tobacco used (dip, chew, or both) was related to personality factors. Significant mean differences were reported between the three user groups and a control group of non-users on 4 of the 16 personality factors of the 16-PF. The results of the Edmundson et al. ( 1988) study need to be viewed with caution. A limitation of the study was the extreme differences in group sample sizes, which varied from 95 non-users to only 11 chewers. Due to the small sample sizes of the user groups, it was again not possible to utilize multivariate models. The significant differences reported in the study were for the global ANOVA of the four groups. Unfortunately, the authors failed to report post-hoc analyses to identify the source of the group differences. If investigators want to determine whether the type of tobacco (snuff vs. chewing tobacco) is a function of personality, much larger sample sizes will be needed to provide the statistical power needed to isolate differences. The merits of such research need to be carefully considered. We are not aware of any viable theory that might be helpful to guide such investigation. In our judgment, the more
Psychological profile of smokeless tobacco users
115
important issue is the identification of the personality profile of the smokeless tobacco user. The results of our study may be generalized to college students. Among the questions that remain to be answered, however, concern the model’s generalizability to other populations. Cross-validation of this profile on other populations would provide evidence similar to the results reported by Spielberger and Jacobs ( 1982) and others (Eysenck et al., 1960; McCrae et al., 1978; Parkes, 1984; Smith, 1970; Spielberger, 1986; Stanaway & Watson, 198 1) on smokers, suggesting that some individuals may be psychologically predisposed to the habitual use of smokeless tobacco products. Of particular interest would be the psychological profile of individuals who quit using smokeless tobacco products (DiLorenzo, Kern, & Pieper, 199 1).
REFERENCES Advisory Committee to the Surgeon General. ( 1986). The health consequences ofusing smokeless tobacco:A report of the Advisory Committee to the Surgeon General, DHHS Publication No. NIH 86-2874. Washington, DC: Government Printing Office. Botvin, G. J., Baker, E., Tortu, S., Dusenbury, L., & Gessula, J. (1989). Smokeless tobacco use among adolescents: Correlates and concurrent predictors. Journal ofDevelopmental and Behavioral Pediatrics, 10, 181-186.
Boyd, G. M., & Clover, E. D. ( 1989). Smokeless tobacco use by youth in the US. Journal of School Health, 59, 189-194.
Chassin L, Presson C. C., Sherman S. J., & Margolis S. (1988). The social image of smokeless tobacco use in three different types of teenagers. Addictive Behaviors, 13, 107-l 12. Chassin, L., Presson, C. C., Sherman, S. J., McLaughlin, L., & Gioia, D. (1985). Psychosocial correlates of adolescent smokeless tobacco use. Addictive Behaviors, lo,43 l-435. Cherry, N., & Keirnan, K. E. (I 978). A longitudinal study of smoking and personality. In R. E. Thornton (Ed.), Smoking behaviors: Physiological andpsychological influences (pp. l2- 18). Edinburgh: Churchill Livingstone. DiLorenzo, T. M., Kern, T. G., & Pieper, R. M. (1991). Treatment of smokeless tobacco use through a formalized cessation program. Behavior Therapy, 22,41-46. Edmundson, E. W., Clover, E. D., Alston, P. P., & Holbert, D. (I 987). Personality traits ofsmokeless tobacco users and non-users: A comparison. International Journal of the Addictions, 22,67 I-683. Edmundson, E. W., Clover, E. D., Holbert, D., Alston, P. P., & Schroeder, K. L. (1988). Personality profiles associated with smokeless tobacco use patterns. Addictive Behaviors, 13,2 19-223. Eysenck, H. J. (1980). The causes and effects of smoking. Beverly Hills, CA: Sage Publications. Eysenck, H. J., & Eysenck, S. B. G. (1986).Manual for the Eysenck Personality Inventory. San Diego, CA: Educational and Industrial Testing Service. Eysenck, H. J., Tarrant, M., Woolf, M., & England, L. (1960). Smoking and personality. British Medical Journal, 1, 1456-1460. Clover, E. D., Edmundson, E. W., Alston, P. P., Holbert, D., & Schroeder, K. L. (1987). Prevalence and patterns of smokeless tobacco use in a southeastern university. Journal ofDrug Education, 17,20 l-2 11. Clover, E. D., Holbert, D., Alston, P. P., White, D. M., & Edmundson, E. W. (1990). Personalitiesofcurrent users and quitters of smokeless tobacco. Health Values, 14,42-45. Clover, E. D., Laflin, M., Flannery, D., & Albritton, D. L. (I 989). Smokeless tobacco use among American college students. Journal QfAmerican College Health, 38,8 l-85. Clover, E. D., Schroeder, K. L., Henningfield, J. E., et al. (1988). An interpretative review of smokeless tobacco research in the United States: Part I. Journal ofDrug Education, 18,285-3 10. Hunter, S. M., Croft, J. B., Burke, G. L., Parker, F. C., Webber, L. S., & Berenson, G. S. (1986). Longitudinal patterns of cigarette smoking and smokeless tobacco use in youth: The Bogalusa Heart Study. American Journal of Public Health, 76, 193- 195. Jacobs, G. A., Neufeld, V. A., Sayers, S., Spielberger, C. D., & Weinberg, H. ( 1988). Personality and smokeless tobacco use. Addictive Behaviors, 13,3 1 l-3 18. Lisnerski, D. D., McClary, C. L., Brown, T. L., Martin, J. P., &Jones, D. R. (1991). Demographic and predictive correlates of smokeless tobacco use in elementary school children. American Journal ef Health Promotion, 5,426-43
I.
McCrae, R. R., Costa, R. T., & Bosse, R. (1978). Anxiety, extraversion and smoking. British JournalqfSocial and Clinical Psychology, 17,269-273. Parkes, K. R. ( 1984). Smoking and the Eysenck personality dimensions: An interactive model. Psychological Medicine, 14, 825-834.
116
JOHN P. FOREYT et al.
Smith, G. M. (1970). Personality and smoking: A review of empirical literature. In W. G. Hunt (Ed.), Learning mechanisms and smoking (pp. 42-6 1). Chicago: Aldine. Spielberger, C. D. (1983). Manualfor the State-TraitAnxiety Inventory.Form Y. Palo Alto, CA: Consulting Psychologists Press, Inc. Spielberger, C. D. (1986). Psychological determinants of smoking behavior. In R. D. Tollison (Ed.), Smoking and society: Toward a more balanced assessment (pp. 89-134). Lexington, MA: D.C. Heath & Company. Spielberger, C. D., & Jacobs, G. A. (1982). Personality and smoking behavior. Journal ofPersonalityAssessment, 44,396-403. Stanaway, R. G., & Watson, D. F. W. (198 1). Smoking and personality: A factorial study. British Journal of Social and Clinical Psychology,20,2 13-2 14. U.S. Department of Health and Human Services. (1988). Nicotine addiction. The health consequences of smoking: A report of the Surgeon General. Washington, DC: Superintendent of Documents.