Adolescent Nonsmokers, Triers, and Regular Smokers' Estimates of Cigarette Smoking Prevalence: When Do Overestimations Occur and by Whom?1

Adolescent Nonsmokers, Triers, and Regular Smokers’ Estimates of Cigarette Smoking Prevalence: When Do

Overestimations Occur and by Whom?’

STEVE SUSSMAN, CLYDE W. DENT, JILL MESTEL-RAUCH,

BRIAN R. HA? c. ANDERSON JOHNSON, WILLIAM B. HANSEN, AND

University of Southern California

The present study tested three assumptions often made by smoking prevention researchers and practitioners regarding the nature and causes of smoking prevalence estimates. Specifically, it is often assumed that (a) adolescent prevalence estimates of all levels of smoking are inflated, (b) youths at different stages of smoking onset all overestimate smoking prevalence, and (c) the predictors of these estimates are the same for youths at different stages of smoking onset. Measures of smoking behavior based on self-reports using a bogus pipeline procedure, estimates of cigarette smoking prevalence (trying behavior and weekly use), and reports of smoking in significant others were obtained by the same youths when they were in eighth and ninth grades. Results indicated that, taken over all youths, estimates of trying cigarettes were accurate, whereas estimates of weekly use were inflated. When responders were defined by their stage of smoking onset, either as a nonsmoker, trier, or regular smoker, analyses revealed that triers were accurate regarding prevalence estimates of trying, whereas nonsmokers made gross underestimations and regular smokers made overestimations of trying. All groups of youths overestimated prevalence of weekly smoking. Analyses of variance revealed that both youths’ smoking status and in- terpersonal smoking environment additively predicted prevalence estimates, and estimates of smoking prevalence were relatively higher in youths who began to smoke, or increased their levels of smoking, within the next year. Thus, previous assumptions were partially confirmed. Future research and prevention program- ming should focus more on regular smoking as the focus of normative smoking distortion effects and correction efforts.

The research reported in this article was supported by grants from the AMA-GLAA, AM-LA, and NCI (R18-CA34622) to B.R. Flay, W.B. Hansen, and C.A. Johnson.

Requests for reprints should be sent to Steve Sussman, Institute for Health Promotion and Osease Prevention Research, University of Southern California, 35 N. Lake Avenue, Suite 200, Pasadena, CA 91101. Brian R. Flay is now at Prevention Research Center, School of Public Health, University of Illinois at Chicago.

537

Journal of Applied Social Psychology, 1988, 18, 7, pp. 537-551. Copyright 0 1988 by V.H. Winston 8 Sons, Inc. All rights reserved.

538 STEVE SUSMAN ET AL.

Several researchers have reported data suggesting that youths per- ceive cigarette smoking as more common than it really is (e.g., American Cancer Society, 1976; Fishbein, 1977; Johnson, 1982; Sherman, Presson, Chassin, Corty, & Olshavsky, 1983). For example, one large sample of midwestern and southwestern adolescents estimated that 40% of teenagers smoke and 60% of adults smoke (Sherman et al., 1983). In reality, only about 20% of American teenagers (American Cancer Society, 1976; Johnson, 1982; Johnston, O’Malley, & Bachman, 1984) and about 35% of American adults (Cullen, Wheeler, Brown, Gritz & Flushman, 1981; Sherman et al., 1983) actually do smoke regularly.

Further empirical evidence indicates that normative overestimates of smoking predicts smoking status, smoking onset, and increases in smoking. In cross-sectional studies, adolescent smokers estimate there to be more adolescents who smoke than do adolescent nonsmokers (Chassin et al., 1981; Presson et al., 1984; Sherman et al., 1983). Two longitudinal in- vestigations found that general smoking prevalence estimates were among the strongest predictors of future smoking onset and increases (Chassin, Presson, Sherman, Corty, & Olshavsky, 1984; Collins et al., 1987).

Fishbein (1977) stated that misinformation of this type could lead to beliefs such as ”my smoking will make me more popular with my friends” and, eventually, to initiation of smoking. Adolescent smoking prevention programs have included intervention components that try to instill accurate smoking prevalence norms, presumably to eliminate potential influence of a general overestimation of smoking prevalence (e.g., Botvin & Eng, 1982; Evans et al., 1978; Flay et al., 1985; Hurd et al., 1980). In theory, by presenting accurate smoking norms and thereby af- fecting a decrease in subjective smoking popularity, the probability of future smoking should decrease. This theory is intuitively appealing and intrinsically logical.

However, smoking prevention researchers and practitioners have apparently operated on at lest three untested assumptions regarding the nature and causes of smoking prevalence estimates. First, prevention program researchers and practitioners have often assumed that all levels of smoking behavior are overestimated by youths. Yet, previous smoking prevalence estimation studies have not investigated judgments made of both trying behavior and regular smoking. Rather, general smoking prevalence questionnaire items were asked in previous studies such as ”Out of every 100 boys in your school, how many of them smoke cigarettes?” (Sherman et al., 1983). The smoking onset process is a stochastic one; 80% to 90% of youngsters have tried at least one cigarette, yet the

JUDGEMENT ACCURACY OF SMOKING PREVALENCE ESTIMATES 539

proportion of youngsters who smoke regularly rarely exceeds 50% of those who have tried (e.g., Palmer, 1970). Thus, in previous studies youths have had to make their own assumptions regarding whether questionnaire items were requesting estimates of trying behavior or of regular smoking.

A second assumption that smoking prevention program researchers and practitioners have often made is that youths at different stages of smoking onset (e.g., triers, regular smokers) all overestimate smoking prevalence. Sherman et al. (1983) reported that both adolescent smokers and nonsmokers (two groups) overestimated smoking prevalence, but, as mentioned before, the items they used did not anchor judgments of prevalence by level of smoking. If youths thought they were estimating trying behavior, both groups of youths actually made underestiinations.

A third assumption often made is that the causes of smoking prevalence estimates are the same for youths at different levels of smoking development. Yet, different influences may affect perceptions and behaviors of youths at different stages of smoking development (Flay, d’Avernas, Best, Kersell, & Ryan, 1983; Leventhal & Cleary, 1980). Non- smokers and triers may be primarily influenced by family and peer models, whereas regular smokers may be primarily influenced by smoking-related experiences and physiological effects of smoking (Flay et al., 1983).

The present study attempted to test these three assumptions to achieve a more complete understanding of the nature of estimates of smoking prevalence. Adolescent nonsmokers, triers, and regular smokers made estimates of the number of peers (out of 100) who have tried smoking and smoke regularly. In addition, they made general estimates of adult smoking prevalence.

Assumption 1: All levels of smoking behavior are overestimated. Three com- peting hypotheses were tested regarding responses made to the peer smoking prevalence items (averaged across all subjects). First, the low ambiguity of the items could have led to accurate estimates of percentage of triers and regular smokers. Second, a selective attention process could have resulted in a smoking overestimation bias across items, even though the items requested specific information. Clearly, cigarette smoking in young teenagers tends to draw attention by others. Individuals who draw attention to themselves may be overrepresented in the sampling of cases, resulting in biased encoding of prevalence information and inflated recall estimates (Crocker, 1981). Third, youths may have viewed smoking as an “all-or-none” phenomenon. Individuals could have accurately encoded instances of trying behavior but, perhaps due to an

540 STEVE SUSSMAN ET AL.

implicit belief that most people who try smoking remain smokers, may have made inflated estimates regarding prevalence of regular smoking.

Assumption 2: Youths at different stages of smoking development all overestimate smoking preualence. In addition to testing between different hypotheses regarding an overall estimation bias, accuracy of estimates made by regular smokers, triers, and nonsmokers were explored. We wanted to examine whether all groups of youths would overestimate prevalence of smoking, in line with that assumption made in current prevention programs.

Assumption 3: The causes of smoking prevalence estimates are the same for youths at different levels of smoking development. Smoking status and the smoking environment (i.e., reported smoking in significant others) were employed as predictors of smoking prevalence estimates. Following from current knowledge of the stages of smoking development, it was hypothesized that exposure to smoking others would be a more important predictor than smoking status of the smoking estimates made by nonsmokers or triers. On the other hand, smoking status would be a more important predictor than exposure to smoking others of estimates made by regular smokers.

Future smoking. Finally, smoking prevalence estimates made by eighth-grade nonsmokers who either did or did not try cigarettes by ninth grade, and estimates made by eighth-grade triers who either did or did not become regular smokers by ninth grade, were calculated to examine whether relatively inflated estimates would be made by those who began or increased their levels of smoking.

Method

Subjects and Questionnaire

A total of 5,610 students from the Los Angeles metropolitan area, who were participating in a longitudinal study of smoking behavior, com- pleted self-report questionnaires in eighth and ninth grades. These grades were selected because much smoking onset occurs during these years of school (Flay et al., 1983). Forty-nine percent of the students were males; 60% of the students were white, 26% were Hispanic, 7% were black, and 7% were Asian. They were assessed in May 1983 and again in May 1984. On each occasion, they indicated their current smoking status, which was validated using a ”bogus pipeline” procedure (Evans, Han- sen, & Mittelmark, 1977).


Smoking status was assessed with two self-report items, one requesting students to indicate whether they had ever had even one puff on a cigarette and the other requesting how much they currently smoked. Stu- dents who reported never having had even one puff of a cigarette were classified as "nonsmokers." Students who reported having had at least one puff on a cigarette, but having smoked only a few times over the last 12 months, wen? classified as triers." Finally, students who reported currently smoking at least one cigarette a week were classified as "regular smokers."

The questionnaire included three smoking prevalence items, which asked "Out of every 100 students your age, how many do you think have tried smoking cigarettes?", "Out of every 100 students your age, how many do you think smoke cigarettes at least once a week?", and "Out of every 100 adults how many do you think smoke cigarettes?" Only one general item addressed adult smoking because levels of smoking are better descriptive of adolescent smoking development behavior than of adult smoking. (Adults tend to be regular smokers or nonsmokers; few are merely "triers.") Responses were given on 5-point scales that ranged from "20 or less" to "between 81 and 100" in intervals of 20. Subjects were also asked to indicate how many of their good friends have tried smoking, how many of their good friends smoke at least once a week, and how many adults they know well that smoke.

Results

Assumption 1: All levels of smoking behavior are overestimated. Taken over all subjects, prevalence estimates regarding number of regular adolescent smokers and adult smokers were greatly inflated (Table 1). However, estimates regarding the number of adolescents who have tried smoking were remarkably accurate! These results were found both at eighth and ninth grades.

Assumption2: Youths at different stages of smoking deoelopment all overestimate smoking prevalence. These prevalence items were re-examined by subject smoking group (Figure 1). Triers were fairly accurate in their estimates of number of triers (a 3% overestimate was made at eighth and ninth grades), whereas nonsmokers grossly underestimated number of triers (by 12% at eighth grade and 11% at ninth grade) and regular smokers overestimated number of triers (by 9% at eighth grade and 6% at ninth grade). These data were quite surprising. Nonsmokers view trying cigarettes as less prevalent than is actually the case (Zs = 22.3 and 14.6 in eighth and ninth grades, respectively, p < .OOOl). On the other hand, all

90

80

70

w

60

z w 2 50

n

P

w4

0

+ I

F 30

w a 3

20

10 0

ES

TIM

ATE

S O

F S

MO

KIN

G P

RE

VA

LEN

CE

: B

Y S

MO

KIN

G S

TATU

S A

T G

RA

DE

EIG

HT

- 74

- 68

TR

Y

55

RE

GU

LA

R

TA

RG

ET

GR

OU

P

70

AD

UL

T

ul

P

h3

v) <

;;I rn

rn

-I >

!-

AC

TU

AL

N

ON

-SM

OK

ER

T

RIE

RS

R

EG

UL

AR

Figu

re 7

.


I I 1 I I I I I 1


Table 1

Actual and Estimated Prevalence of Smoking Behaviors at Eighth and Ninth Grades

Student Student Adults Tried Regular Regular

Grade Eighth Ninth Eighth Ninth Eighth Ninth

Actual 65 70 9 12 33 33 Estimated 63 69 37 46 63 62

Difference -2 -1 28 34 30 29 Z-score -1.8 -1.7 88.6* 85.6* 103.T 75.8'

~

eP < .oooi, af = 5ooo+. Note: Actual rates of student smoking are based on the self-report srnok- ing behavior items in the student survey. Adult actual rates in the Los Angeles county area are taken from Cullen, Wheeler, Brown, Gritz and Flushman (1981).

three smoking groups overestimated number of regular smokers, with regular smokers giving the most inflated estimates and nonsmokers giving the least inflated estimates.

Assumption 3: The causes of smoking preualence estimates are the same for youths at different levels of smoking development. To test the third assumption, the relative effects of selective exposure and smoking status on smoking prevalence estimates were examined. Six analyses of variance were calculated, three within each grade. Subject smoking status (i.e., nonsmoker, trier, and regular smoker) and one of each selective exposure item (i.e., number of friends who have tried smoking, number of friends who smoke regularly, or number of familiar adults who smoke) were the predictors. Selective exposure responses were classified as "none," "one," or "two or more." One of each respective estimate of smoking prevalence item (i-e., number of teenagers who have tried smoking, number of teenagers who smoke regularly, or number of adults who smoke) was the dependent variable. In all analysis of variance models, both smoking status and selective exposure predicted estimates of smoking prevalence (Table 2). The higher the level of smoking, and the greater the number of


Table 2

Prevalence Estimates and Analysis of Variance Results: Selective Expositre and Smoking Status Predictors

~~ ~ ~

Estimates of Adult Use

Familiar adults who smoke

Smoking Grade 8 status None One Two+ Total

Nonsmoker 56.4 64.0 71.5 63.3 (n=1876)

Trier 59.2 66.4 71.7 65.2 (n=3231)

Regular 62.8 68.9 75.2 69.9 (n=503)

Total 59.5 66.4 72.8 63.3

F(2,5000+) Selective exposure 84 .Oo* Smoking status 9.92+

Grade 9 None One Two+ Total

56.2 62.5 68.9 60.4 (n= 1629)

59.6 64.9 70.1 64.0 (n=3300)

62.4 64.2 71.0 65.9 (n=681)

59.4 63.9 70.0 62.1

64.53* 7.03*

~~ ~~ ~

Estimates of Peers Try

# Good friends who have tried smoking

Smoking Grade 8 status None One Two+ Total

Nonsmoker 45.2 48.2 60.0 53.1 Trier 52.1 59.9 70.6 67.9 Regular x x 74.9 74.2

Total 48.6 54.0 68.5 63.0

F(2,5000+) Selective exposure 25.59+ Smoking status 61.45'

Grade 9 None One Two+ Total

49.7 57.7 64.9 59.4 60.0 64.8 75.3 73.2

X 68.5 76.1 76.0

54.8 63.7 72.1 69.3

67.75* 54.62*


Nonsmoker 30.1 35.3 43.1 33.2 Trier 34.1 37.7 45.3 39.8 Regular 35.3 44.2 56.4 55.0

Table 2 (Continued)

38.5 45.7 50.0 41.7 43.9 44.8 54.0 47.6 45.1 48.2 58.9 54.7

Estimates of Peers Regular

# Good friends who smoke regularly

Smoking Grade 8 Grade 9 status None One Two+ Total None One Two+ Total I

Total 33.2 39.1 48.3 37.1 1 42.5 46.2 54.3 46.4

F(2,5000+)

Smoking status 12.22' Selective exposurc 54.56' 4 9 . w

8.57'

Notes. + means "or more"; * means p < .001; X means cell size was under 20, so such relatively unstable estimates were not analyzed.

significant others who smoke, the greater the prevalence estimate made. There were no significant interactions of selective exposure by smoking status (all ps > .l), and selective exposure was generally the stronger predictor of prevalence estimates.

Future smoking. Finally, smoking prevalence estimates made by eighth-grade nonsmokers who either did or did not try cigarettes by ninth grade, and estimates made by eighth-grade triers who either did or did not become regular smokers by ninth grade, were calculated to examine whether relatively inflated estimates would be made by those who subsequently began or increased their levels of smoking. Z-tests confirmed the expectation that relatively inflated estimates would be made by those more predisposed to smoke in the next year (Table 3).

Discussion

To summarize the results, the first and second smoking prevention program assumptions were disconfirmed, whereas the third assumption was confirmed. Regarding the first assumption, we found that youths, as


Table 3

Prevalence Estimates of Eighth-Grade Nonsmokers Who Do or Do Not Begin Smoking by Ninth Grade

Ninth- grade status

Estimated Estimated Estimated % students % students % adults

-tried- -regular- -regular-

Nonsmokers 50 (N= 1573) Smokers 58 (n=309)

30 58

35 63

~~

Difference

Z-score

8

3.63*

5 ~

5

2.85* 2.89*

Prevalence Estimates of Eighth Grade Triers Who Do or Do Not Become Regular Smokers by Ninth Grade

Ninth- grade status

Estimated Estimated Estimated % students % students % adults

-tried- -regular- -regular-

Triers (N=3053) Regular smokers (n=195)

67

73

~~

38 64

42 68

Difference 6 4 4

Z-score 2.81* 2.74* 2.10*

* p < .05.


a group, accurately estimate smoking trying behavior but overestimate regular smoking. Apparently, youths are capable of making accurate prevalence estimates of trying behavior but not of regular smoking. Thus, of the “low ambiguity,” ”selective attcntion,” and “all-or-none” hypotheses, the ”all-or-none” hypotheses was best supported. Youths may view smoking as a primarily all-or-none proposition. This viewpoint is quite understandable. Youths probably have been exposed to smokers who state that they could not go without cigarettes for a day. Such s t a t e ments are the hallmark of smoking as a problem behavior. On the other hand, probably few youths have witnessed the gradual and stochastic nature of the smoking onset process. The fact that the prevalence estimates of regular smoking were found to be lower than that of trying behavior suggests that youths are aware that not all triers become regular smokers.

Regarding the second assumption, we found that nonsmokers under- estimate prevalence of trying behavior. How could such estimates, in theory, predispose nonsmokers to begin smoking? In fact, gross underes- timates of smoking prevalence would logically tend to negatively influence attitudes regarding the social popularity of smoking and inhibit smoking onset. Consequently, smoking prevention programs may do best not to instill accurate norms regarding trying cigarettes if, by doing so, smoking onset would increase. All smoking groups did overestimate the number of adolescent regular smokers and adult smokers. Perhaps, prevention program efforts should concentrate on presenting accurate norms regarding only regular smoking.

Regarding the third assumption, we found that both selective exposure to others’ and one’s own smoking behavior additively predicts smoking estimates made by youths at all stages of smoking development. In theory, because estimates of the commonness of a behavior are generally based on those instances most readily available in memory (Kahneman & Tversky, 1973), the association with smoking others and one‘s own smoking behavioral instances will tend to lead one to estimate smoking prevalence as relatively common. The predictors of smoking prevalence were the same for youths at all stages of smoking development. Treatment programs might be able to better impart accurate norms by helping youths at all stages of smoking development conceptually separate their own behavior and smoking environment from that of larger smoking environmental contexts.

In addition, we found that inflated estimates, relative to one’s stage of smoking development, were significantly associated with future onset or


relative to one’s smoking status is associated with whether one will become a regular smoker, whereas the accuracy of one’s estimates per se has less predictive significance. For example, all nonsmokers underestimated smoking prevalence; however, nonsmokers who began to smoke within the next year reported higher estimates than did the remaining nonsmokers.

Conclusion

Johnson (1982) and McCaul and Glasgow (1985) provided discussions of untested and erroneous assumptions regarding components of smoking prevention treatment programs. They argued that, although program components are theory-derived and (taken together) greatly inhibit cigarette smoking onset, a paucity of empirical analyses exist regarding the efficacy of different program components. Rather, too great a reliance has been placed on plausible but untested assumptions. The present study provides an empirical analysis of theoretical assumptions underlying one of the smoking prevention program components, namely the discussion of people’s tendency to overestimate smoking prevalence and the presentation of accurate smoking norms to correct such misperceptions. The fact that prevalence estimates of smoking are among the best psychosocial predictors of smoking onset and increases (Chassin et al., 1984; Collins et al., 1987) highlights the importance of empirical research in this area. The present results indicate that youths do not overestimate all stages of smoking onset. Rather, relatively inflated judgments of smoking prevalence, given a youth’s smoking status, best predicts smoking onset and increases in smoking. Finally, all youths do overestimate the prevalence of teenage and adult regular smoking. The provision of accurate norms regarding regular smoking by adolescents and adults might be extremely beneficial to prevention efforts.

References

American Cancer Society. (1976). A study of public school teachers cigarette smoking attitudes and habits. New York: American Cancer Society.

Botvin, G., & Eng, A. (1982). The efficacy of a multicomponent approach to the prevention of cigarette smoking. Preoentive Medicine, 11, 199-211.

550 STEVESUSSMAN ETAL

Chassin, L., Presson, C.C., Bensenberg, M., Corty, E., Olshavsky, R.W., & Sherman, S.J. (1981). Predicting adolescents’ intentions to smoke cigarettes. Journal of Health and Social Behavior, 22,445455.

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, 14,224-243.

Collins, L., Sussman, S., Mestel-Rauch, J., Dent, C.W., Johnson, C.A., Hansen, B.W., & Flay, B.R. (1987). Psychosocial predictors of young adolescent cigarette smoking: A sixteen-month, three-wave longitudinal study. Journal of Applied Social Psychology, 17,554-573.

Crocker, J. (1981). Judgement of covariation by social perceivers. Psychological Bulletin, 90,272-291.

Cullen, J.W., Wheeler, R.C., Brown, M.P., Gritz, E.R., & Flushman, H.O. (1981). Preventing lung cancer in the Los Angeles Basin: Lung cancer prevention workshops. Los Angeles: Division of Cancer Control, UCLA Jonsson Comprehensive Cancer Center.

Evans, R., Hansen, W.B., & Mittelmark, M.B. (1977). Increasing the validity of self-reports of behavior in a smoking in children investigation. Journal of Applied Psychology, 62,521-523.

Evans, R.I., Rozelle, R.M., Mittelmark, M.B., Hansen, W.B., Bane, A.L., & Havis, J. (1978). Deterring the onset of smoking in children: Knowledge of immediate physiological effects and coping with peer pressure, media pmssure, and parent modeling. Journal of Applied Psychology, 8,126-135.

Fishbein, M. (1977). Consumer beliefs and behavior with respect to cigarette smoking: A critical analysis of the public literature. In Federal Trade Commission, Report to Congress: Pursuant to the Public Health Cigarette Smoking Act for the year 1976 (pp. 00-00). Washington, DC: U.S. Government Printing Office.

Flay, B.R., d’Avernas, J.R., Best, J.A., Kersell, M.W., & Ryan, K.B. (1983). Cigarette smoking: Why young people do it and ways of preventing it. In P. McGrath & P. Firestone (Eds.), Pediatric and Adolescent Behavioral Medicine (pp. 132-283). New York: Springer-Verlag.

Flay, B.R., Ryan, K.B., Best, J.A., Brown, K.S., Kersell, M.W., d’Avemas, J.R., & Zanna, M.P. (1985). Are social psychological smoking prevention programs effective?: The Waterloo study. Iournal of Behavioral Medicine, 8,37-59.

Hurd, P.D., Johnson, C.A., Pechacek, T., Bast, L.P., Jacobs, D.R., & Luepker, R.V. (1980). Prevention of cigarette smoking in seventh grade students. Journal of Behavioral Medicine, 3,15-28.


Johnson, C.A. (1982). Untested and erroneous assumptions underlying antismoking programs. In T.J. Coates, A.C. Petersen, & C. Perry (Eds.), Promoting adolescent health: A dialog on research and practice (pp. 149-166). New York: Academic Press.

Johnston, L.D., OMalley, P.M., & Bachman, J.G. (1984). Highlights from drugs and American high school students 1975-1983. Rockville, MD: National Institute on Drug Abuse.

Kahneman, D., & Tversky, A. (1973). On the psychology of prediction. Psychological Review, So, 237-251.

Leventhal, H., & Cleary, P.D. (1980). The smoking problem: A review of the research and theory in behavioral risk modification. PsychologicaZ Bulletin, 88,370-405.

McCaul, K.D., & Glasgow, R.E. (1985). Preventing adolescent smoking: What have we learned about treatment construct validity? Health

Palmer, A.B. (1970). Some variables contributing to the onset of cigarette smoking in junior high school students. Social Science and Medicine, 4, 359-366.

Presson, C.C., Chassin, L., Sherman, S.J., Olshavsky, R., Bensenberg, M., & Corty, E. (1984). Predictors of intentions to smoke: Age, sex, race, and regional differences. The International Journal of the Addictions, 19, 503-519.

Sherman, S.J., Presson, C.C., Chassin, L., Corty, E., & Olshavsky, R. (1983). The false consensus effect in estimates of smoking prevalence: Underlying mechanisms. Personality and Social Psychology Bulletin, 9,

Psychology, 4,361-387.

197-207.

Adolescent Nonsmokers, Triers, and Regular Smokers' Estimates of Cigarette Smoking Prevalence: When Do Overestimations Occur and by Whom?1

Documents