Pierce et al

Copyright 1996 by the American Psychological Association, Inc.
0278-6133/96/S3.00

Health Psychology
19%, Vol. 15, No. 5, 355-361

Validation of Susceptibility as a Predictor of Which Adolescents
Take Up Smoking in the United States
John P. Pierce, Won S. Choi, Elizabeth A. Gilpin,

Robert K. Merritt

and Arthur J. Farkas
University of California, San Diego

Centers for Disease Control and Prevention

Smoking onset has 4 levels, with a "susceptibility" level preceding early experimentation. This
study assessed the predictive validity of smoking susceptibility in a longitudinal study of a nationally
representative sample of 4,500 adolescents who at baseline reported never having puffed on a
cigarette. At follow-up 4 years later, 40% of the sample had experimented with smoking, and 8%
had established a smoking habit. Baseline susceptibility to smoking, defined as the absence of a firm
decision not to smoke, was a stronger independent predictor of experimentation than the presence
of smokers among either family or the best friend network. However, susceptibility to smoking was
not as important as exposure to smokers in distinguishing adolescents who progressed to
established smoking from those who remained experimenters at follow-up.
Key words: smoking initiation, adolescents, exposure to smoking, susceptibility to smoking

Preventing the onset of smoking has been identified as
essential to the public health priority of achieving a rapid
reduction in smoking prevalence (Institute of Medicine, 1994;
U.S. Department of Health and Human Services [USDHHS],
1989, 1994). Since the early 1950s, the proportion of adults
who begin to smoke has declined markedly (Gilpin, Lee, &
Pierce, 1994), so that by 1990, the onset of smoking could be
defined as mainly an adolescent behavior (Centers for Disease
Control and Prevention [CDC], 1991; Lee, Gilpin, & Pierce,
1993; USDHHS, 1994). Although adolescent smoking rates
declined throughout the 1970s and early 1980s, there is little
evidence for any decline since 1985 (Johnston, O'Malley, &
Bachman, 1993; Nelson et al., 1995). However, there has been
a significant decline in the smoking prevalence among Black
adolescents, resulting in a widening of the Black-White
difference in smoking prevalence (Nelson et al., 1995). Because lack of progress has been concurrent with some of the
most extensive public health programs against tobacco (Bal,
Kizer, Felten, Mozar, & Niemeyer, 1990; Community Intervention Trial for Smoking Cessation Research Group, 1995;
Pierce et al., 1994), the Institute of Medicine (1994) has led the
call for research to develop a greater understanding of the
natural history of adolescent addiction to nicotine.
Generally, scholars agree that the onset of smoking is a

John P. Pierce, Won S. Choi, Elizabeth A. Gilpin, and Arthur J.
Farkas, Cancer Prevention and Control Program, Cancer Center,
University of California, San Diego; Robert K. Merritt, Office on
Smoking and Health, Centers for Disease Control and Prevention,
Atlanta, Georgia.
Preparation of this article was supported by funding from the
Robert Wood Johnson Foundation. This work was done during the
tenure of John P. Pierce's established investigatorship from the
American Heart Association.
Correspondence concerning this article should be addressed to John
P. Pierce, Cancer Prevention and Control Program, Cancer Center,
University of California San Diego, La Jolla, California 92093-0901.

355

time-dependent four-level process that includes (a) a preparation period, (b) early experimentation, (c) more advanced
regular but nondaily smoking, and (d) a stable level of
addiction (Elder et al., 1990; Flay, d'Avernas, Best, Kersell, &
Ryan, 1983; Leventhal & Cleary, 1980; Schinke & Gilchrist,
1983; Stern, Prochaska, Velicer, & Elder, 1987; USDHHS,
1994). Progress through these levels is presumed to take at
least 2 years, although the natural history of the smoking onset
process has never been adequately defined (USDHHS, 1994).
Most measures of onset behavior rely on the report of recent
smoking behavior. The preferred measure of adolescent smoking is the report of smoking within the past 30 days (USDHHS,
1994). Experimentation is generally inferred from responses to
questions about ever smoking or age of first cigarette; however,
there is no generally accepted method of identifying adolescents prior to experimentation. As most smoking prevention
programs aim to prevent experimentation (USDHHS, 1994),
such a measure is needed to facilitate both their design and
evaluation. Smoking prevention programs should either prevent target groups from becoming susceptible to smoking or
prevent susceptible adolescents from progressing to experimentation.
Having developed an algorithm for this preexperimentation
phase of the smoking onset process (Pierce et al., 1993; Pierce,
Farkas, Evans, & Gilpin, 1995), we hypothesized that such a
measure should identify which "never smokers" are cognitively
predisposed to smoking. It should include the participants'
intentions (Ajzen & Fishbein, 1980; Conrad, Flay, & Hill,
1992; McNeil! et al., 1988; Sussman, Dent, Flay, Hansen, &
Johnson, 1987) and expectations (Bandura, 1977; Bandura,
1986; Bauman, Fisher, Bryan, & Chenoweth, 1984) for future
behavior. Demonstrating preliminary evidence of the validity
of this preexperimentation measure, we noted an urgent need
to test the predictive validity of this algorithm in a longitudinal
study.
We used a nationally representative longitudinal study of

356

PIERCE, CHOI, GILPIN, FARKAS, AND MERRITT

adolescents to further evaluate the predictive validity of our
measure of preexperi mental ion on a sample who reported that
they were never smokers (or puffers) over 4 years. We
examined the appropriateness of the susceptibility to smoking
algorithm in this population and tested its ability to predict the
next level in the smoking onset process (i.e., experimentation),
taking into account other sociodemographic predictors of
future smoking and exposure to other smokers. We also
addressed the probability that never smokers progress through
the onset process in the 4-year time period to the level denned
as established smoking (having consumed at least 100 cigarettes, current or former smokers).
Method
Participants
The Teenage Attitudes and Practices Survey (TAPS) was designed
to provide information on adolescent smoking behavior and was
developed under the direction of the National Center for Health
Statistics and the Office on Smoking and Health, Centers for Disease
Control and Prevention (Allen, Moss, Giovino, Shopland, & Pierce,
1992). The TAPS interviewed adolescents who had responded to the
1989 National Health Interview Survey (NH1S), an annual household
interview survey of the civilian, noninstitutionalized population of the
United States. TAPS I was conducted in 1989, and the follow-up,
TAPS II, was conducted in 1993.
Nine thousand nine hundred sixty-five adolescents aged 12-18 years
were interviewed by TAPS 1 in 1989 by either telephone or mail
questionnaire. This represented a response rate of 82% of the original
sample of 12,097 adolescents who had responded to the NH1S in 1989.
Of the 9,965 adolescents in TAPS I, only the respondents reached by
telephone (N = 9,135) were eligible for follow-up in 1993. The 830
adolescents who were not reached by telephone responded through
mail questionnaire and were not eligible for the follow-up in 1993.
The follow-up telephone survey, TAPS II, was completed in 1993 by
87% of the eligible TAPS I respondents (N = 7,960), who by that time
were aged 15 to 22 years (CDC, 1994). This article focuses on the 4,500
participants from the longitudinal sample who, at baseline, reported
never having experimented with smoking.

Measures of Smoking Onset
In both surveys we identified current smokers with the standard
question, "Think about the last 30 days. On how many of these days
did you smoke?" Experimentation with cigarettes was defined as a
positive response to either of two questions: "Have you ever smoked a
cigarette?" and "Have you ever tried or experimented with cigarette
smoking, even a few puffs?" Two negative responses required classifying a respondent as a never smoker. Only those respondents so
classified according to 1989 data were included in this analysis. A
positive response to the any cigarette question led to an additional
question about whether the individual had ever smoked 100 cigarettes.
In 1993, we classified respondents who reported having smoked 100
cigarettes as meeting the criteria for established smoking (either
current or former smokers).
The survey included all three questions on the algorithm (Pierce et
al., 1995) for classifying a respondent as susceptible to smoking,
although one question had slightly different wording from that of the
previously published algorithm. To be classified as not susceptible to
smoking, a respondent had to answer "no" to the question, "Do you
think that you will try a cigarette soon?" and "Definitely not" to the
questions, "If one of your best friends were to offer you a cigarette,

would you smoke it?" and "Do you think you will be smoking cigarettes
1 year from now?"
In our published question set, this last question was worded, "Do
you think that you will smoke a cigarette in the next year?" Rewording
appeared to reduce by about 7% the proportion of never smokers
classified as susceptible. We obtained this estimate by analyzing the
1990 and 1993 Youth Attitudes and Practices Survey section of the
California Tobacco Surveys, which contained the two different versions of the "smoking in the next year" question.

Other Predictors of Smoking Status
Sociodemographic information obtained for this study included date
of birth, gender, and race or ethnicity (non-Hispanic White, Hispanic,
Black, Asian, or other). The TAPS sample was drawn from households
that had recently completed the NHIS, making adult-provided NHIS
data available on family income and the educational attainment of the
responsible adult who gave permission for the adolescent to be
interviewed. Respondents reported family income in 27 income
categories (from < $1,000 to $50,000+) that for the purposes of this
analysis were reduced to four categories (< $16,000, $16,000-$29,999,
$30,000-$49,999, and $50,000+). Because of the association between
school performance and smoking prior to completion of formal
education (Parrel & Fuchs, 1982), we analyzed how respondents
thought they were performing at school, compared with the average
student. We include this reported relative performance.
Exposure to smokers in the social network is a strong and consistent
predictor of smoking initiation (Ary & Biglan, 1988; Bauman et al.,
1984; Best, Thompson, Santi, Smith, & Brown, 1988; Flay et al., 1983).
Both adolescent surveys sought detailed smoking status for each older
member of the household and immediate family members not living at
home. Respondents also reported the smoking status for each of their
designated four best male and four best female friends. Prior detailed
analysis of these exposure data led to the development of a single
four-level variable that reflected minimal exposure to smoking (i.e., no
exposure from family or best friends), exposure through family
members only, exposure through best friends only, and exposure
through both family and peer networks (Pierce et al., 1995).

Statistical Analysis
The NHIS uses a multistage sample design to provide national
estimates of the civilian, noninstitutionalized population. It is a
complex sample design that involves both clustering and stratification.
The stratification variables were race (Black and non-Black), sex, and
age categories (10-14 years, 15-17 years, 18-19 years, and 20-22
years). The multistage NHIS sample design requires a Taylor series
approximation to estimate variance on the basis of the NHIS weighting
procedures (Allen et al., 1991). We used the SUDAAN program for all
statistical analyses (Research Triangle Institute, 1989). All percentages were weighted and adjusted for sampling design and nonresponse.
We used two nested logistic regressions to identify which specified
variables predicted any change in use or experimentation with cigarettes in the interval between the two surveys and which predicted
established smoking at follow-up. Both logistic regressions included
only adolescents who were never smokers in 1989. All percentages
(experimentation and established smoking) are based on the initial
cohort of 4,500 never smokers at baseline.
The first logistic regression model attempted to separate adolescents who had experimented with smoking (both experimenters and
established smokers; N = 1,796) from adolescents who remained
never smokers (N - 2,704). The independent variables included age,
sex, race or ethnicity, perceived school performance, adult education

357

VALIDATION OF SUSCEPTIBILITY TO SMOKING
level, household income, exposure to other smokers, and susceptibility
to smoking. AH independent variables were receded and entered as
categorical variables with appropriate design variables, which used
referent groups within each independent variable. We also examined
two-way interactions between susceptibility to smoking and all predictor variables.
The second logistic regression was performed to predict those
adolescents who progressed to established smoking (N = 351) versus
adolescents who remained experimenters (N = 1,445) at follow-up.
This analysis used the same independent variables as those in the first
logistic regression and tested similar two-way interactions with the
susceptibility to smoking variable.
For both logistic regressions, we forced all demographic variables
(age, sex, race or ethnicity, adult education level, and household
income) into the model and examined the significance of the following
variables: susceptibility to smoking, exposure to other smokers, and
perceived school performance. Odds ratios and 95% confidence
intervals were based on standard errors derived from the SUDAAN
procedures.

Results
The Susceptibility to Smoking Measure
A strong univariate association was observed between each
potential susceptibility to smoking question and later smoking
behavior (Table 1). Approximately 17% of adolescent never
smokers who in 1989 thought that they would try a cigarette
soon had become established smokers by 1993, compared with
only 7% of those who did not think that they would try a
cigarette soon (p = .0006). Almost twice as many respondents
had experimented with smoking in the interim (71% vs. 38%,
p < .0001). The vast majority of never smokers indicated that
they would definitely not smoke cigarettes if they were offered
by friends. These respondents were less likely than others to be

Table 1
Never Smoker 1989 Susceptibility Responses as Predictors
of 1993 Experimentation or Established Cigarette Smoking
Cigarette smoking status, 1993
Susceptibility
responses, 1989

AT

Experimentation11
% (95% CI)

Established
%(95%CI)

Do you think you will try a cigarette soon?
Yes
No
p value

199
4,301

71.2(64.7,77.6)
38.4(36.9,39.8)
< .0001

17.1(11.9,22.4)
7.2(6.5,8.0)
.0006

If one of your best friends were to offer you a cigarette,
would you smoke it?
Definitely yes, probably
yes, or probably not
501
Definitely not
3,999
p value

60.3(56.1,64.5)
37.2(35.6,38.8)
<.0001

11.0(8.4,13.7)
7.3(6.4,8.1)
.0105

Table 2
Never Smokers' Rates of Experimentation and Established
Smoking 1993 by 1989 Susceptibility Score
Smoking onset level, 1993
Susceptibility
score, 1989"
0
1
2
3
2+3

N

Experimentation*1
% (95% CI)

Established smoker
% (95% CI)

3,610
623
215
52

35.6 (35.9, 37.2)
52.8 (48.8, 56.8)
64.6 (57.9, 71.4)
74.6 (60.9, 88.3)

6.5 (5.7, 7.3)
12.3 (9.7, 14.9)
11.3(6.9,15.7)
20.6 (9.4, 31.8)

66.5 (60.6, 72.5)

13.1 (8.9, 17.2)

267

Note. N = 4,500. Weighted percentages, adjusted for sampling
design and nonresponse. CI = confidence interval.
a
O = Not susceptible; 1 = one response indicating susceptibility; 2 +
3 = two or three responses indicating susceptibility. bExperimentation group includes adolescents in the established smoker group.

established smokers in 1993 (7% vs. ll%,p = .0105) or to have
experimented with cigarettes (37% vs. 60%, p < .0001). Similarly, the vast majority of respondents indicated "definitely
not" when asked if they thought they would be smoking 1 year
later. Members of this group were also less likely to be
established smokers in 1993 (p = .0003) or to have experimented (;> < .0001).
To develop our a priori additive index of susceptibility to
smoking, we dichotomized each of these questions. To be
labeled "nonsusceptible" (a score of zero on the index), the
respondent had to answer "No" or "Definitely not" to all three
questions (Table 2; Pierce et al., 1993, 1995). The majority of
the 1989 TAPS never smokers (N = 3,610) were in this
nonsusceptible category and, by 4 years later, 35.6% had
experimented with cigarettes and 6.5% were established smokers (i.e., had smoked more than 100 cigarettes). Only 1%
(N = 52) of the 1989 never smokers answered all three
questions in a way that was incompatible with the nonsusceptible definition; at follow-up three quarters of such respondents had experimented, and almost 21% were established
smokers. Although these data suggest a strong relationship
with later experimentation, the small sample sizes of the
categories with a score of 2 or 3 on the index dictated their
combination. The result is a three-level index of susceptibility
to smoking in which each level has nonoverlapping 95%
confidence intervals on the probability of experimentation by
the 1993 follow-up (0 = 35.9%-37.2%; 1 = 48.8%-56.8%;
2 + 3 = 60.6%-72.5%).
The three-level susceptibility to smoking scores were inversely related to age and perceived school performance,
positively related to exposure to other smokers, and were
greater for Hispanics and for adolescents in the lowest adult
education and family income categories.

Do you think you will be smoking cigarettes 1 year from now?
Definitely yes, probably
yes, or probably not
Definitely not
p value

Predictors of Experimentation With Smoking
509
3,991

55.1(50.6,59.6)
37.8(36.3,39.4)
<.0001

13.3(10.3,16.3)
7.0(6.1,7.8)
.0003

Note. N = 4,500. Weighted percentages, adjusted for sampling
design and nonresponse. CI = confidence interval.
a
Experimentation group includes adolescents in the established group.

Table 3 shows the multivariate analysis of which never
smokers had experimented with smoking by 1993. Approximately 40% of those aged 12-16 years had experimented prior
to the second survey. A lower experimentation rate was
observed in those aged 17 (odds ratio [OR] = .80) and 18

358

PIERCE, CHOI, GILPIN, FARKAS, AND MERRITT

Table 3
Predictors of Experimentation Before 1993 Among
1989 Never Smokers
Experimentation before 1993
Independent variables, 1989

n

Age (years)
12
981
807
13
14
741
15
595
16
504
17
477
18
395
Sex
Female
2,297
Male
2,203
Race/ethnicity
White
3,196
Black
769
Hispanic
367
Asian/other
168
Adult education
456
< 12 years
12 years
1,630
1,092
13-15 years
1,322
16+ years
Family income ($)
760
< 16,000
16,000-29,999
1,191
30,000-49,999
1,499
1,050
50,000+
Perceived school performance
Much better than average
922
Better than average
1,781
Average and below
1,797
Exposure to other smokers
2,019
Minimal
Family only
1,200
Best friends only
685
Both family and best friends
5%
Susceptibility
Not susceptible
3,610
Susceptible level 1
623
Susceptible level 2
267

%

OR(95%CI)

40.4
39.8
41.4
42.1
41.1
35.9
35.5

1.00
0.96
1.02
1.03
0.98
0.80
0.81

38.1
41.6

1.00
1.17(1.03,1.33)

42.1
31.6
41.7
33.1

1.00
0.64 (0.52, 0.78)
0.88 (0.68, 1.13)
0.69 (0.49, 0.98)

39.4
39.6
40.6
39.5

1.00
1.01 (0.78, 1.31)
1.07 (0.80, 1.42)
1.05 (0.79, 1.41)

39.7
37.7
39.5
42.5

1.00
0.89(0.71,1.11)
0.99 (0.79, 1.25)
1.17(0.91,1.51)

33.0
37.7
45 .4

1.00
1.16 (0.97, 1.39)
1.56(1.30,1.87)

34.6
39.4
46.8
49.7

1.00
1.25(1.05,1.49)
1.60(1.34,1.90)
1.84(1.49,2.26)

35.6
52.8
66.5

1.00
1.92(1.61,2.30)
3.15 (2.37, 4.17)

(0.79, 1.17)
(0.82, 1.27)
(0.81, 1.31)
(0.76, 1.26)
(0.63, 1.01)
(0.62, 1.07)

Note, N = 4,500. Weighted percentages, adjusted for sampling
design and nonresponse. OR = odds ratios, adjusted for all the
variables in the table; CI = confidence interval.

(OR = .81) years at baseline, although these were not statistically significant. Males were more likely to experiment than
females. Blacks (OR = .54) and Asians (OR = .69) were
significantly less likely to experiment than Whites or Hispanics.
The educational achievement of the responsible adult showed
no association with experimentation, which was not the case
for reported relative performance in school. Respondents who
classified themselves as having school performance at an
average or below-average level were significantly more likely to
experiment than were those who responded that their performance was much better than average (45% vs. 33%).
A strong relationship was also seen for exposure to other
smokers. Thirty-five percent of those respondents who at
baseline reported no exposure by either family or best friends
had experimented with smoking by follow-up. A higher percentage of those exposed to smokers only within their family had

experimented. Those who reported smoking by best friends
had an even higher experimentation rate again. The rate of
experimentation was highest in respondents exposed to smoking by both their family and their best friend network
(OR = 1.84).
Both levels of susceptibility to smoking in 1989 were significant in predicting experimentation by 1993. Respondents with
a lower level of susceptibility to smoking had an odds ratio of
1.92, whereas adolescents classified in the higher susceptibility
to smoking level in 1989 had a slightly higher odds ratio, 3.15.
This susceptibility-experimentation effect had a stronger association than all other predictor variables in the multivariate
analysis. None of the two-way interactions between susceptibility to smoking and the other predictor variables were significant.

Predictors of Established Smoking
Table 4 presents the multivariate analysis of predictors of
progress toward established smoking prior to the 1993 follow-up as compared with those adolescents who remained
experimenters. There were no significant differences among
those aged 12 years compared with those aged 13 to 17 years
who had smoking rates that varied between 6.8% and 10.0%.
However, a significantly lower smoking rate (3%) was observed
for adolescents aged 18 years at the initial interview, compared
with that of 12-year-olds. There were no significant gender
differences in the progression to established smoking. Blacks
(OR = .41) and Asians (OR = .35) were significantly less
likely than Whites to have progressed to the extent of becoming established smokers at follow-up. Hispanics also appeared
less likely than Whites to become established smokers.
Although the educational level of the responsible parent
was not statistically significant in predicting who progressed to
established smoking, there was some suggestion that adolescents from higher-educated households were more likely to be
established smokers. Adults' reported family income level was
not associated with progress toward established smoking at
follow-up. However, reported relative school performance
demonstrated a significant association: Respondents who perceived that they were average or below average in relative
school performance were much more likely to progress to
established smoking than were those who perceived that they
were performing much better than their peers (OR = 1.79,
95% CI, 1.20-2.66).
Exposure to other smokers was an important predictor of
whether adolescents progressed to established smoking or
remained experimenters at follow-up. Approximately 5% of
never smokers who at baseline had minimal exposure to other
smokers had progressed to established smoking at follow-up.
This rate increased among those exposed to smokers in either
the family (OR = 1.85) or best friend network (OR = 1.66).
Adolescents exposed to smokers in both their family and their
best friend network had the highest rate of established
smoking at follow-up, 13.8%.
Susceptibility to smoking was not significant in distinguishing adolescents who progressed to established smoking from
those that remained experimenters during the 4-year study
period.

VALIDATION OF SUSCEPTIBILITY TO SMOKING

359

expectations to define this susceptibility period among never

Table 4
Predictors of Progress to Established
Smoking Versus Experimentation

smokers, provided that nonsusceptibility was defined as the
existence of a determined decision not to smoke (Pierce et al.,
Established cigarette use, 1993

1993). Susceptibility to smoking at baseline was strongly
associated with moving to the next step in the smoking onset

Independent variables, 1989
Age (years)
12
13
14
15
16
17
18
Sex
Female
Male
Race/ethnicity
White
Black
Hispanic
Asian/other
Adult education
< 12 years
12 years
13-15 years
16+ years
Family income ($)
< 16,000
16,000-29,999
30,000-49,999
50,000+
Perceived school performance
Much better than average
Better than average
Average and below
Exposure to other smokers
Minimal
Family only
Best friends only
Both family and best friends
Susceptibility
Not susceptible
Susceptible level 1
Susceptible level 2

n

%

981
807
741
595
504
477
395

7.8
7.8
8.1
10.0

OR (95% CI)

process (defined as experimentation with cigarettes). This
susceptibility to smoking measure was a more independent and
stronger predictor of experimentation than was the existence
of smokers among either the family or the best friend network,
commonly regarded as the strongest predictor of smoking

8.8
6.8
3.0

1.00
1.03 (0.70, 1.52)
1.05 (0.70, 1.57)
1.36 (0.89, 2.08)
1.07 (0.68, 1.67)
0.91 (0.55, 1.50)
0.41 (0.21, 0.80)

2,297
2,203

7.0
8.4

1.00
1.18 (0.91, 1.51)

efforts to deter subsequent smoking behavior in this group are

3,196

9.3
3.4
5.8
2.9

1.00
0.41 (0.24, 0.69)
0.57 (0.35, 0.95)
0.35 (0.14, 0.88)

5.3
8.8
7.7
7.2

1.00
1.63 (0.98, 2.74)
1.41 (0.84, 2.40)
1.55 (0.85, 2.81)

that progressed to become established smokers (i.e., consuming at least 100 cigarettes by 1993) was exposure to other

7.2
7.3
8.0
8.0

1.00
0.76 (0.49, 1.17)
0.84 (0.57, 1.26)
0.83 (0.51, 1.35)

likelihood that he or she will become an established smoker.
However, susceptibility to smoking at baseline was not predic-

4.4
6.9

1.00
1.39 (0.92, 2.10)
1.79(1.20,2.66)

category may have contributed to the insignificant finding. The

1.00
1.85 (1.34, 2.56)
1.66 (1.16, 2.38)
2.46 (1.73, 3.51)

to smoking measure in predicting which adolescents are at
elevated risk for taking up smoking.

onset. However, because the majority of adolescents in absolute numbers who ultimately experimented with smoking came
from the initially large and lowest susceptibility level group,
still needed.
Our data suggest that the predictors of who will experiment

769
367
168
456
1,630
1,092
1,322

760
1,191
1,499
1,050

922
1,781
1,797

10.2

4.7
8.7
9.3

2,019
1,200

685
596

13.8

3,610

6.5

623
267

12.3
13.1

1.00
1.34 (0.99, 1.82)
0.97 (0.63, 1.49)

Note. N = 4,500. Weighted percentages, adjusted for sampling
design and nonresponse. OR = odds ratios, adjusted for all the
variables in the table; CI = confidence interval.

Again, no two-way interactions between susceptibility to
smoking and the independent variables were significant.

Discussion
This article reports natural history information concerning
smoking behavior in a large national sample of adolescents
who reported in 1989 that they had never so much as puffed on
a cigarette. By the follow-up survey 4 years later, approximately 40% of the sample reported having experimented with
cigarettes, and nearly 8% had progressed so far as to be
labeled as established smokers according to the common
definition for defining adult ever smokers in the United States.
We looked for baseline evidence of cognitive susceptibility
to smoking as an initial level in the process of starting to smoke
and found evidence to support the use of both intention and

with smoking and who will progress to become established
smokers differ, indicating that different processes might be at
work at various levels in the onset process. The strongest
predictor of which individuals would make up the small group

smokers. The presence of smokers in the social environment
may reinforce the adolescent who experiments, increasing the

tive in distinguishing established smokers from experimenters
at follow-up. The small sample size in the top susceptibility
strong relationship between susceptibility to smoking and
experimentation demonstrates the validity of the susceptibility

Another important predictor of established smoking was the
level of parental education. Adolescents whose parents had
not completed high school were less likely to become established smokers than were adolescents whose parents had more
education. Analysis of national data on high school seniors and
the prevalence of past-month smoking with respect to parental
education yielded similar results (National Center for Health
Statistics [NCHS], 1993). Among high school seniors whose
parents did not graduate from high school, the prevalence of
past-month smoking decreased from 33% in 1980 to 31% in
1991; among seniors whose parents graduated from high
school, the prevalence of smoking decreased from 34% to 29%
during the same time. However, for high school seniors whose
parents had some postgraduate education, the prevalence of
smoking increased from 24% in 1980 to 27% in 1991 (NCHS,
1993).
Several potential limitations of the current study should be
considered. One is the reliance on self-reported data from
telephone interviews of adolescents in their homes. Studies of
adolescents have shown that there is stability of self-reported
substance use in the adolescent population and that questionnaires provide highly reliable data (Barnea, Rahav, & Teichman, 1987).
Another possible bias may exist because of differences in
reported smoking behavior between adolescents who were
successfully followed up in TAPS II and nonrespondents. In

360

PIERCE, CHOI, GILPIN, FARKAS, AND MERRITT

addition, adolescents who resided in nontelephone households
in 1989 (who were excluded from the baseline sample in TAPS
I) may have been more likely to be smokers. For these reasons,
prevalence estimates from TAPS II may be lower than they
would have been had the entire TAPS I cohort been successfully followed up, and therefore these results may include
some bias that makes interpretation difficult.
This analysis provides evidence of the different levels
involved in the process of taking up smoking. The positive
move toward becoming a smoker appears to be the development of a cognitive susceptibility to smoking. During this
"preparatory" period, the adolescent develops expectations
and beliefs about smoking (Institute of Medicine, 1994).
Further research is needed to identify the environmental
variables associated with becoming susceptible to smoking.
One obvious candidate is tobacco marketing. Over $5 billion
per year is being spent by tobacco companies on the advertising and promotion of cigarette smoking (Federal Trade
Commission, 1995). Unfortunately, federally supported national surveys, such as TAPS, include no questions to measure
adolescent response to tobacco marketing activities.
Other possible influences on susceptibility to smoking are
parental influence, effective school smoking programs, an
effective smoke-free learning environment, and community
smoking norms. Further research may make it possible to
identify what types of preventive efforts will be effective in
preventing adolescents from becoming susceptible to smoking.
Some of this research should focus on the period after
experimentation. Many of the adolescents who experimented
with cigarettes in this study did not progress further in the
smoking onset process, but appeared to become nonsusceptible to future smoking. Identifying the factors that influence
an adolescent to not progress beyond experimentation could
also lead to the development of more powerful smoking
prevention programs.

Bauman, K. E., Fisher, L. A, Bryan, E. S., & Chenoweth, R. L. (1984).
Antecedents, subjective expected utility and behavior: A panel study
of adolescent cigarette smoking. Addictive Behaviors, 9, 121-136.
Best, J. A., Thompson, S. J., Santi, S. M., Smith, E. A, & Brown, L. S.
(1988). Preventing cigarette smoking among school children. Annual
Review of Public Health, 9, 161-201.
Centers for Disease Control and Prevention. (1991). Cigarette smoking among youth: United States, 1989. Morbidity and Mortality
Weekly Report, 40, 712-715.
Centers for Disease Control and Prevention. (1994). Changes in the
cigarette brand preferences of adolescent smokers: United States,
1989-1993. Morbidity and Mortality Weekly Report, 43, 577-581.
Community Intervention Trial for Smoking Cessation Research Group.
(1995). Community Intervention Trial for Smoking Cessation (COMMIT): I. Cohort results from a four-year community intervention.
American Journal of Public Health, 85, 183-192.
Conrad, K. M., Flay, B. R., & Hill, D. (1992). Why children start
smoking: Predictors of onset. British Journal of Addiction, 87,
1711-1724.
Elder, J. P., De Moor, C., Young, R. L., Wildey, M. B., Molgaard,
C. A., Golbeck, A. L., Sallis, J. F., & Stem, R. A (1990). Stages of
adolescent tobacco-use acquisition..4drfl'cmt'Sc/ian'CT5, 74,449-454.
Farrell, P., & Fuchs, V. R. (1982). Schooling and health: The cigarette
connection. Journal of Health Economics, 1, 217-230.
Federal Trade Commission. (1995). Report to Congress for 1992:
Pursuant to the Federal Cigarette Labeling and Advertising Act.
Washington, DC: Author.
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. J. McGrath & P. Firestone (Eds.), Pediatric and
adolescent behavioral medicine: Vol. 10. Issues in treatment. Springer
series on behavior therapy and behavioral medicine (pp. 132-183).
New York: Springer.
Gilpin, E. A., Lee, L., & Pierce, J. P. (1994). Smoking initiation rates in
adults and minors. United States 1944-1988. American Journal of
Epidemiology, 140, 535-543.
Institute of Medicine. (1994). Crowing up tobacco free: Preventing
nicotine addiction in children and youths. Washington, DC: National
Academy Press.
Johnston, L. D., O'Malley, P. M., & Bachman, J. G. (1993). National

References

survey results on drug use from monitoring the future study, 1975-1992:
Vol. 1. Secondary students. Rockville, MD: National Institute on

Ajzen, L, & Fishbein, M. (1980). Understanding attitudes and predicting
social behavior. Englewood Cliffs, NJ: Prentice Hall.
Allen, K., Moss, A., Botman, S., Winn, D., Oiovino, G., & Pierce, J.
(1991). Teenage attitudes and practices survey (TAPS): Methodology
and response rates. Paper presented at the 119th annual meeting of
the American Public Health Association. November, Atlanta, GA.
Allen, K F., Moss, A. J., Giovino, G. A., Shopland, D. R., & Pierce, J.
P. (1992). Teenage tobacco use: Data estimates from the Teenage
Attitudes and Practices Survey: United States, 1989 (Advance data no.
224). Hyattsville, MD: National Center for Health Statistics.
Ary, D. V., & Biglan, A. (1988). Longitudinal changes in adolescent
cigarette smoking behavior: Onset and cessation. Journal of Behavioral Medicine, 11, 361-382.

Drug Abuse.
Lee, L., Gilpin, E. A., & Pierce, J. P. (1993). Changes in the patterns of
initiation of cigarette smoking in the United States: 1950,1965, and
1980. Cancer Epidemiology Biomarkers and Prevention, 2, 593-597.
Leventhal, H., & Cleary, P. D. (1980). The smoking problem: A review
of the research and theory in behavioral risk modification. PsychologicalBulletin, SS, 370-405.
McNeill, A. D., Jarvis, M. J., Stapleton, J. A,

Russel, M. A,

Eiser, J. R., Gammage, P., & Gray, E. M. (1988). Prospective study
of factors predicting uptake of smoking in adolescents. Journal of
Epidemiology and Community Health, 43, 72-78.
National Center for Health Statistics. (1993). Health United States,
1992

(DHHS Publication No. PHS 93-1232). Hyattsville,

MD:

Bal, D. G., Kizer, K. W., Felten, P. G., Mozar, H. N., & Niemeyer, D.
(1990). Reducing tobacco consumption in California. Journal of the

Public Health Service.
Nelson, D. E., Giovino, G. A., Shopland, D. R., Mowery, P. D., Mills,

American Medical Association, 264, 1570-1574.
Bandura, A. (1977). Social learning theory. Englewood Cliffs, NJ:

S. L., & Eriksen, M. P. (1995). Trends in cigarette smoking among
U.S. adolescents, 1974 through 1991. American Journal of Public

Prentice Hall.
Bandura, A. (1986). Social foundations of thought and action: A social

Health, 85, 34-40.
Pierce, 3. P., Evans, N., Farkas, A. J., Cavin, S. W., Berry, C., Kramer,
M., Kealey, S., Rosbrook, B., Choi, W. S., & Kaplan, R. (1994).

cognitive theory. Englewood Cliffs, NJ: Prentice Hall.
Barnea, Z., Rahav, G., & Teichman, M. (1987). The reliability and
consistency of self-reports on substance use in a longitudinal study.
British Journal of Addiction, 82, 891-898.

Tobacco use in California: An evaluation of the Tobacco Control
Program, 1989-1993. La Jolla, CA: University of California, San
Diego.

361

VALIDATION OF SUSCEPTIBILITY TO SMOKING
Pierce, J. P., Farkas, A,, Evans, N,, Beny, C, Choi, W., Rosbrook, B.,
Johnson, M., & Bal, D. G. (1993). Tobacco use in California 1992: A
focus on preventing uptake in adolescents. Sacramento: California
Department of Health Services.
Pierce, J. P., Farkas, A J., Evans, N., & Gilpin, E. (1995). An improved
surveillance measure for adolescent smoking? Tobacco Control, 4,
S47-S56.
Research Triangle Institute. (1989). SUDAAN Professional Software for
SUrvey DAta ANafysis. Research Triangle Institute.
Schinke, S. P., & Gilchrist, L. D. (1983). Survey and evaluation
methods: Smoking prevention among children and adolescents.
NIDA Research Monograph, 48, 96-104.
Stern, R. A., Prochaska, J. O., Velicer, W. P., & Elder, J. P. (1987).

Stages of adolescent cigarette smoking acquisition: Measurement
and sample profiles. Addictive Behaviors, 12, 319-329.
Sussman, S., Dent, C. W., Flay, B. R., Hanseu, W. B., & Johnson, C. A.
(1987). Psychosocial predictors of cigarette smoking onset by White,
Black, Hispanic, and Asian adolescents in Southern California.
Morbidity and Mortality Weekfy Report, 36, 11S-16S.
U.S. Department of Health and Human Services. (1989). The health
consequences of smoking: 25 years of progress. A report of the Surgeon
General (DHHS Pub. No. CDC 89-8411). Rockville, MD: Centers
for Disease Control.
U.S. Department of Health and Human Services. (1994). Preventing
tobacco use among young people: A report of the Surgeon General.
Atlanta, GA: Centers for Disease Control and Prevention, Office on
Smoking and Health.

AMERICAN PSYCHOLOGICAL ASSOCIATION
SUBSCRIPTION CLAIMS INFORMATION

Today's Date:

We provide this form to assist members, institutions, and nonmember individuals with any subscription problems. With the
appropriate information we can begin a resolution. If you use the services of an agent, please do NOT duplicate claims through
them and directly to us. PLEASE PRINT CLEARLY AND IN INK IF POSSIBLE.

PRINT FULL NAME OS KEY NAME OF INSTITUTION

MEMBEROR.CUSTOMERNUMBER(MAYBEFOUND ON ANYPASTBSUE LABEL)

ADDRESS

DATE YOUR ORDER WAS MAILED (OR PHONED)
PREPAID

CHECK
CHARGE
CHECK/CARD CLEARED DATE:.

STATECOUNTRY

CITY

YOUR NAME AND PHOME NUMBER

(If possible, Bond a copy, front and bade, of your cancelled check to help us in our research
of your claim.)
ISSUES:
MISSING
DAMAGED

VOLUME OR YEAR

TITLE

NUMBER OR MONTH

Thank you. Once a claim if received and resolved, delivery of replacement issues routinely takes 4-6 weeks.
———_————

DATE RECEIVED:
ACTION TAKEN: _
STAFF NAME:

(TO BE FILLED OUT BY APA STAFF)

DATE OF ACTION: _
INV.NO.&DATE:
LABEL NO. & DATE:

Send this form to APA Subscription Claims, 750 First Street, NE, Washington, DC 20002-4242
PLEASE DO NOT REMOVE. A PHOTOCOPY MAY BE USED.