Effects of Gender in Social Control of Smoking Cessation
J. Lee Westmaas
State University of New York at Stony Brook
T. Cameron Wild
University of Alberta
University of Toronto and Centre for Addiction and Mental Health
This study of 93 men and 117 women smokers during an ongoing quit attempt examined the roles of
gender and social network influences on quitting. For men, social influences appeared to positively affect
their ability to reduce their smoking but were less effective for women. Specifically, increased reports of
a spouse or partner’s influence, and family and friends’ influence, were associated with greater reductions
in men’s smoking 2 days and 4 months post quit date, respectively. In contrast, for women, greater reports
of spouse or partner influence and of family and friends’ influence were associated with smaller
reductions in smoking. Sex differences in social control strategies and perceived autonomy supportive-
ness of those strategies are discussed as possible explanations for these results.
Key words: gender, quitting smoking, social control
Convincing epidemiological evidence links social network in-
volvement with better health and lowered mortality (Berkman &
Syme, 1979; for a review, see House, Landis, & Umberson, 1988).
Studies documenting the salutary effects of social relationships on
health and mortality, however, have consistently found strong
social ties to benefit men more than women. The marital relation-
ship, in particular, appears to be more advantageous to men’s
health than to women’s (Berkman & Syme, 1979; Blazer, 1982;
Gove, 1973; House et al., 1988; House, Robbins, & Metzner,
1982). One potential explanation for this finding, which this study
investigates, is that female partners are more successful in influ-
encing male partners to change negative health behaviors, for
example smoking, than males are in influencing female partners to
change. We chose to study smoking in a community sample of
men and women because (a) tobacco use claims at least 440,000
lives every year in the United States (Centers for Disease Control
and Prevention, 2002) and is responsible for at least 30% of all
cancer deaths (American Cancer Society, 2002) and (b) research
shows that social network factors consistently predict smoking
status (Conrad, Flay, & Hall, 1992).
Social Control of Health Behaviors
A large literature confirms that the supportive functions of
social networks, such as the provision of instrumental and emo-
tional support, promote health by buffering individuals from the
detrimental effects of stress (for a review, see S. Cohen & Wills,
1985). Sociological theorists have also argued that social networks
serve regulatory or control functions (Durkheim, 1897/1951;
Hughes & Gove, 1981). Social relationships can therefore also
influence health through the regulation of health behaviors by
social network members (Umberson, 1987). Social control of
health behavior can operate directly or indirectly. Avoiding health-
compromising behaviors to fulfill obligations and responsibilities
to one or more significant other(s) constitutes indirect social con-
trol—for example, a mother giving up smoking because of con-
cerns for her children’s health. Social control can also operate
directly—for example, when members of a social network per-
suade an individual to give up health-compromising behaviors and
adopt health-enhancing behaviors. Evidence supports the impor-
tance of both direct and indirect social control in explaining the
relationship between social network involvement and health. For
example, Lewis and Rook (1999) found that, among community
residents, being the target of social control attempts was associated
with a lower incidence of health-compromising behaviors. Addi-
tionally, Umberson (1987) demonstrated that family role obliga-
tions promote regulation of health behaviors linked to mortality.
Gender and Social Control of Health Behavior
Umberson (1992) extended the social control hypothesis by
incorporating theory and findings on gender role socialization. She
J. Lee Westmaas, Department of Psychology, State University of New
York at Stony Brook; T. Cameron Wild, Department of Public Health
Sciences and Centre for Health Promotion Studies, University of Alberta,
Edmonton, Alberta, Canada; Roberta Ferrence, Ontario Tobacco Research
Unit, University of Toronto, Toronto, Ontario, Canada, and Centre for
Addiction and Mental Health, Toronto, Ontario, Canada.
This research was supported by a grant from the Centre for Addiction
and Mental Health to J. Lee Westmaas. We thank William A. Corrigall for
his helpful advice and the research assistants who helped conduct this
study: Linda Couper, Meytal Elhav, Thusi Iswaran, Ailin Oshi, and Jessica
Correspondence concerning this article should be addressed to J. Lee
Westmaas, Department of Psychology, State University of New York,
Stony Brook, New York 11794-2500. E-mail: johann.westmaas@sunysb
2002, Vol. 21, No. 4, 368–376
Copyright 2002 by the American Psychological Association, Inc.
hypothesized that because women’s socialization and experiences
are oriented toward nurturance and caring for others, whereas
men’s are more oriented toward aggression and risk taking, women
would be more likely to exert control over others’ health behavior,
especially that of their spouses. Indeed, a large panel survey
showed that women are more likely than men to be named as
social control agents by both women and men, and by married men
in particular (Umberson, 1992). Moreover, although women re-
ported a greater variety of people who attempted to influence their
health behavior (friends, relatives, etc.), men more often named
their spouses as exclusive agents of social control.
Although these findings suggest that women may influence
others’ health behavior more often than men, little subsequent
research has examined the role of gender differences in social
control processes. In particular, we do not know whether there are
gender differences in the efficacy with which people influence the
health behaviors of spouses or partners. For social control to be a
viable explanation for the greater effect of the spousal relationship
on men’s health compared with women’s health, social control
tactics used by women would have to be fairly successful in
influencing men’s health behavior. Conversely, men would have to
be less successful in influencing their female partners’ health
In this study, we propose that women are in fact more successful
than men in changing their partners’ health behaviors. We hypoth-
esize that gender differences in social behavior and communica-
tion styles result in women having a greater influence on men’s
health behavior than men have on women’s behavior. Specifically,
we predict that greater interpersonal sensitivity, empathy, and
expressiveness among women leads to greater success in attempts
to influence spouses’ health behaviors.
Gender and Interpersonal Behavior
Gender differences in personality and social behavior can be
effectively captured by poles of the agency–communality dimen-
sion (Bakan, 1966; Eagly & Wood, 1991; Lueptow, Garovich, &
Lueptow, 1995). Agency refers to qualities associated with the
male stereotype, such as instrumentality, aggressiveness, and de-
cisiveness, whereas communality refers to qualities associated with
the female stereotype, such as intimacy, emotionality, affection,
and sympathy. In a recent event-sampling study, although gender
differences in agency were not obtained, women were more com-
munal than men, both at home and at work, and were more
communal with other women than men were with other men
(Moskowitz, Suh, & Desaulniers, 1994). A recent meta-analysis of
sex differences in personality reported that women scored higher
on nurturance (tender-mindedness), extraversion, and trust,
whereas men scored higher on assertiveness (Feingold, 1994). The
more communal orientation of women noted in these studies is
consistent with research showing that women smile more than
men, are more nonverbally expressive, are more tactile, and inter-
act more intimately with others (for reviews, see Hall, 1984; Hall
& Halberstadt, 1986). Additional research has demonstrated that
women provide more “person-centered” comforting messages by
listening more, being evaluatively neutral, being more accepting of
another’s distress, and being more empathic (Barbee, Gulley, &
Cunningham, 1990; Derlega, Barbee, & Winstead, 1994; Eisen-
berg & Lennon, 1983). Men, in contrast, produce more instrumen-
tally or problem-focused support messages, such as recommending
escape or diversion strategies. However, both women and men
prefer the person-centered, comforting strategies associated with
the female role (Kunkel & Burleson, 1999).
Gender Differences in Success of Social Control Efforts
Women’s more communal interaction style may be incorporated
into their attempts to influence partners’ health behaviors. For
example, a woman’s attempts to influence might incorporate em-
pathy by acknowledging how difficult it is for her partner to
change and expressing her understanding of the pleasure that the
health-compromising behavior (e.g., smoking) brings to her part-
ner. She may also be more likely to provide emotional support
while the partner is attempting to abstain from the health-
compromising behavior. This may be particularly important for
behaviors such as smoking, in which withdrawal often results in
negatively valenced emotions such as anxiety, depression, and
irritability (Hughes, 1992). If supportive acts are typically incor-
porated into women’s social control efforts, the result might be
greater success in influencing their male partners’ health behaviors
than men might have in influencing their female partners’
Women may also be more successful in attempts to influence
the health behavior of others because of gender role expectations
held by both men and women. Specifically, attempts to influence
others’ health behavior is consistent with the nurturing caregiving
role ascribed to women, especially for family-related health mat-
ters (Guberman, Maheu, & Maille, 1992; Hemard, Monroe, At-
kinson, & Blalock, 1998). In contrast, attempts by men to influ-
ence others’ health behavior are inconsistent with gender role
expectations. Such inconsistency may produce negative reactions
in women partners and may result not only in reduced efficacy by
men in influencing the health behavior of their female partners but
in relationship conflict as well. Finally, women may be more likely
than men to use their children’s health as a reason to influence
their spouses’ health behavior. Given their greater role as gate-
keepers of their family’s health and caregivers of children, women
may be more likely to emphasize their children’s health as a reason
why male partners should change health-compromising behaviors.
Concern over children’s health, as promoted by women, may
represent another way in which women may more powerfully exert
social control over male partners’ health behaviors.
We also wanted to examine the relative success of other direct
and indirect forms of social control, besides partner influence, in
reducing smoking behavior. Earlier studies explored associations
between social influences and smokers’ success at quitting. These
studies, however, have been retrospective or used older data sets
(Gilpin, Pierce, Goodman, Burns, & Shopland, 1992; Halpern &
Warner, 1993), combined various social control influences into
one variable (e.g., Rose, Chassin, Presson, & Sherman, 1996), or
focused exclusively on predictors of social pressures to quit (e.g.,
Royce, Corbett, Sorensen, & Ockene, 1997). Although these stud-
ies implicate social control influences in promoting abstinence
(e.g., advice from a doctor, setting an example for children), we do
not know which of these interpersonal targets is the most important
source of social influence for changing smoking behavior. Because
interactions with a partner or spouse are generally more frequent
than those with other social network members, and thus provide
EFFECTS OF GENDER IN SOCIAL CONTROL OF SMOKING
more opportunities for spouses to engage in efforts at social
control, a partner’s influence should be stronger in changing smok-
ing behavior compared with other influences. Another reason for
the superiority of partners’ influence is that failure could elicit
conspicuous disappointment or criticism from partners. Yet certain
indirect sources of social control may be as important as a part-
ner’s influence. For example, because of women’s role as care-
givers of children, we would expect that indirect sources of social
control, such as wanting to quit for the health of others or to set an
example for children would have a strong impact on women’s
On the basis of our review of the literature, we developed the
1. Men who report that their partners’ desire for them to be
abstinent was important in their decision to quit will be more
successful in reducing their smoking compared with women who
report that their partners’ desire for them to be abstinent was
2. Among direct sources of social control, the influence of a
partner or spouse will be more strongly related to smoking behav-
ior change than more distal sources, such as advice from a doctor
or pressure from family or friends.
3. The more women report trying to quit because of concern
about the effects of smoking on others, or to set an example for
children, the more successful they will be in reducing their smok-
ing. For men, we predict a weaker relationship between intending
to quit because of concern about others or children and reductions
Two hundred ten smokers were recruited for the study through adver-
tisements placed in newspapers serving Toronto, Ontario, Canada, a large
metropolitan area with more than 3 million residents. Twelve advertise-
ments were placed in weekend and weekday issues over a period of 1
month in January 1998. These advertisements generated approximately 500
calls. Questionnaires, along with instructions, were mailed to these callers.
We received completed questionnaires from 210 of these individuals and
compensated them $15 (Canadian) for their participation.
Of the 210 participants, 93 (44%) were men and 117 (56%) were
women. Participants ranged in age from 17 to 71, with a mean age of 41
(SD ? 12) years. Fifty-five percent of participants had at least some
postsecondary education. For 39% of participants, household income was
less than $20,000 per annum, for 19% it was between $20,000 and $30,000,
and for 14.8% it was between $30,000 and $40,000. The remainder
(27.2%) had annual incomes greater than $40,000. (All amounts are in
Canadian dollars.) About a third (34%) of participants were either married
or currently living with a partner, a quarter (25%) were either divorced or
separated, and 37% reported never having been married. Participants
smoked an average of 18 cigarettes per day (SD ? 9). The average age
when participants began smoking was 17 years. The majority had never
sought treatment for smoking (76%), and the average reported number of
previous quit attempts was five (SD ? 7). Responses to a separate question
indicated that 138 individuals (66%) reported having a partner. The ma-
jority of analyses were conducted on this subset of individuals. Of this
group, 35 reported having children who lived with them. Because it is
likely that social influences operate differently for teenagers compared
with adults, two 19-year-old participants were excluded from all analyses.
Newspaper advertisements recruited smokers who were serious about
quitting on their own. The display ads indicated that participants would set
a quit date and report on their quit attempt. A telephone number was listed,
and a tape-recorded message informed respondents of the eligibility re-
quirements. These were that the smoker (a) was planning to quit in the
next 20 days; (b) currently smoked at least 10 cigarettes per day; (c) was
not currently receiving treatment for smoking cessation; and (d) had never
been diagnosed with emphysema, lung cancer, myocardial infarction, or
stroke. Eligible participants were asked to leave their name and phone
number so that a questionnaire packet could be mailed to them. The
recorded message informed smokers that the study did not provide treat-
ment for smoking cessation.
On their receipt of the questionnaire packet, participants were asked to
complete measures that assessed sociodemographic variables, smoking
dependence, smoking history, the importance of various reasons for quit-
ting, and other smoking-related variables not used in the current analysis.
In the letter accompanying the packet, participants were asked to select a
quit date that would occur within 20 days of receiving the packet. They
were asked to complete specific measures included in the packet at the end
of the 1st and 2nd days of their attempted quit. The 2 days immediately
following the quit attempt were chosen because the majority of nicotine
withdrawal symptoms are at their peak during this period (Hughes, 1992).
Thus, the effects of our social control variables on smoking behavior
should be more apparent within this period. From here on, we refer to this
period as the 2-day post-quit-date period and to the period prior to that as
the pre-quit-date period.
When we received the entire set of completed measures, we mailed a
check for $15 (Canadian) to each participant. Four months later, we sent
each participant a letter asking how many cigarettes per day they currently
smoked (response rate 76%).
Single items were used for the majority of sociodemographic variables
assessed (age, sex, marital status, income, education, and number of
children). Smoking dependence was assessed using the six-item Fager-
strom Test for Nicotine Dependence (Heatherton, Kozlowski, Frecker, &
Fagerstrom, 1991). Additional single items assessed age of initiation for
smoking, number of previous quit attempts, previous treatment for smok-
ing cessation, self-efficacy for quitting, motivation to quit, parental status,
and presence of children in the home.
On five separate 5-point scales, participants were asked
to rate the importance of direct and indirect sources of social influence in
their decision to quit. Items for direct sources were “Pressure from family
and friends,” “Advice from a doctor,” and “Spouse or partner wanted me
to quit.” Items for indirect sources were “Concern about the effects of my
smoking on other people” and “To set an example for children.” The
importance of each of these influences was rated on a single Likert scale
ranging from 1 (not at all important) to 5 (very important).1These items
have been used in previous investigations (Gilpin et al., 1992; Halpern &
Warner, 1993) and have shown differential relationships with success in
quitting (e.g., Halpern & Warner, 1993).
1A principal-components analysis (varimax rotation) confirmed the
distinction between direct and indirect forms of social control. Two factors
with eigenvalues greater than 1 emerged from the analysis of the five
sources. On the first Direct Sources factor, which accounted for 49% of the
variance, loadings were greater than .64 for direct sources and less than .28
for indirect sources. On the second Indirect Sources factor, which ac-
counted for 21% of the variance, loadings were greater than .89 for indirect
sources and less than .19 for direct sources.
WESTMAAS, WILD, AND FERRENCE
Reduction in smoking level.
variable rather than point-prevalent abstinence because any significant
main or interactive effects would be more readily detectable with a nor-
mally distributed continuous dependent variable than with a dichotomous
dependent variable. In fact, according to power analysis using procedures
recommended by J. Cohen (1988), our current sample size has minimal
power (.30) to detect even a small main effect of gender on abstinence
coded dichotomously (assuming a .05 significance criterion). In contrast,
with our current sample size, we have good power (.80) to detect a
medium-to-large interaction effect using a continuous measure of cigarette
intake. Additionally, smoking level has been used as an end point in
previous studies (Aubin, Laureaux, Tilkete, & Barrucand, 1999; Becona,
Vasquez, Fuentes, & Lorenzo, 1999), and reduction in cigarette consump-
tion is considered a valid measure of smoking behavior change (e.g.,
Bollinger, 2000). Moreover, prior research has demonstrated that in non-
intervention studies, self-report of smoking is a valid indicator of smoking
level (for a review, see Velicer, Prochaska, Rossi, & Snow, 1992).
Smoking level at the end of each of the first 2 days of the quit attempt
was assessed by a single item “What was the total number of cigarettes you
smoked today?” The total number of cigarettes smoked over the 2-day
period was summed and divided by 2 to yield the average daily number of
cigarettes smoked in the 2-day post-quit-date period. Smoking level at the
4-month follow-up was assessed by follow-up letters that asked partici-
pants how many cigarettes they currently smoked per day. More than three
quarters of participants (76%) replied to this question, with 19% of these
indicating abstinence. Assuming conservatively that nonresponders at the
4-month follow-up were not abstinent, the total abstinence rate for the
entire sample, on the basis of self-reported abstinence for the two follow-up
periods, was 12%.2Among 4-month responders who reported continued
smoking, the mean number of cigarettes smoked per day was 16 (SE ? 12).
For each participant, reduction in smoking level was calculated by
subtracting daily smoking level during the 2-day post-quit-date period from
the pre-quit-date daily smoking level. The resulting values indicate the
magnitude of reduction in number of cigarettes smoked per day immedi-
ately following the participants’ quit attempt. Similarly, for each partici-
pant, we subtracted the average daily smoking level at the 4-month
follow-up from the pre-quit-date daily smoking level. This yielded the
magnitude of reduction in participants’ daily smoking level 4 months after
their quit attempt. Twenty-four participants were abstinent at both the
2-day and 4-month post-quit-date periods.
We used smoking level as the dependent
There were no significant gender differences in preexisting level
of tobacco dependence (on the basis of Fagerstrom Test for Nic-
otine Dependence scores), age, age of smoking initiation, number
of previous quit attempts, self-efficacy, or motivation to quit (see
Table 1). Men and women also did not differ on number of
cigarettes smoked per day before the quit attempt or at the 2-day
follow-up. At the 4-month follow-up, there was a tendency for
women to report smoking fewer cigarettes, although the difference
was not statistically significant (p ? .06).
Chi-square analyses were conducted to examine potential gen-
der differences in sociodemographic and smoking-related vari-
ables. These analyses revealed no significant gender effects for
education, marital status, income, parental status, or history of
depression. Women, however, were more likely to report having
received treatment for smoking cessation (29% of women vs. 17%
of men), ?2(1, N ? 210) ? 4.01, p ? .04. Women were also more
likely to report currently living with children (30% of women vs.
16% of men), ?2(1, N ? 210) ? 5.24, p ? .02. Women and men
did not differ in how influential they rated direct and indirect
sources of social control in their decision to quit.
Direct social control.
social control relate to having a partner, we restricted our analyses
to individuals who reported having a partner (n ? 138). In addi-
tion, when reduction in smoking level at the 4-month follow-up
was the dependent variable, only responders to the 4-month
follow-up were included.3Age, income, and Fagerstrom depen-
dence scores were covaried in all analyses if they were signifi-
cantly related to the dependent variable under consideration. Our
hypothesis that social control from a partner would be more
strongly related to reduced smoking in men than in women trans-
lates to positing an interaction between gender and participants’
ratings of partner influence. Hierarchical regressions were thus
conducted in which gender and ratings of partner influence were
Because our hypotheses regarding direct
2This proportion of abstinence is comparable to that found in studies of
self-quitters. For example, Hughes et al. (1992), using biochemical verifi-
cation, found that 11% of 630 self-quitters were abstinent at 3 months.
3The distribution for reduction in smoking level at the 4-month
follow-up among responders revealed an outlier that was more than 20
units greater than three standard deviations above the mean. This outlier
was excluded in analyses because extreme cases can exert an undue
influence on results and should be removed or rescored (J. Cohen & Cohen,
1983; Tabachnick & Fidell, 1983).
Gender Differences in Sociodemographic, Smoking History,
and Smoking Level Variables
(n ? 93)
(n ? 116)
Fagerstrom Test for Nicotine
Age of smoking initiation
Previous quit attempts
Motivation to quit
No. of cigarettes smoked/day
No. of cigarettes smoked/day
(48 hr post-quit-date)
No. of cigarettes smoked/day
(4 months post-quit-date)
EFFECTS OF GENDER IN SOCIAL CONTROL OF SMOKING
entered in Step 1 and the multiplicative interaction term was
entered in Step 2.4
No significant main effects of gender and social influence
ratings were obtained for reduction in smoking for the 2-day
post-quit-date period, but the increment in variance attributed to
the interaction was significant, ?F(1, 122) ? 4.59, p ? .03 (see
Table 2). For male smokers, greater partner influence predicted
greater reductions in daily smoking level (B ? 0.20, SE ? .53;
Figure 1). In contrast, for women, the greater the partner influence,
the smaller the reduction in smoking rate (B ? ?1.40, SE ? .54).
We also regressed reduction in daily smoking at the 4-month
follow-up on our independent variables. These results revealed a
significant main effect of reports of partner or spouse influence,
with greater reports of partner influence predicting greater reduc-
tions in smoking (? ? ?0.23, p ? .02; see Table 2). The inter-
action between gender and reports of partner influence was not
significant for reduction in daily smoking at the 4-month
To examine the extent to which pressure from family and
friends predicted reductions in smoking, we reconducted analyses
substituting these ratings for ratings of spouse or partner influence.
For reduction in daily cigarette intake in the 2-day post-quit-date
period, no main effects of gender or influence ratings, or of their
interaction, were obtained. For reduction in daily cigarette intake 4
months post quit date, however, a significant interaction with
gender (Figure 2) was obtained, ?F(1, 96) ? 4.35, p ? .04 (see
Table 3). This result indicated that for men, the greater the reported
influence of family and friends, the greater the reductions in
smoking (B ? 3.00, SE ? 1.12). In contrast, for women, greater
family and friend influence predicted smaller reductions (B ?
?0.30, SE ? 1.10). Analyses using ratings of a doctor’s influence
indicated no main effect of this variable or of its interaction with
gender on smoking reduction at either post-quit-date periods.
Indirect social control.
We also hypothesized that greater con-
cern about the effects of smoking on others or wanting to set an
example for children would predict greater reductions in smoking
for women than men. For our results to be relevant to the possi-
bility that women (or men) may use the health of their children as
a way to influence their partners to quit, our analyses were first
restricted to participants who reported having a partner (whether
married or cohabiting) and who also had children living with them.
This resulted in a final subsample of 35 individuals. In this
subsample, the distributions for both indirect social control vari-
ables were skewed, with the majority of participants (over 75%)
giving a rating of 4 or 5. Among participants who scored 4 or 5, we
dichotomized ratings for the two influence variables by creating
two groups for each variable: individuals who scored 4 (“low”)
versus individuals who scored 5 (“high”). We then conducted
2 ? 2 analyses of variance that tested the interaction of gender
with each indirect social influence variable. These analyses indi-
cated a significant interaction between gender and concern about
the effects of smoking on the reduction in smoking 2 days post quit
date, F(1, 25) ? 5.32, p ? .03. Means inspection indicated that
among men, a higher rating on concern about effects of smoking
on others was associated with a reduction in smoking of approx-
imately 10 cigarettes from the pre-quit-date period to 2 days post
quit date, t(10) ? 2.7, p ? .02. Among women, however, cigarette
intake was higher the lower the rating on concern about effects of
smoking on others (Figure 3).
Previous population-based studies have examined the role of
indirect social influences without restricting analyses to partici-
pants with children (Gilpin et al., 1992; Halpern & Warner, 1993;
Orleans, Jepson, Resch, & Rimer, 1994). This is because individ-
uals can be concerned about the effects of smoking on others who
are not part of one’s family or because individuals may want to set
an example for children in general. For comparison purposes we
conducted analyses on the whole sample of individuals (but ex-
4To test for spurious moderator effects when a significant interaction
effect was obtained (Lubinski & Humphreys, 1990), we reconducted the
analysis with a quadratic term (in Step 2) derived from importance ratings.
In these analyses, the significance of the increment in variance attributed to
the interaction term did not change, and the quadratic term was never
significantly related to the dependent variable. The quadratic term was thus
dropped from the final models.
Effects of Gender and Partner Influence on Reduction in
Smoking Level at 2-Day and 4-Month Follow-Ups (n ? 128)
Fagerstrom Test for Nicotine
Spouse or partner influence
Gender ? Spouse or Partner Influence
Spouse or partner influence
Gender ? Spouse or Partner Influence
* p ? .05. *** p ? .001.
quit date, as a function of spouse or partner influence and gender.
Reduction in daily cigarette consumption, 2 days following the
WESTMAAS, WILD, AND FERRENCE
cluded teenagers and nonresponders for the 4-month follow-up
analysis). Our hierarchical linear regressions indicated no signifi-
cant main effects of these indirect sources. The interaction between
gender and quitting smoking to set an example for children,
however, was significant for the reduction in smoking at the
4-month follow-up, ?F(1, 150) ? 4.79, p ? .01 (see Table 4). It
was men, not women, however, for whom quitting smoking to set
an example for children predicted greater reductions in smoking 4
months after the quit attempt (B ? 1.57, SE ? .73; Figure 4).
Among women, quitting to set an example for children decreased
the likelihood of reducing smoking (B ? ?0.57, SE ? .65).
The present study demonstrated that reports of a spouse or
partner’s influence independently, and interactively with gender,
predicted reduced smoking both in the long and short terms,
suggesting the importance of a spouse or partner in producing
behavioral change. However, influences from the immediate social
network (spouse or partner, family and friends) appeared to be less
effective in helping women reduce their smoking both in the short
and long terms, whereas for men, these social influences were
positively related to the ability to reduce their smoking. These
results suggest that female partners of male smokers may be a
valuable source of influence in cessation interventions (although
the success of engaging female partners may depend on whether
they themselves are nonsmokers). Moreover, these results point
out the need to further understand how women’s social networks
are involved in a smoking cessation attempt, because greater
influences from the network appeared to detrimentally affect
women’s ability to reduce their smoking.
Results for indirect sources of social control indicated that it was
men, rather than women, for whom greater concern about the
effects of smoking on others predicted greater reductions in smok-
ing. When analyses were not restricted to individuals with chil-
dren, a similar pattern emerged—that is, 4 months following the
quit date, it was men, rather than women for whom greater reports
of wanting to set an example for children predicted greater reduc-
tions in smoking.
To explain these findings, it may simply be the case that among
men, concern about the effects of one’s smoking on others and
wanting to set an example for children (not necessarily their own)
is indeed intrinsically motivating. Another possibility is that mem-
bers of men’s social networks, including the spouse or partner,
encourage men to be concerned about how their smoking is af-
fecting their children. Men’s female partners may also encourage
them to change their smoking behavior in order to set an example
for children in general. This would be consistent with the nurtur-
ing, care-giving role ascribed to women (Guberman et al., 1992;
Hemard et al., 1998).
The finding that for women, greater concern about the effects of
smoking and wanting to set an example for children predicted
smaller reductions in smoking was unexpected. One potential
explanation is that female smokers’ social networks, including the
spouse, may be using the idea that women should be concerned
Effects of Gender and Pressure From Family and Friends on
Reduction in Smoking Level at 4 Months (n ? 101)
Pressure from family and friends
Gender ? Pressure From Family and Friends
* p ? .05.
Effects of Gender and Quitting to Set an Example for Children
on Reduction in Smoking Level at 4 Months (n ? 154)
Quitting to set an example for children
Gender ? Quitting to Set an Example
* p ? .05.
the quit date, as a function of pressure from family and friends and gender.
Reduction in daily cigarette consumption, 4 months following
quit date, as a function of concern about the effects of smoking on others
and gender. Subgroup ns (left to right) are 5, 5, 7, and 9. Error bars
represent standard errors.
Reduction in daily cigarette consumption, 2 days following the
EFFECTS OF GENDER IN SOCIAL CONTROL OF SMOKING
about the effects of their smoking on others, or should set an
example for children, to motivate them to quit. Although women
may endorse this idea fully, they may react negatively to such
influence attempts because of the implication that by smoking they
are somehow deficient in fulfilling the stereotyped female role of
being concerned about others and children. This may be perceived
as an additional psychological burden by women that negatively
affects their ability to reduce their smoking. If this is indeed the
case, it would suggest that although men’s smoking may benefit
from the social control efforts of the social network to increase
concern about the effects of smoking on others, women’s ability to
reduce their smoking may be detrimentally affected by such ef-
forts. There may also be other uninvestigated factors that might
explain these results (e.g., women’s health status), and that should
be explored in future research.
Previous research supporting gender differences in agentic and
communal social interaction styles (Feingold, 1994; Lueptow et
al., 1995; Moskowitz et al., 1994) has suggested that the nature of
men’s social influence attempts may also be responsible for the
results obtained for women. Indeed, in an experimental study of
husbands’ and wives’ social influence strategies (Brown & Smith,
1992), men’s attempts to persuade spouses were accompanied by
greater anger, hostility, and a “coldly assertive” interpersonal style
Men’s attempts to influence their partners’ smoking may have
incorporated more instrumentally oriented advice—for example,
recommending escape or diversion strategies. In general, these
strategies are more likely to be suggested by men in their attempts
to help their partners cope with stressors (Barbee et al., 1990;
Derlega et al., 1994), but such strategies may be inappropriate for
cessation if women experience a high level of negative affect when
they abstain. Indeed, in our study, women were significantly more
likely than men to indicate that coping with negative affect was a
reason for smoking (p ? .008), suggesting that abstaining from
cigarettes may have produced more negative affect in this sample
of women than in the men. For these women, the provision of
emotional support may have been extremely important in helping
them abstain. Inadequate emotional support, coupled with per-
ceived social pressure from men to quit, may have increased
women’s difficulty in resisting smoking. We do not mean to imply
that the influence techniques and support actions that men use are
not important, but rather that the context in which they are applied
may render them ineffective and possibly detrimental. The process
of quitting smoking requires both behavioral and cognitive coping
strategies (Curry & Marlatt, 1985; Shiffman, Read, Maltese, Rap-
kin, & Jarvik, 1985), and for some individuals, diversion or escape
strategies favored by men (Barbee et al., 1990; Derlega et al.,
1994) might work. Possibly these are strategies that men them-
selves might use to quit. For a partner’s quit attempt, however,
men might benefit from instruction on how to provide helpful
support to partners who they would like to see abstinent. These
strategies may need to be more communal, with perhaps a greater
focus on being sympathetic or empathic, communicating under-
standing of the difficulty inherent in a quit attempt, and providing
affection. This may be particularly important if other stressors are
experienced simultaneously. The extent to which partners or others
can provide resources to help women appraise and cope with
stressors would likely play an important role in women’s ability to
reduce their smoking. Indeed, stress-buffering support has been
theorized as a possibly important factor in actual cessation and
early maintenance (S. Cohen et al., 1988).
Future research should examine smokers’ perceptions of the
supportiveness of men’s and women’s social influence strategies.
More generally, perceptions of a partner’s support for quitting
have long been associated with success at quitting (Mermelstein,
Lichtenstein, & McIntyre, 1983; Ockene, Benfari, Nuttall, Hur-
witz, & Ockene, 1982; West, Graham, Swanson, & Wilkinson,
1977; for an exception, see Pollak & Mullen, 1997). Relationships
observed between support for quitting and smoking behavior
change, however, have been based on data from small, unrepre-
sentative samples of smokers—for example, individuals in treat-
ment programs or enrolled in worksite programs. Moreover, al-
though direct and indirect social pressures to quit smoking have
been conceptualized as social support (S. Cohen et al., 1988),
whether smokers perceive social pressures to quit or social control
efforts as support for quitting is unknown. Research is thus needed
to closely examine relationships among the nature and frequency
of influence attempts from male and female partners, perceptions
of support, and success in reducing or quitting smoking. Such
research will help elucidate why social control influences are
differentially associated with smoking behavior change as a func-
tion of gender and has the potential to highlight the importance of
gender in a theoretical model explaining the role social relation-
ships play in health.
The present study did not find associations between the moti-
vational influences of a doctor and the reduction in smoking level
from baseline to the 2-day or 4-month follow-up. This surprising
finding should be regarded cautiously, however, because of the
absence of information in this study on the nature and extent of
participants’ interactions with health care providers. In research
designed explicitly to test the impact of a doctor’s influence,
physicians’ messages about smoking’s harmful effects have been
found to increase abstinence rates among self-quitters by up to
10% (Fiore et al., 2000).
Biochemical validation of smoking was not conducted, so there
was no way to ascertain smokers’ true level of smoking after their
the quit date, as a function of quitting to set an example for children and
Reduction in daily cigarette consumption, 4 months following
WESTMAAS, WILD, AND FERRENCE
quit date. Nevertheless, because the study required minimal con-
tact with investigators, the likelihood that participants lied about
their smoking level is substantially diminished (Velicer et al.,
The present study did not independently verify the occurrence
of social control efforts by social network members. We would
expect that participants’ ratings of the influence from these sources
of social control would be correlated with actual influence at-
tempts reported by the network members themselves. Future re-
search should include both participants’ and social network mem-
bers’ reports of influence attempts in order to verify this.
Although we did not collect data on nonparticipants who re-
sponded to advertisements, other analyses suggested that our sam-
ple was comparable to population-based or randomized samples of
smokers. For example, smokers in our full sample had an average
age of 41 years, smoked an average of 18 cigarettes per day, and
began smoking, on average, at age 17. In a random sample
of 8,013 enrollees in a large HMO, Curry, Grothaus, and McBride
(1997) reported an average age of 41.1 years for the 1,127 indi-
viduals who reported current smoking. Their sample reported an
average smoking rate of 17.3 cigarettes per day, and an average
age of 18.2 years for beginning smoking. The gender composition
of our sample is also similar to other studies in which just over half
of all smokers participating were women (e.g., Curry et al., 1997;
Mermelstein, Cohen, Lichtenstein, Baer, & Kamarck, 1986). Al-
though the similarities on these characteristics suggest a fair de-
gree of comparability, replicating our results with a randomized
sample would strengthen confidence in the pattern of results we
obtained. In addition, following participants for a longer period of
time and verifying smoking level for all participants at all
follow-up time points would increase the generalizability of our
findings. Some degree of confidence can be attributed to our
results at the 4-month follow-up, however, because when pre-quit-
date smoking levels were assumed for nonresponders, results re-
mained the same.
Overall, our results suggest gender differences in the effects of
social control influences on smoking behavior change. We have
suggested here that this might be due to differences in influence
strategies that men and women use. Differences in influence strat-
egies among men and women are garnering some research atten-
tion (e.g., Carli, 1999), but such differences have not been studied
in a context that embraces both personal relationships and health.
Within the domain of health behavior change, self-determination
theory (Deci & Ryan, 1985) asserts that the provision of an
autonomy-supportive environment (e.g., by emphasizing choice,
being less controlling, and acknowledging feelings) promotes sus-
tained health behavior change, a proposition that has been sup-
ported by empirical research (Williams, Grow, Freedman, Ryan, &
Deci, 1996; Williams, Rodin, Ryan, Grolnick, & Deci, 1998). The
possibility that there may be gender differences in ability to
provide an autonomy-supportive environment to a partner attempt-
ing to change health behavior has not been examined, however.
Conducting research into the social influence efforts of men and
women should enrich understanding of how social relationships
influence health and why social ties appear to be more protective
of men’s health than of women’s.
American Cancer Society. (2002). Cancer facts and figures. Atlanta, GA:
Aubin, H. J., Laureaux, C., Tilkete, S., & Barrucand, D. (1999). Changes
in cigarette smoking and coffee drinking after alcohol detoxification in
alcoholics. Addiction, 94, 411–416.
Bakan, D. (1966). The duality of human existence: An essay on psychology
and religion. Chicago: Rand McNally.
Barbee, A. P., Gulley, M. R., & Cunningham, M. R. (1990). Support
seeking in personal relationships. Journal of Social and Personal Rela-
tionships, 7, 531–540.
Becona, E., Vasquez, F. L., Fuentes, M. J., & Lorenzo, M. (1999). Anxiety,
affect, depression, and cigarette consumption. Personality and Individ-
ual Differences, 26, 113–119.
Berkman, L. F., & Syme, L. S. (1979). Social networks, host resistance,
and mortality: A nine-year follow-up study of Alameda County resi-
dents. American Journal of Epidemiology, 109, 186–204.
Blazer, D. (1982). Social support and mortality in an elderly community
population. American Journal of Epidemiology, 115, 684–694.
Bollinger, C. T. (2000). Practical experiences in smoking reduction and
cessation. Addiction, 95(Suppl. 1), S19–S24.
Brown, P. C., & Smith, T. W. (1992). Social influence, marriage, and the
heart: Cardiovascular consequences of interpersonal control in husbands
and wives. Health Psychology, 11, 88–96.
Carli, L. L. (1999). Gender, interpersonal power, and social influence.
Journal of Social Issues, 55, 81–99.
Centers for Disease Control and Prevention. (2002). Annual smoking—
attributable mortality, years of potential life lost, and economic costs—
United States, 1995–1999. Morbidity and Mortality Weekly Report, 51,
Cohen, J. (1988). Statistical power analysis for the behavioral sciences
(2nd ed.). Hillsdale, NJ: Erlbaum.
Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation
analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.
Cohen, S., Lichtenstein, E., Mermelstein, R., Kingsolver, K., Baer, J. S., &
Kamarck, T. W. (1988). Social support interventions for smoking ces-
sation. In B. H. Gottlieb (Ed.), Marshaling social support: Formats,
processes, and effects (pp. 211–240). Newbury Park, CA: Sage.
Cohen, S., & Wills, T. A. (1985). Stress, social support, and the buffering
hypothesis. Psychological Bulletin, 98, 310–357.
Conrad, K. M., Flay, B. R., & Hall, D. (1992). Why children start smoking
cigarettes: Predictors of onset. British Journal of Addiction, 87, 1711–
Curry, S. J., Grothaus, L., & McBride, C. (1997). Reasons for quitting:
Intrinsic and extrinsic motivation for smoking cessation in a population-
based sample of smokers. Addictive Behaviors, 22, 727–739.
Curry, S. J., & Marlatt, G. A. (1985). Unaided quitters’ strategies for
coping with temptations to smoke. In S. Shiffman & T. Wills (Eds.),
Coping and substance abuse (pp. 243–265). Orlando, FL: Academic
Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-
determination in human behavior. New York: Plenum.
Derlega, V. J., Barbee, A. P., & Winstead, B. A. (1994). Friendship,
gender, and social support: Laboratory studies of supportive interactions.
In B. R. Burleson, T. L. Albrecht, & I. G. Sarason (Eds.), The commu-
nication of social support: Message, interactions, relationships, and
community (pp. 136–151). Newbury Park, CA: Sage.
Durkheim, E. (1951). Suicide: A study in sociology (J. A. Spaulding & G.
Simpson, Trans.). New York: Free Press. (Original work published in
Eagly, A. H., & Wood, W. (1991). Explaining sex differences in social
behavior: A meta-analytic perspective. Personality and Social Psychol-
ogy Bulletin, 17, 306–315.
EFFECTS OF GENDER IN SOCIAL CONTROL OF SMOKING
Eisenberg, N., & Lennon, R. (1983). Sex differences in empathy and Download full-text
related capacities. Psychological Bulletin, 94, 100–131.
Feingold, A. (1994). Gender differences in personality: A meta-analysis.
Psychological Bulletin, 116, 429–456.
Fiore, M. C., Bailey, W. C., Cohen, S. J., Dorfman, S. F., Fox, B. J.,
Goldstein, M. G., et al. (2000). Treating tobacco use and dependence: A
clinical practice guideline (Agency for Healthcare Research and Quality
Publication No. 00-003). Rockville, MD: U.S. Department of Health and
Gilpin, E., Pierce, J. P., Goodman, J., Burns, D., & Shopland, D. (1992).
Reasons smokers give for stopping smoking: Do they relate to success in
stopping? Tobacco Control, 1, 256–263.
Gove, W. (1973). Sex, marital status, and mortality. American Journal of
Sociology, 79, 45–67.
Guberman, N., Maheu, P., & Maille, C. (1992). Women as family care-
givers: Why do they care? Gerontologist, 32, 607–617.
Hall, J. A. (1984). Nonverbal sex differences: Communication accuracy
and expressive style. Baltimore: Johns Hopkins University Press.
Hall, J. A., & Halberstadt, A. G. (1986). Smiling and gazing. In J. S. Hyde
& M. C. Linn (Eds.), The psychology of gender: Advances through
meta-analysis (pp. 136–158). Baltimore: Johns Hopkins University
Halpern, M. T., & Warner, K. E. (1993). Motivations for smoking cessa-
tion: A comparison of successful quitters and failures. Journal of Sub-
stance Abuse, 5, 247–256.
Heatherton, T. F., Kozlowski, L. T., Frecker, R. C., & Fagerstrom, K.
(1991). The Fagerstrom Test for Nicotine Dependence: A revision of the
Fagerstrom Tolerance Questionnaire. British Journal of Addiction, 86,
Hemard, J. B., Monroe, P. A., Atkinson, E. S., & Blalock, L. B. (1998).
Rural women’s satisfaction and stress as family health care gatekeepers.
Women and Health, 28, 55–77.
House, J. S., Landis, K. R., & Umberson, D. (1988, July 19). Social
relationships and health. Science, 241, 540–545.
House, J. S., Robbins, C., & Metzner, H. L. (1982). The association of
social relationships and activities with mortality: Prospective evidence
from the Tecumseh Community Health Study. American Journal of
Epidemiology, 116, 123–140.
Hughes, J. R. (1992). Tobacco withdrawal in self-quitters. Journal of
Consulting and Clinical Psychology, 60, 689–697.
Hughes, J. R., Gulliver, S. B., Fenwick, J. W., Valliere, W. A., Cruser, K.,
& Pepper, S. (1992). Smoking cessation among self-quitters. Health
Psychology, 11, 331–334.
Hughes, M., & Gove, W. (1981). Living alone, social integration, and
mental health. American Journal of Sociology, 87, 48–74.
Kunkel, A. W., & Burleson, B. R. (1999). Assessing explanations for sex
differences in emotional support. Human Communication Research, 25,
Lewis, M. A., & Rook, K. S. (1999). Social control in personal relation-
ships: Impact on health behaviors and psychological distress. Health
Psychology, 18, 63–71.
Lubinski, D., & Humphreys, L. G. (1990). Assessing spurious “moderator
effects”: Illustrated substantively with the hypothesized (“synergistic”)
relation between spatial and mathematical ability. Psychological Bulle-
tin, 107, 385–393.
Lueptow, L. B., Garovich, L., & Lueptow, M. B. (1995). The persistence
of gender stereotypes in the face of changing sex roles: Evidence
contrary to the sociocultural model. Ethology and Sociobiology, 16,
Mermelstein, R., Cohen, S., Lichtenstein, E., Baer, J. S., & Kamarck, T.
(1986). Social support and smoking cessation and maintenance. Journal
of Consulting and Clinical Psychology, 54, 447–451.
Mermelstein, R. J., Lichtenstein, E., & McIntyre, K. O. (1983). Partner
support and relapse in smoking cessation programs. Journal of Consult-
ing and Clinical Psychology, 51, 465–466.
Moskowitz, D. S., Suh, E. J., & Desaulniers, J. (1994). Situational influ-
ences on gender differences in agency and communion. Journal of
Personality and Social Psychology, 66, 753–761.
Ockene, J. K., Benfari, R. C., Nuttall, R. L., Hurwitz, I., & Ockene, I. S.
(1982). Relationship of psychosocial factors to smoking behavior change
in an intervention program. Preventive Medicine, 11, 13–28.
Orleans, C. T., Jepson, C., Resch, N., & Rimer, B. K. (1994). Quitting
motives and barriers among older smokers. The 1986 Adult Use of
Tobacco Survey revisited. Cancer, 74(Suppl. 7), 2055–2061.
Pollak, K. I., & Mullen, P. D. (1997). An exploration of the effects of
partner smoking, type of social support, and stress on postpartum smok-
ing in married women who stopped smoking during pregnancy. Psy-
chology of Addictive Behaviors, 11, 182–189.
Rose, J. S., Chassin, L., Presson, C. C., & Sherman, S. J. (1996). Prospec-
tive predictors of quit attempts and smoking cessation in young adults.
Health Psychology, 15, 261–268.
Royce, J. M., Corbett, K., Sorensen, G., & Ockene, J. (1997). Gender,
social pressure, and smoking cessation: The community intervention
trial for smoking cessation (COMMIT) at baseline. Social Science and
Medicine, 44, 359–370.
Shiffman, S., Read, L., Maltese, J., Rapkin, D., & Jarvik, M. E. (1985).
Preventing relapse in ex-smokers: A self-management approach. In
G. A. Marlatt & J. R. Gordon (Eds.), Relapse prevention (pp. 472 –520).
New York: Guilford Press.
Tabachnick, B. G., & Fidell, L. S. (1983). Using multivariate statistics.
New York: Harper & Row.
Umberson, D. (1987). Family status and health behaviors: Social control as
a dimension of social integration. Journal of Health and Social Behav-
ior, 28, 306–319.
Umberson, D. (1992). Gender, marital status, and the social control of
health behavior. Social Science and Medicine, 34, 907–917.
Velicer, W. F., Prochaska, J. O., Rossi, J. S., & Snow, M. G. (1992).
Assessing outcome in smoking cessation studies. Psychological Bulletin,
West, D. W., Graham, S., Swanson, M., & Wilkinson, G. (1977). Five-year
follow-up of a smoking withdrawal clinic population. American Journal
of Public Health, 67, 536–544.
Williams, G. C., Grow, V. M., Freedman, Z. R., Ryan, R. M., & Deci, E. L.
(1996). Motivational predictors of weight loss and weight-loss mainte-
nance. Journal of Personality and Social Psychology, 70, 115–126.
Williams, G. C., Rodin, G. C., Ryan, R. M., Grolnick, W. S., & Deci, E. L.
(1998). Autonomous regulation and long-term medication adherence in
adult outpatients. Health Psychology, 17, 269–276.
WESTMAAS, WILD, AND FERRENCE