Beyond modeling: Parenting practices, parental smoking history, and adolescent cigarette smoking

Addictive Behaviors 29 (2004) 17 – 32

Beyond modeling: Parenting practices, parental smoking history, and adolescent cigarette smoking Molly Middlecamp Kodl*, Robin Mermelstein Department of Psychology, University of Illinois at Chicago, USA Health Research and Policy Centers, University of Illinois at Chicago, 850 W. Jackson Blvd., Suite 400, Chicago, IL 60607, USA

Abstract This study explored parental beliefs and behaviors designed to convey an antismoking message across levels of self-reported parent and adolescent smoking behavior. Parental self-efficacy, beliefs about smoking, the family relationship, antismoking messages, reactions to smoking, and household smoking rules were explored. Participants were 345 6th, 8th, and 10th graders (50% female; 93% White) and their parents (86% mothers). Beliefs about smoking, parental efficacy, and rules varied across levels of both parental and adolescent smoking. Parents with a history of smoking and parents of adolescents who had tried smoking were less efficacious, held weaker antismoking beliefs, and less often reported household smoking rules. Overall, examining parental behavior across levels of both parent and youth smoking is important. Similarly, parental efficacy, not previously studied in relation to parenting about smoking, may be important to target in future interventions. D 2003 Elsevier Ltd. All rights reserved. Keywords: Smoking; Adolescents; Parenting; Parent – child relationships

1. Introduction Parents and the family context play important roles in youth smoking. Both active attempts by parents to deter cigarette smoking with antismoking messages or clear consequences for smoking and more passive familial factors such as the nature of the

* Corresponding author. Health Research and Policy Centers, University of Illinois at Chicago, 850 W. Jackson Blvd., Suite 400, Chicago, IL 60607, USA. Tel.: +1-312-996-1940; fax: +1-312-996-2703. E-mail address: [email protected] (M.M. Kodl). 0306-4603/$ – see front matter D 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0306-4603(03)00087-X

18

M.M. Kodl, R. Mermelstein / Addictive Behaviors 29 (2004) 17–32

family relationship have consistently been associated with youth smoking (e.g., Biglan, Duncan, Ary, & Smolkowski, 1995; Chassin, Presson, Sherman, Montello, & McGrew, 1986; Cohen & Rice, 1997; Cohen, Richardson, & LaBree, 1994; Doherty & Allen, 1994; Henriksen & Jackson, 1998; Kandel & Wu, 1995; Radziszewska, Richardson, Dent, & Flay, 1996). Adolescents themselves also acknowledge the important role that parents play in providing smoking-related messages (Mermelstein, 1999). Although the evidence for the link between parenting practices and youth smoking is increasing, less is known about how parenting practices vary by parental smoking experiences or by adolescents’ own patterns of experimentation. The purpose of this study was to examine how parenting practices and smoking-related attitudes, along with the family relationship, vary by parental smoking experience and by stage of adolescent cigarette use. The present study goes beyond previous work by considering multiple stages of use for both parents and adolescents, collecting data about parenting practices directly from the parents, rather than examining parental practices based solely on child report (e.g., Henriksen & Jackson, 1998), and simultaneously examining several key family variables. This study also includes the smoking behaviors of adolescents, rather than focusing on early adolescents just entering the early stages of experimentation (e.g., Fearnow, Chassin, Presson, & Sherman, 1998; Jackson & Henriksen, 1997). There are many ways in which parents can attempt to influence their child’s decision to smoke. Parents can attempt to convey antismoking messages to their child overtly, such as by providing antismoking messages, or more covertly, for example, by establishing household antismoking policies. Both overt and covert parental practices have been found to be effective. For example, children and early adolescents who report parental antismoking and antidrug use statements are less likely either to use substances or to report future intentions to smoke cigarettes (Bailey, Ennett, & Ringwalt, 1993; Chassin, Presson, Todd, Rose, & Sherman, 1998; Henriksen & Jackson, 1998; Jackson, 1997; Jackson & Henriksen, 1997). Even general communication, in the context of a positive parent–child relationship, is associated with lower rates of alcohol and tobacco use (Cohen et al., 1994; Kafka & London, 1991). Establishing household rules about cigarette smoking and consequences for using cigarettes is another way to convey antismoking messages to youth. Recent studies have found that home smoking bans are associated with decreased rates of youth smoking (Chassin, Presson, Todd et al., 1998; Farkas, Gilpin, White, & Pierce, 2000; Wakefield et al., 2000). Established consequences of smoking are also important. Jackson (1997) and Jackson and Henriksen (1997) found that a lower expectation of being punished for smoking was related to increased rates of smoking initiation in children. Overall, these studies support the importance of household smoking policies and established consequences for smoking. Parents’ behaviors, such as setting household smoking rules or making antismoking statements, are likely to be a function of their own experiences with smoking, their smokingrelated beliefs and attitudes, their confidence in their ability to influence their child’s behavior (i.e., their self-efficacy), and the overall family environment. Self-efficacy for parenting, both in general and specific to conveying information about smoking, may affect adolescent smoking directly and affect a parent’s antismoking socialization practices. Parental self-

M.M. Kodl, R. Mermelstein / Addictive Behaviors 29 (2004) 17–32

19

efficacy encompasses the mastery of the attitudes, the skills, and the behaviors essential to exerting control over various parenting and family roles (Schunk, 1990). A parent’s sense of efficacy to influence his or her child’s smoking by, for example, discussing the issue is likely to be related both to a parent’s own smoking experiences and to the types of parenting they engage in. To date, no study has examined parental self-efficacy in relation to antismoking socialization and to child smoking. We expected, however, that parental efficacy would be influenced by parental smoking history and would also influence child smoking. It is also important to consider the family environment because it provides an overall context or background within which parent–child interactions take place and has also been independently related to adolescent smoking behavior. Low family cohesion, poor familial functioning, and family disunion have been related to increased rates of adolescent cigarette use (Doherty & Allen, 1994). Inadequate parental monitoring (Biglan et al., 1995; Radziszewska et al., 1996) and perceived parental permissiveness (Cohen & Rice, 1997; Kandel & Wu, 1995) have also been consistently associated with higher rates of adolescent substance use. One of the limitations of much of the work on parenting practices and youth smoking has been the lack of consideration of variations in parental smoking history. For example, some studies have examined parenting practices and parent smoking based on child reports of parental smoking (Henriksen & Jackson, 1998; Jackson & Henriksen, 1997), or examined only recent smoking by parents (e.g., in the last year) (Kandel & Wu, 1995). The consideration of different levels of parental smoking across the lifetime may help to account for some of the discrepancies in the literature about the link between parent and child smoking, which may have been the result of different conceptualizations of parent smoking, indirect assessments of parent smoking, or only measuring current use (Distefan, Gilpin, Choi, & Pierce, 1998; Landrine, Richardson, Klonoff, & Flay, 1994; McNeill et al., 1988). However, considering different levels or stages of parental smoking across the lifetime may be important. Both Bauman, Foshee, Linzer, & Koch (1990) and Bailey et al. (1993) found that lifetime parental smoking, even when parents have not smoked during the child’s lifetime, is an important predictor of adolescent smoking. Given these findings, it may be reasonable to hypothesize that levels of parental use may influence parental attitudes, efficacy, and antismoking socialization behaviors. For example, parents who have a history of youthful experimentation and who did not progress to regular smoking may feel that adolescent smoking is a ‘‘rite of passage’’ and may not feel as concerned or set rules about what they perceive as their child’s harmless experimentation. Parents who are ex-smokers or former experimenters may not feel confident about their ability to set rules or consequences for their child’s smoking given their own history. If this is the case, then it may be important to target these parents in intervention efforts. The present study was designed to test the hypothesis that parenting practices and attitudes would vary in an ordered fashion by parental smoking history. We considered four levels of use: never smoker, former experimenter, former regular smoker, and current smoker. We hypothesized that parents who were current smokers would report the lowest levels of antismoking beliefs, self-efficacy, and antismoking socialization practices, followed in order by former smokers, former experimenters, and never smokers. We

20

M.M. Kodl, R. Mermelstein / Addictive Behaviors 29 (2004) 17–32

expected that the more general variables of the family relationship and parental monitoring would similarly vary by parental smoking. This study goes beyond prior work in this area by also examining how parental factors vary by stage of adolescent smoking. Adolescent smoking is frequently considered to progress through stages. Mayhew, Flay, and Mott (2000) outline six stages of smoking onset: (1) nonsmoking with no intention to smoke; (2) nonsmoking with an intention to smoke; (3) trier; (4) experimenter; (5) regular user; and (6) established/daily user or dependent smoker. This study focused on adolescents in the earlier (i.e., up to and including regular users) stages of use. Of interest was whether parenting practices varied by stage of adolescent smoking. We hypothesized that parental attitudes and practices, as well as the family environment, would fall in an ordered manner across the stages of use with parents of nonsmokers/nonintenders having the highest antismoking attitudes and practices, followed in order by parents of youth in each of the stages.

2. Method 2.1. Procedure The overall design of this study was a cross-sectional survey of 6th, 8th, and 10th grade students and their parents. Adolescents completed brief in-class surveys administered by research staff. Written parental permission for the survey was obtained for the 6th and 8th grade participants. For 10th graders, a waiver of written parental consent was granted, and parents who did not want their child to participate notified the research team (2% declined). Parent data were collected by mailed questionnaires. Parents of students who participated in the in-class survey received a letter requesting their participation in the project, the parent questionnaire, and a stamped return envelope. Parents who did not initially respond were sent a second questionnaire. No substantive differences existed between those parents who responded to the first mailing (n = 291) and those who responded to the second (n = 54) so data were combined. Parents received a US$10 coupon to a local supermarket for their participation. 2.2. Participants Self-report surveys were completed by 573 6th, 8th, and 10th students. These completion rates represented 63% of all 6th and 8th grade students in one school and 94% of all 10th grade students in one high school. The difference in response rates between schools is most likely a reflection of the difference in the parental consent procedures used. From these participants, 44 students were excluded due to missing data or for having a sibling in the study. Parent questionnaires were then mailed to the remaining 529 students who completed the in-class survey. The mail survey was returned by 345 parents. Considering only parent–student dyads with complete data, 46% of potential 6th and 8th graders and 83% of 10th graders were included in

M.M. Kodl, R. Mermelstein / Addictive Behaviors 29 (2004) 17–32

21

the study sample. Thus, the final sample for the study was 345 adolescents and a matched parent. Of parents, 96% were non-Hispanic white, 2.6% were Asian, and 1.4% were other ethnicities. Parents averaged 45.4 years of age and 86% of the parent responders were mothers. The parent sample was well educated and of middle- to upper-middle class background. Sixty-seven percent (n = 226) had received at least a college degree, 21% completed some college, and 12% had only a high school degree. Sixty percent had a yearly household income of at least US$85,000. The student sample of 345 adolescents was comprised of 72 sixth graders, 57 eighth graders, and 216 tenth graders, with equal numbers of males (n = 171) and females (n = 172); two students did not provide gender information. Students averaged 15.1 years in age. Approximately 93% were non-Hispanic white (2.3% were Asian; Black, Latino, American Indian, Arabic, and Bi or Multi-racial students each made up less than 1% of the sample). Most (84%) lived with both parents. 2.3. Measures 2.3.1. Adolescent measures 2.3.1.1. Smoking status. Survey questions addressed whether students had tried cigarettes, patterns of smoking in the past 30 and 90 days, and intentions to smoke in the future. Child smoking behavior was classified into five categories: (1) Never users, not susceptible to smoking in the future. Based on Pierce, Choi, Gilpin, Farkas, and Merritt (1996), students in this group had never tried smoking a cigarette, answered ‘‘No’’ to the question, ‘‘do you think that you will try a cigarette soon,’’ and answered ‘‘Definitely Not’’ to the questions, ‘‘if one of your best friends were to offer you a cigarette, would you smoke,’’ and ‘‘at any time during the next year, do you think you will ever smoke a cigarette.’’ (2) Never users, susceptible to smoking in the future. Again, based on criteria used by Pierce et al. (1996), students in this group had never tried smoking a cigarette (said ‘‘No’’ to ‘‘have you ever tried smoking a cigarette’’), but did not respond with ‘‘No’’ to the question, ‘‘do you think that you will try a cigarette soon,’’ or with ‘‘Definitely Not’’ to the questions, ‘‘if one of your best friends were to offer you a cigarette, would you smoke,’’ and ‘‘at any time during the next year, do you think you will ever smoke a cigarette.’’ (3) Former triers. Former triers were adolescents who had tried cigarette smoking in the past, but not during the preceding 90 days (responded with ‘‘0 days’’ to the question, ‘‘think about the last three months. How many days did you smoke or try a cigarette.’’ (4) Current experimenters. Students in this group had tried cigarette smoking and responded with ‘‘1 to 5 days’’ on the question, ‘‘think about the last three months. How many days did you smoke or try a cigarette.’’ (5) Regular users. Regular users were those students who had tried smoking and who reported at least weekly smoking over the past month (responded ‘‘6 to 9 days’’ or more days to the question, ‘‘think about the past 30 days. How many days did you smoke or try cigarettes.’’

22

M.M. Kodl, R. Mermelstein / Addictive Behaviors 29 (2004) 17–32

2.3.2. Parent measures 2.3.2.1. Smoking status. Parents’ smoking behavior was classified into the categories of never smokers, former experimenters, former regular smokers, and current smokers based on responses to several questions regarding lifetime smoking history. (1) Never smokers. Parents who answered ‘‘No’’ to the question, ‘‘have you ever tried smoking a cigarette, even one or two puffs,’’ were classified as never smokers. (2) Former Experimenters. Experimenters were those parents who endorsed having tried a cigarette but who responded either that they had smoked fewer than 100 cigarettes (said ‘‘No’’ to the item, ‘‘have you smoked 100 or more cigarettes in your lifetime,’’) or had never smoked on a regular basis (said ‘‘No’’ to the question, ‘‘have you ever smoked cigarettes on a daily basis,’’). (3) Former regular smokers. Former smokers had smoked a minimum of 100 cigarettes in their lifetime, endorsed smoking on a daily basis in the past (said ‘‘Yes’’ to the question, ‘‘have you ever smoked cigarettes on a daily basis), and said ‘‘No’’ to the question, ‘‘do you currently smoke cigarettes on a regular basis.’’ Of the former smokers, 70% quit more than 10 years ago and 20% quit between 5 and 10 years prior. (4) Current smokers. Current smokers were parents who endorsed current smoking (said ‘‘Yes’’ to the question, ‘‘do you currently smoke cigarettes on a regular basis,’’) and who had smoked a minimum of 100 cigarettes. 2.3.2.2. Parental self-efficacy. Self-efficacy for parenting was measured using 11 questions from the general parental efficacy scale developed by Elder, Eccles, Ardelt, and Lord (1995). Efficacy for parenting and influencing child behavior in different domains (e.g., helping their child stay out of trouble in school) is emphasized. Additionally, seven items assessing parents’ efficacy to influence their child’s smoking-related behavior were developed for this study (e.g., confidence in their ability to prevent their child from smoking, enforce house rules about smoking, and talk to their child about smoking). Item response options ranged from 1 (not at all confident) to 10 (extremely confident). Scale scores were the mean of each respondent’s answers to the items, and higher scores were representative of greater efficacy (Cronbach’s a=.91). 2.3.2.3. Beliefs about youth smoking. Parents’ beliefs about youth smoking were assessed with nine items drawn from Chassin, Presson, Rose, and Sherman (1998) and rationally developed for this study. Items assessed parents’ (1) health related beliefs about youth smoking and (2) beliefs about the normative nature of youth smoking. Item responses ranged from 1 (strongly disagree) to 5 (strongly agree). Higher scores were indicative of stronger pro-smoking beliefs. Scale scores were the mean of each respondent’s answers to the items (Cronbach’s a=.79). 2.3.2.4. The family relationship. The positive nature of the parent–child relationship was assessed with four items based on those used by Cohen et al. (1994) (e.g., ‘‘I hug my child,’’). Family involvement was assessed with three items developed by Biglan et al.

M.M. Kodl, R. Mermelstein / Addictive Behaviors 29 (2004) 17–32

23

(1995) that measured the degree to which parents agreed with statements such as, ‘‘our family members support one another.’’ Responses for all items ranged from 1 (strongly disagree) to 5 (strongly agree) and scale scores were the mean of responses for all seven items. Higher scores were indicative of a more positive parent – child relationship. Cronbach’s a for the overall scale was .85. 2.3.2.5. Explicit messages about smoking. Eleven items (based on Henriksen & Jackson, 1998, and rationally derived) were used to measure whether and how often parents made verbal statements about smoking to their adolescent (e.g., ‘‘smoking cigarettes is bad for you,’’ ‘‘smoking is addictive’’). Item responses ranged from 1 (no, never) to 4 (many times). Parents’ scale scores were the mean of responses to all items, and higher scores indicated a greater frequency of messages (Cronbach’s a=.88). 2.3.2.6. Reactions to youth smoking. Nineteen items developed from our own work and from Chassin, Presson, Todd et al. (1998) were used to assess parental reactions to youth smoking. After a series of factor analyses were conducted, two subscales were retained. The first factor, behavioral consequences, was comprised of four items (e.g., ground child, take away privileges) and had a coefficient a of .73. The second factor, emotional responses, was comprised of four items (e.g., be disappointed, be worried) and had a coefficient a of .81. Responses ranged from 1 (not at all likely) to 5 (very likely). Scores on these factors were based on the mean of parental responses, and higher scores indicated that parents were more likely to react in the ways described. 2.3.2.7. Parental monitoring. Parental monitoring of their child’s activities was measured using five items developed by Griesler and Kandel (1998) and Biglan et al. (1995). Items included how often children go where they are told not to and how well parents know their child’s friends. Responses ranged from 1 (strongly disagree) to 5 (strongly agree). Higher scores indicated that parents monitored more, Cronbach’s a=.74. 2.3.2.8. Rules. Parents indicated whether they had any of six household rules about smoking. Responses were classified into two categories for analysis: whether parents had household smoking rules, scored as 1 (e.g., ‘‘no one may smoke in the home’’), or did not have rules about smoking in the house (e.g., ‘‘anyone may smoke in the home’’), scored as 0.

3. Results 3.1. Comparison of responders and nonresponders In order to determine whether parents who responded to the mail survey differed from nonresponders, their child’s responses to an item on the in-class survey asking about parent smoking were compared. Parents who smoked, as reported by their children, were less likely

24

M.M. Kodl, R. Mermelstein / Addictive Behaviors 29 (2004) 17–32

to respond to the mailings (55% returned a survey) than were parents who did not smoke (69% returned a survey), v2(1, N = 529) = 9.06, P < .01. Parental response rates also varied by child smoking experience. Of those students who had tried smoking, fewer (57%) of their parents responded than did parents of children who had not tried smoking (72%), v2(1, N = 530) = 14.20, P < .001. Thus, the current study sample under represents parents who smoke and parents of children who have tried smoking. 3.2. Relationship between child and parent smoking As expected, child smoking behavior varied across parental smoking status, v2(12, N = 345) = 33.73, P < .01, see Table 1. Among parents who had ever smoked regularly (current and former smokers), 49% had a child who had ever smoked compared to only 32% of the children of parents who never smoked regularly. Interestingly, parents were largely unaware of the smoking behavior of their child. For example, only 36% of the parents (n = 8) of adolescent regular smokers identified their child as smoking. Only one other parent, a parent of an experimenter, identified their child as smoking. 3.3. Correlations among parent variables Based on scale distributions, two parent scales, beliefs and family relationships, were transformed using either inverse of scores or square root. These transformations changed the direction of interpretation such that after transformation, high scores meant stronger antismoking beliefs and less positive family relationships, respectively. Correlations among study variables can be seen in Table 2. As expected, ratings of parental self-efficacy and beliefs about youth smoking were significantly positively correlated, and both were significantly associated with more positive ratings of the family relationship, providing more messages, the presence of household rules, and more often reacting, both emotionally and behaviorally, to youth smoking. More positive ratings of family relationship were also significantly related to more frequent parental messages and monitoring of their child’s activities. Surprisingly, reactions, monitoring, and rules were unrelated. Table 1 Child smoking status by parental smoking status Child status

Nonsusceptible Susceptible Former trier Experimenter Regular user

Parent smoking status Never (n = 54)

Experimenter (n = 150)

Former (n = 118)

Current (n = 23)

Total n

29 (53.7%) 11 (20.4%) 7 (13.0%) 5 (9.3%) 2 (3.7%)

57 (38.0%) 41 (27.3%) 24 (16.0%) 23 (15.3%) 5 (3.3%)

39 (33.1%) 23 (19.4%) 27 (22.9%) 20 (16.9%) 9 (7.6%)

8 2 1 6 6

133 77 59 54 22

v2 (12, N = 345) = 33.73, P < .01.

(34.8%) (8.7%) (4.3%) (26.1%) (26.1%)

M.M. Kodl, R. Mermelstein / Addictive Behaviors 29 (2004) 17–32

25

Table 2 Intercorrelations among parental practices Variable

1

2

(1) (2) (3) (4) (5) (6) (7) (8)

–

.36** –

Self-efficacy Beliefs Family relations Messages Emotional reactions Behavioral reactions Monitoring Rules

3

4 .46** .29** –

5 .16** .20** .13* –

6 .20** .14* .02 .13** –

.17** .20** .07 .28** .43** –

7

8 .44** .32** .47** .04 .01 .07 –

.23** .21** .09 .11 .10 .10 .02 –

Correlations are two-tailed Pearson correlations. N ranged from 341 to 344. *P < .05. **P < .01.

3.4. Variations by parental smoking status To examine differences in ratings of parental beliefs and behavior by parental smoking status, a MANCOVA with the dependent variables of parental efficacy, beliefs about youth smoking, ratings of the family relationship, messages, emotional and behavioral reactions to youth smoking, monitoring practices, and household smoking rules and an independent variable of parental smoking status was conducted.1 Child grade and parental level of education were used as covariates since these variables were related either to parental smoking or the dependent variables.2 We hypothesized that parent variables would fall in an ordered manner from current smokers to never smokers. For example, we hypothesized that never smokers would have the highest efficacy scores, followed by experimenters, former smokers, and finally by current smokers, who we expected to have the lowest levels of efficacy. Analyses indicated a significant multivariate effect for parent smoking status, Pillai’s F(24,984) = 4.86, P < .001. Step down procedures were used to further examine the significant main effect of parental smoking status, controlling for child grade, and parental education. Variables were entered in the following order: parental efficacy, beliefs, ratings of the family relationship, messages, emotional reactions, behavioral reactions, monitoring, and rules. Parental efficacy, beliefs, 1

Given the significant correlations among the dependent variables of interest, MANCOVA procedures were used to test for variations by parent and child smoking history. To further examine significant main effects, conservative step down F tests were used. Step down procedures test the highest priority dependent variable through ANOVA and test remaining dependent variables through a series of ANCOVAs that control for higher priority variables. Theoretically, self-efficacy, beliefs, and family relationship are more global constructs that were expected to account for some of the variation in parental behavior. Consequently, these variables had the highest entry priority. Additionally, these variables, especially efficacy, were highly correlated with the other variables of interest. It is important to note that while this was a conservative procedure, results obtained for each dependent variable were similar to those obtained when examining between-group differences using one-way ANOVAs. 2 We also examined potential relationships between child and parent gender on the independent and dependent variables. Neither parent nor student gender was significantly related and they were not included as covariates.

26

M.M. Kodl, R. Mermelstein / Addictive Behaviors 29 (2004) 17–32

ratings of the family relationship, emotional reactions, and household smoking rules were all related to parental smoking status, after controlling for higher priority dependent variables and covariates, see Table 3. Parental messages, F(3,330) = 2.25, ns, behavioral reactions, F(3,328) < 1, ns, and monitoring, F(3,327) = 1.29, ns, were not related to smoking status after controlling for higher priority dependent variables and covariates. To further explore our hypotheses about the patterns of responses by level of parental smoking, three contrasts were conducted for each dependent variable: (1) those with a lifetime smoking history versus those without (nevers and experimenters versus former and current smokers), (2) never smokers versus all others, and (3) current smokers versus all others (see Table 3). As expected, those with a lifetime history of smoking were lower in efficacy, had weaker antismoking beliefs, were less likely to react in emotional ways, and were less likely to have household smoking rules, as compared to never smokers and experimenters. Additionally, those who were current smokers were the least efficacious and the least likely to have household smoking rules. 3.5. Variations in parental behavior by child smoking status To examine differences in ratings on parent variables by child smoking behavior, a MANCOVA with the dependent variables of parental efficacy, beliefs about youth smoking, ratings of the family relationship, messages, reactions to youth smoking, monitoring practices, and household smoking rules and an independent variable of child smoking status was conducted. Child grade, parental level of education, and parental smoking status were used as covariates.2 We hypothesized that the parent variables would differ in an ordered manner by child smoking status, such that the parents of more regular adolescent smokers would report the lowest levels of parenting practices. Analyses indicated a significant multivariate effect for child smoking status, Pillai’s F(32,1308) = 2.90, P < .001.

Table 3 Adjusted means of parental beliefs and behaviors by parental smoking status Variable

Never smoker

Experimenter

Former

M (n = 54)

M (n = 150)

M (n = 118)

Efficacy Beliefs Family relations Emotional reactions Rules

8.54 0.40 1.24 4.18 0.80

8.04 0.37 1.27 4.06 0.78

7.82 0.37 1.21 3.73 0.76

Current M (n = 23)

Step down F

Significant contrasts

7.24 0.36 1.21 3.71 0.20

7.18*** 6.75*** 3.14* 3.54* 14.49***

1, 2, 3 1, 2 1, 2 1, 2, 3

Means were adjusted for higher priority variables and for parent education and child grade. The following contrasts were conducted: (1) never smokers and experimenters compared to formers and currents smokers, (2) never smokers versus all others, and (3) current smokers versus all others. Contrasts significant at P < .05 or better are listed. Degrees of freedom ranged from (3,333) to (3,326). *P < .05. ***P < .001.

M.M. Kodl, R. Mermelstein / Addictive Behaviors 29 (2004) 17–32

27

Table 4 Adjusted means of parental beliefs and behaviors by child smoking status Variable

Nonsusceptible

Susceptible

Former trier

Current experimenter

Regular user

M (n = 133)

M (n = 77)

M (n = 59)

M (n = 54)

M (n = 22)

Efficacy Beliefs Monitoring Rules

8.42 0.38 4.19 0.79

8.14 0.37 4.10 0.65

7.80 0.37 4.04 0.81

7.58 0.37 3.89 0.67

6.74 0.35 3.62 0.54

Step down F

Significant contrasts

8.78*** 2.77* 5.59** 3.28*

1, 1, 1, 1,

2, 3 3, 4 2, 3 4

The following contrasts were conducted: (1) regular smokers compared to all others, (2) current experimenters versus regular smokers, (3) never triers (nonsusceptible and susceptible) versus those who have tried, and (4) nonsusceptible versus susceptible students. Contrasts significant at P < .05 or better are listed. Degrees of freedom ranged from (4,331) to (4,324). *P < .05. **P < .01. ***P < .001.

Step down procedures were used to further examine the significant main effect of child smoking status, controlling for child grade, parental education, and parental smoking. Variables were entered in the following order: parental efficacy, beliefs, ratings of the family relationship, messages, emotional reactions, behavioral reactions, monitoring, and rules. Table 4 presents the variables that differed significantly by child smoking status—parental efficacy, beliefs, monitoring, and rules. Parental ratings of the family relationship, messages, and reactions were unrelated to child smoking (means not presented) after controlling priority dependent variables and covariates. As can be seen in Table 4, there was an ordered relationship between each of the significant dependent variables and child status. To further examine this ordering, four contrasts were conducted for each dependent variable: (1) regular adolescent smokers compared to all others, (2) current experimenters versus regular smokers, (3) never triers (nonsusceptible and susceptible) versus those who have tried, and (4) nonsusceptible versus susceptible students. As expected, the parents of regular smokers reported less efficacy, more pro-smoking beliefs, less monitoring of their child’s activities, and fewer household smoking rules than did parents of other adolescents. Regular smokers also differed significantly from current experimenters in terms of parental efficacy and beliefs. Students who had ever tried smoking (e.g., regular users, current experimenters, and former triers) also differed from those who had not tried (nonsusceptible and susceptible). Parents of nontriers reported higher levels of efficacy, fewer pro-smoking beliefs, and more monitoring of their child’s activities than did parents of triers.

4. Discussion One of the strongest predictors of adolescent smoking has been parental smoking (e.g., Flay et al., 1994; Rowe, Chassin, Presson, & Sherman, 1996). Not surprisingly, in the present study, children who had at least one parent who was a current smoker were two times more

28

M.M. Kodl, R. Mermelstein / Addictive Behaviors 29 (2004) 17–32

likely to have experimented with smoking and two and a half times more likely to go beyond initial experimentation. Even children whose parents did not currently smoke, but who were former smokers, had an elevated risk for smoking. These findings add to those of Bauman et al. (1990) and Bailey et al. (1993) supporting the importance of parental lifetime smoking, not just current use, as an influential factor in youth smoking. The primary purpose of the present study was to go beyond merely documenting the link between parent and child smoking and examine how parental behaviors designed to convey an antismoking message vary according to levels or patterns of parent and adolescent cigarette use. Overall, we found significant differences in parental behavior and smokingrelated beliefs across patterns of both adult and youth smoking. As expected, both parental ratings of self-efficacy and beliefs about youth smoking were related to parent and child smoking status. Parents with a lifetime history of smoking, even at low experimental levels, were less likely to feel efficacious in their parental roles and held weaker antismoking beliefs than parents who had never smoked. Similarly, even after controlling for parental smoking, children who had tried smoking and those who were in more regular stages of use had parents who felt less efficacious and held weaker antismoking beliefs. The relationships between parental beliefs, efficacy, and smoking status is important because it suggests that parents’ values and views of their abilities are related not only to their own behaviors, but also to the behavior of their children. These findings differ from those of Chassin, Presson, Rose et al. (1998) who found that maternal beliefs were related to maternal smoking, but that maternal beliefs did not influence either their child’s own health beliefs or smoking. In terms of parental self-efficacy, considering the correlations among parental efficacy and parental behavior (e.g., providing messages, establishing consequences for smoking, setting household rules), it may be that efficacy is only indirectly related to child smoking and that parental behaviors (e.g., rule setting and setting consequences for smoking) actually mediate the relationship between parental efficacy and child smoking behavior. Due to the cross-sectional nature of the present study, we were unable to appropriately test this possible mediational link. The presence of household smoking rules also differed by both parent and child smoking behavior. Similar to other researchers (e.g., Farkas et al., 2000; Wakefield et al., 2000), we found that adolescents who were more regular smokers, as compared to all other adolescents, lived in households that were less likely to have established rules about smoking. Additionally, parents with any lifetime history of regular smoking were less likely to have household smoking rules then were parents without a history of regular smoking. Thus, household smoking rules were related to the smoking behavior of both adults and children. In terms of other familial factors, the patterns that emerged in relation to parent smoking status differed from those observed for child smoking behavior. For example, parents with a lifetime history of regular smoking, and not just current smokers, were less likely to say that they would have strong emotional reactions to their child’s smoking. Thus, a parent’s smoking history influences his/her reaction to smoking by his/her child. However, parental reactions to a child’s possible smoking were unrelated to the child’s smoking status. Considering that only a minority of the parents were aware of their child’s smoking, most parents’ responses to the questions of how they would respond to their child’s smoking were

M.M. Kodl, R. Mermelstein / Addictive Behaviors 29 (2004) 17–32

29

hypothetical in nature. Thus, parents may not have conveyed potential consequences of smoking to their child. It is also possible that a parent’s emotional reactions do not deter a child’s experimentation. Parental monitoring differed according to level of child smoking but did not differ by parental history of smoking. While parents’ smoking history did not seem to impact their monitoring practices, differences in monitoring behavior were noted across levels of child smoking. Parents of both regular smokers and ever triers reported lower levels of monitoring than parents of adolescents who had never tried smoking. Generally, the perception of supportive, engaged parenting, including monitoring, has been associated with reduced rates of youth smoking and experimentation (Cohen & Rice, 1997; Glendinning, Shucksmith, & Hendry, 1997; Jackson, Henriksen, Dickinson, Messer, & Robertson, 1998; Kandel & Wu, 1995; Radziszewska et al., 1996). Thus, the findings of the present study are in line with the conclusions of previous studies. One surprising finding was that the frequency or type of messages that parents reported providing to their adolescent did not differ by parent smoking behavior. Chassin, Presson, Todd et al. (1998) similarly found that self-reported antismoking messages did not differ by parent smoking status. In contrast to the results of other studies (e.g., Chassin, Presson, Todd et al., 1998; Jackson & Henriksen, 1997), however, the present study did not find an association between parental antismoking statements and child smoking behavior. The methodology of the present study, which examined parental reports of antismoking communication rather than adolescents’ perceptions, could be one reason for the disparity. Child perceptions of parent’s antismoking socialization may well differ by child smoking (Henriksen & Jackson, 1998), and future studies should incorporate both child and parental perspectives. The relatively low rates of parental smoking and the high prevalence of reported antismoking messages from parents in the present study could also help to explain discrepancies with other studies. Self-reported parenting practices may also be subject to social desirability bias. It is important to note several potential limitations of this investigation. First, adult smoking rates in the current sample are much lower than national averages, reflecting the lower response rates from smoking parents and the high SES of the sample. The lower prevalence rate of smoking in the present study, coupled with a relatively high level of parental education and lack of ethnic and cultural diversity, suggests that results may not be generalizable across socioeconomic status and ethnicity. This is especially likely given the differences in antismoking socialization, the family environment, and the potential differences in parental efficacy across ethnicity (Clark, Scarisbrick-Hauser, Gautam, & Wirk, 1999; Griesler & Kandel, 1998). The cross-sectional design of the present study also prevented an examination of reciprocal effects between child smoking and the development of parental behaviors or vice versa. The data in this study were not sufficient to examine whether parental behaviors were precursors to adolescent smoking or whether adolescent experimentation with smoking served as a catalyst for antismoking parenting. A longitudinal study of parental behavior, parental smoking, and youth smoking is necessary to address the important issue of bidirectionality of influence. Furthermore, our data came primarily from mothers. Although parental behavior did not vary by parental gender in the present study, examining parental smoking and parenting practices by all parents who play a role of an adolescent’s life is important.

30

M.M. Kodl, R. Mermelstein / Addictive Behaviors 29 (2004) 17–32

Contradictory behaviors and messages within families could account for discrepancies between parental reports of parenting and child smoking. Overall, the methodology in the present study offered an advantage over previous studies by assessing parental attitudes and behavior through parental self-reports while subsequently examining self-reported smoking behavior among adolescents. This methodology is in contrast to some previous studies (e.g., Henriksen & Jackson, 1998; Jackson & Henriksen, 1997) that have only asked children about their perceptions of parental behavior or examined only parental projections about their future behavior (Fearnow et al., 1998). Additionally, this study examined differences in parental behavior across multiple levels of both parent and adolescent smoking. There are several important findings from this study. First, it appears that any history of regular smoking among parents is important. Parents with a lifetime history of regular smoking hold different beliefs and engage in different behaviors than parents without this history, as expected. Similarly, despite the fact that few parents accurately described their child’s smoking behavior, there were differences in the actions that parents took and the beliefs they held across child smoking history. Patental behavior was related to adolescent smoking. This suggests that family factors and household conditions can be protective against cigarette experimentation among youth. Importantly, this study failed to find any support for the conclusions of previous studies that parental verbal, explicit antismoking messages, in and of themselves, influence adolescent smoking behavior, at least by parental self-report. It appears that parental messages cannot be separated from other parental factors, such as smoking-related beliefs, parental selfefficacy, and other parental behavior. Consequently, smoking prevention programs directed at parents should focus on more than just teaching parents to tell their children that smoking is ‘‘bad’’ for them. Finally, this study highlighted the potentially important role of parental selfefficacy in the domain of adolescent substance use prevention. Future investigations and interventions should consider parental self-efficacy when working with parents because of its direct and indirect links to both adolescent and parent smoking. Acknowledgements This research was funded in part by grant CA80266 from the National Cancer Institute and by a grant from the Tobacco Etiology Research Network, funded by the Robert Wood Johnson Foundation. Portions of this research have been presented at the annual meeting of the Society for Research in Nicotine and Tobacco and the biennial meeting of the Society for Research in Adolescence. References Bailey, S., Ennett, S., & Ringwalt, C. (1993). Potential mediators, moderators, or independent effects in the relationship between parents’ former and current cigarette use and their children’s cigarette use. Addictive Behaviors, 18, 601 – 621.

M.M. Kodl, R. Mermelstein / Addictive Behaviors 29 (2004) 17–32

31

Bauman, K., Foshee, V., Linzer, M., & Koch, G. (1990). Effect of parental smoking classification on the association between parental and adolescent smoking. Addictive Behaviors, 15, 413 – 422. Biglan, A., Duncan, T., Ary, D., & Smolkowski, K. (1995). Peer and parental influences on adolescent tobacco use. Journal of Behavioral Medicine, 18, 315 – 330. Chassin, L., Presson, C., Rose, J., & Sherman, S. (1998). Maternal socialization of adolescent smoking: Intergenerational transmission of smoking-related beliefs. Psychology of Addictive Behaviors, 12, 206 – 216. Chassin, L., Presson, C., Sherman, S., Montello, D., & McGrew, J. (1986). Changes in peer and parental influence during adolescence: Longitudinal versus cross-sectional perspectives on smoking initiation. Developmental Psychology, 22, 327 – 334. Chassin, L., Presson, C., Todd, M., Rose, J., & Sherman, S. (1998). Maternal socialization of adolescent smoking: The intergenerational transmission of parenting and smoking. Developmental Psychology, 34, 1189 – 1201. Clark, P., Scarisbrick-Hauser, A., Gautam, S., & Wirk, S. (1999). Anti-tobacco socialization in homes of AfricanAmerican and White parents, and smoking and nonsmoking parents. Journal of Adolescent Health, 24, 329 – 339. Cohen, D., & Rice, J. (1997). Parenting styles, adolescent substance use, and academic achievement. Journal of Drug Education, 27, 199 – 211. Cohen, D., Richardson, J., & LaBree, L. (1994). Parenting behaviors and the onset of smoking and alcohol use: A longitudinal study. Pediatrics, 94, 368 – 375. Distefan, J., Gilpin, E., Choi, W., & Pierce, J. (1998). Parental influences predict adolescent smoking in the United States, 1989 – 1993. Journal of Adolescent Health, 22, 446 – 474. Doherty, W., & Allen, W. (1994). Family functioning and parental smoking as predictors of adolescent cigarette use: A six-year prospective study. Journal of Family Psychology, 8, 347 – 353. Elder, G., Eccles, J., Ardelt, M., & Lord, S. (1995). Inner-city parents under economic pressure: Perspectives on the strategies of parenting. Journal of Marriage and the Family, 57, 771 – 784. Farkas, A., Gilpin, E., White, M., & Pierce, J. (2000). Associations between household and workplace smoking restrictions and adolescent smoking. Journal of the American Medical Association, 284, 717 – 722. Fearnow, M., Chassin, L., Presson, C., & Sherman, S. (1998). Determinants of parental attempts to deter their children’s cigarette smoking. Journal of Applied Developmental Psychology, 19, 453 – 468. Flay, B.R., Hu, F.B., Siddiqui, O., Day, L.E., Hedeker, D., Petraitis, J., Richardson, J., & Sussman, S. (1994). Differential influence of parental smoking and friends’ smoking on adolescent initiation and escalation of smoking. Journal of Health and Social Behavior, 35, 248 – 265. Glendinning, A., Shucksmith, J., & Hendry, L. (1997). Family life and smoking in adolescence. Social Science & Medicine, 44, 93 – 101. Griesler, P., & Kandel, D. (1998). Ethnic differences in correlates of adolescent cigarette smoking. Journal of Adolescent Health, 23, 167 – 180. Henriksen, L., & Jackson, C. (1998). Anti-smoking socialization: Relationship to parent and child smoking status. Health Communication, 10, 87 – 101. Jackson, C. (1997). Initial and experimental stages of tobacco and alcohol use during late childhood: Relation to peer, parent, and personal risk factors. Addictive Behaviors, 22, 685 – 698. Jackson, C., & Henriksen, L. (1997). Do as I say: Parent smoking, antismoking socialization, and smoking onset among children. Addictive Behaviors, 22, 107 – 114. Jackson, C., Henriksen, L., Dickinson, D., Messer, L., & Robertson, S. (1998). A longitudinal study predicting patterns of cigarette smoking in late childhood. Health Education and Behavior, 25, 436 – 447. Kafka, R., & London, P. (1991). Communication in relationships and adolescent substance use: The influence of parents and friends. Adolescence, 26, 587 – 598. Kandel, D., & Wu, P. (1995). The contributions of mothers and fathers to the intergenerational transmission of cigarette smoking in adolescence. Journal of Research on Adolescence, 5, 225 – 252. Landrine, H., Richardson, J., Klonoff, E., & Flay, B. (1994). Cultural diversity in the predictors of adolescent cigarette smoking: The relative influence of peers. Journal of Behavioral Medicine, 17, 331 – 346. Mayhew, K., Flay, B., & Mott, J. (2000). Stages in the development of adolescent smoking. Drug and Alcohol Dependence, 59, S61 – S81.

32

M.M. Kodl, R. Mermelstein / Addictive Behaviors 29 (2004) 17–32

McNeill, A., Jarvis, M., Stapleton, J., Russell, M., Eiser, J., Gammage, P., & Gray, E. (1988). Prospective study of factors predicting uptake of smoking in adolescents. Journal of Epidemiology and Community Health, 43, 72 – 78. Mermelstein, R. (1999). Explanations of ethnic and gender differences in youth smoking: A multi-site, qualitative investigation. Nicotine and Tobacco Research, 1, 591 – 598. Pierce, J., Choi, W., Gilpin, E., Farkas, A., & Merritt, R. (1996). Validation of susceptibility as a predictor of which adolescents take up smoking in the United States. Health Psychology, 15, 355 – 361. Radziszewska, B., Richardson, J., Dent, C., & Flay, B. (1996). Parenting style and adolescent depressive symptoms, smoking, and academic achievement: Ethnic, gender, and SES differences. Journal of Behavioral Medicine, 19, 289 – 305. Rowe, D., Chassin, L., Presson, C., & Sherman, S. (1996). Parental smoking and the ‘‘epidemic’’ spread of cigarette smoking. Journal of Applied Social Psychology, 26, 437 – 454. Schunk, E. (1990). Introduction to the special section on motivation and efficacy. Journal of Educational Psychology, 82, 3 – 6. Wakefield, M., Chaloupka, F., Kaufman, N., Orleans, C., Barker, D., & Ruel, E. (2000). Effect of restrictions on smoking at home, at school, and in public places on teenage smoking: Cross sectional study. British Medical Journal, 321, 333 – 337.