Health Education Research Advance Access published July 14, 2004 HEALTH EDUCATION RESEARCH Theory & Practice
Pages 10
Anti-smoking socialization beliefs among rural Native American and White parents of young children Michelle C. Kegler1,3 and Lorraine Halinka Malcoe2 Abstract
1
Department of Behavioral Sciences and Health Education, Rollins School of Public Health, Emory University, Atlanta, GA 30322 and 2Department of Family and Community Medicine, Masters in Public Health Program, University of New Mexico, Albuquerque, NM 87131, USA 3 Correspondence to: M. C. Kegler; E-mail:
[email protected]
Introduction Teen smoking rates vary significantly by race/ ethnicity (US Department of Health and Human Services, 1994, 1998). In 2000, 16.8% of AfricanAmerican high school students reported smoking within the past 30 days, compared with 31.8% of White students, 22.6% of Hispanic students and 20.6% of Asian American high school students (Centers for Disease Control and Prevention, 2001). National rates of smoking among Native American youth are not as readily available, but studies suggest that Native American rates are comparable or even higher than White rates (US Department of Health and Human Services, 1994, 1998; Bachman et al., 1999). Among adults, 36.0% of NativeAmericans smoked in 2000 compared with 24.1% of White adults (Centers for Disease Control and Prevention, 2002). Recent studies suggest that anti-smoking socialization by parents may also vary by race/ethnicity (Koepke et al., 1990; Alexander et al., 1999; Clark et al., 1999; Mermelstein et al., 1999; Crawford et al., 2001; Kegler et al., 2002). As described by Jackson et al., anti-smoking socialization includes ‘factors that could influence children’s expectations regarding
Health Education Research Ó Oxford University Press 2004; All rights reserved
doi:10.1093/her/cyg107
Downloaded from http://her.oxfordjournals.org/ by guest on June 9, 2013
This study assesses similarities and differences in anti-smoking socialization beliefs of White and Native American parents in a low-income, rural population in northeastern Oklahoma. Data are from a population-based, cross-sectional children’s environmental health study in which inhome interviews were conducted with 356 parents (56.2% White, 43.8% Native American), primarily mothers, of young children. Approximately 65% of the participants had a high school education or less and over 50% smoked. The Native American participants represented numerous tribes and did not live on reservations. Multivariate logistic regression models were used to examine associations between race/ethnicity, education, smoking status and six antismoking socialization beliefs. Results showed that White and Native American parents in this study were very similar in their anti-smoking socialization beliefs, with the one exception that Native American parents were less likely to believe that schools are better than parents in teaching children about the dangers of cigarette smoking. Parental education was significantly associated with the beliefs that all children will
try smoking and that forbidding children to smoke will only make them want to smoke more, with less-educated parents more likely to share these beliefs. Findings suggest that interventions to promote anti-smoking socialization beliefs among parents with high school education or less may be important in low-income, rural communities with high smoking rates.
M. C. Kegler and L. H. Malcoe
2 of 10
(Jackson and Henriksen, 1997) found that several anti-smoking socialization constructs (e.g. talking to the child about not smoking, the child believed parents would know if s/he smoked, etc.) were related to lower rates of smoking onset in AfricanAmerican and White children. Similarly, Sargent and Dalton (Sargent and Dalton, 2001) found that adolescents who thought their parents would be very upset about their smoking were significantly less likely to start smoking than youth who thought their parents were less disapproving. Further, Dalton and Sargent (Dalton and Sargent, 2001) found that parents who smoked and communicated to their children that they were very much against the child smoking were just as influential as non-smokers in keeping their children from smoking. Similarly, Jackson and Henriksen (Jackson and Henriksen, 1997) found that children whose parents engaged in anti-smoking socialization had lower rates of smoking onset, even when the parents smoked. Previously, it was widely assumed that parents who smoked had little credibility in delivering anti-smoking messages (Jackson and Henriksen, 1997; Dalton and Sargent, 2001). Clark et al. (Clark et al., 1999) examined similarities and differences in parental anti-smoking socialization by smoking status, and found that smokers, more often than non-smokers set ground rules for smoking for their children, but also were more likely to believe all kids will try tobacco and that punishing children is not likely to keep them from trying again. The primary purpose of the current study is to report Native American parents’ anti-smoking socialization beliefs from one region of the US, and to examine whether rural White and Native American parents have similar anti-smoking socialization beliefs. A second purpose of this study is to examine anti-smoking socialization beliefs by education levels and smoking status of the parents.
Methods Study participants Population-based data were collected as part of an outcome evaluation for a children’s environmental
Downloaded from http://her.oxfordjournals.org/ by guest on June 9, 2013
the feasibility, acceptability and consequences of smoking cigarettes’ [(Jackson et al., 1998), p. 437]. In a study of African-American and White parents, Clark et al. (Clark et al., 1999) found that AfricanAmerican parents more often engaged in anti-smoking socialization at home than did White parents. Specifically, African-American parents more often set and communicated ground rules regarding tobacco use for children. They were less likely than White parents to believe that all kids will try tobacco and less likely to believe that forbidding teens to use it will make them want it more. These findings are supported by a multi-site focus group study of 1175 teens that explored ethnic and gender differences in teen smoking (Mermelstein et al., 1999). Mermelstein et al. (Mermelstein et al., 1999) found that African-American, Asian/Pacific Islander and Hispanic youth reported strong anti-smoking messages from their parents. In a deeper analysis of family influences using the same focus group data, researchers noted that several aspects of anti-smoking socialization in the home appeared similar for Native American and White youth, relative to youth from other racial/ethnic backgrounds (Kegler et al., 2002). For example, at least some White and Native American youth discussed that their parents felt it was the teen’s decision as to whether or not to smoke. At present, there is no body of research demonstrating that differing levels of anti-smoking socialization contribute to racial/ethnic differences in teen smoking, although it is an intriguing hypothesis. Evidence is accumulating, however, to support the importance of anti-smoking socialization as a vehicle to prevent youth smoking (Jackson and Henriksen, 1997; Sargent and Dalton, 2001; Jackson and Dickinson, 2003). Until recently, much of the research on smoking onset and parental influences examined the relative influence of peers and parents at various points in the initiation process (Bauman et al., 1990; Flay et al., 1994; Fergusson et al., 1995). Direct modeling of smoking by parents was the primary parent-related construct under examination (Sargent and Dalton, 2001). More recently, researchers are examining other mechanisms through which the family may influence smoking initiation. Jackson and Henriksen
Anti-smoking beliefs among Native American and White parents
Measures Demographics Information was collected on the respondents’ race/ ethnicity (coded White = 0 and Native American =
1) and education level (coded 1 = less than high school, 2 = high school graduate, 3 = some college or vo-tech and 4 = college degree). Additional demographic information included age, gender and marital status. Household income and the number of adults and children supported by this income were also obtained, as was information on whether the respondent received various types of federal or state assistance. Percent of federal poverty level was calculated based on the 2000 poverty thresholds. Tribal enrollment was assessed by asking those who described their race as Native American or American Indian if they were an enrolled member of any American Indian tribe. Those who responded yes were then asked which tribe or tribes they were enrolled with.
Smoking status Smoking status was assessed through two questions. Respondents were first asked, ‘Have you smoked at least 100 cigarettes in your entire life?’. Those who responded yes were then asked, ‘Do you now smoke cigarettes everyday, some days, or not at all?’. Those who had not smoked 100 cigarettes in their lifetime were classified as nonsmokers; those who reported smoking at least 100 cigarettes, but not smoking now were classified as ex-smokers. Parents who reported smoking some days or everyday were classified as smokers. For most analyses, non-smokers and ex-smokers were combined.
Anti-smoking socialization beliefs The anti-smoking socialization belief measures were based on those used by Clark et al. (Clark et al., 1999). The six items were assessed with a four-point Likert scale ranging from strongly disagree to strongly agree. Items included, ‘All kids will try smoking, it’s part of growing up; Punishing children for trying smoking is not likely to keep them from trying it again; Children who see their parents smoking are likely to try smoking; Parents who smoke have difficulty telling their children not to smoke; Schools are better than parents in teaching children about the dangers of 3 of 10
Downloaded from http://her.oxfordjournals.org/ by guest on June 9, 2013
health intervention study (Kegler et al., 2000). Because the primary focus of the evaluation study was childhood lead poisoning, eligible study participants included Native American and White children aged 1–6 years and their caregivers living in or near a Superfund site in rural northeastern Oklahoma. The study area included the largest town in the county (2000 population 13 704), and several additional small towns and unincorporated rural areas. With the exception of the larger town, all residences in the study area were visited by two-person teams to identify and recruit eligible families (i.e. those with a child aged 1–6). In the larger town, city blocks were randomly selected from census block groups. At least three repeat visits were made to each residence if no one was home. Additional strategies were employed to recruit Native American participants, such as attending tribally sponsored local events, actively recruiting through health and social service clinics, and visiting all tribal housing residences. Staff visited 6686 residences in the summer and fall of 2000 with the goal of recruiting 400 eligible families to participate. Staff identified 474 eligible families largely through door-to-door recruitment. A total of 87 caregivers refused to participate, most often due to the blood lead-testing component of the study, leaving a response rate of 81.6%. Trained interviewers conducted in-home interviews with primary caregivers of all families who agreed to participate. The caregiver interview took 45–60 min to administer, and covered knowledge, beliefs and behaviors related to lead-poisoning prevention. There were also questions on other environmental factors that may affect the health of the participant’s child, including exposure to second-hand smoke, and beliefs about second-hand smoke and antismoking socialization. Only Native American and White parents are included in the analyses presented in this paper; the larger sample includes some parents who were not White or Native American and some non-parent caregivers.
M. C. Kegler and L. H. Malcoe cigarette smoking; and If parents forbid children to smoke cigarettes, they will only want to smoke more’. These items were not intended to comprise a scale.
Analysis
Results Study participants The majority of participants were mothers (90.4%), with fathers comprising only 9.6% of the sample (Table I). Because of the inclusion criteria, all participants had at least one child aged 1–6 years. The mean age of the parents was 28.0 years and about 60% were married. The median annual
Table I. Demographic characteristics of study participants Characteristic Gender (N, %) women men Age (mean, SD) Annual household income (US$)(median, range)
Total
Whites
322 (90.4) 34 (9.6) 28.0 (6.78) 16 800 (1 800–48 000)
Native Americans
182 (91.0) 18 (9.0) 28.3 (6.61) 16 800 (1 800–48 000)
140 (89.7) 16 (10.3) 27.6 (7.00) 16 800 (2 400–48 000)
Percent of federal poverty level in 2000 (N, %) 185
181 (52.9) 119 (34.8) 42 (12.3)
Education level (N, %) less than high school high school graduate some college/vo-tech college graduate
92 141 93 30
(25.8) (39.6) (26.1) (8.4)
55 75 48 22
(27.5) (37.5) (24.0) (11.0)
37 66 45 8
(23.7) (42.3) (28.9) (5.1)
50 216 55 33 2
(14.0) (60.7) (15.3) (9.3) (0.6)
28 122 31 19 0
(14.0) (61.0) (15.5) (9.5)
22 94 24 14 2
(14.1) (60.3) (15.4) (9.0) (1.3)
Marital Status (N, %) living with someone married divorced/separated single widowed Smoking status (N, %) current smoker ex-smoker non-smoker Tribal enrollment (N, %) yes no a
Not applicable.
4 of 10
191 (53.8) 38 (10.7) 126 (35.5) –a –a
103 (54.2) 68 (35.8) 19 (10.0)
113 (56.5) 20 (10.0) 67 (33.5) –a –a
78 (51.3) 51 (33.6) 23 (15.1)
78 (50.3) 18 (11.6) 59 (38.1) 103 (66.0) 53 (34.0)
Downloaded from http://her.oxfordjournals.org/ by guest on June 9, 2013
All data were entered into EpiInfo and then validated (entered again and compared) to minimize data entry errors. SAS (version 8) was used for data analysis (SAS Institute, 1999). Bivariate associations between anti-smoking socialization belief variables and race/ethnicity, education, and smoking status were examined using linear trend tests (Cochran–Armitage) to take advantage of the ordinal nature of the belief variables (Agresti, 1990). Multiple logistic regression models were used to
examine the relative contributions of race/ethnicity, education and smoking status to each of the antismoking socialization beliefs. Responses to the belief items were dichotomized (agree versus disagree) for the multiple logistic regression analyses.
Anti-smoking beliefs among Native American and White parents
Bivariate analyses Anti-smoking socialization beliefs by parental race/ethnicity Table II presents parental anti-smoking socialization beliefs by race/ethnicity and tribal enrollment of the parent. Overall, Native American and White parents had very similar beliefs about socializing children against cigarette smoking. Roughly 60% (58.8% of Whites; 57.7% of Native-Americans) kind of agreed or strongly agreed that all kids would try smoking. Similarly, about 60% (59.6% of Whites; 64.9% of Native Americans) agreed that forbidding children from smoking cigarettes would only make them want to smoke more. Native American and White parents also had similar views on the efficacy of punishment as a deterrent to smoking; about 75% felt that punishing children for trying cigarettes was not likely to keep them from trying again. Parents’ views on the impact of parental smoking on children were also similar by race/ethnicity: close to 90% felt that children who see their parents smoking are likely to try it themselves. In addition, about 70% felt parents who smoke have a difficult time teaching their children about the dangers of smoking. The only significant difference between Native American and White parents was in their views about schools being better than parents in teaching children about the dangers of cigarette smoking. White parents were more in agreement with this statement than Native-
American parents (P = 0.037). No significant differences were found between Native Americans enrolled in a tribe and those not enrolled, but Native American parents who were enrolled tribal members tended to believe more strongly than non-enrolled parents that parents who smoke have difficulty telling their children not to smoke (P = 0.051).
Anti-smoking socialization beliefs by parental education level Anti-smoking socialization varied more by education level of the parent than by parent’s race/ ethnicity (Table III). Parent’s education was strongly associated with the belief that all kids will try smoking. Over 60% percent of parents with a high school education or less agreed with this statement compared with 26.7% of parents with a college education (P < 0.001). Striking differences by education were also evident in the percent of parents who felt that forbidding children from smoking would make them want to smoke more. About three-quarters of parents with less than a high school education agreed with this statement, in contrast to only 40.0% of parents with a college degree (P < 0.001). Less-educated parents were also more likely to believe that punishing children for smoking would be ineffective (P = 0.013).
Anti-smoking socialization beliefs by parental smoking status All but one of the anti-smoking socialization beliefs varied significantly by smoking status of the parent (Table 3). Based on tests of linear trend, smokers more than non-smokers somewhat agreed or strongly agreed that all kids will try smoking (P = 0.019), that forbidding children from smoking would be more likely to make them want to smoke more (P = 0.048) and that punishing children for trying smoking was not likely to keep them from trying again (P = 0.007). Interestingly, non-smoking parents were more likely than parents who smoked to believe that smokers have a difficult time telling their children not to smoke (P < 0.001). Associations between parental smoking status and anti-smoking socialization beliefs were also examined separately by parent’s race/ethnicity (data not 5 of 10
Downloaded from http://her.oxfordjournals.org/ by guest on June 9, 2013
household income was $16 800 and 53% lived below the 2000 federal poverty level. Approximately one-quarter of participants had not completed high school and 8.4% had a college degree. A relatively high proportion (53.8%) were current smokers and 10.7% were ex-smokers. Forty-four percent of the respondents described themselves as Native American and, of these, 66% were enrolled members of a tribe. Native American participants were enrolled in 19 different tribes and none currently lived on a reservation. There were no significant differences between Native-American and White parents in smoking status or any of the sociodemographic variables.
M. C. Kegler and L. H. Malcoe Table II. Associations between parental anti-smoking beliefs, race/ethnicity and tribal enrollment Belief
Total Race/ethnicity of parent (n = 356)
Tribal enrollment (Native Americans only)
White Native American Pa (n = 200) (n = 156)
No (n = 53)
Yes (n = 103)
0.731
Pa
23.6 17.6 32.2 26.6
25.0 17.3 24.4 33.3
18.9 20.8 26.4 34.0
28.2 15.5 23.3 33.0
0.473
If parents forbid children to smoke cigarettes, they will only want to smoke more strongly disagree kind of disagree kind of agree strongly agree
14.5 23.6 31.8 30.1
14.7 25.8 32.8 26.8
14.3 20.8 30.5 34.4
0.230 15.1 26.4 26.4 32.1
13.9 17.8 32.7 35.6
0.411
Punishing children for trying smoking is not likely to keep them from trying again strongly disagree kind of disagree kind of agree strongly agree
15.2 10.4 30.6 43.8
15.5 10.0 33.5 41.0
14.7 10.9 26.9 47.4
0.538 18.9 9.4 20.8 50.9
12.6 11.7 30.1 45.6
0.786
Children who see their parents smoking are likely to try smoking strongly disagree kind of disagree kind of agree strongly agree
5.3 7.6 27.0 60.1
5.0 7.5 26.5 61.0
5.8 7.7 27.6 59.0
0.678
9.4 9.4 26.4 54.7
3.9 6.8 28.2 61.2
0.165
Parents who smoke have difficulty telling their children not to smoke strongly disagree kind of disagree kind of agree strongly agree
19.7 11.5 25.3 43.5
19.5 14.0 24.0 42.5
19.9 8.3 26.9 44.9
0.555 26.4 13.2 22.6 37.7
16.5 5.8 29.1 48.5
0.051
Schools are better than parents in teaching children about the dangers of cigarette smoking strongly disagree kind of disagree kind of agree strongly agree
33.8 19.0 25.6 21.6
29.6 17.9 29.1 23.5
39.1 20.5 21.2 19.2
0.037 45.3 22.6 17.0 15.1
35.9 19.4 23.3 21.4
0.148
a
Based on the linear trend test.
shown). Associations with parental smoking were very similar for Native American and White parents.
Multivariate analyses Results from the multiple logistic regression analyses of the dichotomized anti-smoking beliefs are presented in Table IV. Two-way interaction terms 6 of 10
were included in the initial models, but were not significant and are not reported here. Education level of the parent remained associated with two beliefs in multivariate models. Education was significantly associated with the belief that all kids will try smoking, it’s part of growing up [odds ratio (OR) = 0.62, 95% confidence interval (CI)
Downloaded from http://her.oxfordjournals.org/ by guest on June 9, 2013
All kids will try smoking, it’s part of growing up strongly disagree 24.2 kind of disagree 17.5 kind of agree 28.7 strongly agree 29.6
Anti-smoking beliefs among Native American and White parents
Discussion Our study was the first to examine similarities and differences in anti-smoking socialization beliefs among rural Native American and White parents with young children. Overall, Native American and White parents in our study had very similar antismoking socialization beliefs, with the only exception being that Native Americans were less likely to believe that schools are better than parents in teaching children about the dangers of cigarette smoking. Previous qualitative research with Native
American and White teens had suggested that they may experience similar anti-smoking environments in the home, with Native American and White parents feeling lower levels of parental efficacy in keeping children from smoking cigarettes compared with African-American parents (Mermelstein et al., 1999; Kegler et al., 2000). Our findings are generally consistent with Clark et al.’s (Clark et al., 1999) data on White parents, suggesting that both Native American and White parents may have less stringent anti-smoking socialization beliefs than African-American parents. Consistent with findings from other studies, we found that smokers and non-smokers had different beliefs about anti-smoking socialization (Clark et al., 1999). Our bivariate results showed that smokers were less likely than non-smokers to believe that punishing children for smoking or forbidding children to smoke would have the intended results. In multivariate analyses, these associations remained borderline significant. Further, smokers in our study were less likely than nonsmokers to believe that parents who smoke have difficulty telling their children not to smoke. Because recent research shows that parents who smoke and engage in anti-smoking socialization practices are influential in decreasing their children’s susceptibility to smoking, encouraging smokers to act on this belief (i.e. smokers can tell their children not to smoke) presents a significant intervention opportunity (Jackson and Henriksen, 1997; Sargent and Dalton, 2001; Jackson and Dickinson, 2003). Another intervention opportunity stems from our finding that parents with less education are the least likely to have strong anti-smoking beliefs. After controlling for race/ethnicity and smoking status, education level of the parent was associated with the beliefs that all kids will try smoking and that forbidding children from smoking will only make them want to smoke more. Interventions designed to promote anti-smoking socialization beliefs and practices among parents with low levels of education are needed. Lastly, the finding that Native American parents were less likely than Whites to believe that schools were better than parents at teaching children about the dangers of smoking, suggests that 7 of 10
Downloaded from http://her.oxfordjournals.org/ by guest on June 9, 2013
0.48–0.79]. As parental education level increased, the estimated odds of having this belief decreased after controlling for race/ethnicity and smoking status. Education was also significantly associated with the belief that children will only want to smoke more if their parents forbid them to smoke. As education level of the parent increased, the estimated odds of having this belief decreased (OR = 0.65, 95% CI 0.51–0.84). After controlling for parental education and smoking status, race/ethnicity remained significantly associated with the belief that schools are better than parents in teaching children about the dangers of cigarette smoking (OR = 0.60, 95% CI 0.39–0.93). Native American parents were less likely than White parents to have this belief. After controlling for parental education and race/ ethnicity, smoking status only remained significantly associated with the belief that parents who smoke have difficulty telling their children not to smoke. Smokers agreed with this belief less than non-smokers (OR = 0.30, 95% CI 0.18–0.50). In the full model, none of the predictor variables were significantly related to the belief about the effectiveness of punishment in deterring children from trying to smoke again, although smokers tended to agree with this belief more than non-smokers (OR = 1.53, 95% CI 0.93–2.52). Likewise, there were no significant associations with the belief that children who see their parents smoke are likely to try it themselves, but smokers tended to agree with this belief less than non-smokers (OR = 0.56, 95% CI 0.29–1.09).
M. C. Kegler and L. H. Malcoe Table III. Associations between parental anti-smoking beliefs, parental education level and parental smoking status Beliefs
Education level Less than high school
Smoking status High school graduate
All kids will try smoking, it’s part of growing up strongly disagree 14.3 24.1 kind of disagree 17.6 12.1 kind of agree 22.0 34.0 strongly agree 46.2 29.8
Some college/vo-tech
College graduate
31.2 18.3 31.2 19.4
33.3 40.0 16.7 10.0
Pa
Smoker
28.7 17.1 31.1 23.2
20.0 17.9 26.8 35.3
