ARCHIVED - Chronic Diseases in Canada


Volume 30, no. 3, June 2010

Effects of socialization in the household on youth susceptibility to smoking: a secondary analysis of the 2004/05 Canadian Youth Smoking Survey

A. S. H. Schultz, PhD (1); J. Nowatzki, MSc (2); D. A. Dunn, MSN (3); E. J. Griffith, PhD (2, 4)

Author References

  1. Faculty of Nursing, University of Manitoba, Winnipeg, Manitoba
  2. Epidemiology and Cancer Registry, CancerCare Manitoba, Winnipeg, Manitoba
  3. Community and Health Studies, Kwantlen Polytechnic University, White Rock, British Columbia
  4. Department of Community Health Sciences, Faculty of  Medicine, University of Manitoba, Winnipeg, Manitoba

Correspondence: Dr. Annette Schultz, Helen Glass Centre for Nursing, University of Manitoba, 89 Curry Place, Winnipeg, Manitoba  R3T 2N2; Tel.: (204) 258-1311; Fax: (204) 233-7214; E-mail:


Objective: To determine associations between younger youths’ susceptibility to smoking and four household variables related to tobacco socialization: parental and sibling smoking, restrictions on smoking in the home and exposure to smoking in vehicles.

Methods: A secondary analysis of the 2004/05 Canadian Youth Smoking Survey used logistic regression to investigate the relationships between youth susceptibility to smoking, gender, and four household variables related to tobacco socialization. Susceptibility to smoking was operationalized by three levels of smoking experience and intention: non-susceptible non-smoker, susceptible non-smoker and experimenter/smoker. The national survey included 29 243 grade 5 to 9 students from randomly sampled public and private schools in ten provinces.

Results: For non-smokers, the odds of being susceptible to smoking increased with having a sibling who smokes, a lack of a total household smoking ban and riding in a vehicle with a smoker in the previous week, when adjusting for all other variables in the model. These variables also increased the odds of being an experimenter/smoker versus a susceptible non-smoker. Parent smoking status was not significant in these models.

Conclusion: Denormalization messages, through enforced home and vehicle smoking bans, appear to support youth in maintaining a resolve to not smoke, regardless of parental smoking status.

Keywords: youth; smoking susceptibility, socialization; home and vehicle smoking bans; sibling smoking, Canadian Youth Smoking Survey 2004/5.


Preventing youth from smoking continues to be a public health priority.1a-6a Research has typically focused on factors associated with the self-reported smoking behaviour of adolescents 14 years and older. Since most adult smokers report having smoked their first cigarette by the time they were 14 years old,3b thoughts about experimenting with smoking are likely to develop before then. A longitudinal study by Pierce et al. suggests that youths’ perceived susceptibility to try smoking—defined as the degree of resolve to remain a non-smoker versus the likelihood of smoking a puff or a whole cigarette in the future—was strongly associated with future smoking behaviours.6b Therefore, it is important to focus on a younger age group to examine factors associated with both future intentions regarding smoking and smoking behaviours that have already occurred.

The 2004/05 Youth Smoking Survey (YSS) utilized data collected from 29 243 Canadian youth in grades 5 to 9 (age 11 to 15 years).7a There were few established smokers (1.9%) and experimenters (13%) in these grades; of the 85% of youth who were non-smokers, 64% were categorized as being non-susceptible and 36% were categorized as susceptible to experimenting with smoking in the future. Investigation of the differences between non-susceptible non-smokers, susceptible non-smokers, and experimenters/smokers gives us opportunities to explore factors associated with smoking intentions and behaviours.

This study is designed to explore the associations between younger youths’ susceptibility to smoking and household variables related to tobacco socialization. We chose to focus on the household environment because caregivers’ anti-smoking messages, rules and behaviour are particularly critical during this developmental phase when youths’ resolve to remain a non-smoker is being shaped.3c,6c

Living with a parent or sibling who smokes is associated with adolescents (aged 12 to 17 years) establishing smoking behaviour.4b,5b,8a-11 Pierce et al. proposed that parental smoking may be one of the factors that shape children’s cognitive maps regarding the acceptability of smoking.6d As such, it is plausible that family members’ smoking shapes youths’ attitudes towards smoking before smoking initiation.

Caregiver-defined rules that restrict smoking in the home are on the rise,12,13 possibly as a result of evidence of negative health effects from exposure to environmental tobacco smoke (ETS). Not only do anti-smoking actions, such as smoking bans in the home, protect youth from ETS, they are inversely related to smoking rates of older youth, even in homes with a parent who smokes.6e,14,15 Household rules seem to influence youths’ decisions related to experimenting with smoking,8b,16a and youth in homes with total bans were significantly less susceptible to ever trying smoking.16b Thus, home smoking bans are an important factor to consider when examining youth susceptibility to smoking.

In addition to exposing youth to ETS, exposure to smoking in vehicles, much like exposure to smoking in homes, may be sending messages that smoking is acceptable. Thus, parent-enforced vehicle smoking bans are an additional anti-smoking strategy that could influence youths’ decisions regarding tobacco use. This strategy may be particularly useful with children under the age of 15 because they are most likely to be passengers in vehicles driven by their parents or by drivers known to their parents. We found no studies that explored the relationship between youths’ susceptibility to smoking and parental rules related to smoking in vehicles or youth exposure to smoke in vehicles.

Given the importance of the household environment, the current study focuses on associations between youth susceptibility to smoking and the following variables: parental* smoking, sibling smoking, restrictions on smoking in the home and exposure to smoking in vehicles.


This secondary analysis of the 2004/05 YSS, a Health Canada–sponsored national survey of grade 5 to 9 youth, was conducted using logistic regression analysis to determine the association between variables in the household environment and youths’ susceptibility to smoking. Data used to calculate variables were obtained from the Public Use Microdata files for the 2004/05 YSS.7b

Sampling procedure

The YSS study design was based on sampling in two stages and included public and private schools from ten provinces.17a The first stage involved sampling school boards within each province, using stratification based on corresponding adult smoking rates. At stage two, schools were sampled from the selected school boards. A random sample of private schools was also selected from each province. If a board or school declined to participate in the study, a replacement was chosen from a predetermined substitute list. All students in the selected schools were eligible to participate as long as parental consent was given.


The independent variables in our analysis were gender, grade and four household variables related to tobacco socialization: parent smoking status, sibling smoking status, home smoking restrictions and exposure to smoking inside a vehicle. For parent (mother, father, guardian, or other caregiver) smoking status, respondents were classified as being from a family where no parent smokes, one parent smokes, or both parents smoke. For sibling smoking status, respondents were classified as having no siblings who smoke or having at least one sibling who smokes. Home smoking restrictions were defined by responses to the question “What are the rules about smoking in your home?” with possible responses being a total ban (“no one is allowed to smoke in my home”), some restrictions (“only special guests are allowed to smoke in my home” or “people are allowed to smoke only in certain areas in my home”) and no restrictions (“people are allowed to smoke anywhere in my home”). Exposure to people smoking inside a vehicle was determined from responses to “During the past 7 days, on how many days did you ride in a car with someone who was smoking cigarettes?” with possible responses falling into two categories: none (“0 days”) and 1 or more days (“1 or 2 days,” “3 or 4 days,” “5 or 6 days,” or “All 7 days”).

The dependent variable used in this study was youth susceptibility to smoking operationalized through three levels of youth smoking uptake. Measurement of the dependent variable was constructed from responses to two YSS–derived variables on smoking intentions and behaviours. This measure was based on work on youth smoking uptake by Wakefield et al. that used responses about past and current smoking behaviour and intentions regarding future smoking behaviour.16c In the current study, non-susceptible non-smokers had never tried a cigarette and intended not to smoke in the future; susceptible non-smokers had never tried a cigarette and had weak intentions regarding future non-smoking, or they had tried a few puffs of a cigarette and had strong intentions not to smoke in the future; and experimenters and smokers include those youth who had tried a few puffs of a cigarette and had weak intentions regarding future non-smoking or those who, regardless of intentions regarding future smoking, had smoked a whole cigarette or more.

Statistical analysis

Logistic regression analysis was conducted using Stata/SE version 9.2 for Windows. Like Wakefield et al.,16d we performed a threshold of change analysis that allowed independent variables to have varying effects at different threshold levels.18 As such, a multivariate logistic regression was performed at each threshold between the different sequential levels or degrees of smoking susceptibility. (This does not mean that youth will progress through the levels in a direct sequence.) The first regression compares non-susceptible non-smokers to susceptible non-smokers, and the second, susceptible non-smokers to experimenters and smokers.

The weighting of the individual participants compensated for the complex sampling strategy used in the YSS study design. The weights took into consideration the board selection scheme, school selection and student non-response. Individual weights were also calibrated to the province, gender and grade distribution. A set of bootstrap weights were generated to attach to the data.17b


A total of 29 243 grade 5 to 9 students participated in the 2004/05 YSS. The overall response rates for this national survey were as follows: 78% of selected school boards participated, 55% of selected schools participated and 58% of eligible youth in those schools completed a survey. The sample included slightly fewer students in grade 9 than in other grades; there were approximately the same proportion of male and female students responding; and the majority of respondents reported that no parent(s) smoked, no sibling(s) smoked, there was a total ban on smoking in their home, and they had not been exposed to smoking in vehicles in the previous seven days. With regard to smoking susceptibility, 85% reported being non-smokers, with almost two-thirds non-susceptible non-smokers and one-third, susceptible non-smokers. Table 1 shows the characteristics of the respondents.


Table 1
Characteristics of 2004/05 Canadian Youth Smoking Survey respondents
N = 29 243
  n %
Grade at school  
5 5881 20.1
6 6657 22.8
7 5894 20.2
8 5864 20.1
9 4947 16.9
Male 13 911 47.6
Female 15 332 52.4
Number of parents who smoke
0 19 280 66.6
≥ 1 9652 33.4
Number of siblings who smoke 
0 25 717 88.5
≥ 1 3331 11.5
Household smoking restrictions 
Total ban 18 578 65.5
Some restrictions 6551 23.1
No restrictions 3223 11.4
Days exposed to smoking in vehicles in the previous 7 days
0 21 041 72.9
≥ 1 7827 27.1
Stage of smoking susceptibility 
Non-susceptible non-smokera 15 855 54.2
Susceptible non-smokerb 9036 30.9
Smoker or experimenterc 4344 14.9


Abbreviations: N, overall sample size; n, sub-sample size; %, percent; ≥, equals or greater than.

a Non-susceptible non-smokers had never tried a cigarette and intended to not smoke in the future.

b Susceptible non-smokers have never tried a cigarette but nevertheless had weak intentions regarding future non-smoking, or had tried a few puffs of a cigarette and had strong intentions not to smoke in the future.

c Experimenters have smoked a whole cigarette or more, or had tried a few puffs of a cigarette and had weak intentions regarding future non-smoking.



Logistic regression
In preliminary analyses, Spearman’s correlations were calculated for each pair of independent variables. There were statistically significant positive correlations between all of the four household variables related to tobacco socialization. Because of the high correlations, tests were conducted to check for collinearity amongst the independent variables. The variance inflation factors for all independent variables considered for the model were acceptable. However, including grades in the analyses reduced the overall stability of the model below acceptable levels, so grade level was not included in the final model. Table 2 shows the Spearman’s correlations between the independent variables included in the final model.


Table 2
Spearman’s correlations between independent variables

Gender Parent
Household restrictions Days exposed to smoking in vehicles in the previous 7
Parent smoking -.02*      
Sibling smoking -.03** .20**    
Household smoking restrictions  -.0* .50** .15**  
Days exposed to smoking in vehicles in the previous 7 days -.04** .56** .20** .41**


* p < .05 

** p < .01

Variables were defined as follows: Gender: 0 = females, 1 = males; Parent smoking: 0 = no parents/guardians smoke; 1 = 1 parent/guardian smokers; 2 = 2 parents/guardians smoke; Sibling smoking: 0 = no siblings smoke; 1 = 1 or more siblings smoke; Household restrictions: 0 = total ban, 1 = some restrictions, 2 = no restrictions; Days exposed to smoking in vehicles in the last 7 days: 0 = 0 days, 1 = 1 or more days.



Results suggest that the variables that increase the odds of being a susceptible non-smoker versus a non-susceptible non-smoker, when adjusting for the other variables in the model, include having at least one sibling who smokes, having only some restrictions or no restrictions on smoking in the home, and having been recently exposed to smoking in vehicles. Gender and parental smoking are not significantly associated with susceptibility to take up smoking for non-smoking youth when adjusting for all other variables in the model. The same variables increase the odds of being an experimenter or smoker versus a susceptible non-smoker, when adjusting for the other variables in the model. Once again, gender and parental smoking are not significant at this threshold when adjusting for all other variables in the model. Table 3 shows the logistic regression results for each of the youth smoking susceptibility thresholds.


Table 3
Logistic regression thresholds of change analysis
  Susceptible non-smoker thresholda
OR (CI)c
Experimenter/smoker thresholdb
OR (CI)c
Female 1.00 1.00
Male 1.04 (.92-1.16) 1.02 (.85-1.22)
Number of parents who smoke  
0 1.00 1.00
1 1.09 (.90-1.32) 1.02 (.84-1.24)
2 .99 (.77-1.27) 1.12 (.86-1.46)
Number of siblings who smoke
0 1.00 1.00
≥ 1 1.61** (1.24-2.10) 3.10** (2.55-3.77)
Household smoking restrictions
Total ban 1.00 1.00
Some restrictions 1.60** (1.34-1.92) 1.26* (1.06-1.50)
No restrictions 1.46** (1.13-1.90) 1.70** (1.31-2.21)
Days exposed to smoking in vehicles in the previous 7 days
0 1.00 1.00
≥ 1 1.43** (1.18-1.73) 2.34** (1.84-2.97)


Abbreviations: CI, confidence interval; OR, odds ratio; p, p-value.

a Threshold between non-susceptible non-smoker level and susceptible non-smoker level.

b Threshold between susceptible non-smoker level and experimenter/smoker level.

c 95% confidence interval.

* p < .05;

** p < .01




This analysis of the Canadian 2004/2005 YSS has provided some unique evidence concerning youths’ susceptibility to smoking and the smoking socialization that occurs in the household environment: evidence of sibling smoking, some or no household smoking restrictions, and riding in a vehicle with a smoker were each associated with increased odds of being at a higher level or degree of smoking susceptibility when adjusting for the other variables in the model.

There are at least two plausible socialization mechanisms that could provide an explanation for these results.6f,10b First, exposure to smokers may model smoking behaviour, provide positive images of smoking and ease access to cigarettes. Second, total smoking bans in the home and in vehicles may provide denormalization messages that indicate that smoking is neither appropriate nor socially acceptable.

Parental smoking has been linked with higher rates of youth experimenting with smoking,4c,9b,10c though our findings suggest no significant relationship between parental smoking and youth smoking at either of the smoking susceptibility thresholds when adjusting for the other variables in the model. There are at least three points to consider related to this unexpected finding. First, the influence from parental smoking might be exerted through riding in a vehicle with a smoker who is a parental figure in the youths’ life. Second, parental smoking status at younger ages (i.e. grade 3) influences smoking behaviour during adolescence (in grade 12).1b Thus, parental smoking status during adolescence may not be a complete reflection of the influence parental smoking has on children. Third, a total home smoking ban appears to be more strongly associated with youth smoking uptake than parental smoking status,16e evidence for which is supported by our study findings. Perhaps parental anti-smoking actions, such as a home smoking ban, provide strong denormalization messages about smoking irrespective of parental smoking status.

Exposure to older youth who smoke appears to influence the decisions youth make about smoking.4d,5c For example, youths aged 14 to 16 years with siblings who smoked were more likely to report being smokers themselves.4e In addition, Leatherdale et al. found that grade 5 and 6 youth who attended a school with higher smoking rates among grade 8 students were more likely to have smoked a whole cigarette.5d In our study, sibling smoking was linked with respondents’ expectations about future smoking and current smoking behaviours. Though the age of the study participants’ siblings is unknown, based on the young age and the low smoking rates of the YSS participants, it is reasonable to assume that the siblings who smoked were older. Regardless, findings from this study suggest that sibling smoking has a strong association with youths’ smoking behaviours and intentions when adjusting for the other variables in the model.

This study demonstrates that a lack of household smoking restrictions is associated with a higher degree of smoking susceptibility at both tested thresholds when adjusting for the other variables in the model. In other words, living in a home without a full smoking ban or only some smoking restrictions means non-smoking youth are more likely to be susceptible to future smoking and susceptible non-smokers are more likely to be experimenters or smokers. These findings concur with previously reported associations between home smoking bans and youth smoking uptake.8c,16f

Exposure to smoking in vehicles is also strongly associated with being at a higher level of smoking uptake at both thresholds when adjusting for the other variables in the model. This suggests that youth who are passengers in cars where others smoke are more likely to have begun to consider smoking in the future (susceptible non-smoker) or even to experiment with smoking. This new finding raises several questions that require further investigation. Who is smoking in the vehicles—drivers or passengers? Are the drivers the parents of the youth or licensed drivers known to them? What beliefs exist around the effects of smoking in vehicles? A qualitative study of low-income White and Black families reported that decisions to ban smoking in cars was commonly not discussed regardless of the smoking status of adults within the family.19 For families with no smokers, the common underlying rationale for this lack of discussion was that smokers rarely rode in their cars. For families with smokers, there was lack of clarity concerning risks associated with exposure to second-hand smoke in cars. Regardless, banning smoking in vehicles is another anti-smoking measure parents can use to influence their children’s decision related to taking up smoking, even if the parents themselves are smokers.

As such, banning smoking in cars might be more than simply an issue of protecting children from second-hand smoke; it may indicate that smoking is neither acceptable nor appropriate and thus help prevent children from considering or experimenting with smoking in the future. This finding provides evidence that supports legislation that bans smoking in cars when children are passengers; such legislation is currently being implemented or considered for implementation in various places in Canada, the United States and Australia, among others.20


Findings from this study are subject to several limitations. First, cross-sectional survey data cannot provide causal conclusions but can provide evidence to support or dispute identified associations. Findings from this study reflect a comparison between three levels of smoking uptake at two thresholds of change and not a progression along a trajectory of smoking acquisition. Further investigation using longitudinal survey design would refine our understanding of independent variable influences and causation. Next, secondary data analysis commonly raises questions not covered by original survey items. For example, it would have been useful to know if those siblings who smoked were older, if they lived at home, and if they smoked in vehicles in which the respondents were passengers. A third limitation is that the data are based on youth self-reports. In a related study, youth responses were compared to available parent responses. Although the majority of parent and youth responses were in agreement, when discrepancies occurred many produced a non-random pattern of disagreement.21 Therefore, when interpreting the results of the present study, it is important to keep in mind that household variables are measured from the perspective of the youth.


This secondary analysis of the 2004/05 YSS has deepened our understanding of youths’ smoking intentions and behaviours and the associations with socialization in the household. Youths’ attitudes towards smoking and their decisions to smoke will likely be shaped by many factors in their environment, including such things as peer smoking behaviours, school smoking bans, and community smoking norms. Findings from this study suggest that strong denormalization messages in the household environment may serve as a protective mechanism for youth. Smoking bans in the home and in vehicles may provide a means of socialization that support youth in their resolve to remain smoke free, regardless of the parents’ smoking status. Further research is needed to explore the mechanisms underlying the relationships between household factors and youth smoking susceptibility. Developing prevention strategies that focus on influencing actual smoking behaviour of youth along with their susceptibility to smoke in the future would help address the identified public health priority of decreasing youth smoking rates. Health promotion messages directed at parents and guardians could highlight the potential potency of home and vehicle smoking bans.


This research was funded through a Health Canada short contract (#4500142783) within the Office of Research, Surveillance & Evaluation Tobacco Control Programmes. All authors have no competing interests to declare.


* “Parent” signifies a parental figure or caregiver who is the father/mother or guardian or other caregiver.


  1. a,b  Bricker JB, Peterson AV, Andersen MR, Leroux BG, Rajan KB, Sarason IG. Close friends’, parents’, and older siblings’ smoking: reevaluating their influence on children’s smoking. Nicotine Tob Res. 2006;8:217-26.
  2. ^  Eisenberg ME, Forster JL. Adolescent smoking behavior: measures of social norms. Am J Prev Med. 2003;25:122-8.
  3. a,b,c  Elders JM, Perry CL, Eriksen MP, Giovino GA. The report of the Surgeon General: preventing tobacco use among young people. Am J Public Health. 1994;84:543-7.
  4. a,b,c,d,e Komro KA, McCarty MC, Forster JL, Blaine TM, Chen V. Parental, family, and home characteristics associated with cigarette smoking among adolescents. Am J Health Promot. 2003;17:291-9.
  5. a,b,c,d  Leatherdale ST, McDonald PW, Cameron R, Brown KS. A multilevel analysis examining the relationship between social influences for smoking and smoking onset. Am J Health Behav. 2005;29:520-30.
  6. a,b,c,d,e,f  Pierce JP, Choi WS, Gilpin EA, Farkas AJ, Merritt RK. Validation of susceptibility as a predictor of which adolescents take up smoking in the United States. Health Psychol. 1996;15:355-61.
  7. a,b  Health Canada. Youth Smoking Survey 2004-2005: Public Use Microdata. Ottawa (ON): Health Canada, 2006.
  8. a,b,c  Andersen MR, Leroux BG, Bricker JB, Rajan KB, Peterson AV. Antismoking parenting practices are associated with reduced rates of adolescent smoking. Arch Pediatr Adolesc Med. 2004;158:348-52.
  9. a,b  Bricker JB, Leroux BG, Andersen MR, Rajan KB, Peterson AV Jr. Parental smoking cessation and children’s smoking: mediation by antismoking actions. Nicotine Tob Res. 2005;7:501-9.
  10. a,b,c  Conley Thomson C, Siegel M, Winickoff J, Biener L, Rigotti NA. Household smoking bans and adolescents’ perceived prevalence of smoking and social acceptability of smoking. Prev Med. 2005;41:349-56.
  11. ^  Jackson C, Dickinson D. Enabling parents who smoke to prevent their children from initiating smoking: results from a three year intervention evaluation. Arch Pediatr Adolesc Med. 2006;160:56-62.
  12. ^  Ashley MJ, Cohen J, Ferrence R, Bull S, Bondy S, Poland B, Pederson L. Smoking in the home: changing attitudes and current practices. Am J Public Health. 1998;88:797-800.
  13. ^   Levy DT, Romano E, Mumford EA. Recent trends in home and work smoking bans. Tob Control. 2004;13:258-63.
  14. ^  Clark PI, Schooley MW, Pierce B, Schulman J, Hartman AM, Schmitt CL. Impact of home smoking rules on smoking patterns among adolescents and young adults [Internet]. Prev Chronic Dis. 2006 April [cited 2010 Mar 2];3:A41. Available from: pcd/issues/2006/apr/05_0028.htm
  15. ^   Farkas AJ, Gilpin EA, White MM, Pierce JP. Association between household and workplace smoking restrictions and adolescent smoking. JAMA. 2000;284:717-22.
  16. a,b,c,d,e,f  Wakefield MA, Chaloupka FJ, Kaufman NJ, Orleans CT, Barker DC, Ruel EE. Effect of restrictions on smoking at home, at school, and in public places on teenage smoking: cross sectional study. BMJ. 2000;321:333-7.
  17. a,b  University of Waterloo. Youth Smoking Survey (YSS): 2004-05 Microdata User Guide.Waterloo (ON): Centre for Behavioural Research and Program Evaluation, 2006.
  18. ^  Hedeker D, Mermelstein RJ. A multilevel thresholds of change model for analysis of stages of change data. Multivariate Behav Res. 1998;33:427-55.
  19. ^   Kegler MC, Escoffery C, Butler S. A qualitative study on establishing and enforcing smoking rules in family cars. Nicotine Tob Res. 2008;10:493-7.
  20. ^  DeRosenroll M, Cunningham R. Laws banning smoking in vehicles carrying children – international overview. Toronto (ON): Canadian Cancer Society;2009 [cited 2010 Jan 19]. Available from:
  21. ^  Nowatzki J, Schultz ASH, Griffith EJ. Discrepancies between youth and parent perceptions of their household environment relevant to smoking: a secondary analysis of the 2004/05 Canadian Youth Smoking Survey. Chronic Dis Can. 2010;30:78-84.
Report a problem or mistake on this page
Please select all that apply:

Thank you for your help!

You will not receive a reply. For enquiries, contact us.

Date modified: