Health Education Research Advance Access originally published online on August 2, 2006
Health Education Research 2006 21(5):674-687; doi:10.1093/her/cyl071
Smoking status moderates the contribution of social-cognitive and environmental determinants to adolescents' smoking intentions
Research Group for Stress, Health and Well-being, Department of Psychology, University of Leuven, Belgium
*Correspondence to: A. Victoir. E-mail: An.Victoir{at}psy.kuleuven.be
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionConflict of interest statementAppendixReferences |
|---|
In this study, it was tested whether attitudes, self-efficacy, social influences and the perception of the school and home environments had different associations with intentions for adolescent non-smokers, occasional smokers and daily smokers. A regression model allowing for separate slopes of social-cognitive and environment variables accounted for 72% of the variation in intentions. For non-smokers, ease of refusing to smoke (ß = –0.06) and social influences favouring smoking (ß = 0.05) were linked to intentions. Occasional and daily smokers' intentions were associated with health consequences (ß = –0.05 and ß = –0.06, respectively) and ease of smoking/buying cigarettes (ß = 0.05 and ß = 0.24, respectively). Social influences favouring smoking (ß = 0.10) were also associated with intentions in daily smokers. In an extended model for current smokers (adjusted R2 = 0.45), context-cued nicotine cravings (ß = 0.27) were linked to daily smokers', but not occasional smokers' intentions. The results suggest that motivating adolescents to abstain from or to quit smoking implies working on different combinations of determinants in non-smokers, occasional smokers and daily smokers. Interventions for daily smokers should supplement motivational techniques with stratagems that allow smokers to reduce the number of cravings they experience in specific contexts.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConflict of interest statementAppendixReferences |
|---|
Adolescent smoking is an important concern to public health. Many young people start smoking at an early age [1] and continue to smoke as adults. Therefore, smoking prevention among adolescents is a major public health target.
Personalized interventions (e.g. computer-tailored messages) can prevent smoking initiation [2] and help smokers to quit smoking [3], but are costly in terms of development and dissemination [4]. Alternatively, adolescents who share characteristics can be targeted by interventions matched to their profile [5]. For example, if daily smokers are not confident they can quit smoking [6], interventions in this group could include easy-to-implement tips on quitting. A first step to this approach would be to identify different types of smokers and non-smokers, either by clustering adolescents into groups with similar psychological and/or smoking-specific variables [7, 8] or by testing whether pre-defined groups differ on variables that predict uptake, maintenance or abstinence of smoking.
With respect to the latter, a variety of social and intra-personal influences have been linked to adolescent smoking [9]. The relationship between these influences and their impact on smoking intentions and behaviour are represented in models, such as the Theory of Planned Behaviour [10]. In this model, behaviour-specific determinants are organized in two tiers. Attitudes, norms and perceived behavioural control contribute to an intention to smoke or abstain from smoking. Intentions are most proximal to actual behaviour, but can be overruled by barriers (actual control).
There are two ways in which individuals or groups vary with regard to behavioural determinants. First, individuals may assume different positions on social-cognitive dimensions. For example, attitudes can range from extremely negative to extremely positive. Groups can have identical scores on a particular variable, but should have unique profiles when all determinants are considered. This condition is central to stage theories of behavioural change [11, 12], but also applies to categorizations of people who would benefit from different interventions. Groups defined by current smoking effectively occupy different positions on social-cognitive dimensions. Adolescent non-smokers, occasional smokers and regular smokers have different smoking attitudes, perceive different norms and differ in the control they have over (not) smoking [13]. For example, young smokers are less likely to think that smoking is addictive than non-smokers, more likely to think that smoking hurts only if you inhale and more likely to state that parents and friends approve of smoking [14, 15].
Second, the contribution of a determinant to intentions and behaviour may vary for people who belong to different groups. In the Theory of Planned Behaviour, the relative importance of attitudes, norms and perceived control can vary across behaviours, situations and persons [10, 16].
There are several methods to determine the relative importance of determinants in different groups. The strongest test would be to establish in intervention studies which groups benefit from manipulations that focus on different determinants. However, most smoking prevention programmes for adolescents are multi-focused and have minor effects in general [17, 18]. It is possible that some subgroups of the target population benefit more from specific programme components than other groups, but effect evaluations usually are not that fine grained.
Alternatively, longitudinal research can reveal which determinants predict the course of a smoking career. Becoming an experimenter (as opposed to merely trying cigarettes) is linked to friends' influence, smoking prevalence estimates, smoking intentions, school grade and use of other substances. Becoming a regular smoker (as opposed to experimenting) is linked to parental smoking and family conflict [19]. Transitions from one group to another are also associated with specific combinations of determinants, as was shown for progression to smoking in respondents who had different smoking intentions at baseline [20], respondents with different smoking frequencies [21] or respondents in different stages of change [22, 23]. Wetter et al. [24] used yet another method. They predicted smoking status 4 years after they had measured social-cognitive determinants, and did this for three separate categories: baseline never smokers, baseline occasional smokers and baseline daily smokers. In occasional smokers, long-term smoking was predicted by positive reinforcements from smoking, appetite/weight control and smoking to control affect. In daily smokers, strategies for affect control and number of cigarettes smoked per day were predictive of smoking at follow-up. Non-smokers' behaviour was linked to sibling smoking. Hedeker et al. [25] focused on intentions, and showed that the relative contribution of attitudes and subjective norms to the prediction of intentions varied between individuals.
Finally, determinants can be investigated cross-sectionally to determine which variables are linked to group membership or which variable has the strongest link with intentions at that point in time. Tyc et al. [26] looked at adolescent non-smokers with low or high intentions to start smoking, and smokers with low or high intentions to quit. The odds of ending up in adjacent categories were predicted by partially unique patterns of variables, suggesting that social-cognitive determinants do get different weights in these groups.
In summary, social-cognitive factors may relate in different ways to intentions and ultimate behaviour in different groups. However, it is difficult to conclude which determinants have the highest impact in specific groups. The reasons are 2-fold. First, the criteria used to define groups vary. Typologies refer to future orientations, such as motivation to reduce or quit smoking [26, 27], to susceptibility to smoking [28], time of initiation [29] or current smoking [30]. Some authors combine criteria, such as the combination of past smoking and future smoking plans [21] or plans for change and actual change efforts [22, 31]. Moreover, nominally identical categories may differ and categories with different labels can be very similar. Second, a determinant may assume different shapes and functions [32]. For example, peer influences may translate to external pressure and modelling in non-smokers and occasional smokers, while for regular smokers, peers figure as ‘stimuli’ that elicit nicotine craving. It is however important to investigate variations in determinant weight in order to design effective interventions.
Hypotheses and aim of the study
Our main hypothesis was that adolescents who differ in current smoking status would accord different weights to determinants of future smoking intentions. The categories selected were non-smokers, occasional (i.e. less than daily) smokers and daily smokers, because they represent common patterns of smoking in adolescence. Data from the 2001/2002 Health Behaviour of School-Aged Children survey revealed that across Europe, 84% of the 11–15 year olds do not smoke, 9% smoke less than daily and 7% smoke daily [1]. The Flemish 2002 data showed that most 11–12 year olds are non-smokers. About 13% of students aged 13 or older smoke occasionally. Daily smoking increases from about 5% in 13–14 year olds to 
30% in 17–18 year olds [33].
As regards the relative importance of determinants over groups, we expected attitudes, peer influences and perceived ease of refraining from smoking to be more strongly linked to smoking intentions in non-smokers. In contrast, we expected the strength of those relationships to be lower in occasional smokers, for whom we expected parental smoking models [19] and perceived ease of buying and smoking cigarettes [34] to be important. For daily smokers, we anticipated that social-cognitive determinants would explain relatively little variability in smoking intentions. Frequent smokers are well aware of the habitual and addictive aspects of their smoking behaviour [35]. Regular smoking is deterred by regulations at home [36] and at school [37] and confined to locations where users cannot be sanctioned, such as the school parking lot [38]. The consistent linkage between environmental cues and nicotine administration leads to context-specific cravings [39, 40]. We expected daily smokers' intentions to reflect the cue-reactivity of their smoking pattern.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConflict of interest statementAppendixReferences |
|---|
Procedure
Three schools, located in three different medium-sized towns in Flanders (Belgium), participated in this study. Two schools were asked to participate by school health counsellors. One school enlisted in the study in order to plan a smoking prevention initiative. In return, all schools received school-specific reports on pupils' behavioural and determinant profiles.
Data were gathered in collective sessions supervised by school health counsellors. Teachers were asked not to attend or—if their presence was required for practical reasons—to remain inconspicuous. Participation rates were not recorded in detail because it was agreed with students that name-based data (including participation) would not be collected. Participation was estimated by the health counsellors and school staff to exceed 90%. Responses were collected anonymously; neither teachers nor counsellors had access to data. Most respondents completed one questionnaire, but 631 respondents filled out a second questionnaire on drug use. Completing a single questionnaire took 40 min. Pupils who filled in two questionnaires did so in one session lasting 100 min.
Participants
Data were collected from 1233 secondary school students (335 boys, 885 girls, 13 subjects did not indicate gender) aged 13–22 years (M = 16.1 years, SD = 1.92). Respondents were enrolled in general (n = 205), technical (n = 737) and vocational education (n = 277); 14 respondents did not indicate type of education. The technical and vocational courses prepared students for professions in health and bodily care, fashion and office management, which may explain the gender imbalance in the sample. Relative to the student population in Flanders (26% general, 21% technical and 18% vocational education enrolment [41]), students from technical education were oversampled.
The sample was comprised of 736 non-smokers (60.7%), 199 occasional smokers (16.4%) and 278 daily smokers (22.9%). Twenty respondents did not indicate current smoking behaviour. Relative to a representative sample of Flemish 13–18 year olds (with 69.4% non-smokers, 6.3% occasional smokers and 18.0% daily smokers [33]), we counted more occasional and daily smokers (P < 0.001). In the non-smokers group, 281 (38.2%) respondents indicated they had tried a cigarette at least once. Daily smokers had on average 9.4 cigarettes a day; occasional smokers smoked on average 4.7 cigarettes a week.
Measures
Smoking status
Groups were identified by responses on two items measuring current smoking and one item measuring past smoking experience. Respondents could indicate if—at the time of the study—they considered themselves to be non-smokers (I do not smoke), occasional smokers (I smoke occasionally/I smoke, but not every day) or daily smokers (I smoke every day). The item referred to cigarettes only. Occasional smokers indicated how many cigarettes they smoked per month or per week; daily smokers wrote down how many cigarettes they smoked per day. Past experience was binomial: respondents indicated if they had ever smoked a cigarette (or part of one). For additional analyses, the group of non-smokers was split into never smokers (who had never touched a cigarette, n = 455) and experimenters (who had at least smoked once, n = 281).
Social-cognitive measures
The social-cognitive variables were measured with a questionnaire that was conceptually based on the Theory of Planned Behaviour [42]. Attitudes, self-efficacy (ease/difficulty of performing an action) and prescriptive norms (what parents and peers would like the respondent to do) were included as antecedents of intentions.
Measures of the environment
A second part of the questionnaire was devoted to adolescents' perception of the environment. In line with the attitude–social influence–self-efficacy model, an extension of the Theory of Planned Behaviour, social influences were considered to extend beyond prescriptive norms [43]: smoking by parents and peers was also measured. Also, perceived school smoking policy and smoking prevalence on school premises, a factor associated with increased risk of smoking [44], were registered. For current smokers, environmental cues eliciting a craving to smoke were estimated.
Scale construction and data handling
Items measuring antecedents of smoking intentions that were filled in by all respondents were analysed via exploratory principal component analysis. Twelve factors (eigenvalue > 1.00) emerged. All factors were interpretable and consistent with the theoretical framework. Items with varimax rotated loadings >0.70 were aggregated in scales; items with loadings >0.50 were included if they increased internal consistency. Items with cross-loadings >0.30 were removed. Two scales were discarded because of insufficient internal consistency. Items filled in by smokers only (measuring context-specific urges to smoke) were examined separately and were found to represent a single factor. In total, 10 multi-item scales were composed (Table A-I): 3 referring to attitudes (health consequences, social benefits and social penalties of smoking), 1 referring to social influences favouring smoking, 2 referring to self-efficacy (ease of refusing a smoke and ease of buying/smoking cigarettes) and 4 referring to the environment (parental smoking, smoking prevalence in ‘risky’ and ‘safe’ school locations and contexts in which smokers experience an urge to smoke). School smoking policy was included as a one-item measure.
Data analysis
Mean responses for the three smoking status groups were compared via univariate analyses of variance and the Tukey post hoc test for unequal sample sizes. Correlations between the social-cognitive/environment measures and intentions were calculated for the three groups separately. Because this test involved the estimation of 32 parameters, the probability level was set at a conservative P = 0.0016 (=0.05/32). Between-group differences in correlations and regression analyses per group indicated whether separate-slopes analysis would be appropriate.
The separate-slopes regression analyses were performed to test group differences in association between social-cognitive/perceived environment variables and intentions. In a first test, the intention to smoke was predicted for all three groups. The independent variables were interaction terms between the categorical variable (smoking status group) and the 10 continuous variables measuring social-cognitive/environmental influences. Age and education were included as main effects, because they correlated with intentions in univariate tests. The continuous predictors were standardized by subtracting scores from the mean and dividing by the standard deviation. In this model, the minimal P value per interaction term was set at 0.005 to account for the number of tests. If an interaction was significant, beta values were compared in order to test in which group the association between the predictor and intentions was strongest. The minimal P value for a beta parameter to be declared different from zero was set at 0.016 (adjusting for three tests per parameter). A beta value in one group was declared significantly different from its counterpart in another group if the 98.4% confidence intervals (CIs) for the groups did not overlap. Collinearity statistics were acceptable, the lowest tolerance level (0.17) was observed for self-efficacy in buying/smoking cigarettes in daily smokers. In a second model, we added one predictor, the interaction between smoking status and contexts eliciting craving. This variable was measured only in occasional and daily smokers, so only these two groups were compared. The minimal P value per interaction term was now set at 0.0045. The P value for beta values was set at 0.025 and 97.5% CIs were calculated. For this analysis, all collinearity statistics were well within the acceptable range. Again, age and education were included as main effect predictors.
In a hierarchical regression, it was tested whether the inclusion of interaction terms (smoking status by continuous predictors) increased the proportion of explained variability in intentions over and above the proportion explained by group membership and demographic variables alone.
Finally, it should be noted that all analyses were also run on four groups (never smokers, experimenters, occasional and daily smokers).
To perform correlation and regression analyses with sufficient power, missing values were substituted by mean scores (calculated over the entire sample) provided the total percentage of missing data did not exceed 10 (over the sample) and the respondent had filled in at least 90% of the questions. No substitutions were made for intentions.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConflict of interest statementAppendixReferences |
|---|
Mean values
Univariate analyses of variance revealed differences between non-smokers, occasional smokers and daily smokers on social-cognitive measures and measures of the environment (Table I). Post hoc tests showed differences in the expected direction. Non-smokers were more convinced of the health adverse consequences of smoking than smokers, were more likely to indicate that smoking leaves your clothes and hair smelling ‘dirty’ and were less convinced of the social benefits. Social influences favouring smoking and intentions to smoke in future were lowest in non-smokers, with gradual increases in occasional and daily smokers. Daily smokers found buying and smoking cigarettes considerably easier than non-smokers did. Occasional smokers differed from daily smokers in the number of environments that elicit a craving and in social influences favouring smoking.
|
Group-specific associations between social-cognitive/environment measures and intentions
Bivariate correlations
The relationship between determinants and intentions varied across groups (Table II). Difference tests showed that the correlation between intentions and social penalties of smoking was higher for occasional than for daily smokers (P = 0.015). Self-efficacy in refraining from smoking was more closely linked to intentions for non-smokers than for occasional (P = 0.031) or daily (P = 0.022) smokers. Finally, the correlation between intentions and self-efficacy in buying/smoking cigarettes was lower for non-smokers than for occasional smokers (P = 0.038).
|
Within-group regressions
In regression analyses performed for each group separately, a significant proportion of variance in intentions was explained in all groups (adjusted R2 = 0.17, F(12, 682) = 12.90, P < 0.001 for non-smokers; adjusted R2 = 0.19, F(12, 176) = 4.82, P < 0.001 for occasional smokers and adjusted R2 = 0.09, F(12, 245) = 3.12, P < 0.001 for daily smokers). In each group, different predictors were significant (betas are not reported here). So, both the within-group correlations and regression analyses indicated that a separate-slopes model (that takes the interaction between group membership and continuous predictors into account) would be appropriate.
Multivariate estimates for the whole sample
The separate-slopes model for the whole sample (Table III) explained 72% of the variance in smoking intentions (adjusted R2 = 0.72, F(35, 1106) = 83.93, P < 0.001). Univariate tests showed four significant interactions (upper row lines). The predictive value of health consequences, social influences favouring smoking, self-efficacy in refraining from smoking and self-efficacy in buying/smoking cigarettes differed between groups.
|
Inspection of the beta coefficients showed that slopes that were significant in one group were not always so in another group. Non-smokers' intentions were linked to self-efficacy in refraining from smoking: those who found it difficult to refuse a smoke were more likely to indicate that they might smoke in the future than non-smokers who were convinced that they could easily refuse cigarettes. Social influences favouring smoking were also linked to higher smoking intentions in the non-smoking group. Occasional smokers' intentions were linked to health consequences. Respondents who thought that smoking would affect their health were less inclined to indicate that they would continue smoking. A second significant predictor in this group was ease of buying/smoking cigarettes. Both variables were also associated with intentions in the daily smokers group. Finally, social influences favouring smoking were also associated with daily smokers' intentions.
Inspection of the CIs for the slopes revealed that the contribution of ‘ease of buying/smoking cigarettes’ to the prediction of intentions was stronger in daily smokers than in occasional smokers. CIs for other predictors ‘shared’ by two groups had an overlap, indicating that no reliable differences in strength were found.
Multivariate estimates for occasional and daily smokers
The extended model for occasional and daily smokers explained 45% of variance in intentions (adjusted R2 = 0.45, F(26, 420) = 15.02, P < 0.001). One interaction term was significant (Table III, lower row lines). Daily smokers who encountered many contexts that elicited a craving had higher intentions to continue smoking. The CIs however show that the slope of this predictor may be of a similar elevation in occasional smokers.
Variance explained by group-specific slopes
It was tested whether the separate-slopes models performed better than a model with only group membership, age and educational type as predictors of intentions. In hierarchical regressions, these three variables explained 67% of the variance in intentions for the whole sample and 34% of the variance in current smokers. The increment in explained variance gained from adding group x continuous predictor interactions was significant in both groups (Fincrement = 16.48 for the whole sample and 9.73 for smokers, both P < 0.001), indicating that the continuous predictors (weighed by group membership) explained a sizeable proportion of the variance in smoking intentions.
Never smokers versus experimenters
All analyses were run again on four groups (never smokers, experimenters, occasional smokers and daily smokers). Never smokers and experimenters who were non-smokers at the time of the study had different mean values on all social-cognitive variables. All differences pointed to experimenters having less ‘healthy’ cognitions and intentions regarding future smoking. However, never smokers and experimenters ‘behaved’ identically in all correlational tests. A separate-slopes analysis performed on these four groups showed that associations between the continuous variables and smoking intentions were identical for never smokers and experimenters. Therefore, these results were not detailed.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConflict of interest statementAppendixReferences |
|---|
Non-smokers, occasional smokers and daily smokers gave different responses on measures of variables that are considered antecedents of adolescent smoking. Not all variables showed a ‘dose-response’ relationship, with non-smokers scoring lowest (or highest) and daily smokers scoring highest (lowest). This is a first indication that the three groups may benefit from interventions that focus on different determinants of smoking.
A second indication that these groups are fundamentally different is the fact that the relationship between smoking determinants and smoking intentions differed across groups (Table IV). Hedeker et al. [25] demonstrated this variability for individual respondents, whereas we demonstrated it at group level.
|
Our second hypothesis, concerning the relative importance of specific determinants, was only partially confirmed. In the group of non-smokers, ease of not smoking and friends' influences and intentions were linked to intentions, as expected. Attitudes, on the other hand, were not associated with non-smokers' intentions. In contrast, there was a significant association between attitudes and smoking intentions in the group of occasional smokers, where we had expected attitudes to play a lesser role than in non-smokers. For occasional smokers, we further observed the expected link between ease of smoking/buying cigarettes and intentions, but not the link between parental smoking models and intentions. Finally, in the group of daily smokers, we found three significant associations between intentions and social-cognitive predictors, which were contrary to our expectations. Daily smokers who perceived more health adverse consequences were less likely to indicate that they intended to keep on smoking. The strong link between self-efficacy in smoking/buying cigarettes and intentions in the daily smokers group can possibly be explained by the fact that daily smokers face the largest risk of getting caught while buying and smoking cigarettes, in order to satisfy their smoking needs. The social influence factor may have been important for daily smokers, because friends do more than approve of or model smoking behaviour. Often, they also provide a ‘context’ for smoking, which in the long term make them likely elicitors of nicotine cravings. A further analysis of the difference between occasional and daily smokers confirmed that specific contexts play a considerable part in the prediction of daily smokers' intentions. Most likely, a history of smoking in these contexts transformed them into discriminative stimuli for smoking, leading to context-specific smoking urges [40].
These findings imply that prevention programmes for adolescents should take profiles of relevant antecedents to intentions into account (Table IV). Refusal skill training (e.g. building skills needed to recognize and resist social pressure) may be effective for non-smokers, but may not significantly change occasional smokers' plans. All smokers may benefit from reinforcing non-smoking policies on school premises [45] and restricting access to cigarettes [46]. Providing information about health risks does not deter adolescents from taking up smoking or reduce smokers' consumption [46], but a more balanced appraisal of pros versus cons of smoking has been linked to stronger motivations to quit [47]. Interventions aimed at daily smokers should be less focused on the ‘why’ of smoking and incorporate features addressing the ‘where’ and ‘when’ of cigarette consumption. Behavioural techniques, such as exposure to smoking cues, while preventing smokers to light up, can decrease cue-reactivity [48]. Nicotine replacement therapy can help sever the links between contexts and nicotine cravings [49], especially if the medication offers rapid release from withdrawal symptoms [50]. Another option is to enforce stricter no-smoking policies. Even regular smokers experience fewer urges to smoke in contexts where smoking is impossible [51]. Social influences can be countered by creating social environments supportive of non-smoking and by behavioural techniques (insofar as smoking peers are cues for smoking).
Many interventions and policy approaches are effective in youth tobacco control [46]. Our data suggest that it is not advisable to lump them together in ‘one size fits all’ programmes offered to class groups with heterogeneous smoking careers. ‘Targeting’, i.e. designing interventions that reflect group-specific characteristics [3], seems an effective way of combining advantages of collective strategies with the strength of interventions matched to respondents' personal profile.
It must be noted that the predictive weight of the social-cognitive and environment variables was rather low. This can happen when predictors contribute redundant information. However, inspection of the tolerance indices and within-group correlations (not reported) ruled out multicollinearity as an explanation. Alternatively, the categories may not have been homogeneous. Non-smokers in our study may never have tried a cigarette, tried one or a few or may have been frequent smokers in the past, which would lead to within-group differences [14, 52, 53]. Our distinction between never smokers and experimenters was perhaps too coarse to pick up these differences. Occasional smokers who used to smoke daily are different from those who never took up daily smoking [54], smokers who smoke less than weekly differ from those who smoke at least once a week [19] and tobacco chippers, who smoke <5 cigarettes per day, differ from nicotine-dependent smokers [55]. Finally, within groups, there may be respondents who do not want to change their behaviour and respondents who consider a change in the near or distant future. The fact that health consequences of smoking predicted intentions in daily smokers may be indicative of this: smokers contemplating change see more adverse consequences of their behaviour and are more responsive to health warnings than smokers who do not contemplate a change [47]. It would have been difficult to include intentions or future plans in the classification algorithm, as our purpose was to study associations with intentions. We must conclude, however, that the group distinction in this study may be too simple to detect all differences in determinant weights. The phase models by Flay et al. [19] and Kremers et al. [56] seem promising in that respect, but the current study lacked the necessary measurements and power to test these classifications.
The smoking intentions measure differed from those commonly used in studies of Theory of Planned Behaviour variables. Usually, respondents indicate their position on a bipolar continuum ranging from ‘I definitely intend to smoke’ to ‘I definitely intend not to smoke’, combining direction and strength of future plans. Our intention measure also included a direction component (I intend to smoke in future versus I do not intend to smoke), but no strength component. Instead, we asked respondents with what frequency they intended to smoke (the intention not to smoke amounts to a zero frequency). Intention measures that refer to substance use with different frequencies or quantities have also been used by Conner and McMillan [57, 58], because it is important to know what kind of ‘smoking career’ adolescents have in mind. Respondents who plan to smoke every day can be said to have different intentions from respondents who plan to smoke once a week, even if all of these respondents have ‘strong’ intentions to smoke in future. When a separate-slopes analyses were performed on a binary transformation of intentions (respondents who indicated they did not want to smoke in future, we categorized as ‘intends not to smoke’ and all other respondents as ‘intends to smoke’; results were not detailed), the group-specific covariates of intentions were similar to the ones found in the original analysis. It remains to be tested whether a traditional intention–strength multi-item scale would show a similar pattern of associations.
The data presented are cross-sectional. We consider the exploration of cross-sectional patterns important, both for model development and planning of prevention initiatives. It should however be noted that we cannot say which determinants will prove important in the movement from one category to another [59].
We do not suggest that smoking status is an explanatory variable of the same nature as the social-cognitive variables. The latter can explain the ‘why, how, when and where’ of adolescents' intention to smoke or abstain from smoking. Current smoking status explains variations in intention, but only in the strictest statistical sense.
Finally, the sample was not representative of the Flemish student population. The proportion of non-users in our sample was smaller than the estimated proportion for Flanders [33]. Female students and students from technical courses were overrepresented. For further studies, stratification with respect to sex, educational type and age groups would be advisable.
|
Conflict of interest statement |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConflict of interest statementAppendixReferences |
|---|
None declared.
|
Appendix |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConflict of interest statementAppendixReferences |
|---|
|
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConflict of interest statementAppendixReferences |
|---|
1. Godeau E, Rahav G, Hublet A. Tobacco smoking. In Currie C, Roberts C, Morgan A (Eds.), et al. Young People's Health in Context. Health Behaviour in School-Aged Children (HBSC) Study: International Report From the 2001 /2002 Survey.Copenhagen, Denmark: World Health Organization, Regional Office for Europe 2004 pp. 63–72.
2. Ausems M, Mesters I, van Breukelen G, et al. Short-term effects of a randomized computer-based out-of-school smoking prevention trial aimed at elementary school children. Prev Med 2002 34:581–9.[CrossRef][ISI][Medline]
3. Strecher VJ. Computer-tailored smoking cessation materials: a review and discussion. Patient Educ Couns 1999 36:107–17.[CrossRef][ISI][Medline]
4. Dijkstra A and De Vries H. The development of computer-generated tailored interventions. Patient Educ Couns 1999 36:193–203.[CrossRef][ISI][Medline]
5. Kreuter MW and Skinner CS. Tailoring: what's in a name? Health Educ Res 2000 15:1–4.
6. Hines D, Fretz AC, Nollen NL. Regular and occasional smoking by college students: personality attributions by smokers and non-smokers. Psychol Rep 1998 3:1299–306.
7. Patton D, Barnes GE, Murray RP. A personality typology of smokers. Addict Behav 1997 22:269–73.[CrossRef][ISI][Medline]
8. Poland BD, Cohen JE, Ashley MJ, et al. Heterogeneity among smokers and non-smokers in attitudes and behaviour regarding smoking and smoking restrictions. Tob Control 2000 9:364–71.
9. Petraitis J, Flay BR, Miller TQ. Reviewing theories of adolescent substance use: organizing pieces in the puzzle. Psychol Bull 1995 117:67–86.[CrossRef][ISI][Medline]
10. Ajzen I. The theory of planned behavior. Organ Behav Hum Decis Process 1991 50:179–211.[CrossRef][ISI]
11. Armitage CJ and Arden MA. Exploring discontinuity patterns in the transtheoretical model: an application of the theory of planned behaviour. Br J Health Psychol 2002 7:89–103.[CrossRef][ISI][Medline]
12. Sutton S. Interpreting cross-sectional data on stages of change. Psychol Health 2000 15:163–71.
13. Hill JA, Boudreau F, Amyot E, et al. Predicting the stages of smoking acquisition according to the theory of planned behavior. J Adolesc Health 1997 21:107–15.[CrossRef][ISI][Medline]
14. Pederson LL, Koval JJ, McGrady GA, et al. The degree and type of relationship between psychosocial variables and smoking status for students in grade 8: is there a dose-response relationship? Prev Med 1998 27:337–47.[CrossRef][ISI][Medline]
15. Johnson CC, Myers L, Webber LS, et al. Profiles of the adolescent smoker: models of tobacco use among 9th grade high school students Acadiana Coalition of Teens against Tobacco (ACTT). Prev Med 2004 39:551–8.[CrossRef][ISI][Medline]
16. Ajzen I. Nature and operation of attitudes. Annu Rev Psychol 2001 52:27–58.[CrossRef][ISI][Medline]
17. Thomas R. School-based programmes for preventing smoking (Cochrane Review) DOI@10.1002/14651858.CD001293. The Cochrane Library.Chichester, UK: John Wiley & Sons 2003.
18. van Sluijs EMF, van Poppel MNM, van Mechelen W. Stage-based lifestyle interventions in primary care: are they effective? Am J Prev Med 2004 26:330–43.[ISI][Medline]
19. Flay BR, Hu FB, Richardson J. Psychosocial predictors of different stages of cigarette smoking among high school students. Prev Med 1998 27:A9–18.[CrossRef][ISI][Medline]
20. Kremers SPJ, de Vries H, Mudde AN, et al. Motivational stages of adolescent smoking initiation: predictive validity and predictors of transitions. Addict Behav 2004 29:781–9.[CrossRef][ISI][Medline]
21. Kremers SPJ, Mudde AN, De Vries H. Model of unplanned smoking initiation of children and adolescents: an integrated stage model of smoking behavior. Prev Med 2004 38:642–50.[CrossRef][ISI][Medline]
22. Dijkstra A, Tromp D, Conijn B. Stage-specific psychological determinants of stage transition. Br J Health Psychol 2003 8:423–37.[CrossRef][ISI][Medline]
23. Mudde AN and De Vries H. Het ASE-model en het voorspellen van motivatieverandering bij stoppen met roken. Gedrag Gezond 1998 26:121–34.
24. Wetter DW, Kenford SL, Welsch SK, et al. Prevalence and predictors of transitions in smoking behavior among college students. Health Psychol 2004 23:168–77.[CrossRef][ISI][Medline]
25. Hedeker D, Flay BR, Petraitis J. Estimating individual influences of behavioral intentions: an application of random-effects modeling to the theory of reasoned action. J Consult Clin Psychol 1996 64:109–20.[CrossRef][ISI][Medline]
26. Tyc VL, Hadley W, Allen D, et al. Predictors of smoking intentions and smoking status among nonsmoking and smoking adolescents. Addict Behav 2004 29:1143–7.[CrossRef][ISI][Medline]
27. Dijkstra A and De Vries H. Subtypes of precontemplating smokers defined by different long-term plans to change their smoking behavior. Health Educ Res 2000 15:423–34.
28. Filice GA, Hannan PJ, Lando HA, et al. A period of increased susceptibility to cigarette smoking among high school students. J Sch Health 2003 73:272–8.[ISI][Medline]
29. Robinson M, Berlin I, Moolchan ET. Tobacco smoking trajectory and associated ethnic differences among adolescent smokers seeking cessation treatment. J Adolesc Health 2004 35:217–24.[ISI][Medline]
30. Holmen TL, Barrett-Connor E, Holmen J, et al. Adolescent occasional smokers, a target group for smoking cessation? The Nord-Trøndelag Health Study, Norway, 1995 –1997. Prev Med 2000 31:682–90.[CrossRef][ISI][Medline]
31. Norman GJ, Velicer WF, Fava JL, et al. Cluster subtypes within stage of change in a representative sample of smokers. Addict Behav 2000 25:183–204.[CrossRef][ISI][Medline]
32. Mayhew KP, Flay BR, Mott JA. Stages in the development of adolescent smoking. Drug Alcohol Depend 2000 59:S61–81.[CrossRef][ISI][Medline]
33. Vereecken C and Maes L. Voorlopig Rapport Jongeren En Gezondheid 1990–2002.Available at: http://users.ugent.be/cvereeck/hbsc/: Accessed: 16 January 2006.
34. Lambert M, Hublet A, Verduyckt P, et al. Gender Differences in Smoking in Young People.Brussels, Belgium: Flemish Institute for Health Promotion 2002.
35. Bancroft A, Wiltshire S, Parry O, et al. "It's like an addiction first thing... afterwards it's like a habit": daily smoking behaviour among people living in areas of deprivation. Soc Sci Med 2003 56:1261–7.[CrossRef][ISI][Medline]
36. Mott JA, Crowe PA, Richardson J, et al. After-school supervision and adolescent cigarette smoking: contributions of the setting and intensity of after-school self-care. J Behav Med 1999 22:35–58.[Medline]
37. Pinilla J, Gonzales B, Santana Y. Smoking in young adolescents: an apporach with multilevel discrete choice models. J Epidemiol Community Health 2002 56:227–32.
38. Sussman S, Stacy AW, Ames SL, et al. Self-reported high-risk locations of adolescent drug use. Addict Behav 1998 23:405–11.[CrossRef][ISI][Medline]
39. Solomon RL. The opponent-process theory of acquired motivation: the costs of pleasure and the benefits of pain. Am Psychol 1980 35:691–712.[CrossRef][Medline]
40. Ramsay DS and Woods SC. Biological consequences of drug administration: implications for acute and chronic tolerance. Psychol Rev 1997 104:170–93.[CrossRef][ISI][Medline]
41. Department of Education. Schoolbevolking gewoon secundair onderwijs. [School Population in Regular Secondary Education.].Available at: http://www.ond.vlaanderen.be/onderwijsstatistieken/: Accessed: 29 July 2005.
42. Ajzen I and Madden TJ. Prediction of goal-directed behavior: attitudes, intentions, and perceived behavioral control. J Exp Soc Psychol 1986 22:453–74.[CrossRef][ISI]
43. de Vries H, Backbier E, Kok G, et al. The impact of social influences in the context of attitude, self-efficacy, intention, and previous behavior as predictors of smoking onset. J Appl Psychol 1995 25:237–57.[Medline]
44. Ellickson PL, Bird CE, Orlando M, et al. Social context and adolescent health behavior: does school-level smoking prevalence affect students' subsequent smoking behavior? J Health Soc Behav 2003 44:525–35.[CrossRef][ISI][Medline]
45. Moore L, Roberts C, Tudor-Smith C. School smoking policies and smoking prevalence among adolescents: multilevel analysis of cross-sectional data from Wales. Tob Control 2001 10:117–23.
46. Lantz PM, Jacobsons PD, Warner KE, et al. Investing in youth tobacco control: a review of smoking prevention and control strategies. Tob Control 2000 9:47–63.
47. Velicer WF, Norman GJ, Fava JL, et al. Testing 40 predictions from the transtheoretical model. Addict Behav 1999 24:455–69.[CrossRef][ISI][Medline]
48. Rohsenow DJ, Niaura RS, Childress AR, et al. Cue reactivity in addictive behaviors: theoretical and treatment implications. Int J Addict 1991 25:957–93.
49. Cummings KM and Hyland A. Impact of nicotine replacement therapy on smoking behavior. Annu Rev Public Health 2005 26:583–99.[CrossRef][ISI][Medline]
50. Niaura R, Sayette M, Shiffman S, et al. Comparative efficacy of rapid-release nicotine gum versus nicotine polacrilex gum in relieving smoking cue-provoked craving. Addiction 2005 100:1720–30.[CrossRef][ISI][Medline]
51. Dols M, van den Hout M, Kindt M, et al. The urge to smoke depends on the expection of smoking. Addiction 2002 97:87–93.[CrossRef][ISI][Medline]
52. Pomerleau OF, Pomerleau CS, Namenek RJ. Early experiences with tobacco among women smokers, ex-smokers, and never-smokers. Addiction 1998 93:595–9.[CrossRef][ISI][Medline]
53. Pomerleau CS, Pomerleau OF, Snedecor SM, et al. Defining a never-smoker: results from a nonsmokers survey. Addict Behav 2004 29:1149–54.[CrossRef][ISI][Medline]
54. Gilpin E, Cavin SW, Pierce JP. Adult smokers who do not smoke daily. Addiction 1997 92:473–80.[CrossRef][ISI][Medline]
55. Shiffman S and Sayette MA. Validation of the nicotine dependence syndrome scale (NDSS): a criterion-group design contrasting chippers and regular smokers. Drug Alcohol Depend 2005 79:45–52.[CrossRef][ISI][Medline]
56. Kremers SPJ, Mudde AN, de Vries H. "Kicking the initiation": do adolescent ex-smokers differ from other groups within the initiation continuum? Prev Med 2001 33:392–401.[CrossRef][ISI][Medline]
57. McMillan B and Conner M. Using the theory of planned behaviour to understand alcohol and tobacco use in students. Psychol Health Med 2003 8:318–28.
58. Conner M and McMillan B. Interaction effects in the theory of planned behavior: studying cannabis use. Br J Soc Psychol 1999 38:195–222.[Medline]
59. Engels RCM, Knibbe RA, Drop MJ. Predictability of smoking in adolescence: between optimism and pessimism. Addiction 1999 94:115–24.[CrossRef][ISI][Medline]
Received on March 23, 2005; accepted on June 22, 2006
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||