the impact of social influence on adolescent intention to ... · fishbein, 1980; fishbein &...

British Journal of Health Psychology (2009). 14. 681-699 MigaM <i ' tv© 2009 The British Psycho/of/ca/Society * * * i ^ bociety

681

TheBritishPsychological

www.bpsjournals.co.uk

The impact of social influence on adolescentintention to smoke: Combining types andreferents of influence

Paulo D. Vitoria'*, M. Fátima Salgueiro^ Silvia A. Silva'and H. De Vries^' Faculdade de Ciencias da Saude da Universidade da Beira Interior, Covilhâ, Portugal^Instituto Superior de Ciencias do Trabalho e da Empresa, UNIDE, Lisboa, Portugal^Instituto Superior de Ciencias do Trabalho e da Empresa, CIS, Lisboa, Portugal'^Department of Health Promotion, Maastricht University, Maastricht,The Netherlands

Objectives. Theory and research suggest that the intention to smoke is the maindeterminant of smoking initiation and emphasizes the role of cognitive and socialfactors on the prediction of the intention to smoke. However, extended models such asthe I-Change and results from published studies reveal inconsistencies regarding theimpact of social influence on the intention to smoke. Possible explanations for this maybe the definition and measurement of the constructs that have been used.

Design and methods. The current study was designed with two main goals: (i) totest a measurement model for social influence, combining different types of socialinfluence (subjective norms, perceived behaviour, and direct pressure) with variousreferents of influence (parents, siblings, peers, and teachers); (ii) to investigate theimpact of social influence on adolescent intention to smoke, controlling for smokingbehaviour. LISREL was used to test these models. The sample includes 3,064Portuguese adolescents, with a mean age of 13.5 years, at the beginning of the seventhschool grade.

Results. The hypothesized measurement model of social influence was supported byresults and explained 29% of the variance of the intention to smoke. A more extendedmodel, including attitude and self-efficacy, explained 55% of the variance of the intentionto smoke. Perceived behaviour of peers, parental norms, and perceived behaviour ofparents were the social influence factors with impact on adolescent intention to smoke.

Conclusions. Results suggest that different referents exert their influence throughdistinct types of social influence and recommend further work on the definition andmeasurement of social influence.

* Correspondence should be addressed to Paulo D. Vitoria, Faculdade de Ciencias da Saude, Universidade da Beira Interior,Covilhä, Portugal (e-mail: [email protected]).

DOI: 10.1348/135910709X421341

682 Paulo D. Vitoria et ai

Smoking is the single most important preventable cause of morbidity and mortality indeveloped countries (World Health Organisation, 1997, 1999). Worldwide, deathsattributable to tobacco are near 5 million (Lopez, Mathers, Ezzati, Jamison, & Murray,2006). In Portugal, between 1950 and 2000, about 2,00,000 deaths are attributed tosmoking, and nowadays more than 8,000 deaths per year are caused by tobacco, half ofthem in middle age (Peto, Watt, & Boreham, 2006). During the last 10 years, Portugueseoverall smoking prevalence has been decreasing (from 29 to 24%, EuropeanCommission, 2003, 2007), as has the prevalence of tobacco 'ever smokers' and 'dailysmokers' among Portuguese adolescents (8.5-5.0%, Matos etal., 2006). In spite of thesepositive trends, smoking figures in Portugal are still high and the consequences ofsmoking remain severe, underlining the need for more research to improve knowledgeand control of this very harmful behaviour

An important objective for controlling tobacco use and damage is to promoteeffective smoking prevention programmes. This implies to reach youth early in theiradolescence since more than 90% of smokers had their first puff during adolescenceand the peak of initiation is between 11 and 15 years (e. g. Azevedo, Machado, & Barros,1999; Santos & Barros, 2004; US Department of Health and Human Services, 1994).These findings are supported by the results of recent studies claiming thatadolescents felt the first symptoms of dependence after just a few cigarettes andbefore smoking regularly (Colby, Tiffany, Shiffman, & Niaura, 2000; DiFranza et al.,2000, 2002). Consequently, prevention activities should reach adolescents beforesmoking initiation.

Several studies suggest that youngsters start to take up smoking because of directpressure exerted by others (e.g. Evans etaL, 1976, 1978; US Department of Health andHuman Services, 1994). Although, other sources of infiuence are recognized (e.g.parents and mass media), Evans' work resulted in a major emphasis on the direct role ofpeer pressure as the main determinant of smoking onset, combined with theadolescents' inability to refuse cigarettes (Evans et al., 1978, 1988). At the same time,Bandura (1977, 1986) proposed that social influence could also have an indirect effecton youth when exposed to the behaviour of others (mere exposure) and/or byobserving the positive outcomes of behaviour acquired by others (vicarious learning).Furthermore, several studies outlines the importance of parental behaviour (e.g. Bricker,Peterson, Leroux et al, 2006; Bricker, Peterson, Sarason, Andersen, & Rajan, 2007;Engels, Knibbe, De Vries, Drop, & Van Breukelen, 1999; O'Loughlin, Paradis, Renaud, &Gomez, 1998), and sibling behaviour as determinants for adolescent smoking initiation(Brickereiö/., 2006; Rajan etal, 2003; Slomkowski, Rende, Novak, Lloyd-Richardson, &Niaura, 2005), and some research suggests that teachers could also play a role in thisprocess (Barnett et al, 2007; Bente, Torsheim, Currie, & Roberts, 2004; Bewley,Johnson, & Banks, 1979; Wold, Torsheim, Currie, & Roberts, 2004).

The theory of planned behaviour (TPB; Ajzen, 1988, 1991; Ajzen & Madden, 1986),an extension of the theory of reasoned action (TRA; Fishbein & Ajzen, 1975), is themost applied model in studies on health-related behaviours (e.g. Armitage & Conner,2001; Rivis & Sheeran, 2003). This theory explicitly acknowledges social infiuence, viasubjective norm, as a proximal predictor of intention, and behaviour, as well as attitudetowards the behaviour and perceived behavioural control (Ajzen, 1988, 1991; Ajzen &Fishbein, 1980; Fishbein & Ajzen, 1975). However, the theoretically predicted role ofsocial infiuence as a main determinant of intention, and behaviour is not clearlyconfirmed by published research on health-related behaviours (e.g. Armitage &Conner, 2001). This inconsistency between theory propositions and research results

Impact of social influence on intention to smoke 683

suggests the need of an extension of the social influence component of the TPB. Directpressure, the concept proposed by Evans to explain smoking onset (Evans, 1976) andmodelling, a central concept in the social learning theory (Bandura, 1977, 1986), werenot included in TPB. Namely, the unique contribution of direct pressure, andmodelling (perceived behaviour of important others or descriptive norms) wasdemonstrated in an elaboration of the TRA, the ASE-Model (De Vries, Backbier, Kok, &Dijkstra, 1995), and its successor the Integrated Model of Behaviour Change or theI-Change model (De Vries, Mudde et al., 2003; Holm, Kremers, & De Vries, 2003;Vitoria, Kremers, Mudde, Pais Clemente, & De Vries, 2006). More examples found inliterature of efforts to expand the social influence component of TPB were moralnorms (Godin, Conner, & Sheeran, 2005), social support (Rhodes, Jones, & Courneya,2002), self-identity (Hagger & Chatzisarantis, 2006), and prototype perceptions(Gibbons, Gerrard, & Lane, 2003). Among these approaches two main streams emerge:one more focused on normative processes (e.g. Subjective norms and descriptivenorms) and other more focused on direct interpersonal processes (e.g. Direct pressureand social support).

All the above-referred theories agree on one point: a significant part of humanbehaviour can be predicted from the intention to perform it and, furthermore, intentionmediates influence on behaviour from other proximal variables (Ajzen & Fishbein,1980; Fishbein & Ajzen, 1975). Results of many studies have supported this claim (e.g.Armitage & Conner, 2001). A recent meta-analysis concluded that a medium-to-largechange in intention leads to a small-to-medium change in behaviour (Webb & Sheeran,2006). This result indicates that the impact of intention on behaviour is not as big asprevious meta-analyses based on correlational studies have concluded but, on the otherhand, confirms that intention continues to be the main predictor of behaviour.Intention to smoke was also empirically confirmed in many studies as the mainpredictor of adolescent smoking behaviour (e.g. Andrews, Tildesley, Hops, Duncan, &Severson, 2003; Choi, Gilpin, Farkas, & Pierce, 2001; Vitoria et al, 2006; Wakefieldet al., 2004). Considering these results and the need to address youth with smokingprevention activities during early adolescence, preferentially before the firstexperiences with tobacco, the intention to smoke should play a central role in anywork aimed to prevent smoking initiation and to improve the results of smokingprevention programmes.

As referred before, the expected role of social influence as a main determinant ofsmoking onset has not been confirmed by published research. Moreover, the role ofimportant referents has been also inconsistent. Peer influence was considered thebest predictor of adolescent smoking initiation in many studies (e.g. Conrad, Flay, &Hill, 1992; Kobus, 2003; Tyas & Pederson, 1998), but the results of recent researchsuggest that family members, mainly parents but also siblings, may have an importantinfluence on adolescent smoking behaviour, particularly in early adolescence(e.g. Bricker, Peterson, Andersen et al, 2006, Bricker, Peterson, Leroux et al, 2006;De Vries, Engels, Kremers, Wetzels, & Mudde, 2003; McMillan, Higgins, & Conner,2005; Nebot, Tomás, Ariza, Valmayor, & Mudde, 2002; Nebot et al, 2004; Rajan et al,2003). These conflicting research results may be caused by inconsistencies in theconceptualization of social influence, the operationalization, and measurement of socialitifluence concepts, the statistical approaches adopted in the investigation anddifferences in the age of the study sample (e.g. Avenevoli & Merikangas, 2003; De Vries,Candel, Engels, & Mercken, 2006; De Vries, Engels etal, 2003; Hagger & Chatzisarantis,2005; Kobus, 2003; McMillan etal, 2005).


The purpose of the current study was to further analyse the role and impact of socialinfluence on adolescent intention to smoke testing a more comprehensiveoperationalization of social influence. Previous studies have investigated either therelative contribution of different types of social influence (e.g. De Vries et al., 1995;McMillan et al., 2005) or the relative contribution of different referents (e.g. Bricker,Peterson, Leroux et al, 2006, Bricker et al, 2007; Kobus, 2003). However, one cansimultaneously conceive the existence of differences in the impact on youngsters'intention to smoke from norms of parents or norms of peers or perceived behaviour ofbest friend or direct pressure of siblings. Thus, the main contribution of the currentstudy is to acknowledge both the existence of different types and referents of socialinfluence and to combine them in a new measurement model. The proposed socialinfluence measurement model, with three scales (three types of social influence), eachwith four factors (the four groups of important others as sources of influence) wastested and used to investigate the impact of social influence on adolescent intention tosmoke, firstly on its own, and secondly together with the other proximal variables of theI-Change model: attitudes and self-efficacy (see Figure 1, forgetting for now the figures).

We have chosen intention to smoke as the dependent variable for several reasonsalready exposed in this introduction: (i) because research indicates that preventionactivities must start preferentially before the first experience with tobacco (Colby et al.,2000; DiFranza et al, 2000, 2002); (ii) because of the low age of participants in thestudy, and (iii) because intention is appointed by the theory as the best predictor ofbehaviour (e.g. Ajzen, 1988,1991; Fishbein & Ajzen, 1975) and this theoretical claim hasbeen confirmed by research results (Andrews et al, 2003; Choi et al, 2001, Conradetal, 1992; McMUlan etal, 2005; Vitoria etal, 2006; Wakefield etal, 2004).

To operationalize social influence, we have adopted the three concepts (types) ofsocial influence proposed by De Vries et al (1995); (i) Subjective norms, which is themost used definition of social influence in research on smoking behaviour since it is oneof the core proximal variables in TPB (Ajzen, 1988, 1991; Ajzen & Madden, 1986);(ii) the perceived smoking behaviour of important others, inspired by Bandura'sconcept of indirect, or vicarious learning (Bandura, 1977, 1986), and (iii) the directpressure to smoke from important others, suggested by (Fvans etal, 1976). Regardingapproaches with the purpose of expanding the social influence component of TPB,among these three variables, one represents the normative stream (subjective norms),other represents the direct interpersonal stream (pressure), and the third (perceivedbehaviour) is in-between these two main streams of efforts to extend the socialinfluence component in TPB-based models. The three types of social influence werecombined with referents or important others for adolescents, namely (i) parents (father,mother); (ü) siblings [brother(s), sister(s)]; (iii) peers (best friend, friends, people in thesome year), and (iv) teachers (see also De Vries et al, 1995; Holm et al., 2003; Vitoriaet al, 2006).

Method

ParticipantsThe sample comprises 3,064 seventh grade students with a mean age of 13.5 years old(SD— 1.1) 50.9% were girls, and 13% were repeating the seventh grade. Regardingnationality, 82% w ere Portuguese and most of the remainders were from the Portugueseformer colonies in Africa.

Impaa of social influence on intention to smoke 685

Considering smoking behaviour, 71.3% reported never having smoked ('neversmokers'), 24% smoked occasionally ('occasional smokers'), and 4.7% smoked weeklyor more frequently ('regular smokers'). Regarding intention to smoke, 85% did notintend to smoke in the coming year and 72% did not intend to smoke in the future.

The percentage of smokers among important others were 23.8% of mothers, 43.5%of fathers, 10.5% of brothers, 6.8% of sisters, and 8.9% of best friends (informationreported by adolescents).

Procedure

The data collection took place in 1998 and was conducted as part of the EuropeanSmoking prevention Framework Approach (ESFA; De Vries, Mudde et al., 2003). In thefirst phase of this research, 64 public schools, with seventh grade students, from fivemunicipalities near Lisbon, were invited to participate in the project. Twenty-fiveschools confirmed participation by sending a letter of participation and by appointingan official school contact teacher. The distribution, in terms of urbati/rural location andsocio-economic level, did not differ when comparing schools that accepted with thosewho refused to participate in this study. Schools that refused participation indicatedbeing too busy. Questionnaires were filled in during classes, under teacher supervision.Measurement conditions w ere optimized by assuring respondents the strictconfidentiality of their responses. A pilot of an initial version of the questionnaire,made with Portuguese adolescents, led to the final version of the questionnaire.

Out of the 3,134 questionnaires received, 3,064 (97.8%) were completedsatisfactorily (questions on smoking behaviour were answered) and were included inthe analysis.

Measures

All variables were assessed using a questionnaire developed and validated (De Vries &Kok, 1986; De Vries, Dijkstra, & Kuhlman, 1988, 1995; Markham etal., 2004; Vitoriaet al., 2006), based on the theoretical models I-Change (De Vries, Mudde et al., 2003)and ASE (De Vries & Mudde, 1998; De Vries et al, 1988; De Vries, Backbier, & Kok,1994, 1995). The questionnaire assesses smoking behaviour, intention to smoke, socialinfluence, attitude, and self-efficacy.

Intention to smoke was measured by two questions on adolescents' intention tosmoke in the future and within the following year. Answering categories ranged from'Definitely not' (1) to 'Definitely yes' (7). The sample polychoric correlation betweenthe 2 items (intention to smoke in the future and in the next year) is .79.

Social influence was measured using three different scales assessing subjectivenorms, perceived behaviour, and direct pressure to smoke from parents, siblings, peers,and teachers.

The subjective norms scale assesses adolescents' perceptions of whether importantreferents think they should smoke or not and comprises eight items [father, mother,brother(s), sister(s), best friend, friends, people in the same year, and teachers]measured on a seven-point Iikert-type scale. For example, 'My best friend thinks that. . . ' 'I don't know' or 'I don't have' (1), 'I definitely should not smoke' (2) to 'I definitelyshotild smoke' (7)'.

The perceived behaviour scale assesses adolescents' perception on the smokingbehaviour of important others and included five items [father, mother, brother(s).


sister(s), and best friend] measured on a three-point scale [e.g. 'Does your father smoke?';'No' (1), 'I don't know' or 'I don't have' (2), 'Yes' (3)] and three items (friends, people inthe same year, and teachers) measured on a six-point scale [e.g. 'How many of your friendssmoke?'; 'I don't know' (1), 'Almost nobody' (2) to 'Almost all of them' (6)].

The direct pressure scale measures the adolescents' perception of being directlypressured to smoke by important others evaluated by eight items [father, mother,brother(s), sister(s), best friend, friends, people in the same year, and teachers]. All theitems were answered using a 6-point scale [(e.g. 'Have you ever felt pressure from yourmother to smoke?', 'I don't know' or 'I don't have' (1), 'Never' (2) to 'Very often' (6)].

Attitude measures the adolescents' beliefs on the disadvantages of smoking andincludes three items [('If I smoke, or were to smoke, I consider my behaviour. . . ''Wrong', 'Very stupid', and 'Very bad for my health')], measured on a seven-point scale:'Very wrong' / 'Very stupid' / 'Very bad for health' (1) to 'Very right' / 'intelligent' / 'goodfor my health' (7). Answers 'I don't know' were coded in the middle of the scale (4). Thesample polychoric correlation between the three items are: .71 ('wrong' - 'stupid'), .65('stupid' - 'bad for health'), and .68 ('wrong' - 'bad for health').

Self-efficacy measures the adolescents' perception of their ability to refrain fromsmoking using three items: 'When with others who smoke', 'When feel upset', and'When on the way home from school'. These items were answered using a seven-pointscale: 'I am sure I won't smoke' (1) to 'I am sure I will smoke' (7). Answers 'I don'tknow' were coded in the middle of the scale (4). The sample polychoric correlationsbetween the three items are: .81 ('with others who smoke' - 'when feel upset'), .80('with others who smoke' - 'when on the way home from school'), and .86 ('when feelupset' - 'when on the way home from school').

Analysis

Besides obtaining descriptive statistics and polychoric correlations, data analysisincluded performing confirmatory factor analysis (CFA), as a first step, to confirm eachof the three measurement models of social infiuence (subjective norms, perceivedbehaviour, and direct pressure to smoke), each with four factors (the referents: parents,siblings, peers, and teachers).

As a second step structural equation modelling (SEM) was used to calculatecorrelations between factors of each scale of social influence and intention to smokeand to estimate the impact of social influence on adolescents' intention to smoke. Thisgoal was accomplished firstly by testing each of the three social models separately,secondly by testing all the three types of social influence at the same time, and thirdly bytesting the social influence together with attitudes and self-efficacy, the other proximalvariables of the I-Change model (see Figure 1, ignoring for now the figures). Pastsmoking behaviour was controlled by performing the last analysis using a samplewithout 'regular smokers'.

All items were treated as variables measured on an ordinal scale, and LISREL 8.80(Jöreskog & Sörbom, 2006) was used to calculate polychoric correlations and to performCFA and SEM. A robust maximum likelihood estimation procedure implemented inLISREL was used to estimate all models.

Model-data fit was assessed using the following measures of fit: the Satorra-Bentlerscaled chi-square statistic, the expected cross-vaUdation index (ECVI), the Akaikeinformation criterion (AIC), the root mean square error of approximation (RMSEA), the

Inipact of social influence on intentíon to smoke 687

Standardized root mean square residual (SRMR), the comparative fit index (CFI), andnon-normed fit index (NNFI).

Results

Testing social influence measurement modelsThe first goal of the current study w as to test the structure of the three social influencescales (three types of social infiuence), each with four factors, distinguishing the fourgroups of important referents (parents, siblings, peers, and teachers), as theoreticallyhypothesized (see Figure 1, forgetting for now the figures). To accomplish this goal, foreach social influence scale, three different measurement models w ere tested using CFA:a first-order factor analysis model with four factors w as compared to a first-order single-factor model and to a second-order factor analysis model, as suggested by several authors(e.g. Byrne, 1998; Kline, 1998).

Subjective norms CFA. Results of the four factors model revealed factor loadings thatwere statistically significant and adequate (all greater than .66, on a standardizedsolution - see Table 1). The four factors scale measurement model was confirmed byacceptable fit indices, and revealed to perform better than the single factor model(see Table 2). LISREL estimated correlations between the four factors ranged between.42 and .66 (Table 3). These correlations bet\veen factors of subjective norms were highenough to suggest the use of a second-order model. However, the AIC value obtainedfor the estimated second-order FA model (142.62) was higher than that obtained forthe estimated first-order FA model (120.61), thus suggesting a poorer model-data fit(Table 2). Consequently, the first-order factor analysis model with four factors seems tobe the most appropriate. The obtained factors composite reliability coefficients wereabove the commonly used threshold value for acceptable reliability (Hair, Anderson,Tatham, & Black, 1998): .82 (parents), .75 (siblings), and .78 (peers; teachers wasmeasured with only one item). The average variance extracted for each factor wasalso acceptable (Hair et al, 1998): .70 (parents), .59 (siblings), and .54 (peers).

Perceived behaviour CFA. All factor loadings were statistically significant andadequate (>.4O) supporting the four factors structure (Table 1). Moreover, theoverall goodness-of-fit indices lend substantial support for this four factors structureover the single-factor structure (see Table 2). LISREL estimated correlations betweenthe four factors range between — .01 and .27 (Table 3). In this case, correlationsamong the four factors were low, suggesting a poor fit of a second-order model. Thefit indices confirm that the four factors model is the most appropriate measurementmodel (Table 2). The factors composite reliability coefficients Ç64 for parents, .63for siblings, and .65 for peers), and the average variance extracted for each factor(.48 for parents, .46 for siblings, and .40 for peers) were somewhat below therecommended limit. Despite these values, regarding that our research question is toinvestigate the specificity of types and sources of social influence, we have decidedto accept the four factors model. Even though the obtained coefficients are notfar from the proposed standard limits, w hich the literature does not impose asabsolute (e.g. Hair et al, 1998), the interpretation of our results should be donewith some caution.

Direct pressure to smoke CFA. The scale with four referents •was confirmed in a CFAwith factor loadings that were statistically significant and relatively high (>.55),resulting in a reasonable model-data fit (Table 1). This measurement model also revealed

688 Paulo D. Vitoria et al.

3Vif/i<U

d.

T3CCD

L T

3O

I

-o

SI/)

O

SL_D(/I

WU

bb

o

•g(U

Io

(U

(2

i/5

J3

GO >O 00t ^ vO 1 ^


(U

-5

<uO

-5

ooo

I

ZZ

u

en

to

:

r^ ON o>ON ON ON

00 ro roo o o (N to— o

00 -^ T(- — TT LOo o o — o o

— — rs— Ö «N^ rs ^

lOO

o o oo o o

rs t^ r-.p ON ONON r~^ TTi ^ ro NO1 ^ — —

lO lO LO(N o o

o o oo o oV V V V V V

( ^ - ^ LO•r (N —Ö LO NO(N

lO lO NO

— o o

o o o

vO — -^p — pLO ^O NO

lO (N <NO O O

h mO ro 00o rs —

O —p rs

— — rNi• ^ vO vO

s

O lo 1(N — —

LO NO

ON r s

p — pÖ 00 rJrs — fS

<uX)o

il.S b "H .g 8 bU Í5 O OJ U Mni u -Q rt P¿-.

io

II lu 2ÎÖ o

£ 7, «11 ' - ' * ^ i« i l ' - '

f ^ 8. l f = 8^1Ç o o S Ç o a i t j Ç

t o t i . t o y t o i i - i / ) a in

c >- ^

.£ o(/) Ll.


X I

tu-a3

ai

-5caiai

Vi

OCi

8T3

(2

DP-

DP-

DP-

DP-

PB-

PB-

PB-

PB-

I

Zl/l

z

SN-

«tit

ude

Teac

hers

/

Pee

rsS

iblin

gsP

aren

tsTe

ache

rsP

eers

Sib

lings

Par

ents

Tea

cher

sP

eers

slin

gs

in

Par

ents

Inte

ntio

nIte

m

<7>O

vO

•r 00 (NLO - ^

I 00 00 00 LO 00LO -<r -«r o o

, o 1^ n 00 (N (NI — p — — p pI I

I — o -^ ON t^ t^ 00IN o o TT o — IN

, - ^ I N L O N O — O O l O ^I I N p p L O C N p — IN

I I

Oh^TT N O I N V O O T N O

I <N — O O O O O O O

- ^ N O I N I ^ - ^ L O O O r O N O L OO O O O O O O — O O

, o o o — O N O o o r ^ L o o o i N v oI L O O O l N O O O l N O O —

^h <N NO r^ ^^ ^^ ^^ ro l/l ^^ NÛI N O - ^ p l N — p —;r^|-N4 — p

I I

, N O ^ — IN — r o i N - 5 - L O I N N O r ^ l - ^

I N O N O U - l p p p p p p — p p —I I I

i N o o o o o ^ r o o o L o c ^ r ^ o c N j L O N OI N I N — o — — ^ O O — r O O L O N O

í V,tu "" ö l ^ 12 i

Í5 1 s S ^ 3 g S Í5 ;5 S « -o Sgo . i o û . H - £ i ; 5 a ; f 2 : o . L o o . h - S "yZ Z Z Z C Û C Û C Û C Q D | - Q - U L D L ' B ^o o L O L O L o Q L . a . a . a . Q Q Q Q < o o

ai3

Q."Q3O

1

X)ai>a)

i.CQ'Q.

Eo

ai_

3vt

to

i

Impaa of sodal influence on intention to smoke 691

to have a more adequate fit than the single-factor structure (see Table 2). LISRELestimated correlations values among the four factors ranged from .48 to .63 (Table 3).These values were high, suggesting a second order FA model. Although, the fit indices ofthis model were similar to those obtained by the four factors model, in line withprevious decisions about subjective norms and perceived behaviour, and consideringthe purpose of the current study, we decided to adopt the four factors scalemeasurement model. The three factors composite reliability coefficients were .67(parents), .56 (siblings), and .79 (peers), and the average variance extracted for eachfactor were .92 (parents), .39 (siblings), and .55 (peers). The coefficient for siblings isbelow the proposed standard limit but, for the reasons presented before, we havedecided to continue with the four factors model. As stated before, this set of valuesimpose some caution in the interpretation of our results.

In summary, the goodness-of-fit results and the measurement model assessmentsprovided enough support for the confirmation of the proposed model with three socialinfluence scales (the three types of social influence) each with four factors (the referents).

Testing the impact of social influence on adolescents' intention to smokeThe second goal of this study was to investigate the role of social influence in predictingintention to smoke using the measurement models already tested. Table 3 shows theestimated correlations between intention to smoke and the remaining factors of themodel as well as the intercorrelations among the factors.

Analysing correlations between the intention to smoke and the factors of the socialinfluence model, higher scores were found for perceived behaviour of peers (r = .48),peers' direct pressure to smoke (r = .30), and parental norms (r = .22). Theintercorrelations among factors of different scales were diverse, ranging from r = .73(between siblings norms and siblings direct pressure) to r = —.56 (between siblingsdirect pressure and perceived behaviour of siblings).

To investigate the impact of social influence on adolescents' intention to smoke themodels of subjective norms, perceived behaviour, and direct pressure were first testedseparately.

The subjective norms model revealed adequate goodness-of-fit indices: X' (26) = 98.56,p < .001; AIC = 156.56; RMSEA = .03; SRMR = .02; CFI = 1.00; NNFI = .99. Subjectivenorms explained 5% of the variance of adolescents' intention to smoke. In thismodel parents were the only factor with a significant impact on adolescents' intentionto smoke, with a standardized effect of .15 QS < .05), suggesting that parental expressionof anti-smoking expectations contributed to a low er intention to smoke.

The perceived behaviour model showed also adequate goodness-of-fit indices:X (26) = 245.99, p < .001; AIC = 303.99; RMSEA = .05; SRMR = .06; CFI = .97;NNFI = .95. Perceived behaviour explained 24% of the variance of the intention tosmoke. Referents with significant impact were peers (standardized effect of .48,p < .05) and parents (standardized effect of .07, p < .05), suggesting that adolescentsperceiving more frequent smoking behaviour of their peers, and of their parents, had ahigher intention to smoke.

The direct pressure to smoke model also demonstrated adequate goodness-of-fitindices: x^(26) = 29.12,/? = .31; AIC = 87.12; RMSEA = .01; SRMR = .04; CFI = 1.00;NNFI = 1.00. Results revealed that direct pressure to smoke explained 15% of thevariance of the intention to smoke. Significant impacts with intention were found frompeers (standardized effect of .46, p < .05) and from teachers (standardized effect of


— 32,/J < .05). However, the impact from teachers was in the opposite direction. Thislast effect was unexpected considering the absence of correlation between teachers'direct pressure to smoke and intention to smoke (see Table 3). This effect is likely to becaused by the high intercorrelations between this factor and some others included inthe model, thus, causing a suppressor effect (Tabachnick & Fidell, 1996). Hence, we willconsider pressure from peers as the unique factor with a significant impact on intentionto smoke, suggesting that adolescents feeling higher pressure from peers towardssmoking present a higher intention to smoke.

The second step to investigate the impact of social influence on adolescents' intentionto smoke was to test the overall social influence model, combining the three types of socialinfluence with the four groups of referents. This model resulted in an adequate model-datafit: x\225) = 678.41, p < .001; AIC = 930.41; RMSEA = .03; SRMR = .05; CFI = .99;NNFI = .99 and explained 29% of the variance of adolescents' intention to smoke.Significant estimates (p < .05) of intention to smoke were perceived behaviour of peers(standardized effect of .43), parental norms (standardized effect of .20), and perceivedbehaviour of parents (standardized effect of .08). However, peers' direct pressure tosmoke was found to have a borderline influence (standardized effect of .15; j5 < .10).

Finally, to verify if the effects observed with the proposed social influence modelremain significant when tested in a larger model, we have included in the modelattitudes, and self-efflcacy. These three constructs (social influence, attitudes, and self-efficacy) are the proximal variables of intention to smoke in the I-Change model.Correlations between attitude, self-efficacy, and intention to smoke are higher thancorrelations between social influence scales and intention to smoke (ranging between.52 and .66; see Table 3). Correlations between attitude and social infiuence scales arelow, and range between - .06 and .17, with the exception of direct pressure to smokefrom peers which is higher (37). Correlations between self-efficacy and social influencerange between .02 and .28, with the exception of direct pressure to smoke from peers(.47). The estimates of the model in a completely standardized solution are presented inFigure 1. The model has an adequate model-data fit: x^(363) = 912.79, p< .001;AIC = 1,242.79; RMSEA = .02; SRMR = .04; CFI = .99; NNFI = .99, and explained 55%of the variance of adolescents' intention to smoke.

Even when tested together with attitudes and self-efficacy, the impact of socialinfluence on intention to smoke remained significant, and the factors with a significantimpact on intention to smoke were confirmed: perceived behaviour of peers(standardized effect of .20, p < .05), parental norms (standardized effect of .16,p < .05), and perceived behaviour of parents (standardized effect of .07, p < .05). Thismodel also revealed that attitudes (standardized effect of .25, p < .05) and self-efficacy(standardized effect of .40,p < .05) had a significant impact on adolescents' intention tosmoke which were larger than the impact showed by social influence.

An alternative model was also considered, including the three types of socialinfluence as second-order factors, besides attitudes, and self-efficacy, as explanatoryvariables of intention to smoke. The following fit indices were obtained:X (423) = 1,466.26, p < .001; AIC = 1,676.26; RMSEA = .03; SRMR = .07;CFI = .99; NNFI = .98. Results confirmed those previously obtained with the proposedsocial influence model. A total of 53% of the variance on intention to smoke wasexplained. Perceived behaviour (standardized effect of .17, p < .05) and subjectivenorms (standardized effect of .10, p < .05) had a significant impact on intention tosmoke, along with attitude (standardized effect of .28, p < .05), and self-efficacy(standardized effect of .51,p < .05).


Because past smoking behaviour showed to be an important determinant of intentionto smoke and present smoking behaviour (e.g. Rivis, Sheeran, & Armitage, 2006), wehave also decided to test the proposed model excluding from the sample the subset of'regular smokers' (N = 143). Considering the sample without 'regular smokers'CN = 2,921), the model [x^(367) = 851.25,p< .001; AIC = 1,15325; RMSEA = .02;SRMR = .06; CFI = .99; NNEI = .99] explained 45% of the variance of adolescents'

Next year

Future

Figure I. Path diagram of the proposed model to investigate the impact of social influence on

adolescents' intention to smoke with LISREL estimates in a standardized solution.

694 Paulo O. Vitoria et al.

intention to smoke. In this model, regarding social influence, only parental norms(standardized effect of .15, p < .05) showed to be significant. Perceived behaviour ofpeers (standardized effect of .09) and perceived behaviour of parents (standardizedeffect of .05) had a borderline influence (p < .10). Attitude (standardized effect of .25)and self-efficacy (standardized effect of .39) remained both significant (p < .05).

In summary, results show an impact of social influence on adolescents' intention tosmoke although it is smaller than the impact showed by attitudes and self-efficacy.Results also provide some support to the hypothesis that different referents exertedtheir influence through distinct types of social influence. Among the various factors ofsocial influence tested, parental norms seem to be the most stable in influencingintention. Perceived behaviour of peers and perceived behaviour of parents appear tohave also an impact on intention to smoke. However, the impact of perceived behaviourweakened when past smoking behaviour was controlled for in the analysis.

Discussion

Results provide support to the hypothesized measurement model of social influence,which combines the three types of social influence (subjective norms, perceivedbehaviour, and direct pressure to smoke) with the four groups of important others(parents, siblings, peers, and teachers). This model explained 29% of adolescents'intention to smoke variance, figure that is similar to the result obtained in a previousstudy (De Vries et al, 1995). The extended version of the model, including socialinfluence, attitudes, and self-efficacy, explained 55% of the variance of adolescents'intention to smoke, outcome which is also aligned with results obtained in previousstudies (e.g. Armitage & Conner, 2001).

Considering the three types of social influence, results of the current study suggestthat perceived behaviour, and subjective norms are related with adolescents' intentionto smoke. Direct pressure to smoke did not show a significant impact when the threetypes of social influence were tested simultaneously but, when tested alone, had animpact grater than subjective norms. If past behaviour was controlled for in the analysis,only subjective norms maintain the significant effect on intention to smoke. Perceivedbehaviour, which had the strongest effect among social influence scales when using thesample with all the participants, lost strength, and significance when participants withsmoking experience were removed from the analysis. This observation suggests thatperceived behaviour influence is effective only with adolescents which have alreadysmoking experience.

Regarding the four groups of referents, results show an impact of peers, and parentson adolescents' intention to smoke. Siblings and teachers did not show a significantassociation with adolescents' intention to smoke. Peers had the strongest associationwith intention to smoke. However, when past behaviour was controlled for, peersinfluence became non-significant, and only parents maintained its influence.

An important result of the current study is the suggestion that different referentsexerted their influence on adolescents' intention to smoke through distinct types ofsocial influence. Peers' influence was exerted through perceived behaviour and parentalinfluence was mainly exerted through subjective norms. Hence peers, and parentalinfluences appear to be exerted or experienced by adolescents in different ways.Furthermore, these results suggest that intention to smoke of adolescents in differentpositions in the smoking initiation process was influenced by different types of socialinfluence and referents. These conclusions should be considered with some caution


regarding that some values of factors reliability did not reach the recommendedstandards. This warning suggests the need for more research in this field.

In brief, considering Portuguese youngsters in early adolescence, parental norms arethe only social influence factor with a clear impact on adolescents' intention to smoke.Results of the current study also reveal parents as a consistent and peers as aninconsistent source of influence on adolescents' intention to smoke. These resultscontradict the common position that peers are the strongest source of influence onadolescents' smoking initiation (e.g. Conrad et al., 1992; De Vries, Engels et al., 2003;Kobus, 2003; Tyas & Pederson, 1998) and support more recent research resultsindicating parents as an important source of influence on smoking initiation, particularlyin early adolescence (e.g. Bricker, Peterson, Andersen et al, 2006, Bricker, Peterson,Leroux et al, 2006; De Vries, Engels et al, 2003; McMillan et al, 2005). On the otherhand, the current study has confirmed subjective norms as the type of social influencewith a more consistent impact on adolescents' intention to smoke, wliich is in line withthe classic approach, that gives to subjective norms the role of representing the socialinfluence component in TPB based models (Ajzen, 1988, 1991; Ajzen & Fishbein, 1980;Fishbein & Ajzen, 1975). Moreover, agreeing with the conclusion of a previous study byDe Vries etal (1995), the current results provide some support for the role of perceivedbehaviour in influencing intention to smoke of adolescents, especially of those whohave already smoking experience.

These results sustain the position that social influence is a complex construct andentail the need for further work on the understanding of social influence role inexplanation of intention and behaviour (e.g. Avenevoli & Merikangas, 2003; De Vriesetal, 2006; De Vries, Engels etal, 2003; Hagger & Chatzisarantis, 2005; Kobus, 2003;McMillan et al, 2005). Further longitudinal and cross-sectional research, includingadolescents with different ages, is needed to test the usefulness of this model ininvestigating the impact of social influence on the intention to smoke and on smokinginitiation and the role of these factors as predictors of intention, and smoking. Theoutcome of future studies on the intention-behaviour gap (e.g. Armitage & Conner,2001; Webb & Sheeran, 2006) is also important for research on smoking initiation givingto intention the nuclear role.

The present study has some limitations. The generalization of our results is limitedsince participants were Portuguese and comparisons between Portuguese adolescentsand adolescents from other nationalities have showed noticeable differences (De Vries,Mudde et al, 2003). Another limitation is the indirect gathering of data on parents,siblings, friends, and teachers, based on adolescents' own reports. However, studies inwhich independent reports from adolescents, and referents were obtained have showedthat adolescents appear to be well aware of their parents' and friends' smokingbehaviour and smoking positions (Engels, Finkenauer, Meeus, & Dekovic, 2001; WUks,Callan, & Austin, 1989). Finally, the current study was based on a cross-sectional design,thus, precluding conclusions on the direction of relations between social influence andthe intention to smoke. There is also some evidence that cross-sectional designsunderestimate parental and overestimate peers' influence on smoking intention andonset (e.g. De Vries, Engels et al, 2003; Engels, Knibbe, Drop, & De Haan, 1997).

Nevertheless, the current study has implications: results suggested that adolescents'extended exposure to more smoking (or less anti-smoking) expectations expressed byparents and, somehow, adolescents' exposure to peers, and parental smoking behaviourimply a higher intention to smoke. These findings have some relevant implications forsmoking prevention. Parents' expression of disagreement on their children present or

696 Paulo D. Vitoria et aL

future Smoking behaviour could contribute to adolescents' lower intention to smoke.Another useful preventive strategy could be raising the awareness of smokers, especiallyparents and adolescents, on the pernicious influence of their smoking behaviour ontheir offspring and colleagues. Helping smokers to quit, especially parents and youth,could have a positive impact on preventing smoking intention and initiation.

AcknowledgementsThe authors would like to thank Carlota Simôes Raposo and Filipa Peixoto for their assistance inthe collecting of data and all teachers, students, and others involved in the ESFA projectmanagement and implementation. Paulo Vitoria work was partially supported by a PhD grant(FCT: SFRH/BD/34503/2006). The ESFA project was financed by a grant from the EuropeanCommission (The Tobacco Research and Information Fund; contract 96/IT/13-B96SOC96201157).

ReferencesAjzen, L (1988). Attitudes, personality, and behavior. Buckingham: Open University Press.Ajzen, I. (1991). The theory of planned behaviour. Organizational Behavior and Human

Decision Processes, 50, 179-211.Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behaviour.

Englewood Cliffs, NJ: Prentice-Hall.Ajzen, I., & Madden, T. J. (1986). Prediction of goal-directed behavior: Attitudes, intentions, and

perceived behavioral controV Journal of Experimental Social Psychology, 22, 453-474.Andrews, J. A., Tildesley, E., Hops, H., Duncan, S. C, & Severson, H. H. (2003). Elementary school

age children's future intentions and use of substances. Journal of Clinical Child andAdolescent Psychology, 32, 556-567.

Armitage, C. J., & Conner, M. (2001). Efficacy of the theory of planned behaviour: A meta-analyticreview. British Journal of Social Psychology, 40, 471-499.

Avenevoli, S., & Merikangas, K. R. (2003). Familial infiuences on adolescent smoking. Addiction,98(X), 1-20.

Azevedo, A., Machado, A. P., & Barros, H. (1999). Tobacco smoking among Portuguese high-schoolstudents. Bulletin of the World Health Organization, 77, 509-514.

Bandura, A. (1977). Social learning theory. New York: General Learning Press.Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory.

New York: Prentice-Hall.Bamett, T. A., Gauvin, L., Lambert, M., O'Loughlin, J., Paradis, G., & McGrath, J. J. (2007). The

influence of school smoking policies on student tobacco use. Archives of Pediatrics andAdolescent Medicine, 161(,9X 842-848.

Bente, W, Torsheim, T., Currie, C, & Roberts, C. (2004). National and school policies onrestrictions of teacher smoking: A multilevel analysis of student exposure to teacher smokingin seven European Countries. Health Education Research, 19, 217-226.

Bewley, B. R., Johnson, M. R., & Banks, M. H. (1979). Teachers' smoking./owrw«/ of EpidemiologyCommunity Health, 33OX 219-322.

Bricker, J. B., Peterson, A. V, Andersen, M. R., Rajan, K. B., Leroux, B. G., & Sarason, 1. G. (2006).Childhood friends who smoke: Do they influence adolescents to make smoking transitions?Addictive Behaviors, 31, 889-900.

Bricker, J. B., Peterson, A. V, Leroux, B. G., Andersen, M. R., Rajan, K. B., & Sarason, L G. (2006).Prospective prediction of children's smoking transitions: Role of parents' and older siblings'

^. Addiction, 101, 128-136.


Bricker, J. B., Peterson, A. V, Samson, I. G., Andersen, M. R., & Rajan, K. B. (2007). Changes in theinfluence of parents' and close friends' smoking on adolescent smoking transitions. AddictiveBehaviors, 32, 740-757.

Byrne, B. M. (1998). Structural equation modelling with IJSREL, PREUS and SIMPLIS: Basicconcepts, applications, andprogratnming. New Jersey: La^vrence Eribaum.

Choi, W. S., GUpin, E. A., Farkas, A. J., & Pierce, J. P. (2001). Determining the probability of futuresmoking among adolescents. Addiction, 96, 313-323.

Golby, S. M., Tiffany, S. T., Shiffman, S., & Niaura, R. S. (2000). Are adolescent smokers dependenton nicotine? A review of the evidence. Drug and Alcohol Dependence, 5%S,uçp\ 1), S83-S95.

Conrad, K. M., Flay, B. R., & Hill, D. (1992). Why children start smoking cigarettes: Predictors ofonset. Britishjournalof Addiction, 87, 1711-1724.

De Vries, H., Backbier, E., & Kok, G. (1994). A Dutch social influence smoking preventionapproach for vocational school students. Health Education Research, 9, 365-374.

De Vries, H., Backbier, E., Kok, G., & Dijkstra, M. (1995). The impact of social influences in thecontext of attitude, self-efficacy, intention and previous behaviour as predictors of smokingonst^i. Journal of Applied Social Psychology, 25, 237-257.

De Vries, H., Candel, M., Engels, R., & Mercken, L. (2006). Challenges to the peer influenceparadigm: Results for 12-13 year olds from six European countries from the Europeansmoking prevention framework approach study. Tobacco Control, 15, 83-89.

De Vries, H., Dijkstra, M., & Kuhlman, P. (1988). Self-efficacy: The third factor besides attitude andsubjective norm as a predictor of behavioural intentions. Health Education Research, 3,273-282.

De Vries, H., Engels, R., Kremers, S., Wetzels, J., & Mudde, A. (2003). Parents' and friends' smokingstatus as predictors of smoking onset: Findings from six European countries. HealthEducation Research, 18, 627-636.

De Vries, H., & Kok, G. J. (1986). Erom determinants of smoking behaviour to the implications fora prevention programme. Health Education Research, 1, 85-94.

De Vries, H., Mudde, A., Leijs, I., Charlton, A., Vartiainen, E., Buijs, G., etal (2003). The EuropeanSmoking Prevention Framework Approach (EESA): An example of integral prevention. HealthEducation Research, 18, 611-626.

De Vries, H., & Mudde, A. N. (1998). Predicting stage transitions for smoking cessation applyingthe attitude-social influences-efficacy model. Psychological Health, 13, 369-385.

DiEranza, J. R., Rigotti, N. A., McNeill, A. D., Ockene, J. K., Savageau, J. A., St Cyr, D., etal (2000).Initial symptoms of nicotine dependence in adolescents. Tobacco Control, 9, 313-319.

DiEranza, J. R., Savageau, J. A., Rigotti, N. A., Fletcher, K., Ockene, J. K., McNeill, A. D., etal. (2002).Development of symptoms of tobacco dependence in youths: 30 month follow up data fromthe DANDY study. Tobacco Control, 11, 228-235.

Engels, R. C. M. E., Einkenauer, C, Meeus, W., & Dekovic, M. (2001). Parental attachment andadolescents' emotional adjustment: The associations with social skills and relationalcompetence./oMmß/ of Counselling. Psychology, 48, 428-439.

Engels, R. C. M. E., Knibbe, R. A., De Vries, H., Drop, M. J., & Van Breukelen, G. J. P (1999).Influences of parental and best friends' smoking and drinking on adolescent use: Alongitudinal study./owrw«/ of Applied Social Psychology, 29, 337-361.

Engels, R. C. M. E., Knibbe, R. A., Drop, M. J., & De Haan, Y T. (1997). Homogeneity of cigarettesmoking within peer groups: Influence or selection? Health Education Behaviour, 24,801 -811.

European Commission (2003). Smoking and the environment: Actions and attitudes - specialeurobarometer 183. Brussels: European Commission.

European Commission (2007). Attitudes of Europeans towards tobacco - spedat eurobarometer272c. Brussels: European Commission.

Evans, R. I. (1976). Smoking in children: Developing a social psychological strategy of deterrence.Joumai of Preventive Medicine, 5, 122-127.


Evans, R. I., Dratt, L. M., Raines, B. E., & Rosenberg, S. S. (1988). Social influences on smokinginitiation: Importance of distinguishing descriptive versus mediating variables. Journal ofApplied Social Psychology, 18, 925-943.

Evans, R. L, Rozelle, R. M., Mittelmark, M. B., Hansen, W. B., Bane, A. L, & Havis, J. (1978).Deterring the onset of smoking in children: Knowledge of immediate physiological effects andcoping with peer pressure, media, and parent modeling. Journal of Applied SocialPsychology, 8, 126-135.

Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction totheory and research. Reading, MA: Addison-Wesley.

Gibbons, F. X., Gerrard, M., & Lane, D. J. (2003). A social-reaction model of adolescent healthrisk. In J. M. Suis & K. A. Wallston (Eds.), The handbook of social-health psychology. Oxford:Blackwell.

Godin, G., Conner, M., & Sheeran, P. (2005). Bridging the intention-behaviour 'gap': The role ofmoral norm. British Journal of Social Psychology, 44, 497-512.

Hagger, M. S., & Chatzisarrantis, N. L. D. (2005). First- and higher-order models of attitudes,normative influence, and perceived behaviour control in the theory of planned behaviour.British Journal of Social Psychology, 44, 513-535.

Hagger, M. S., & Chatzisarrantis, N. L. D. (2006). Self-identity and the theory of planned behaviour:Between- and within-participants analyses. British Journal of Social Psychology, 45, 731-757.

Hair, J. E, Anderson, R. E., Tatham, R. L., & Black, W. (1998). Multivariate data analysis(5th ed.). Englewood Cliffs, NJ: Prentice-Hall.

Holm, K., Kremers, S. P J., & De Vries, H. (2003). Why do Danish adolescents take upsmoking? European Journal of Public Health, 13, 67-74.

Jöreskog, K. G., & Sörbom, D. (2006). LISREL 8.80 for Windows [Computer Software].Lincolnwood, IL: Scientiflc Software International.

Kline, R. B. (1998). Principles and practice of structural equation modelling. London: GuilfordPress.

Kobus, K. (2003). Peers and adolescent smoking. Addiction, 5S(Suppl 1), 37-55.Lopez, A. D., Mathers, G. D., Ezzati, M., Jamison, D. T., & Murray, G.J. L. (2006). Global and regional

burden of disease and risk factors, 2001: Systematic analysis of population health data. Lancet,36, 1747-1757.

Markham, W A., Aveyard, P, Thomas, H., Gharlton, A., Lopez, M. L., & De Vries, H. (2004). Whatdetermines ñiture smoking intentions of 12- to 13-year-old UK African-Garibbean Indian,Pakistani and white young people? Health Education Research, 19, 15-28.

Matos, M. G., SimÖes, G., Tomé, G., Gaspar, T., Gamacho, L, & Diniz, J. A. (2006). A saúde dosadolescentes portugueses hoje e em 8 anos - Relatório preliminar do estudo HBSC 2006.Lisboa: Faculdade de Motricidade Humana.

McMillan, B., Higgins, A. R., & Gönner, M. (2005). Using an extended theory of planned behaviour tounderstand smoking amongst schoolchildren. Addiction Research and Theory, 13, 293-306.

Nebot, N., Tomás, Z., Ariza, G., Valmayor, S., López, M. J., átjuárez, O. (2004). Factors associated withsmoking onset: 3-Year cohort study of schoolchildren. Archivos Bronconeumologia, 40,495-501.

Nebot, M., Tomás, Z., Ariza, G., Valmayor, S., & Mudde, A. (2002). Factores asociados con laintención de fumar y el inicio del habito tabaquico en escolares: Resultados del estudio ESFAen Barcelona. Gaceta Sanitaria, 16, 131-138.

O'Loughlin, J., Paradis, G., Renaud, L., & Gomez, L. S. (1998). One-year predictors of smokinginitiation and of continued smoking among elementary schoolchildren in multiethnic, low-income, inner-city neighbourhoods. Tobacco Control, 7, 268-275.

Peto, R., Watt, J., & Boreham, J. (2006). Deaths from smoking; An electronic resource. Geneva:International Union Against Gancer (UIGG).

Rajan, K. B., Leroux, B. G., Peterson, A. Y, Jr., Bricker, J. B., Andersen, M. R., Kealey, K. A.,etaL (2003). Nine-year prospective association between older siblings' smoking and children'sdaily smoking./owma/ of Adolescent Health, 33, 25-30.


Rhodes, R. E., Jones, L. W., & Courneya, K. S. (2002). Extending the theory of planned behavior inthe exercise domain: A comparison of social support and subjective norm. Research Quarterlyfor Exercise and Sport, 73(2), 193-199.

Rivis, A., & Sheeran, P (2003). Descriptive norms as an additional predictor in the theory ofplanned behaviour; A meta-analysis. Current Psychology, 22, 218-233.

Rivis, A., Sheeran, P., & Armitage, C. J. (2006). Augmenting the theory of planned behaviour withthe prototype/willingness model; Predictive validity of actor versus abstainer prototypes foradolescents' health-protective and health-risk intentions. British Journal of HealthPsychology, 11, 483-500.

Santos, A. C, & Barros, H. (2004). Smoking patterns in a community sample of Portuguese adults,1999-2000. Preventive Medicine, 38(1}, 114-119.

Slomkowski, C, Rende, R., Novak, S., Uoyd-Richardson, E., & Niaura, R. (2005). Sibling effects onsmoking in adolescence; Evidence for social influence from a genetically informative design.Addiction, 100, 430-438.

Tabachnick, B. G., & FideU, L. S. (1996). Using multivariate statistics (3rd ed.). New York; HarperCollins.

Tyas, S. L., & Pederson, L. L. (1998). Psychosocial factors related to adolescent smoking; A criticalreview of the literature. Tobacco Control, 7, 409-420.

US Department of Health and Human Services (1994). Preventing tobacco use among youngpeople: A report of the surgeon general. Washington, DC; US Department of Health andHuman Services.

Vitoria, P. D., Kremers, S. P J., Mudde, A., Pais Clemente, M., & De Vries, H. (2006). Psychosocialfactors related \vith smoking behaviour in Portuguese adolescents. European Journal ofCancer Prevention, 15, 531-540.

Wakefield, M., Kloska, D. D., O'Malley, P M., Johnston, L. D., Chaloupka, F., Pierce, J., etal (2004).The role of smoking intentions in predicting future smoking among youth; Findings frommonitoring the future data. Addiction, 99, 914-922.

Webb, T. L., & Sheeran, P. (2006). Does changing behavioral intentions engender behavior change?A meta-analysis of the experimental evidence. Psychological Bulletin, 132, 249-268.

Wilks, J., Callan, V. J., & Austin, D. A. (1989). Parent, peer and personal determinants of adole-scent drinking. British Journal of Addiction, 84, 619-630.

Wold, B., Torsheim, T., Currie, C, & Roberts, C. (2004). National and school policies onrestrictions of teacher smoking; A multilevel analysis of student exposure to teacher smokingin seven European countries. Health Education Research, 19(5}, 217-226.

World Health Organisation (1997). Tobacco or health: A global status report. Geneva; WorldHealth Organisation.

World Health Organisation (1999). The world health report. Geneva; World Health Organisation.

Received 10 April 2008; revised version received 26 November 2008

Copyright of British Journal of Health Psychology is the property of British Psychological Society and its

content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's

express written permission. However, users may print, download, or email articles for individual use.

the impact of social influence on adolescent intention to ... · fishbein, 1980; fishbein &...

Documents