Elsevier

Social Science & Medicine

Volume 200, March 2018, Pages 249-257
Social Science & Medicine

A longitudinal study into the reciprocal effects of identities and smoking behaviour: Findings from the ITC Netherlands Survey

https://doi.org/10.1016/j.socscimed.2017.12.006Get rights and content

Highlights

  • Smoking-related identities and smoking behaviour are reciprocally related.

  • Intention to quit and smoking behaviour are important for identity change.

  • Quitter identity is more important for smoking cessation than smoker identity.

  • Socio-economic status does not moderate relations between behaviour and identity.

Abstract

Objective

Although it has been found that identity constructs related to smoking are associated with changes in smoking behaviour, the direction of causal associations is as yet unclear. This study aims to clarify the nature and direction of these associations.

Methods

In this longitudinal study we examined the reciprocal relations between identity constructs (i.e., smoker self-identity, quitter self-identity and smoker group-identity), intention to quit and smoking and quitting behaviour among a sample of 1036 smokers and ex-smokers, using cross-lagged structural equation modelling. Moreover, we tested whether these relations differed by socio-economic status (SES).

Results

Identity and smoking behaviour were reciprocally related in that in intention to quit and smoking behaviour consistently predicted identity change, and identity predicted (changes in) intentions to quit and smoking behaviour. Behaviour appears more important for identity change than identity for behaviour change. Furthermore, quitter self-identity appears more important than smoker self- and group-identity. Relationships did not differ significantly between SES-groups. The findings were replicated using a cross-validation sample.

Conclusion

Results imply that changing smoking behaviour may be a vehicle to change smoking-related identity. Moreover, strengthening identification with quitting is more crucial for quit success than decreasing smoker identities. The finding that behaviour may be more important for identity than vice versa, if replicated, may call for additions to identity theories.

Introduction

People are motivated to behave in line with their identity. PRIME theory (PRIME stands for plans, responses, impulses, motivation and evaluation) defines identity as thoughts, feelings and images we have of ourselves. It is one of several theories that recognise that identity can be a particularly powerful influence on behaviour (West, 2006). Identity can be based on behaviours, such that particular behaviours are important for the way that people perceive themselves (i.e., self-identity). In addition to identification with behaviours, the social identity approach states that people may derive an important part of who they are from their memberships in groups or social categories (Tajfel and Turner, 1979, Turner et al., 1987), that is, their social identity (or group-identity). People are likely to behave according to the group's social norms when their group identification is strong (Tajfel and Turner, 1979, Tajfel and Turner, 1986). People not only hold perceptions of the self in the present, but in addition have views on who they may become in the future (Barreto and Frazier, 2012).

Research on smoking and identity typically examines “self-identity” and “group-identity”. Self-identity in relation to smoking refers to the importance of behaviours such as smoking and quitting for how individuals perceive themselves (e.g., ‘Smoking is important for who I am’). Whereas group-identity is very similar to the construct of social identity, self-identity can be seen as a part of personal identity as defined in the social identity approach (i.e., an individuals' perception of the self as a unique person that is different from others). Although it has been shown that both self- and group identity are related to smoking behaviour and smoking cessation, the (causal) direction is as yet unclear. Therefore, applying a three-wave cross-lagged model, we examined whether these identities affect smoking behaviour, or vice versa, or that identity and smoking (cessation) behaviour are reciprocally related.

Most studies on smoking (cessation) and identity focused on identity as a precursor of behaviour. This work has clearly shown that identity is important for quit intentions (a key predictor of quitting; Smit et al., 2014, Vangeli et al., 2011) as well as smoking and quitting behaviour, even when controlling for important factors such as nicotine dependence (Hertel and Mermelstein, 2012, Høie et al., 2010, Meijer et al., 2015, Meijer et al., 2016, Moan and Rise, 2005, Moan and Rise, 2006, Tombor et al., 2013, Van den Putte et al., 2009). Smokers who identify more with smoking as a behaviour or with the group of smokers have weaker quit intentions, are less likely to quit, and may even increase their smoking. In contrast, smokers who identify more with quitting, non-smoking, or non-smokers have stronger quit intentions and are also more likely to attempt to quit. In line with the above findings, the social identity model of cessation maintenance (Frings and Albery, 2015) and the social identity model of recovery (Best et al., 2015) propose that stronger (social) identification as ‘recovering addict’ facilitates recovery from addiction. Notably, whereas relations between identity and smoking behaviour are typically examined in prospective designs, allowing for interpretations regarding directionality, associations between identity and quit intention are usually investigated cross-sectionally, such that the direction remains unclear (Meijer et al., 2015). In sum, previous work suggests that identity affects smoking behaviour (see West, 2006).

However, other studies suggest a reversed causal order: people base their self-conceptualizations on behaviours that they frequently engage in, such that the behaviour is perceived to show who they are (Bem, 1972). With regard to smoking, two studies indeed suggest that smoking behaviour affects smoking-related identities. Specifically, after participating in a smoking cessation program, successful ex-smokers came to perceive themselves more as non-smokers and less as smokers (Shadel et al., 1996). Furthermore, increases in smoking behaviour are associated with subsequent increases in smoker self-identity among adolescent smokers (Hertel and Mermelstein, 2016).

Finally, retrospective qualitative studies showed that smoking became increasingly less important to the way ex-smokers perceived themselves as they learned to live without smoking (Brown, 1996; Luck and Beagan, 2015, Vangeli and West, 2012), suggesting that identity change and smoking behaviour change go hand in hand (cf. identity shift theory; Kearney & O'Sullivan, 2003). Similarly, identity theory states that people act in line with their identity, but at the same time identity may change to match behaviour (Stets and Burke, 2003). Moreover, the social identity model of recovery (Best et al., 2015) acknowledges that successful behaviour change may reinforce recovery identities.

Smoking is more prevalent and persistent among those with lower socio-economic status (e.g., Bricard et al., 2016, Pisinger et al., 2011, Reid et al., 2010). Evidence also suggests that identity and identity dynamics differ with SES. A large-scale longitudinal study showed that lower-SES smokers (vs. middle and higher-SES) and lower-SES ex-smokers (vs. middle-SES) identify more with smoking (Meijer et al., 2017). In addition, higher-SES smokers and ex-smokers move away from a smoking identity and toward a quitting identity more quickly than their lower-SES counterparts. Correspondingly, other work showed that lower-SES smokers have more difficulty picturing themselves as non-smokers than higher-SES smokers, whereas the relation between non-smoker self-identity and quit intention was stronger among lower-SES than higher-SES smokers (Meijer et al., 2015). This suggests that non-smoker self-identities may be particularly key for smoking cessation among lower-SES smokers, although SES did not moderate relations between identity and quit intention in another study (Meijer et al., 2016).

In sum, previous work showed that identity is important for smoking behaviour and vice versa, and that other variables such as SES may possibly influence this relationship. However, it is as yet unclear how identity changes and behaviour changes over time are associated, and what the role of quit intention is. The current longitudinal study examined and compared unique relations between identity constructs at the self- and group-level (i.e., smoker self-identity, quitter self-identity and smoker group-identity), quit intention and smoking behaviour among a large sample of smokers and ex-smokers. Cross-lagged structural equation modelling was applied to investigate and compare these relations and cross-validation was used to assess generalizability of results. The following research questions were addressed (RQs):

  • 1.

    Do smoker self-identity, quitter self-identity and smoker group-identity predict changes in smoking behaviour over time (RQ1)?

  • 2.

    Does smoking behaviour predict changes in smoker self-identity, quitter self-identity and smoker group-identity over time (RQ2)?

  • 3.

    Do quitter self-identity, smoker self-identity and smoker group-identity predict changes in quit intention over time (RQ3)?

  • 4.

    Does quit intention predict changes in quitter self-identity, smoker self-identity and smoker group-identity over time (RQ4)?

  • 5.

    Do identity constructs and quit intention uniquely predict smoking behaviour one year later (RQ5)?

  • 6.

    Are relations between identity (intention) and smoking behaviour mediated by intention (identity; RQ6)?

  • 7.

    Do associations over time between identity, quit intention, and behaviour differ between lower, middle and higher-SES groups (RQ7)?

Section snippets

Participants

This study is part of the International Tobacco Control Policy Evaluation Project (www.itcproject.org) (Fong et al., 2006). Data used for the current study were collected annually in the International Tobacco Control (ITC) Netherlands Survey from 2009 to 2014 (from now waves 1-6, respectively). The ITC Netherlands Survey is the only ITC Survey that assesses identity. The same larger dataset was used in one other study that examined a different set of questions and variables, namely psychosocial

Preliminary analyses

Correlations between the variables that were used in the models were examined first (see Supplement F). Almost all correlations were significant and in the expected direction. Smoker self- and group-identity correlated positively, and both smoker identity constructs correlated negatively with quitter self-identity. Furthermore, quit success -where higher scores indicate longer abstinence-correlated negatively with smoker identities and positively with quitter self-identity. Stronger quit

Discussion

This large-scale longitudinal study examined relations between identity (i.e., smoker self-identity, quitter self-identity and smoker group-identity), quit intention and quit success among smokers and ex-smokers, and tested whether these relations differ with socio-economic status (SES). Cross-lagged structural equation modelling was used because it allowed for examination of the research questions, and cross-validation was used to test generalizability of the findings. Indeed, results held up

Conclusions

The results have important implications. The finding that behaviour may be more important for identity than vice versa, if replicated, may call for additions to identity theories. Moreover, changing smoking behaviour may be a vehicle to change smoking-related identity, for example through smoking cessation counselling. Furthermore, quitter self-identity appeared more important for quit intentions and smoking behaviour than smoker identities. Future research should therefore investigate ways to

Funding

Financial support for this research was provided by the Netherlands Organization for Health Research and Development (ZonMw) (121010008, 200130002, 70000001).

RW has undertaken consultancy for, and received research funding and travel expenses from, companies that develop and manufacture smoking cessation medications (Pfizer, GSK, and J&J). He is an unpaid advisor to the UK's National Centre for Smoking Cessation and Training. His salary is funded by Cancer Research UK. The other authors declare

References (52)

  • I. Tombor et al.

    Positive smoker identity as a barrier to quitting smoking: findings from a national survey of smokers in England

    Drug Alcohol Depend.

    (2013)
  • M.L. Barreto et al.

    Coping with life events through possible selves

    J. Appl. Soc. Psychol.

    (2012)
  • D. Best et al.

    Overcoming alcohol and other drug addiction as a process of social identity transition: the Social Identity Model of Recovery (SIMOR)

    Addiction Res. Theor.

    (2015)
  • D. Bricard et al.

    Educational inequalities in smoking over the life cycle: an analysis by cohort and gender

    Int. J. Publ. Health

    (2016)
  • J.M. Brown

    Redefining smoking and the self as a non-smoker

    West. J. Nurs. Res.

    (1996)
  • G.J. Burkholder et al.

    An illustration of a longitudinal cross-lagged design for larger structural equation models

    Struct. Equ. Model.

    (2003)
  • J.E. Cameron

    A three factor model of social identity

    Self Ident.

    (2004)
  • G.T. Fong et al.

    The conceptual framework of the international Tobacco Control (ITC) policy evaluation Project

    Tobac. Contr.

    (2006)
  • O. Fotuhi et al.

    Patterns of cognitive dissonance-reducing beliefs among smokers: a longitudinal analysis from the International Tobacco Control (ITC) Four Country Survey

    Tobac. Contr.

    (2013)
  • A.W. Hertel et al.

    Smoker identity and smoking escalation among adolescents

    Health Psychol.

    (2012)
  • A.W. Hertel et al.

    Smoker identity development among adolescents who smoke

    Psychol. Addict. Behav.

    (2016)
  • M. Høie et al.

    An extended version of the theory of planned behaviour: prediction of intentions to quit smoking using past behaviour as moderator

    Addiction Res. Theor.

    (2010)
  • L.T. Hu et al.

    Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives

    Struct. Equ. Model.: A Multidisciplinary Journal

    (1999)
  • M.H. Kearney et al.

    Identity shifts as turning points in health behaviour change

    West. J. Nurs. Res.

    (2003)
  • K. Luck et al.

    Occupational transition of smoking cessation in women: “You’re restructuring your whole life”

    J. Occup. Sci.

    (2015)
  • H. Markus et al.

    Possible selves

    Am. Psychol.

    (1986)
  • Cited by (21)

    • What kind of smoking identity following quitting would elevate smokers relapse risk?

      2021, Addictive Behaviors
      Citation Excerpt :

      Identification as a smoker trying to quit may form a label that is less committed to cessation and an absence of rules providing motivation to prevent relapse. Past research has found that individuals who identify as a ‘smoker who does not smoke’ find quitting to be challenging and that some element of non-smoking needs to be incorporated into identity to reach stable abstinence (Meijer et al., 2018). However, our finding seems to contradict this given the low relapse rate observed among former smokers who described themselves as a smoker but committed to not smoking again which suggests that identity label is less important than level of commitment to a non-smoking lifestyle and having the non-smoking rules established in maintaining abstinence.

    • Social identity differentiation predicts commitment to sobriety and wellbeing in residents of therapeutic communities

      2019, Social Science and Medicine
      Citation Excerpt :

      At T2, participants reported a reduction in the proportion of their social network comprised of substance users along with an increase in the proportion of their social network that was comprised of other people in recovery (Bathish et al., 2017). Previous research has focused on the influence that a substance using identity (e.g., smoking group identity in Meijer et al., 2018), or recovery identity (e.g., Mawson et al., 2015) have on recovery trajectories. In this study, we examine identity differentiation – defined as the extent to which respondents endorse a recovery identity more strongly than a user identity when the two are measured simultaneously – in relation to six-month outcomes.

    View all citing articles on Scopus
    View full text