Elsevier

Appetite

Volume 53, Issue 1, August 2009, Pages 66-75
Appetite

Research report
Adolescent soft drink consumption, television viewing and habit strength. Investigating clustering effects in the Theory of Planned Behaviour

https://doi.org/10.1016/j.appet.2009.05.008Get rights and content

Abstract

Clustering refers to the co-occurrence of behaviour and may be particularly relevant in light of the present obesity epidemic. Since evidence regarding clustering of motivational and habitual constructs within the framework of the Theory of Planned Behaviour (TPB) is limited, clustering effects of TPB cognitions and habit strength regarding soft drink consumption and television viewing were studied in a sample of Dutch adolescents (n = 312; mean age = 14.62; SD = 1.62) using cross-sectional data. Results showed that not only soft drink consumption and television viewing cluster (r = .42), but also their intentional (r = .36) and habitual (r = .37) constructs. Furthermore, unmediated effects were found between habit strength and its respective behaviour, whereas habit strength was associated with its clustered behaviour through decreased perceptions of controllability. Our findings suggest that interventions that aim to change habitual soft drink consumption and television viewing may need to incorporate an environmental component, as well as explore the potential usefulness of synergistic effects of incorporating multiple clustered behaviours, as well as their corresponding beliefs and habits in health behaviour change interventions.

Introduction

The prevention of adolescent overweight and obesity is an important public health issue, because of their relation with various health risks (Goran et al., 2003, World Health Organization, 1997). Additionally, overweight adolescents are more likely to become overweight and obese adults (Magarey et al., 2003, Wang et al., 2008). While genetic factors are related to the individual onset of weight gain, these factors are unlikely to be responsible for the current obesity epidemic (Kumanyika, Jeffery, Morabia, Ritenbaugh, & Antipatis, 2002; World Health Organization, 1997). Weight gain results from a positive energy balance, in which energy intake (through diet) exceeds energy expenditure, which is largely modifiable through changes in physical activity. Soft drink consumption and sedentary behaviours, such as television viewing, are important behaviours in adolescents that exert pressure to a positive energy balance (Crespo et al., 2001, Henderson, 2007, Kremers et al., 2005a, Rennie et al., 2005). Consequently, interventions to change these behaviours in a more healthy direction are urgently needed (Kremers et al., 2005a, Sirard and Barr-Anderson, 2008).

Effective intervention design necessitates the theory-based identification of important and modifiable proximal determinants of the behaviour to be targeted in an intervention. One of the most commonly used theoretical frameworks is the Theory of Planned Behaviour (TPB) (Ajzen, 1991). According to the TPB, any given behaviour results from the intention to perform this behaviour. Intention is, in turn, influenced by three social-cognitive concepts, which are attitude, subjective norm, and perceived behavioural control (PBC). Attitude is the degree to which performing this behaviour is positively or negatively valued, while subjective norm refers to the degree to which important others feel that this behaviour should be performed. Finally, PBC measures the extent to which behaviour is easy or difficult, as well as under one's control or not. When perceived behavioural equals actual control, the TPB also postulates that PBC influences behaviour directly.

Traditionally, determinant research within the framework of the TPB has focused on understanding single health behaviours (Ajzen, 1991). Subsequently, health behaviour change interventions have often been designed to change single behaviours, including sedentary behaviours (Simon et al., 2004) and soft drink consumption (French, Hannan, & Story, 2004). However, increased scientific interest is turning towards the investigation of the occurrence of multiple behaviours simultaneously (Keller et al., 2008, Kremers et al., 2005a, Kremers et al., 2004, Schuit et al., 2002). Recent empirical evidence indeed indicates that health behaviours in youth tend to cluster (Kremers et al., 2004, Rosenberg et al., 2007). Clustering refers to the co-occurrence of behaviours that is more prevalent than can be expected based on the prevalence of the separate behaviours (Schuit et al., 2002) and can be analyzed through formal cluster analyses or prevalence odds ratios (Everitt, 1993, Schuit et al., 2002). Clustering has also been studied using correlational and regression analysis (De Vries et al., 2007, Kremers et al., 2004, Kremers et al., 2007, Rosenberg et al., 2007).

However, research that has simultaneously looked at television viewing and soft drink consumption has largely overlooked the possible clustering of these two behaviours. While the evidence for clustering of adolescent soft drink consumption and television viewing is therefore limited, a recent study amongst Dutch adolescents (Kremers et al., 2007) did report clustering of soft drink consumption and television viewing, with those who spent more time watching television also drinking more soft drinks (β = .28, p < .001). Such findings warrant further research into the combined investigation of behavioural causes rather than a focus on single behaviours (Kremers et al., 2005b, Kremers et al., 2007). Importantly, evidence that provides insight into clustering of behavioural determinants may not only have theoretical implications, but synergistic effects between motivational constructs may also allow for more effective intervention development (Kremers et al., 2004). For instance, interventions that are successful in inducing positive changes in cognitions and intentions in one behaviour may lead to positive changes in a clustered behaviour, without the latter actually being targeted in an intervention.

Next to the investigation of clustering effects regarding health behaviours and their motivational constructs, scientific efforts have also turned towards the study of automatic or habitual processes in the explanation of health behaviour (de Bruijn et al., 2007a, de Bruijn et al., 2007b, de Bruijn et al., 2008, Kremers et al., 2007). While the TPB assumes that health behaviour is the resultant of cognitive evaluations and planned intentions, everyday behaviours, such as television viewing and soft drink consumption are a natural part of adolescents’ everyday life that probably do not require much intentional effort to be set in motion (Aarts et al., 1997, Kremers et al., 2007). Rather, such behaviours may have become habitual. Recent determinant studies have therefore begun incorporating a survey-based measure of habit strength (Verplanken & Orbell, 2003) and results have consistently shown that habit strength adds significantly to the amount of explained variance in health behaviour across various age groups (de Bruijn et al., 2009, de Bruijn et al., 2008), including adolescents (Reinaerts, De Nooijer, Candel, & De Vries, 2007). Such findings illustrate that health behaviour may indeed become partially habitual.

To date, however, there has been little scientific effort to study the role of habit strength from a clustering perspective within the framework of the TPB. That is, studies on clustering have often focused on covariations between behaviours (Burke et al., 1997, De Vries et al., 2007, Kremers et al., 2004, Rosenberg et al., 2007) and TPB variables (De Vries et al., 2007, Kremers et al., 2004), but have often failed to investigate the potential clustering of behavioural habits. Some evidence, however, points to the usefulness of incorporating habit strength when studying clustering effects of adolescent television viewing and soft drink consumption. Kremers et al. (2007) reported positive associations between soft drink consumption habit strength and television viewing habit strength (r = .50), indicating that a quarter of the variance in soft drink consumption habit strength can be explained by television viewing habit strength. Perhaps more importantly, this study also found that more time spent on television viewing was associated with a stronger habit towards soft drink consumption (β = .30, p < .01). These results underline the potential importance of studying the role of habit strength from a clustering perspective, because it allows for the identification of clustering processes at a more automatic level in addition to cognitive and motivational clustering processes.

Nevertheless, the reasons for the associations between habit strength and both its respective and clustered behaviour remain unclear. However, some experimental evidence points to the apparent difficulty in suppressing and controlling habitual behavioural responses (Aarts and Dijksterhuis, 2000, Bargh, 1994, Shiffrin and Schneider, 1977). For instance, Aarts and Dijksterhuis (2000) found that, in response to the activation of a travel goal (i.e. going to university), the habitual response (e.g. going by bicycle) was more difficult to suppress than the non-habitual response (e.g. going by car). Because controllability of behaviour (i.e. PBC) is acknowledged in the TPB as an important antecedent of intention and behaviour (Ajzen, 2002), strong habits may influence behaviour through decreased perceptions of controllability (e.g. a stronger habit towards television viewing influences actual television viewing through decreased perception of controllability regarding limited television viewing). Furthermore, because television viewing and soft drink consumption tend to co-occur in a similar context (Kremers et al., 2007, Lemish, 1987), when adolescents have difficulties to control television viewing behaviour, they may also have diminished perceptions of controllability regarding behaviours that co-occur with television viewing, such as soft drink consumption. Testing these assumptions may yield potentially important theoretical knowledge (i.e. can behaviourally specific motivational constructs and habits impact on co-occurring behaviours?), but may also allow for more effective health behaviour change interventions.

At present, however, no empirical evidence exists that has explicitly modelled PBC as a potential mediator in the habit strength–health behaviour relationship. A conceptual model combining the theoretical relations in the TPB and the above-cited experimental evidence was therefore formed (see Fig. 1) in order to explore this issue in more detail. We hypothesized that PBC would be the key mediator in the habit strength–health behaviour relationship with habit strength and PBC showing inverse relationships. Furthermore, in line with earlier evidence (Kremers et al., 2004, Kremers et al., 2007), positive clusters of television viewing and soft drink consumption, as well as positive clusters between their respective TPB-concepts and habit strength, were also expected.

Section snippets

Subjects and procedures

Data for the present study were available from 312 students of a combined vocational and secondary school in Gouda, the Netherlands. Gouda is a city of approximately 70,000 inhabits, located in the western part of the Netherlands. Distribution of gender, age, and ethnicity in Gouda is similar to that of the Netherlands (Statistics Netherlands, 2002). Students were asked to complete a self-administered questionnaire during school hours in April 2008. A total of 312 students (65.3% girls; n = 203)

Descriptives and bivariate associations

Table 1 shows mean scores for, and bivariate associations between, study variables. Mean soft drink consumption was 476.52 ml per day, while, on average, respondents watched television for 434.99 min per week. Respondents had a negative intention towards limited television viewing and soft drink consumption, but a positive attitude towards limited television viewing and soft drink consumption. Mean scores for PBC regarding limited soft drink consumption and limited television viewing were above

Discussion

In line with the TPB-relations, univariate results from the present study showed that PBC and intention were the strongest TPB-correlates of both soft drink consumption and television viewing. In turn, PBC was also the strongest TPB-correlate of intention to limit either soft drink consumption or television viewing. Additionally, attitude was a significant correlate of intention to limit television viewing, but not for intention to limit soft drink consumption. The latter findings were also

Acknowledgements

The authors wish to thank Pauline Rambonnet en Silvester van de Sande for their efforts in the data-gathering process.

References (53)

  • E. Reinaerts et al.

    Explaining school children's fruit and vegetable consumption. The contributions of availability, accessibility, exposure, parental consumption and habit in addition to psychosocial factors

    Appetite

    (2007)
  • K.L. Rennie et al.

    Behavioural determinants of obesity. Best practice and research

    Clinical endocrinology and metabolism

    (2005)
  • D.E. Rosenberg et al.

    Covariation of adolescent physical activity and dietary behaviors over 12 months

    Journal of Adolescent Health

    (2007)
  • J.R. Sirard et al.

    Physical Activity in Adolescents: From Associations to Interventions

    Journal of Adolescent Health

    (2008)
  • L.Y. Wang et al.

    The Association Between Body Mass Index in Adolescence and Obesity in Adulthood

    Journal of Adolescent Health

    (2008)
  • H. Aarts et al.

    Physical exercise habit. On the conceptualization and formation of habitual health behaviours

    Health Education Research

    (1997)
  • I. Ajzen

    Perceived behavioral control, self-efficacy, locus of control, and the theory of planned behavior

    Journal of Applied Social Psychology

    (2002)
  • J. Arbuckle et al.

    AMOS 4 user's reference guide

    (1999)
  • C.J. Armitage et al.

    Efficacy of the theory of planned behaviour. A metaanalytic review

    British Journal of Social Psychology

    (2001)
  • J.A. Bargh

    The four horsemen of automaticity. Awareness, efficiency, intention and control in social cognition

    (1994)
  • K.A. Bollen

    Structural equations with latent variables

    (1989)
  • R.J. Budd

    Response bias and the theory of reasoned action

    Social Cognition

    (1987)
  • J. Cohen

    Statistical power analysis for the behavioral sciences

    (1988)
  • C.J. Crespo et al.

    Television watching, energy intake, and obesity in US children. Results from the third national health and nutrition examination survey, 1988–1994

    Archives of Pediatrics and Adolescent Medicine

    (2001)
  • G.J. de Bruijn et al.

    Does habit strength moderate the intention-behaviour relationship in the Theory of Planned Behaviour? The case of fruit consumption

    Psychology and Health

    (2007)
  • G.J. de Bruijn et al.

    Associations of social-environmental and individual-level factors with adolescent soft drink consumption. Results from the SMILE-study

    Health Education Research

    (2007)
  • Cited by (0)

    View full text