Objective: Multiple studies have identified clusters of co-occurring healthrelated behaviours. Little is known, however, about factors associated with such clusters. This study aims to identify these factors and to assess whether their effects are in accordance with the Theory of Triadic Influence (TTI). Design: A cross-sectional study using a representative sample (N = 3497) of the Dutch population aged 19-40. Main Outcome Measures: Our data concerned 18 health-related behaviours combined in three clusters (Health, Alcohol and Delinquency) and 30 non-behaviour-specific (i.e. ultimate or distal) cultural, social and intrapersonal factors. The three clusters were used as outcomes in regression analyses. Results: Descriptive Norms of Friends and Gender were associated with all three behaviour clusters. Furthermore, Having Parents who Smoke or Consume Alcohol was associated with, respectively, the Health and Alcohol clusters. Self-Control and past Parental Monitoring were associated with the Health and Delinquency clusters. Effect sizes were moderate to large (r 2 : 0.05 to 0.22). Conclusion: Factors with a moderate to large association with several behaviour clusters were identified. These factors were located within the social and intrapersonal stream of the TTI, not within the cultural stream.Keywords: multiple risk behaviour; theory of triadic influence; self-control; delinquency; descriptive norm; parental monitoring; determinant; cluster
IntroductionDuring the past decade, health-related behaviours have frequently been shown to co-occur (e.g. Conry et al., 2011;Faeh, Viswanathan, Chiolero, Warren, & Bovet, 2006;Lippke, Nigg, & Maddock, 2012;Poortinga, 2007;Schuit, Van Loon, Tijhuis, & Ocke, 2002;Van Nieuwenhuijzen et al., 2009;Wiefferink, et al., 2006 investigated co-occurrence of health behaviours in several ways. Some studies looked at joint prevalences, i.e. the prevalence of combinations of risk behaviours (e.g. Faeh et al., 2006;Poortinga, 2007;Schuit et al., 2002). Some studies identified clusters of persons who show the same behaviours (e.g. Conry et al., 2011). Other studies identified clusters (also called factors) of behaviours closely associated with each other. The latter type of clusters will be used in this study. The following clusters were found: a cluster of healthy eating and physical activity (Lippke et al., 2012), a cluster of smoking and alcohol abuse (Wiefferink et al., 2006), three separate clusters of, respectively, health-promoting behaviours, alcohol/sexual behaviour, and delinquent behaviours (Van Nieuwenhuijzen et al., 2009).The existence of health behaviour clusters has triggered some researchers to search for factors associated with these clusters (e.g. Brooks, Magnussen, Spencer, & Morgan, 2012;Prochaska, 2008). If such factors exist, then there would be support for the development of more integrated approaches to promoting healthier lifestyles. In the words of Prochaska (2008) 'Multiple Health Behaviour Research represents the future of preventive medicine' (p. 281). Some stud...