The current study assessed associations of and interactions between friend and parent influences with schoolchildren’s EBRBs across Europe. Results showed that both friend and parental norm and modelling were significantly associated with children’s EBRBs. In addition, we found evidence that friend and parental factors interact. In general this study showed that friend influences were stronger when no family rules were in place. Furthermore, the combination of unfavourable friend and unfavourable parent influences was associated with high soft drink consumption, while the combination of favourable friend and parent influences was associated with a higher probability for daily breakfast consumption.
Previous studies and reviews already indicated that social influences are important for EBRBs in youth [9, 44, 45], and that these social influences can have different forms and sources [21, 34, 46]. However, only a few studies looked at parental and friend factors simultaneously [25, 30–34]. In contrast to findings of Kirby et al. , in our sample parental influences remained significantly associated with EBRBs in a multivariate model that included the friend influences. Although the study by Kirby et al. involved the same age group as the current study, it is more often found that in primary schoolchildren parents do influence physical activity levels of their children. A review by Edwardson and Gorely  found that parents played an important role in children aged 9–11 years especially by being active role models. This association was less clear for adolescents aged 12–18 years , also in line with the conclusion of the review by Uijtdewilligen and colleagues who found insufficient evidence for the prospective association between parental PA and adolescent’s PA . The study sample of the ENERGY-project is in the transition to adolescence, but it is still in the age group where parents are expected to have a regulatory role which may explain the significant association of family rules, parental norms and modelling with their children’s EBRBs, even when friend influences are taken into account.
In addition, the current study suggests that parental and friend influences interact and may strengthen or weaken each other’s effect. Salvy et al. suggested that parents may have an inhibitory influence on unhealthy eating habits . In our study we also found that favourable parental factors (i.e. rules in place, parents thinking that drinking soft drink is bad and low parental modelling) tempered the association of unfavourable friend norms and high friend modelling with high soft drink intake. It may be that having family rules in place reflects a more restrictive and controlling parenting style such as an authoritative or authoritarian parenting style, and that in families with authoritative or authoritarian parenting styles others outside the family system have less influence.
Furthermore, also for daily breakfast consumption, we found that if children reported high parental modelling, they were very likely to consume breakfast daily, even if they perceived unfavourable friend norms and friend modelling. That parental norms and modelling were stronger associated with daily breakfast consumption than friend norms and modelling seems plausible, as children usually do not have breakfast together with their friends. In addition, we found that friend norms and modelling were associated with sport participation and TV time. However, significant interaction between parental and friend influences was found in 10 out of the 20 potential interactive associations, indicating that associations of friend norms and modelling with EBRBs might depend on the parental norms and modelling. The current study included a selection of potential behavioural correlates and the strength of associations and interactions may depend on what variables are studied. In the current study the clearest pattern was found for the interaction between family rules and friend modelling. This may be explained by the fact that having family rules in place was a clearer concept for the participating children than parental modelling or parental norms, resulting in more valid answers (as shown in ) and thus a more straightforward pattern.Nevertheless, in general our results indicate that in this age group parental as well as friend influences are important for engagement in specific EBRBs and that both joint effects and moderating effects are important (Figure 1) depending on the specific parent and friend factors and behaviour.
Social influences on human behaviour and health have been well established and are an important topic for further study in social psychology and health promotion research [34, 49]. Leading theories of determinants of health behaviour such as Social Cognitive Theory , the Theory of Planned Behaviour  and social-ecological models of health behaviour [51, 52] all presume that social influences such as norms and/or modelling are important determinants of health behaviours. However, interactions between such influences from different ‘sources’, i.e. in the present study parents and friends, has not been studied before. If the present results are supported in further research, such interactions should be taken into account in theory and models of determinants of health behaviour for youth.
The major strength of the current study is its sample size. Furthermore, this is one of the first studies looking at interactions between parent and friend influences on four different EBRBs. A limitation of the study is its cross sectional design which restrains drawing conclusion on the direction and causality of the relationships. As previously suggested  it may very well be that parents react on their children’s behaviours by setting rules and initiating favourable example behaviour and communicating favourable norms and beliefs. Furthermore, children may seek friends that are similar to them, also with respect to EBRBs and beliefs regarding the EBRBs . Finally, the current study used self-reported data, which may suffer from recall bias and social desirability. However, previously assessed reliability of the items was mostly good – excellent. In addition, validity of the EBRB items was mostly good to excellent, except for the items assessing norms and modelling. The latter is most likely due to the general difficulty to assess construct validity of cognitions, especially in this age group, and the lack of a gold standard.
Taking these limitations into account, the results suggest that both parental as well as friends norms and modelling are associated with engagement in EBRBs among 10–12 year-olds, and that parental norms, modelling and rules may also moderate friends’ influences. Therefore, parental as well as friend norm setting and modelling should be considered when planning health promotion interventions. These results confirm the importance of involving parents in school-based interventions . More insight is still needed in how parental and friend involvement can be improved, how modelling and norm setting occurs and how this can be used to promote a healthy energy balance.