Crystal Smit

Chapter 6 140 water fountains in schools (Muckelbauer et al., 2009), as a possible way to improve the impact on children’s water consumption. This combined approach is largely consistent with the Capability, Opportunity, Motivation, Behavior (COM-B) model (Michie, Van Stralen, & West, 2011). This model posits behavior as the result of an interaction between having the physical and social opportunities, motivation, and the capability to perform the behavior (Atkins & Michie, 2013). A next step for future research could be to apply the entire COM-B model to improve healthy drinking behaviors in children, thus not only focusing on physical (i.e., installing water fountains) and social opportunities (i.e., peer influences) and motivation (i.e., increasing intrinsic motivation), but also on the capability of children (i.e., increasing knowledge of drinking water; Atkins & Michie, 2013). The Offline and Online Social Networks of Children The findings of this dissertation can serve as a basis for translation to other social networks of children. Aside from the classroom networks, the Share H 2 O intervention approach can also be applied in, for example, children’s sports, music, scouting, theater, and other clubs. Researchers could then investigate how the behavior spreads across these different social networks by, for example, tracking the interactions between children via a Bluetooth signal on their smartphone devices (van Woudenberg et al., 2020) and measuring the behavior of the person they have spent time with. It can also be fruitful to target children’s online social networks. Today’s children spend a lot of time using social media platforms (Chassiakos et al., 2016; Valkenburg & Piotrowski, 2017) and are often permanently connected to their online social networks through these platforms (Boyd, 2014). However, we recommend social network intervention developers to not translate the entire Share H 2 O intervention into an online environment. A recent social network intervention in which influence agents were trained online instead of face-to-face (as in the current dissertation) showed no effect on the behavior of the targeted adolescents (van Woudenberg et al., 2018). The researchers argued that this less personal approach resulted in lower involvement of the influence agents. It is, therefore, advisable to maintain the training of the influence agents face-to-face.