Crystal Smit

General Discussion 6 135 should drink more water, supporting them by filling their bottle with water, and offering rewards may have caused peers to feel that their classmates expected them to drink more water (i.e., injunctive norm; Cialdini et al., 1991). This may have prompted some of the targeted peers to drink more water and less SSBs in order to make a good impression on others (Vartanian, 2015). Contrary to our findings, other social network interventions have actually found an increase in descriptive norms (Latkin et al., 2013; Mellanby, Newcombe, Rees, & Tripp, 2001; van Woudenberg et al., 2020). A reason for these contradicting findings could be that in these interventions the influence agents mainly focused on modeling and providing information about the desired behavior. These influence mechanisms probably responded to the perceived descriptive norm of the children. Thus, it seems that the way influence agents promote the behavior, as well as how it is interpreted by their peers (Lapinski & Rimal, 2005), has consequences for the effectiveness of social network interventions. It would be relevant for future research to gain more insight into the relationships between perceived social norms and behavior in social network interventions as well as the functioning of the underlying influencing mechanisms (i.e., social modeling, impression management, or social facilitation). LIMITATIONS Self-Reported Data In all studies in this dissertation, beverage consumption was measured by self-report; children were asked to report how much water they drank the day before. Although these types of measurements are generally found to be reliable (Vereecken & Maes, 2003), there is a chance that children under- or overreport their own beverage consumption (Collins et al., 2010; Lally et al., 2011). The studies in this dissertation attempted to reduce this potential problem by assessing the intake of children multiple times during each data collection (i.e., every other day). Nevertheless, future studies can improve the methodology by using additional and more direct measures, such as observations in schools (Beets, Tilley, Weaver,