Crystal Smit

Chapter 2 42 Aronson, 1999; Miller & Wozniak, 2001; Mussweiler & Neumann, 2000) and the self-determination theory (i.e., supporting the individuals’ need for autonomy by providing choices; Deci & Ryan, 1985; Soenens & Vansteenkiste, 2010) in the training of the influence agents. It might be the case that if the influence agents were not asked whether they actually would like to take on this role, that they may have felt that the role of an influence agent was imposed on them. This could have led to them not promoting the desired behaviors, because drinking water is presumably not considered a popular beverage among this age group (Drewnowski et al., 2013). Reactance might have occurred (Brehm, 1966), leading the influence agents to reject water drinking as beneficial. Our findings suggest that the training influenced them positively, given the reported increase in their water consumption. Nevertheless, these explanations remain speculative and future research is necessary to examine whether the training increased influence agents’ intrinsic motivation and level of self-persuasion immediately after the training. However, it is important to note that, overall, the influence agents reacted enthusiastically at the end of the training and expressed their willingness to help make drinking water a trend. Some limitations should be addressed in interpreting the findings of this pilot study. First, although the reported increase in water consumption (and decrease in SSB consumption) was significant, it is important to mention that the found effects were small e with an average increase of water consumption of less than half a serving. Related to this, beverage consumption was assessed by self-report measures only. An additional methodology would have been to use a more direct measurement, such as the use of flow meters attached to the schools’ water fountains to determine the amount of water dispensed from these fountains (Loughridge & Barratt, 2005). Unfortunately, this was not possible in our study. Future studies should seek to replicate our findings using additional means to evaluate beverages consumption. Second, the sample was relatively small, future research is needed to replicate this study in a larger and more diverse sample. This is especially important when it comes to analyzing the data without the influence