Crystal Smit

Social Network Intervention vs. Mass Media Intervention 4 81 water consumption, there were no statistically significant differences between the three conditions. Moderating Effects of Norms on Water Consumption Descriptive Norms The second structural path model examined the potential moderating role of descriptive norms on the effectiveness of the social network intervention. This model showed a good fit to the observed data, RMSEA = .02, CFI = .99 and normed c 2 = 1.16. At T2, the main effect of descriptive norms emerged as statistically significant, but this effect was qualified by a significant interaction between descriptive norms and the difference among the social network intervention and control condition on water consumption, adjusting for water consumption at T1 ( b = -.38, SE = .16, β = -.12, p = .028, 95% CI [0.23, -0.01]; see Table 4.4). To interpret this interaction, we conducted simple slope analysis. Figure 4.3a presents the significant interaction, with water consumption at T2 (adjusted for T1) on the y -axis, conditions on the x -axis and separate regression lines for participants with high (+1 SD ) and low (-1 SD ) descriptive norms. This figure indicates that there was a positive relation between conditions and water consumption at T2 (adjusted for T1) for high descriptive norms ( b = -.60, SE = .42, p = .154) and a negative relation for low descriptive norms ( b = .25, SE = .29, p = .381), but neither slope significantly differed from zero. Thus, there is some evidence to suggest that children reporting higher descriptive norms consumed more water in the social network intervention and less water in the control condition compared to those with lower norms. While the simple slopes differed in valence, this interpretation is made cautiously considering the lack of statistically significant simple slopes.