Spreading the word! Effects of a randomized normative informational campaign on residential water conservation

This paper reports the direct and spillover effects of a norm-based informational campaign on residential water usage. I follow a two-stage randomized saturation design to produce and measure interference within utilities. A percentage of households within utilities is targeted to receive periodic reports comparing their usage to neighbors' average, while the rest are left untreated to measure spillover effects. I find targeted and spillover households reduced their water usage compared to the control group, with average reductions of 8.2 % and 5.6 %, respectively. I also find that the campaign's effects depend on the number of targeted households within a utility, the feedback frequency, baseline water usage, and the proximity to other directly treated households. The findings shed light on the efficacy of leveraging social influence and norms to foster environmentally responsible behaviors.