rennykefs/Impact-of-Social-Nudges-on-Peak-Electricity-usage-in-Kenya

A/B test, Difference-in-Differences panel analysis evaluating behavioral demand-side energy interventions. # Impact-of-Social-Nudges-on-Peak-Electricity-usage-in-Kenya A/B test, Difference-in-Differences panel analysis evaluating behavioral demand-side energy interventions. ## Project Overview This project evaluates whether social comparison and cost-equivalent messaging can reduce peak-hour electricity usage in Kenya. Using a panel dataset of 1,000 households over 6 months, we implement a randomized treatment design and estimate causal effects using Difference-in-Differences and Two-Way Fixed Effects models. ## Research Question Does social comparison + cost messaging reduce household peak electricity usage relative to generic reminders? ## Experimental Design - 1,000 households (500 control, 500 treatment) - 6-month panel data - Months 1–2: Baseline - Months 3–6: Intervention - Outcome metric: Peak Energy Waste Index (PEWI) ## Methodology - Parallel trends validation - Difference-in-Differences estimation - Household fixed effects - Month fixed effects - Robust standard errors ## Key Findings Treatment × Post = -0.0217 (p < 0.001) The intervention reduced peak electricity share by 2.17 percentage points (~5.4% relative decline). Estimated savings: - 3.25 kWh per household per month - Scalable to 0.325 GWh monthly at 100,000 households