Behavioral, Information, and Monetary Interventions to Reduce Energy Consumption in Households: A Living Systematic Review and Network Meta‐Analysis

Description

Methods

Data Structure

The synthesized data is structured around three primary dimensions: Interventions, Study Characteristics, and Outcome Measures.

• Interventions: The review classifies policy instruments into five core categories:
1. Monetary: Pricing, rebates, and financial rewards.
2. Information: Energy audits, tips, and awareness campaigns.
3. Feedback: Smart meters, home displays, and redesigned utility bills.
4. Social Comparison: Benchmarking household use against peers (e.g., Home Energy Reports).
5. Motivation: Goal setting, commitments, and gamification.
• Methodological Framework: The data includes 663 effect sizes harmonized to Cohen’s d. We use a multilevel random-effects model to account for dependencies within studies and an NMA to evaluate the relative efficacy of intervention "packages".
• Risk of Bias (ROB): Each study is coded for quality using a modified framework from the Center for Environmental Evidence (CEE), covering internal and external validity.

What can I use this data for?

This dataset is designed for researchers, climate policy analysts, and practitioners to:

• Evaluate Policy Efficacy: Compare the average effectiveness of different intervention types (e.g., Monetary vs. Behavioral).
• Design Intervention Packages: Identify which combinations of interventions (e.g., combining information with social comparison) yield the highest energy savings.
• Assess Evidence Quality: Filter findings based on the Risk of Bias to ensure policy recommendations are built on robust evidence.
• Track Living Evidence: As a "living" review, this dataset serves as a pilot for continually updated climate policy evidence.

Series information

Summary Figures

The record includes key visualizations derived from the Network Meta-Analysis:

• Rankograms: Visualizing the probability of an intervention being the most effective.
• Network Geometry Plots: Showing the relationships and evidence loops between different intervention categories.
• Forest Plots: Displaying average effect sizes (Cohen's d) across different subgroups and ROB levels.

Code Availability

The analysis was performed using the metafor package in R.

• Repository: https://github.com/tarunMkhanna/LSR-HouseholdEnergy
• Main Scripts: Includes code for the fixed-effects model, NMA implementation, and machine learning screening logs.

Changelog

• v1.2 (Current): Update includes literature published through January 2025 and January 2026, adding roughly --- new documents to the initial pool and ....(more).
• v1.1: Update includes literature published through (month) 2020?? and December 2024, adding roughly 53,000 new documents to the initial pool and nearly doubling the number of effect sizes used for synthesis.
• v1.0: Initial meta-analysis based on Khanna et al. (2021) methods, covering 122 studies.