Article type
Abstract
Background:
Efficient energy usage is critical for limiting global warming to below 2°C (climate action) and achieving other Sustainable Development Goals (SDGs), including SDG7 (affordable and clean energy) and SDG3 (healthy lives). The recent energy crisis triggered by the Russian invasion of Ukraine has underscored the need for energy savings and the critical role of behavioral interventions, emphasizing the importance of timely availability of comprehensive evidence for policy and practice.
Objectives:
Use a machine-learning (ML) enhanced systematic review methodology to continuously assess the efficacy of monetary incentives, information provision, social comparison, feedback, and goal setting in reducing household energy demand and associated CO2 emissions;
Introduce the living evidence concept to climate change mitigation, addressing methodological challenges and promoting evidence-based decision-making;
Enable future living evidence synthesis in the Campbell Collaboration by testing the methods and reflecting on specific challenges for applications outside of the medical field
Methods:
We conduct an ML-assisted systematic review and meta-analysis to assess behavioral interventions’ efficacy in reducing residential energy demand. This review involves several methodological firsts: developing a statistical stopping criterion for prioritized title and abstract screening in living evidence applications, resolving some of the statistical challenges in updating network meta-analysis, and setting up guidelines for updating policy recommendations based on a living review.
Results:
Our meta-regression finds that both monetary and nonmonetary interventions reduce household energy consumption, with monetary incentives showing a more pronounced effect. Specific combinations of interventions enhance effectiveness, while others may diminish it. We estimate a small annual carbon emissions reduction potential of these interventions at about 0.35 GtCO2 yr-1, with greater reductions possible through deploying the most effective intervention packages. Continuous updates will refine these findings and deepen understanding of intervention impacts.
Conclusions:
This living systematic review pioneers the application of the living evidence model to climate change mitigation, whereby we introduce it to the Campbell community and start a critical cross-community learning effort. While further methodological innovation is necessary, we demonstrate how statistical stopping criteria can assist in safely stopping dynamic screening efforts in living evidence synthesis.
Efficient energy usage is critical for limiting global warming to below 2°C (climate action) and achieving other Sustainable Development Goals (SDGs), including SDG7 (affordable and clean energy) and SDG3 (healthy lives). The recent energy crisis triggered by the Russian invasion of Ukraine has underscored the need for energy savings and the critical role of behavioral interventions, emphasizing the importance of timely availability of comprehensive evidence for policy and practice.
Objectives:
Use a machine-learning (ML) enhanced systematic review methodology to continuously assess the efficacy of monetary incentives, information provision, social comparison, feedback, and goal setting in reducing household energy demand and associated CO2 emissions;
Introduce the living evidence concept to climate change mitigation, addressing methodological challenges and promoting evidence-based decision-making;
Enable future living evidence synthesis in the Campbell Collaboration by testing the methods and reflecting on specific challenges for applications outside of the medical field
Methods:
We conduct an ML-assisted systematic review and meta-analysis to assess behavioral interventions’ efficacy in reducing residential energy demand. This review involves several methodological firsts: developing a statistical stopping criterion for prioritized title and abstract screening in living evidence applications, resolving some of the statistical challenges in updating network meta-analysis, and setting up guidelines for updating policy recommendations based on a living review.
Results:
Our meta-regression finds that both monetary and nonmonetary interventions reduce household energy consumption, with monetary incentives showing a more pronounced effect. Specific combinations of interventions enhance effectiveness, while others may diminish it. We estimate a small annual carbon emissions reduction potential of these interventions at about 0.35 GtCO2 yr-1, with greater reductions possible through deploying the most effective intervention packages. Continuous updates will refine these findings and deepen understanding of intervention impacts.
Conclusions:
This living systematic review pioneers the application of the living evidence model to climate change mitigation, whereby we introduce it to the Campbell community and start a critical cross-community learning effort. While further methodological innovation is necessary, we demonstrate how statistical stopping criteria can assist in safely stopping dynamic screening efforts in living evidence synthesis.