A field experiment using rebates and machine learnings to promote energy-saving behavior.
How to cite
Ishihara, Takunori; Kido, Daido; Sasaki, Shusaku; Ida, Takanori. 2020. "A field experiment using rebates and machine learnings to promote energy-saving behavior.." AEA RCT Registry. AEARCTR-0006139. https://doi.org/10.1257/rct.6139-1.2000000000000002
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
We study the impact of rebate treatments on electricity conservation behavior. Furthermore, by linking household demographic information to electricity data, and using machine learning algorithm we will identify which households should be treated by rebates.
Intervention
Rebate: 100 yen per kWh for the amount of energy saving from the previous month's average energy consumption between 5pm and 9pm in the winter and 1pm and 5pm in the summer.
Experimental design
[Survey 1]
In conjunction with the power company, we will conduct a survey of its customers to obtain household demographics such as gender, age, and income of household members, as well as their use of home appliances. The survey can be completed either online or by mail. At the beginning of the survey, we will ask them for their consent to participate in this research experiment. Only those who have given their consent to participate will be included in the following experiments. In addition, we will pay a reward to the participants of the experiment for their participation in this survey and the following experiments and surveys.
[Experiment]
We randomly divide the participants into three groups. One will be the control group and will not offer a rebate. The other two will be the intervention group. One will offer a compulsory rebate and the other will offer a rebate at the participant's self-selection. Participants have one week to qualify for the rebate. During that period, we give a rebate of 100 yen per kWh for the amount of energy saving from the previous month's average energy consumption between 5pm and 9pm in the winter and 1pm and 5pm in the summer.
[Survey 2]
For those who did not opt for the rebate in the treatment group and those assigned to the control group, we will conduct a survey asking them if they are home or not home on a particular day. We will pay 500 yen to the respondent as a reward for this survey. The purpose of this study is to construct a model for predicting home and absenteeism from electricity consumption using machine learning methods based on electricity consumption data and information on home and absenteeism.
We will conduct Experiment and Survey (2) once every six months. Each participant will participate in the experiment or survey 2 at least four times. We also plan to regularly review the groupings and randomly group them again.
Design details
Cited by
No citation data available for this registration yet.