Editorial
Dear Reader, welcome to the first newsletter of the Horizon 2020 bEhaVioral Insights anD Effective eNergy policy acTions (EVIDENT) project. EVIDENT is EU Funded Research and Innovation Project and this second newsletter provides a description of the project’s use cases. We hope you will find the contents of this newsletter interesting, and your comments and suggestions are always appreciated.
EVIDENT use cases
The key components of the EVIDENT project are the experiments and quasi-experiments which will seek to examine the impact of behavioural interventions on energy efficiency. A description of all use cases to be employed in the EVIDENT project are discussed in this newsletter, with reference to experimental design, research question, participants, tools, and any ethical or quality factors.
The goal of use cases 1 and 2 is to provide insights on whether self-consumption and peer-comparison feedback affect customers’ energy efficiency. In addition, EVIDENT will try to estimate the indirect factors that impact consumers’ behaviours and biases by providing answers to the following questions:
- Are there crowding out effects between different message framings?
- Do consumers know about pricing differentials between peak and off-peak consumption?
- Can consumers accurately estimate the energy use of everyday household appliances?
- Does consumption feedback through consumer ranking affect energy efficiency?
We will carry-out three field experiments that will last for one year, using Public Power Corporation’s (PPC) and Checkwatt’s (CW) datasets. Within the first experiment we plan to send information material through e-mail to PPC’s customers to acquire self-consumption feedback (use case 1) and engage with peer comparison feedback (use case 2). Consumers will receive this information once or twice a month for a predetermined period of time. The experiment aims to estimate the impact of both the information content and the message frequency on energy conservation. During the second experiment, we plan to repeat the first experiment for CW customers taking care of the similarities and the differences between the provided data and customers characteristics. The third experiment regards CW’s customers and concerns shifting consumption between peak and off-peak time periods. The objective of this experiment is to estimate whether changing prices impact decision making.
Summary of statistics for Public Power Corporation’s (PPC) 1st experiment dataset
CW’s customers’ distribution in heating area class and heating type electricity
Use case 3 seeks to examine the role of big data in assessing the impact of behavioural insights in energy consumption. Thus, the following questions will be explored:
- Can a data-driven machine learning (ML) analysis on pre-treated data efficiently assist the design of a randomized control trial?
- Can ML data-driven approaches identify heterogeneity in treatment effects that were not specified in the pre-analysis plan?
- Do ML methodologies outperform traditional econometric energy consumption and production forecasting models?
The procedures that will be followed in use case 3 are quite different from the other use cases since no field experiment will be implemented.
In the context of use case 3, two experiments will be designed and implemented. During the first one, we plan to implement a data-driven approach to partition the data into subpopulations that differ in the magnitude of their treatment effects. The pre-analysis plan will be based on data provided by CW and PPC containing historical observations on customers’ energy consumption. The analysis will try to estimate the treatment effect heterogeneity and to test hypotheses about the differences between the effects in different subpopulations. Several ML methodologies will be leveraged to reveal relationships between attributes not easily observable by using traditional approaches. For the second experiment, the EVIDENT consortium plans to leverage ML models to predict and evaluate the energy performance of each household. The data provided by CW and PPC (house construction year, internal environmental conditioning, peak operation hours, size of the house, building insulation etc.) will be supplemented with environmental data (environmental temperature, humidity, daylight aperture, etc.).
Procedure for energy consumption and production prediction for optimal decision-making
This use case seeks to determine the impact of energy related financial literacy, demographic factors, and behavioural intention/attitude, on decisions to repair or replace household appliances across different resident types. Specifically, the following questions are asked:
- What impact do financial literacy, energy literacy, environmental concern or resident type have on the decision whether to repair or replace a broken household appliance?
- What impact does additional financial information have on the decision to repair or replace?
- What type of information impacts willingness to pay for a repair or replacement of an appliance? Specifically, what impact has financial, anticipated lifecycle or environmental framing on willingness to pay for a repair?
- Does providing tips and information related to financial literacy enhance consumers’ ability to make better choices?
For the purposes of this use case, we will develop a serious game, and two individual surveys. One survey before the serious game play through and the other afterwards.
The serious game will consist of a series of decision points in which the user is tasked with choosing whether to repair or replace an appliance and how willing they are to pay for this. When an appliance breaks in the game, three opportunities to choose between repair or replace will be presented with the information provided at each choice differing depending on the condition to which the individual is assigned.
The purpose of the first survey is to determine the participants’ energy related financial literacy, their environmental literacy level and socio-demographic characteristics. A follow-up survey will be sent serious game players six to twelve months after their participation. This survey will be used to determine the generalisation of the serious game effects to actual behaviour in real life circumstances. This survey will consist of a series of questions which will seek to establish whether the participant had an opportunity to repair or replace an appliance, and their understanding and application of financial literacy as it applies to energy efficiency.
The flow of the serious game
The fifth use case of the EVIDENT project seeks to examine the impact of energy related financial literacy levels, demographic factors, environmental literacy levels and willingness to pay for more efficient household appliances. The following questions are asked:
- What impact do financial literacy, energy literacy, environmental concern have on implicit discount rates?
- What impact do factors such as financial information (purchase price, operating cost), risk reduction, energy discounts and loans have on implicit discount rates for home appliances?
- Does providing more financial information impact implicit discount rates and willingness to purchase more efficient home appliances?
- For consumers who choose to purchase more efficient appliances, what impacts do direct rebound rates have when choosing an appliance?
A survey will be developed to examine the impact of a number of specific factors on the decision to purchase energy efficient appliances. This survey will consist of a discrete choice experiment to determine the impact of the salience of factors such as financial information (purchase price, operating cost), risk reduction, energy discounts and loans on implicit discount rates for home appliances. It should be noted that final measures employed within this use case seek to examine the available measures related to the areas of financial literacy, environmental literacy and energy consumption. Recommendations on best measures for each of these areas will also be considered.
An instance of the gameplay of the serious game
About EVIDENT
EVIDENT has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 957117. The information contained in this newsletter reflects only the authors’ view. EC is not responsible for any use that may be made of this information.
EVIDENT website: https://evident-h2020.eu/
EVIDENT partners
This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No