Having skin in the energy game: The Impact of social norms on Energy Regime Changes

Last registered on April 08, 2020

Pre-Trial

Trial Information

General Information

Title
Having skin in the energy game: The Impact of social norms on Energy Regime Changes
RCT ID
AEARCTR-0005685
Initial registration date
April 06, 2020
Last updated
April 08, 2020, 11:56 AM EDT

Locations

Region

Primary Investigator

Affiliation
Georgia State University

Other Primary Investigator(s)

Additional Trial Information

Status
Completed
Start date
2017-10-15
End date
2020-03-15
Secondary IDs
Abstract
We present a survey study that explores two experimental analysis: first, a cross-country norms effect on petition signing, and second, a cross-state comparison, focusing on clean energy adaptation instead of fossil fuel energy. In the first study, we use energy consumption information from the US, EU, and China. Using the first study, we redesigned the second experiment by providing energy utilization in Arizona and New Mexico to validate our norms. The results reveal that the attendants are more likely to sign a petition—by seven percent—in favor of clean energy act when they have been provided with additional information.
External Link(s)

Registration Citation

Citation
Farhidi, Faraz. 2020. "Having skin in the energy game: The Impact of social norms on Energy Regime Changes." AEA RCT Registry. April 08. https://doi.org/10.1257/rct.5685-1.0
Experimental Details

Interventions

Intervention(s)
Intervention Start Date
2017-10-15
Intervention End Date
2020-03-15

Primary Outcomes

Primary Outcomes (end points)
The results reveal that the attendants are more likely to sign a petition in favor of clean energy act when they have been provided with additional information.
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
Experiment 1: We designed a petition (as a survey study), calling for energy reform, moving away from fossil-based energy to renewables—such as the wind, hydro, solar, nuclear, and thermal energy. This would be funded by either an increase in sales tax by one percent to subsidize fossil-based energy producers to adopt other technology or charging fossil-based energy producers with the carbon tax (ten percent), and then subsidizing the producers who want to generate other energy sources.
The experimental method includes four petitions (one control group and three treatment groups). We have designed four different survey links for each group, which was sent via email. Since the approved target population is twelve thousand faculty, staff, and students at Georgia State University, we randomly assigned three thousand to each group in such a way that six thousand would be randomly selected from the faculty and staff pool, and six thousand from the students’ pool. In the survey, after asking some general information such as gender, income level, native or non-native to the US, and occupation, we provided the information about the effect of the carbon emission on the environment and human lives, including the carbon reduction by switching energy-based fuels.
The survey contains energy usage—based on types—in the US at the national level. More than eighty percent of the US energy consumption is supplied by fossil-based energy, which produces more than 15 billion metric tons of carbon dioxide. The forests required to sequester the produced carbon every year in the US is more than 15 times of the existing forests in the US. Carbon emissions from coal are about 25 times more than solar PV to produce the same amount to generate electricity; and more than double about natural gas; and still around one-fifth of the total energy produces by coal because it is marginally cheaper and available, excluding the environmental damages it causes.
The US uses more fossil fuel than European countries, but less than China . Therefore, by providing energy information about the countries that are utilizing more clean energy compared to the US, we hoped to nudge households to support clean energy adaptation. Thus, in the first treatment, we added the comparison between European countries and the United States as a descriptive norm in which the US uses about 82 percent of her energy from fossil fuel resources. In contrast, this number is around 45 percent for European countries. The second norm is the comparison between the US and China, while China allocates 89 percent of her energy needs from fossil fuel-based resources. Since China uses more fossil-based energy, it might be useful to verify the possible downturn effect of the social norm; in this case, people might think there is another country which is worse when concerning the environment. Participants would ask: “why should we care?” And the third is the comparison between the US, European countries, and China, all together, to verify the impact of the full exposure of the information. At the end of the petition, we asked participants if they are willing to sign the petition or not; if they agreed to sign, then we would follow with another question, asking whether they prefer sales tax increment or carbon tax reform on fossil-based energy producers to cover the costs. While the first treatment would directly affect the household’s costs, the latter increases individual’s living costs indirectly.
The purpose of this design is to test two hypotheses. The first hypothesis is whether providing additional information on country energy consumption—as a descriptive norm—can influence the households’ decision to support a petition in favor of clean energy . The second hypothesis is having the households’ support for subsidizing clean energy. The respondents would choose the carbon tax rather than the sales tax. Revealing this hypothesis may enlighten legislators to proceed to provide the required resources for such subsidy—by imposing a carbon tax on energy producers which indirectly impact the consumers’ consumption prices, instead of increasing a sales tax which directly affects the prices. We speculate that the respondents would choose the carbon tax rather than the sales tax since they comprehend the immediate price effect. We think it would be crucial if the subjects believe that it is not a hypothetical survey and have actual impacts. Therefore, we included a paragraph in the petition that states that we plan to submit the outcome of this petition to Governor Deal. Since there is a high-cost associated with the petition, we thought that less likely people would sign it. This assumption gives me a powerful tool (since subjects realized that engaging this activity comes with the costs) to identify the effectiveness of social behavior.

Experiment 2: The first experiment was performed at the university level—where respondents achieved a higher education compared to the average individuals in the US. A more plausible extension of this work could be a field experiment executed outside a university campus, where subjects’ educational attainment would not be upwardly biased, and therefore, the results would be a better prediction of the society’s aggregate understandings and willingness to participate in environmentally friendly reforms. Another challenge that should be taken into account in a similar study is that a survey must design in such a way that can isolate any respondents’ prior beliefs about the compared countries versus US to unbiasedly determine any treatment effects that arise from providing new information on energy concepts.
Based on the results of the pilot study, we found three major caveats: First, the analysis was under-power; second, there was an educational bias since the surveys were distributed at the university level (among faculty, staff, and students); and third, subjects prior beliefs/judgments about other countries (possibly) made the norms ineffective. In the current experiment, we try to correct for those as follow:
Instead of having three treatment groups (and a control group), we plan to reduce it to two treatment groups. We will collect the data among a general population across different states and regions. We can also provide some incentives for the respondents to participate. Instead of country-level comparison, we intend to perform a state-level comparison. We also want to add a question about their political affiliation to control for possible prior beliefs. We intend to execute the same experiment—at the same time—in two different states, one Democrat and one Republican, but the same region, and with the same geographical boundaries. We selected Arizona (as a red state) and New Mexico (as a blue state). Arizona uses 9 percent of her energy as renewable, and New Mexico utilizes only 8 percent. To control even more about subjects’ judgment about the norm, we plan to compare the energy use from each of the selected states, with two different pioneer states in using renewable, one blue, and one red, with the same percentage usage. We chose South Dakota with 38 percentage and Maine with 36. Hopefully, with the current design, we can rule out the subjects’ preferences and beliefs about the states they are going to compare with. And this time, we are not going to test the possible downturn effect of the social norm, so we would not compare the targeted states with the ones which use less renewable energy. And also, at the end of the petition, we will not ask participants to choose the channel of subsidy (carbon tax versus sales tax); only provide these two mechanisms as possible resources to serve the purpose.
In general, all the surveys contain energy usage—based on types—in the US at the national level. Using the results of the pilot study, we can carry out a relevant power analysis to compute the required sample size for such an experiment with two treatment groups and a control group (three treatments in total). The total sample displays the number of the required respondents in both compared groups. The alpha (α) represents the Type I error, and the power shows the one minus Type II error (1-β) in this statistical test. The table below shows the power analysis of the different comparisons among the three survey groups, based on the pilot study (T1-3 are for the different treatment groups). In this analysis, the percentage of the respondents in the control group who would sign the petition is considered 63%, while the amount is 78% for the first treatment and 56% for the second treatment. Based on our calculation, the total required number of the observations should be equal to 1022 for both T1 & T3 (511 each), and 94 to T2 (the largest number associate with T2 is 188, which includes both T1 and T2, where half goes to T2); therefore, the required subjects for each state would be: 1,116, which makes the total equal to 2,232.

Experimental Design Details
Randomization Method
The first randomization, has done by the university officials itself since they were sending the emails. The second experiment has been randomized by Qualtrics itself since they were collecting the data using their pool.
Randomization Unit
Individual level.
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
Experiment 1: 1
Experiment 2: 2 (states)
Sample size: planned number of observations
Experiment 1: 1,000 faculty, staff, and students Experiment 2: 1,200 residents
Sample size (or number of clusters) by treatment arms
Experiment 1: 250 each treatment
Experiment 2: 400 each treatment
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB

Institutional Review Boards (IRBs)

IRB Name
IRB Approval Date
IRB Approval Number

Post-Trial

Post Trial Information

Study Withdrawal

There are documents in this trial unavailable to the public. Use the button below to request access to this information.

Request Information

Intervention

Is the intervention completed?
Yes
Intervention Completion Date
March 15, 2020, 12:00 +00:00
Data Collection Complete
Yes
Data Collection Completion Date
March 10, 2020, 12:00 +00:00
Final Sample Size: Number of Clusters (Unit of Randomization)
First experiment (used as a pilot): 1 cluster, university
Second experiment: 2 clusters, Arizona and New Mexico states
Was attrition correlated with treatment status?
No
Final Sample Size: Total Number of Observations
First experiment (used as a pilot): 665 individuals (faculty, staff, and students)
Second experiment: 2,460 individuals
Final Sample Size (or Number of Clusters) by Treatment Arms
First experiment (used as a pilot): T1: 147, T2:158, T3:214, T4:146 Second experiment: 410 in each treatments
Data Publication

Data Publication

Is public data available?
No

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials