A large-scale population survey experiment on preferences, beliefs, and motivated reasoning regarding climate policy instruments

How to cite

Jarke-Neuert, Johannes; Frondel, Manuel; Perino, Grischa; Flörchinger, Daniela. 2022. "A large-scale population survey experiment on preferences, beliefs, and motivated reasoning regarding climate policy instruments." AEA RCT Registry. AEARCTR-0009305. https://doi.org/10.1257/rct.9305-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Johannes Jarke-Neuert* , Manuel Frondel , Grischa Perino , Daniela Flörchinger

* Primary investigator

Registry ID
AEARCTR-0009305
DOI
https://doi.org/10.1257/rct.9305-1.0
First registered
Last updated
Status
In development
Trial dates
2022-07-18 – 2022-08-15
Countries
Germany (National)
JEL codes
C93, D02, D83, D91, H41, Q54, Q58
Registry record
http://www.socialscienceregistry.org/trials/9305

Abstract

There is empirical evidence in the economics, political science, social psychology, and neuroscience literature that humans tend towards motivated reasoning. This implies that they discount new information—or scientific evidence—that runs counter to their prior beliefs and attitudes. This experiment aims at providing causal evidence for the existence and the extent of motivated reasoning in the context of climate action. Specifically, two questions are addressed: (1) How do beliefs on effectiveness of climate policy instruments affect individual preferences? (2) Is there evidence for motivated reasoning?
We compare one abstract mitigation option, the reduction of the cap in the EU ETS, with a concrete and intuitive option, the reductions of emissions from a coal-fired power plant in Germany, and with a linear combination of the two. The conjecture is that motivated reasoning is more likely for options people are emotionally attached to, or that are part of their identity. Such emotional links are more likely to exist for the coal phase-out than for the EU ETS.
Each subject makes two sequential choices between the alternatives, whereas the informational condition is manipulated within-subjects. The details of this manipulation are changed across four between-subjects conditions. Three main hypotheses are tested with this design: (1) Provision of information on the effectiveness of mitigation options induces participants to adjust their choices in the second decision in line with the information received. (2) Under baseline, subjects that state to have participated in protests relating to phasing-out coal or extracting coal (prior to making any decision on mitigation options) respond less to information on the relative ineffectiveness of directly reducing emissions by coal-fired power plants. (3) Under a "name and shame" treatment, provision of information on climate effectiveness is less likely vs. basline to induce participants to adjust choices in line with the information provided. Seven further auxiliary hypotheses are also tested.

Intervention

We compare an abstract mitigation option, the reduction of the cap in the EU ETS, with a concrete and intuitive option, the reductions of emissions from a coal-fired power plant in Germany, and with a linear combination of the two. The participants in our experiment are free to choose between the following four options:
• A: Reducing the number of allowances in the EU ETS by 10 tons of CO2.
• B: Reducing the emissions from a coal-fired power plant in Germany by 10 tons.
• C: Mixed option with 5 tons each via A and B .
• D: No climate action.
Ten tons of CO2 are roughly equivalent to the carbon footprint of an average German citizen. Participants are informed that decisions are real in that the choices of randomly selected participants will be implemented with the help of the operator of a coal-fired power plant and an NGO that retires EU ETS allowances. Furthermore, we elicit participants’ beliefs about the effectiveness of all options in reducing total GHG emission in the EU.
Each subject makes two sequential choices between alternatives A, B, C and D, whereas the informational condition is manipulated within-subjects between first decision (d=1) and the second decision (d=2). The details of this manipulation are changed across four between-subjects conditions. In the baseline condition BSL (code z=0), subjects choose in d=1 between alternatives A, B, C and D without further information on the true effect of each alternative on total CO2 emissions given, whereas information on the true effect of each alternative on total CO2 emissions under the current rules of the EU ETS are given in d=2. In the Regulatory Uncertainty condition RU (code z=3), information on the true effect of each alternative on total CO2 emissions under the current rules of the EU ETS are given in d=1, and information on the true effect of each alternative on total CO2 emissions under the proposed reform rules of the EU ETS are given in d=2.
The remaining two conditions are identical to BSL (z=0) except a different framing of the decisions, implemented by an additional sentence in the instructions, respectively. In the Environmental Impact Frame condition FREI (code z=1) the sentence stresses that the EU ETS is an instrument restricting markets (as opposed to calling it a “market-based” instrument) and provides the government a strong control on total emissions. In the Name and Shame Frame condition FRNS (code z=2) the sentence emphasizes the importance of coal combustion for CO2 emissions and states how emissions from coal-fired power plants have increased in 2021.
See the supporting document for further details.

Experimental design

See supporting document.

Design details

Randomization method
Randomization done at survey start by a computer (pseudo-random number generator).
Randomization unit
Individual-subject-level randomization.
Clusters
1,800 individuals subjects.
Observations
1,800 individuals subjects.
Arms
The target sample sizes are 600 subjects in BSL (z = 0), and 400 subjects in FREI (code z = 1), FRNS (code z = 2) and RU (code z = 3), respectively.
Primary outcomes
A four-point categorial variable that codes the individual choices d=1 and d=2 for each subject: The participants in our experiment are free to choose between the following four options: A: Reducing the number of allowances in the EU ETS by 10 tons of CO2. B: Reducing the emissions from a coal-red power plant in Germany by 10 tons. C: Mixed option with 5 tons each via A and B . D: No climate action. See supporting document for further details.
Secondary outcomes
Participants’ beliefs about the effectiveness of all options in reducing total GHG emission in the EU.
Sponsors
Name: Deutsche Forschungsgemeinschaft (DFG) Location: Bonn, Germany Url: https://www.dfg.de/en/index.jsp Name: Bundesministerium für Wirtschaft und Klimaschutz Location: Berlin, Germany Url: https://www.bmwi.de/Navigation/EN/Home/home.html
Partners
Name: forsa Gesellschaft für Sozialforschung und statistische Analysen Partner_type: private_company Url: https://www.forsa.de/1/ Name: ForTomorrow gGmbH Partner_type: private_company Url: https://www.fortomorrow.eu/ Name: STEAG GmbH Partner_type: private_company Url: https://www.steag.com/en/

Cited by

No citation data available for this registration yet.

Keywords