How does motivated reasoning moderate the effect of information on attitudes towards wind farms?

Participants are provided with four pieces of information about wind farms. The information received varies in congeniality and source type (academic sources, newspaper articles, tweets).

2024-09-16

2024-09-27

There are three primary outcomes in this study. First is "Attitude_Post" which is one's attitude towards wind farms after the information treatment has been provided.

The second outcome of interest is whether or not the participant states that they signed a petition which calls on the Secretary of State for Energy Security and Net Zero of the United Kingdom, Edward Miliband, to “allocate more resources to support the transition to renewable energy generation.”

The third outcome of interest is the proportion of potential raffle winnings to “Power for People” which campaigns for the UK to rapidly transition to 100% clean energy and for this to benefit local communities. Participants are told that upon completion of the survey they will be entered into a raffle worth £100. They are asked how much they would like to donate should they win this raffle.

"Attitude_Post" which is determined by agreement with a set of statements about tradeoffs involving wind farms, measured on a 6 point likert scale. The average value of their answers from the likert scale is used.

Perceived reliability of information.

After each information treatment, participants are asked to rate their perceived reliability of the information on a 5 point likert scale.

After providing basic information about wind farms and administering a set of five questions to elicit participants’ pre-existing attitudes about wind farms, respondents were provided with one of fifteen information treatments. These treatments vary in source type (academic journal article, newspaper or tweet). The participants will not be given detailed information about the source type to prevent confounding effects between the tone of the information and the source of the information. However they will be told whether the source is an academic journal article, newspaper or tweet.

The treatments also vary in the portrayal of wind farms' economic impacts, ranging from arguments emphasising their benefits to those highlighting potential drawbacks. This will be referred to as the congeniality of the information, which is a categorical variable that can have one of five values: 4P (indicating four pieces of positive information about wind farms), 3P1N, 2P2N, 1P3N, or 4N (indicating four pieces of negative information about wind farms). This non-binary approach was chosen to account for the potential non-linear effects of the treatments, allowing for the identification of any such non-linearity. Seeing as there are three source types and five levels of congeniality, this creates a 3x5 factorial design.

Each respondent was provided with four pieces of information and as such, they could receive four pieces of “pro-wind farm” information, four pieces of “anti-wind farm” information or anything in between. The source type would be consistent for each respondent throughout. For instance, the respondent would either receive all of their information from academic journal articles, newspapers or tweets. Respondents would be asked questions to ensure that they are absorbing the information, before being asked to rate their perceived reliability of the information.

After respondents received the information treatment, they are asked to rate their agreement with a different set of statements to elicit their post-treatment attitudes. Agreement with each statement is indicated on a seven-point Likert scale. Participant i’s attitude after the treatment, referred to as “Attitude Post” is the average value of their answers across this set of post treatment statements. In the case of a participant indicating that they “don’t know” whether they agree or disagree with a given statement, this answer is not considered when calculating the average value.

We also gave participants the option to sign a petition which called on the Secretary of State for Energy Security and Net Zero of the United Kingdom, Edward Miliband, to “allocate more resources to support the transition to renewable energy generation.” Participants will be asked, upon returning from the petition, whether they have signed the petition. The response will be used as the second dependent variable.

Finally, we informed participants that they had been entered into a raffle for participating in the survey. They were given the option to donate a portion (of their choosing) to the not-for-profit organisation “Power for People” which campaigns for the UK to rapidly transition to 100% clean energy and for this to benefit local communities. Whether the participant donates and the amount that they choose to donate will also be used as dependent variables.

An important aspect of this study is to identify whether there is heterogeneity across factors that affect one’s propensity to engage in motivated reasoning. In order to assess heterogeneity, we construct subgroups using the following variables which were identified as relevant using Hart et al. (2009)’s meta-analysis on selective exposure. These include a measure of closed-mindedness, perceived knowledge of wind farms (Berkowitz, 1965; Brechan, 2002), the strength of one’s pre-existing attitudes and the perceived and actual quality of the source of information. Additionally, actual knowledge is measured and considered a relevant factor for assessing heterogeneity (Sylvester, 2021).

At the beginning of the survey, participants are asked to state which day of the month they were born (1-31) and their answer to this question is used to randomly assign them to one of 15 treatment groups.

Individual

No

N/A

1,629

The study uses a 5x3 factorial design and therefore uses 15 treatment levels. Each treatment should have roughly 108 individuals. However, given the method of randomisation this may vary.