Preempting Polarization: An Experiment on Formation of Opinions about Net Neutrality

Participants in the study are provided with both partisan and non-partisan information about net neutrality, an issue they are unlikely to be familiar with.

The partisan information consists of texts, images, and tweets revealing the history of net neutrality policy changes in the United States. The materials highlight Democrat's support for net neutrality along with the opposition of Republican politicians to net neutrality.

The non-partisan information is based on two videos -- one discussing pros of net neutrality, and the other focusing on the cons.

We randomize the order in which the partisan information and the non-partisan information are presented. In AFTER condition, the partisan information is presented after the non-partisan information. In BEFORE condition, the partisan information is shown prior to the non-partisan information.

Note that at the point of measuring donations (primary outcome) the overall informational content is the same for both conditions -- only the order differs. However, for the secondary outcome (pro vs. con), the AFTER group participants do not possess partisan information (just the non-partisan information), whereas those in the BEFORE group possess both.

2022-07-07

2022-09-30

Amount donated (out of $0.5 bonus payment) to the Electronic Frontier Foundation, a charity that supports net neutrality

Heterogeneity:
We will look at heterogeneity with respect to whether the participants reported voting for Donald Trump in the 2020 presidential election.

For the heterogeneity analysis, we will use a Prolific prescreening question "Who did you vote for in the 2020 US presidential election?".

The side of the argument (pro vs. con) that the participant chooses when evaluating non-partisan information from two videos

We will perform the same heterogeneity analysis as for the primary outcome.

The experimental design is hidden until the end of the trial.

We recruit participants on Prolific. We restrict the pool of eligible participants to those who answered "Republican" on Prolific's US political affiliation question and those who answered the question about their 2020 vote (regardless of the answer). Participants are required to complete a 15-minute Qualtrics survey about net neutrality.

After collecting demographics and a brief introduction to the issue of net neutrality, we randomize participants into two treatment groups: BEFORE and AFTER.

Participants in the AFTER group will see the PRO and CON videos on net neutrality immediately after we introduce the issue. After both videos play, participants are shown the "PRO vs CON question". At the bottom of the page (so they can re-watch any videos, as needed) they are asked to explain (in at least 20 words) which argument they found most convincing, pro or con. To incentivize engagement, they are told the top 25% of answers (based only on the clarity of their answer, not which side they chose) will receive a bonus payment. On the following page, they indicate whether the most convincing argument was from the PRO video (in favor of net neutrality) or the CON video (against net neutrality).

Participants in the BEFORE group watch the PRO and CON videos after learning that their Republican party opposes the issue of net neutrality and the Democratic party supports the issue. When they watch the video content, they can interpret it with this knowledge from Partisan Information in mind. When they answer the "PRO vs CON question", they have knowledge of both Partisan Information and Non-partisan Information.

In both treatment groups, after participants have received both Partisan and Nonpartisan Information, they are told that they now have a chance to support net neutrality. They are told about the Electronic Frontier Foundation, a charity that supports net neutrality. They are offered a $.50 bonus payment, and they are asked how much of that bonus they would like to donate to the net neutrality charity, EFF.

Both for the primary and the secondary outcome, to test for the impact of the intervention on the outcome, we will use a two-sided t-test for difference in means between the two experimental conditions (BEFORE and AFTER).

Qualtrics randomization

Individual

No

N/A

800 individuals. This includes 200 individuals from the pilot study that we conducted prior to the pre-registration. We decided to pool the pilot with the subsequent data due to the very limited number of eligible Republicans on Prolifc. This should give the study sufficient power to detect effects close to 0.2 s.d.

400 individuals per group