A Proposal for Understanding Social Media’s Impact on Elections: Rigorous, Peer-Reviewed Scientific Research

By Talia Stroud (UT Austin), Joshua A. Tucker (New York University), Annie Franco (Facebook) and Chad P. Kiewiet de Jonge (Facebook)

In the aftermath of the 2016 US Presidential election, the country was rocked by one troubling revelation after another about social media and new threats to democracy: the prevalence of false news, the activity of Russian trolls, and the Cambridge Analytica scandal. At the center of the world of social media stood Facebook, and speculation abounded about Facebook’s impact on the election and the quality of democracy in general.

Academic research on how social media generally, and Facebook specifically, affect democracy and elections has proliferated since 2016. Two critical problems have prevented as much progress as we would have liked. First, increasing public concern and legal obligations related to data privacy led social media companies to restrict access to data previously used by external researchers.

The second problem was that it is hard to conduct a rigorous scientific study on social media’s impact after the fact. With the US 2020 election upon us, it is clear we need a new approach. For this election, we are integrating Facebook researchers’ intimate knowledge and access to internal Facebook data with the expertise of independent scholars to conduct scientifically rigorous research that allows these academics to ensure that the research asks the right questions and that the questions are answered with the right methods.

With these goals in mind, Facebook today announced a new project that is bringing together a team of approximately two-dozen Facebook researchers and outside scholars to study Facebook’s and Instagram’s impact on four key outcomes that have dominated public and academic attention: political participation, political polarization, knowledge and misperceptions, and trust in US democratic institutions.

We are four of the team leaders of this process — two Facebook researchers and two outside academics — and we are participating in this project because we believe it is absolutely crucial that there be rigorous scientific assessments of Facebook’s and Instagram’s roles that are available to the public, journalists, and policy makers alike.

Such a large scale research partnership with Facebook is a new undertaking, and one that we hope may serve as a model for others in the future. To guide the effort in a way that ensures ethical practices, transparency, and scientific rigor, we established the following ground rules:

• Facebook will not have any right of pre-publication approval, and will only be entitled to check that papers do not violate legal or privacy obligations.
• Facebook did not select the academic team members. The current academic team has been independently formed by those of us who Chair the North America (Stroud) and Electoral Integrity (Tucker) advisory committees of Social Science One.
• All papers proposed by the academic teams will have academics as lead authors, with final say over the text of the papers.
• Although Facebook is covering the costs associated with running the study itself (e.g. paying the survey vendor), none of the academic team will be financially compensated for their participation in the project.
• To address privacy obligations, only Facebook employees will be able to “touch” the raw data, but both teams will work together to devise appropriate monitoring systems for assuring the scientific integrity of the research.
• Transparency in the research process is paramount, and includes:

- A rapporteur with access to all researchers who will document the research process
- Pre-registration of study plans before data collection that will be made public simultaneously with the publication of results
- Explicit informed consent of study participants whose individual-level data will be collected and analyzed  
- Listing all researchers who have contributed to the project as co-authors on papers, with clear delineation of their contribution to that paper
- An attempt to publish all pre-registered study designs in peer reviewed journals; those that cannot be published will be posted in public scholarly archives
- Processes to provide journals publishing the papers with the ability to replicate the analyses in the papers 
- Publication of all papers accepted by peer reviewed journals in an Open Access format, which means they will be freely available to the public
- As much of the data generated from the project as is possible -- given privacy limitations -- made available to scholars for further study in the future.

Given the crucial importance of understanding Facebook’s and Instagram’s influence on the democratic process in our country, we have committed ourselves to providing the public and policy makers with scientifically sound answers to critical questions about these platforms’ impact. We know everyone will be anxious to see the results. As we will need time to analyze the data, we anticipate that findings will be ready to be shared in the summer of 2021 at the earliest. We hope that the effort will be judged as worthwhile once the studies, designed to make sense of what has been a remarkable transformation of the political process in the digital information age, are complete.