Industry/Academic Research on Facebook Content and Use in Election 2020

Session Submission Type: In-Person Roundtable

Session Description

This roundtable will discuss a pathbreaking collaboration among over two dozen independent (that is, not paid by Facebook) academics and a team of Facebook researchers. Since early 2020, this group has worked together to evaluate the role of Facebook and Instagram in the 2020 U.S. presidential election. The project involved a multi-wave panel survey beginning in August 2020 and continuing through the aftermath of the 2020 general election. Survey participants were selected in two ways: a stratified random sample of Facebook and Instagram users recruited on the platforms and an independent probability sample, stratified based on whether people used these platforms. The survey focused on four key outcome variables: (1) political participation, (2) political polarization, (3) knowledge and misperceptions, and (4) beliefs about democratic norms. For those who gave their consent, the survey data were paired with on-platform data. We also use platform-wide aggregate statistics to understand the content to which users were exposed on Facebook and Instagram, such as exposure to various types of news and political content, misinformation, targeted political ads, and other similar content. A series of different classifiers were used to categorize election-relevant platform content.

The collaboration was structured such that all of the academics and Facebook researchers contributed to the common goods by, for example, helping to design the survey and develop classifiers. Teams of academic and Facebook researchers then worked to design and pre-register analyses that focused on different aspects of the election and platforms. The academics are lead authors on the publications resulting from these collaborations, with control rights over the details; Facebook agreed to no pre-publication approval. The aspiration is a model of industry/academic collaboration that can be replicated in other contexts.

The collaboration has, as of now, resulted in over a dozen pre-registered designs for academic research papers. As such, we are proposing two separate roundtables to different APSA sections which, if accepted, will be our first public presentations of study designs to the academic community. The other roundtable we have proposed will focus on the research designs of the experimental analyses, and this roundtable will focus on observational designs related to analyzing content on the Facebook and Instagram platforms during the 2020 election campaign and its extended aftermath, as well as how platform content relates to people’s political attitudes, beliefs, and behaviors. In addition to sharing research designs, panelists will reflect on the experience of this unique collaboration, including its strengths and limitations.

The panelists include academics with extensive experience in studying political uses of social media and the key dependent variables for this initiative, including Shanto Iyengar, David Lazer, Jennifer Pan, Jaime Settle, and Rebekah Tromble. They will be joined by Facebook researchers with many years of experience in conducting research on the effects of social media on people and society: Chad Kiewiet de Jonge, who was one of the project leads at Facebook, and Devra Moehler, who helped to lead the survey effort and has experience both from within academia and Facebook. From the academic side, Joshua Tucker and Talia Stroud led the project, and will serve as moderators for the session.

Sub Unit

• Division 40: Information Technology, & Politics

Chair

• Joshua A. Tucker New York University

Discussant

• Natalie Jomini Stroud University of Texas, Austin

Presenters

• Devra Coren Moehler Facebook

• Chad Kiewiet de Jonge Facebook

• Rebekah Tromble George Washington University

• Jaime E. Settle College of William & Mary

• Jennifer Pan Stanford University

• David Lazer Northeastern University

• Shanto Iyengar Stanford University