Using Facebook Messenger versus Groups for News Engagement

Abstract

For online news organizations trying to improve audience engagement strategies, Facebook Groups and Messenger chats constitute promising avenues. We explore whether these meso news-spaces, with different discourse architectures and group sizes, affect the substance of the discussion and people’s impressions. In this study, we experimentally tested how training and intimate forms of news engagement in small-group discussions on Facebook Messenger compare with larger conversations in a Facebook Group. The study draws from a real-world experiment in collaboration with Vox Media and its popular Facebook Group for the political podcast The Weeds. Results show that participants perceive the Messenger group as more civil and respectful and report being less prone to self-censor in the Messenger group. Comments in the Messenger group, however, are less relevant to The Weeds podcast and participation in the Messenger group leads people to have less favorable views of the large Facebook Group. The ways in which discourse architecture and group size affect digital discussions provide theoretical and practical insights.

Keywords:

• Audience engagement
• discourse architecture
• news comments
• experiment
• Facebook messenger
• Facebook groups

Acknowledgements

The authors thank Gabrielle Chavez and Natalie Deller for coding messages. We are particularly indebted to Lauren Katz at Vox Media, and to Eve Pearlman, Adriana Garcia, and Erica Anderson at Spaceship Media for their collaboration on this project. We would also like to thank The Weeds community for their willingness to participate in the research. The Center for Media Engagement at the Moody College of Communication at the University of Texas at Austin is grateful for funding that supported this research from the News Integrity Initiative at the Craig Newmark Graduate School of Journalism at CUNY, the Democracy Fund, and the John S. and James L. Knight Foundation.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

1 See Popular Science, https://www.popsci.com/science/article/2013-09/why-were-shutting-our-comments/, and The Courier Journal, https://www.courier-journal.com/story/opinion/2021/05/13/why-were-leaving-comments-on-courier-journal-com/5066101001/.

2 In theory, it is possible that these trainings also could have affected the outcomes. In practice, however, the evidence is thin. We asked respondents on the post-wave survey to evaluate the trainings; the average response was ambivalent (M = 4.44, SD = 1.09, Range = 1 to 7 with larger values indicating more positive assessments). We also ran correlations between people’s impressions of the training and the four outcome variables; only one was significant (social capital in the small group chat, r = .29, p = .04).

3 To minimize differences across groups, we ran 50 different formations of the six group chats. We ran ANOVAs and crosstabulations to examine pre-existing differences among the group formations. We chose the formation with the lowest number of differences in our key variables and major demographics across group chats.

4 In the week of April 8, 2020, some members of one of the group chats did not receive an email from the Vox moderator telling them the episode to which they should listen. They still received the prompt in the group chat and exchanged messages about the topic at hand (coronavirus’ impact on healthcare).

5 Different from Ferrucci et al.’s (2020) study, all six factors loaded on the same factor. We used principal components analysis with varimax rotation; all factors loaded on the first factor, which accounted for 61.93% of the variance in the pre-wave and 61.28% of the variance in the post-wave.

6 Some members of the group chat and commenters on the Facebook posts in the large group sent or posted several messages in a row. In order to see if this style of sharing had an outsized impact on our analysis, we computed our results with these messages separated and with these messages condensed into one. The results of the content analysis hold in every case except for how often a message was a reply to someone else’s message or comment.