Carnegie Mellon University

Center for Informed Democracy & Social - cybersecurity (IDeaS)

CMU's center for the study of disinformation, hate speech and extremism online

IDeaS Center for Informed Democracy & Social-cybersecurity

AI Generated Image

April 24, 2024

Comparing How Different News Types Are Shared on Different Platforms

By Christine Lepird

Article published in Social Media + Society, https://journals.sagepub.com/doi/10.1177/20563051241245950 

Paper Citation:

Lepird, C. S., Ng, L. H. X., Wu, A., & Carley, K. M. (2024). What News Is Shared Where and How: A Multi-Platform Analysis of News Shared During the 2022 U.S. Midterm Elections. Social Media + Society, 10(2). https://doi.org/10.1177/20563051241245950

Keywords: multi-platform; pink slime; low credibility news; politics; news sharing; elections

 

For the past 5 years, automated partisan news masquerading as “local” news without local reporters has been infiltrating Americans’ news diets. It’s referred to as Pink Slime, and during the 2020 U.S. Presidential Election, 3.7% of American adults visited at least one Pink Slime site, with 17.7% of these sites coming from Facebook referrals. All research into the spread of Pink Slime during elections has been conducted on data from the 2020 U.S. Presidential Election, so we set out to analyze how this news type (in relation to others) spread during the 2022 U.S. Midterm elections on Facebook, Reddit, and Twitter. After gathering 1.3 million news link posts to these social media platforms pertaining to key elections in swing states, we classified the news links as Real News (high credibility, national scope), Local News (high credibility, local scope), Low Credibility News (low credibility, national scope), or Pink Slime (low credibility, local scope) to answer our research questions.

How do the three social media platforms differ by news type shared?

Each of the four news types was present in each of the three social media platforms, per the table below. This finding runs counter to previous research that found that Pink Slime was not present on Reddit. The previous study authors believed that since Pink Slime sites were not present in the largest political subreddits that they would not be on the site; however, larger subreddits may have more stringent moderation rules in place. Furthermore, the subreddits where we observed the Pink Slime posts were smaller and many had a focus on local communities.

All of the platforms had a majority of their posts sharing Real News. Reddit had almost 90% of its posts going to Real News sites which may be due to its integrated moderation system. Of the three platforms, Twitter had the lowest percentage of posts going to high credibility news (Real News and Local News); however, this may be due to Elon Musk’s acquisition of the platform. Under his new leadership, Twitter aggressively slashed headcount in early November 2022 which may have included vital content moderators (inside this research’s data collection window).

 christine-blog-1.png

Breakdown of News Types Shared on the Three Platforms, as Percentages of the Amount of Each News Type Site as a Total of the Number of Sites Shared Within Each Platform

When looking at users who shared one news type and then another, by platform, we found that the most likely course of action would be for a user to continue to share the same news type as they previously had. However, we see (per the image below) that those sharing Real News on Twitter did not post Pink Slime after having posted Real News. The opposite is true for Reddit.

christine-blog-2.png

Likelihood of sharing one news type based on previous news type shared by platform. Nodes represent news type; links between two nodes represent users shared the two news types in consecutive posts; link thickness represents the likelihood of sharing two news types in sequence. Image generated by Christine Sowa Lepird

How do engagement of the different news types shared differ?

Comparing engagement of the news types proved challenging. On Facebook, the majority of Real News posts were shared to the largest quartile of Facebook Pages. However, Low Credibility News and Pink Slime had 22% and 21% of their posts present in the smallest quartile of these groups, respectively. Comparing absolute engagement by news type would not take into consideration these differences. In order to normalize the engagement, we established a Relative Engagement metric which takes the engagement a post received and divided it by the size of the audience (numbers of followers of a Facebook Page). The result, in the image below, shows Pink Slime with the highest relative engagement and Real News with the lowest, a worrying trend.

christine-blog-3.png

Logarithmic distribution of relative engagement measure by news type for Facebook. Image generated by Christine Sowa Lepird.

How do classes of users differ in their news type sharing behavior?

In order to understand the sharing behavior of users sharing Pink Slime, we made a network of all the users (from all the platforms) who shared Pink Slime and included all of the other news domains they shared. The resulting image, below, paints an interesting picture. While we would expect to see clusters of news based on geographic region (i.e. a cluster sharing news pertaining to the Pennsylvania Senate race), instead we find two clusters split along political lines. The left cluster includes Pink Slime sites owned by left-leaning organizations while the right cluster includes Pink Slime sites owned by right-leaning organizations. This indicates that, despite marketing themselves as local news, these sites are shared along partisan lines.

christine-blog-4.png

News sources shared by users (including all platforms) who shared pink slime domains. Pink Slime sites are labeled and given a pink node coloring, local news sites are green nodes, real news sites are blue nodes, and low credibility news sites are red nodes. Gray nodes represent the users themselves. Links are formed between users and the sites they share. Image generated by Christine Sowa Lepird

When we analyzed the differences in news sharing by human and bot accounts, we found that there was a higher proportion of bot accounts sharing Pink Slime then Low Credibility News; more than twice as many humans shared Pink Slime then Local News compared to bots, which may indicate humans have a harder time differentiating between these news types.

In summary:

  1. The news types lower in credibility (Low Credibility News and Pink Slime News) are shared the most on Twitter, and these news types have higher Relative Engagement on Facebook Pages compared to high credibility news types (Real News and Local News).
  2. Instead of in geographic clusters, Pink Slime News is shared along political leaning, lending credence to the assertion that this news type is politically-driven as opposed to geographically-driven.
  3. Bots share the Lower Credibility News while humans share proportionately more Local News after sharing Pink Slime.