A new study finds that Facebook is not censoring conservatives despite their repeated attacks
Right-leaning pages consistently earn more interactions on Facebook than left-leaning pages
A new nine-month Media Matters study of Facebook pages that regularly post about American political news looked at two different measures of performance and found that right-leaning pages outperformed left-leaning pages or performed similarly to them. Right-leaning pages consistently earned more average weekly interactions than left-leaning pages, while both types of pages earned similar engagement rates -- a measure of performance that accounts for interactions as well as numbers of page likes and posts.
Key findings include:
- Right-leaning pages earned more interactions than left-leaning and nonaligned pages. Between January 1 and September 30, right-leaning Facebook pages tallied more than 6 billion interactions (reactions, comments, shares), or 43% of total interactions earned by pages posting about American political news, despite accounting for only 26% of posts.
- Pages without a political alignment drew fewer interactions even though they created the most total posts. Our group of nonaligned pages, many of them from traditional media outlets, posted more than 3.2 million times -- roughly 59% of total posts -- yet accounted for just 30% of total interactions.
- Left-leaning pages were less active than right-leaning or nonaligned pages. These pages posted the least, had the fewest number of page likes, and earned the fewest interactions.
- Right-leaning and left-leaning pages had similar interaction rates, a performance metric that measures engagement of a Facebook page in relation to the number of page likes and how frequently it shares posts. In other words, although right-leaning pages drew more interactions overall, they had similar performance to left-leaning pages when accounting for left-leaning pages' fewer page likes and posts.
Without any evidence, conservatives repeatedly claim that they are being censored
For years, right-wing media and conservative politicians have claimed that social media and tech companies are biased against them and censor their content, despite copious data proving otherwise. In a new study, Media Matters again found no evidence of this alleged censorship. In fact, we found that right-leaning pages consistently earn more interactions than left-leaning or ideologically nonaligned pages.
One recent example of how right-wing media deploy these censorship claims came on October 14, when Twitter and Facebook limited the circulation of a controversial and unverified New York Post story about Hunter Biden (son of Democratic presidential nominee Joe Biden) that was ostensibly reported from hacked materials. President Donald Trump, Republican lawmakers, and right-wing pundits immediately started alleging that social media companies were showing systematic bias against them by limiting their ability to share the story.
The argument isn’t new. For years, conservatives have claimed that Facebook scrubs right-leaning content from its site and that its algorithm directs traffic away from pages with a conservative bent. As Media Matters and others have demonstrated again and again, no such bias exists.
In fact, recent reporting suggests the opposite might be true. Facebook’s algorithm is known to amplify content that evokes an emotional response, reward sensational and partisan content, and contribute to polarization. When Facebook tweaked its algorithm to reduce partisan content in 2017, the platform reportedly throttled traffic away from progressive news sites like Mother Jones while explicitly avoiding taking similar actions against conservative outlets like The Daily Wire and other “right-wing ‘junk sites.’”
New study finds that right-leaning pages are actually outperforming left-leaning and nonaligned pages in terms of average weekly interactions
In this latest study, Media Matters used CrowdTangle data to compile and analyze millions of posts from right-leaning, left-leaning, and ideologically nonaligned Facebook pages that were posted between January 1, 2020, and September 30, 2020. All the pages in this study regularly post about American political news.
The study found that right-leaning pages and left-leaning pages consistently outperformed ideologically nonaligned pages in terms of average weekly interactions. Right-leaning pages consistently earned more average weekly interactions than either left-leaning or ideologically nonaligned pages. These findings are consistent with findings from previous Media Matters studies that analyzed the performance of Facebook pages that post about American political news.
Right-leaning pages earned more total interactions than left-leaning or nonaligned pages, despite accounting for only a quarter of posts
Between January 1, 2020, and September 30, 2020, right-leaning pages earned over 6 billion interactions, which is 43% of all engagement on posts from the Facebook pages in our study. In comparison, left-leaning Facebook pages earned over 3.5 billion interactions, and nonaligned pages earned over 4.2 billion interactions.
Even though right-leaning pages garner consistently high engagement on posts, they do not post nearly as much content as nonaligned pages. Right-leaning pages shared over 1.4 million posts, or almost 26% of all posts from Facebook pages posting about U.S. politics in our study. Nonaligned pages shared over 3.2 million or nearly 60% of all posts.
Left-leaning pages are less active than either right-leaning or nonaligned pages. Posts from left-leaning pages accounted for less than 16% of posts from the pages in our study and earned only 26% of all interactions.
The discrepancy in average weekly interactions — with right-leaning pages higher than left-leaning and nonaligned pages — was consistent over the nine-month study period, with the exception of five weeks in which nonaligned pages narrowly earned more interactions. Nonaligned pages also generally earned more average interactions each week compared to left-leaning pages.
Right-leaning and left-leaning pages consistently had higher interaction rates than nonaligned pages
Interaction rate is a performance metric that measures engagement of a Facebook page in relation to the number of page likes and how frequently it shares posts. Between January 1, 2020, and September 30, 2020, right-leaning pages had an average interaction rate of 0.64%. Left-leaning pages had a slightly higher rate of 0.75%, but this higher interaction rate was not maintained over the full nine-month period. A likely explanation for this higher rate is the smaller overall size and activity level of left-leaning pages relative to the dominance of their right-leaning counterparts.
These interaction rates for ideologically aligned pages are three times higher than the average interaction rate (0.21%) of nonaligned pages over the same time period. The disparity between average interaction rates for partisan and nonpartisan pages was consistent throughout the studied time period, with right-leaning and left-leaning pages maintaining similar weekly average interaction rates while nonaligned pages had much lower rates.
Using CrowdTangle, Media Matters compiled a list of 1,773 Facebook pages that frequently posted about U.S. politics from January 1, 2020, to August 25, 2020.
For an explanation of how we compiled pages and identified them as right-leaning, left-leaning, or ideologically nonaligned, see the methodology here.
The resulting list consisted of 771 right-leaning pages, 497 ideologically nonaligned pages, and 505 left-leaning pages.
Using CrowdTangle, Media Matters compiled weekly leaderboard data for the largest 400 pages for each ideology for a nine-month period that started on January 1, 2020, and ended on September 30, 2020.
We reviewed the data for these 1,200 pages, including total interactions (reactions, comments, and shares), total posts, pages likes, and interaction rate. CrowdTangle calculates the interaction rate of a group of pages by dividing the total number of interactions (reactions, comments, and shares) earned on all posts from the pages by the total number of posts from the pages, and then dividing by average page likes for the pages. For this calculation, average page likes is based on the page likes for each page at the end of the time frame.