There has been lots of noise about ideologically motivated censorship of politicians on Facebook this year. We looked at the data to determine the facts of the matter.
At the beginning of 2018, there’s been a good deal of controversy about unfairly censored voices from politicians. It came up in Mark Zuckerberg’s Senate testimony back in April, and was raised again by Nigel Farage when he spoke to the European Parliament this week.
We wanted to see if we could clear up the confusion, and whether this is a genuine phenomenon, by looking at the Facebook Pages of the most senior politicians in the U.S. and the U.K.
Given the most recent controversy began with Nigel Farage’s questioning in the European Parliament, let’s begin in the U.K.
Farage vs the rest
For the purposes of this study we’re looking at average engagements, so that the amount posted doesn’t affect the rankings. We’ve also started this in August to exclude all the content from around the general election, which skews the data somewhat.
We looked at Farage alongside Jeremy Corbyn and Theresa May, the respective leaders of the Labour and Conservative parties, alongside the three main parties in the U.K.
Since August, Farage has had higher average engagements per post than either of the two main party leaders, and higher than the official Facebook Page of the three main parties, though the latter is the case for the parties as well. The politicians tend to get more engagement than their parties.
So that’s the how many, but what about the when? The claim has been that this has all got worse since the algorithm change, which came into official effect on January 11th of this year, but that does not necessarily seem to be the case.
As we can see, there has been a general pattern for all three of these politicians, with a general trend downwards since August before picking back up again. Corbyn’s worst month came in November before consistently rising through April, while May hit a low point in February before improving towards the end of our analysis.
Farage was more erratic in the engagements he received, but the average engagements to his page in April 2018 were the second highest of any of the months that we looked at, aside from December which was skewed somewhat by a particularly viral Christmas post.
The same general downward skew is also visible when we look at the three parties in isolation. One thing to note here is how much Labour and the COnservatives exceed the performance of the Lib Dems. Here, the Labour Party started with higher average engagements in August, but the gap appears to have closed over the past nine months, with the Conservatives actually having a higher average engagement in April.
For the US, as there was no defining event in the past twelve months, we looked at the full year, from May 1st 2017 to April 30th 2018.
We looked at a list of active U.S. politicians from both side of the aisle, which contained Donald Trump, Bernie Sanders, Elizabeth Warren, Kamala Harris, Ted Cruz, Rand Paul, Paul Ryan, Nancy Pelosi, Marco Rubio, Chuck Schumer, and Mitch McConnell. We did not include Hillary Clinton in the analysis as she is not a currently active politician.
Donald Trump is clearly ahead of the rest in terms of average engagements, at well over 2.5x what his closest rival Bernie Sanders achieves. That is to be expected though, given he has roughly 3x more followers than the Vermont Senator.
Beyond those two Elizabeth Warren is the clear standout, with Kamala Harris also performing better than average. These are all politicians that are known to resonate better with younger people, and tend to post relatively provocatively on social, which perhaps explains their overperformance here. According to a recent Morning Consult poll, Bernie Sanders is the most popular Senator in the country.
The other politicians have largely similar levels of engagement, with party alignment seemingly unimportant to the numbers. Republican politicians Ted Cruz, Paul Ryan, and Rand Paul all outperformed Democratic Congressional minority leaders Chuck Schumer and Nancy Pelosi in terms of average engagement for the year, for example.
Sanders, Trump, and Warren
We just looked at Sanders’ relative success alongside Elizabeth Warren, but if we look at their performance over time it gives a very different picture, especially when looked at alongside President Trump’s figures.
There has been a clear downward trend in average engagements per post since May of last year, with Bernie Sanders’ average engagements by month nearly quartering in size, and Elizabeth Warren’s dropping by a factor of eight or nine since May of last year. This has been a lengthy drop too, not just something that had occurred since the algorithm change in January.
Something similar also appeared to have been happening to President Trump since October, but some recovery appears to have occured in the last month. It remains to be seen if that is something that continues going forward.
In terms of the two parties in the US, the Democrats have had a slightly higher average engagement for most of the months in the last year, but the gap appears to have been narrowing since August, to the point that in three of the last four months there has been almost no difference in average engagements.
In conclusion, there does not appear to have been a one-sided censoring of political content. Average engagements have declined generally for politicians across the board, especially for US politicians with high name recognition such as Bernie Sanders and Elizabeth Warren. Even Donald Trump has seen a decline in his engagements, though that trend may be beginning to reverse.
For full disclosure, we don’t have full insight into how Facebook’s algorithm surfaces content, but this is what our data has reflected.
See the Facebook posts that are driving the most engagements in real time, with NewsWhip Spike.
Benedict Nicholson is the Managing Editor at NewsWhip. An Englishman in New York, he is interested in the intersection of PR, brands, and journalism, and the trends and innovation around that.
Email Benedict via firstname.lastname@example.org.