The Facebook Effect on Impeachment?

Stefan Hankin
6 min readOct 22, 2019


With the 2020 election only about a year away, Democrats and Republicans are preparing for a fight unlike anything we’ve seen before and it appears that the Republicans are entering from much higher ground. Now more than ever presidential campaigns revolve around the digital fight, and the Trump campaign is not messing around. According to a New York Times article out this weekend, the Trump campaign has spent the last three years building a digital campaign apparatus that combines sophisticated targeting techniques with pathos-driven messaging. The Democrats, on the other hand, are “not even fighting last year’s war,” according to political consultant and chief executive of Tovo Labs, David Goldstein. If the Democratic Party is supposedly the party of change and embraces the modern world we live in, how did we end up in the situation: coming from behind on a platform we once dominated? And more importantly, how can we push ahead as 2020 draws near?

Digital media first played a major role in the 2008 election with the Obama campaign making extensive use of online social media platforms such as Facebook and YouTube. While this feels like the norm now, at the time it was pushing the envelope, and Obama owes a decent chunk of his victory to his campaign’s ability to take advantage of this “new” technology. Eight years later, the Trump campaign and his supporters (both domestic and international) were able to use social media platforms even more effectively. The combination of heavy spending on Facebook ads and message testing, the presence of bots from countries such as Russia, the emergence of right-wing news sources and blogs publishing false news, and the influence of alt-right Internet users spreading false messages created a perfect environment for Republican success.

The decrease in support for Clinton is often attributed to the Comey letter that was released within two weeks of the election but, according to our data, a decline in support for Clinton was happening well before this event. In fact, as shown in the chart below, her support had flat-lined for weeks before the letter, while Trump’s numbers had been growing since early September.

If we look at our overall score of measures (a combination of support and opposition for both candidates coupled with the intensity of these measures), we see that in Wisconsin the race moved from a likely Clinton win in early September to a likely Trump win by November 1st. The Comey letter clearly had an effect, but it just accelerated a pattern that was already in place.

In 2016, Facebook did not release information on what ads were run, where they were run, and to whom they were targeted, so it is tough to pinpoint the effects of the Trump campaign’s efforts versus that of outside interference. That being said, the drop in the likelihood of Clinton winning and the increase in Trump support does occur at the same time that online foreign and alt-right political activity began increasing in earnest (according to most estimations). As more articles with false information and messages that dug into voters’ fears were released, Trump support increased and support for Clinton weakened.

In a recent op-ed in the New York Times, Thomas B. Edsall argues that Trump is already winning the digital campaign and will likely continue to do so unless the DNC improves its data operation and online messaging. Not only do we agree with this conclusion, but we may already be seeing the effects of this advantage when it comes to the defensive strategy by Republicans in the fight over impeachment. Over the two weeks following Nancy Pelosi’s impeachment inquiry announcement, Trendency saw an increase in total allocation for support of impeachment. However, in a pattern similar to the support for Clinton in 2016, support for impeachment has dropped off quite a bit since news broke about Rudy Guiliani’s associates being arrested at Dulles Airport — an event which one would think would increase the support for the inquiry. Instead, we saw a four-day decline in support. Overall, the trend since Trump publicly asked China to investigate Vice President Biden has been a negative one. It is impossible for us to definitively declare that this is because of Facebook and other social media platforms, however, the pattern is very similar to 2016.

As mentioned before, Facebook is not exactly forthcoming with information on ads being run. However, if we are correct that the main source of movement towards the President is because of online efforts, we should see changes in the data that fit the pattern of Facebook use. According to a study at Pew Research, Facebook has the highest concentration of users who are female, older, white, and do not have a college degree.

According to our data, we see that non-college white women are the only group of voters behaving in an unexpected matter. As shown in the chart below, from the early stages of this story (the time when rumors about the Ukrainian phone call had started, through Trump publicly calling on China to investigate the Biden’s), all four groups moved more in favor of the inquiry and ended this period at a higher net than where they started. The difference is that non-college white women were the only group to move in a positive direction from September 30th through October 2nd.

If we extend the data through the middle of the month, we again see throughout earlier last week that three of the four groups follow similar patterns. The one group which does not move as expected is non-college-educated white women. In fact, since Guiliani’s associates were arrested on October 10th, this cohort is the only one that moves in a positive direction. White non-college women are the only group which ended this three week period at a higher level than when we started closely tracking this question.

If we completely smooth out these lines, we see a clear view on the difference between these four cohorts. In fact, the average movement among the other three groups over this period is a 15-point net gain in support for impeachment. Non-college women moved in the opposite direction by five points.

Based on what we are seeing, it feels like a reasonable assumption that the President’s campaign and its supporters are getting an early start on their 2020 playbook on digital media.

The lack of data made available to us makes this a game of assumptions, but the patterns are a little too close for comfort (if you are on the side of opposing Trump) and should certainly be a warning sign for Democrats. This means that not only is Edsall correct in his opinion, but the GOP advantage is already playing out and the Times article from this weekend should not be ignored.

If Democrats and the DNC want to be successful in taking back the White House next year, waiting for a nominee is going to make this all too little too late. The 2020 election is not going to be won or lost on TV, and it won’t be won in the news cycles. It is going to come down to who can control the online narrative, and right now it seems that the Democrats are running from behind leading into the year of the election.



Stefan Hankin

President of Trendency Research and Lincoln Park Strategies Research. The status quo is not a strategy nor a solution.