Just like television shaped the Nixon-Kennedy race, the social media platform shows that new media matters in politics. Communication professor Drew Margolin studies human dynamics through social media and has been tracking how people react to presidential candidates on Twitter since the beginning of the election cycle. Margolin says this will be the campaign where we learn that Twitter matters.

How did this research start?

Four years ago, we thought it would be fun to see how users on Twitter would react to the presidential debates. Initially, we focused on the reactions of people “loyal” to one side or the other. As this election’s primary campaign kicked off a year ago, there was a lot of division within the parties – in particular in the Republican Party where you had 16 or 17 different candidates running. So, we thought it would be interesting to look not only at reactions of Twitter users between the parties, but within the parties as well. We started tracking people who were loyal to just one of the candidates and tracked over time whether they switched their loyalties or stayed with them.

What’s different about Twitter?

Twitter provides a few advantages. First of all, it records people’s immediate reaction in real time but without the unnatural settings of focus groups. These are people in their natural environment, on the record for the things that they say to their friends and followers. Their reactions are also time-stamped, and we can measure the audience and resonance over time. It is almost a natural experiment, especially during the debates when nobody knows what will be said. There is a lot of bias in how people recall things after the fact and after listening to pundits, but via Twitter we can look at language as it is created and dispersed through the world in a fresh form.

What are some of the through lines in this campaign?

This has been a really negative campaign; it’s been more about jeering than cheering. Early on, at the conventions both candidates had trouble generating positive attention even from among their own followers. During the debates, we’ve found that whoever is being talked about more on Twitter is generally generating more negative sentiment. Twitter has reflected the severe polarization of the country: Republican groups spew invectives at Clinton; Democratic groups spew invectives at Trump. No one has gotten much positive attention. The other through line is that such negativity is not exactly equivalent between the candidates: Trump is really an outlier. On numerous evenings he has drawn negative attention from across the board.

What will be remembered about these debates?

This will be the campaign where we learn that Twitter matters, in the way that television mattered for the Nixon-Kennedy debates.

NOTE: More data and analysis from the beginning of the campaign season can be found at http://cornellcals.tumblr.com/tagged/Election-2016-Tweets. To monitor how our groups are responding throughout the campaign, check out debatemeter.com.