Drew Margolin, a professor of communication at Cornell University who studies human dynamics through social media, isn’t just tracking how the electorate is reacting to candidates in one single moment via Twitter, but how they have been reacting since the beginning of the primaries. This historical data gives Drew and his collaborator Yu-Ru Lin, University of Pittsburgh, a unique window into changing sentiment within a party about candidates and topics, as well as across party affiliations.
For example, how are early Trump or Hillary supporters reacting differently from more general Republicans and Democrats who originally favored another candidate? What about those who may have been #NeverTrump or #NeverHillary? And what does it mean for Election 2016?
Margolin will be dissecting the data from these randomly selected computational focus groups during both National Conventions, the Presidential Debates, and the National Election.
Insights from the night first and second nights of the RNC convention in Cleveland, the make up of their computational focus groups, and general thoughts on using Twitter as a sentiment barometer can be found here. Coverage of Ted Cruz’s non-endorsement can be found here, and analysis of Donald Trump’s acceptance speech here.
Margolin analyzed the first three nights of the Democratic National Convention in Philadelphia, including enthusiasm for Bernie Sanders on day one, the effect of Bill Clinton’s speech on day two, and hints of a unifying party on night three.
Analysis of the fourth and final night, including Clinton’s acceptance speech below …
Democratic National Convention: Final Night
The conclusion of the Democratic National Convention on Thursday was Hillary Clinton’s night, and her task was clear. On Wednesday her colleagues generated unity and enthusiasm for the party. On Thursday she had to draw that enthusiasm to her candidacy. According to our data, this is exactly what she did.
Hillary Delivers
Across the whole evening, Hillary was able to attract tweets with relatively positive sentiment from 5 out of 6 groups. As I mentioned in my previous post, this consistency across groups has only been matched by Joe Biden and Ivanka Trump.
Perhaps more importantly, the dynamics of this enthusiasm could be directly tied to her speech. Prior to her speech (which, like with Trump and Ivanka, I count beginning when Chelsea took the stage at 10:02 p.m. ET), the party was fairly unified — 5 out of 6 groups tweeted negatively about Trump, and the anti-Hillary hashtags were virtually nonexistent, receiving less attention than they had from each group the night before.
This is much the way Biden, Kaine, and Obama had left things last night. But prior to her speech tonight, Hillary was only drawing relatively positive sentiment from the Defectors to Hillary, her most loyal group throughout the convention, and was in fact drawing significant negative attention from the Hillary Dumpers, who were circulating a “we hate hillary” campaign.
From the time her speech began, however, this trend changed. Hillary received relatively positive tweets from 5 out of 6 groups during and after her speech. Among these were the previously recalcitrant Hillary Dumpers. In fact, oddly, the only group that did not show a significant tendency to use more positive language when talking about her was the loyal Defectors to Hillary (they also sent a small number of tweets during the speech, perhaps because they were focusing on her instead of their second screen). This effect could not be attributed to Chelsea who, unlike Ivanka last week, drew relatively positive tweets from only half of our groups. In particular, the groups more loyal to Bernie (the Hillary Dumpers and the Bernie Holdouts) did not mention Chelsea very much and did not show significant sentiments toward her one way or the other. The change in sentiment thus appears to be a response to Hillary herself.
How does this compare with effect of Trump’s speech last Thursday night? As I mentioned last week, night 4 of the RNC looked very much like night 1 for our groups. All groups were negative on Hillary, and positive sentiment toward Trump came mainly from the Defectors to Trump, including during his speech.
The DNC shows a different pattern. While unity against Trump was a constant each night, Hillary was able to draw positive tweets from only 1 group on night 1, and only 2 groups on night 2. This is quite a contrast to her 5 out of 6 performance during her speech on Thursday.
Similarly, #NeverTrump remained at around 2% for Trump Dumpers, and about 1% for Trump Avoiders most of the RNC week, even during Trump’s speech. By contrast, #BernieOrBust and #NeverHillary began at a combined 8% on Monday for the Hillary Dumpers but was reduced to less than 0.2% during Hillary’s speech.
Both candidates received substantial attention during their speeches, being mentioned in 25%+ of group members tweets. This is quite substantial considering that we include all tweets sent by group members, not just those to a hashtag about politics. But analysis of the texts of the tweets indicates that the content of their speeches did not receive equivalent attention. Specifically, many tweets mentioned key lines from Hillary’s speech. Most prominent was her “a man you can bait with a tweet is not a man you can trust with nuclear weapons.” Portions of this phrase appeared in about 1 out of 70 tweets by each of 4 groups, including the Bernie Holdouts and the Hillary Defectors (1 in 90). Her statement that I “believe in science” also received similar attention. By contrast, Trump’s top phrases — “build the wall” (1 in 145) and “make America great again” (1 in 165) — received less attention from even his most loyal supporters (Defectors to Trump). His most attended to phrase from the Trump Dumpers was “I alone can” (1 in 130) which, as I described last week, was viewed negatively.
Consistent with Hillary’s “better together” theme, an inspection of the most used phrases indicates that another speech resonated with our Twitter groups. Khizr Khan, the father of a Muslim soldier who was killed in the line of duty, challenged Donald Trump by asking him “have you ever even read the Constitution?” This phrase was cited in between 1 in 25 (Defectors to Hillary) to 1 in 120 (Bernie Holdouts) tweets across all groups. This rate of citation is very high in absolute terms — rivaling Clinton’s rate of citation and outpacing the citations of Trump’s speech — and even more impressive considering that the speech was not in prime time nor featuring a known public figure. Khan also drew mentions in about 3-4% of tweets sent before Hillary’s speech, a large number for a speech not in prime time.
What Does It Mean?
The pattern observed in these data is consistent with what one would expect from a “strong” convention. According to the patterns of tweets sent by those who initially favored Hillary and those who initially resisted her, the fractures in the party healed over the past week and the Democrats are now unified behind Hillary.
As I wrote last week, these patterns were absent from the Republican convention, leading me to estimate that Trump would not get a large bounce, at least among Republicans. While polls are still being taken, the consensus is that Trump got a moderate bounce of 4-6 points. Thus, the data here would suggest that Clinton would get a bounce larger than that. However, these analyses are designed to provide some insight into how people react to events while they happen. How these reactions translate into longer term feelings about who people plan to vote for is not yet well understood.
See you at the debates!
- Drew
For media interviews contact:
Kathleen Corcoran
Office: 202-434-8036
Cell: 607-882-3782
kmc327 [at] cornell.edu
