I just wanted to quickly follow up on AL’s earlier post regarding DNI Clapper’s remarks that:
“Of course the Russian efforts affected the outcome. Surprising even themselves, they swung the election to a Trump win. To conclude otherwise stretches logic, common sense, and credulity to the breaking point. Less than eighty thousand votes in three key states swung the election. I have no doubt that more votes than that were influenced by this massive effort by the Russians.”
Was there active collusion between the Trump campaign — or the candidate himself — and Russian proxies or agents? Clapper does not go that far because he doesn’t have proof. But what he calls Trump’s “aggressive indifference” to the intelligence community’s detailed presentation of Russian activities is, in his view, damning enough. “Allegations of collusion and the results of the election were secondary to the profound threat Russia posed — and poses — to our system,” Clapper writes, and he does a fair job explaining why.
The National Bureau of Economic Research (NBER) has just released a working paper entitled: Social Media, Sentiment and Public Opinions: Evidence from #Brexit and #USElection. Here’s a link to a pdf of the report, which I’ll also attach to the bottom of the post. I want to excerpt this bit from the introduction to the paper (emphasis mine):
We find that information about the Brexit and the 2016 U.S. Presidential Election is disseminated and absorbed among Twitter users within 50-70 minutes. This suggests that information rigidity could be very low for critically important issues with wide coverage or that news cycles in social media are short-lived. We also observe the differential impact of tweeting activities by user type. For example, “remain” supporters in the Brexit Referendum respond stronger and faster to messages created by other “remain” supporters when compared with the reaction to messages from “leave” supporters. Furthermore, human tweeting activity could be influenced by bots. The degree of influence depends on whether a bot provides information consistent with the priors of a human. For instance, a bot supporting the “leave” campaign has a stronger impact on a “leave” supporter than a “remain” supporter. Similarly, Trump supporters are more likely to react to messages spread by pro-Trump bots. Further examination shows that the sentiment of tweets plays an important role in how information is spread: a message with positive (negative) sentiment generates another message with the same sentiment. These results provide evidence consistent with the “echo chambers” effect in social media; that is, people tend to select themselves into groups of like-minded people so that their beliefs are reinforced while information from outsiders might be ignored. Therefore, social media platforms like Twitter could enhance ideological segmentation and make information more fragmented rather than more uniform across people. Finally, we provide a quantitative assessment of how bots’ traffic contributed to the actual vote outcomes. Our results suggest that, given narrow margins of victories in each vote, bots’ effect was likely marginal but possibly large enough to affect the outcomes.
And this section from Section E of the findings on p. 20 of the report (emphasis mine):
But again, even this small difference could have played an important role in the outcome of these close-call elections. Specifically, our analysis in Section 2.5 suggests that a percentage point increase in the share of pro-Trump tweets in total tweets is associated with a 0.59 percentage point increase in the share of actual pro-Trump votes. Therefore, the observed difference between actual and counterfactual pro-Trump tweet shares suggests that 3.23 percentage points of the actual vote could be rationalized with the influence of bots.
This is one econometric analysis of the effects of the Russian active measures and cyberwarfare campaign against the US in the 2016 election and the UK during the Brexit referendum. It is an important piece of unclassified, open sourced supporting analysis to DNI Clapper’s conclusions. But the research shows that there is a strong correlation between the Russian active measures and cyberwarfare campaign and shifts in voting in the US and the UK. This research is not conclusive. It does not and cannot put an end to the suspicions or concerns, but it is important as part of the larger explanation of what happened in the 2016 presidential election and the 2016 Brexit referendum.
Open thread!