Early snap polls on last night’s presidential debate suggest a narrow victory for Clinton. Here is what we can learn from Twitter, via analysis by the Data Face, of tweets about Clinton and Trump during the debate.
The Data Face sampled almost 270,000 tweets and scored them in terms of their tone from -1 (most unfavorable to a candidate) to +1 (most favorable). This graph plots the average for each candidate over time:
In general, Twitter users were more favorable to Clinton than Trump — in line with the snap polls. This was particularly true when she addressed the tape of Trump’s lewd comments about women and in the half-hour after the debate.
Trump did best when he addressed the question about the video but worse when he explicitly disavowed the position his running mate, Mike Pence, took on Syria.
This contrasts a bit with the first debate, when Clinton’s advantage on Twitter was larger and more durable. This in turn was mirrored in the post-debate polls, in which many more Americans said she was the winner. Here is a similar graph from tweets during the first debate.
Of course, Twitter users are not a random sample of the American public. The ultimate verdict on the debate will be influenced by any consensus among political reporters and observers. My forecast for what the conventional wisdom will be:
Forecast for debate CW: Trump started bad, finished better. Not enough to help him. Why the CW actually matters: https://t.co/9sttEUdUrv
— Good Authority (formerly TMC) (@goodauth) October 10, 2016
