Who won the Presidential Debates and what do Canadians think? (Updated)

There have been three US Presidential debates. There have been women in binders, horses and bayonets and some rather confusing comments about foreign policy.

The big question is, who won? Some polls say Romney, some say Obama but let’s see what the social space said.

We looked at the data from September 1 to today (53 days) and it looks like Obama took the prize but barely, scraping by with 51 per cent of share of voice. Search terms included “win,” “won,” “debate” and the candidates’ names.

When we break it down, you can see that in some social spaces such as blog and forums, both men tied. Obama still had the small lead in News and on Twitter.

In the Popularity Comparison over the last 53 days, we can see the spikes  – October 4 and 18 – the days after the first two debates and October 22, the day of the third debate. Once again, the race was tight, but you can see that Romney won the first debate, Obama took the second debate and the third debate is too close to call.

When we look at just Twitter, the winner is more obvious:

It’s clear that on Twitter, Obama won the second and third debates. That’s a good sign for Obama supporters. Now we just wait and see what happens next month. While we wait, why not check out this video on the Electoral College:


Mark Blevis took a look at what Canadians thought about the debate.

Canadian political enthusiasts love their US political theatre. The digitally-engaged among us have actively contributed to the online chatter surrounding the three US Presidential debates. Our digits have been going digital, sharing quips, thoughts and quotes while glued to the tube.

How busy have we been?

We’ve issued 420,498 debate-related tweets, collectively, on the dates of the three debates. And that traffic offers some very interesting insight.

The first discovery that jumped out at me is Canadians had more to say about the second debate (153,070 tweets) than either of the other two (136,614 and 128,814 respectively). It’s worth noting (and I’m relying on my memory here) that I tracked roughly 114,000 election-related tweets on May 2, 2011, the date of Canada’s last general vote.

Note the swell on October 11 coincides with the debate between Joe Biden and Paul Ryan.

This differs from the collective international Twitter traffic which favoured the first debate (11.2 million tweets) over the second (10 million) and third (7 million).

Last night, a majority of participants came and went in a single tweet (81%) — roughly 17% issued 2-4 tweets and 2% issued 5-7 tweets. The most engaged (those who issued 8 or more tweets) were statistically insignificant.

Canadians mentioned Mitt Romney more often than Barack Obama on all three debate days.


Romney’s popularity in Canadian tweets might have a lot to do with him providing some of the most tweetable moments. The most popular theme for Canadians during the first debate was the Governor’s plan to cut funding to PBS (evidenced in the popularity of tweets mentioning Big Bird and Sesame Street). In the second debate it was his “binders full of women.” The most tweeted theme in the third debate was Obama’s lessons to Mitt Romney on the modern military (evidenced in terms such as horses, bayonets and fewer).

Both candidates had a roughly 60/40 split of tweets issued by men and women respectively. If there’s enough interest, I can prepare follow-up analysis which compares popular themes based on gender.

On average, Ontario issued the greatest number of debate-related tweets (58%) followed by BC (12%), Québec (11%) and Alberta (9%). The remaining provinces and territories combined for 10% of Canadian tweets about the debate.

A cursory (emphasis on cursory) analysis of sentiment revealed that, overall, Canadians favoured Barack Obama. The math is a bit confusing at first glance so let me explain it this way. On average, 6% of all mentions of President Obama had a positive tone and slightly fewer than 11% had a negative tone. Contrast that with Mitt Romney whose averages were just under 5% positive and nearly 13% negative. I need to emphasize that a deeper dive into context, language and tone would be necessary to arrive at more credible/meaningful sentiment analysis. I just don’t think I’ll have the time to review over 400,000 tweets anytime soon.


Analysis performed using Marketwire/Sysomos MAP.

Leave a Reply