Media Leading Polls by the Nose?

By Melanee Thomas on Sep 24, 2008

Today, I went to a talk given by Dr. Stuart Soroka entitled, "I know what's going to happen five days from now." Soroka, one of the founding members of the Media Observatory at the McGill Institute for the Study of Canada and his research team investigate how news media affects poll results. He argues that there's about a four to six day lag between when a story breaks in the news and when it shows up in public opinion polls. Based on data from the current election campaign, Soroka's willing to bet that both the Liberals' and Conservatives' polling numbers go up over the weekend.

Of course, the five days he can't predict are at the end of the campaign, where it appears as though all bets are off. 

One of the main elements the team focuses on is the tone of media coverage (whether parties and leaders are covered positively or negatively). They started the week before the writ was dropped and continue to track coverage for each day of the campaign. The Conservatives and the Liberals seem garner roughly the same amount of news coverage (26% of all articles) save for the first week where the Conservatives were the first party mentioned in nearly 35% of election articles. 

The NDP appears to be the party with the momentum, steadily increasing the number of stories where they are the first party mentioned, and increasing the number of stories where Layton is the first leader mentioned. Dion is on a bit of a first-mention upswing as well, while Harper is trailing off a bit. 

News articles, unlike opinion or editorial pieces, are rarely cast in a negative or positive tone. As a result, the positive-negative treatment of the parties and leaders don't differ much from each other except for a few important exceptions. The Bloc's net tone yo-yos (though the newspapers examined are primarily English-language), and the Greens are getting glowing coverage. 

Of the leaders, Harper's coverage has been getting more positive as the campaign progresses while Dion's coverage yo-yos. Last week was quite positive for Layton, and Duceppe's been on a downward slide as his coverage becomes more and more negative. May too appears to be slipping: her coverage, at first positive, is slipping into neutral territory. 

I can't help but wonder, though, how much of poll results rise from a self-fulfilling prophesy: if they're (not) covered, they're (ir)relevant. If a party is covered as slipping in the polls, the slide will likely continue, and vice versa. If a leader is positively covered, their numbers will go up. There's a chicken-and-egg problem that needs to be resolved here that isn't quite addressed by this data.

That said, it sure is interesting.  

Syndicate content