Wednesday, July 14, 2010

Let's talk about polls...

CBC reports that Environics has a three-point spread between the Conservatives and the Liberals.

Libloggers go wild.

Not to burst the bubble, but Environics is kind of the best pollster for the Liberals. Maybe we shouldn't read too much into this? I suspect our support is hovering around where Ipsos-Reid puts us - 28-30ish percent.

But, I digress. The CBC story is much more interesting than the Environics (and Leger) poll mentions. Here we have mention of a clear study done on the effects of polling methodologies on party support. To sum it up:

The results of that poll were based on live telephone interviews and the question focused on choosing a party's candidate.

But a recent EKOS poll, which suggested the Conservatives have a 10-point lead over the Liberals, used interactive voice recognition technology (IVR), meaning respondents punched their preferences into a keypad, rather than speak to an operator.

The poll question also differed in that it asked party preference: If a federal election were held tomorrow, which party would you vote for?

A Leger poll suggested the Tories had an 11-point lead over the Liberals. That poll was conducted over the internet and included the names of the party leaders in the voter intention questions alongside their party names.

In other words: focusing on the party + candidate generates more Liberal enthusiasm; IVR can lead to funny results; party + leader is probably not a good combination for us, at the moment anyways.

No comments:

Post a Comment