Beyond The Commons

Aaron Wherry covers all the goings-on in and around Parliament Hill. Follow Aaron on Twitter: @aaronwherry

Categories: Beyond The Commons, Capital Read

Get out the vote

by Aaron Wherry on Friday, July 16, 2010 11:30am - 0 Comments

Eric at 308 takes a shot at adjusting polls for voter turnout trends.

First, I looked at the last week of polling in the three last elections. In 2008, the average result of that last week was 34.2% for the Conservatives, 26.6% for the Liberals, 19.5% for the NDP, 9.6% for the Bloc Québécois, and 9.6% for the Greens. The actual election results were 37.7%, 26.3%, 18.2%, 10.0%, and 6.8%, respectively.

In other words, on election day the Tories had 110% of their average polling result, compared to 104% for the Bloc, 99% for the Liberals, 93% for the NDP, and 71% for the Greens.

Tags: election, poll, polling, Polls, voter turnout

More by Aaron Wherry

Cats

OUCH!

Warren Kinsella says the census was intrusive and he refused to do it !!

Way to step all over the narrative you've been building for the last week eh lefty partisan Wherry ?

Cats caught a mouse!

http://intensedebate.com/people/WDM WDM

I'm strongly considering changing my user name to Dog.

http://intensedebate.com/people/David0M David_M.

Dogs have an exellent attutude with regards to Cats.

Michael

I am no statistician, but does this imply that the average poll leans to the left?

Michael

I made my first comment before I read the actual article. 308 seems to be more interested in the actual time when a voter makes up his/her mind and the actual statistical process and information used. 308 readily admits that the data is far from ideal. As much as I would like to fear monger about the left spin machine, I don't think it is justified in this case.

http://intensedebate.com/people/Stewart_Smith Stewart_Smith

Not so fast, it is after all fairly well known that 308 is a fetid hive of Liberal toadies and hacks using their ill-gotten statistical evidence to assist the Taliban, promote gay pride parades and bust mass murderers out of federal prisons.

Marion

No, it means that the Tories are better are getting the people who plan to vote for them to actually cast their ballots.

http://intensedebate.com/people/sourstud sourstud

And that the left is as lazy as the stereotype portrays them to be.

Bob

Or the polls were just wrong.
http://twitter.com/AlisonLoat @AlisonLoat

Does it not mean that the Conservatives are better at getting out the vote than the other parties?

http://intensedebate.com/people/AJR79 AJR79

Only in 2008. In 2004 and 2006 they underpreformed the polling due to "hidden agenda" and fear of a CPC majority.

I don't think the "hidden agenda" meme will be very effective next election, so I believe the way this was adjusted is pretty useless.
http://intensedebate.com/people/Stewart_Smith Stewart_Smith

I think with all parties, there is the issue of getting out the vote, i.e. which party has the more motivated, effective volunteer base.

There is also the issue of the vote coming out. I suspect that when someone perceives their side is doing well at the wire, they are more likely to vote. (I know the other argument can be made, but it is inconsistent with my observations of human behaviour.)

With the Greens and to a lesser extent the NDP, there is the plug your nose factor. If you expressed intent to vote for them as a protest, at the last minute there can be a tendency to make a pragmatic shift especially if you perceive it will make a difference in your riding.

http://intensedebate.com/people/tobyornotoby tobyornotoby

I'm not sure these last week polls show voters changing or maknig up their minds so much as measuring which ones actually follow through on their choice. A major weaknesss in electoral polls is that respondents lie about their actual intention to vote, because they don't want to appear irresponsible (even over the phone with a stranger).

I saw one the other day in which 8% reported they wouldn't bother to vote, when we know that it will be more like 35-40% of eligible voters who don't show on Election Day. The Green Party seems to be the major "beneficiary" of the phantom voter responses.

http://intensedebate.com/people/LynnTO LynnTO

It's also that people who don't vote are also less likely to respond to telephone or online surveys.

http://intensedebate.com/people/AJR79 AJR79

Nice shout out Aaron.

That being said, I'm not a big fan of the way that was done. There are some very well reasoned arguements against it, in the comments under the post.
http://intensedebate.com/people/Stewart_Smith Stewart_Smith

The correction being explored is interesting, but I would be more interested in a routine measure of the volatility of the electorate. There is always those that are expressly undecided (although the pollsters always want to hide that number like it was dirty laundry.), but I suspect a much more robust number should be possible.

http://intensedebate.com/people/LynnTO LynnTO

It's not hidden because it's dirty laundry, it's hidden because undecided voters can't be counted on to show up to the polls.

And, time after time, decided vote tends to render closer to the actual distribution of votes.

http://intensedebate.com/people/Stewart_Smith Stewart_Smith

Understood, but it would still be interesting to distinguish between broad but soft support that has occured in the past and the highly polarized situation today. As an amateur, a followup question such as "Are you likely to change your mind between now and an election?" might catch that (when routinely collected over a period of time). If nothing else it would provide the junkies will a little more juice.

http://intensedebate.com/people/LynnTO LynnTO

I agree, I think those concepts and questions would be great to see in a media poll. Though, as I'm sure you're already aware, Canadians don't see most of the polls that are conducted in Canada, and the ones that are published? They may not see the full slate of results. There's lots that could be being asked that we don't hear about.

But man, would I love to sink my teeth into more deliberative questions like the ones you mention…

hosertohoosier

I don't think there is evidence of a consistent last-week poll bounce (the last week turned against Harper in 2004 and 2006). I do think there is stronger evidence that pre-election polls underestimate Reform/Alliance/Conservative support (but not because of a get out the vote effect, since the bounce occurs shortly after the election starts).

In 1993 Reform outperformed pre-election polls by 9-10%. In 1997, Reform did roughly the same. In 2000, the Alliance was polling at 17% at its inception, and about 21% when the election was declared. They won 26% of the vote. In 2004, the newly merged Tories were at the mid-20s even post-adscam. They won 29%. In 2006 polls in November and the cooler phase of the election in December had the Tories well behind the Liberals (people kept saying "what is the point of an election, if it will just bring the same results?"). Harper won 36% of the vote. Additionally, if you look at polls in August 2008, Harper was only ahead of Dion by about 2%. In fact Dion led in 3/7 polls. One other way to gauge this is to look at election scares. There too, Harper has consistently gained any time there has been a threat of an election.

I hypothesize that this is related to a different factor. There are some Canadians that see themselves as being "not very political", who decide who to vote for a month before an election. These voters are more likely to support the Conservatives, and less likely to register in polls outside of election season because they don't follow political developments very closely. These folks are precisely the kind of people that the politics of the Reform party (and to a lesser degree, the Tories) appealed to.