On A Roll: Two Election Calls in a Row

Elections are always an anxious time for public opinion researchers.  During the campaign we easily get wrapped up in all the data in attempts to explain for the public what is happening and why.  But on election night, things get serious.  It’s the one time we’re held to account for our numbers.

The provincial election in Ontario was only the second election that Abacus Data did regular polling and released a final poll on the eve of the campaign.  We did some polling during the New Brunswick provincial election but did not survey at the end of the campaign.

In May, we did quite well at forecasting the federal election and last night, our record for accuracy continued.  The numbers tell the story:

Our performance and those of other pollsters last night supports the following:

1. Online research is a legitimate and accurate method of measuring public opinion and vote intentions.  When done right, it is as or more accurate than traditional telephone surveys.

2. Our non-traditional vote intention question uncovered the potential swings in this campaign and allowed us to anticipate the late swing to the Liberals that our final poll measured.  Those “torn” voters who ended up voting swung heavily to the Liberals opting for stability and the status quo over change.

3. Most public opinion researchers did well last night and that is good news for the industry as a whole.  While there will always be some polls that are not as accurate as others, the aggregate performance of the industry last night demonstrates that we can accurately measure public opinion using a range of methodologies even in challenge electoral environments.

I’d like to thank our media partner, Sun News, for their support throughout the campaign and those who gave us an opportunity to explain our methodologies and try new things.

The whole team at Abacus Data will continue to provide a different perspective on public affairs and consumer behaviour.  We will continue to challenge orthodoxies and we’ll never stop trying new and innovative ways of measuring public opinion and behaviour.

Don’t just take it from us, here’s what the Globe and Mail wrote about the performance of pollsters in the Ontario election