Monday, 6 May 2019

Let's Compare the 338 Projection Averages with the Latest Polls

The 338 federal electoral projection was published yesterday in Maclean's and it shows the Conservatives with a six-and-a-half point lead over the Liberals in the popular vote, but a substantial lead in the seat projection (averages of 174 for the CPC and 111 for the LPC). See all the numbers and full map of the projection here.

Several comments I received last night were suggesting the latest Angus Reid poll, which measured the CPC with a 13 point lead, may have skewed the projection too much in favour of the Conservatives.

It is a legitimate concern that one poll (possibly an outlier) could skew the 338 numbers and therefore wrongly influence the week's narrative, but one that rarely applies here because the model actually detects outliers when it sees one. A poll that strays too far from the current averages will see its weight automatically reduced in the calculations. However, if said poll is subsequently confirmed by another polling firm, it regains the weight it was supposed to have in the first place.

And not one, but two polling firms measured the Conservatives with a 13 point lead in recent days: the aformentioned Angus Reid Institute poll and the latest Leger/Canadian Press poll (Léger had the CPC at 40% nationally).

In fact, let's compare the five latest polls which were considered for yesterday's 338 update with the projection averages:

You may notice that all dots but three lie within the confidence intervals (the exceptions would be Angus Reid LPC at 25%, Leger NDP at 12% and Ipsos NDP at 19%).

I believe this graph above clearly shows the 338 numbers are close to the current averages. I hope this answers the concerns that were directed at me last night.

Philippe J. Fournier is the creator of Qc125 and 338Canada. He teaches physics and astronomy at Cégep de Saint-Laurent in Montreal. For information or media request, please write to

Philippe J. Fournier est le créateur de Qc125 et 338Canada. Il est professeur de physique et d'astronomie au Cégep de Saint-Laurent à Montréal. Pour toute information ou pour une demande d'entrevue médiatique, écrivez à