A Warning About Ipsos/Global News Poll From Earlier This Week

The Ipsos / Global News poll released Tuesday is the only one of four polls this week to measure the Conservatives at 40% and the NPD at 20%, recent highs for both parties. I'll write it again and repeat it: watch out for aberrant numbers. They make good clicks, but rarely represent reality.

Nevertheless, polls are statistical sets that contain noise, so it is quite normal for some numbers to come out of the current averages and spreads, especially with the current turbulence in Ottawa.

In the Global News article, the CEO of Ipsos says the CPC has taken a considerable lead in 905 region and could, with these numbers, sweep this crucial region.


Here is a direct quote:

"Crucially, the Tories enjoy a leadership in the vote-rich 905 region surrounding Toronto. "The Conservatives have over 20-point lead in the 905," said Bricker. "With that kind of lead, they're set to sweep the 905. If they sweep the 905, they probably win the election." 

In Ipsos' tables (you can find them here), it is written that the Conservatives are in the lead in Ontario with 40% of decided voters, with only 31% supporting the Liberals.

The unweighted Ontario sample is 288 respondents. With such a small subsample, the margin of error for a probabilistic survey would be about ±5.5%.

The 905 region contains about one-quarter of Ontario's total population (eg, 28 seats out of 121).

So, what would be the Ipsos sample for the 905? Somewhere between 65 and 80 respondents? With 80 respondents, the margin of error explodes to almost ±11%. Even if Ipsos had collected 100 respondents from 905 (and then weighted the results), the margin of error for such a sample would still be ±9.5%.

Admittedly, the margins of error given above are only valid for random samples and therefore do not apply to internet panels. Nevertheless, the results of internet panels still contain uncertainty!

So unless the CEO of Ipsos has figures that he has not published or shared with the public (which is possible), he can not at all declare such statements with such certainty.

But hey, it generates a lot of clicks.

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 info@Qc125.com.

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 info@Qc125.com.