The two Quebec voting intent polls were released on March 21 by political commentator Jonathan Trudeau on a morning show hosted by Paul Arcand at Montreal’s Cogeco station.
They have surprised many observers both in their similarities and differences, but above all because of what they suggest.
First, the elephant in the room, a Mainstreet Research poll that gave the Conservative Party of Quebec Eric Duhem (PCQ) second place! This study was commissioned by an oil and gas company that wants to develop shale gas in the St. Lawrence Valley, Utica Resources. Participants were asked questions about energy with questions such as: “Quebec imports all the oil and gas it consumes. Does this concern you? and “Do you think Quebec should produce natural gas for export to Europe to replace Russian gas?” The poll was taken at a time when the National Assembly was considering Bill 21 to ban all exploration and development of hydrocarbons in Quebec, a piece of legislation strongly opposed by the study’s sponsor.
So far no problem.
However, the sponsor took the opportunity to also measure respondents’ voting intentions. And the results are surprising to say the least: 36% for Coalition Avenir Québec (CAQ) versus 24% for PCQ. The rest of the opposition parties are far behind.
However, the recent average of PCQ support favorable to shale gas development in Quebec is hovering around 14%. How to explain what seems to be true statistical aberration? Mainstreet Research polls are conducted via automated phone calls (known in the industry as IVR). interactive voice response) and thus are truly “probabilistic”, meaning that the sample of respondents is random (in this particular case with an error of ±3%, 19 times out of 20). The probabilistic nature of this type of survey makes them subject to statistical fluctuations, but not of that order.
Mainstreet Research President Quito Maggi also expressed doubt about the results of his study. “We have been tracking the growing support for PCQ for over a year now, mainly in the Capitale-Nationale region. This data, however, seems far outside the current norm compared to what we’ve been measuring in recent months and shows us some volatility, especially among younger voters. As with federal PPC in 2021, PCQ support is likely inflated due to pandemic fatigue. »
We are faced with a classic example of selection bias, because the survey would never have made it to the media if PCQ had not been beneficial. Imagine an actor – interested and partisan – who conducts opinion polls while the latter is on the move, with everything that happens in the news. The first survey gives 14%, the second – 18%, the third – 24%. Because the only data offered to the public is the highest, so we are faced with a biased perception of reality.
To illustrate this point, here is a small plot of 500 points representing random data with a value of around 18 (with a standard deviation of 3). The vast majority of scores are just below or above 18, with only a few further from the average (arrows point to 12 and 24, six points from the average).
Now look at exactly the same distribution, but with less than 20 points subtracted from the chart.
The perception is different, isn’t it?
And it is for this reason that the poll in question was considered partisan in the list of Qc125 polls (see full list here). It’s not that it was compromised or falsified, but it was made public only because it suited the parties involved.
The second survey comes from Synopsis Research Marketing. The national figures for the Liberal Party, Parti Québécois and Québec Solidaire are about the same as those for Mainstreet, but for the CAQ and PCQ the data show real discrepancies.
According to Synopsis, CAQ is in first place with 44%, the highest share of support for the François Legault formation since last fall. As for PCQ, it sits at 16%, which is statistically equal to PLQ and QS (but still above its current average).
However, this study was commissioned and provided by… CAQ. Moreover, only partial information was disclosed to us, namely national data and breakdown by mother tongue and age groups.
Naturally, the CAQ wouldn’t give a reviewer an unfavorable poll. Again, this creates a selection bias. That is why he was also tried on a party basis. This has nothing to do with the performance or reliability of the poll, but with the source that leaked the poll.
The same was with Léger poll spent this winter riding Marie-Victorin, resulting in a tie between CAQ’s Shirley Dorismond and PQ’s Pierre Nantelle in the April 11 by-election. It was commissioned and then donated by the Party of Quebec. Would it have been made public if he had measured Pierre Nantel’s gap by 10 points? Of course not. The same goes for CAQ and his home review Synopsis.
I often say that one must be careful when interpreting political polls, especially those that are filtered for partisan purposes, published not for the purpose of informing, but for the purpose of persuasion. That’s why I think it’s critical to re-emphasize the importance of keeping the focus on the entire landscape of public opinion, not just on a tree in the forest.
* * *
All Quebec voting intent polls can be found on the Qc125 webpage here.