samedi 18 mars 2017

A Brief Analysis of Léger's Latest Numbers

A new Léger poll on Quebec voting intentions was published early this morning in Le Devoir and it contains several interesting observations.

First, despite low levels of satisfaction towards the current government (Leger quantifies at 65% those who are unsatisfied with the Liberal government), the Couillard Liberals are still ahead in voting intentions in Quebec. Whereas the gap between the LPQ and the PQ was at three points according to Leger's January poll, it grew to a stunning nine points according to this new poll.

Here are the Leger numbers published this morning:

LPQ 34% (+2% with respect to January 2017)
PQ 25% (-4%)
CAQ 22% (-1%)
QS 14 % (+5%)
CPQ 2% (-1%)
GPQ 1% (--)
ON 1% (--)
Autre 2%

It surely seems like Gabriel Nadeau-Dubois has had an immediate effect on Quebec solidaire's standings in the poll. Add the arrival of G. Nadeau-Dubois to Manon Massé's much publicized actions to save her Sainte-Marie-Saint-Jacques riding and you have optimal conditions for a poll boost. Both events have been well documented in the province's media.

However, a note of caution here: Quebec solidaire isn't the first party to get a boost in the first poll following the arrival of a star candidate, so it would be foolish to draw hasty conclusions on QS's standings based on this poll alone.


* * *

However, let's have a little fun with this poll: for the sake of this post, I ran the simulator using only Leger's latest numbers. Let's see what the seat distributions and regional breakdown look like.

I really want to emphasize this point: the following numbers and graphs do not constitute a new Qc125 projection, but merely what the simulator outputs with this poll only as input. For official projections, I use a weighted average of several polls spanning many months (and published by more than a single polling firm...).

Oh, and a friendly warning to PQ members: this one is going to sting. But yet again, it's only one poll, so let's not read too much into it, shall we?

These numbers are the results of 10 000 simulations. Here is the statistical distribution of the popular vote:


We can clearly see that the Liberals, with a 9% lead in this Leger poll, pull ahead of the pack. Indeed, the LPQ wins the popular vote in every single simulation. The PQ is still ahead of the CAQ, even though their respective distributions overlay one another.

Here is the seat distribution with respect to the popular vote over the 10k simulations. Again, the Liberals are clearly ahead. Many of those red dots are above the 63 seat majority line.


With a stunning 14% of the popular vote, Québec solidaire could realistically dream of winning no less than five seats on the island of Montreal, and even more than that in some of simulations.

On average, the Liberals win 60,1 seats. The PQ is a distant second with 35,2 seats. The CAQ is at 25 seats and QS, 4,6 seats.

Let's take a look at the regional distribution of seats. Consider the following figure:

Using Leger's latest numbers only, the model has the Liberals leading in 64 ridings, including 9 swing ridings. The PQ would be reduced to only three seats in all of Montreal and Laval.

Basically, both the PQ and the CAQ would barely fare better than they did at the 2014 election.

What about the odds of winning the election? Let's take a look:


Out of ten thousand simulations, the Liberals won 9 980 of them, a stunning winning probability of 99,8%. The PQ wins 0,1% of simulations and there is an equal probability of a LPQ-PQ tie.

I will end of this: to ignore polls is just as misguided and lazy as giving a single poll too much credit. PQ supporters should indeed worry about those results, but the QS surge might very fade away in the coming months, like it so often happens when a new party leader is chosen (or, in this case, a party spokeperson). It is always more prudent and wise to calculate a projection based on several polls and not a single (bad) one.

And again, this is why these numbers here are not a new Qc125 projection, even though this poll will weigh its share in the next projection, which I will publish at the end of the month, probably once CROP releases its own March numbers.

That being said, I wish all of my readers a great weekend!


* * * * * * *

Qc125 sur Twitter: @Qc_125.

Qc125 sur Facebook: @Qc125

Qc125 sur Google+