Polls have been relatively stable this election. Since the writ was dropped, we can really only identify very small trends, namely the small drop of the Conservatives and the slight increase of the Liberals.

If the election were tomorrow, these projections (of a few days ago) would still be valid

However, there was one poll last week that stood out and generated some discussions. I'm of course talking of the Forum poll that had the New Democrat at 40%, the Liberals at 30% and the Conservatives at only 23%. This is a stunning result. And one that is particularly off the poll average.

Before I continue, I want to say that I actually quite like Forum. They poll all the time, they provide details and their track-record is pretty good. It's easy not to be wrong when you never poll. So this post shouldn't be seen as an indication that I believe Forum's methodology to be fundamentally flawed. Not at all.

Getting variations between polls is not only expected, it's desirable. After all, this shows that the sampling process is truly random or at least close to it. It's when polls all show the exact same numbers that this is suspicious. With that said, there is a limit in how much variation we should observe. Especially given that federal polls have sample sizes of at least 1000 observations. In the case of the Forum poll, it was actually 1440 respondents.

When you poll, you can get results off the mark if you are unlucky when sampling. Like you got too many NDP voters compared to how they are truly in the population. But you can be unlucky when you ask (say) 20 people, not when asking 1440. If you are truly selecting the people randomly, it's incredibly unlikely. If you want an analogy, it's like when you play the lottery. Yes you can win in theory, but the odds are definitely against you.

So let's look at how "unlucky" Forum got possibly. The recent (completely un-adjusted) poll average excluding this poll is:

NDP: 34%

CPC: 29%

LPC: 27%

So let's assume (for now) that this poll average is representative of the true voting intentions in the population in this country. The question we have now is: how likely is it to randomly select 1440 Canadians and get that 40% of them would vote NDP, 30% Liberals and only 23% Conservatives?

To answer the question, I simply used my usual simulations. Specifically, I ran 10000 simulations where I was randomly sampling 1440 respondents from a population where there were 34% of NDP, 29% of CPC and 27% of LPC.

Here below you have the graph of the 10000 simulations for the level of support of the NDP. As you can see, in average, we indeed get the true support of 34%. But depending on how unlucky we got in our sampling process, we sometimes get the NDP much higher or much lower. Most of the times though, we are between 32 and 36 (if we round up). This is what the typical margins of error are telling us.


So, how many situations do we have where the NDP is as high as 40% and the Tories as low as 23%? The answer: none! That's right, out of 10000 simulations, there wasn't a single case where we managed to recreate the numbers of Forum. The highest for the NDP was 38.8% and the lowest for the CPC was 24%. And we are talking of one simulation out of 10000.

If you're wondering why, this is because 1440 isn't that small of a sample size. If you prefer, think of this analogy: you can flip a coin 5 times and get "head" 5 times. Unlikely (actually less than 5% chances) but not impossible. But if you flip a coin 1440 times, chances are you will have got mostly 50% of "Heads" and 50% of "Tails". This is a similar logic here.

Maybe the odds are actually so small we need more than 10000 simulations. So I tried with 100,000. The result: one case with the NDP above 39.6% and the lowest for the CPC at 23.5%. So quite close, but let's remember it's 1 out of 100000.

So, what does it mean? There are essentially three possibilities:

1. Forum was the only one to pick up something. What I mean here is that the average above (the 34-29-27) is actually not valid. Maybe the NDP is indeed on the rise and the Conservatives falling fast.

Of course the follow up question is: how come only Forum picked that up? After all, Forum was on the field August 23rd and 24th. Others were in the field during that time and some (Ipsos ,Innovative and today's Abacus - not included in the poll average yet though) actually polled after Forum.

2. Forum is actually wrong. And I'm not talking 1 out of 20 times type of wrong. I'm talking 1 in more than 100,000. So either they got incredibly unlucky when selecting the sample, or they didn't select it correctly.

People tend to think that pollsters simply randomly choose 1440 people across Canada when in fact we are talking of a complex process with different layers (they first decide to poll a specific number in each province, etc) and weighting. If you look at the poll by province, you see the Conservatives very low in Ontario (26%), Qc (11%!) as well as in the Prairies (tied for 2nd at 28%) and even Alberta ("only" 42%). The NDP on the other hand was very high in Quebec (54%!!) and the Prairies (first at 41%). So it does seem some provinces were simply off and caused the results overall to also be wrong or incorrect.

Forum polls a lot and fast. Their polls do have more volatility than others (we have observed this before). Earlier this campaign, Forum already had a one day poll (conducted on a Sunday between 11am and 4pm) with the NDP at 38%. And let's not forget about Brandon-Souris.

At the end of the day, I'd simply classify this poll as an outlier and move on. The NDP is not at 40% and the CPC is not at 23%.
Le dernier Crop-La Presse d'hier est venu confirmer presque chiffres pour chiffres le plus récent Abacus. Au Québec, le NPD y est largement en tête avec 47% des voix, une hausse de 4.1 points par rapport à 2011. Le plus proche poursuivant se retrouve loin derrière, le PLC est en effet à seulement 20% (une hausse de près de 6 points cependant par rapport à la dernière élection). Le PCC est à 13% (en baisse de 3 points environ) et le Bloc est à 16% (baisse de 6.8 points). Comme d'habitude avec Crop, pas de pdf détaillé. Cette firme devrait avoir honte.

Une fois ajouté à la moyenne des sondages, ce Crop ne la change pas fondamentalement. Mais cela fait deux sondages de suite où le NPD est à 47%. Si ce parti avait réussi à remporter 59 sièges (sur 75 à l'époque), on peut se demander si la formation de Thomas Mulcair n'aurait pas une chance de remporter tous les sièges cette fois-ci. Après tout, le mode de scrutin récompense déjà fortement le parti qui termine premier, imaginez quand celui-ci est proche d'une majorité des votes!

En utilisant le simulateur, on obtient les projections suivantes (différences par rapport à 2011; souvenez-vous qu'il y a maintenant 78 sièges):

NPD: 68 (+9)
PLC: 7 (-1)
PCC: 4 (-1)
Bloc: 0 (-4)

Tous les partis perdraient des sièges, en particulier le Bloc qui se verrait rayé de la carte. La chute du PLC peut être surprenante compte tenu du fait que ce parti enrégistre une hausse de 6 points en termes de votes. Comment expliquer cela? Dans Bourrassa, malgré une belle victoire dans la partielle, le départ de Denis Coderre fait un peu mal et permet au NPD de passer devant (les résultats de la partielle sont ajustés en fonction des intentions de vote provinciales au moment de l'élection. Pour le PLC, la victoire aurait dû être plus importante et le départ de Coderre coûte environ 5 points). C'est la seule perte. Quant à l'absence de gain, le fait que le NPD est aussi en hausse n'aide évidemment pas.
En passant, avec ces pourcentages, Trudeau aurait encore 92% de chances de gagner son comté. Et c'est sans compter sur un effet "chef du PLC" (par contre, vu que Trudeau s'était présenté dans ce comté avant, l'effet personnel est inclus).

Le NPD est aussi garanti de remporter le plus grand nombre de sièges au Québec. Ce n'est même pas proche.

68 sièges c'est bien, mais en même temps, on pourrait imaginer davantage. Dans les faits, la seule raison que le PCC et PLC gagnent des sièges est en raison de la concentration de leurs votes.

J'ai fait mes habituelles simulations mais juste pour le Québec. Voici les minimums et maximums absolus (voyez ces scénarios comme les cas super extrêmes où les sondages auraient sous- ou surestimé massivement un parti et ce dernier aurait eu un vote incroyablement efficace ou non).

NPD: 54-76
PLC: 1-16
PCC: 0-8
Bloc: 0-4

Ainsi, un balayage complet du Québec ne semble pas possible. Le maximum pour la formation orange est 76 sièges. Et il y a une circonscription que le NPD ne remporte dans aucune simulation: Mont-Royal. En même temps, avoir un seuil à 54 sièges devrait être suffisant pour satisfaire Thomas Mulcair.

Ci-dessous, vous avez la distribution de sièges pour le NPD. Sur l'axe vertical, vous avez le nombre de simulations. Et oui, c'est pas le plus beau graph au monde!

Le NPD aurait quand même près de 15% de chances de remporter 70 sièges et plus. C'est assez fou. Bien sûr, tout cela est basé sur un sondage (ou deux en incluant l'Abacus). Je ne pense pas que le NPD soit vraiment à 47% au Québec. Mais je crois que ce parti est en effet largement devant. Et il est possible qu'il soit aussi élevé. Nous verrons dans les prochains sondages. Cependant, si la tendance se maintient, le PCC et PLC devraient vraiment s'inquiéter. Il est difficile de battre un parti qui a une province aussi grande que le Québec voué à lui et garantissant 50-60 sièges. Surtout qu'il est fort possible que cette élection ne nécessite que 120-130 sièges au total pour former le gouvernement.

Aussi, il semble que le retour de Duceppe et son effet sur le Bloc n'ait vraiment été que très temporaire. Il se peut que le Bloc serait encore plus bas avec Mario Beaulieu, mais il reste que la formation souverainiste pourrait bien se retrouver rayée de la carte le 19 octobre au soir. Nous verrons si le Bloc change de stratégie à un moment donné. Remarquez que je ne suis pas sûr qu'il y ait quoique ce soit que Duceppe pourrait faire pour retourner le Bloc au-dessus des 25%. Mais bon, l'électorat Québécois est connu pour sa volatilité.

In this article, I look at whether the Green Party can realistically expect to elect more than one MPs, in particular in Victoria.
Week two and we finally, finally have our first true, complete poll. This is from Ipsos and you can find the details at the end of this article. As a reminder, until now, we've had one Forum poll done on a Sunday between 10am and 4pm, another Forum for the GTA-only done on a Friday during the same hours, and one Mainstreet poll for BC only. We also got a Nanos, but they do a weird 4-week average which makes the numbers pretty much useless. So it's nice to finally have a Canada-wide poll, done over multiple days. The fact we have less polls during this campaign than we did before the writs were dropped is puzzling, but let's move on.

The new Ipsos numbers show a close three-way race with a slight edge for NDP, at least in terms of votes. At 33%, the NDP has a 2 points lead over the incumbent Conservatives, and a 5 points gap ahead of the Liberals.While it might not look like it at first sight, Justin Trudeau and his party are actually the ones making gains. Indeed, compared to the most recent Ipsos poll, they are up 3 points while the NDP and Tories are both down. Given the sample sizes, this is actually a significant change.

Moreover, this poll seems to confirm a small trend observed before where the Liberals appear to be slowly rising. It's way too early to tell, but here are the evidence so far. The above mentioned GTA-only poll had the Liberals slightly higher than expected. Sure, this poll had serious limitations, but we can't completely discard it. Using the Mainstreet numbers, we get the Liberals at 27% in BC (among people who watched the debate though), higher than the previous poll average and 8 points higher than in the previous poll from the same firm. On top of this, we now have had three polls with information regarding the Maclean's debate and the performance of each leader. In all three, Justin Trudeau is seen as the winner by most. The table below shows my usual "debate index" where I divide the % of people picking one candidate as the winner by the % of people who want to vote for this party. Therefore, a score above 1 means the leader managed to convince people outside of his party support while a score below 1 indicates the opposite. The idea of the index is to remove the bias caused by the fact that a lot of voters will see "their" leader as having done the best (i.e: trying to remove some subjectivity). It's also there to see which leader might have done the most in order to convince other voters to switch party.

Forum (GTA only)
Mainstreet (BC only)
Ipsos (Canada wide)
Stephen Harper / CPC
0.84
0.96
0.75
Justin Trudeau / LPC
0.96
1.10
1.00
Thomas Mulcair / NDP
0.82
0.68
0.84
Elizabeth May / Green
5.13
4.84
5.32

As you can see, Elizabeth May is clearly the "winner" using this measure. She's by far the leader who managed to reach outside of her party base. Of course, this index is usually quite favorable to small parties as they can't really do worse (plus, let's face it, a lot of people might judge the leader of a small party differently. For instance because she doesn't "have" to look like a Prime Minister). Nevertheless, May can be very satisfied with her performance.

But among the main three parties, Trudeau is the closest to 1 while Mulcair is the furthest in average. We aren't talking about a debate performance with a knockout or anything, but Trudeau seems to have done at least relatively better than his two direct opponents. Even if you don't like my index, Trudeau was chosen as having had the best performance by the most voters in all three polls.

What I'm trying to say here is that so far, there hasn't really been any bad news for the Liberals. While the polls were not Trudeau's friends this summer, the Liberal leader seems to have at least stopped the decline. The only negative was that the LPC was technically down this week in the Nanos poll, but the sample size is so small that it's hard to really take anything from it.

Of course, being (slightly) up is good, but the Liberals were starting third with virtually no chance of winning the most seats. So they had a long way to go. Let's see the projections we would get using the latest Ipsos numbers (given the lack of polls, we might as well use this survey only. No need for a poll tracker when there aren't any poll).

Voting intentions; Seat projections with 95% confidence intervals; Chances to win the most seats.



First thing first, yes, the Conservatives have more chances to win the most seats despite being projected with less wins. This can happen because the distributions of seats are not symmetric. In other words, the average of the projections is not the same as the projections of the average voting intentions. If you prefer, just interpret this as if the polls are slightly wrong (like underestimating the CPC), then the total number of seats of this party could increase quickly. On the other hand, the NDP wouldn't gain as much of an small underestimation. Also, and this is important, in case of a tie between the two parties, the model gives the win to the incumbent Stephen Harper (as is the rule of our system, at least in theory). If we look into the 5000 simulations, we actually find 2662 cases where the Conservatives win the most seats, the exact same number of scenarios where it's the NDP, and 171 cases where it's a tie.
At the end of the day though, don't read too much into this. The situation with this poll is that it's a close race between Tories and New Democrats with the Liberals far behind.

Also, you may have noticed the percentages for the voting intentions are slightly different from the published Ipsos poll. This is because I allocate undecided differently. I give more of them to the CPC because the incumbent is often underestimated (and small parties receive none). It's not a major change but it's enough to make a small difference. If you don't think we should do that and you trust the polls (or you really distrust them), remember you can always use your own numbers in the simulator.

But speaking of the Grits, they are at least back to a situation where it's possible for them to win. That wasn't the case at the beginning of this campaign. It's still a long shot, but it's better than nothing. It does show that Trudeau and his party are having a relatively good start of a campaign. Had it been otherwise (as well as win by Mulcair at the debate), the Liberals were at risk of becoming irrelevant very early on in this election.

In conclusion, nothing major has happened so far in this campaign. But if we had to pick one party/leader with a better start, it seems that would be Justin Trudeau and the Liberals. We are far from a come back or surprise win, but the slow and continuous decline observed in the recent months seems to have been stopped and even inverted. We'll hopefully get more polls soon to see if there is really a trend.


Details about the poll:
Ipsos poll for Global News. Conducted between August 7th and 10th among 2022 respondents from the Ipsos online sample. Margins of error for a probabilistic sample of this size would be 2.03% 19 times out of 20 for the NDP and less for the other parties. The question asked was: "Thinking of how you feel right now, if a federal election were held tomorrow, which of the following parties' candidates would you, yourself, be more likely to support?". More details available here.

One day after the debate, we have (finally) a new poll from Forum. Although this one was only conducted in the GTA. Still, given this is one of the most important regions in term of seats (be honest, how many times have you heard this election will be decided in the GTA?), it is actually quite useful. And the sample size is perfectly decent at 1016 respondents. However, as it was the case with the previous poll from Forum, this one was conducted between 10am and 3PM today! It's obviously a serious limitation as it means that "normal" people who go to work from 9 to 5 were harder to reach.

You can find all the details of this poll at the bottom of this post, complying with Election Canada's guidelines.

With this limitation behind us, let's take a look at the poll. The Liberals are first in the GTA with 37% of the vote. The Conservatives are second at 31% and the NDP is third with 28%. How does this compare to 2011 and the recent polls?

Compared to the last general election, this is a drop of around 10 points for the Tories (who, if you remember, probably created the biggest surprise by winning some seats not only in the suburb of Toronto, but also in the city itself). The Liberals are up about 5 points, so is the NDP.

Is this poll different from other recent ones? Well, most polls don't include a breakdown for the GTA only since the sample size isn't big enough for it. However, if I look at the projections based on the recent polls, I had the Liberals at 34% in the GTA, the CPC at 32% and the NDP at 31%. This is well within the margins of error then. It does seem though that I was slightly overestimating the NDP but it's most likely due to the previous Forum poll that had this party really high. Overall, I'd say this new poll is in line with what we would expect to see in the GTA.

How bad would it hurt Stephen Harper in term of seats? My model doesn't specifically use the number for the GTA. Instead, it uses provincial ones and adjust them using regional coefficients. As mentioned above, the percentages in the GTA by doing that are quite close to what Forum is showing in this poll. Looking at the projections, we see that Prime Minister could see his 40 seats from 2011 cut by half, while the Liberals would move from 10 to 26 and the NDP would gain 4 seats. If we remember this party was slightly too high in my projections, it means the gain in this region for Thomas Mulcair could be very limited. I mentioned before how Ontario isn't the best potential source of seats for this party.

In conclusion, this poll in the GTA shows results that are to be expected. The Liberals are slightly higher (and NDP slightly lower) but it remains to be seen if this is indicative of a bigger trend (are the Grits up in Ontario?) or if it was just this poll. What it also shows is that the GTA is indeed a battleground with all three parties being competitive, although the NDP would have a hard time winning the most seats there. For Justin Trudeau, it must be nice to finally have one poll showing some good news.

By the way, this poll also included a question regarding the debate of last night. I'm covering this in my other post.




Details about the poll mentioned in this article:

Firm: Forum Research
Data collected on: August 7th, between 10am and 4pm.
Method: IVR.
Question asked: "A federal election has been called for October 19. Which party are you most likely to vote for in this election?" followed by "Even though you may not have made up your mind, which party are you leaning towards at this time?"
Population sampled: citizen aged 18+ in the GTA.
Sample size: 1019.
Margins of error: + or - 2.97% for the Liberals, less for the other parties, 19 times out of 20.


First of all, let's face it, this is a very slow start to this election. We are almost a full week into it and the only poll we got was a snap poll from Forum, done Sunday between 10am and 4pm. Not exactly a lot of data to work with.

I'm currently setting up a poll tracking page, with all the recent polls, projections, etc. As it stands, my projections of the beginning of the week for the Huffington Post would still be valid. What's the point of a poll tracker if there is only one poll to track? So more frequent updates will come. Also, yes, I'm writing for the Huffington Post about twice weekly. I'll post the links for their articles on the blog, but not copy the texts themselves. Yesterday, I looked at how the NDP could win without actually winning Ontario.

But for now, the only actual event of this early campaign has been yesterday Maclean's debate. And as usual, the question is: was there a winner?

Before I actually try to answer this questions, I'd like to point out that quite often the pundits and analysts conclude there was no clear winner. Or that there wasn't any knockout punch. In recent memories, the only modern debates I can thin of where everybody agreed right away who lost and who won is the first presidential debate in the 2011 US election. In this case, Romney clearly outperformed a tired Obama. But otherwise, it's never fully clear cut. With that said, I often disagree that one debate won't have an impact. I think we've had plenty of times where a debate had a more subtle, slower impact on the voting intentions. Not a dramatic boost (again, like Romney in 2011), but slow adjustments. All that to say that I find pundits and experts too often downplay the impact of a debate. Of course, one could wonder if a debate taking place 10 weeks prior to the election can even have an impact.

So, was there a winner? Let's do a quick review of the press.

Mulcair turning a restrained, and some might say uncomfortably smiley performance from the Allisson Jones of the Canadian Press (for the National Post).

Justin Trudeau walks out with the win according to Michael Den Tandt from the NP.

Gilbert Lavoie from Le Soleil declares Mulcair as winner, with Trudeau surprising.

Oh, Google and Twitter seem to indicate people were searching a lot for Justin Trudeau. Not sure if that means he won or not. After all, they mention many looked for "Who does Trudeau consider to be middle class?"

May and Mulcair outshine Harper according to The Guardian.

And the Canadian politics sub Reddit doesn't seem to agree on anything, but their straw poll has Mulcair winning ahead of May. Please remember that a sub reddit is not a representative sample. But hey, the opinion of one analyst isn't either.

So, all in all, clearly no consensus. And unfortunately, not true poll either. Darrel Bricker from Ipsos told me on Twitter he thought asking 500 people could be a good idea, but a viewership expected to be low would make it difficult. Although it seems the debate didn't do that bad (without doing well either, especially compared to other federal debates that can attract 3x more viewers. But hey, it's the middle of the summer).

So, who won the debate? It seems nobody really agrees. Although written newspapers seem to think Trudeau did quite well. But it's not a landslide win.

Edit:

Actually, here is a Forum poll about the debate. The results? Justin Trudeau as the winner for 23%. But "nobody" is actually ahead with 25% of the votes. This poll is only for the GTA. And as usual for  the opinion on the winner of a debate, we need to remember that partisans tend to think their leader won. In other words, if you are a Liberal, you are much more likely to rate the performance of Trudeau as the best. In order to correct for this, I always do my debate index where I divide the % of people who declare one individual as the winner by the % of people who declare wanting to vote for this party. A score above 1 means you convinced people even outside of your party, a score below 1 indicate that some of your own partisans didn't think you were the best. If we do this, we get the following index (excluding the "others" and "don't know":

Index
Harper
0.84
Trudeau
0.96
Mulcair
0.82
May
5.13

It's not surprising for a leader of a small party to do well with this index. Plus, May did objectively relatively well yesterday. As for the others, honestly, it's really pretty much a wash and it seems everyone mostly convinced his own partisans. The Forum poll actually breaks down the question by political preferences and we indeed see that it was the case. Remember, the debate index isn't there so much to declare a winner but to see if a debate could have a potential impact on the voting intentions. This first debate certainly doesn't look like it will. But hey, we never know.


Edit 2:

A second poll, conducted by Mainstreet research in BC finds that Trudeau won with 26% of respondent choosing the Liberal leader as the winner. Mulcair is the choice of 23% while May receives 22%. Stephen Harper is right behind at 20%. So again, it's pretty much a draw here as well. However, if we do the same index, given the voting intentions in this poll (37% NDP, 26% LPC. 23% CPC and 5% Green, with 8% undecided), we find there again that Green party leader Elizabeth May clearly convinced the most people out of her own party. I think with the Forum and Mainstreet polls, we can safely conclude that May probably "won". Still in BC, Mulcair is definitely the one with the lowest index score. That was also the case with Forum. We aren't talking of a horrible performance, but definitely a disappointment.

In conclusion, no knockout punch but May seems to have done really well, so did Trudeau. Mulcair didn't bomb but was probably the biggest disappointment while Harper did what was expected.


So let's compare that to this site's poll here below