France 2012: the official thread

[big bad fab]

0.8 0.6 0.4 0.2 but it's not so artificial because we lost anyway 4 polls at once.
So the changes are mainly the fact of Bayrou really rising, not of my new weighting §

And there is another OpinionWay poll tonight where he is at 14... (I'll publish it tomorrow on my blog).
And Hollande is at 27 in this poll...
Don't worry, I think Sarkozy is more at risk because of Bayrou, than Hollande.
But it's beginning to be interesting

[big bad fab]

2012 Big Bad Tracker #34 - 26 December 2011      
            
Each opinion poll is weighted with its sample and loses a ponderation of 20% each week: it will be the case until mid-January.            

I haven't given a ponderation based on the quality of pollsters, as big changes have occurred in their teams, without any certainty that the 2007 rating is still OK.      
And this tracker is built more out of fun than out of scientific seriousness...       

      

Hollande   29,60
Sarkozy   25,13
Le Pen   17,22
Bayrou   10,87
Mélenchon   6,69
Joly   4,08
Villepin   2,20
Morin   0,65
Lepage   0,53
Dupont-Aignan   0,88
Boutin   0,62
Nihous   0,00
Arthaud   0,48
Poutou   0,38
Chevènement   0,67
   
Hollande   57,36
Sarkozy   42,64

Hollande is back to pre-primary levels.
Bayrou is really surging, now above 10%.
It's not very bright for Sarkozy but he is more or less steady.

[big bad fab]

Fab, I strongly encourage you to adopte a more moderate position regarding ponderation. I've had a look at the program you present on your blog, and I'm a bit worried you're getting too fast in the way you want to calculate the tracker after January. Personally, I'd wait for mid-February to switch to 25%, for April to switch to, say 33% and wouldn't go as far as you go by counting only the week's polls in the very end. I know things can move faster in the very last week, but overestimating the predictive power of final polls might be a mistake. Don't get me wrong, all your ideas are really great and I perfectly understand your reasoning. I hope you'll just take my objections into account.

Also, I've counted and this tracker is actually the 34th.

I don't know where I've made a mistake in the number of trackers
Will try not to forget to change it on Monday.

As for weighing the polls, I'm happy to have some feedback from you.
Well, that's only my current plan. I'll see how many polls we have each week from January. Maybe I'll make this a bit slower.
In fact, I've set this based on the 2007 campaign, which is very fresh in my memory.
This time, Sarkozy will try to have a slower flow... err... wait, are we sure ?
And Hollande, theoretically far better organized than Royal, will have his own agenda.
And Marine is far more professionnal than Big Daddy.
And Bayrou will be more active because he really believes in himself this time (and it's the last occasion for him...).
So, I think the pace will be quicker than we expect for the moment.
Hence my scheme. But I'll change it if necessary.

Just remember that you repeated me that I should change the pace during the last summer and through the autumn... I was rather right not to change too fast then. Now, it's the contrary: maybe you'll be right.

Anyway, this campaign may well be more suspenseful than anticipated, after all

[big bad fab]

2012 Big Bad Tracker #35 - 2 January 2012   

            
Each opinion poll is weighted with its sample and loses 20% of its weight each week: it will be the case until mid-January.            


      

Hollande   29,31
Sarkozy   24,95
Le Pen   17,21
Bayrou   11,36
Mélenchon   6,59
Joly   3,98
Villepin   2,35
Morin   0,65
Lepage   0,54
Dupont-Aignan   0,89
Boutin   0,64
Nihous   0,00
Arthaud   0,47
Poutou   0,39
Chevènement   0,67
   
Hollande   57,17
Sarkozy   42,83

No new poll during this holiday week. So Bayrou is the only winner.

[big bad fab]

We are Friday and still no new poll
I was convinced they'd go back to "work" as soon as the 2nd of January...
They seemed hungry enough to be interested in this "good school-boy" start from Hollande (he is really lucky to have Sarkozy against him...); so, why can't they "create" the event by polling NOW ?

French medias are really big, big sh***

[big bad fab]

Thanks for the info.

But this is what I've said: they all poll at the same time and then they are all unemployed at the same time...
My next Monday tracker will again take into account absolutely no new poll

Sure, it's the las time it occurs, I think. But still...
Hope YouGov will land in France soon

[big bad fab]

2012 Big Bad Tracker #36 - 9 January 2012   

            
Each opinion poll is weighted with its sample and loses 20% of its weight each week: it will be the case until mid-January.            


      

Hollande   28,49
Sarkozy   25,23
Le Pen   17,8
Bayrou   11,94
Mélenchon   6,33
Joly   3,59
Villepin   2,51
Morin   0,78
Lepage   0,53
Dupont-Aignan   0,78
Boutin   0,61
Nihous   0,00
Arthaud   0,50
Poutou   0,49
Chevènement   0,44
   
Hollande   55,96
Sarkozy   44,04

The new IFOP poll is included (all the numbers on my blog).

[big bad fab]

As long as Antonio doesn't physically harm himself as a result....Sarkozy actually winning would be hillarious

I think at this point I would just give up on partisan politics forever and focus only on data and analysis.

Take some comfort with the last BVA poll.
I'll try to publish it tonight.
Well, I must acknowledge BVA is a bad pollster this cycle and they seem to be always late in the trends...

[big bad fab]

France is now a great country:
http://www.parismatch.com/Actu-Match/Presidentielle-2012/la-presidentielle-en-temps-reel.html

Each day, at 18:00 PM (CET), you'll have a daily poll from IFOP, with a panel of 300-350, rolling on 3 days, so that we have a total 900-1000 voters...

My own tracker will take into account every new complete poll, i.e. each 3 days, I'll put the result into the database.
IFOP will become more important than others, but it's deserved and maybe OpinionWay, IPSOS and Harris will have their own trackers or, at least, more polls.

Details (in French) here : http://sondages2012.wordpress.com/


Well, a great day for polling, an awful day for us, UMPers

[big bad fab]

2012 Big Bad Tracker #37 - 16 January 2012   

            
Each opinion poll is weighted with its sample and loses 20% of its weight each week: it will be the case until mid-January.            


      

Hollande   28,2
Sarkozy   24,66
Le Pen   18,25
Bayrou   12,88
Mélenchon   6,91
Joly   3,21
Villepin   2,55
Morin   0,54
Lepage   0,44
Dupont-Aignan   0,44
Boutin   0,52
Nihous   0
Arthaud   0,49
Poutou   0,45
Chevènement   0,45
   
Hollande   56,28
Sarkozy   43,72

[big bad fab]

I've just published on my blog a short study of transfers between 1st and 2nd rounds.
As there are some graphs and if you know that "vers" means "towards", you'll be able to understand the essential part of it

For those interested, but not fluent in Victor Hugo's language...

And this article will give some comfort to all of you, "liberals", who are so worried of a Sarkozy comeback

[big bad fab]

BBF I really like much your blog, but everytime, it"s irritate me about your heavy remanence of MARION "Marine" Le pen.


As I know, if you come this way, you have no reason to call Martine Aubry "Aubry" or Ségolène Royal "Ségolène".

There is one BIG difference between Panzergirl and the others: her motto includes "marine", which is a fake first name. She has never had Marine in her official identity. Never. This is not even her fifth first name or something. She is just "Marion Anne Perrine Le Pen".
Of course, "la vague Perrine" or "la vague Marion", it's less cool
Sarkozy, Royal, Aubry et alii have not created anything.

But if other readers are irritated, I may take it into account

[big bad fab]

2012 Big Bad Tracker #38 - 23 January 2012   

            
Each opinion poll is weighted with its sample and loses 20% of its weight each week: it will be the case until mid-January.            


      

Hollande   28,33
Sarkozy   24,07
Le Pen   18,80
Bayrou   13,11
Mélenchon   7,19
Joly   3,22
Villepin   2,41
Morin   0,56
Lepage   0,33
Dupont-Aignan   0,54
Boutin   0,32
Nihous   0,00
Arthaud   0,50
Poutou   0,28
Chevènement   0,34
   
Hollande   56,72
Sarkozy   43,28

These numbers are boring.... Can't Bayrou really surge ? Can't Sarkozy fight back ? Can't even Mélenchon climb above 10% ?
Bah....

[big bad fab]

You're right, I'll change it for next week (have not done so before by being purely lazy )

[big bad fab]

Tonight on TV, Hollande wasn't too bad. Juppé looked old.
Only some weaknesses on his personality, but other wise good on his program...
He will be very difficult to beat, even by Bayrou.

[big bad fab]

2012 Big Bad Tracker #39 - 30 January 2012   

            
Each opinion poll is weighted with its sample and loses 20% of its weight each week: it will be the case until mid-January.            


      

Hollande   28,45
Sarkozy   24,04
Le Pen   18,52
Bayrou   13,26
Mélenchon   7,54
Joly   3,24
Villepin   2,00
Morin   0,49
Lepage   0,37
Dupont-Aignan   0,59
Boutin   0,36
Nihous   0,09
Arthaud   0,46
Poutou   0,24
Chevènement   0,37
   
Hollande   57,05
Sarkozy   42,95

Mélenchon is up but it doesn't hurt Hollande, who is even stronger.
Bayrou is more or less stalled.

[big bad fab]

2012 Big Bad Tracker #39 - 30 January 2012   

            
Each opinion poll is weighted with its sample and loses 20% of its weight each week: it will be the case until mid-January.            


      

Hollande   28,42
Sarkozy   23,91
Le Pen   18,65
Bayrou   13,27
Mélenchon   7,57
Joly   3,22
Villepin   1,97
Morin   0,52
Lepage   0,37
Dupont-Aignan   0,58
Boutin   0,33
Nihous   0,09
Arthaud   0,46
Poutou   0,25
Chevènement   0,38
   
Hollande   57,04
Sarkozy   42,96

Mélenchon is up but it doesn't hurt Hollande, who is even stronger.
Bayrou is more or less stalled.

UPDATED: sorry, I've forgotten to add the last daily IFOP...

[big bad fab]

2012 Big Bad Tracker #39 - 30 January 2012   

            
Each opinion poll is weighted with its sample and loses 20% of its weight each week: it will be the case until mid-January.            


      

Hollande   28,38
Sarkozy   23,92
Le Pen   18,78
Bayrou   13,20
Mélenchon   7,56
Joly   3,20
Villepin   1,97
Morin   0,52
Lepage   0,34
Dupont-Aignan   0,62
Boutin   0,30
Nihous   0,08
Arthaud   0,46
Poutou   0,28
Chevènement   0,39
   
Hollande   57,04
Sarkozy   42,96

Mélenchon is up but it doesn't hurt Hollande, who is even stronger.
Bayrou is more or less stalled.

UPDATED: sorry, I've forgotten to add the last daily IFOP...

UPDATED AGAIN ! You won't believe it, but this sh*** pollster, IFOP (yes I'm angry !), has in fact released other polls which are different from their daily ones...
So, another big one is in, but, fortunately, with no great effect.
It's really a mess in their daily rolling poll, because they don't give the daily sample and they change the sample each night !!!
And for their other polls , they don't say if it's different or not from their daily one and you're forced to entail it from their foggy comments...
They seemed the most reliable this cycle, but frankly, they are essentially here to make big money... and they don't care about political geeks...

[big bad fab]

It's really a mess in their daily rolling poll, because they don't give the daily sample and they change the sample each night !!!
And for their other polls , they don't say if it's different or not from their daily one and you're forced to entail it from their foggy comments...
They seemed the most reliable this cycle, but frankly, they are essentially here to make big money... and they don't care about political geeks...

Actually they are polling all the day long, and use the data from day-1 at noon to day-3 at noon so it's a little messy because at the beginning they were using all the day-1, all the day-2 and all the day-3 (hope it's crystal clear, but I'm not sure). But you can infer the daily sample by substracting the daily sample to the one of the day before, right?
Concerning the other online poll they released earlier today, it was done during only two days. They used the daily tracking sample (about 660 people), and added about 700 interviews in order to have a large sample.

Well, clearly they are good at trying new things (even if I'm a bit conservative - on this aspect - and prefer telephone polls than online survey) but completely messy. I definitely prefer less vocal but more reliable pollsters like TNS Sofres and IPSOS.

Unfortunately no, because, one day it's day D-3 plus day D-2 plus day D-1 plus day D.
The next one, it's the same plus day D+1. and then it's only D-1, D, D+1 and D+2. And then, only 3 days, etc.
It's completely erratic and you can't infer some daily samples because the total samples are never calculated with the same number of daily samples in them.

[big bad fab]

But one day the sample is 866, the next one 945 and the one after that 1176. And then back to 960 or 877...
And as they don't publish it on Saturdays and Sundays...
It's really impossible to infer daily samples.

I hate them: it's like promising to give you chocolate and then be obliged to only look at it through a window

[big bad fab]

2012 Big Bad Tracker #40 - 6 February 2012   

            
Each opinion poll is weighted with its sample and loses 25% of its weight each week: it will be the case until mid-February.            


      

Hollande   29,93
Sarkozy   24,29
Le Pen   18,05
Bayrou   12,72
Mélenchon   7,8
Joly   3,05
Villepin   1,66
Morin   0,39
Lepage   0,31
Dupont-Aignan   0,63
Boutin   0,18
Nihous   0,08
Arthaud   0,42
Poutou   0,2
Chevènement   0,28
   
Hollande   57,28
Sarkozy   42,72

Leftists of the world, rejoice !
Hollande is unbeatable and Sarkozy is doomed !

[big bad fab]

Wasn't it a bit hasty to switch from 20 to 25% in weighting ? The rythm of polls doesn't seem to have accelerated much since last month.

Oh, it has in fact: I've still 3 polls to publish on my blog
22 polls are currently in the tracker, with a total sample of 13,904 RVs !!!

And, what is more, we have now one more event each week. So the rhythm of the campaign justifies this new weighting.

And it will change again on the 20th of February, which will be quite fine, since Sarkozy will probably have declared himself at this date (or will be about to do it). So, I'm pretty satisfied of my weightings so far

[big bad fab]

Oh, it has in fact: I've still 3 polls to publish on my blog

Damn you, why didn't you publish them already !

Anyways, stay prudent. 25% might be fine, but if you raise it further the tracker could be subject to heavy and meaningless swings. Let's keep some correction based on past results until the very last days of the campaign.

Published yesterday night

Well, I also need not to have excessive delays in swings or new trends. I'll see, but for the moment, I stick to my changes in weightings.


As for Hash's numbers, well, they are impressive, but not quite expected and not very original.

The only numbers that can be surprising are:

- for hunters, the leftism is even greater than expected, and might explain that Nihous is so low in polls. Sure, hunters are really not from Sologne

- for Moroccans and Tunisians, it's probably just statistical noise, but it can be explained by the "chiraco-gaullism": being anti-US (or viewed as such) is probably popular with Arabs (at least with Arabs in more "enlightened" context, hence the bad result for Villepin among Algerians, who anyway can't be fans of gaullism...)

- for Polish and Portuguese, Sarkozy can probably thank the weakness of Boutin and the fact that there is no candidate able to carry Catholic values.

- the difference between Pieds-Noirs and their children is no surprise, but is very interesting.

- among civil servants, territorial ones are the more leftist and that's again no surprise, but it's interesting to see it confirmed.

[big bad fab]

Probably not Porte d'Eau materiel.

GAME CHANGER

I don't think so, frankly: remember how many times Royal's flat was "visited" ?
Only the loonies face break-in, it seems

[big bad fab]

When you get new poll results, I'd be grateful if you could publish them in these thread as soon as possible. It is understandable that you need time to write a full analysis to post on your blog, but here you'd only need to give us the rough numbers.

I don't have enough time. Usually, they are published quite quickly on my blog (and readable even for our non-French friends ).