Élection présidentielle française de 2007 (French Presidential Election) 1er Tour: Your Choice?

[Douai]

Who is your choice for the Élection présidentielle française de 2007's premier tour?

[Douai]

Why is there a huge discrepancy between my poll and Ali_reza's :confused: ?In my poll Bayrou has 6 votes, and Royal has 2 votes.In Ali_reza's thread Sarkozy has 2,Royale has 2,Bayrou has 3, and Le Pen has 1.Could it be because I have the party affiliation in my poll?

[Nanbawan]

Why is there a huge discrepancy between my poll and Ali_reza's :confused: ?

If it's not strictly the same persons who voted and should logically have made the same choices, it's a little bit normal that on such a limited sample (less than 10 persons participating) you have results that do not converge. It's still pretty random.

If you throw a perfect die millions of times, you should get statistical results that are very close to (well, if it's millions of time, we can consider it equals) 1/6 for each number. Throw it in just ten attempts and you'll get a random statistical result. You'd have -I dunno - 4 times one dot, twice two, etc...

You'd have a point in saying that it's not exactly the same with opinions but the core principle is the same, you need a certain amount of people to make it representative.

Scientific polls use a specially constituted sample of around a 1000 persons that are supposed to be a good representation of the French society. Opinion poll are just...opinions normally. It's just a photograph of what people are keen to state at the moment they were interrogated. The media usually fail to stress that point.

That's for the theory because polling companies do not issue those bare figures ; they inject data from past election (difference between polls and actual results with regard with all the different categories of voters) to have a 'better projection'. So, in the end, it's a gloubiboulga somewhere between the original opinion poll and an attempt to predict the future result (and to me it's somewhere near intellectual fraud). Yet what's the validity of that if you fail to include properly those who are undecided and the amount of turn out ? They have patterns and data for that as well but it's obvious and easy to understand why the whole thing gets so fuzzy...