Do polls "predicting" François Hollande defeat in first round 2017 French presidential elections make any sense?

Ah political polls. Don't you just love 'em?

The frequency with which they're commissioned and published in France would have you believe the French do...well at least the country's media does when the news schedule is slack or journalists feel like a good old job of "professional" political speculation.

The latest "nonsense" poll to be published is one carried out by OpinionWay for Le Figaro and LCI telling us that if the 2017 presidential election were to take place today (well, you know how these things work) François Hollande would not make it past the first round.

He would only win 18 per cent of the vote in the first round, trailing both the far-right Front National (FN) leader Marine Le Pen (25 per cent) and the (presumed) candidate for the centre-right Union pour un Mouvement Populaire (Union for a Popular Movement, UMP) Nicolas Sarkozy (29 per cent).

In other words the presidential second round in 2017 would be between Le Pen and Sarkozy.

(screenshot OpinionWay poll of voting intentions)

"Allô ! Non mais allô, quoi," to quote a great modern day French thinker.

What's this all about.

Seriously - forecasting results three years hence, based on a poll taken today is...well, misleading to say the least.

Of course it's probably one of the drawbacks of the "quinquennat" or the five-year presidential mandate passed by Jacques Chirac in 2000 and first used in 2002 to replace the previous seven-year term in office.

No sooner has a president been elected in France, than attention seems to focus on what might or could happen five years down the line.

Of course Hollande is unpopular at the moment. We know that because...well the polls keep telling us and the media delights in repeating it.

But predicting that Hollande might not even make it past the first round in 2017 when he's not even halfway through his term in office is...well surely complete and utter nonsense.

In fact it's a non story and one of pure fiction.

Sure it feeds into the widely-held (according to those very same opinion polls) belief that Hollande is incompetent, lacks clear vision and was the major reason for his Socialist party's defeat in last month's local elections,

But in and of itself, the survey says nothing about the likely outcome in 2017. Rather it's just a snapshot of current opinion and the image those polled have of Hollande.

After all, if a week is proverbially "a long time in politics", what the heck does that make three years?

Not convinced? Then just take a look at what a poll, taken at a similar stage during Sarkozy's term in office, predicted for the first round of the 2012 election - two years before the Dominique Strauss-Kahn affair hit the headlines.

Sarkozy followed by Martine Aubry and François Bayrou.



(screenshot La Nouvelle Edition, Canal +)



The actual result (just in case you needed a reminder) Hollande 28.63 per cent, Sarkozy 27.18 per cent and Le Pen 17.90 per cent.

French presidential election 2012 - too many opinion polls - says survey

Now there's a headline that's bound to shock.

How could anyone imagine for one moment that the French would feel that way about opinion polls?

After all in the run-up to the first round in this year's presidential elections there were only 375 according to the Commission des sondages, the regulatory body which, as its name suggests, oversees opinion polls.

One of many, many opinion polls (screenshot BFM TV)

That figure is a record (surprise, surprise) far outstripping the total number in both rounds during recent presidential elections; 293 in 2007, 193 in 2002 and 111 in 1981.

And the commission sure has its work cut out with newspapers, television and radio constantly turning to the likes of BVA, CSA, Harris, Ifop, Ipsos, Opinion Way and TNS-sofres to question French voting intentions.

In a world that's far from being perfect those conducting opinion polls seem to be...well even less perfect.

Oh yes they might be congratulating themselves at the moment on getting it "almost right" but several of them underestimated by a couple of percentage points the support for the far-right candidate Marine Le Pen and others overestimated for the far-left candidate Jean-Luc Mélenchon.

The numbers aren't so dramatic according to pollsters who always give themselves that all so important "margin of error" but that doesn't stop the French from being fed up with the frequency of published polls and the perceived impact they have on voting intentions.

That's according to - of all things of course - a poll.

You just can't get away from them can you?

Even though talking to anyone on the street in France would probably give you much the same result, that would only be anecdotal of course and lacking the "objectivity" of the poll conducted by Ifop.

Anyway, according to this, in a manner of speaking, "poll of polls" 63 per cent (of those questioned) think the media publish too many of them and 60 per cent believe they have an influence on the way people vote.

But here's the thing.

By and large those questioned only consider polls can influence the way other people vote; only 15 per cent say their choice can be swayed.

As far as Frédéric Dabi, the general deputy director of Ifop is concerned, that's proof that polls have a value without distorting the outcome.

(you might need to read the following quote a few times because it seems like a classic case of doublespeak)

"Even if that percentage (believing polls can influence the way people vote) isn't negligible, the fact that the overwhelming majority believes that the surveys do not affect their vote undermines the whole discussion about the influence of polls," he says.

"It's the sort of debate that occurs every time a party or a candidate is in trouble."

Right. Understood.

So 375 opinion polls which reflect (more or less) voting intentions without having an impact on the outcome in the first round and more - many more - to follow in the second.

Whoopee!