Thursday, April 27, 2017

Trust the Polls: Marine Le Pen is Not Favored to Win (You Know. Probably).




Image Credit: CNN
You guys have all probably heard by now that Marine Le Pen has made it into the French presidential run-off election and is set to square off against Emmanuel Macron in a couple of weeks. The far right candidate has repeatedly made headlines for her controversial social stances, anti-EU rhetoric, and her party’s racially charged and antisemitic past. That’s a bit of an understatement considering that her father was in charge during that racially charged and antisemitic part, but c’est la vie. She received approximately 21 percent of the popular vote to Macron’s 24 percent meaning that neither would automatically win the election. The instant that it became evident that it would be a run-off between the centrist (or at least establishment) candidate and a right-wing populist there were comparisons being drawn between now, Brexit, and the 2016 US Presidential election. People have been comparing her to President Trump and asking if she has a serious chance at being France’s next president, drawing inferences from what is made to seem like an emerging trend in global politics.

First and foremost, the fact that she has even made it this far means that unequivocally, yes, she has a “serious” shot. She and Macron are the two candidates vying for the position and people should take them both seriously. If you dislike one or the other and you’re in a position to vote go out and do it. That said, there’s a difference between “serious” and likely. And the polls strongly suggest that her victory is not particularly likely provided things remain constant.

“But Peter,” You may be thinking, “Trump won the election despite what the polls said. You yourself said that it was a humbling experience — so how can you be so sure?”

Simple. It may come as a shocker to some, but the United States and France are actually two separate countries.
Coulda fooled me.




Comparing the two events is like apples and oranges.

First and foremost, France does not employ an electoral vote system; they rely on the popular vote. I can not say this emphatically enough but Trump lost the popular vote — which the polls accurately called. There’s the fact that most people tend to be politically moderate (which Macron is) and that Macron’s policy positions are closer to those held by the majority of candidates expelled from the race — evidenced by the fact that the two of the top three performing candidates ejected have already endorsed Macron. If you can’t get what you want, you tend to go with the closest available option or at least do everything to avoid your absolute least preferred option. Which is Le Pen for a number of French voters. But, and I don’t mean to beat a dead horse, there are also the polls explicitly showing that Macron is more generally favored in a head-to-head with Le Pen.

I discussed a few of the reasons why we should generally trust the polls in a recent YouTube video but I want to go a bit more in depth here. First and foremost, as Nate Silver has pointed out, overcoming a twenty point gap would be a helluva feat. People may be inclined to make allusions to Sanders’ upset over Clinton in the Michigan primary last year, but there’s a difference between infrequently polling a populace that is hard to pin down and is guaranteed to have low turnout versus frequently polling a populace with high turnout and stable aggregate preferences. There were more French polls in the last 7 days than in the entire month leading up to the Michigan primary. We have a lot more data to draw from.

But not only do we have more data but we have more high quality data. Every poll regarding the French election over the last four months boasts a higher sample size than the largest one ever done for the Michigan primary and largely come from reputable agencies. (You may imply what that intimates for the Michigan polls what you will). These polls are packed tightly together and, since most eligible French voters actually vote, it’s really easy to draw solid samples of likely voters. I’ve mentioned before that polls are a probabilistic endeavor rooted in a kind of statistics reliant on repeated trials for validation. Well, we are getting a crud-ton of them and they are independently arriving at largely the same figures: Macron at about 60 percent and Le Pen at 40 percent. Even more significantly, it’s been hovering at those values since January implying that these preferences are pretty stable. Stable opinions are harder to change; once people make up their mind they’re pretty much set for the election. So although Le Pen is drawing attention to Macron’s wife and trying to nab the “floating” voters of those who have been ejected from the race, there probably aren’t that many voters who are indecisive between the two given just how much attention has been paid to this race.

So why are people still skeptical of the polls given all of this evidence? I suspect it’s for two reasons. First: predictions of recent historical contests pitting the globalist status quo against populist right were portrayed with way too much certainty than what was appropriate considering how close the races were in the models. This was largely due to how these stories were reported on but also in the kind of confidence that surveyors were stating. Second: a Le Pen victory in the face of polls would perpetuate a really engaging narrative that is either explicitly endorsed or tacitly referenced in articles on the topic. You know, the story of haughty left-wing elites using their best methods to try and understand the machinations of the society only to be bested and proven wrong by a silent, neglected conservative majority who will no longer stand idly by as their society abandons them and what they believe.

The only problem with the story is that it isn’t true. We’re great at seeing patterns and weaving narratives; it’s how our species have passed down wisdom for literal eons. Our brains are probably wired for it and we’re subsequently apt to find patterns that aren’t actually there. Yes there is a significant plurality of people who are fed up, who are leery, and who are riled-up and ready to show it. It’s important that we don’t neglect them and their concerns in the policy arena and that we do everything we can to make the most valid measures possible to understand their positions and sample appropriately so that we don’t exclude them. But, by and large, they are not underrepresented in the samples. They are not unduly shy in expressing their preferences. The measures worked and the predictions were accurate even if the substantive outcomes were unexpected. So although polling can never be 100 percent certain, unless something huge happens in the next couple of weeks Macron is in a great position to be the next president of France.

You know, probably.




This post first appeared on Medium.


Peter Licari Student in Political Behavior and Elections

No comments:

Post a Comment