Eszter-Petronella SOÓS PhD

54. If you do not understand it, you better ignore it?

The editorial staff of Aujourd’hui en France / Le Parisien decided that they would simply stop commissioning opinion polls ahead of the 2017 French presidential election. They say  – inter alia – that polls and surveys failed to predict the results of the Brexit referendum and of the US presidential election. I am utterly outraged by this argument, not to mention by the fact that other media shared this as relevant news…

First of all, there are better and worse analyses; and all analysts also have better and worse days. That’s a fact. If one is not entirely imprisoned by their preconceptions and ideological madness, they are going to try and pay attention to those analysts who say more right than wrong, and they are going to even try to learn their methods (see an interesting and important book about this here).

For example: I paid most attention to Nate Silver’s FiveThirtyEight project during the US presidential election process even when the US left-wing media loudly criticized it because it had not given a “high enough chance” to Hillary Clinton’s expected victory. Silver´s staff clearly indicated for example that based on numbers there was a chance that Clinton would win the majority of the popular vote, but fail to secure the majority in the electoral college (and that is what happened).

After the election, they explained: we “strongly disagree with the idea that there was a massive polling error. Instead, there was a modest polling error, well in line with historical polling errors, but even a modest error was enough to provide for plenty of paths to victory for Trump.[…]”

Of course, in order to understand these complex forecasting models, we need to know, for example, that a forecast predicting a 90% chance of Clinton presidency also means a Trump presidency, on average, one out of ten times. During the last week of the campaign, according to the Silver project, Clinton’s chance was around 70%, which also meant that three out of ten times… you understand.

I do not say that all these methodological issues related to opinion polls are crystal clear to me. This is not my area of expertise and I have no daily practice in it. I also know that everything can be criticized, and methodology is not almighty either.However, it is a matter of elementary intellectual humility to realize and recognize what one knows and what one does not know. And – Goodness! – when one realizes one´s challenges, one may ask questions, read and learn…. 

I have just recently realized that statistics, probability theory are almost as important university subjects as opinion polling or the practical use of SPSS are… It does not hurt to know what the program is doing when it throws out stupid things… (. Therefore, now I am very interested in probability theory, so I know I also need to learn some analysis – and yes, I am struggling with it, but I am also making efforts to acquire the mathematical background necessary to political forecast (for now I would prefer not to tell the details of my fight, but I can assure you I try my best ().

I am extremely upset when a leading daily newspaper’s editorial staff decides not to make any effort to learn, to understand research critically, or to apply available knowledge. No, it decides to throw everything out in the trash. And this staff is proud to do so, even when, ideally, their task is to educate not only themselves but also the readers who also have a hard time interpreting opinion polls, probability and methodological notes…
No, here, the password is: “if you do not understand it, you better ignore it!” This intellectual levity brings shame to all intellectuals, and as an intellectual, I feel I should apologize instead of these – ” journalists”.

(Featured image: