Why I (almost) don't Believe in Polls

Why I (almost) don’t Believe in Polls

Some people affirm unhesitatingly today that we are governed by polls. And others ask why polls were so wrong about Brexit or Trump…

Why polls anyway?

So are polls truly able to foresee the results of elections? Can they really tell us what a given population thinks? Even though pollsters themselves, as well as journalists at times, remind us that polls are but a snapshot, a reflection, they are nevertheless used by political analysts and everyone as information for the impending results of the elections or on the assumed opinion of the population.

So two questions come up: do polls really give us accurate information on the results of elections? and are they really a reflection, a snapshot of public opinion? The answer to both these questions is negative.

Pollsters themselves do not claim any more that polls are tools for forecasting. Especially these days…

Moreover, even if polls can inform us about the opinions of the persons polled and their evolution in time, they in no way reflect a faithful image, a snapshot of public opinion (if at all that exists).

The reason is straightforward: polls are essentially based on linear and/or temporal projections as:

  • If so many among the cross-section of people polled behave in this way, then x percent of the population will behave the same way
  • If the people polled provided us in the past with certain information on the overall population, we can advance the same hypothesis, we can even make the desirable corrections in the current results, keeping the past in mind

Wrong tools

Well, all this cannot hold any longer and rightly so! The world is no longer linear, it has become chaotic, it has become fractal.

And here too, we have several examples:

  • The choice itself of the sample of those polled raises a problem. How do you project results that were obtained from a sample of people as they were got in the past, when more and more people do not have a landline (those polled are even today often interviewed over their landline), especially among younger voters?
  • More important still, we are seeing the emergence of what we might compare to positive feedback: those polled choose to answer in conformity with their desire to manipulate the result of the poll!
  • And let us not forget the phenomenon of the influence of the observer on the thing observed. So it is about the choice of pollsters to conduct polls on possible future scenarios of the second round in presidential election with candidates that should not have gone beyond the first round as per their own polls!

So, useless?

However, these polls should not be taken to be pointless. It is the way they are used that is excessive.

Today, in a non-linear world that has become fractal, these polls are probably more relevant than ever. But the probabilities and statistics must now be in consonance with the theory of chaos.

So, polls should highlight two things that are essential in rightly observing and understanding a fractal world:

  • Polls allow us to follow the changes in time: What are the answers given to the same questions by a sampling of people living in different periods? This is especially helpful to track the evolution of people opinion on subjects such as death penalty, same sex mariage, right to abortion, etc.
  • Polls are also an invaluable tool in providing us with different possible scenarios. Rather than foresee one and only one possible future reality, namely the most probable, polls allow us to work on different possible outcomes.

In a linear world, it is true, we can make projections, but in a chaotic and fractal world, we can paint scenarios, we can see different possible paths. And even more important we can create new narratives. And for that purpose polls can be helpful.

How to use polls?

It would be a pity to throw the baby out with the bathwater! We need more than ever analytic and projection tools. We need to unify the science of probability with the science of chaos in our methods of polling, as is happening already in the field of meteorology and finance, for example.

So, let us learn how to use polls for what they can really bring us, and a little less for our bar-room discussions. Let’s use them to create new narratives for a better world 🙂

You may also like

Leave a comment