The US elections are around the corner and we will, hopefully, soon know who the US President is for the next 4 years. The polls currently have Biden winning, but we all remember that Clinton was also winning 4 years ago. So, should we believe the polls this time around? Have they learned?

An interesting article in a Dutch newspaper argues that they have fixed the main issues and that we can be more confident in the predictive nature of the current polls:

  1. They are now better representing the different population segments in the polls, such as the lower educated segment of the population that gave Trump a boost in the election 4 years ago.
  2. The polls now focus less on national popular vote, but rather take into account the US voting system, where results should be measured by State. In the past, State-level polls were of worse quality, but that has improved, according to the article.

Although not necessarily related to the quality of the polls, this time around the share of undecided voters is lower, making it easier to predict the results.

So yes, it seems that the industry has reflected and fixed some of the issues that caused them to incorrectly predict a Clinton victory. Naturally, there is no guarantee, but it does seem that the results of the polls should be closer to reality.

Source (NRC Handelsblad): Biden gaat aan kop, maar kloppen de peilingen?

Mastodon