The problem with polling 2020

Contrary to the prevailing notion, especially from conservative ranks, the polls in 2016 were actually not that far off. With one or two exceptions, Wisconsin being the most significant, the pollsters didn't stink up the joint as much as it might appear at first glance. Even the aggregate national test between President Trump and Hillary Clinton was within a point and a half. My final projected vote percentage was Hillary 49.2% and Trump 45.8%, a projected victory margin of 3.4% for the former first lady. Once all the votes were tallied, she ended up winning the popular vote by 2 points, 48.0-46.0%. Most elections, the pollsters would have been applauded for such an accurate prediction.

But one doesn't win the White House without a majority of the Electoral Votes, and those come via majorities in the states rather than nationally. So how did pollsters fare there? Well, again, not as bad as one might expect given how convincingly, it was reported, Clinton was favored to win. Here's a view of all states decided or predicted to be decided by 5 points or less. I've colored the result in green if the polls got within 4 points of the actual result, dark red otherwise, and I've included the candidate which the polling error favored.

Read Full Article:

https://electionprojection.com/the-problem-with-polling-2020.php