Recently I wrote an article on the New York Post’s Twitter account still being locked. This is still the case, and it made me brainstorm many other topics associated with elections . This thinking session brought up the topic of polls and predictions. How accurate are they? Why were many polls so far off in 2016? Why do polls seem to tighten the closer to election day? 
Many polls are predicting Biden to win by a large margin, and many of the points I make below indicate factors that could make the margin much smaller. The one factor that could cause a greater margin than predicted would be an increase in Black and Hispanic voter participation. This is due to the fact that Biden still holds majority over those demographics so an increase in the total number of voters would benefit him the greatest.
Here’s what I think the polls could be missing
Narrowing party registration gaps, especially in swing states.
There are indications that Republicans have significantly closed the registered voter gap, especially in swing states. Trump narrowly won the 2016 election in many of these states, so a lesser difference between registered Democrats and registered Republicans could indicate Trump holding his lead in swing states. Of course, non-affiliated voters still hold a large chunk (often 20%+) in swing states, so nothing is set in stone.
Unwillingness for people to truthfully poll, especially for a candidate like Donald Trump.
Some studies have shown that people are unwilling to share their vote choice with polls. The study linked states that around 11.7% of Republicans don’t share their truthful choice with polls, 10.5% of non-afiiliated, and 5.4% of Democrats. That’s certainly not insignificant.
Shifting demographics in the Black and Hispanic vote.
If the Black and Hispanic vote turnout is the same or less than 2016, this will be a net negative impact for the Democratic party. Trump closed the polling gap from 2016 between both these demographics (although still doesn’t poll above 50% for either demographic). So if the total pie of votes for Black and Hispanic voters doesn’t increase, Trump takes a bigger portion of a pie the same size as 2016.
Increasing distrust of media and technology amongst conservatives.
If a person doesn’t trust the news, believes they’re being censored, or is generally less trusting in institutions, why would they answer truthfully to polls? Why would they participate in the polls in the first place? Are polls even further off than the “shy” voters study linked in the truthful poll section above?
Unknown voter turnout for the Black and Hispanic vote.
2016 saw a decrease in participation with these two demographics. If participation is higher than anticipated, it will significantly benefit Biden and result in a landslide win.
Wrapping it up
While many polls try to adjust for factors like these, it’s impossible to accurately measure all the variables associated with elections. For example, every time that Florida has seen under a 4% party registration spread, Republicans have won the state, anything above 4%, the Democrats have won. Right now the spread is under 2%. But mail-in voting will be at an all time high, so will overall participation increase or are only active voters shifting their voting method?
Many polls will correctly predict the election within their margin of error. The only issue is that it’s not useful in states consistently decided by less than the typical margin of error (like Pennsylvania and Florida).
Enjoy the content? Subscribe below to get notifications of new posts and access to subscriber giveaways.