By Vynateya Purimetla (’21)
In the 2016 election, Secretary of State Hillary R. Clinton was favored with a 3% national lead over challenger Donald J. Trump, and was predicted to win with 300 electoral votes. In a shocking upset, razor-thin margins in swing states like Florida, Ohio, Iowa, Pennsylvania, Wisconsin, and Michigan all flipped to Trump. In margins of 0.3% in Michigan and 0.7% in Wisconsin, Trump secured the 270 electoral votes needed to carry the Presidency. Following this upset, national confidence in polling was shot.
Many explanatory theories were cooked up: the secret Trump vote, the likely voters vs. registered voters oversampling, and the inaccuracy of location sampling. However, in the 2018 midterm elections, the polling was generally accurate. Those projected to win the House and Senate did so. This created a disconnect. Those who were left reeling from the 2016 elections lost faith in the institution of polling and were biting their nails awaiting the 2020 results. Those who were confident that pollsters adjusted their formulas and corrected their mistakes to result in accurate 2018 projections were confident that the 2020 reporting was accurate.
However, both of these groups were correct (and incorrect) about the accuracy of polling in 2020. Although Biden won the Presidency with 306 electoral votes, he was projected to win by anywhere from 345-383. Furthermore, although his final national vote tally was expected to be a 5-6% lead, he will finish with a respectable 3.5%.
So… wasn’t the polling right? It projected that Biden would win and he won. No harm no foul- right? Well, not exactly. Although Biden won comfortably, he was predicted to win in a landslide. The polling was off by a 3% margin. Though this doesn’t shift the race in 2020, it made all the difference in 2016. So, it seems like the changed models and the lessons learned by pollsters in 2016 were not fully adopted.
Furthermore, there does seem to be tangible harm enacted by faulty polling. Based on the projections from polling, hundreds of millions of hard-earned dollars were poured into the South Carolina, North Carolina, Kentucky, Maine, and Iowa races. Perhaps these funds could have been better appropriated to closer races if not for the false hope created by polling.
Although polling’s errors were not too grave this cycle, the predictions for the House and Presidency have been reasonably accurate, it’s plunders have resulted in harms. Without its inaccurate projections, less Americans would have not poured their time and money into flipping Texas blue or swaying the South Carolina race. One thing’s for certain… they still have many more corrections to make.