Why the polls were wrong
By Ralph Dickerson
Prior to this year’s presidential election, polling data and pundits predicted Democrat Joseph Biden to win in a landslide. National polling data showed Biden with a 51-percent to 43-percent lead on Republican Donald Trump. The results on November 3rd surprised pundits, and showed the national polls erred significantly predicting Trump’s support.
Though several states are still counting all of the votes, current results show that on Election Day, Biden garnered 50.75-percent of the vote, which is basically in line with polling data. Trump on the other hand significantly out-performed the polling data as he garnered 47.57-percent of the vote, a result almost 5-percentage points higher than predicted. What caused this discrepancy?
Many theories abound about why Trump performed better than expected. Some say he performed better because in the last few days before the election undecided voters swung to Trump. Others say he performed better than expected because many people polled did not want to really go on record as saying they supported Trump, and lied to the pollsters. I do not agree with any of those theories.
I think the polls were inaccurate because pollsters did not accurately sample all of the demographics that make up this country. No matter the ethnic makeup of a community, three demographics define this country: urban dwellers, suburban dwellers and rural residents.
When pollsters conduct their polls, I think they basically interview only urban and suburban dwellers, which tend to favor Democrats. I think that is why the election results for Biden basically matched the polls. He was predicted to garner 51-percent of the vote, and he received 50.63 on Election Day. The discrepancy happened in Trump’s poll numbers.
Prior to the election, polls put him at only 43-percent in the polls, yet on Election Day he garnered 47.7-percent of the vote, a difference of 4.7-percentage points. In rural states, Trump strongly out-performed Biden. For example In North Dakota and Oklahoma Trump garnered 65-percent of the vote, in Idaho he took almost 64-percent of the vote, and in Arkansas and Kentucky Trump earned 62-percent of the vote. I think national polling data underrepresented the rural support for Trump, which skewed the polls strongly toward Biden, when in fact the race was much closer than the polls showed.
In the future when pollsters conduct their polls, they either need to ensure their polling data includes the proper number of rural residents in their formulas, or they need to add four-percentage points to the Republican candidate’s polling numbers to get the accurate strength of each candidate’s support from the electorate.
