Why pollsters didn't see Donald Trump's victory
Steve EsackContact Reporter
Call Harrisburg Bureau
ris Borick, Muhlenberg College political science professor and pollster, reflects on how many public opinion polls failed to foresee Republican Donald Trump's chances of winning Pennsylvania and a majority of other vital states on his way to collecting the minimum 270 electoral college votes needed to become the 45th president of the United States.
Q: For weeks and months, virtually all public opinion polls showed Democrat Hillary Clinton beating Republican Donald Trump by varying degrees on Election Day. For example, The Morning Call/Muhlenberg College poll, conducted Oct. 30 to Nov. 4, had her up by 4 percentage points with third-party candidates included. What happened?
A: A number of factors could have led to the discrepancy. First, turnout in rural areas and small towns across Pennsylvania was exceptional. Our likely voter model appears to have underestimated this very high turnout in many regions outside the southeastern corner of the state. In our last poll we also had a large share of voters (7 percent) indicate they planned to vote for [Libertarian] Gary Johnson. His final numbers in the state were about 2.5 percent of the vote. It is quite possible that many of those reporting intent to vote for Johnson were ultimately Trump voters who were either undecided at that point or reticent to express their true intent to vote for Trump.
Continued: http://www.mcall.com/news/local/elections/mc-pa-chris-borick-on-presidential-polling-20161111-story.html
This focuses on both the national level and state level, Pennsylvania.
