November 6, 2018
By Vani Sharma and Tony Sjodin

During I Block on Wednesday, Oct. 17, Spencer Kimball, an Assistant Professor at Emerson College, spoke to Lexington High School students about the political polling process. The event catered to students studying statistics and government, but was open to any student interested.
Kimball discussed the idea of “polling literacy” and how journalists and the public often read polls incorrectly.
“That’s why we come and speak to campuses, because we want you to be polling literate. We want you to be able to look at these results and have critical thoughts about what they really mean,” Kimball said.
Kimball explained what steps pollsters take to design questions, collect responses, and analyze data. Toward the end of the I Block, students asked him questions about his work.
“Seeing the relevance of what we are learning ... and the practical applications of it for predicting election outcomes, or for that matter predicting product launch and people’s opinions, is what statistics is trying to teach students to be able to do,” Christopher Doucette, a statistics teacher at LHS, said.
Kimball has visited LHS in previous years as one of the political speakers that Larry David, a U.S. History and U.S. Government and Politics teacher at LHS, has invited to speak to students.
“We’ve learned a lot about polling in AP Gov class, but I think the I Block really hit home the two main challenges that pollsters face: choosing a good sample and making an unbiased poll. Minute little details like the order of a poll’s questions or the voice used to read the questions can have a large effect on a poll’s data,” David Arena, a senior who attended the I Block, said.
The 2016 election was among the topics discussed, including questions about how so many pollsters predicted the results incorrectly.
“One interesting thing I learned from the talk is that most polls are susceptible to ‘herding,’ where pollsters try to match their findings to previous results so they can seem more accurate. Perhaps this is why Trump’s win came as such a surprise: pollsters initially predicted a Clinton victory, and most polls after that tried to lean towards Clinton to match,” Arena said.
Kimball spoke about the margin of error in polling samples, and explained how collecting and analyzing data can affect the margin of error.
“I hope that students take away an interest for the variety and power of things that statistics can do. He’s focusing in this example specifically on politics and giving people the advantage of knowing what the hot-button issues for voters are, but the power of asking good questions and getting responses that are reflective of the population’s attitude can give people an advantage in business, in politics, and lots of other situations,” Doucette said.
Comentarios