"Polling peoples’ opinions give office holders a better sense of public priorities and helps to explain positions on various issues.”
- Brian Schaffner
During the fall 2010 elections, Schaffner’s group of UMass undergraduate students conducted the only statewide exit poll, which proved highly useful in the post-election analysis of health care reform. After such high-quality results from the first class, Schaffner decided to offer it as a regular fall course during election years. Now he is gearing up for the exciting political season ahead.
The undergraduate political science course is framed around designing an exit poll study, taking the sample on Election Day and analyzing the data in an intensive final paper. In the months leading up to November, the pollsters-in-training gather boardroom style in the classroom to learn about analytic tools and public opinion research. Most of them are eager to be in the class because the word has spread—previous students have found the class incredibly rewarding.
“They felt like they were out actively generating knowledge, which was exciting for them. In my view, it’s exactly the kind of thing students should be doing at a research university like UMass Amherst,” Schaffner explains.
The students talk excitedly, citing the hands-on aspect as a major motivator in their decision to take the class—many are enthusiastic about learning “real-life skills” they can take with them into their professional careers. Each student designs his/her own survey based on the topic they plan to cover in their final analysis. A previous student, for instance, chose to ask voters if they had read the voter information packet sent out by the state and concluded that those who did were more informed on the issues. With Schaffner’s guidance, the group refines their research questions and learns how to take unbiased samples. To prepare for the main event, Schaffner provides them with a foundation of public opinion literature and teaches them to use statistics software. On Election Day, 25 students spread out across the state to poll 12 precincts. The data they obtain is then incorporated into each of their papers, which Schaffner plans to help them publish as co-authors.
In educating students, Schaffner emphasizes scientific methodology. He explains that any minor detail can affect survey results, down to tonality, phrasing and “skip pattern.” A skip pattern refers to the premeditated pattern with which each student decides to approach exiting voters (every tenth person, for example). Since they cannot possibly survey every voter, the established pattern ensures the study remains controlled and neutral.
For Schaffner, objectivity is extremely important. In order for a poll to provide the desired civic service, it must accurately represent public opinion. In an ideal democracy, public opinion drives public policy. In that equation, polls are a way to test whether elections are serving their purpose and provide a deeper understanding of voter patterns.
“Elections are a really blunt instrument for making elected officials accountable. In an election, somebody wins and somebody loses, but it’s not clear why one side won and one side lost or what the people want to have happen. Polling peoples’ opinions give office holders a better sense of public priorities and helps to explain positions on various issues,” Schaffner says.
On Election Day, Schaffner reports the results-in-progress to the media, which is all-too anxious to get accurate data quickly—the students’ results have been used by several local news outlets and even the Huffington Post. And with a heated presidential election, medical marijuana, right-to-die, and the most-watched senate race in the country looming in Massachusetts, Schaffner is confident the press will watch the polls closely once again.
Amanda Drane ‘12