Southeast Missouri State University student publication

Former Southeast assistant professor conducts political polls

Monday, October 22, 2012
Dr. Will Miller is a former assistant professor of political science at Southeast Missouri State University. Submitted photo

Dr. Will Miller is a former assistant professor of political science at Southeast Missouri State University, and he now works as an assistant professor of public administration at Flagler College in St. Augustine, Fla. He has worked on multiple political campaigns and currently conducts polls, as well as advising and management.

Q: What exactly is the kind of work that you do or have done during political campaigns?

A: I have served a variety of roles for various campaigns in my life. I started out mainly doing opposition research, which unfortunately many view as the dirty side of politics. I became very good at finding out about every aspect of a candidate's life and beliefs and determining strategic ways by which to utilize that information. Eventually that grew tiresome, and I began working in polling. Today, I mainly work as a pollster and provide strategic analysis for candidates and issue groups based on that data. But I also will still provide overall campaign management for candidates I truly believe in.

Q: What kinds of polls do you conduct, and what do these polls tell you?

A: I've conducted a number of different polls. The primary type of poll we conduct at the local level is a benchmark survey where we try to gather as much information as possible for a candidate regarding voter attitudes and opinions. Most local officials can only afford one or two polls, so we pack as much information as possible into them. These polls help establish priority issues and potential marketing strategies. With higher stakes races, we focus on tracking polls. These are shorter and occur regularly and are used to look at trends in approval and support. We also will conduct quick polls after major speeches or campaign events to track what voters learned and how they respond. Off the campaign side, I do community polls where nonprofit and civic groups buy questions to determine citizen awareness and attitudes toward their programs and goals.

Q: How do you conduct a poll to ensure that it gives you accurate information?

A: To begin, as campaigns, we use different polling companies than you hear about in the news every day. Companies like Gallup and Rasmussen are doing commercial polls for general interest and news organizations. Campaigns hire myself and other pollsters to get information for only their use. We use various, different call centers and phone research companies to accomplish the goals of the candidate. To ensure accurate information, we pretest our surveys. In other words, we field some sample calls and make sure the questions we ask are actually measuring what we want to. Further, we are extremely careful and deliberate with sample selection. Major news polls you read about today -- especially the major news organizations -- simply allow respondents to state they plan to vote and count that as a likely voter. We, on the other hand, actually go through a battery of questions and determine from that whether someone is actually going to vote. We also call to speak to an individual. Most major news polls call a phone number or address and speak to a member of the household. We know, however, that there can be vast differences within a household regarding political opinions. If we call to talk to Brian Asher, it's because demographically, we want to talk to you.

Q: How can an interested person determine whether or not poll results shown in the media are accurate?

A: Most major media sources have become better about giving citizens a glimpse into their procedures. First off, always pay attention to margin of error. If the difference is less than the margin of error, it's technically a toss-up. Second, if there isn't information provided about how the sample was selected, I'd generally disregard the poll. Most importantly, know the perceived bias of the polling agency. Rasmussen, for example, tends to oversample conservative voters while CNN tends to underestimate them. It doesn't make their results untruthful; it just means you have to read between the lines a little bit more.