Gauging Public Opinion

On Super Tuesday 2016, Democrats and Republicans in Virginia and 10 other states headed to the election polls to cast a ballot with their choice for a party nominee.

Dr. Harry Wilson, a professor in Roanoke College's Department of Public Affairs, watched the results closely, for two reasons. First, Wilson is a senior political analyst at the Roanoke Valley's CBS affiliate, WDBJ7, and would be called upon that evening to provide punditry about what it all meant. Second, the Virginia primary marked the first test of 2016 polls conducted by the College's Institute for Policy and Opinion Research (IPOR).  

This election year is shaping up to be IPOR's biggest so far in its history, and Wilson, who serves as the institute's director, is bubbling with excitement.  

"So many things in this election I am not happy about, so it's difficult for me to say I'm excited; 'fascinated' is a better word," Wilson says.  

Along with IPOR's primary poll, which showed Hillary Clinton and Donald Trump with substantial leads in the Democratic and Republican primaries, the institute plans to conduct five general election polls: one in August and two each in September and October.  

Additionally, IPOR conducts quarterly consumer sentiment and real estate polls of Virginia, delivering index numbers that measure confidence in the economy compared to in the past.  

As political observers and economists continue to increasingly rely on a steady diet of polls showing how people feel about the world, Roanoke College and IPOR have taken on a crucial role in Virginia and beyond. 

1983: BEGINNINGS

IPOR began as the Center for Community Research in 1983, when it was run by sociology Professor Gregory Weiss. Each fall, the center partnered with what was then The Roanoke Times & World-News for an annual community survey. Each spring, it took on service-oriented research projects, working with government agencies and non-profit organizations on program analyses and surveys.  

The community survey was offered as a fall class for students, who spent three months in the field going door to door to speak to Roanoke Valley residents. The Roanoke Times & World-News devoted the front page of its Sunday Horizon section to publishing the results, presented as a snapshot of the valley's life.  

Wilson arrived at Roanoke College in the fall of 1986, three years into the life of the polling center. "Dr. Weiss was moving onto other things and I was interested in this work," says Wilson, who earned his doctorate at Rutgers University, which maintained a polling operation. Intrigued by the idea of directing a polling center, "it was a great opportunity that I may not have had at another institution," he says.  

By the early 1990s, The Roanoke Times (by then, the paper had dropped "World-News") had re-thought its arrangement with the Center for Community Research and discontinued its partnership. Wilson took the loss of the center's sponsor as an opportunity to try its hand at something new: political polling.  

In 1994, when Col. Oliver North, a Republican, challenged former governor and incumbent U.S. Sen. Charles Robb, Wilson moved to make his idea a reality. That first year of political polling ran on a shoestring.  

The Roanoke College center conducted only one poll that year, and Wilson was not pleased when he saw the results, which showed Robb with a lead over North, though it fell within the margin of error. Every other poll showed North with a lead. Wilson suggested to his research partner, John Keyser, that perhaps they shouldn't release their results.  

"John sat me down and said, 'Can you tell me any single thing that we did wrong? Tell me logically. Where did we screw up?'" Wilson says. "I looked at him and said, 'I don't think we did screw up.' He looked at me, shrugged his shoulders and said, 'Well, then let's release it.'"  

The poll's release was met with skepticism, from the campaigns of North and independent Marshall Coleman. The North campaign spokesman said it didn't match the GOP's internal polls, and Coleman's camp dismissed it altogether, with his spokesman telling The Roanoke Times, "I don't think any of that is true. There's nothing to show that kind of movement. It just isn't there."  

Roanoke College was vindicated by a series of polls that followed it, however, as well as the final result, when Robb narrowly won re-election with 46 percent of the vote to North's 43 percent.  

"There were raised eyebrows, but we were correct," Wilson says. "Other polls fell in line behind us afterward. I was really pleased. It told me we really did know what we were doing."  

Roanoke College polls became an institution over the following years, tracking Virginia as it moved closer and closer to the center of the national conversation.

Sara Gubala '98, a political science/Spanish major, while at Roanoke, got involved with the center through a class. She joined during the 1997 gubernatorial race.

"Initially, it was frightening," Gubala says. "Calling people was not something I enjoyed doing. I distinctly remember Dr. Wilson telling me, 'Either call and complete surveys or your grade will suffer,' so that made it very real to me. Once I got past the initial awkwardness, it was OK."  

Gubala participated in several more surveys. Her involvement at the center led her to become more involved in politics, attending the Virginia Republican convention and interning with U.S. Rep. Bob Goodlatte of Roanoke County. Today, she teaches political science at Lamar University in Beaumont, Texas, and she still carries her experience at Roanoke College with her.  

Recently, "I was on a local political show. I was asked about a poll for the Republican presidential candidates, and I was able to give some insight into this poll, which had been incorrect in another state," Gubala says. "It seemed like suddenly, everything had come full circle. Here I was talking about a poll in the same way Dr. Wilson talks about polls."  

2008: EXPANSION

The Center for Community Research dabbled in the 2004 presidential election, but in 2008 it jumped in with both feet. That presidential election completed Virginia's transformation into a much-coveted battleground state, where either party can win. Victory there may mean winning nationally, so candidates devote considerable time, money and energy into courting voters. Battleground states also command the attention of national media outlets, which crave polls.  

With the growth of Virginia's influence and the subsequent growth in the number of polling operations watching the state, Wilson decided that the Center needed to scale up and branch out. He approached Roanoke College President Michael C. Maxey and Dr. Richard Smith, vice president and dean of the College.  

"President Maxey had always seen the value of polling and of IPOR, so when he became president I went to him with a proposal to expand," Wilson says. "Dean Smith was also supportive. It fit in well with his philosophy of what a liberal arts college could do. Both were and continue to be very strong supporters."  

President Maxey says he recalls that meeting as a moment where we saw clearly Roanoke's role in adding value and insight into national affairs.  

"IPOR gave us the outlet for relevant value in that arena," he says. "Of course, we had great confidence in Dr. Wilson and other faculty here to deliver a high quality, nonpartisan polling endeavor."  

With a beefed-up budget, Wilson began expansion in 2009. The following year, he hired Dr. David Taylor, associate professor and chairman of the College's department of mathematics, computer science and physics, to become the center's associate director for systems support and analysis.  

Taylor came on board for his mathematical prowess, but as the center scaled up its software, his programming expertise played an important role as well. The strengths of Wilson and Taylor complemented one another, adding a new dimension.  

"We work together well," Taylor says. "He's been doing political science for years. He's got books out there, and he knows issues and how to talk about issues well. He has the connections and the name to go along. What I bring is not any of that; it's the opposite. I have a strong math background, I know enough stats that I can work with the numbers and not worry about them."  

The center hired a call center supervisor, and the call center was moved from a room in West Hall to the former Farmer's National Bank Building down the street, where it was outfitted with 20 stations, computers and automatic dialers.

With all of the changes, a new name seemed in order, too, and in 2010, the Center for Community Research thus became the Institute for Policy and Opinion Research, or IPOR for short.  

"It's dramatically different today than it was from where we were back in '89," Wilson says. "It's a very different operation. Today, it's more like a small business." 

2011: ECONOMICS

"Business" not only describes how the institute functions, but a whole new focus of its survey operations.  

As part of his expansion planning, Wilson cold-called Don Levy, director of the Siena College Research Institute (SRI) in Loudonville, N.Y. Both Roanoke College and Siena College are small, religiously affiliated liberal arts colleges, so Wilson saw a kinship and similarity in mission. Siena is positioned in a much different atmosphere, given its geographic proximity to Marist College and Quinnipiac University, both of which are known for their polling operations.  

Wilson visited Siena College on an overnight trip, gathering ideas and information. Levy suggested that Roanoke College might begin measuring consumer sentiment in Virginia. Thus began a relationship that also includes Rutgers University in New Jersey.  

"Our relationship with Roanoke College and Rutgers is really an exciting academic relationship," Levy says. "We're fully transparent with each other. We share questionnaire development. In many ways, it's the best of what the academy [higher education] ought to offer-peers working with one another to enhance each other's work and generate important information not only for the academy but the general public."  

To launch that survey at Roanoke College, however, Wilson needed an economics expert. He turned to Dr. Alice Kassens, Roanoke College's John S. Shannon Professor of Economics.  

Starting in November of 2011, IPOR conducted twice-a-year consumer sentiment and real estate surveys using a real estate index developed by Siena College and a national consumer sentiment model created by the University of Michigan. Over a series of 50 standardized questions, respondents are asked how they feel about their economic status in the present and how they anticipate it will change in the future. Their responses generate an index number that is used by economics and market analysts around the country.  

Kassens remembers that, like Wilson, she felt nervous during that first run in 2011. "I was worried we'd make a mistake and get wacky numbers, but we didn't," she says.  

To double-check her accuracy, Kassens later ran an analysis of those who took the longest to respond, concerned that they may have answered differently from those who responded immediately. To her relief, she found no statistical difference between the two, lending evidence that the methodology was sound.  

In 2013, IPOR expanded its consumer sentiment and real estate surveys from twice-a-year to quarterly, with the eventual goal to run them monthly.  

Additionally, IPOR works with its partners at Siena College and Rutgers University to take coordinated consumer sentiment polls in Virginia, New York and New Jersey, respectively.  

Every October, Kassens presents results from the consumer sentiment surveys to Virginia Gov. Terry McAuliffe's Joint Advisory Board of Economists, of which she is a member, to help the state administration begin the process of building a budget.  

Kassens has taken on a bigger role in regional media, too, appearing on WVTF radio and elsewhere to discuss the state of the economy. That, in turn, has led her to hone her public relations and communications skills.  

"I find myself always thinking about what questions people will ask me," Kassens says. "I try to have an answer for them. They're usually simple but excellent questions, things like 'What does this mean for me?' You need to get away from all the economic jargon, and just get down to basics. It's been fun, and it's certainly made me a better writer and economist."  

2016: THE FUTURE

As IPOR heads into 2016, its economic surveys have become a crucial component of its work. Its political polls continue to drive media interest, however, particularly in an atmosphere where analysts and observers crave up-to-the-moment information on how voters are responding to candidates. Unlike the economic surveys, the political campaigns eventually do result in a final, definitive figure, measured on Election Day. Although the polls capture a moment in time prior to the election, observers often measure them against that final figure.  

Wilson and IPOR suffered a bump in the road during the 2012 election, when its modeling showed Republican presidential candidate Mitt Romney with a lead over Barack Obama. When Obama won Virginia and the presidency, critics came down on the institute. Wilson acknowledges that he made a couple of mistaken assumptions about the shape of the electorate that year.  

For 2016, Wilson and Taylor have tweaked their approach to incorporate political parties-a challenge, though not impossible, since there is no party registration in Virginia.   "We don't just sit back and do the same things over and over again," Taylor says. "It's not just about calibrating results."  

The changes seemed to work. IPOR's Super Tuesday poll showed Clinton and Trump with leads, and each finished first in their respective primaries. As candidates move through primary season toward the party nominations and general election, IPOR continued to refine its approach.  

The process provides a real-time, hands-on education for student workers such as Marie Brown '18, a political science major.  

"I participate in both the consumer sentiment and political polls," Brown says, but "I prefer to survey people with the political poll because I love hearing what the people in my state think about presidential candidates or different policies that are being debated."