Tom Foley’s double-digit Republican primary victory over John McKinney last week wasn’t a huge surprise, particularly to those who follow the Quinnipiac Poll.
Since June 2013, Hamden’s Quinnipiac University Polling Institute correctly identified Foley as the overwhelming favorite for the Republication nomination. The Greenwich businessman’s comfortable 57-43 percent victory showed Quinnipiac had taken the correct pulse of the Connecticut electorate.
Accuracy is a hallmark for any reputable polling organization, particularly Quinnipiac, which has risen to national prominence in recent years for its proclivity to correctly predict election outcomes and public sentiment on key political issues.
Frequently cited by media outlets, public officials and researchers, the Q Poll’s findings not only drive the national debate at times, but can be the spark that changes political campaign strategies.
“It has become the gold standard,” said Chris Healy, a Connecticut political consultant and former chairman of the state Republican party. “It has been pretty accurate over the last couple of election cycles … It is obviously one poll that a lot of people look to.”
The Quinnipiac poll was founded in 1988 by marketing research professor Paul Falcigno, originally only examining Connecticut issues and political races. Since then, it has expanded its polling to nine states — New York, New Jersey, Ohio, Florida, Pennsylvania, Colorado, Iowa, and Virginia — which were strategically selected so the poll could accurately predict presidential elections, said Doug Schwartz, the Quinnipiac poll director.
The polling institute, with its 10 full-time employees, receives all of its funding from Quinnipiac University and does not accept outside financial support to maintain its political independence, Schwartz said.
“I view what we do as a public service,” Schwartz said. “It is important that we are an independent poll as a counter to the partisan polls out there.”
Tracing its path to notoriety
The Q poll first gained national attention in 1998 after accurately predicting Chuck Schumer’s New York U.S. Senate victory over three-term Republican incumbent Al D’Amato. Two years later, Quinnipiac correctly called Hillary Clinton’s New York U.S. Senate victory.
But the poll didn’t become a national player among the political chattering class until 2001, when it started polling presidential elections and concentrated on a new way to predict outcomes.
Its strategy, which was shaped by George W. Bush’s 2000 presidential election victory over Al Gore, was to focus on the swing states of Florida, Ohio, and Pennsylvania. Looking at the history of national elections, if a presidential candidate won two of those three swing states, he won the race.
“The reporters and the campaigns are really focusing on a certain number of states,” Schwartz said. “Polling swing states is something we feel separates us. That really helped us in terms of our national reputation.”
The Q poll’s focus is not just on elections. Schwartz regularly works with former journalists in each of the nine states it tracks to poll local issues, ranging from the legalization of marijuana in Colorado to tracking Ohioans’ public sentiment on LeBron James’ return to the Cleveland Cavaliers. The aim is to keep the Quinnipiac brand relevant year-round, Schwartz said.
“We poll regularly in all of our states,” Schwartz said.
Roy Occhiogrosso, managing director of Global Strategies Group in Hartford, which ran Gov. Dannel P. Malloy’s 2010 gubernatorial campaign, said issued-based polling is particularly important for lawmakers, some of which believe their decisions and votes must reflect the majority of their constituents’ opinions.
“For a lot of elected officials, it is harder for them to take a position that is not popular in the polls,” Occhiogrosso said.
Therefore, Quinnipiac’s polls can help government bodies like the Connecticut General Assembly make decisions, Occhiogrosso said. An issue like gay marriage, for example, which wasn’t a popular cause among state legislative bodies 10-15 years ago, is more accepted today because polling shows a majority of voters in many states are okay with it.
With a myriad of polling groups competing for attention across the country, the ones that stand the test of time and gain true notoriety must be accurate on a consistent basis, Occhiogrosso said.
“Quinnipiac is clearly the most prominent one that is out there,” Occhiogrosso said.
The Q Poll strategy
To maintain its accuracy, Quinnipiac takes steps like calling people’s mobile phones to ensure it gets a true sampling of voters, Schwartz said. Polls that don’t call cell phones are secluding 40 percent of the population, particularly skewing against young people.
“You have to give everyone in the population equal chance of being selected,” Schwartz said. “That is the theory that underlies everything we do.”
At its Hamden office, the Quinnipiac poll has 150 work stations, where interviewers make calls between 6-9 p.m. Most of the interviewers, who are paid an hourly wage, are Quinnipiac students, although some are local community members who have been doing the work for 10 years or more.
Each poll surveys between 1,000-1,200 people, which gives the results a 3 percent margin of error. It usually takes five to seven days to complete a survey, although respondents are asked about more than one topic, so Quinnipiac typically issues two or three results per survey.
Respondents are selected at random. A phone bank dials the numbers, and the interviewers request to speak to whomever in the household has the next birthday. Once the interviewer reaches the right person, which can be a timely and difficult task, the survey’s cooperation rate is about 50 percent (it’s 60 percent in Connecticut where Quinnipiac is more well-known).
Interviewers will make 150-200 calls per three-hour shift and typically perform five or six interviews during that time. They will call a phone up to four times over the course of several days until receiving a “yes” or “no” from the interview target.
To maintain its accuracy, Quinnipiac studies demographics — gender, age, and education levels — of each state it polls. If a state is 52 percent female, for example, then the poll respondents should be 52 percent female.
“We have been very accurate. We have a good track record,” Schwartz said.
Polls, however, can have their downside in the political process, Healy said. Voters can look at poll results, see who the leaders are, and decide not to participate in an election because they feel their votes won’t influence the outcome.
That is why Quinnipiac’s independence and accuracy is tantamount, Schwartz said. If polls can sway public opinion and campaign outcomes, then results need to come without bias.
“We have no axe to grind, and this is an important check on groups that sponsor polls,” Schwartz said.
This year, the Quinnipiac poll has four contested U.S. Senate races and seven governor’s races it will track in its nine states, including the gubernatorial rematch between Foley and Malloy in Connecticut.
“It will be a good race. It will be a close race,” Schwartz said.
