Eagleton Institute has polled and analyzed New Jerseyans’ opinions since 1971

The Eagleton Center for Public Interest Polling, home of the Rutgers-Eagleton Poll, was the nation’s first university-based state survey research center when it was established in 1971. As it releases its milestone 200th poll, the center’s dedication to reporting objective, scientifically sound information about public opinion on politics and policy in the Garden State is widely recognized. Equally important is its legacy of involving students through an internship program that allows students to create survey questionnaires, conduct interviews and analyze data for public release under faculty supervision.

Four decades since the first Rutgers-Eagleton Poll (originally, the New Jersey Poll), polling is ingrained in the fabric of American politics –  especially as polls now help determine who will participate in nationally televised presidential primary debates. David Redlawsk, professor and center director, and Ashley Koning, assistant director, discuss with Rutgers Today polling's role in the political process, challenges facing pollsters in an increasingly digital era and when polls should not carry the weight they do.

Polling
The Rutgers Eagleton Center for Public Interest Polling is releasing its milestone 200th poll since it was established in 1971.

How does polling work, and is it more art or science?

Ashley Koning: Polling is a scientific, quantifiable way of understanding public opinion. It is not meant for predictive purposes but rather to capture a snapshot in time about what people are thinking. It works because of what we call “random sampling,” which allows us to generalize to an entire population by interviewing just a fraction of it. Results are stated within a margin of sampling error – the “+” or “-” number that describes the range within which the actual number falls if we had interviewed the entire population. While polling is rooted in statistics, it is also very much an art. Pollsters must make subjective decisions about the wording and order of their questions, and how they collect, code, analyze and interpret data. Their interaction with respondents, and the personal characteristics of each, can also influence opinion.

How can average Americans become knowledgeable consumers of polls? How can they trust a poll’s accuracy?

David Redlawsk: Legitimate scientific polls are accurate subject to a range of possible errors (like sampling, measurement, nonresponse and coverage errors). Smart consumers should look for polls that fully disclose details about what they do, how they do it and their sponsors. If there is no disclosure, you can ignore the poll. The American Association of Public Opinion Research has established a taskforce to emphasize transparency on these matters. Moreover, consumers should understand that any one poll is just an estimate. In many ways, looking at a range of polls on a single topic gives a better sense of public sentiment on an issue.

What challenges does the polling industry face today?

Koning: The increasing prevalence of modern technology, including the Internet and cell phones, and the declining rates of willing respondents, are important issues for the survey research field. Online polls are cost-effective, but concerns persist about their coverage (not everyone uses the Internet); demographic skew (users are younger, more educated and more affluent); and representativeness (there is no way to randomly sample email addresses). With the increased popularity of cell phones and the decline in landline usage, telephone polling has become more difficult, expensive and time-consuming. Cell phone-only users also tend to be younger and less affluent, which can skew results. Finally, while researchers have not yet encountered significant problems in data quality due to falling response rates – now in the single digits – they are concerned about what this decline means for the ability of polls to give us results that can be generalized to the public at large in the future.

Polls seem to play an increasingly prominent role in today’s political environment. Are polls designed for this?

Redlawsk: Using national polls to determine participants in the various Republican primary presidential debates is a terrible idea. Polls assume that people are in fact thinking about the topic being polled. But the presidential nomination process is not a national process, and there is no current national campaign, except for in a few states like Iowa, New Hampshire and South Carolina, where campaigns have been intense and voters are engaged. In most of the country, a large percentage of respondents are paying little attention and are unengaged this early on. National polls therefore give a warped sense of what is happening, and using polls to determine “main stage” candidates misrepresents the state of the campaign. Nevertheless, despite the current controversy over polls and the GOP debates, polling still is a vital check on the democratic process, not only for tracking election trends but also for many issues that never come up for a vote.

Media Contact: Steve Manas, 848-932-0559, smanas@ucm.rutgers.edu