Polling Young Voters - Difficulties, Best Practices, and Consequences

Cross posted at Daily Kos as part of the Kossacks Under 35 weekly diary series.

Between the Obama campaign's questionable assertions of a commanding lead among young Iowans, and competing claims by the campaigns and the press corps that young voters are either a powerful but hidden voting block or an inconsequentially miniscule portion of the electorate, polling numbers on young voters have become a common topic on this blog and within the political press. Youth turnout has been increasing for three straight years, and the sheer size of the Millennial Generation make today's young voters a demographic that cannot be ignored. Yet young voters remain the least polled segment of the electorate, and their views are highly underrepresented in both state and national polling data.

There are a number of reasons for this. Polling methodologies, lifestyle barriers, and, even in what is shaping up to be a $200 billion dollar campaign cycle, prohibitively high costs all present barriers that preclude the opinions of young voters from what is the most common method of gaging public opinion in our political debate.

Methodologically, young voters are often casualties of "likely voter screens," which seek to narrow the polling sample down to those most likely to turnout at the polls. What determines a "likely voter" can change from poll to poll, as different pollsters deploy different methods to filter out portions of the electorate. In some instances this is determined by asking questions about how closely a respondent is tracking the current race, how determined they are to vote, and even the strength of their support for one of the current candidates. In other polls, a voter cannot be a "likely voter" unless they have already voted in an election or participated in a caucus. For many 18 - 21 year olds, the upcoming election is often their first, a situation which automatically disqualifies them should a pollster come calling.

These methodological barriers assume that a pollster can even get through to a young voter. According to Jim Gerstein of Democracy Corps, most pollsters call in the early evening, at a time when younger voters are least likely to be home. Even that assumes that young voters have a landline, something that is frequently not the case as more and more young voters are using cell phones as their primary (and only) means of communication. According to research by the Pew Center for Internet and American Life, American's under 30 are more likely than those over 30 to use their cellphone instead of securing a landline. A study by the New Politics Institute, suggests that young Hispanics do so at twice the rate of white youth.

Not only are the voices of these voters being shut out of this important part of the political process, but this situation creates a challenge for pollsters. Pollsters determined to survey young voters may opt to purchase lists of cell phone numbers, but this is often a more expensive option than random digit dialing (the normal polling method). Aside from the money spent on the list, the rates of refusal among cell phone users are high, meaning that a pollster needs to call many more numbers before attaining an appropriately representative sample size (typically 1,000 respondents between 18-29 in order to accurately gage youth opinion). That, too, drives up costs.

One solution to the problem of rising costs is to collect data via the web, however this presents troubling digital divide issues that can be just as exclusionary than RDD polling.

There are a number of best practices emerging to overcome these challenges in the field of youth polling. Gerstein notes that Democracy Corps is now employing a hybrid polling model (pdf) that incorporates both random digit dial in conjunction with cell phone polling and web surveys. These types of hybrid polls are designed to avoid the challenges and biases presented by RDD polling alone, while avoiding the digital divide issues that plague web-only polling methods. Unfortunately, Hybrid polling is also an expensive project, and not many pollsters are willing to take up the costs or challenges associated with such an involved survey.

Campaigns seeking a snapshot of the youth electorate frequently aggregate 5-10 individual surveys to attain statistically reliable information about young voters, a process that is also employed by organizations like Young Voter Strategies/Rock The Vote, which issues bi-monthly reports on youth opinion culled from multiple polling sources. That type of third-party analysis by groups like RTV or RT Strategies is also becoming the norm. With the costs of reliable youth polling increasing, it may fall to these third party companies and nonprofits to provide the kind of polling data that campaigns and media outlets can no longer afford.

It's worth noting that these are not just young voter problems. As more and more Americans switch abandon their landlines and use their cellphone as their primary means of communication, it will become more expensive to poll the entire electorate. Young voices are being shut out of the debate now, but in 5, 10, 15 years, as Millennials age into the electorate, this is going to become an issue for all polling and all demographics.