Do Polls Undercount Obama Supporters?
If they fail to keep up with changes in telephone usage, they might, says one pollster.
The result, he explained, is that some compromises may be made. There are firms, such as Rasmussen Reports, that still conduct polls without including cellphone users, while others do primarily landline polls with some cellphone polling mixed in and then weight the results accordingly to draw an educated conclusion about the overall population surveyed. The problem, of course, with both of these methodologies is that they are primed to undercount two populations: young people and people of color.
Low-income people and renters are more likely than others to live in a cellphone-only home. This includes a number of black Americans. Hispanics are also more likely than other populations to live in cellphone-only homes. And then there are young people, who are by far the largest demographic group living in cellphone-only homes.
These are the demographic groups that essentially put President Obama in the White House. African-American and Hispanic voter turnout increased by 2 million people in the 2008 election, an increase credited with making the difference in swing states like North Carolina, Nevada and others. Turnout among young voters was up by 2 percent over the 2004 election -- an increase credited with making the difference in states like Indiana.
Four years later, there are countless news reports about the decrease in enthusiasm among young voters for the Obama campaign. Yet in the last four years, the number of cellphone-only-using young adults has increased significantly, but the world of polling has struggled to keep up with them. This means that polling data may not fully reflect the attitudes of young Americans this election, an election in which their votes could determine the outcome.
Paleologos, the director of the Suffolk University Political Research Center, noted that for the center's recent poll with USA Today, 35 percent of those contacted were on cellphones. The average number for the center and other institutions hovers somewhere between 20 percent and 25 percent.
Paleologos wanted to make sure the poll accurately reflected the American population, including young people. As he explained, "You have to go where people are when you're a pollster, and young people are on cellphones." He quickly added, "But that's not where they're going to stay. There are always new forms of communication. We have to keep up with technology."
We won't know until Election Day whether the polls have failed to accurately do that this election cycle, particularly where young voters of color are concerned.
Keli Goff is The Root's political correspondent.