Pew Sees 66% More Voters As Likely to Vote than Actually Voted in 2006

Just starting to dig through the latest Pew poll, but this preliminary statistic really jumped out at me:

The survey by the Pew Research Center for the People & the Press was conducted Aug. 25-Sept. 6 among 2,816 registered voters, including 2,053 voters considered the most likely to vote on Nov. 2.

Crunching those numbers, Pew is characterizing 72.9 percent of registered voters as likely to vote -- a fairly staggering number.

Back in 2008, MSNBC estimated 184 million Americans to be registered to vote. An AP estimate put the number at 187 million voters. Based on these figures, and Pew's projections, somewhere between 134 million and 136 million voters might turn out in November. For reference, only about 131 million voters filled out ballots for President in 2008, and 123 million filled out ballots in House elections that year. The past two midterms saw just 81 million voters and 75 million voters in 2006 and 2002, respectively. Putting this all together, turnout based on this projection would be about 66 percent higher than in 2006 and and about 80 percent higher than in 2002.

Considering that the MSNBC and AP numbers may overestimate registered voter totals -- the estimates were based on reports from Secretaries of State offices around the country, which tend to include some duplication -- perhaps it would be better to look at numbers from the Census. According to the bureau (.xls), there were roughly 146 million registered voters in 2008. Plugging in Pew's projection to these numbers, there are about 106 million likely voters. This number would be less than voted in 2008 -- but still 31 percent higher than voter turnout in 2006 and 42 percent higher than voter turnout in 2002.

Perhaps somewhere between about a third and two-thirds more voters are going to turnout this November than turned out during the last midterm (which would represent an even greater increase from the midterm before that). But it's also possible that pollsters are making educated guesses about voter turnout that aren't exactly on the mark, classifying some eager respondents as likely voters when they are actually just the types of people willing to sit for 10 or 15 minutes on the phone with a stranger.

For what it's worth, if the electorate on November 2 looks anything like the pool of registered voters (as opposed to those deemed by Pew to be likely voters), the Democrats should have an advantage -- 47 percent to 44 percent, to be precise.