The Truth on Voter Turnout?

Voter turnout this primary season is shattering records, right, especially among the youth cohort? No so fast says Carl Bialik at the Wall Street Journal:

Turnout has been high this primary season, particularly in Democratic races. Voting experts agree on that much. What’s fodder for debate is whether high turnout is surprising these days. Academics disagree on whether there really was a steady decline in turnout rates since 1972, when Americans aged 18, 19 and 20 first got the right to vote. That was the conventional wisdom until 2001, when a paper questioned the belief. But some turnout experts maintain that turnout fell between 1972 and 2000, with a sharp uptick in 2004 amid strong feelings among members of both major parties toward President Bush.

My print column this week examines the debate, which was spurred in part by a recognition that non-citizens are a growing portion of the population, and shouldn’t be included in turnout calculations because they aren’t eligible to vote. “The real thing to look at here is that the population of the country has a larger and larger proportion of people not eligible to vote because they are not citizens,” Ray Wolfinger, professor emeritus of political science at the University of California, Berkeley, told me. This has many political implications besides just turnout rates, of course. One of these is that voters in states with relatively high proportions of non-citizens — such as California, 16% non-citizens in 2000, compared with 7% nationwide — have disproportionate clout, because Congressional seats and presidential electors are apportioned by population regardless of citizenship.

The turnout debate is also spurred in part by uncertainty in these numbers. While the Census Bureau takes pains to ensure respondents are honest about their citizenship status, immigrants may remain wary of government questionnaires (I wrote about the fuzziness of immigration stats in 2006). Turnout rates also must be corrected for felons who have lost their right to vote, another group that I’ve written is hard to count. Meanwhile, the Census Bureau doesn’t count Americans who live abroad. Michael McDonald, a turnout researcher at George Mason University, has tried, using information from consulates and the military, but says federal data in the area has grown more sparse.

What do you think? Have turnout rates declined? Are there long-term trends at play, or merely unique features of each election? Will turnout increase or decrease in November, compared to 2004? Please let me know in the comments.

Further reading: Mr. McDonald publishes his turnout estimates here. Walter Dean Burnham, a pioneer in the field, published his electorate numbers in December in the Journal of the Historical Society, which has promised to share the numbers free on its Web site. Curtis Gans, another leading turnout expert, told me he has nearly completed his own archive of turnout figures.

Hmm, what do I think?

I think, for one thing, that states with high rates of illegal immigration are getting a boost in their congressional and Electoral College representation based on their disproportionate numbers of resident aliens, AND this should be a big topic in the immigration debate but it hasn't. It reminds me of the Three-Fifths Compromise at the Consitutional Convention, whereby blacks in slaveholding states would be counted in total population tallies at three-fifths their total number for purposes of representation in Congress. The practice was inherently undemocratic, and it took a Civil War to fully eliminate all aspect of that peculiar bargain.

Besides that, Bialik doesn't quite say it here, but how is voter turnout being measured? The proportion of the eligible electorate that votes on election day?

Bialik actually does address this in his print column:

Turnout rates are easy to define: You divide the number of people who voted into the number of people who could have voted. But they're hard to calculate in the U.S. The 50 states and the District of Columbia run the show, and they generally don't keep an accurate count of eligible voters.

In other words, turnout numbers are inherently unreliable because of the potential flaws and variability in state level counting.

Not only that, Bialik's got a full passage on the unrepresentativeness of Census Burueau counting methods:

Up through the 2000 elections, researchers generally divided total votes into the U.S. Census Bureau's count of the voting-age population. That number, however, excludes Americans living abroad, who can vote. And it includes those declared mentally incompetent and felons, who are not allowed to vote in some states. Most important, it also includes noncitizens, a nonvoting population that has grown much faster than the general population -- to 18.6 million. That's more than a fivefold increase from 1970, compared with an overall population growth of just 38%.

What this signifies, essentially, is that the numbers on turnout that political science professors are throwing out every semester to their students in American Government 101 are essentially meaningless, or at least, the data requires the accompanyment of a lot of methodological disclaimers.

As long as the same data counting methods are used from election to election, then we should have some comparability. If turnouts increased over the years, the numbers would have to be adjusted for the proportion of eligible voters, felons, and what not, but other than that, not bad. Right?

Wrong.

Things don't sound so good for voter turnout studies - we might as well be looking at crystal balls (it's not just illegals, who are today a larger share of the population than in the 1970s, which means we can't compare turnout rates over time withhout factoring out those who're ineligible; and don't even get going about those in prison, paroled, or what have you, who are 1 in 100 of the population, many of whom are disenfranchised).

This means that rather than using hard, rigorous statistics, professors will have to refer to all the Obamaniacs the Illinois Senator's managed to turnout out in the Iowa and Nevada caucuses.

Good grief, Charlie Brown!!