Home | Opinion | Columnists | Mark Mellman

Registration and turnout 1st things 1st

Most campaigns relegate voter-registration efforts to second-class status.

“Can’t afford it. Need to put the money into GOTV.” “New registrants don’t vote.” These are among the conversation-enders I’ve heard on the subject.

Most campaigns relegate voter-registration efforts to second-class status.

“Can’t afford it. Need to put the money into GOTV.” “New registrants don’t vote.” These are among the conversation-enders I’ve heard on the subject.

The truth is we know damn little about what works in campaigns. Most of what passes for evidence in this business is nothing more than dimly remembered anecdote or thinly disguised salesmanship.

Our firm tried to redress this failing some years ago, suggesting to a number of Democratic entities (which shall remain nameless) a program of carefully designed off-off-year experiments to determine what worked and what did not in the get-out-the-vote sphere. None wanted to spend the money. A hundred million and more on programs, they said in effect, but nary a dime to determine how to maximize the effectiveness of those dollars.

Since then, professors Alan Gerber, Donald Green and their colleagues have produced a series of compelling experiments that measure the impact of various forms of contact on turnout, turning the results into an invaluable manual.

They estimate that the untold hundreds of millions of dollars spent on GOTV efforts in 2004 added 3-4 million votes, or about 2.5 to 3.2 percent of the total — potentially enough to make a decisive difference.

But how do registration efforts compare to GOTV programs in cost-effectiveness? Some recently released census data, supplemented by our own analyses of voter files, hints that registration might be more important than campaigners have assumed.

A March census report suggests that to be a registrant is to be a voter. Nearly 89 percent of those who were registered cast ballots in 2004 — a three-point increase over 2000.

However, the census data are survey-based, with all the attendant misreporting problems. In addition, it is conceivable that all the new registrants are among the 12 percent who didn’t vote, which would translate into little impact from registration efforts.

Voter-file analysis can help obviate those problems. In these data, too, at least in ’04, to be a registrant was to be a voter. That was true in battleground states like Ohio, where 72.6 percent of registrants voted; Iowa (75.3 percent); Nevada (74.4 percent); and New Mexico (77.5 percent), but also in states like California (75.3 percent), where there was relatively little GOTV activity.

Again, though, were the new registrants concentrated among the small number who did not vote? Not really. In Ohio, 59.3 percent of new registrants voted; in Iowa, it was 74.5 percent of new registrants; in Nevada, 71 percent; in New Mexico, 55.4 percent; and in California, 73.7 percent.

A comparison of the census report and the Green, Gerber et al. analysis suggests that increased registration accounted for three to four times as many additional voters as did GOTV activity. Both were likely significant.

The issue is particularly acute for Democrats because the census reports that our core constituency is much less likely to be registered. Similar numbers of registered whites and African-Americans cast ballots (89 percent and 87 percent, respectively). However, whites were more likely to be registered (74 percent) than African-Americans (69 percent). Thus the registration gap is twice as large as the turnout gap. The difference is even greater for Latinos. Eighty-two percent of registered Latinos voted, but just 58 percent were registered.

These data are hardly definitive. Voters registered by “programs” may be the ones who do not turnout, while the self-motivated new registrants do cast ballots. But that is exactly my point. There is a lot we don’t know. We had better figure it out.

Mellman is president of The Mellman Group and has worked for Democratic candidates and causes since 1982, including Sen. John Kerry (D-Mass.) in 2004.