Two studies, two standards

As the vote on Question — a referendum for which a Yes vote supports the 38-year practice of election day voter registration — draws near, the No on 1 campaign has modified its tone and focus.

Although the only issue voters are faced with is election day registration, arguments from No on 1 and letters to the editor supporting its position are focusing on another issue: voter identification.

There certainly are questions about why they are choosing that focus instead of the one that’s actually on the ballot.

While that emphasis is a matter of Maine politics, this post looks at an odd set of contradictions.

The focus: a research critique endorsed by opponents of Question One raises issues that apply to these opponents’ own study — two studies, two research standards.

Here’s what’s being examined.

1. The critique endorsed by opponent of Question One major advocate for limiting access to the ballot on election day endorses the view of the Heritage Foundation’s Hans von Spakovsky regarding the studies conducted by the Brennan Center of New York University School of Law that found that many people are disenfranchised by the new voter identification rules. It’s worth noting that the Brennan Center has released an extensive and convincing rebuttal to these criticisms.

(Those who have followed the politics of voting have certainly heard of Mr. von Spakovsky, as in 2007 seven attorneys and one political geographer from the Department of Justice contended he had politicized the operations of the DoJ’s voting section. When he recently gave testimony to a Senate committee, his analysis of the data was flawed by a very simple statistical error, the conflation of correlation and causation.)

2. The election day registration opponents’ own study: The document laying out various criticisms that range widely from public opinon data to information about clerical errors in the voting rolls to fraud potential.

A couple of points of contradiction:

Von Spakovsky and Ingram criticize the use of a survey of “987 individuals to estimate the number of Americans without valid documentation.” They contend “the report contains no information on how the survey determined whether a respondent was actually an American citizen” and claim that information is missing about survey methodology.

The Brennan Center disputes this, saying, “the survey was, in fact, essentially limited to eligible voters, since it focused exclusively on U.S. citizens over the age of eighteen, the main determinants of voter eligibility across the country.” Moreover, they point to transparency about the survey’s methods and note that the survey instrument developed and fielded by the respected Opinion Research Corporation (which conducts the well-regarded CNN poll) and uses live operators.

But, in any case, what of the survey done by No on 1 advocates?

True, it did not pose the same sort of questions posed by the Brennan Center.

But what sort of sample did it use?

Its sample was about half of the 987 included in the survey criticized by von Spakovsky and Ingram — only 500.

Since it was a robopoll, it’s hard to tell what sort of person actually decided to answer the questions. Question wording is a very important element of polling. And the firm used, Pulse, provides a very inexpensive service because clients write their own questions. Moreover, to this day, it is unclear how respondents were classified as likely voters and thus included in the reported data.

While the von Spakovsky and Ingram case against the Brennan Center study appears to have holes, these Maine No on 1 advocates argue the criticisms are valid, even as those critiques apply to its own report.