…but for the wrong reason.
Ohio Republicans sued to have the results of comparisons between new voter registrations and existing state databases (e.g. motor vehicle) made available to local officials. The Supreme Court rejected the suit on the grounds that the Ohio Republican Party had no standing to sue under the law (not commenting on the merits of the suit).
Source: cnn.com, et al.
At first glance it seems a public good to prevent vote fraud, and that this may be done through weeding out bogus voter registrations, and comparisons to state databases sounds like a good tool to help with the weeding. There are two things wrong with this idea:
- Voter registration fraud has little effect on vote fraud. This is because just having a name on a voter role doesn’t get you into vote. When Mickey Mouse shows up to vote, he might claim he left his ID in his other pants, but he won’t get to vote. No vote, no foul. If he shows a genuine ID, then he IS is some database, whether it matches or not.
- Matching is error prone. This is how the 2000 election was stolen in Florida, not by butterfly ballots and hanging chads. Florida election officials purged the Florida rolls selectively using matching with national convicted felon databases–both felons who committed no crimes in Florida and by law could still vote, and non-felons who had names similar to convicted felons. Bad matching goes both ways.Consider the hypothetical case of “Washington George”. Mr. George, was mis-keyed in the state DMV database (as “George Washington”), but not the Voter Registration database. If Washington George shows up to vote he may be challenged and have to cast a provisional ballot because of an error.
In my day job, I work with state databases and I know they are in pretty bad shape across the board. I also know that most database matching techniques in the real world are abysmally bad. This doesn’t mean matching cannot be good, but that it usually isn’t.