[Image courtesy of growabrain]
The Longmont (CO) TimesCall recently ran an editorial that was, in my opinion, remarkably insightful on the subject of voter registration data:
When a report showed that thousands of registered voters in Florida were, in fact, dead, it played directly into the suspicions that some have about the prevalence of voter fraud.
You see, some people -- on both sides of the political debate -- believe the only way their opponents can win is through cheating.
The interesting supplemental fact, however, to the dead voter report is the fact that no cases of voter fraud have been documented directly from the revelation.
When Florida lawmakers authorized the expansion of database searches for voter registration purposes, they put into place a mechanism that would produce exactly this result: finding voters who had died in other states but whose passing was not previously noted by local elections officials. In Florida, that's an all-too-common occurrence, because many older voters live in other states during the summer months but have Florida addresses come November, when Election Day arrives.
What the showing from Florida reveals is that increased linking of databases will help to clean up voter registration rolls much more efficiently for cases where people have multiple addresses or part-time residences. States such as Arizona and Texas also face many instances of split residencies.
The discovery that some people on a voter registration roll are no longer living or are newly naturalized citizens should not be taken as direct evidence of fraud; in fact, the discoveries should be celebrated for the fact that it means election officials are staying on top of their duties to ensure those who are eligible to vote may do so, and those who would be tempted to cheat have far fewer means to do so.
For all the tut-tutting about the accuracy of allegations of fraud on the rolls - not just in Florida but also Michigan and South Carolina - I think the TimesCall has it exactly right. Voters are so mobile, and their life circumstances change so frequently, that relying on a single database (in this case the voter rolls) to keep track of them is simply unrealistic. Moreover, matching just two files together often raises more questions than it answers, just as a person with two clocks is never entirely sure of the correct time.
Rather, election officials can - and should - look to a variety of data sources to ensure that the rolls include as many eligible voters - and only eligible voters - as possible. This modern, data-focused approach is the one championed by Pew's elections team to bring the registration process into the 21st century. [UPDATE: This editorial from the Columbus Dispatch about Ohio's participation in a multistate data initiative is another example of how this process works.]
Consulting multiple data sources on registration is like consulting multiple clocks for the time - providing greater confidence in, and evidence for, the accuracy of the rolls. Kudos to the TimesCall for its nuanced understanding of this issue.