It's academic

Update: the recount results are now complete and certified (available here). Waukesha had a total of 87 votes that were changed for Prosser and Kloppenburg resulting in a 0.07% error in the original count (not including the 14,000 vote election night snafu).

As the governor is poised to sign the controversial voter ID bill in Wisconsin, what seems to be lacking on both sides of the discussion is evidence of voting irregularities or fraud. An article from yesterday’s Milwaukee Journal Sentinel cites 20 prosecutions of voter fraud from the November 2008 election (none of which involved voting in someone else’s name), but citizens have been presented with little additional evidence of voter fraud.

Of far greater concern for voter confidence is the state’s ability to count votes correctly. Here, we do have evidence from the recount currently underway from the Supreme Court election in April. Certified results from every county except for Waukesha are available here from the Government Accountability Board. Adding up the total number of changed votes for each candidate reveals 1,271 votes that were missed or incorrectly counted in the original vote. Here’s what the distribution of changed votes looks like across the state:

Obviously, the totals are larger in more highly populated counties. Below is another map showing the percentage of the total votes cast — highlighting counties within the state with the highest proportion of discrepancies. The highest percentage was 0.76% change in Waupaca County — almost every one out of a 100 votes was counted incorrectly.

If we’re looking to instill greater confidence in the electoral process, shouldn’t we start (as Senator Lena Taylor noted) by getting the count correct? Surely in our age of technology we could do substantially better.

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License.

As the Wisconsin Senate debates the proposed voter ID bill (SB 6), which requires a valid photo ID to vote, I was interested in the spatial distribution of disenfranchisement that might occur. Unfortunately, there’s little detailed data on registered voters and drivers licenses that’s publicly available for a comprehensive analysis. As a proxy, I’ve compared the number of valid drivers licenses from the Department of Transportation (available here for 2009) to the number of registered voters from the Government Accountability Board (available here for 2010).

Interpretation of this map requires caution for several reasons: 1) it doesn’t include all forms of ID that can be used (e.g. military IDs) for which detailed data isn’t available, 2) many drivers have licenses without a current address, 3) many drivers are not of voting age, and 4) the data are not from the same year.

Several counties (Dane, Milwaukee, Ozaukee, and Door Counties) have more voters than drivers, meaning that currently registered voters within these counties will have to obtain new IDs to vote. The total voters without drivers licenses within these counties, 20,162 voters, is well within the margin of recent elections. The largest numbers are from Dane and Milwaukee Counties which have traditionally voted heavily for the Democrats. Will there be a differential impact on future elections?

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License.