May 04 2006
Election Day "Receipts"
One issue that came up in the March primary was the lack of a “receipt” given to voters on election day. I heard this question on election day, saw comments about it on the blog, was asked about it at election judge schools, and finally received another question on WDWS when I was on Penny for your Thoughts recently.
Prior to the advent of purely electronic or touch screen voting systems, no one really cared about a “receipt”. Many systems never had anything at all (like lever systems) or had something that was not really a receipt. Some people seemed to classify the stub on your punch card ballot as a receipt. It had no information on it, and quite frankly, you could have grabbed a stub from any punch card in America and it would not be distinguishable from the one in Champaign County. I compare the punch card stub to a grocery sack. It might demonstrate that you were in the store, but it does nothing to prove that you bought anything.
The confusion seems to stem from the evolution of the debate about electronic voting. When electronic voting began, people were complaining about a lack of a paper trail. Quite simply, the integrity of the election was in the hands of computer programmers. Even extensive testing could fail to reveal the designs of a master programmer.
What most election integrity advocates wanted was a paper trail. They wanted votes to be cast on an actual piece of paper, much like punch card does, so that any programming manipulation could be counterbalanced by a hand count of ballots. The idea of a receipt for electronic voting is a half way solution to a problem. The paper trail, as exists with the system we have in Champaign County, is far more preferable than a “receipt.”
People have also suggested that they be provided a piece of paper showing how they voted. Once again, that might look good, but it would do nothing to prove that the computer is actually tallying the votes correctly. Additionally, that piece of paper would facilitate vote fraud in a way that the political machines could only dream of.
The hand count of ballots in County Board district 9, which matched the computer count, demonstrates clearly the value of having a paper trail as provided by the optical scan equipment. However, there is no guarantee in any election that a hand count will happen. As such, I’m going to implement a change in future elections.
State law currently requires that we retabulate 5% of the precincts after election day (that means six in Champaign County) and prior to the canvass of votes. The State Board of Elections randomly selects those precincts. This year, when those six precincts were selected, representatives of the Democratic, Republican, and Green Parties ran the ballots through a tabulator and compared results with those from election day. In the future, I will randomly assign two precincts to each of the three established political parties and ask them to select one race in each of those precincts to hand count. Then, we will have a team of individuals from the parties conduct a hand count of ballots. With this policy, we will have a guarantee of a hand count of ballots in each election that will give voters in Champaign County absolute confidence that the tabulating machines are actually counting the ballots in the same way that the ballots are being voted.
I haven’t heard of this happening in other jurisdictions in the country, but I’d be interested in your thoughts and suggestions on this idea.
That sounds like a great way to bolster confidence in the optical scan machines, always verifying it with a hand count of a few precincts done by party representatives.
If I understand your plan, you will pick two precincts randomly, and the parties will each pick any race from each of the two (resulting in a total of six races). Rather than have each party count the races they picked, you may want to have all three parties count all six of the resulting races. This gives you a 3-way check on the hand count (which is more error prone than the machines) and makes sure that you don’t get in a situation where one party rep notices a discrepancy in “their” race, but keeps quiet if the discrepancy isn’t in their favor. In a big precinct I can easily see it taking a couple of go-arounds to get the hand counts to match, due to human error. If that’s too many votes to count, let the parties each pick a race in each precinct, and then randomly choose which one of the three choices gets counted in the first precinct, then randomly pick one of the other parties’ choice in the second precinct.
Another idea: when you randomly pick the two precincts, you might want to pick one from among the 50% of precincts with the most votes, and one from among the 50% of the precincts with the least votes, just to make sure you don’t end up with two huge-turnout precincts (a lot of work to count) or two tiny-turnout precincts (not good statistical validation).
I assume you’ll publish the results of the hand counts (or at least a statement that they matched) along with the official election results. What will you do if there’s a discrepancy? Will you just note it and move on, and let a candidate ask for a recount if they want one, or is there some defined process that you will follow if you find a problem?
Thanks for a great post. The six precincts will be picked by the State Board of Elections. We will have no ability to avoid larger precincts. We’ll just hope that they are moderately sized.
Any discrepancies will be available for perusal. I’m confident that the only potential problem, and one which I’ve seen, is when a voter uses and X or check mark. If that causes a change in any total, that would be available for someone to use in a discovery but would not officially change a result.
The key in my mind will be that any discrepancy can be explained. Any discrepancy that can’t be explained needs to be investigated.
We had a high success rate on the hand counts for the six precincts in County Board district 9. When you are just focusing on one race, and two others are with you, it helps. I would imagine that we would use a similar procedure this fall.
Thanks for your input. It sounds like this idea has at least one Yes vote.
Mark, it is a good idea. As an election judge for Ludlow 2, any ideas which can help hard-working judges are appreciated. In a non-partisan comment: You have served effectively and well, and it is an honor to be an election judge for Champaign County.
Bob Menge
Rantoul (judge in Ludlow 2)
As elaborated by anonymous, I think that this is an excellent idea and should be pursued. I would also like to have the results of this “check” posted on the website.