There's no such thing as fail-proof pre-election security. That's why we need audits after each election.
In November 2018, for the first time in Wisconsin history, some of our computer-tabulated election results will be verified by hand-counting paper ballots.
But the audits ordered by the Wisconsin Elections Commission for this November are far from the audits recommended by election-security experts and being implemented in other states.
Here's a run-down of the benefits and limitations, of how far Wisconsin has come and how far we have to go.
First, the benefits:
To effectively prevent miscounted preliminary results from being declared final, audits need to:
- be completed early in the canvass period (the two weeks following each election) to allow time to correct any miscounts with a full hand count; and
- hand count enough ballots, chosen the correct way, to provide statistical confidence that the right winners had been identified.
First, timing: This is the most important improvement. The state elections agency has had authority to order voting-machine audits ever since 2006, but in previous years allowed local clerks to delay them until after they had certified final election results. This, of course, eliminated the possibility that miscounts could be corrected and with it any fraud-deterrence value the audits might otherwise have had.
The WEC has ordered this year’s voting-machine audits to be completed by November 28, before the results in the statewide races are declared final. This gives the audits some deterrence value, because if the local clerks don’t wait until the very last minute, there still might be time to correct any miscounts before anyone is officially declared the winner.
Second, size of the sample: This year's audits will include more machines that in previous years (five percent of the voting machines, or 183 machines statewide out of more than 3,500). That's almost twice as many machines as in previous years, and--believe it or not--auditing 5% of the state’s voting machines will put Wisconsin in the top dozen states for the size of the post-election verification effort. Many other states require samples of only 1% or 3%.
In addition, the WEC decided to select the sample in a way that ensures at least one voting machine in every county will be audited. Again, this is more protective than it might sound. The vote-counting code for statewide races is typically created for an entire county at the same time, and then copied into each voting machine. As a result, mis-programming is a county-level risk, and auditing only one machine has a good chance of detecting any such problems.
Now, the limitations:
1. These voting-machine audits won't guarantee the right winner was identified.
These are designed to be voting-machine audits, not election audits. Here's the difference:
Election audits answer the bottom-line question: “Have we identified the right winners?” Election audits hand-count as many randomly selected ballots or entire precincts as are needed to provide confidence that the correct winner has been identified. If the first sample does not confirm a winner, the sample must be increased and increased again, until either the Election-Night winner is confirmed or a full hand count has identified the real winner.
But under current state law, the WEC has no authority to order election audits. The sample size for these audits won't be big enough to verify the winner but the widest-margin results, and the WEC is not requiring the municipalities to expand the audit to more machines if one is found to have miscounted.
Voting machine audits answer a much smaller question: “Did this voting machine work as it was supposed to work on Election Day?” The auditors will NOT be looking to see whether the voting machines correctly identified and counted voter intent. They will be looking to see whether the machines operated as they were designed and programmed to operate--and, sadly, our election laws do not require that those two things to be the same thing.
So, hold onto your hat; I’m not making this up: A voting-machine audit might deem even a serious miscount to be no problem. A real-life example can illustrate this. A vendor who programmed the voting machines for the City of Medford (Taylor County, Wisconsin) once disenfranchised one-third of that city's voters by programming the voting machine to ignore all straight-party-ticket votes. If a voting-machine audit had detected that, it would NOT have been flagged as a problem. Remember: a voting-machine audit asks only whether the voting machine operated as it was supposed to. In the Medford miscount, the answer is “Yes; the voting machine ignored all the votes it was programmed to ignore.”
2. The voting-machine audits will be done only for one election every two years.
WEC's authority to order voting-machine audits is limited to November elections in even-numbered years. Miscounts in other elections will continue to go undetected and uncorrected. Fraud will not be deterred in any other election.
If voters want more security than that--and we do--we're going to have to work on our individual county clerks to do the right thing, voluntarily, after every election.
3. Oversight must be provided by the voters.
Observers are needed at every audit. Transparency is a vital component of a valid audit, and transparency is impossible if no one shows up to observe. It’s not saying anything worse than “Election officials are normal human beings” to point out that they are less likely to cut corners and more likely to follow the procedures carefully when observers are present. WEC does not have the staff--no where near enough staff--to observe these audits to make sure they are done well.
In addition, because of the weird requirement that miscounts will not be considered ‘errors’ if they can be explained by human error or intervention, observers need to be present to make sure that anomalies are noted and acted upon. The WEC might not require municipal clerks to correct any miscounts they find, but voters should certainly insist upon it.
There will be a voting-machine audit somewhere in your county between Monday, November 12 and (probably) Thanksgiving.
Chances are, however, it won't be well-publicized. You will need to contact your municipal clerk to find out when and where it will take place.
BE THERE to observe. Heck, offer to volunteer as a vote-counter if you’ve got a receptive, public-welcoming municipal clerk. If the municipal clerk follows WEC instructions and common sense, the audit won’t take more than a few hours. Depending on the size of the precinct, you could be done before lunch.
No one will know which municipalities will be conducting voting-machine audits until the day after the election. That’s done to increase the audit’s value for deterring fraud. The list will be posted on the WEC website, this website, and on our Facebook page, and we’ll be sending email notices to anyone whose home county we know. (If you want an email telling you where the voting-machine audit(s) in your county will be conducted, email us at WiscElectionIntegrity@gmail.com and tell us what county you live in.)