In 1987, the Markle Foundation was concerned about the accuracy and effectiveness of voting and voting machines. The foundation wanted to fund an evaluation, but had trouble finding an organization to do the work. Consumers’ Union declined the offer but directed Markle to ECRI Institute. We accepted the job and hired an engineer and a computer expert to add to a lawyer heading the project.
Our team reviewed all the available information on voting and voting machines. We contacted the voting machine manufacturers and got them to provide test units for our testing, just like we do in Health Devices. We found most voting systems were good but some, like the Votomatic, were error prone.
Votomatic used pre-punched ballots to make it easy for voters to punch out the chad (the small piece of paper in the hole) to cast a vote. The pre-punched ballots also made it easy for un-voted chads to fall out during vote counting; we found thousands of chad in a ballot reader where they should not be. Our final report was sent to each U.S. county election official—more than 3000, in all. Despite our findings, nothing was done about the potential problem because of the costs involved in change.
Fast forward to the 2000 presidential election. CBS 60 Minutes asked ECRI Institute to examine the Florida Votomatic ballots to determine if there was a difference between the Bush and Gore chads. After extensive testing, we saw no significant differences between the chads on the ballots. Nevertheless, as our original testing showed, the problem was most likely caused by the pre-punched chads falling out of the ballots during recounts and handling. What would have happened if our report did cause change?