1/Like this: Ballot marking devices have essential security flaws. Like all computers they can be hacked, misprogrammed, misconfigured & misused. Because printed ballots from a compromised machine cannot be trusted as an expression of voter intent, audits cannot detect cheating.
2/Why can't BMD printed ballots be trusted as an expression of voter intent? Because voters won't notice fraudulent outputs of compromised BMDs. It's the most widely replicated experiment in election security dating back a decade or more.
3/...as summarized in the NAS "Securing the Vote" report: without complete record of votes expressed, voters would be unable to recall all of their prior choices. nap.edu/25120
4/Isn’t it just a question of better ballot or user interface design? No. Most voters won’t check, but even if all checked machine-marked ballots for errors, they would be unable to recall their prior choice because there are fundamental cognitive limitations on human memory.
5/Hundreds of studies over 100 yrs point to fundamental limitations of human memory that would inhibit ballot verification by voters. There is virtually no controversy about the science: “Accepted facts about memory are remarkably firm.” amazon.com/Oxford-Handboo…
6/All of which means that cheating is impossible to detect because the original expression of voter intent has been lost & replaced by machine-markings. If a machine has been compromised, fraudulent ballots will be swept into the audit trail and counted as if they were valid.
7/A more detailed argument is in the updated and revised version the paper “Ballot Marking Devices (BMDs) Cannot Assure the Will of the Voters,” by Andrew Appel, Philip Stark and me. The expanded and revised paper was recently posted on SSRN: dx.doi.org/10.2139/ssrn.3…
8/@threadreaderapp unroll
