Midterm Recap with Voting Booth's Steven Rosenfeld

But, stepping back from Georgia, it's important to note that on a national level, it has never been easier to register and to vote. Even Florida just reenfranchished 1.5 million ex-felons in a vote where more than 60 percent was needed to pass. That vote alone could change the state's political complexion in the near future, if even a small percentage of those folks vote.

JB: I was gratified by Florida's decision regarding ex-felons. That's truly big. Did any other state also adopt that course? I know that ex-felons have long been effectively disenfranchised all over the country, not just in the South. Tell us more about that, please.

SR: On felon disenfranchisement, Florida was the nation's worst offender. Its anti-felon laws literally predated Jim Crow; they went back to post-Reconstruction. I'm not away of any other state that took such a big step this fall. But, notably, the Democratic governor of Virginia also re-enfranchised more than 100,000 ex-felons (if my memory serves me) recently; but he had to make two efforts to do that. Why? The state's GOP-majority legislature opposed him.

The most intriguing aspect of Florida's vote to me was how it won by large numbers in a state that's still fairly conservative once you get outside of the biggest cities. Sometimes it is a bit too easy for progressives to forget how conservative much of the country still is. Or maybe, it's not quite correct to call it conservative--but pushed by right-wing media to vote against Democrats, even if Democrats have an agenda that could help their communities. Nonetheless, the biggest selling point of the Florida initiative was redemption: everybody deserves a second chance. It was notable how that religious message also dovetailed with a social justice message.

JB: A long-awaited 'Amen' to that. One key component of elections should be justifiable voter confidence that our votes were counted as cast. Have we made any strides in that department since the infamous hanging chads of Florida, 2000? And if so, how?

SR: I think there have been some significant strides. First, there's been an about-face by the county's election officials on using entirely paperless voting machinery, which was what the country bought into post-Florida (as a reaction to the not-quite-perforated computer punch cards that resulting in dangling strips of paper known as chads). Today, something like 80 percent of the national electorate votes on ink-marked paper ballots. The reason behind that shift is a bit layered.

Going back a few years, there were all the visible problems with DREs (direct recording electronic) systems functioning properly; i.e., poor calibration means a voter picks candidate X and the machine reads candidate Y. Those kinds of snafus, plus the way the voting machine industry created a black box around its software (always a trade secret), and made local and state governments dependent on private sector technicians, added up to a lack of transparency and fueled speculation that counts could be pre-programmed, or rigged. At any rate, the country started moving back to ink-marked paper ballots, and then 2016's Russian hacking (of state voter registration databases and private voting machinery vendors) reinforced the conclusion that paper-based voting was more secure and transparent.

At the same time this return to paper has been occurring, there have been two different new approaches surrounding the process to measure vote count accuracy--generally called audits. This is increasingly important, because most of the election machinery bought after Florida is at the end of its useful life and will be replaced in the next year or so. Thus, whatever systems are acquired will be with us for the 2020s (unless smartphones take over, but let's not digress).

So two kinds audits have emerged. The first is called a risk-limiting audit. That is a statistical exercise after Election Day. All the ballots are assembled in one place. Then you pull a random number to see if the previously reported results were 95 percent correct. This was developed by some noted academic statisticians, who created the probability formula. When election results are very close you have to keep pulling out more ballots to hit that 95 percent threshold. This approach was first deployed statewide in Colorado, where there's voting by mail and central counting operations. Typically, these audits aren't done before election results are certified, which means they won't affect who is declared the winner.

The second audit approach involves a two-step process where after paper ballots are initially counted (counties and states use optical scanners and related software), they are put through a second scan by a completely different computer system. That second scan creates a detailed digital image of every ballot card and analyzes all of the ink marks on the paper. (It can assess how confidently ovals were filled out and allow anyone to zoom out to look at the surrounding pages when there are questionable marks). This second and independent process ends up doing a few things.

First, you get an accessible library of every ballot (for manual review if needed) and you can then check how accurate the initial count was. This second scan can be very granular, i.e., scanning ballots in batches tied to the specific machine used in precincts or counting centers. The likelihood of hacking the vote is virtually eliminated because the only thing in common between these two systems is paper, which doesn't carry computer viruses. Finally, this audit can be done before results are officially certified. That final point is very important, because that means an audit is not an exercise with no electoral consequences.

These two audit approaches are really different. One is like an exit poll, where you are told, after a mathematical analysis, to believe the results. The other gives you a complete accounting of all the votes before the results become official. Ironically--at least to me--many progressives favor the first of these approaches, risk-limiting audits, even though I don't think it is accurate to call them transparent and verifying the vote count. Yet, some well-known groups have pushed this approach. It's been promoted by influential editorial boards. In contrast, the ballot-image audit hasn't received the same attention. It's newer and also doesn't have all of the voting machine manufacturers behind it (as it exposes their machinery's shortcomings).

Without going on endlessly here, the ballot-image audits were developed in Florida by the same officials who were prevented from completing a recount by the U.S. Supreme Court in the 2000 presidential election. Seven Florida counties, Maryland, and others use it. But it was not used in the recent Florida recounts, because the state didn't license it for that purpose. Had it been, the recounts likely would have proceeded very quickly and differently-- as Broward County has it. Instead, what we saw in Florida was the limitations of big counties with handling the volume of paper in a more outdated recount process, and the limitations of some older scanners.

Perhaps I've gone too deep into the weeds here. The emerging audit choice is do you want a statistical sample akin to an exit poll to tell you to trust the results, or a complete accounting of all the votes based on a separate secondary scan with searchable images and a ballot library?

Next Page  1  |  2  |  3  |  4

(Note: You can view every article as one long page if you sign up as an Advocate Member, or higher).