The image will forever be burned in my head. A lady with a thick pair of prescription glasses squints at a hanging chad. It was the 2000 Florida recount, the worst nightmare for Democrats and Republicans alike—a presidential election that came down to a pile of illegible half-punched papers.
Since that ever-controversial election, Caltech and MIT have teamed up in a joint research initiative called the Voting Technology Project (VTP). Their goal? To bring cross-discipline researchers to analyze and scrutinize the voting process.
"Our research indicates that the human factor is probably the biggest source of problems in elections," VTP co-director Michael Alvarez explains. "Generally speaking, those problems are relatively low magnitude and low scale."
Things Have Improved
The VTP’s most recent report is a rare piece of positive news in an otherwise negatively toned election. Between 2000 and 2008, voting in the U.S. became far more reliable, thanks largely to new voting machines as well as statewide computerized voting registration. Such measures produced a 50% drop in lost votes from 2000 to 2008.
"If you roll back the clock to 2000, Florida was an interesting microcosm. They used all of the different voting systems found throughout the nation. When we analyzed the election, we found that there were voting technologies that were clearly inferior," Alvarez explains. "What’s happening in many cases was the way ballots were designed, voters were making simple mistakes. They were overvoting or they were undervoting. We recommended these inferior technologies be phased out, which they were, and be replaced with technologies that allow voters to check their ballots."
The project used a simple indicator to measure voting mistakes called residual votes—they’re the difference between total votes counted for an office and the total votes cast. What residual votes indicate is the number of people who punched a card, pulled a lever or otherwise filed a vote on election day that in the end was not officially tallied.
In the days of 2000, errors caused by old lever machines and punchcards created a residual voting rate of 2%, meaning 2% of voters’ votes just didn’t count. Touch-screen voting systems and optical scanners subsequently reduced the residual voting rate to just 1% in 2008. Alvarez credits much of this reduction to simple human improvements. Most notably, newer electronic systems allow voters to see who they voted for (and if they accidentally voted for more than one candidate), so voters can catch their own mistakes when they inevitably happen. And these more reliable electronic systems can be found in just about every county in the U.S.
So does Alvarez have any concerns for the coming election? Absentee voting, he says. One confounding variable is the general reliability of mail during a time when the USPS is struggling. But the other? Again, it’s human error.
"Voters who vote by mail are not able to use ballot checking technologies," he explains. "They may make a simple mistake. There may be a stray ink mark. The ink may smear when it’s stuck in the envelope."
And a simple smear of ink can be enough to have a vote thrown out altogether.
What Should Still Change?
All of this said, if Alvarez could make one change to the current voting process, with no limitations of funding, regionality, or bureaucracy, it wouldn’t be in voting technology but voting registration. That burden of registration is almost entirely on the back of voters who need to preplan their participation. Couple that burden with the fact that Americans are expected to vote quite frequently (up to a few times a year, to keep registration current), many well-meaning citizens can be frozen out of the voting process.
"Some states have same-day registration, and they tend to see higher registration and turnout, which I think is a good thing," Alvarez laughs. "In other countries, the government decides who’s eligible, and they’re added to the roster. If you were to move in that direction, it could possibly lead to higher rates in registration and higher rates of turnout, too."
But my takeaway from our conversation was a decidedly heartening one. Electoral college notwithstanding, in less than 10 years, the process has evolved significantly to better the odds of every vote truly counting.
"We can always do better, but our election officials throughout the country do a pretty good job," he says. "And when elections are tight, when questions arise, we have processes for solving those."