Monday, November 27, 2006

The easy choice: prevent the loss of 175,000 votes

If you were purchasing voting machines in New York City (or anywhere else in the State, for that matter), and you knew that one choice would likely result in the loss of hundreds of thousands of votes more than the other -- in every election -- would it take you long to figure out which one to buy?

On Sunday, the Daily News again endorsed optical scan machines for New York (second editorial), noting that the Brennan Center had done the same last Tuesday. As the News noted, the Brennan Center's endorsement came down to a very simple, but incredibly important reason: optical scans lose less votes than any other system New York is currently considering.

New York will be choosing its replacement for the old lever machines in a matter of months. The choice is between two types of machines: Precinct Count Optical Scans (where a voter marks a ballot by fillling in ovals, much the way she would an SAT exam) and "full face" touchsreens or "DREs" (choices are listed on a computer screen; voters make their choices by touching the part of the screen that lists their favorite candidates).

Usability experts have long argued that "full face" DREs -- which list every candidate and every race on a very large computer screen -- are inherently confusing. There's just too much information presented to the voter on a computer screen at once. These experts have long predicted that this confusing interface would lead to voters skipping races and accidently choosing the wrong candidates.

The empirical evidence is now in. And guess what? The experts were right. Professor David Kimball (who worked with the Brennan Center on its voting system usability studies) examined the "lost vote" rates in over 2,000 counties in 2004. For the Presidential race, there were substantially more "lost votes" (i.e., where no vote was recorded) in jurisdictions that used full face DREs than those that used optical scan machines. That difference increases as we examine races further down the ballot. In fact, 15.4% of voters who used full face DREs did not have votes recorded on state ballot initiatives (generally at the bottom of the ballot), compared to only 8.8% who used optical scan ballots. The higher lost vote rate on full face DREs applied to every single model of full face DRE used in 2004.

Putting this in perspective, the difference in lost votes between full face DREs and optical scans on state ballot initiatives is roughly 7%. That easily represents 175,000 voters in New York City in a high turnout year.

In other words, if New York City purchases full face DREs instead of Optical Scans, it may well cost itself tens or (more likely) hundreds of thousands of votes on every state ballot initiative, every year. And every county in the state that chooses full face DREs over optical scans is likely to cost itself significant political power.

Call us crazy, but these facts seem to make for a very easy choice.

Categories: General, Voting

No comments: