Florida Non-Votes with New Equipment
For all the bad press (much of it deserved) that Florida's election system has received since 2000, it's a pleasure to report some good news. The number of non-votes declined dramatically from 2000 to 2004, going from 2.9% to 0.4%, as reported in this story from the Palm Beach Post and this analysis by the Florida Department of State.

The decline in non-votes or "residual votes" -- the sum of overvotes (for which more than one vote was cast) and undervotes (for which not vote was recorded) -- is largely attributable to the change in voting equipment. In 2000, Florida voters used a variety of different types of equipment, including central-count optical scan and the infamous punch card. There were over 180,000 residual votes statewide, many times the margin of victory as you'll no doubt recall.

This year, Florida voters used either precinct-count optical scan or electronic touchscreen voting systems. Both systems provide "second chance" voting, meaning that voters are given the opportunity to correct errors. With the touchscreen systems, inadvertant overvotes aren't allowed and voters have the chance to see their choices, and thereby check their work, before voting. With precinct-count optical scan (in contrast to its central-count cousin) voters put their ballots through a counter at the polling place, which will notify them if they've overvoted.

Both systems registered minuscule non-vote rates: around 0.42% with electronic machines and 0.40% with precinct-count optical scan. This is about the lower limit of what it's possible to accomplish with technology. That's because the remaining residual votes are probably intentional undervotes -- that is, voters who intentionally abstained from the presidential race. Surveys in prior elections show that the rate of intentional undervoting is around 0.3-0.7%, varying slightly by race and income levels.

An especially noteworthy improvement occurred in Palm Beach County, where in which 6.4 percent of ballots cast by punch card recorded no vote for President in 2000. This year, only 0.4 percent of votes recorded no vote, using the county's touchscreen voting machines. We know that in 2000, a disproportionate number of those who didn't have their votes recorded were racial minorities. While the reports released thus far doesn't include an assessment of race, expect the so-called "racial gap" in uncounted votes to have declined dramatically once such analyses are conducted -- as was the case in Georgia, when it converted to an electronic voting system.

The bottom line: Despite all the security concerns surrounding electronic voting, and some of them are real, moving to electronic and precinct-count optical scan voting technology has been a very good thing for Floridians. More of their votes, many more, were counted in 2004 than was the case four years ago.
- posted by Dan Tokaji @ 11:53 PM