Are We Condemned to Repeat It?

By Daniel P. Tokaji

It’s election year in Ohio, likely to be pivotal in the presidential contest. Everyone expects a close race. Yet there’s great concern about whether the state’s election infrastructure can hold up to the pressure that will be upon it. Of special concern is the voting equipment to be used, particularly in the state’s largest and most diverse county. Another worry is provisional ballots, upon which the state increasingly relies for registration problems, voters who lack proper ID, and those who’ve moved. If the election is close enough, the two major parties could wind up fighting over which provisional ballots should count. The spectre of litigation thus hangs heavy over the state. To top it all off, there have been repeated accusations of partisanship by Ohio’s chief election official, the Secretary of State, from the opposing party.

Any of this sound familiar?

When the nation’s attention focuses more intently on battleground states later this year, many will no doubt scratch their heads and wonder what the State of Ohio has been up to since 2004. In reality, there have been plenty of changes, some of them for the better. But many of the same issues remain. This comment discusses three big ones: voting equipment, provisional ballots, and allegations of partisanship. It remembers the past while looking ahead to the future, in light of some brand new information from the March primary.

Voting Equipment

This is one area in which there have been significant improvements since 2004, when most voters throughout the state still used “hanging chad” punch card voting systems. The result was that tens of thousands of ballots didn’t register a vote for President. Elsewhere in the country, voters used newer voting equipment that provided voters with notice and an opportunity to correct errors. Such “notice” voting equipment combined, with better procedures, saved about one million votes that would otherwise have been lost in 2004. By 2006, Ohio counties had finally switched to notice voting technology, of either the electronic touchscreen or precinct-count optical scan variety. Both types of systems allow voters to check for overvotes, and thus reduce the number of uncounted votes. The bad news is that Cuyahoga County, the state’s biggest county which includes Cleveland, had well-documented problems implementing the Diebold touchscreen system it decided to buy. Among the problems was that somewhere around 20% of the paper records generated by the system were damaged or unreadable, something that’s especially problematic given that Ohio law makes paper the official ballot of record.

As a result, Cuyahoga County switched again, using a non-notice optical scan ballot system in the March 2008 primary. The ACLU sued, arguing that votes would predictably be lost due to the switch, but a federal district judge declined to order the county to use a notice system so close to election day. (Disclosure: I consulted with the ACLU on that case.)

The result was that Cuyahoga County used a non-notice system in the March primary. So how did things go? At first glance, it didn’t look so bad. The county initially reported 818 overvotes at the time of its unofficial count. (This page now seems to have been removed from the Secretary of State’s website.) That’s more than any other county but one, but still relatively low when you consider the total turnout of 436,609 in that county.

Some brand new information explains the suspiciously low number of overvotes initially reported. The earlier figure didn’t include ballots that would have been overvotes, but had been “remade” so as to avoid being rejected by tabulating equipment. The practice of remaking, as the name suggests, involves marking a new ballot that replicates what election officials believe to be the voter’s intended choices, when a ballot is mismarked. Until today, there wasn’t publicly available information on how many ballots had been “remade” but I’ve just received information from the Cuyahoga Board of Elections that that there were really over 4000 overvotes:

During the Unofficial Count, the M650 Optical Scan units were programmed to accept and count overvotes, alllowing all ballots to scan through the machines without stopping. A total of 4,117 overvotes were recorded within the 406,450 votes counted in the Unofficial Canvass, equaling 1.01% of the total votes cast.

… All ballots containing overvotes were examined by a BOE Management Team to determine whether the voter clearly intended to vote for a single candidate. Of the 4,117 overvotes, it was determined that 1,240 contained a clear indication of the voter’s intent. Those ballots were remade …, reducing the overvote total to 2,877

The translation is that there were a lot more overvotes than the county originally reported (though “406,450 votes” should, I think, be “406,450 ballots,” making the percentage smaller than reported). The number of overvotes was reduced from 4,117 by remaking ballots, in accordance with what the board’s management team deemed the “clear indication of the voter’s intent.”

Does anyone remember Palm Beach County?

There are of course good reasons why election officials might want to remake ballots, but it introduces an element of subjectivity into the process. In the event of a close election, one could easily imagine litigation over whether and how to remake those ballots, analogous to those which took place in Florida over whether and how to count punch card ballots in 2000.

At the time of the ACLU lawsuit earlier this year, Cuyahoga County had expressed its intent to use a notice system — probably precinct-count optical scan ballots — in the November 2008 election, but has since decided that it would prefer to stay with it’s non-notice, central-count system. But an Ohio statute passed earlier this year (SB 286) forbids counties from using central count systems after the March 2008 primary. Sec. 3506.21(D)(1). There’s also a reasonable argument that remaking ballots to avoid overvotes violates Sec. 3506.21(B)(3) of that law, which provides:

If automatic tabulating equipment detects that more marks were made on an optical scan ballot for a particular office, question, or issue than the number of selections that a voter is allowed by law to make for that office, question, or issue, the voter’s ballot shall be invalidated for that office, question, or issue.

If Cuyahoga or other counties continue the practice of remaking ballots before feeding them into the scanner, we could expect an argument that this practice violates Ohio law.

The bottom line is that Cuyahoga County will have to switch to a notice-based system, probably precinct-count optical scan, unless the Ohio legislature changes the law. At this point, it looks like there won’t be any such change. In my view this is a good thing, since precinct-count optical scan systems provide voters with notice of overvotes and therefore reduce errors. It also avoids the legally problematic practice of remaking ballots, which also increases the risk of post-election litigation over the result. But it will mean that Cuyahoga County will have a lot of work to do, getting a new system up and running in less than six months.

Provisional Ballots

Provisional ballots are required by the Help America Vote Act of 2002 (HAVA), for voters who arrive at their polling place to find their names not on the rolls or who don’t have required identification. In 2004, attention focused on the question of whether provisional ballots should be counted if cast in the wrong precinct. Ohio’s then-Secretary of State, Republican Ken Blackwell, said no. This led to litigation by the Democratic Party and voting rights groups, in which Blackwell’s position prevailed. A subsequent study by a Cleveland advocacy group found that, of 11,600 provisional ballots rejected for this reason in 2006, 32% were actually at the right polling place but the wrong precinct. (Many polling places in Ohio have multiple precincts voting at them.)

Just after the election, in the early morning hours of November 3, 2004, it looked like Ohio’s provisional ballots might determine the outcome of the race between President Bush and Senator Kerry. Bush led by 136,484 votes in Ohio, with what turned out to be 155,000 provisional ballots left to be verified and counted. At the end of the day, when those provisional ballots were counted, President Bush’s margin of victors shrunk by around 18,000 votes to 118,601.

Provisional ballots can make the difference between victory and defeat in a close race. Several states were closer than Ohio in 2004, including swing states of Iowa, New Mexico, and Wisconsin. In all these states, the margin of victory was less than the roughly 18,000 votes that Kerry “made up” through provisional ballots in 2004. Provisional ballots thus provide fertile ground for post-election disputes — it’s one of the main ways that someone on the losing end can make up ground. It follows that heavy reliance on provisional ballots increases the “margin of litigation,” within which election results can be disputed and litigated after the fact.

Now here’s the really troubling part: It looks like Ohio’s reliance on provisional ballots is actually increasing, rather than decreasing. In our report last year on the election systems of five midwestern states, my colleagues Steve Huefner, Ned Foley, and I found that the percentage of Ohio voters casting provisional ballots increased between the 2004 and 2006 general elections, going from 2.77% to 3.09% of total ballots cast (see Chapter 3, p. 32).

In the 2008 primary, the percentage of provisional ballots was higher still. According to recent data from the Secretary of State’s office, there were 123,432 provisional ballots, out of 3,603,523 overall. That means that 3.43% of Ohioans voted provisionally in this year’s primary. Once those provisionals were counted, Senator Obama picked up two delegates from Ohio’s March 4 primary, a possibility I’d noted here and here.

It’s hard to say for sure why this is happening, since there are a number of reasons voters cast a provisional ballot. It could be that there are problems with the new statewide registration database required by HAVA, resulting in many voters’ names not being on the rolls when they appear at the polls. It could be that voters and/or pollworkers are unfamiliar with Ohio’s rather confusing identification requirements, enacted for the 2006 election. It could be that a lot of voters have moved from one precinct to another between elections and are casting provisional ballots at their new polling place, as state law allows. What we know is that more provisional ballots increases the risk of post-election litigation.

Allegations of Partisanship

In 2004, Democrats harshly criticized Secretary of State Blackwell, for what they perceived to be partisan decisions. Foremost among them was the decision not to count provisional ballots cast in the wrong precinct, and to require that voter registration applications be on heavy-stock “80 pound” paper to be accepted (the latter order was ultimately rescinded).

In 2006, Ohio elected a Democrat, Jennifer Brunner, to the Secretary of State’s office. She too has proven controversial, most notably with the recommendations contained in her office’s EVEREST report late last year, which included some drastic, ill-considered, and impractical recommendations — including a return to central-count voting systems. She’s also come under fire for her efforts to remove local voting officials who’ve criticized her, a point on which I’ve expressed concern. Republicans accuse her of a “Culture of Intimidation, Bullying.”

The truth is that there’s plenty of blame to go around, when it comes to the problems in Ohio’s election system. Brunner has done some good things, like commissioning research on the security of all Ohio voting systems and clarifying state rules on voter ID. But she’s also made some bad decisions, which have damaged her credibility with the legislature, the media, and the public. And like any other elected official, she’s probably motivated to some extent by the interests of her party.

At the same time, Republicans are partly responsible for many of the problems with the state’s election system, particularly their enactment of massive changes in the state’s election laws in 2006, many of them confusing and otherwise problematic. In addition, they’re no doubt motivated by the hope of capturing the Secretary of State’s office in 2010, which brings with it not only control of the state’s election system but — probably even more significant, from the parties’ perspective — what could well be the pivotal vote on Ohio’s reapportionment board, which will redraw state legislative districts after the next census. This partly explains the heated rhetoric flying between Brunner and the GOP.

From Katherine Harris to Kevin Shelley to Ken Blackwell to Jennifer Brunner, it’s become abundantly clear that partisanship in the administration of elections is a recurring subject of controversy that won’t go away by itself. It’s important to look beyond personalities, and recognize that there’s an institutional problem with having a state’s chief election official chosen through partisan elections — as a sizeable majority of states still do. It’s not just that the umpire has a stake in the game; he or she is actually a player in that game — however much election officials may talk about not “wear[ing] a jersey of one of the teams.” This is a structural problem that demands structural reform. One good model is Wisconsin’s Government Accountability Board, discussed here.

Of course, we won’t see such institutional reform between now and November 4, 2008. Nor are we likely to see major changes in the voting equipment that’s used. What we can do is focus on procedures that will ensure voting system integrity, consistent implementation of voter ID and provisional voting rules, and poll worker recruitment and training. And for those of you wondering what you can do, volunteering as a poll worker or in election protection efforts (like this one) is a great place to start.

So come see us in Ohio this fall. It’s sure to be a lively election season. Again.