As part of our commitment to conduct a scholarly investigation of the electoral process in Ohio this year, Election Law @ Moritz is beginning the process of identifying the questions we should ask and the evidence we should look for. We are conducting this investigation, not with any expectation that it will cast doubt on the result of the presidential election in this state, but to better understand the extent to which there were problems with the process. This understanding could contribute, we would hope, to proposals for improving the process in advance of the next major election in the state, in 2006. We will report what we find, without regard to whether those findings might be considered more advantageous to one party or another.
As a preliminary matter, these are some of the questions we anticipate asking:
1. Long lines: the extent of the disenfranchisement they caused. How many Ohioans left polling places without casting a ballot because they could not stand in line any longer? How many Ohioans decided not to go to the polls in the afternoon or evening because they heard about how long the lines were? (Conceptually, these questions seem distinct in assessing the disenfranchisement effect of the lines, because one can never be sure that the person who stayed home really would have gone to the polls, whereas someone who had to leave after standing in line obviously did.) It may be difficult, if not impossible, to get a good answer to these questions, based on "hard" evidence, but they are nonetheless worth pursuing.
2. Long lines: why they occurred. To what extent were they caused by insufficient numbers of voting machines, insufficient numbers of poll workers, inadequate training of poll workers, a disproportionate percentage of first-time voters whose unfamiliarity with the process required additional time, or other factors? On what basis were machines and workers assigned to each precinct? What was the ratio of machines to registered voters in each precinct? Machines to expected turnout? There has already been some reporting on the number of votes cast per machine, but there needs to be further examination of this issue. For one thing, as several colleagues have observed, the number of votes cast per machine doesn't necessarily capture the length of a line at a particular precinct, since the line may be moving more quickly there than elsewhere.
3. Long lines: how long in various precincts. We know from widespread news reports that the lines were excessively long in many precincts in Ohio. But we don't yet have a systematic survey of exactly how long the lines were in different counties – and different precincts within particular counties. If possible, it would be useful to know, at least for various sample precincts, how long the line was at different times throughout the day on November 2.
4. Provisional ballots: their availability. There have been some word-of-mouth stories about individuals unable to cast provisional ballots, either because a precinct temporarily ran out or because poll workers failed to provide them to individuals who were eligible under HAVA to receive them. At this point, we have not seen evidence that either confirms or refutes these stories – or begins to show the magnitude of this alleged violation of HAVA. Because one of the most important innovations of this federal statute was its guarantee that individuals who believed themselves eligible to vote at a particular precinct would not be turned away without having had the opportunity to cast a provisional ballot, it is important to investigate these allegations and attempt to quantify the extent this problem occurred.
5. Provisional ballots: evaluation for voter eligibility. This process is still occurring and may not be complete in some counties until Wednesday, November 17. Because of the novelty of HAVA-mandated provisional ballots and some uncertainty on how they are to be evaluated to determine eligibility, it is important to develop a good understanding of how this determination is being made in the various counties throughout the state. We will learn what percentage of provisional ballots are determined to be eligible, but it will be important to know how many ballots were subjected to divided votes among the county officials who examined them, how much deliberation occurred with respect to these disputed ballots, and what evidence officials considered in making their eligibility determinations.
6. Provisional ballots: reasons why they were cast. It is anticipated that the 155,428 provisional ballots will fall into a variety of categories, in terms of the reasons why individuals were required to cast them rather than regular ballots. Expected to be among the most likely reasons are: (1) failure to appear on registration rolls because of incomplete registration forms; (2) failure to appear on registration rolls because of lost or undelivered registration forms; (3) failure to appear on registration rolls because of purging of rolls; (4) failure to provide ID for those required to do so by HAVA; (5) decision to cast provisional ballot at wrong precinct even after being told ballot won't count; (6) change of address, as permitted under pre-HAVA state law.
It is too early to know the extent to which actual practices will bear out these expectations. Interestingly, however, the Toledo Blade published a story containing some preliminary results from Lucas County. Only 709 of the 1,548 provisional ballots examined thus far by the county board of elections were determined to be eligible: 45.8%. The ballots found to be ineligible fell into the following categories:
|no affirmation of belief of eligibility||57|
|signature different from registration form||29|
|empty provisional ballot envelope||41|
This break-down, while useful, does not tell us what percentage were "not registered" because their registration forms were incomplete (lacking a driver's license number, for example) rather than never received – or never sent in the first place. Likewise, regarding the 29 provisional ballots where the county officials found a mismatch between the signature at the polling place and the signature on the registration form, was there any dispute among the officials in this evaluation and how obvious was the discrepancy to someone untrained in handwriting analysis?
7. Voting machine errors. It has been widely reported that one voting machine initially recorded an extra 3893 votes for President Bush. What steps have been taken to make sure that no comparable errors occurred elsewhere?
8. Punch card problems. The Columbus Dispatch has reported that that there were 76,068 punch card ballots for which no presidential vote was recorded. Our Election Law @ Moritz Dan Tokaji has explained why a manual recount of these ballots would be unlikely to yield more than about 20,000 extra votes for Kerry. In any event, for historical reasons, it would be interesting to conduct an analysis of those ballots, to determine what percentage actually reveal a choice for a candidate, according to Ohio's "two-corner" standard.
We are aware that additional questions could be asked about the voting process in Ohio this year, and we welcome suggestions of what inquiries to add to our list. (Please email them to firstname.lastname@example.org.) We also welcome thoughts about how best to obtain evidence that would help answer these questions.
The investigation of what happened on Election Day 2004 in Ohio is likely to be a lengthy one. But it is under way.