A View from the Recount Room

In one of the closest congressional elections in modern memory, Republican Mariannette Miller-Meeks defeated Democrat Rita Hart in the 2020 race to represent Iowa’s Second Congressional District in the U.S. House of Representatives.1 In the first canvass conducted after the polls closed, Miller-Meeks led by just 47 votes.2 At Hart’s request, officials recounted the ballots in all of the district’s 24 counties.3 After the recount, the margin of victory was even narrower: Miller-Meeks received 196,964 votes and Hart received 196,958, giving Miller-Meeks the win by a mere six votes.4

I had the honor of serving on the recount board for Johnson County, the second-largest county in the district.5 Under Iowa law, a county’s recount board consists of “[a] designee of the candidate requesting the recount,” “[a] designee of the apparent winning candidate,” and “[a] person chosen jointly by the [two candidates’ designees].”6 I held the third of those three seats. After recounting Johnson County’s 84,197 ballots—a task that took about sixty hours, running from November 17 until November 24—the board concluded that Hart had added an additional five votes to her total and Miller-Meeks had added an additional two, for a net gain of three votes for Hart in Johnson County.7

In Part I of this Essay, I briefly explain how we performed the recount. I particularly focus on the role that tabulating machines played in our work, a matter that proved to be contentious districtwide between the two campaigns’ leaders8 and on which the press’s reporting was sometimes inaccurate.9 After describing our methods, I suggest that—for the benefit of recount boards in future elections—policymakers should more clearly identify the options available to those charged with recounting large numbers of ballots in short periods of time. In Part II, I turn to Iowa laws that require recount boards to disregard votes that appear on ballots containing certain types of markings, as well as laws that limit the types of markings that recount boards are permitted to count as votes. Using examples from ballots we encountered in Johnson County, I argue that some of those rules should be amended and others should be abandoned altogether.

Throughout this Essay, I speak only for myself. I do not speak for the recount board or any of the board’s other members, for the Hart campaign or the Miller-Meeks campaign, or for any of the public servants who cheerfully and tirelessly assisted us during our eight days of work. I do not distinguish in this Essay between board decisions that were unanimous and those that were made by two-to-one votes. We made decisions of both varieties, but I describe all of them here simply as decisions of the board. When I make factual assertions about the recount and do not provide supporting citations, I am speaking based on my personal experience during the recount process.

I.     Recount Mechanics

The Hart campaign hoped the recount would reveal that, in enough instances to make a difference, what the tabulating machines counted as undervotes or overvotes in the initial canvass were actually votes for Hart.10 An undervote occurs when a voter chooses fewer than the permissible number of options for a given race, such as when a voter does not choose anyone at all in a congressional contest, while an overvote occurs when a voter chooses an impermissibly large number of options for a given race, such as when a voter chooses both candidates in a congressional race despite instructions stating that the voter may select no more than one.11 In the Hart versus Miller-Meeks contest, the machines reported there were 3,851 undervotes and 56 overvotes in Johnson County alone,12 and the Hart campaign believed there were approximately 18,000 undervotes and 200 overvotes districtwide.13 The Hart team believed that, if the machines had failed to detect what human eyes could see in even a small fraction of those cases, a recount might swing the race in Hart’s direction.

Whether human eyes in each county would indeed scrutinize the ballots, however, was a matter that each county’s recount board had to decide for itself. So far as recount mechanics are concerned, Iowa law states only that “the board shall undertake and complete the required recount as expeditiously as reasonably possible”14 and that, “[i]n counties with electronically tabulated optical scan ballots [such as those used in Johnson County], the recount board may request that the ballots be recounted by machine, may count the ballots by hand, or may do both.”15

A machine recount would do little or nothing to reveal whether the machines were failing to count humanly perceptible votes for Hart or Miller-Meeks. A hand count, in contrast, would enable a recount board to dig beneath the machines’ tallies of undervotes and overvotes, but that approach would present drawbacks of its own. Manually counting the more than 80,000 ballots in Johnson County, for example, would be extraordinarily time consuming, but state law only gives recount boards a narrow window of time to do their work.16 Even if the necessary time were available, manually counting that many ballots would risk replacing a set of machine tabulation deficiencies with a set of human ones. As I discovered during my time on the board, for example, the tabulating machines are highly reliable counters of documents: If one of those machines reports that it has processed X number of paper ballots, you can almost bet your life that it has processed X number of paper ballots. Human hands, in contrast, can sometimes struggle to detect that a pair of ballots are sticking together like new dollar bills. Fatigued eyes, hands, and minds can make other types of counting mistakes as well, particularly as one long day of sorting and counting ballots rolls into the next. For those and other reasons, the accuracy of each board member’s tabulations ought to be verified by at least one other board member, a fact that only adds to the time and human resources required to complete the task.

So, how should a recount board manage those tradeoffs? During the districtwide recount, the press reported that, in Scott County (the district’s largest17), the recount board was using machines to separate out ballots with overvotes and write-in votes, as well as “unclear ballots,” so that board members could focus their attention on just those documents, trusting the machines to accurately read and count the rest.18 The press reported that we in Johnson County were taking a comparable approach, “using the machines to separate out ballots with under votes and over votes [and] spot-checking those ballots.”19 Whatever the accuracy of the press’s description of the process in Scott County, its description of our process in Johnson County was inaccurate. Even if our board had wanted to use the machines to separate out ballots with overvotes or undervotes so that we could focus our energies on them, we could not have done so. The Johnson County Auditor and his staff advised us that their machines had not been programmed to perform that sorting function prior to Election Day, and state law does not permit anyone to reprogram the machines during a recount.20 Our machines could read the ballots and tabulate the votes using their Election Day programming, but they could not physically sort the ballots into separate voting groups.

After discussion, our board voted to proceed on a precinct-by-precinct basis, using machine recounts for some precincts and hand counts for others. The county’s precincts fall into two categories, the first spatial in nature and the second temporal. Johnson County has 57 geographically identified precincts: Solon, University Heights, Iowa City’s separately numbered precincts, and so forth.21 On Election Day, more than 23,000 Iowans cast ballots at those precincts’ polling locations.22 These were the “Election Day precincts.” Most Johnson County voters, however—nearly 61,000 of them—opted to vote early using absentee ballots.23 In accordance with Iowa law,24 we treated all of the absentee ballots as coming from a single precinct of their own, no matter where in the county they originated. This was the “Absentee precinct.”

For each of the Election Day precincts, we began with a machine recount and then shifted to a hand count if we found that, within a given precinct, there was reason to question the machine-reported totals. Proceeding one precinct at a time, the Auditor’s staff ran the ballots through the machines and then presented them to us with a printed, machine-generated tabulation of the votes in each precinct. We examined all the individual ballots—with more than one board member looking at each—searching for any reason to believe the machines might not have tabulated voters’ choices correctly. Suppose, for example, we found a ballot with a completely filled-in oval next to one candidate’s name and a dot of ink in the oval next to the other. A machine that detects both markings would count this as an overvote. A human observer, however, could dismiss the ink dot as a mere “hesitation mark” and count the ballot as a vote for the candidate with the fully darkened oval, as permitted by Iowa law.25 In 53 of the 57 Election Day precincts, we found no such difficulties and so accepted the machine totals. We did, however, find a few ballots in the other four Election Day precincts that called the machine totals into question. In each of those instances, we first decided how the problematic ballots should be treated and then tabulated that precinct’s ballots by hand.

Given its enormous size, the Absentee precinct presented a tougher challenge. By the time our work brought us to the foot of that mountain, I had seen that voters do sometimes mark their ballots in ways that a human can interpret more reliably than a machine. Hesitation marks of the type I just described are one case in point, but there are others. If a voter fills in the ovals next to two competing candidates’ names and then scribbles one of them out, for example, the machines will detect two ink-marked ovals and count the ballot as an overvote, while a human can perceive that the voter did not intend to cast a vote for the candidate with the scribbling.26 Moreover, there was reason to believe we might encounter such issues at a greater rate in the Absentee precinct than in the Election Day precincts. When you make a mistake on your ballot at a polling place, it is easy to obtain a replacement ballot.27 Poll workers might even remind you at the outset that this option is available. But if you make a mistake when filling out an absentee ballot at home, you might not have the time or inclination to obtain a replacement, assuming the possibility even occurs to you.28

Manually counting nearly 61,000 ballots within the short period of time allotted by law, however, seemed risky. When we performed hand counts in four Election Day precincts, the relatively small numbers with which we were dealing made it easy to confirm that our totals were reliable. We could readily count and recount our sorted stacks of ballots as many times as necessary, and we could use the machine tallies as benchmarks against which to compare our work.29 Tackling a precinct containing nearly 61,000 ballots was a different story. Suppose, for example, that we spent days manually sorting and counting those ballots and concluded that there were, say, 60,924 of them, and suppose the machines—machines that count documents with exceptional accuracy—told us that they had counted 60,934. What would we make of the ten-vote difference? Might we have overlooked a stack of ten ballots somewhere?30 Might we have failed to detect instances in which ballots were stubbornly stuck together? Even if we had another set of days to manually count all those ballots again, there would be no guarantee that our second count would produce results more reliable than the first or that we would have enough time to complete the job. And there were other risks, as well. Given the volume of ballots with which we were dealing, it was not inconceivable that we would mistakenly place at least one or two ballots in the wrong stacks when doing our sorting and counting. In a race as close as this one, any such mistakes could undermine the county’s effort to help determine the proper districtwide winner.

After consulting with the Johnson County Attorney—who, in turn, consulted with counsel in the Secretary of State’s office to ensure the board would be proceeding within the bounds of what the Secretary of State deemed legally permissible—the board opted to handle the Absentee precinct using a machine-assisted hand-count procedure that worked as follows. The Auditor’s staff ran the ballots through the machines in small batches of about 200 ballots each and then delivered them to us with machine-generated reports of the vote totals in each batch. We then examined each of the ballots individually, applying all the ballot-evaluation rules that apply to hand counts and looking for issues of any kind that might prompt us to question or reject the machine totals.31 Although our study of the ballots entailed more than just looking for ballots that the machines had problematically tabulated as undervotes and overvotes,32 we did manually count the undervotes and overvotes in each batch and then compare those numbers to the machines’ own tallies, since deviations between those two sets of numbers would be one indication that the machine tallies might be unreliable. (If we counted 15 undervotes in a batch and the machines had counted 16, for example, we would know that our eyes might be detecting something the machines had missed.) When we spotted issues that might prompt us to reject the machine totals, we decided how to resolve those issues and then, if appropriate, we manually tallied all of the ballots in that batch to ensure we had reached conclusions that conformed with our judgments.

In my view, that approach to the Absentee precinct worked well. Indeed, I would recommend it to future boards that are tasked with recounting large numbers of ballots and do not wish to do so simply by running the ballots through the tabulating machines a second time and then accepting whatever results the machines alone produce. It bears emphasizing, however, that our board devised its methodology (1) amidst publicized disagreements between statewide leaders of the Hart and Miller-Meeks camps about what the law does and does not permit recount boards to do33 and (2) after the board consulted with the County Attorney and the County Attorney consulted with counsel in the Secretary of State’s office. Going forward, it would be better if the state’s published regulations made it unambiguously clear at the outset that this approach is among the options from which recount boards may choose.34 Such clarity could help speed the process along when boards are charged with recounting large numbers of ballots in a short amount of time, and it might also help to depoliticize the public’s perception of some of the decisions that recount boards make.

*  *  *  *

As I have indicated, evaluating ballots that the machines might have inaccurately counted as undervotes or overvotes was one component of the work that we did, but our judgments extended more broadly than that. Iowa law contains rules under which a recount board must disregard votes cast on ballots bearing certain types of markings, as well as rules limiting the types of ballot markings that a recount board may count as valid votes. Because we opted to visually inspect all the Johnson County ballots rather than simply run them through the machines a second time and accept the resulting tabulations, our work brought us face-to-face with those rules on numerous occasions. Those encounters sometimes led to legally required outcomes that would surprise many Iowans and that I believe are difficult to defend on policy grounds. It is to those rules that I now turn.

II.     Recount Rules

The overwhelming majority of ballots in Johnson County presented no difficulties whatsoever: Almost all voters cleanly darkened the ovals corresponding to their choices and made no other markings on their ballots. A handful of voters, however, filled out their ballots in ways that raised issues for the recount board to resolve. I discuss three such sets of issues here: one concerning voters’ use of “identifying marks,” one concerning voters’ ballot markings that deviate from the ballot’s instructions, and one concerning voters’ use of the space for write-in votes.

A.     Identifying Marks

Imagine you are filling out your ballot, whether at home or at a polling place. You clearly mark your choice in the congressional race but then, when making your selection in another contest, you mistakenly fill in the oval next to a candidate whom you oppose. You could obtain a replacement ballot,35 but that possibility either does not occur to you or strikes you as a hassle. You decide simply to scribble out the mistake, place your initials next to the scribbling to assure anyone who sees it that the scribbling is yours, and fill in the oval next to the name of your preferred candidate. Or perhaps you fill out your ballot flawlessly but then sign it at the bottom because you want election officials to know it comes from an eligible voter. Or maybe you write your contact information on the ballot so that election officials know how to reach you if the ballot presents any difficulties. Or—shifting gears—perhaps you mark your choice in the congressional race but then are irked to see that, in some of the other contests, the only listed candidates are members of, say, the Democratic Party. You are a devout Republican and so, in a fit of pique, you write “Any Republican!” on the write-in lines assigned for those races.

The Johnson County recount board encountered varieties of each of those examples. In each instance, the tabulating machines had read and reported the votes in the Hart versus Miller-Meeks race. Sometimes, however, we concluded that we were legally obliged to disqualify the ballots on which those votes appeared. The laws necessitating those disqualifications have deep roots in American and Iowa law. And they ought to be abandoned.

1.     The Law of Identifying Marks

For ballots of the sort I have described, the difficulty lies in what the law calls “identifying marks.” To understand these laws’ origins, we need to return briefly to the raucous world of electoral politics in nineteenth-century America. During much of the 1800s, governments neither prepared the ballots that voters used nor provided voters with secrecy when voting. At the beginning of that century, voters often made their own ballots at home, then brought those ballots with them to the polling place on Election Day.36 Before long, however, party leaders and others hoping to round up votes saw an opportunity. They could prepare stacks of ballots bearing their desired selections, print those ballots on paper that was distinctively colored or otherwise identifiable from a distance, distribute those ballots among potential supporters, and even try to press those ballots into the hands of voters on Election Day as they approached the polling place.37 Because the ballots were easily spotted from afar and voters did not submit their ballots in secret, everyone—employers, incumbents, candidates, neighbors, friends—could watch to see which visibly distinguishable ballot a person opted to use.38 This opened the door for rampant voter coercion,39 and it also enabled people to sell their votes to the highest bidder: After striking the deal, the buyer could simply watch to be sure the seller followed through.40

Elections in Iowa were no exception. Emory English, who served several terms in the Iowa legislature just after the turn of the century,41 recalled that Iowa voters in the late 1800s often found themselves surrounded by aggressive vote-seekers:

[T]he voters . . . were bewildered and annoyed by the multiplicity of tickets offered them, some being thrust into their hands even when they were in the act of voting. All manner of pressure and intimidation were practiced, and high-handed methods of voting and counting of votes were indulged in and tolerated. . . .

    . . . .

    Voting conditions [became] intolerable; the imposition and intimidation practiced were unbearable . . . .42

Other democratic countries were afflicted with comparable problems until officials in Australia unveiled a very different approach to conducting elections.43 Under the new Australian model, governments themselves would print ballots listing the various contests and the options available for each.44 They then would bar voters from submitting ballots of any other kind and would provide voters with a private setting in which to make and submit their selections.45 American states flocked to the Australian approach,46 with Iowa joining them in 1892.47

Even with secretly cast votes on government-provided ballots, however, people could still find buyers for their votes. A voter could tell a prospective buyer that he (the voter) would cast his ballot for whomever the buyer liked. Of course, a sensible buyer would want proof that the voter did what he was paid to do. To provide the buyer with this assurance, the seller could pledge to mark his ballot in some distinctive way. By looking for the voter’s identifying mark on the submitted ballots, the buyer could later determine whether the voter had held up his end of the bargain.48

That possibility did not escape the attention of authorities in Iowa and elsewhere.49 As part of its 1892 reforms, the Iowa legislature established criminal penalties for any voter who “mark[s] his ballot . . . for the purpose of identifying said ballot.”50 The Iowa Supreme Court shortly thereafter determined that ballots marked in violation of the 1892 legislation “should be rejected.”51 The legislature embraced that view just a few years later, declaring in the 1897 Code that “[a]ny ballot marked by the voter in any other manner than as authorized in this chapter, and so that such mark may be used for the purpose of identifying such ballot, shall be rejected.”52 As the Iowa Supreme Court explained in 1916’s Fullarton v. McCaffrey,53 the central purpose of these laws was to make it harder for people to find buyers for their votes:

The distinguishing mark prohibited by law is one which will enable a person to single out and separate the ballot from others cast at the election. It is something done to the ballot by the elector designedly and for the purpose of indicating who cast it, thereby evading the law insuring the secrecy of the ballot. . . . The design of the Australian Ballot Law is not only to free the voter from intimidation by making his way of voting known only to himself, but to close the door securely against making merchandise of his vote as nearly as human ingenuity can.54

Iowa’s high court gave the same explanation half a century later, writing that “[t]he statute barring identifying marks originated from a desire to guard against the possibility of a vote seller indicating to a vote buyer in advance how his ballot could be distinguished so the buyer could determine if the bribed voter had carried out his agreement.”55

Iowa law today continues to bar the placement of identifying marks on ballots. One statute declares that “[a] ballot shall be rejected if the voter used a mark to identify the voter’s ballot.”56 Another states that, if a voter fills out a ballot at a polling place and then hands it to an election official, “[a]n identifying mark or symbol shall not be endorsed on the voter’s ballot.”57 Another brings the force of criminal law to bear, stating that “[a] person commits the crime of election misconduct in the third degree if the person willfully . . . [m]ark[s], or caus[es] in any manner to be marked, on a ballot, any character for the purpose of identifying such ballot.”58

The Iowa Administrative Code builds upon that statutory foundation, stating that a ballot must be rejected if it “includes an identifying mark.”59 The code defines an identifying mark as (1) “[a] comment or statement that indicates the identity of the voter either individually or as a member of a group” or (2) “[i]nitials, a printed name or a signature placed on the ballot in any place other than on the lines intended for write-in votes or intended for the initials of the election official who issued the ballot.”60 The administrative code thereby even goes further than the statutes, insofar as it bans counting ballots with identifying marks regardless of the purposes for which those marks were made. Another administrative code provision permits voters to correct mistakes on their ballots, but only “if the correction does not include an identifying mark.”61 Lest there be any doubt on the matter, the code provides an illustration in which a voter has darkened an oval, placed an “X” through it to cancel it, placed his or her initials next to the change, and then filled in the oval corresponding to the candidate of her choice.62 “This example does not show a vote,” the administrative code states, because “[t]he initials next to the correction identify the voter.”63 The code elsewhere provides another illustration to reiterate that a ballot is disqualified if the voter writes something on it that “identifie[s] the political affiliation of the voter” or amounts to a statement of “political identification.”64

2.     Identifying Marks in Our Recount

Pursuant to the legal rules we were obliged to apply, our recount board disqualified ballots on these grounds: ballots with initials, ballots with names or signatures, and ballots with statements of political affiliation. In my judgment, the law on these matters was clear. In each of those instances, however, it also seemed clear that the cure was far worse than the disease. Although perhaps it is conceivable that an Iowa voter today would use an identifying mark in a bid to facilitate a vote-selling scheme, it seems exceedingly unlikely—certainly not sufficiently likely to justify trying to prevent it by disqualifying all ballots bearing identifying marks. Ballots today are ordinarily handled with great care, by very small numbers of people, under secure conditions. They are stored in boxes or bags with sealing labels signed by multiple people65 and they usually are counted using machines rather than human eyes.66 In Johnson County, the three members of the recount board were probably the first individuals—apart from the voters themselves—to see that some of the county’s ballots did indeed have identifying marks. Given the difficulty of using such marks to further a vote-selling scheme under modern voting conditions, I have no doubts about the innocence of the Johnson County voters whose ballots got trapped in Iowa policymakers’ nets. If we are concerned about vote-selling today, we ought to tackle the problem directly using anticorruption legislation aimed at those whose behavior is demonstrably corrupt, and not sweep up innocent voters in the process.67

Let me briefly provide just a few examples to illustrate the point. On one Johnson County ballot, we discovered that a woman had elegantly penned her full name, her address, and her telephone number. This surely was a voter who wanted to be certain that election officials knew she was an eligible voter and would know how to reach her if they had any difficulty processing her ballot. If any ballot could ever be deemed to bear an identifying mark, it was this one. The law compelled us to disqualify her vote. As we did each time we made such a decision, we walked the ballot over to the recount monitors from the Hart and Miller-Meeks campaigns and to the monitors that the majority and minority parties in the U.S. House had sent. The bipartisan groan that emerged from that group was, to my ears, the same sympathetic “ohhh” that one utters when hearing a story about a person who unwittingly stumbles into trouble when trying to do the right thing. Here was a woman whose determination to be sure her voice was heard prompted her to mark her ballot in a manner that required us—so far as the congressional race was concerned—to silence her voice altogether.

Although most of the identifying marks we saw were lesser degrees of the one I just described, a few echoed the language of partisan politics. On many Johnson County ballots, there were three races—state representative, auditor, and sheriff68—for which the only named candidates were Democrats. For these races, at least one voter used the write-in line to write in “Republican” and then darkened the oval next to that line. A couple of other voters wrote in “Republican” or “Any Republican!” on those lines but did not darken the accompanying ovals. Bearing in mind the Iowa Administrative Code’s explanation that statements of political identification or affiliation are identifying marks,69 should these ballots be disqualified? Were these statements of political identification or were they simply declarations that, for those particular offices, Johnson County would be better served by Republicans? Building on the rule that something written on a write-in line counts as a write-in vote only if the accompanying oval is marked,70 we did not disqualify the ballot on which the oval had been darkened, but we did disqualify the ballots on which those ovals were unmarked. One rationale for that distinction was that, in the latter instances, the text on the write-in lines was functionally no different from text that a voter could write in any white space on the ballot, and the words “Republican” or “Any Republican!” elsewhere on the ballot would be an identifying mark under Iowa law. In my judgment, that was a workable distinction serving no good objective other than obeying legal rules. With or without darkened ovals, those words did nothing to facilitate vote-selling or any other form of corruption. The November 2020 election ought to be the last in which this legal trap awaits the unsavvy voter.

B.     Prescribed Marks

In the November 2020 election, Johnson County’s ballots instructed voters to use a pen to “completely fill in the oval next to” their choices.71 In the language of election law, a darkened oval was thus the “prescribed mark” by which voters were to indicate their selections, and the oval that voters were to darken was the “voting target.”72 The vast majority of Johnson County voters used the prescribed mark when making their choices. Other voters, however, took more idiosyncratic paths. Some of those paths led to trouble.

The Iowa Administrative Code contains a provision that, when read in isolation, suggests voters have significant latitude to choose how to mark their votes: “A vote for any office or question on a ballot shall not be rejected solely because a voter failed to follow instructions for marking the ballot,” the code states, and “[i]f the choice of the voter is clear from the marks for any office or question, the vote shall be counted as the voter has indicated.”73 But the code also contains more specific rules that recount boards must follow when determining what is and is not a valid vote.74 Two of those rules read as follows:

“If a voter uses both the prescribed mark and other marks, only the prescribed marks shall be counted as votes.”75

“If a voter does not use the mark prescribed in the voting instructions but consistently uses some other mark, the mark shall be counted as a vote if the mark is [in or near the voting target or is near a candidate’s name].”76

Some Johnson County voters in the November 2020 election filled out their ballots with something other than the prescribed mark, but they did so in ways that did not put their votes in jeopardy. Some, for example, placed an “X,” a check mark, or a small but perceptible dot in the ovals corresponding to their selections. These were not prescribed marks but, because they appeared within the voting targets, the machines were able to tally them as votes for the designated candidates. And because the voters used these marks consistently when signaling their selections, we were able to count these as votes as well.

A handful of other voters, however, marked their ballots in more problematic ways. A couple of individuals, for example, used fully darkened ovals to register some of their choices, but then placed an “X” in the ovals for other races on the ballot, including the ovals for the Hart versus Miller-Meeks contest. The machines read these as votes because the machines simply look for ink markings within the voting targets. Pursuant to the rule stating that only prescribed marks count as votes when a voter uses both the prescribed mark and some other mark, however, the board concluded that these ballots had to be moved to the undervote category, the category of ballots containing no selections at all in the congressional contest.

A couple of voters placed an “X” through the voting targets for all of their apparent selections—so far, so good. But then they also darkened the ovals through which some of those “X”’s had been drawn, while leaving other “X”-marked ovals—including the ones for our congressional race—undarkened. The machines counted these as votes because they were able to detect that markings had been made in the voting targets. But what was the recount board to do? On the one hand, the voters had fully darkened some of the voting targets elsewhere on the ballot, but not in the targets for our race. Did this trigger the rule that, if the prescribed mark is used for some selections but not others, then none of the markings that deviate from the prescribed mark count as votes? On the other hand, did these voters indeed use the prescribed mark anywhere on their ballots? Rather than simply fill in some of the ovals, they filled in those targets and drew an “X” through them. So should the board rule that, because the voter had consistently used an “X” throughout the ballot, a mere “X” was sufficient to register a vote even if it was not accompanied by a fully darkened oval? The board took the latter position and counted the votes.

Another ballot was (for me, at least) even more disquieting. This voter, who came to us from one of the Election Day precincts, used the prescribed mark to register his or her selections in nearly all the races and questions that appeared on the ballot. In our congressional race, however—and only in our congressional race—the voter’s mark conspicuously missed the voting targets entirely: The voter’s oval-sized mark sat to the right of a voting target, midway between the target and its accompanying candidate name. In my judgment, the likeliest explanation for the mark’s placement was that the voter simply made a ballot-marking mistake: He or she had missed the voting target but intended to cast a ballot for the candidate whose target and name stood directly to the left and right, respectively, of the voter’s mark.77 Because the mark sat entirely outside the voting target, however, it was not the prescribed mark, and under the rules a non-prescribed mark cannot count as a vote unless the voter uses that mark “consistently.” If this voter had consistently signaled his or her choices by placing marks between voting targets and candidate names, the machines would not have read those as votes, but the recount board could have counted the mark as a vote for the congressional candidate whose name the mark adjoined.78 Our race, though, was the only one in which the voter had marked the ballot in this fashion. The board concluded that the rules did not permit it to count the mark as a vote, and so we moved the ballot to the undervote column.

We certainly do need standards for discerning voter intent in order to ensure that all ballots cast in a race are evaluated on an equal basis. It would not suffice to tell 24 recount boards in 24 counties that they each are to discern voter intent by whatever standards they independently deem appropriate.79 But there is considerable distance between (1) taking that approach and (2) telling recount boards that they cannot count a non-prescribed mark as a vote unless the voter consistently uses that same mark throughout the ballot. There are middle-ground positions available, and I will suggest one of them here. The following proposal would not have helped the Election Day voter who made one wayward mark, but it would have provided us with clear authority to count the other controversial ballots I have described.

When a voter makes a mark of any kind within a voting target, the law ought to presume that the voter intended the mark to count as a vote, unless a recount board finds that the mark was either a hesitation mark80 or a mark the voter tried to cancel.81 That, after all, is the general rule that the tabulating machines apply: They read marks of any kind in the voting targets and they tabulate election results accordingly. In most cases, those results stand without challenge. Why should Iowans be at risk of having those machine-tabulated votes taken away if there is a recount and the recount board opts to examine all the ballots one by one? Suppose, for example, a voter darkens the voting targets in some races and places an “X” in the voting targets for others. Why did the voter do this? We don’t know. Perhaps he or she is obsessively devoted to the Emersonian proposition that “[a] foolish consistency is the hobgoblin of little minds.”82 Because we can only speculate about the voter’s thinking, we need rules that align with the most plausible explanation. The likeliest explanation, in my view, is that the voter deployed two different methods of indicating her choices.

Votes are sacred things. We celebrate when individuals take the time necessary to add their own drops to an ocean of ballots. With that in mind, it seems far better to tell voters that they make unintended marks within voting targets at their peril—and so, for example, one should be cautious about making hesitation marks that could lead to machine-tabulated overvotes that a recount board might never have the opportunity to cure83—than to tell voters that they act at their peril if, within voting targets, they intentionally make marks that a recount board might later deem invalid on grounds of inconsistency.

C.     Write-In Votes

When a ballot lists the names of candidates for an office and then provides space for a write-in vote, I am not sure why anyone who wishes to vote for one of the listed candidates would disregard the voting target next to her name and opt to write that name on the write-in line instead. But if you pull together a group of 80,000 voters, there evidently will be some who feel drawn to that path. In Johnson County, several voters wrote in Hart’s or Miller-Meeks’ name, rather than simply fill in the oval next to the candidate’s listed name. Perhaps for these voters it felt like an especially unambiguous, emphatic way to declare their preferences. If that explanation is right, then it also is sadly ironic because, for some of these voters, their decision to take the write-in route led to their apparent vote attempts being disregarded altogether.

In a few instances, Johnson County voters filled in the voting target next to Hart’s or Miller-Meeks’ listed name, but then also filled in the oval next to the write-in line and wrote in the same candidate’s name. With two darkened voting targets, the machines counted these ballots as overvotes rather than as votes for the doubly identified candidates. Because we reviewed each ballot, we were able to reclassify these as votes for the designated candidates.84 There were other ballots, however, that the law obliged us to treat in a more problematic fashion. In these instances, the voters wrote Hart’s or Miller-Meeks’ name on the write-in line but then did not mark any voting target—neither the one next to the write-in line nor the one next to the candidate’s already-listed name. Because these voters did not darken any of the voting targets for the congressional race, the machines classified these as undervotes—i.e., as failures to cast any vote at all. If there had been no recount, these voters—with respect to the congressional race, at least—would have had no impact. But once the recount board spotted these ballots, what were we to do? Should we join the tabulating machines in treating these as undervotes benefitting no candidate, or should we count them as votes for the candidates whose names the voters wrote in?

If left to my own devices, I would count these as votes for the written-in candidates. To be sure, the ballots are clouded in some uncertainty; I cannot be sure what these voters were thinking. But if asked to provide the likeliest explanation, I would say these voters handwrote the names of the candidates for whom they intended thereby to vote, but they neglected to mark the oval next to the write-in line. That seems a far better explanation than the hypothesis that these voters either (a) handwrote the names of the candidates they opposed or (b) took the time to handwrite the names of tentatively preferred candidates but then had second thoughts and so left the accompanying voting target blank in order to avoid casting any vote at all.

In any event, the law did not leave us free to choose from among those possibilities. The legislature has said that writing in a person’s name without marking the accompanying voting target counts as a vote only if “the voter is using a voting system other than an optical scan voting system.”85 But if a precinct is using optical scan equipment, then the vote counts only if the voter has also marked the oval next to the write-in line.86 Does this rule also apply to recounts? Yes. The Iowa Administrative Code states that “[t]he precinct election officials and recount board members shall count a write-in vote only if the voting target is marked.”87

The rule requiring voters to mark the write-in line’s voting target when attempting to register a write-in vote makes perfect sense when what one is talking about is the initial vote canvass. If a precinct uses optical-scan technology so that election officials can count votes quickly and with a very high degree of accuracy, it would make no sense to require them to sift through all of the ballots to find write-in votes lacking a marked oval. This is particularly true, for heaven’s sake, when it comes to write-in votes for candidates whose names already appeared on the ballot. But when it comes to recount boards—boards that are called into action only in very close elections—the rule ought to be different. If a voter has, for whatever reason, disregarded a name that already appears on the ballot and has opted instead to write in that name on the write-in line, then a recount board should be free to count it as a vote for the named candidate. The law should favor the view, in other words, that the voter probably intended to cast a ballot for that candidate but simply forgot to mark the accompanying voting target. Such a rule would do nothing to help these voters in ordinary elections—and so the rule itself would not incentivize voters to take this ill-advised path—but it would help them when recount boards are constituted and choose to examine each individual ballot with care.

III.     Conclusion

As the United States Supreme Court observed more than half a century ago, “[n]o right is more precious in a free country than that of having a voice in the election of those who make the laws under which, as good citizens, we must live.”88 Though the various people in Johnson County’s recount room had different roles to play in the November 2020 election, their shared commitment to fundamental democratic ideals was, for me, a wonderfully bright light in a politically acrimonious, pandemic-darkened year. The experience also persuaded me, however, that there are ways in which the law could serve those democratic ideals more fully. In addition to spelling out more explicitly the recount methodologies that recount boards are permitted to deploy in their effort to do quick and reliable work,89 policymakers should abandon the rule that disqualifies ballots bearing identifying marks,90 loosen the rule concerning prescribed marks,91 and permit recount boards to count written-in names of ballot-listed candidates as votes for those candidates regardless of whether the accompanying voting targets have been darkened.92

The number of voters who were adversely affected by the rules I have questioned here was small—tiny, in fact, when taken as a proportion of the overall number of ballots cast. But the 2020 election in Iowa’s Second Congressional District reminds us that tiny numbers can sometimes make all the difference. And even more important than the outcome in any particular race is the principle that, when elections are exceptionally close and recount boards are called into action, every eligible Iowan’s vote deserves to be counted unless there are weighty reasons to disregard it. Some of the election rules in place in Iowa today fall short of meeting that mark.

  1. Tom Barton, Miller-Meeks Wins Iowa Congressional Race by 6 Votes. A Legal Challenge Is Likely., Quad-City Times (Dec. 15, 2020), [] (reporting that Miller-Meeks had won by “the slimmest margin in any congressional race since 1984, when Indiana’s 8th Congressional District was decided by four votes”). The seat previously had been held by Democrat Dave Loebsack, who stepped aside after fourteen years of service in the House rather than run for reelection. See Laura Barrón-López, Democratic Rep. Dave Loebsack to Retire in 2020, Politico (Apr. 12, 2019, 5:45 PM),
    retire-2020-1273465 [].

         On December 22, 2020, Hart filed a petition with the U.S. House of Representatives, asking that body to declare that she was the rightful winner. See Hart v. Miller-Meeks, Notice of Contest Regarding the Election for Representative in the One Hundred Seventeenth Congress from Iowa’s Second Congressional District at 3–6 (H.R. Dec. 22, 2020),
    wp-content/uploads/sites/45/2020/12/IA-02-Notice-of-Contest-and-Appendix-Scan1.pdf [https://]. See generally U.S. Const. art. I, § 5, cl. 1 (“Each House shall be the Judge of the Elections, Returns and Qualifications of its own Members . . . .”); L. Paige Whitaker, Cong. Rsch. Serv., RL33780, Procedures for Contested Election Cases in the House of Representatives (2016), [https://] (describing the procedures and standards that apply to such challenges). Hart’s petition remains pending at the time of this writing. By stating above that Miller-Meeks defeated Hart, I do not mean to take any position on the merits of Hart’s petition. I mean simply to describe the election results as they currently stand.

  2. [2]. See Tom Barton, Mariannette Miller-Meeks Leads by 47 Votes Ahead of Recount in House 2 Race: Recount All but Certain in Race for Who Succeeds Dave Loebsack, Gazette (Nov. 11, 2020), [].

  3. [3]. See Tom Barton, Hart Campaign Outlines Next Steps in Iowa 2nd Congressional District Recount, Quad-City Times (Dec. 1, 2020), []; see also Iowa Code § 50.48(1)(a) (2020) (stating that a “county board of canvassers shall order a recount of the votes cast for a particular office” upon the request of any person whose name appeared on the ballot for that office). Iowa’s second congressional district consists of twenty-four counties in the southeast portion of the state. See District Maps & Data (Legislative & Congressional), Iowa Legislature, [] (providing links to district maps).

  4. [4]. See Iowa Sec’y of State, Election Canvass Summary 31 (2020),
    elections/pdf/2020/general/canvsummary.pdf [].

  5. [5]. See QuickFacts, U.S. Census Bureau,
    /clintoncountyiowa,scottcountyiowa,johnsoncountyiowa,IA [] (stating that the population of Scott County is approximately 173,000 and the population of Johnson County is approximately 151,000). Clinton County is a distant third, with a population of approximately 46,000. See id.

  6. [6]. Iowa Code § 50.48(3)(a). If the candidates’ two designees cannot agree on the third member, the chief judge of the judicial district must choose a person to fill the board’s final seat. See id. § 50.48(3)(b).

  7. [7]. See November 3, 2020 General Election, Johnson Cnty.,
    dept_auditor_elections.aspx?id=27268 [] [hereinafter 2020 General Election] (reporting the total number of ballots cast in Johnson County, the results of the initial November 10 canvass, and the results of the November 24 recount).

  8. [8]. See Tom Barton, Miller-Meeks Campaign Alleges Missing Votes, ‘Illegal’ Method in Scott County Recount, Quad-City Times (Dec. 8, 2020),
    070-a333-5db19e058b30.html [] (reporting on the disagreement between the two campaigns); Zachary Oren Smith, ‘Not Looking Pretty Right Now’: 2nd District Recount Continues, Campaigns Fret Approach, Iowa City Press-Citizen (Nov. 20, 2020, 7:00 AM),
    2nd-district-recount-hart-miller-meeks-us-house-race/3770666001/ [] (same).

  9. [9]. See infra note 19 and accompanying text.

  10. [10]. See Smith, supra note 8 (reporting that the Hart campaign wanted recount boards to take a close look at ballots that the voting machines regarded as undervotes and overvotes).

  11. [11]. See Iowa Admin. Code r. 721-26.1(49) (2018) (defining “[o]vervote” and “[u]ndervote” (emphasis omitted)); see also Iowa Code § 49.93 (2020) (“If a person votes for more than the permitted number of candidates, the vote for that office shall not count.”).

  12. [12]. See 2020 General Election, supra note 7 (reporting the total number of undervotes and overvotes in the Miller-Meeks/Hart race, and reporting that there were 68 write-in votes, as well).

  13. [13]. See Michaela Ramm, Fact Checker: Did Mariannette Miller-Meeks Seek to ‘Limit’ the Vote Count?, Gazette (Dec. 7, 2020), [] (reporting this claim by the Hart campaign).

  14. [14]. Iowa Code § 50.48(4)(a) (2020).

  15. [15]. Iowa Admin. Code r. 721-26.105(2).

  16. [16]. See Iowa Code § 50.48(3)(b) (establishing deadlines that ensure the recount board is formed “not later than 5:00 p.m. on the eleventh day following the canvass”); id. § 50.48(4)(c) (“The recount board shall complete the recount and file its report not later than the eighteenth day following the county board’s canvass of the election in question.”). The fact that each county gets one three-member recount board to do the work necessary to meet the statutory deadline, even when there are enormous differences in the number of ballots that each county must recount, is an issue that the legislature might consider when contemplating possible reforms.

  17. [17]. See QuickFacts, supra note 5.

  18. [18]. See Barton, supra note 8.

  19. [19]. See Tom Barton, Mariannette Miller-Meeks’ Lead Increases a Bit in Scott County Recount: She Picks Up 3 Votes, Rita Hart Picks Up 1, in Nation’s Closest Congressional Race, Gazette (Nov. 18, 2020),
    a-bit-in-scott-county-recount-20201118 []; see also Zachary Oren Smith, IA-02: Three Counties Use Voting Machines to Assist Hand Recount, Defying Secretary of State Opinion, Iowa City Press-Citizen (Nov. 24, 2020, 9:24 AM),
    com/story/news/politics/elections/2020/11/23/scott-johnson-clinton-county-defies-secretary-state-machine-assisted-hand-recount/6386162002 [] (inaccurately reporting that Johnson County was “defying” the Secretary of State and that the board was focusing only on ballots containing undervotes, overvotes, and write-in votes).

  20. [20]. See Iowa Code § 50.48(4)(a) (“If automatic tabulating equipment was used to” conduct the initial ballot count, “[t]he same program used for tabulating the votes on election day shall be used at the recount unless the program is believed or known to be flawed.”).

  21. [21]. See November 3, 2020 General Election Polling Places and Sample Ballots, Johnson Cnty., [
    /B66D-2BR4] (listing the 57 precincts and their November 2020 polling locations).

  22. [22]. See Johnson Cnty., Official Canvassed Results (2020), https://www.johnson []. The number of ballots in each precinct averaged just over 400, ranging from a low of around 120 to a high of more than 1,000. See id. For reflections on absentee ballots and the legal issues that arise when states restrict their availability, see Sal H. Lee, Note, Judicial Review of Absentee Voting Laws: How Courts Should Balance State Interests Against the Fundamental Right to Vote Going Forward, 105 Iowa L. Rev. 799 (2020).

  23. [23]. See Johnson Cnty., supra note 22; see also Stephen Gruber-Miller, Monday Is the Last Day to Vote Early in Iowa. Here’s What You Need to Know., Des Moines Reg. (Nov. 2, 2020, 11:49 AM), [] (reporting that Iowans “have been taking advantage of early voting in record numbers this year as the coronavirus pandemic continues”).

  24. [24]. See Iowa Code § 53.23 (2020) (providing for the treatment of absentee voters as members of a single precinct); Iowa Admin. Code r. 721-26.2(2) (2018) (same).

  25. [25]. See Iowa Admin. Code r. 721-26.1(49) (“‘Hesitation mark’ means a small mark made by resting a pen or pencil [tip] on the ballot.”); id. r. 721-26.17(49) (stating that hesitation marks “shall not count as votes”).

  26. [26]. See id. r. 721-26.21(49) (“A vote for an office or question shall be counted if the voter has marked the ballot in a manner that will be counted as an overvote by automatic tabulating equipment but the voter has indicated in a clear fashion that the voter has made a mistake.”).

  27. [27]. See Iowa Code § 49.100 (2020) (“A voter who spoils a ballot may return the spoiled ballot to the precinct election officials and receive another ballot.”).

  28. [28]. Replacement ballots are indeed available for absentee voters who “spoil” their ballots. See id. § 53.21(3).

  29. [29]. Our handling of the North Liberty 4 precinct illustrates the benchmark point. The machine recount indicated that 501 ballots had been cast in that precinct: 237 votes for Hart, 229 votes for Miller-Meeks, one write-in vote, one overvote, and 33 undervotes. When we examined those ballots, we found the one that we believed the machine had counted as an overvote. In the judgment of the board, that ballot actually contained a vote for Miller-Meeks. To enable us to take account of that judgment, we moved North Liberty 4 from the machine-recount category to the hand-count category and then manually counted all the ballots cast in that precinct. Our hand count indicated that there were 501 ballots: 237 votes for Hart, 230 votes for Miller-Meeks, one write-in vote, zero overvotes, and 33 undervotes. The machine’s own tabulations provided a source of reassurance that we had accurately counted the ballots.

  30. [30]. When performing hand counts, our board followed the “[b]est [p]ractice” advised by the Secretary of State and “sort[ed] the ballots into piles of 10.” Off. of Iowa Sec’y of State, Recount Board Guide 9 (2014) (emphasis omitted) (on file with author).

  31. [31]. As the county attorney explained to us after she consulted with counsel in the Secretary of State’s office, this wholistic examination of each ballot in the Absentee precinct, applying all of the ballot-evaluation rules for hand counts, was necessary to bring our procedure into compliance with the Secretary of State’s guidance. The press evidently did not know the full scope of the legal guidance we were given. See Smith, supra note 19 (“Three county recount boardsare defying a recent legal opinion from the Iowa Secretary of State’s Office and usinga machine to aid the recount of ballots inthe ultra-close 2nd District congressional race.”).

  32. [32]. See, e.g., infra Section II.A (discussing “identifying marks”).

  33. [33]. See Barton, supra note 8; Smith, supra note 19.

  34. [34]. I am assuming here that the state wants to continue to allow counties to make differing recount-method choices. See infra note 79 and accompanying text (observing that this feature of Iowa law is in some tension with equality values).

  35. [35]. See supra notes 27–28 and accompanying text.

  36. [36]. See Burson v. Freeman, 504 U.S. 191, 200–02 (1992) (plurality opinion) (describing this and other nineteenth-century practices recounted in the paragraph above); Spencer D. Albright, The American Ballot 19 (1942) (“Ballots in the early part of the nineteenth century were hand-written ... .”); Jerrold Glenn Rusk, The Effect of the Australian Ballot Reform on Split Ticket Voting: 1876–1908, at 14(1968) (Ph.D. dissertation, University of Michigan) (ProQuest) (noting the early-century practice of preparing one’s own ballot at home).

  37. [37]. See Albright, supra note 36, at 20–21 (briefly describing these developments); Eldon Cobb Evans, A History of the Australian Ballot System in the United States 6–16 (1917) (same); Rusk, supra note 36, at 14–23, 38–39 (same).

  38. [38]. See Evans, supra note 37, at 10–11; L.E. Fredman, The Australian Ballot: The Story of an American Reform 22 (1968).

  39. [39]. See Evans, supra note 37, at 11–13; Fredman, supra note 38, at 24–25.

  40. [40]. See Evans, supra note 37, at 11; Fredman, supra note 38, at 22.

  41. [41]. See Representative Emory H. English, Iowa Legislature,
    legislators/legislator?ga=29&personID=3463 [].

  42. [42]. Emory H. English, Evolution in Iowa Voting Practices, 29 Annals Iowa 249, 254 (1948).

  43. [43]. See Albright, supra note 36, at 23–24 (discussing the Australian model’s influence in England, Belgium, Italy, Norway, and other countries).

  44. [44]. See Evans, supra note 37, at 17 (describing the Australian system); Fredman, supra note 38, at 46 (same). The introduction of the Australian model was not without downsides. Government-issued ballots made it harder for illiterate individuals to vote, for example, because they could no longer rely simply on such things as paper color to be sure they were voting for their preferred candidates. See Andrew Gumbel, Down for the Count: Dirty Elections and the Rotten History of Democracy in America 79 (rev. ed. 2016).

  45. [45]. See Evans, supra note 37, at 17 (describing the Australian system); Fredman, supra note 38, at 46 (same).

  46. [46]. See Fredman, supra note 38, at 31–118 (providing a thorough account of this movement); see also Albright, supra note 36, at 26–30 (discussing these developments); Evans, supra note 37, at 19–21, 27–28 (same); Joseph P. Harris, Election Administration in the United States 19 (1934) (same); Rusk, supra note 36, at 29–36 (same).

  47. [47]. See Act of Apr. 2, 1892, ch. 33, 1892 Iowa Acts 47; English, supra note 42, at 252–54 (discussing these legislative developments). The 1892 legislation was not Iowa’s first enactment concerning ballot secrecy. In 1849, the General Assembly established criminal penalties for “[a]ny judge of election who shall mark the ballot of an elector for the purpose of ascertaining for whom the elector voted.” Act of Jan. 15, 1849, ch. 105, § 6, 1849 Iowa Acts 410, 411.

  48. [48]. See Devine v. Wonderlich, 268 N.W.2d 620, 625 (Iowa 1978) (describing this scheme).

  49. [49]. See Evans, supra note 37, at 65–68 (discussing American states’ decision to ban identifying marks as a means of further addressing the vote-selling problem).

  50. [50]. Act of Apr. 2, 1892, ch. 33, § 27, 1892 Iowa Acts 47, 60. The act also made it a crime for a person to “allow his ballot to be seen by any person with an apparent intention of letting it be known how he is about to vote.” Id.

  51. [51]. Whittam v. Zahorik, 59 N.W. 57, 63 (Iowa 1894).

  52. [52]. Iowa Code § 1120 (1897). The Iowa Supreme Court initially enforced that law in quite a draconian fashion. See Voorhees v. Arnold, 78 N.W. 795, 796 (Iowa 1899) (“If a cross is placed outside the circle or square, instead of being placed substantially in it, as the law requires, or if the word ‘Yes’ is written in the circle, instead of placing a cross there, it may be said, as a matter of law, to be deliberately done, and might be used for identification, and defeats the ballot.”). In later years, the court signaled a willingness to ease up a bit, acknowledging that “[s]ome of the earlier decisions, rendered shortly after the enactment of the Australian Ballot Law in the several states, are somewhat extreme in applying that portion relating to identifying marks, going, as we think, to the verge of infringing on the free exercise of the voting franchise.” Fullarton v. McCaffrey, 158 N.W. 506, 508 (Iowa 1916). But the court still sometimes applied the law with remarkable rigor. See, e.g., Donlan v. Cooke, 237 N.W. 496, 498 (Iowa 1931) (holding that a ballot was disqualified because the voter had written on it “Lets [sic] have it wet,” a Prohibition-related comment that could serve as an identifying mark, and holding that three other ballots were disqualified because they had been “marked with a check mark instead of a cross”).

  53. [53]. Fullarton, 158 N.W. 506.

  54. [54]. Id. at 508.

  55. [55]. Devine v. Wonderlich, 268 N.W.2d 620, 625 (Iowa 1978). Other states’ courts have explained those jurisdictions’ similar laws the same way. See, e.g., Truelsen v. Hugo, 83 N.W. 500, 501 (Minn. 1900) (stating that Minnesota’s ban on identifying marks was “intended to prevent electors from so marking their ballots as to indicate that they had voted according to contract”), overruled by Murray v. Floyd, 11 N.W.2d 780 (Minn. 1943); Sego v. Stoddard, 36 N.E. 204, 207 (Ind. 1894) (“It was believed that if it could be rendered impossible for the buyer or his agent to identify the ballot voted by the purchased voter from a mere indication beforehand how it should be marked, the desired end [of preventing vote-selling] would be reached, because it was believed that, as a general thing, a vote buyer would not risk his money on a vote seller without some assurance other than the mere word of the bribed voter.”).

  56. [56]. Iowa Code § 49.98 (2020).

  57. [57]. Id. § 49.84(1)(c).

  58. [58]. Id. § 39A.4(1)(a)(6); see also id. § 49.92 (“The fact that the voting mark is made by an instrument other than a black lead pencil shall not affect the validity of the ballot unless it appears that the color or nature of the mark is intended to identify the ballot contrary to the intent of section 39A.4, subsection 1.”).

  59. [59]. Iowa Admin. Code r. 721-26.14(50) (2018).

  60. [60]. Id.

  61. [61]. Id. r. 721-26.21(49).

  62. [62]. Id. r. 721-26.21(49) ex. B (providing this illustration).

  63. [63]. Id.

  64. [64]. Id. r. 721-26.20(4) ex. B (giving an example of a voter who impermissibly writes “By the New Party” next to her vote).

  65. [65]. See, e.g., Iowa Code §§ 50.12, 53.30 (2020) (prescribing this security measure).

  66. [66]. See id. § 52.1(1) (“At all elections conducted under chapter 49, and at any other election unless the commissioner directs otherwise pursuant to section 49.26, votes shall be cast, registered, recorded, and counted by means of optical scan voting systems, in accordance with this chapter.”).

  67. [67]. Accord Michael Freiberg, “Anticipating an Evil Which May Never Exist”: Minnesota’s Anachronistic Identifying Mark Statute, 36 Wm. Mitchell L. Rev. 45, 68 (2009) (“A nonexistent threat of an awkward form of corruption does not appear to be a strong rationale for the disenfranchisement of a large number of presumably well-meaning voters.”). Iowa already has such anticorruption laws in place. See Iowa Code § 39A.2(d)(1) (2020) (making it a crime to pay “money or any other thing of value to a person to influence the person’s vote”); id. § 39A.2(d)(3) (making it a crime to “[r]eceive[] money or any other thing of value knowing that it was given in violation of subparagraph (1)”).

  68. [68]. See November 3, 2020 General Election Polling Places and Sample Ballots, supra note 21 (providing links to images of ballots used in each of the county’s precincts; for an example of the three races I reference here, see the ballots in any of the Iowa City precincts).

  69. [69]. See supra notes 60, 64 and accompanying text.

  70. [70]. See infra note 87 and accompanying text.

  71. [71]. See November 3, 2020 General Election Polling Places and Sample Ballots, supra note 21 (providing links to images of ballots used in each of the county’s precincts); see also Iowa Code
    § 49.92 (2020) (“The instructions appearing on the ballot shall describe the appropriate mark to be used by the voter. The mark shall be consistent with the requirements of the voting system in use in the precinct.”).

  72. [72]. See Iowa Admin. Code r. 721-26.1(49) (2018) (defining “[p]rescribed mark” as “the mark shown in the voting instructions as the appropriate way to mark a vote,” and defining “[v]oting target” as “the place designated on a ballot for the voter to mark the voter’s choice” (emphasis omitted)).

  73. [73]. Id. r. 721-26.15(49), 26.15(1); see also id. r. 721-26.104(3) (“A voter’s definite choices shall be counted even if the recount board determines that the voter’s choices differ from the votes as counted by the tabulating device.”). See generally Iowa Code § 49.98 (2020) (“The state commissioner shall, by rule adopted pursuant to chapter 17A, develop uniform definitions of what constitutes a vote.”). The “state commissioner” is Iowa’s Secretary of State. See id. § 47.1(1) (“The secretary of state is designated as the state commissioner of elections and shall supervise the activities of the county commissioners of elections.”).

  74. [74]. See Iowa Admin. Code r. 721-26.104(3) (2018) (“The recount board shall follow the standards for counting votes as prescribed by Iowa Code sections 49.92 to 49.99 and this chapter.”).

  75. [75]. Id. r. 721-26.16(1) (emphasis added).

  76. [76]. Id. r. 721-26.16(2) (emphasis added); see also id. r. 721-26.18(49)(3)–(4) (stating that a mark counts as a vote if it is consistently used and is in or near the voting target or is near a candidate’s name); cf. id. r. 721-26.18(49)(5)–(6) (stating that a mark may be counted as a vote if it consists of “[a] circle around the voting target for all choices” or of “[a] circle around or a line drawn under the name of the candidate for all choices” (emphasis added)).

  77. [77]. To me, that interpretation seemed bolstered by the fact that—in the U.S. Senate race that appeared immediately above the congressional race on the ballot—the voter’s mark had begun to drift to the right; it covered the right half of the selected Senate candidate’s voting target and continued rightward, into the white space that stood between the voting target and the accompanying candidate name. See November 3, 2020 General Election Polling Places and Sample Ballots, supra note 21 (providing links to images of the ballots, on which the Theresa Greenfield/Joni Ernst U.S. Senate race appeared directly above the Hart/Miller-Meeks congressional race on the left-hand side of the first page). Perhaps it bears noting that the machines would easily read that mark as a vote because a portion of the voting target had been inked and a recount board would count it as a vote, as well. See Iowa Admin. Code r. 721-26.18(49)(2) (stating that a mark “shall count as” a vote if it “is a close approximation of the prescribed mark but ... strays outside the voting target or does not completely fill the voting target”).

  78. [78]. See Iowa Admin. Code r. 721-26.18(49) ex. C (providing an illustration showing a permissible vote in the form of check marks consistently placed between voting targets and candidate names).

  79. [79]. See, e.g., Bush v. Gore, 531 U.S. 98, 104–08 (2000) (per curiam) (identifying ways in which Florida’s use of an unelaborated “intent of the voter” standard to count votes in the 2000 presidential election led to unacceptable differences in vote-counting standards in different parts of the state). The fact that counties’ recount boards are permitted to choose differing recount methods, see supra Part I, is itself in some tension with equality values, because it means that the ballots in some counties might be individually scrutinized while ballots in other counties are simply run through the tabulating machines a second time.

  80. [80]. See supra note 25 and accompanying text (discussing hesitation marks).

  81. [81]. See Iowa Admin. Code r. 721-26.21(49) (“A vote for an office or question shall be counted if the voter has marked the ballot in a manner that will be counted as an overvote by automatic tabulating equipment but the voter has indicated in a clear fashion that the voter has made a mistake.”).

  82. [82]. Ralph Waldo Emerson, Self-Reliance (1841), reprinted in Barack Obama, The Inaugural Address 2009, at 47, 62 (2009).

  83. [83]. See supra note 25 and accompanying text (discussing hesitation marks).

  84. [84]. See Iowa Admin. Code r. 721-26.20(3) (“If a write-in vote duplicates an otherwise correctly cast vote for a candidate whose name appears on the ballot, the write-in vote shall be counted. The ballot has been read as overvoted for this office ... .”).

  85. [85]. Iowa Code § 49.99(1) (2020); see also Iowa Admin. Code r. 721-26.51(49) (stating that, when a precinct uses paper ballots without optical scan equipment, “[t]he precinct election officials shall count write-in votes ... without regard to whether the voter has made a mark opposite the candidate’s name”).

  86. [86]. Iowa Code § 49.99(1) (“[W]hen a write-in vote is cast using an optical scan voting system, the ballot must also be marked in the corresponding space in order to be counted.”).

  87. [87]. Iowa Admin. Code r. 721-26.20(1) (emphasis added).

  88. [88]. Wesberry v. Sanders, 376 U.S. 1, 17 (1964).

  89. [89]. See supra Part I.

  90. [90]. See supra Section II.A.

  91. [91]. See supra Section II.B.

  92. [92]. See supra Section II.C.


H. Blair and Joan V. White Chair in Civil Litigation, University of Iowa College of Law.