Reading both Coleman’s and Franken’s appellate briefs—Coleman’s reply is due Friday—has caused me to dig deeper into the factual details concerning this case. I thought I knew this case fairly well from attempting to follow it closely throughout its entirety. But the deeper I dig, the more I learn, and the more I realize I don’t know, or am unsure of, about the record.
Franken’s brief to the Minnesota Supreme Court emphasizes the issue of waiver, arguing repeatedly that Coleman waived the various claims he is attempting to pursue on appeal. Franken may win the waiver issue, in which case the Minnesota Supreme Court will never need to address the merits of the appeal. (Or Franken may partially prevail on the waiver issue, thereby narrowing the scope of the appeal to be addressed on the merits.) But for purposes of this discussion, I will assume that Franken’s waiver arguments are unsuccessful, and thus the court will consider the merits of Coleman’s appeal.
Regarding the merits, Franken’s brief largely paints with a broad brush, treating Coleman’s claims about absentee ballots in bulk, rather than taking on one-by-one the nine separate categories that Coleman presented in his brief. As a litigation strategy, Franken’s approach may be successful, getting the Minnesota Supreme Court to reject Coleman’s federal and state-law claims regardless of the particular factual contexts to which they apply. The particular facts relating to variations among local practices may be the terrain on which Franken feels the least comfortable, and thus his brief may be deliberately avoiding the details of these local variations as much as possible.
In one place, moreover, his brief seems to me to mischaracterize the facts as I understand them. Of course, I may be incorrect, and even if the brief is inaccurate in this respect, it is unlikely to be intentionally so, but the issue is a potentially significant one, and thus it would make sense to clarify the situation at oral argument if possible.
Consideration of this factual issue leads, in turn, to exploration of other related—as well as contrasting—factual issues. Lawyers often say that cases are decided based on their facts, rather than on abstract principles of legal doctrine. That aphorism might also prove true in this case.
Witness Registration and the March 31 Order
The factual issue that I believe Franken mischaracterizes concerns the first of Coleman’s nine categories, the one that Coleman considers the “starkest illustration” of local variation in the treatment of equivalent ballots and the one I discussed at length in my last essay on this case. This category, as readers may recall, is where the witness portion of the certificate on the absentee ballot is complete and unproblematic on its face, containing a Minnesota address for the witness, but it turns out that the witness is not a registered voter. As Coleman alleges, some localities accepted ballots in this situation without checking the Statewide Voter Registration System (SVRS), presuming instead the voter to be registered, while other localities did check and thus rejected any ballot if the SVRS did not show the witness to be registered.
In his brief (at page 37), Franken says that Coleman’s claim about this differential treatment of the same situation among localities is really a non-issue because “the trial court already remedied any alleged problem.” Specifically, Franken contends that the trial court “adopted what Appellants characterize as the more lenient approach, requiring nothing more than a Minnesota address to prove registration.” (Same page, citations omitted.) In other words, Franken seems to be saying that the trial court accepted any ballot previously rejected by the localities that had checked the SVRS and found the witness unregistered—or at least that the trial court would have accepted these ballots upon Coleman’s completion of the necessary additional steps for validating them in other respects (including, for example, showing that its voter did not also vote a regular in-person ballot on Election Day). No need to worry about any disparities among localities on this point, Franken apparently argues, because the trial court itself was willing to review any ballot rejected by counties applying the more stringent standard, and the trial court’s own review would be conducted according to the more lenient standard, thereby creating a single uniform and lenient standard for all ballots statewide.
But Franken’s characterization of the trial court’s position on this issue is not how I understand it. To be sure, in its March 31 order—a ruling that has received less attention than either its February 13 order or its final Findings of Fact & Conclusions of Law (on April 13)—the trial court stated that it “accepted the witness as a registered Minnesota voter if the witness gave a Minnesota address and there was no evidence to the contrary concerning the witness’s registration status.” (Page 17.) This sentence, however, does not do what Franken appears to argue that it does.
It is true that this sentence embraces the lenient standard in a certain respect, just not in the respect relevant to Coleman’s claim with respect to the local variation on this issue. The key language is the last clause of the sentence: “the Court accepted the witness as a registered Minnesota voter if the witness gave a Minnesota address and there was no evidence to the contrary concerning the witness’s registration status.” (Emphasis added.) But in the situation where a locality applied the more stringent standard and examined the SVRS and found the witness was unregistered, there necessarily would be “evidence to the contrary” regarding whether this witness was registered. Thus, the trial court would not accept this ballot for counting, and the consequence of different local standards regarding this situation remains unremedied.
If the same absentee ballot had been cast in a locality that applied the lenient standard on Election Day, the witness would have been presumed registered, the SVRS would not have been checked, the witness’s actual lack of registration would not have been discovered, and the ballot would have been counted. In making this observation, I do not mean to say that Coleman necessarily has a winning Equal Protection claim with respect to the fact that some localities counted ballots on Election Day based on the lenient standard, while others rejected equivalent ballots based on the strict standard. Rather, I’m suggesting here that the Equal Protection claim does not disappear simply on the ground that “the trial court has already remedied [the] alleged problem,” as Franken puts it, if my reading of the trial court’s ruling is correct. Ballots that would have been counted in lenient localities remain rejected under the trial court’s ruling because they were cast in strict localities, and the trial court refused to require these ballots to be treated in the same way as they would have been if they had been cast in lenient localities.
The partial leniency expressed in the trial court’s March 31 order applies only to absentee ballots rejected for reasons other than the one at issue here—that the witness was not registered. In other words, when ruling on a ballot that had been rejected on some other basis (for example, that the voter’s signatures of the ballot envelope and the original application do not match), and having concluded that the ballot was wrongly rejected on this other ground, the court was thus in the position of needing to verify the ballot’s validity in other respects. In this circumstance, the trial court was willing to presume that the witness was registered if the witness had a Minnesota address, and thus the trial court would not itself check the SVRS. In that respect, the trial court was acting like a lenient locality rather than a strict one. But this practice in that circumstance does not negate the fact that where an absentee ballot was rejected by a strict county for the specific reason that the witness was not registered, that rejection remained operative according to the trial court’s ruling—because the strict county refused to apply the lenient presumption and had uncovered evidence that the witness was not registered.
The Implications of the March 31 Ruling on Witness Registration
Assuming I am correct in my understanding of the trial court’s March 31 order, what difference might it make to an analysis of the merits of Coleman’s claim? A number of possibilities arise, but the one that seems most worthy of attention is whether it affects the interpretation of the relevant state statutes along the lines that I undertook in my previous essay. To recall, the key statutory interpretation question seems to be on what basis should local election officials be “satisfied” regarding the “complet[ion]” of the certificate on the absentee ballot envelope: should they be satisfied as long as the certificate on its face is “complete,” listing the witness’s Minnesota address? Or are officials entitled to remain unsatisfied unless and until they check the SVRS to make sure the witness is registered?
The argument might be made that what’s good enough for the trial court itself should be good enough for local election officials. In other words, the trial court deemed itself satisfied—when tasked with the obligation of verifying a ballot’s eligibility and thus acting in the shoes of local election officials—to look only at the completeness of the certificate itself and not check to the SVRS. If the trial court acted properly under the statute, why should local officials be entitled to adopt on their own a more stringent standard of acceptability?
“Limited local resources,” a justification proffered to permit localities to deviate from a uniform state standard, does not apply in this instance. “Limited local resources” would justify not taking the time and expense to check the SVRS; it does not justify taking that extra step when the trial court did not, based on its preference for presumptively trusting absentee voters about the registration status of their witnesses: “Voters overwhelmingly testified that they asked their witnesses whether they were registered voters. Given this testimony and the instructions to the voter,” (emphasis added) the trial court declined to second-guess voters by checking the SVRS. Under the statute, are election officials in some localities entitled to distrust voters more than the trial court did?
If all localities should have acted as the trial court did, and I am correct in my understanding of the limited way in which the trial court applied its standard, then there are ballots that should not have been rejected but were. A separate state-law remedial question remains about what, if anything, should be done in an election contest lawsuit about this violation of the relevant state statutory standard—as I also discuss in my previous essay. But if the correct answer to this remedial question is that the witness’s lack of registration should not now invalidate the ballot (because the witness’s lack of registration would not have been discovered through a SVRS search, if the localities had been acting properly under the statute by confining themselves to an examination of only the completeness of the witness’s certificate on its fact, as described in my previous essay), then the ballot presumably should be counted in the election contest lawsuit as long as it would have been eligible for counting on Election Day in all other respects. And if a proper analysis of state law does not yield this conclusion, there remains the federal Equal Protection issue: when the trial court adopted a lenient standard for its own evaluation of witness certificates, why would the trial court be justified in simultaneously giving operative effect to localities that adopted a stricter standard? There may well be a good and sufficient answer to this question, but it would become necessary to confront that Equal Protection issue if state law tolerates the differential treatment of equivalent ballots, as factually it still seems to be the case on this “witness registration” point. (But, again, waiver, or a failure on Coleman’s part to offer the necessary proof regarding the validity of specific ballots, may mean the Minnesota Supreme Court never needs to worry about whether it must confront this Equal Protection question.)
Incomplete Witness Addresses & the Trial Court’s Moderate Leniency
Franken’s argument about local disparities being cured by the trial court’s conduct, even if inapplicable to the “witness registration” issue specifically, does seem applicable, however, to some of the other eight categories that Coleman identified in his brief. For example, the March 31 order shows that the trial court was also willing to be somewhat lenient on the issue of whether the witness’s address was sufficiently complete to qualify under the statute. Indeed, as a more general point, it is quite remarkable just how lenient the trial court was in its March 31 order, given the reputation it earned for strictness in its own statutory interpretation based on the February 13 order. I don’t think it is well understood, even by those who follow Coleman v. Franken closely, how significantly the March 31 order deviates from the strictness of the February 13 ruling—and how this deviation may have affected the trial court’s handling of particular ballots (either those actually presented to it, or those that could have been presented to it, if litigants had followed all the necessary procedures).
The April 13 Findings of Fact & Conclusions of Law do not highlight the evolution in the trial court’s thinking from February 13 to March 31, but instead appear to revert back to the strictness of the February 13 ruling—yet without repudiating the operative force of the March 31 order in terms of the trial court’s own review of ballots. One of the challenges for the Minnesota Supreme Court may be to determine, based on its review of the entire record before it, just how significant a role the March 31 standards played in the trial court’s evaluation of ballots. To do so would seem to require an appellate examination of the ballots themselves. Neither the March 31 order nor the final April 13 document explains how the trial court’s standards apply to particular ballots.
The March 31 order states (at page 17): “In limited instances, the Court accepted an absentee ballot witnessed by an individual who provided a street address or a post-office box without a city or zip code where the parties also offered into evidence the witness’s voter information look-up from the [SVRS] or the witness’s complete address was evident from the fact of the ballot.” Depending upon what the trial court meant by “limited instances”—does it mean a subset of those envelopes with the described criteria (and if so what subset?), or does the entire set of envelopes with the described criteria define the “limited instances”?—the trial court’s leniency on this issue would seem to go a long way to solving the problem of local variation on the same issue. In his appellate brief (at page 12), Coleman complains that some localities strictly required an absolutely complete address: a missing zip code or city name, even if the other information was present, apparently would invalidate the ballot in these “strict compliance” localities. By contrast, other localities were more lenient, accepting incomplete addresses, rather in the manner of the trial court itself.
Thus, the relevant question becomes: how many ballots were not acceptable under the trial court’s moderately lenient standard but would have been accepted under even more lenient local standards? As a practical matter, this number might be rather small—and significantly lower than the “more than 300 ballots” that were rejected by localities using an absolute “strict compliance” standard, which was much harsher than the trial court’s own standard. Unlike with the “witness registration” issue discussed above, this “witness address” issue is not one for which a local standard stricter than the trial court’s remained functionally operative at the trial court’s stage of ballot review. If the trial court considered the address adequate, the ballot was accepted for counting assuming all other necessary aspects of the ballot’s validation were satisfied. Thus, Coleman could have had any ballot rejected by a locality under too harsh a standard accepted by the trial court’s more lenient standard, if Coleman undertook all the necessary steps to prove the ballot’s validity. Despite this opportunity, if ballots still remain uncounted that would have been acceptable under the trial court’s moderately lenient standard, the responsibility may be Coleman’s for having failed to provide the additional proof necessary to validate the ballots in other respects.
To be sure, Coleman does say that some localities were willing to accept ballots where the witness’s address was entirely “missing,” whereas the trial court clearly required at least a street address. But was the trial court wrong under Minnesota law to require at least this much in order to be “satisfied” that the certificate had been “completed as prescribed” (in the language of section 203B.12? And if not, then would the trial court be obligated under Minnesota law to count ballots that failed to meet this appropriate standard just because some localities had improperly been excessively lenient in willing to accept a ballot with no witness address at all? That remedial obligation seems doubtful as a matter of state law. Nor does it seem likely that federal Equal Protection would compel this conclusion. In this particular context, the improper excessive leniency on the part of some localities does seem a rather minor error that does not rise to the level of a federal constitutional violation (as long as only a few localities strayed from the trial court’s norm, and because this deviation concerned only a relatively inconsequential administrative matter of degree, on which perfect conformity to a moderately lenient statewide standard could not reasonably be expected: just how incomplete a witness’s address could be and still be acceptable—a quintessential “line-drawing” problem). Moreover, there would be no reason to begin to speculate about whether and how Minnesota law might attempt to “uncount” ballots accepted because of this excessive leniency in some localities if it turned out that no more than a handful of ballots were counted that would not have been accepted under the trial court’s standard.
Signature comparisons & the limits of feasible uniformity
The issue of “signature mismatches” is another one for which the trial court’s review would seem to operate as a cure for all, or most, of whatever local variation occurred concerning the standard for comparing signatures on ballot envelopes and ballot applications. The process that the trial court employed permitted voters to testify that the two signatures were theirs and that the local officials had wrongly rejected their ballot on this ground. Thus, if a locality used an excessively rigorous standard for comparing signatures, the trial court’s proceedings permitted the ballot to be validated despite that excessive strictness. If Coleman did not take advantage of this opportunity, the consequence of those ballots remaining rejected is obviously not the trial court’s fault.
Maybe some ballots were accepted as a result of excessively lenient standards regarding signature matching by some localities. But ballots that could not meet the trial court’s procedures obviously should not have been counted, since the voter would have been unable to testify that the two signatures were theirs. If some ballots like this slipped through, that would be wrongful, but before considering what (if any) remedy might be necessary or appropriate under state law, there would need to be in an election contest lawsuit some showing of the magnitude of the problem and how it might have affected the apparent margin of victory. As far as I can tell, there has been no showing of how many ballots might have been counted where the individual who signed the envelope was not in fact the same individual who signed the application.
Nor would it seem that a federal Equal Protection claim would likely have any traction with respect to the specific issue of signature matching. The Bush v. Gore principle, whatever it might mean (and its scope is surely debatable), at a minimum requires that a standard be capable of further precision before demanding that additional precision. The U.S. Supreme Court’s majority opinion in Bush v. Gore made at least this much clear: “in some cases [a] general command . . . is not susceptible to further refinement.” With respect to less-than-fully-dislodged chads, however, the Court concluded that more “specific rules designed to ensure uniform treatment” were “practicable” (emphasis added) and thus required.
Signature matching, however, is much more complicated than deciding where to draw the line between dimpled, punctured, and hanging chads. It is arguably not feasible to give local election officials a set of detailed rules on how to compare signatures, so that there will be no variation in leniency or strictness among all the different set of eyes across a state that will be examining these signatures. It is not at all equivalent, for example, to giving local officials a clear command either to check or not to check the SVRS to see if a witness is registered. Here, Minnesota law permits the opportunity to bring any ballot before the trial court in the election contest lawsuit to revisit whether the ballot should have been rejected for a signature mismatch, permitting voters to corroborate their signatures, with the decision to be made by a single judicial panel. It would seem that state law could do little more to assure uniformity with respect to the ultimate evaluation of all ballots, and thus unlikely that federal Equal Protection would require anything more.
A Couple of Other Categories, Quickly Considered
With respect to one of Coleman’s categories, it is especially unclear (at least to me) how lenient or strict the trial court was. The category involves the situation in which, at the time to accept or reject the absentee ballot envelope, the local election officials cannot find the original ballot application, yet the envelope bears an official sticker from the local election board with the voter’s correct name and address. In this circumstance, it is arguably presumable that the voter properly submitted the application, which prompted the board to send the voter the envelope with the correct sticker, but that the board inadvertently has misplaced the original application. (Such administrative errors regrettably happen.)
One sentence of the March 31 order (on pages 13-14) expresses a willingness to make this presumption: “In certain instances, the Court was willing to find that the voter made an application for an absentee ballot where the absentee ballot return envelope bore a sticker containing the voter’s name and address affixed to the envelope.” But the rest of the same paragraph (as well as the remainder of the same section of the court’s opinion) seems to reject this presumption: “[Coleman] argued that a voter could not obtain an absentee ballot without first completing an absentee ballot application and that the Court could presume the existence of the application. The Court was unwilling to make this presumption.”
One possible way to reconcile these apparently conflicting passages is that the trial court was willing to “excuse” the nonexistence of an application if but only if the voter testified under oath in court that he or she had completed one. Assuming this conjecture to be correct, the trial court’s approach would seem to give Coleman the relief he would need on this point: any ballot rejected by a locality that insisted on finding the application (a standard stricter than applied in other localities) could be reviewed by the trial court—and ultimately counted—by bringing the voter in court to testify to the submission of an application. But I hesitate in reaching even this tentative conclusion on this point, because I’m not sure if that is what the trial court had in mind in this portion of its March 31 order.
One category that Coleman mentions in his appellate brief was actually remedied in the earlier February 13 order. This category is the situation in which the local election board accidently gives the voter the ballot for the wrong precinct. Coleman complains (at pages 17-18) that some localities strictly rejected the ballot in this situation, not even partially counting it for those statewide and other multi-precinct races in which the particular voter was eligible to participate, while other localities were willing to count the ballot partially in this way to rectify their own official error. But in its February 13 order (page 16, footnote 7), the trial court indicated that it would count a ballot in this situation. In other words, on this issue, the trial court adopted the more lenient local approach as the operative statewide rule and thus negated the effect of the rigid rejection policy in some localities. Thus, it is unclear why Coleman would still be complaining about the local variation on this issue in his appellate brief.
Where the Trial Court was Strict—and the Consequences of this Strictness
In contrast to the various categories discussed above, with respect to other categories the trial court—even in its March 31 order—declined to adopt a lenient standard, but instead insisted on strict compliance. For example, one of Coleman’s categories involves the situation in which local election boards incorrectly sent unregistered absentee voters the materials appropriate for previously registered absentee voters. This official error had the consequence of inducing the unregistered voters to fail to submit a registration form at the same time as they submitted their absentee ballot, whereas if the officials had acted correctly, the unregistered voters would have received the registration form with instructions on how to return it in order to have their ballots count.
In this situation, some localities accepted the absentee ballot even though the voter was not registered, reasoning that their own official error was responsible for the voter’s omission. Other localities, by contrast, maintained that even their own official error could not excuse a lack of registration, since registration was an essential prerequisite to counting the voter’s ballot. On this issue, the trial court sided with the localities that took the “rigid rule/no exceptions” approach. As the court stated in its March 31 order (at page 10): “An unregistered voter may not cast a ballot even if the voter’s failure to register to vote is the result of election officials errors or omissions regarding registration.”
The fact that the trial court took the strict approach on this issue would seem to call for a different analysis of the issue on appeal (in comparison to those issues on which the trial court adopted a more lenient approach). If the trial court was wrong as a matter of statutory interpretation to be rigid on this particular issue—maybe the registration requirement should be construed as conditioned on the absence of official error that causes non-registration—then the appropriate recourse (assuming, again, no waiver or other procedural problem on Coleman’s part) would seemingly be to remand for consideration of those ballots that were rejected by localities based on an inappropriately strict standard on this issue.
But if the trial court was correct under Minnesota law to be strict on this issue, then a different sort of inquiry potentially arises. What about ballots that were accepted by localities that applied an inappropriately lenient standard on this issue? How many such ballots were there, and do they have the potential of making a difference in the outcome of this election? The counting of absentee ballots of unregistered voters—assuming, again, that it was improper for localities to count them under any circumstances, even when the localities themselves were responsible for the voters’ failure to be registered—seems no minor matter. It’s not the same as a locality being a bit too generous on how relatively complete or incomplete a witness’s address may be as it appears on the envelope. Would state law really take the position that there never can be any remedy if localities improperly adopt policies to permit the counting of unregistered voters, when state law also takes the position that it is essential that voters be registered for their ballots to count?
On this issue, it might be useful to invoke the “felon voter” analogy used by the trial court (and others) in connection with this case. If some localities deliberately let felons vote absentee ballots, when it is clear that they are not entitled to do so, would it really be true under Minnesota law that there would be no remedy for this violation—assuming (contrary to the situation in Bell v. Gannaway, as discussed in my previous essay) a candidate had no opportunity to challenge the locality’s decision to count these felon ballots prior to their being commingled with all other ballots counted at local precincts on Election Night (including the non-absentee ballots cast in-person at those precincts)? To be sure, Minnesota law might require a candidate in an election contest lawsuit to make some showing that enough of these felon ballots were cast for the opposing (and apparently victorious) candidate, before the contest court would begin to consider what sort of remedy might be feasible, permissible, and appropriate. And, here, Coleman might have failed to make the comparable preliminary showing that enough ballots from unregistered voters were cast for Franken to trigger a consideration of the issue of what remedy might be feasible, permissible, and appropriate in this case. But that procedural repudiation of Coleman’s appeal on this point would be a very different proposition than the trial court’s categorical assertion (paragraph 142 of its final April 13 ruling) that “Minnesota law does not provide a remedy for [a] claim that absentee ballots were wrongfully accepted and counted.”
The trial court’s pronouncement seems overbroad. If 1000 felon ballots were accepted in one locality, and if there were evidence that most of these were voted for a candidate whose apparent margin of victory was less than 100 votes, and if the losing candidate had no chance to challenge the counting of these felon ballots ahead of time, then I would think the Minnesota Supreme Court would at least pause before declaring the situation irremediable. Thus, presumably the Minnesota Supreme Court will want to decide Coleman v. Franken one way or another without committing itself to such a categorical pronouncement.
The same point, by the way, can be made of at least one other issue raised by Coleman. Apparently, some localities chose not to enforce at all the requirement that a non-Minnesota witness be a notary public (or official otherwise authorized to administer oaths). A voter’s failure to comply with this requirement was not induced by official error, and it is evident from the face of an envelope whether the non-Minnesota witness did or did not provide the notary (or other official) seal. Still, some localities counted ballots in this situation. The trial court, however, like other localities, rejected ballots for this evident failure to comply with a basic requirement. It is unclear (again, at least to me) how many ballots were counted by localities that were improperly lenient in this regard. Thus, it would seem to me that the relevant issues on appeal (assuming the waiver hurdle is cleared) are whether Coleman identified enough of these ballots that they might have made a difference and whether he made any showing that these improperly counted ballots were more likely cast for Franken than for himself. As long as state law gave Coleman the opportunity to make that case, there would seem to be no Equal Protection issue to consider regarding this situation, because state law would insist on a single uniform statewide standard—it is impermissible everywhere in the state for a locality to count a ballot in this circumstance—and would provide a remedy to correct improper local deviations from that uniform statewide standard.
Devilish Details and the Ligation Risks of Ignoring Them
Delving into the facts in the way that I have done for this essay confirms my earlier belief that it is necessary for the Minnesota Supreme Court to consider each of Coleman’s nine categories one at a time, or at least to group them into clusters, depending on whether leniency or strictness is the appropriate standard for the category. If nothing else, distinguishing the situations in which state law requires leniency from those in which it requires strictness helps clarify the remedial consequences of a locality’s (or potentially the trial court’s) failure to follow state law properly in this regard. Noncompliance with an obligation to be lenient would seem to call for an opportunity in the election contest lawsuit to count the inappropriately rejected ballot, whereas noncompliance with an obligation to be strict inevitably raises a trickier remedial problem (as I’ve discussed previously). But, after reviewing all nine of Coleman’s categories, there is no reason to think that state law requires a single across-the-board answer of leniency or strictness with respect to all of them.
In this regard, I wonder whether it was strategically wise after all for Franken to lump all nine of Coleman’s categories together in his appellate brief. Even if he focused for the most part on points he believes applicable to all nine, he could have included a section of his brief that consider each one seriatim, explaining why in his view it failed on its own specific facts. That approach would have been in keeping with Franken’s own approach at trial, where he said that every absentee ballot has its own story. At least all equivalent ballots in the same situation have a similar story to tell about that particular situation.
At the same time, however, “too little, too late” may end up describing Coleman’s embrace of specificity with respect to the distinctive factual details of particular ballots and the way in which local officials treated them. It was Coleman who attempted to paint with a broad brush throughout much of the trial, treating his Equal Protection claim as a single undifferentiated whole with respect to any local variations that occurred, and asking early on for all rejected ballots to be accepted. That broad brush may come back to haunt Coleman on appeal if, now having appropriately subdivided his case into nine specific categories, he nonetheless has failed to proffer sufficient factual details with respect to the ballots within each of these categories in order to sustain his overall burden of proof as the one who is contesting Franken’s apparent victory.