RECOMMENDED FOR PUBLICATION
                               Pursuant to Sixth Circuit I.O.P. 32.1(b)
                                      File Name: 20a0331p.06

                   UNITED STATES COURT OF APPEALS
                                 FOR THE SIXTH CIRCUIT



 MEMPHIS A. PHILIP RANDOLPH INSTITUTE; THE EQUITY           ┐
 ALLIANCE; FREE HEARTS; THE MEMPHIS AND WEST                │
 TENNESSEE AFL-CIO CENTRAL LABOR COUNCIL; THE               │
 TENNESSEE STATE CONFERENCE OF THE NAACP;                   │         No. 20-6046
 SEKOU FRANKLIN,                                             >
                                                            │
                              Plaintiffs-Appellants,        │
                                                            │
        v.                                                  │
                                                            │
                                                            │
 TRE HARGETT; MARK GOINS; AMY P. WEIRICH,                   │
                            Defendants-Appellees.           │
                                                            ┘

                        Appeal from the United States District Court
                      for the Middle District of Tennessee at Nashville.
                    No. 3:20-cv-00374—Eli J. Richardson, District Judge.

                            Decided and Filed: October 15, 2020

                Before: MOORE, GIBBONS, and READLER, Circuit Judges.

                                     _________________

                                           COUNSEL

ON BRIEF: Danielle Lang, Jonathan Diaz, Molly Danahy, Ravi Doshi, Caleb Jackson,
CAMPAIGN LEGAL CENTER, Washington, D.C., William L. Harbison, Christopher C. Sabis,
SHERRARD, ROE, VOIGT & HARBISON, PLC, Nashville, Tennessee, for Appellants. Sarah
K. Campbell, Matthew D. Cloutier, OFFICE OF THE TENNESSEE ATTORNEY GENERAL,
Nashville, Tennessee, for Appellees.

    GIBBONS, J., delivered the opinion of the court in which READLER, J., joined.
MOORE, J. (pp. 17–53), delivered a separate dissenting opinion.
 No. 20-6046           Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.          Page 2


                                       _________________

                                             OPINION
                                       _________________

        JULIA SMITH GIBBONS, Circuit Judge. In the midst of the COVID-19 pandemic,
absentee voting has found its way into the spotlight. Record numbers of voters are expected to
vote by mail in the November 2020 election, and the increased interest in absentee voting has
also led to increased interest in the policies and procedures governing how and when voters may
vote absentee. In resolving cases of significant public interest, judges must, as they do in all
cases, reach decision by employing independent, unbiased analysis, based on the law and the
facts of a particular case.

        The plaintiffs in this case consist of individuals and organizations located in Tennessee,
and together they have brought five claims challenging the Tennessee statutory scheme that
governs absentee voting.      One claim challenges the eligibility criteria that Tennessee has
imposed for absentee voters, one claim challenges limits on the plaintiffs’ ability to distribute
unsolicited absentee ballots, two claims challenge Tennessee’s procedures for verifying voter
signatures on absentee ballots, and the final claim challenges a restriction on first-time voters’
ability to vote absentee.     Our decision today deals only with the two claims involving
Tennessee’s signature verification procedures. For the reasons that follow, we AFFIRM the
district court’s order denying the plaintiffs’ requested preliminary injunction on those claims,
although we do so on a basis different from that relied on by the district court.

                                                 I.

        Tennessee gives voters who fall within certain enumerated categories the opportunity to
“vote absentee by mail.” Tenn. Code Ann. § 2-6-201. One such category includes people who
are hospitalized or ill, those with physical disabilities, or caretakers for such persons.

Tennessee has recently interpreted this category to encompass “persons who
have underlying medical or health conditions which render them more susceptible to contracting
COVID-19 or [are] at greater risk should they contract it . . . , as well as those who are caretakers
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.            Page 3


for persons with special vulnerability to COVID-19.” See Fisher v. Hargett, 

,
385 (Tenn. 2020); Appellees’ Br. at 6–7.

         Historically, only about 2.5% of Tennesseans have voted absentee by mail. That number
could be expected to rise for the upcoming election due to the COVID-19 pandemic.

         Since the identity of a person who votes by mail cannot be verified as easily as someone
who votes in person—in-person voters must present photo identification—the legislature has
established procedures and conditions for absentee voting with which it demands “strict
compliance.” Tenn. Code Ann. § 2-7-112(a)(1); § 2-6-101(c); see also City of Memphis v.
Hargett, 

, 110 (Tenn. 2013). The voter must first send a formal request or
application to vote absentee by mail to the county election commission office, which must take
place “not more than ninety (90) and not later than seven (7) days before the election.” Tenn.
Code Ann. § 2-6-202(a)(1). The request must be written, signed, provide certain identifying
information, and establish the voter’s eligibility to vote absentee by mail.

The county administrator of elections reviews requests to vote absentee by mail.

(d). In addition to determining whether the voter has provided the requisite information
and established eligibility to vote absentee by mail, the administrator “shall compare the
signature of the voter [on the request] with the signature on the voter’s registration record.”

f the administrator determines the signatures are “not the same,” then the request is rejected, and
the voter is notified in writing.

If, however, the voter’s
signatures are the same and the voter otherwise qualifies to vote absentee by mail, then the
administrator “shall” mail the voter absentee ballot materials.

who qualify to vote absentee by mail receive (1) an absentee ballot; (2) an inner
envelope in which to place the completed ballot; (3) an outer envelope in which to return those
materials; and (4) instructions.

the inner envelope is an affidavit that the
voter must sign under penalty of perjury to verify that he or she is eligible to vote in the election.

2-6-309. Once the voter has completed the ballot, signed the affidavit, and
placed the materials in the outer envelope, the voter returns the materials by mail to have the
ballot counted.

ballot must be received by no later than the time the polls
close.

     “Upon receipt by mail of the absentee ballot, the
 No. 20-6046               Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.                          Page 4


administrator shall open only the outer envelope and compare the voter’s signature on the
[affidavit1] with the voter’s signature on the appropriate registration record.”

the signatures match, then the ballot is counted.

is the final
verification necessary before the counting board counts the ballots.”).                            If, however, the
administrator determines the signatures do not match, then the ballot is rejected, and the voter is
“immediately” notified in writing.

This method of signature
verification is not a new requirement in Tennessee. See 1994 Tennessee Laws Pub. Ch. 859
(S.B. 2556). Historically, county election officials have quickly notified voters whose absentee
ballots were rejected, including contacting voters by mail, phone, and email.

         The statutory scheme does not provide a voter an opportunity to cure a signature defect
before her absentee ballot is rejected. However, voters whose ballots are rejected may submit a
new absentee ballot or cast a provisional ballot in person (either during early voting or on
election day), provided they do so by close of polls on election day. Tenn. Code Ann. § 2-7-
112(a)(3). Additionally, voters who are concerned that their absentee ballot might be rejected
may cast a provisional ballot even before being notified of a rejection. Tenn. Code Ann. § 2-7-
112(a)(3)(A). If the voter’s absentee ballot is ultimately accepted and counted, the provisional
ballot will be discarded. But if the absentee ballot is rejected, the provisional ballot will be
counted. See Tenn. Code Ann. § 2-7-112(a)(3)(B)(iii)-(v).

         The plaintiffs—a Tennessee voter who wishes to vote absentee by mail in the upcoming
general election2 and organizations engaged in voter outreach with members who wish to vote
absentee by mail—do not dispute Tennessee’s authority to impose a signature verification
requirement for absentee ballots. Nor do they challenge the first signature verification step,
which takes place before election officials send the voter an absentee ballot. Rather, they allege
that Tennessee’s second and final signature verification process is constitutionally inadequate,

         1
           Unaltered, the statute reads: “Upon receipt by mail of the absentee ballot, the administrator shall open
only the outer envelope and compare the voter’s signature on the application with the voter’s signature on the
appropriate registration record.” Tenn. Code § 2-6-202(g) (emphasis added). We agree with the defendants that in
context the statute is referring to the affidavit on the inner envelope. The plaintiffs do not suggest this apparent error
is significant.
         2
         On September 29, 2020, while this interlocutory appeal was pending, the district court dismissed Plaintiff
Kendra Lee, who was not a party to this appeal, without prejudice at her request.
 No. 20-6046              Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.                       Page 5


violating their right to procedural due process under the Fourteenth Amendment and fundamental
right to vote under the First and Fourteenth Amendments.

         Although the statutory scheme is silent as to how election officials are to go about the
business of verifying signatures on absentee ballots, the parties do not dispute the salient features
of Tennessee’s signature verification program, which the defendants—Tennessee’s Secretary of
State and the Coordinator of Elections for the State of Tennessee3—are charged with executing.
In particular, the parties agree that the officials charged with verifying absentee ballot signatures
receive at least some training on signature verification. This training consists of a video prepared
by the Election Division of the Oregon Secretary of State, which is supplemented by directives
from the Division of Elections for the Tennessee Office of the Secretary of State. Among other
things, the Division of Elections directs officials to apply a presumption in favor of the validity
of the signature.       The training video instructs officials that “all but the most obvious of
inconsistent signatures are to be regarded as acceptable.” (R. 46-1, Goins Decl., at ¶¶ 23–24.)
Election officials must compare the questionable signature “with as many exemplars on file as
possible.”

A signature should not be rejected unless three officials, including the
county election administrator, determine that it is inconsistent with the signature on file.

         The parties are also in general agreement as to the number of ballots that have
been reported rejected for inconsistent signatures in the 2016 and 2018 national elections—
around 0.03% and 0.09% respectively—although they disagree about the significance of these
figures.

         Of course, it does not follow from the lack of dispute over the salient features of
Tennessee’s signature verification program that the parties agree on that program’s effectiveness.
The defendants, pointing to the strikingly small rejection rate, insist that the state’s signature
verification procedures are effective and that there is no evidence that the rejections that did
occur were erroneous, rather than proper rejections of invalid ballots.                        In particular, the
defendants point to an absentee ballot cast in 2018 by a voter who had already died as evidence
that the signature-verification process has prevented fraudulent ballots from being counted. The

         3
          Plaintiffs also sued the Shelby County District Attorney General in her official capacity seeking to enjoin
her from enforcing a separate section of the Tennessee law. She is not involved in the signature verification process.
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.          Page 6


plaintiffs, for their part, insist that Tennessee’s training is more likely to produce erroneous
signature verification determinations than to accurately reject ballots that do not have a genuine
signature. This is because, according to the plaintiffs’ expert, a forensic document examiner, it is
particularly difficult without extensive training to tell the difference between the natural
discrepancies in a voter’s signature from signing to signing—attributable to factors “including
age, health, native language, and writing conditions”—and variations attributable to the signature
being forged by a different person. (R. 40-4 Mohammed Decl. I, at ¶¶ 21–23, 28–37.)

       The plaintiffs filed suit on May 1, 2020, one week after Tennessee issued its April 23,
2020 COVID-19 Election Contingency Plan, which had allowed voters to claim the risk from
COVID-19 as a valid reason to vote absentee. On June 12, 2020, one day after the defendants
filed their answer, appellants filed the operative amended complaint and moved for a preliminary
injunction.

       The plaintiffs argue that Tennessee’s signature-verification procedures violate procedural
due process and infringe on the right to vote. The plaintiffs challenged the signature-verification
laws on their face, and not as applied during the COVID-19 pandemic, arguing that the laws
“will have serious consequences on the ability of Tennesseans to vote” due to increased absentee
voting during the pandemic. (R. 39, Am. Compl., at ¶ 65.) The plaintiffs sought an injunction
that would require Tennessee election officials to provide absentee voters notice and an
opportunity to cure signature inconsistencies before rejecting their absentee ballots.

       The defendants opposed the request for preliminary injunctive relief on multiple grounds.
As relevant to this appeal, the defendants argued that the plaintiffs lacked Article III and third-
party standing, that the plaintiffs were unlikely to succeed on the merits of their claims, and that
the harm an injunction would cause to the State and the public interest outweighed the plaintiffs’
alleged harms.

       The district court proceeded on the plaintiffs’ motion—which also sought to enjoin other
facets of Tennessee’s vote by mail procedures—in pieces. It reached the plaintiffs’ signature
verification claims on August 28, 2020 and denied the plaintiffs a preliminary injunction on the
grounds that they had failed to establish a likelihood of success on the merits or that they would
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.              Page 7


suffer irreparable harm if the district court denied the injunction. This interlocutory appeal
followed.

                                                 II.

A.

       “A district court must balance four factors in determining whether to grant a preliminary
injunction: ‘(1) whether the movant has a strong likelihood of success on the merits; (2) whether
the movant would suffer irreparable injury absent the injunction; (3) whether the injunction
would cause substantial harm to others; and (4) whether the public interest would be served by
the issuance of an injunction.’” Am. Civil Liberties Union Fund of Mich. v. Livingston Cnty.,

, 642 (6th Cir. 2015) (quoting Bays v. City of Fairborn, 

, 818–19 (6th
Cir. 2012)). “These factors are not prerequisites, but are factors that are to be balanced against
each other.” Overstreet v. Lexington-Fayette Urban Cnty. Gov’t, 

, 573 (6th Cir.
2002). However, “even the strongest showing on the other three factors cannot ‘eliminate the
irreparable harm requirement.’” D.T. v. Summer Cnty. Schools, 

, 326–27 (6th Cir.
2019) (quoting Friendship Materials, Inc. v. Mich. Brick, Inc., 

, 105 (6th Cir.
1982)). “[T]he party seeking a preliminary injunction bears the burden of justifying such relief.”
Livingston 

(quoting McNeilly v. Land, 

, 615 (6th Cir.
2012)); see also Tenke 

n.2 (“[I]n seeking a preliminary injunction, a
federal plaintiff has the burden of establishing the likelihood of success on the merits.”).

       Whether the movant is likely to succeed on the merits is a question of law, which this
court reviews de novo. Ammex, Inc. v. Wenk, 

, 360–61 (6th Cir. 2019) (citing City
of Pontiac Retired Employees Ass’n v. Shimmel, 

, 430 (6th Cir. 2014) (en banc) (per
curiam)). Otherwise, we review the district court’s “ultimate determination as to whether the
four preliminary injunction factors weigh in favor of granting or denying preliminary injunctive
relief ” for abuse of discretion. 

. Under the abuse-of-discretion
standard, this court “may reverse the district court if it improperly applied the governing law,
used an erroneous legal standard, or relied upon clearly erroneous findings of fact.”

A. Philip Randolph Inst., et al. v. Hargett, et al.            Page 

.

       While the district court did not directly consider the question, we begin our analysis with
whether the plaintiffs have standing. Without standing, we lack subject matter jurisdiction over
the claims before us. Lyshe v. Levy, 

, 857 (6th Cir. 2017). Thus, a “party who fails
to show a ‘substantial likelihood’ of standing is not entitled to a preliminary injunction.” Waskul
v. Washtenaw Cnty. Community Mental Health, 

, 256 n.4 (6th Cir. 2018) (quoting
Food & Water Watch, Inc. v. Vilsack, 

, 913 (D.C. Cir. 2015)). “However, an
inability to establish a substantial likelihood of standing requires denial of the motion for
preliminary injunction, not dismissal of the case.” 

.

       The doctrine of standing arises from Article III of the Constitution, which gives federal
courts jurisdiction over cases and controversies. U.S. CONST. art. III § 2; see also Lujan v.
Defenders of Wildlife, 

, 559–60 (1992). To establish standing, a plaintiff must show
an injury in fact that is fairly traceable to the defendant’s conduct and is likely to be redressed by
a favorable judicial decision.

To win declaratory or injunctive relief, a plaintiff
“must show actual present harm or a significant possibility of future harm.” Grendell v. Ohio
Supreme Court, 

, 832 (6th Cir. 2001). An organization may have standing either in
its own right, Northeast Ohio Coalition for the Homeless v. Husted, 

, 624 (Cir.
2016), or on behalf of its members “when its members would otherwise have standing to sue in
their own right, the interests at stake are germane to the organization’s purpose, and neither the
claim asserted nor the relief requested requires the participation of individual members in the
lawsuit,” Friends of the Earth, Inc. v. Laidlaw Envtl. Servs. (TOC), Inc., 

, 181
(2000). During the pleading stage, the burden remains on the plaintiffs to clearly allege facts that
demonstrate each element of standing. Spokeo, Inc. v. Robins, 

, 1547 (2016).

       An injury in fact must be concrete, particularized, actual, and imminent. Friends of the
Earth, Inc. v. Laidlaw Envtl. Servs. (TOC), Inc., 

, 180 (2000). Put another way, the
“threatened injury must be ‘certainly impending’ to constitute injury in fact, and ‘[a]llegations of
possible future injury’ are not sufficient.” Clapper v. Amnesty Int’l USA, 

, 409
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.           Page 9


(2013) (internal quotations omitted) (quoting Whitmore v. Arkansas, 

, 158 (1990)).
When the plaintiffs’ allegations of future injury are based on past human errors, the plaintiffs
face a high bar to demonstrate standing. See O’Shea v. Littleton, 

, 495–96 (1974)
(“Past exposure to illegal conduct does not in itself show a present case or controversy regarding
injunctive relief . . . if unaccompanied by any continuing, present adverse effects.”). In City of
Los Angles v. Lyons, the plaintiff sought an injunction against the City of Los Angeles to ban the
use of chokeholds by law enforcement officers in most circumstances. 

, 97–98
(1983). The plaintiff argued that he had standing to seek the injunction because a Los Angeles
officer had previously put him in an illegal chokehold.

The Court found that the
threat of future unlawful conduct by some law enforcement officers did not establish standing.

The Court reasoned that standing would require proof either “(1) that all police
officers in Los Angeles always choke any citizen with whom they happen to have an encounter
. . . or, (2) that the City ordered or authorized police officers to act in such manner.”

Otherwise, the plaintiff’s allegations failed to demonstrate an imminent risk of harm sufficient to
seek injunctive relief.

applied Lyons in another case challenging Tennessee elections
procedures. See Shelby Advocates for Valid Elections v. Hargett, 

(6th Cir. 2020),
cert. denied, --- S. Ct. ----, 

(2020). In Shelby Advocates, one organizational
plaintiff and four individual plaintiffs alleged “a variety of election administration problems,”
including that “election workers [were] poorly trained, sometimes distributing the wrong
ballots . . . , sometimes recording the wrong address when registering a voter, and once
distributing a poll book without redacting voters’ personal information.”

see also

(“The complaint’s allegations with respect to injury all boil down to prior system
vulnerabilities, previous equipment malfunctions, and past election mistakes.”) The Shelby
Advocates court found that the plaintiffs failed to allege imminent harm because there was no
evidence that “Shelby County election officials always [made] these mistakes” or that “the
government entities ordered the election workers to make any such mistakes.”

The
court held that policies like the ones challenged by the plaintiffs which only “add risk to the ever
present possibility that an election worker will make a mistake” do not, without more, create a
threat of imminent injury.

A. Philip Randolph Inst., et al. v. Hargett, et al.          Page 10


       Here, the plaintiffs have clearly not demonstrated that they face an actual, concrete,
particularized, and imminent threat of harm. The plaintiffs’ allegations involve two layers of
speculation about the upcoming election. First, they argue that based on historical rejection
rates, which were 0.03% in 2016 and 0.09% in 2018, that some absentee ballots will be rejected
for inconsistent signatures. Second, the plaintiffs claim that an unknown number of the ballots
that are rejected will be erroneously rejected because of human error, thereby infringing on the
plaintiffs’ members’ constitutional rights. The plaintiffs do not cite any official data to support
their theory that some of the absentee ballots will be incorrectly rejected. They also do not allege
that one of their members has had an absentee ballot erroneously rejected in the past. Instead,
they rely on the expert opinion of a forensic document examiner. After reviewing the Tennessee
statutes and academic literature, the plaintiffs’ expert concluded that it was highly likely that the
Tennessee officials will erroneously reject some absentee ballots in the upcoming election. The
expert explained that even forensic examiners make mistakes when verifying signatures, so it is
likely that Tennessee election officials will make mistakes as well.

       While the dissent contends that the forensic examiner’s opinion is unrefuted, in the
district court the defendants argued that the expert’s opinion was based on a misunderstanding of
Tennessee election safeguards.       Specifically, the defendants provided declarations from
Tennessee election officials who detailed the procedures they have put in place to protect against
human errors in the signature verification process. For example, election officials watch a
mandatory training video, where they are instructed to accept all but the most obviously
inconsistent signatures and to compare each signature to as many examples on file as possible.
Officials are required to start from the presumption that signatures are valid and look for ways to
accept rather than reject each ballot. Before any ballot is rejected for inconsistent signatures,
three trained election officials, including an administrator, must agree that the signature on the
absentee ballot does not match the signature in the voter registration records. Given the training
and protections practiced by Tennessee officials, it is far from inevitable that an absentee ballot
will be incorrectly rejected due to an inconsistent signature.

       Moreover, the dissent’s reliance on the forensic document examiner’s opinion goes
beyond the bounds of the examiner’s expertise. A forensic document examiner is a forensic
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.            Page 11


scientist who gives expert opinions about the authenticity of particular documents. Perhaps such
experts are qualified to tout generally their own expertise in verifying documents over that of a
lay examiner. But that is not the approach of the plaintiffs’ expert. Instead, he speculates that
lay election workers, in the face of increased absentee ballots, will not do as well as they have
done in the past in verifying signatures. This opinion relates to the personal characteristics of
Tennessee election workers, suggesting that they will be less diligent than they have been in the
past if they are faced with more work. This is an area in which the document examiner has no
expertise and one in which his opinion amounts to pure speculation. His opinion on this point
fails to support an inference of imminent harm.

       Furthermore, even if an individual’s ballot is erroneously rejected as part of the signature
verification process, the individual may still have an opportunity to vote through another means.
Under Tennessee law, officials are required to notify individuals “immediately” if their ballot is
rejected due to an improper signature, and officials go to great lengths to promptly notify
affected voters. After they are notified that their absentee ballot has been declined, voters are
able to either send a second absentee ballot or cast a provisional ballot in person. Many voters,
therefore, will likely have an opportunity to cure any errors in their initial absentee ballot.
Tennessee also has procedures for disabled individuals who are unable to write their signature or
mark; they may go to their voting precinct and cast a ballot with the assistance of an election
administrator. Tenn. Code Ann. § 2-7-112(b). Thus, even in the remote possibility that someone
requests an absentee ballot and in the interim suffers an injury rendering him unable to replicate
his previous signature, Tennessee provides safeguards for that individual’s right to vote.

       In sum, the plaintiffs have failed to meet their burden of establishing that they are at risk
of a concrete, imminent injury. Their alleged injury is even more remote than the allegations
brought by the plaintiff in Lyons. In Lyons, the plaintiff had actual evidence of past injury.
Here, the plaintiffs cannot cite with certainty or specification any past erroneous rejection of an
absentee ballot. If concrete evidence of past harm was not enough to establish standing in Lyons,
then the speculative allegations of past and future harm in this case are certainly insufficient.
Accordingly, the plaintiffs have failed to make a substantial showing of standing because they
 No. 20-6046             Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.                  Page 12


have failed to demonstrate that they are facing an actual, concrete, particularized, and imminent
injury.

          Despite the dissent’s insistence, this case is not controlled by Sandusky Cnty. Democratic
Party v. Blackwell, 

(6th Cir. 2004). In Sandusky, plaintiffs challenged Ohio’s
provisional ballot system, which they argued violated federal election laws. See

There, this court held that the plaintiffs had standing even though they could not identify which
future voter would be erroneously turned away at her polling place on election day.

The court explained that:

          [A] voter cannot know in advance that his or her name will be dropped from the
          rolls, or listed in an incorrect precinct, or listed correctly but subject to a human
          error by an election worker who mistakenly believes the voter is at the wrong
          polling place. It is inevitable, however, that there will be such mistakes.

the plaintiffs in Sandusky, the plaintiffs here have failed to show
that human errors are inevitable. We are not, as the dissent implies, asking the plaintiffs to
predict the future and specifically identify which absent ballots will be erroneously rejected. We
are simply asking for the plaintiffs to show, as is required under Sandusky, that such errors surely
will happen. They have not done so. Instead, the plaintiffs’ allegations boil down to fear of “the
ever present possibility that an election worker will make a mistake.”                     Shelby 

. As was the case in Shelby Advocates, “[a]ny analogy to Sandusky falls short.”

that the plaintiffs have not demonstrated an injury in fact, they cannot
show either direct organizational standing or representative standing on behalf of their
members.4 Even if the dissent is correct that the plaintiffs have significantly shifted their
operations, activities, and strategies in response to the COVID-19 pandemic, that would not
overcome the plaintiffs’ imminence problem. “An organization can no more spend its way into
standing based on speculative fears of future harm than an individual can.” Shelby 

(citing 

); see also Fair Elections Ohio v. Husted,

          4
          The plaintiffs make the barest of arguments that the remaining individual Plaintiff, Sekou Franklin, is
likely to have standing of his own accord. Because this argument was not raised below and was only raised on
appeal in a footnote in the plaintiffs’ reply brief, we decline to address it.
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.           Page 13

, 460 (2014). Therefore, under any theory, the plaintiffs have failed to show a
substantial likelihood of standing because they have not shown a threat of actual, imminent
harm. This alone is enough to affirm the district court’s denial of plaintiffs’ motion for a
preliminary injunction.

       2.

       Given our conclusion that the plaintiffs have failed to show a substantial likelihood of
establishing standing, we need not go further. However, we will provide limited commentary to
guide the district court in the ultimate resolution of various issues in this case. On the merits, the
plaintiffs advance two theories of how the Tennessee signature verification procedure violates
the constitutional rights of their members. First, Tennessee’s failure to provide pre-rejection
notice and the opportunity to cure a signature mismatch before the ballot is rejected violates
procedural due process. Second, the current procedure violates the fundamental right to vote
because the failure to provide pre-rejection notice and an opportunity to cure will result in “the
absolute deprivation of eligible voters’ right to vote.” (CA6 R. 28, Appellant Br., at 12.)

       When considering whether a challenged state action violates procedural due process, we
first consider whether there is a protected liberty interest. Johnson v. Morales, 

,
922 (6th Cir. 2020). If there is a protected liberty interest, we consider “what process is due.”

that there is a state-created liberty interest in voting absentee and having
one’s absentee ballot counted and argue, therefore, that this court should determine what process
is due under the factors established in Mathews v. Eldridge:

       First, the private interest that will be affected by the official action; second, the
       risk of an erroneous deprivation of such interest through the procedures used, and
       the probable value, if any, of additional or substitute procedural safeguards; and
       finally, the Government’s interest, including the function involved and the fiscal
       and administrative burdens that the additional or substitute procedural
       requirement would entail.

Mathews v. Eldridge, 

, 335 (1976). The defendants argue that this court should not
conduct a separate analysis under Mathews because the Anderson-Burdick balancing test
“provides a single standard for evaluating challenges to voting restrictions.” (CA6 R. 29,
Appellees’ Br., at 18–19.) Under Anderson-Burdick:
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.          Page 14


       A court considering a challenge to a state election law must weigh “the character
       and magnitude of the asserted injury to the rights protected by the First and
       Fourteenth Amendments that the plaintiff seeks to vindicate” against “the precise
       interests put forward by the State as justifications for the burden imposed by its
       rule,” taking into consideration “the extent to which those interests make it
       necessary to burden the plaintiff ’ s rights.”

Burdick v. Takushi, 

, 434 (1992) (quoting 

). The nature
of the burden the signature verification process places on voters determines the standard of
review. Rational basis review applies if the requirements are “reasonable nondiscriminatory
restrictions” and “the State’s important regulatory interests are generally sufficient to justify the
restrictions.” 

(internal citation omitted). Strict scrutiny applies if the
signature verification process presents “severe restrictions, such as exclusion or virtual exclusion
from the ballot.” Thompson v. Dewine, 

, 808 (6th Cir. 2020). And if the signature
verification process falls somewhere between a “reasonable nondiscriminatory restriction” and a
“severe restriction,” this court weighs the burden imposed against “the precise interests put
forward by the State as justifications for the burden imposed by its rule, taking into consideration
the extent to which those interests make it necessary to burden the plaintiff ’ s rights.” Id.
(quoting 

) (internal quotation marks omitted).

       The Sixth Circuit has not clearly answered “whether procedural due process claims are
viable in voting rights cases outside the Anderson-Burdick framework.” League of Women
Voters of Ohio v. LaRose, No. 2:20-CV-3843, 

, at *12 (S.D. Ohio Sept. 27,
2020) (determining that such procedural due process claims were viable and analyzing claim
under Mathews); see also New Georgia Project v. Raffensperger, No. 20-13360-D, --- F.3d ----,

, at *3 (11th Cir. Oct. 2, 2020) (rejecting the application of Mathews to voting
laws as doing so “would stretch concepts of due process to their breaking point”). While Obama
for America announced a “single standard for evaluating challenges to voting restrictions,” that
case did not specifically address procedural due process claims. Obama for Am. v. Husted,

, 430 (6th Cir. 2012). In another voting rights case, this court considered the
plaintiffs’ procedural due process claim without suggesting that such a claim lacks viability
outside Anderson-Burdick. Ne. Ohio Coal. for the Homeless v. Husted, 

, 637 (6th
Cir. 2016); see also Schmitt v. LaRose, 

, 642 (6th Cir. 2019) (same). Given our
 No. 20-6046            Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.         Page 15


conclusion that we lack standing to hear this case, we do not now resolve the question of whether
Anderson-Burdick’s “single standard” encompasses procedural due process claims, and we
mention it only to highlight the ongoing uncertainty in this circuit regarding the viability of these
claims.

          But regardless of whether Anderson-Burdick governs the plaintiffs’ procedural due
process claim, it certainly governs their claim that the signature verification procedure violates
the fundamental right to vote. On that claim, the plaintiffs contend that the burdens imposed on
their members’ right to vote are substantial because they could result in ballots being improperly
rejected and votes not being counted. See 

(“Under Anderson-Burdick, we
first look to the burden the State’s regulation imposes on the right to vote.”). Plaintiffs also
believe that the State’s interests in maintaining its current procedure are low because the State’s
current procedure actually diminish election integrity. See 

(discussing the
second element of the Anderson-Burdick test which weighs the State’s interest against the burden
on voting). The defendants, in contrast, claim that the current procedure imposes a minimal
burden on the right to vote because voters have methods of ensuring that their ballots will be
counted, such as casting a provisional ballot, and the state’s interest in preventing election fraud
and promoting finality is high.

          Determining whether a violation of the fundamental right to vote is likely to occur under
Anderson-Burdick involves careful balancing of the burden on the voters with the state’s
legitimate interests. Given that we have already concluded that the plaintiffs likely lack standing
to pursue their claims, we need not engage in that balancing here.

C.

          Finally, we note, briefly, that the plaintiffs have also failed to demonstrate that they are
likely to suffer irreparable harm. Irreparable harm is an “indispensable” requirement for a
preliminary injunction, and “even the strongest showing” on the other factors cannot justify a
preliminary injunction if there is no “imminent and irreparable injury.” D.T. v. Summer Cnty.
Schs., 

, 326–27 (6th Cir. 2019). “To merit a preliminary injunction, an injury ‘must
be both certain and immediate,’ not ‘speculative or theoretical.’”

(quoting Michigan
 No. 20-6046            Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.      Page 16


Coalition of Radioactive Material Users, Inc. v. Griepentrog, 

, 154 (6th Cir. 1991)).
“A restriction of the fundamental right to vote . . . constitutes irreparable injury.” Obama for

.

       Our focus during our inquiry into irreparable harm “is the group for whom the law is a
restriction, not the group for whom the law is irrelevant.”           Planned Parenthood of S.E.
Pennsylvania v. Casey, 

, 894 (1992)). If a ballot is improperly rejected and the
voter has no ability to cure that rejection, that might amount to a restriction of a constitutional
right. But, as we discussed during our analysis of whether the plaintiffs have standing, the
plaintiffs are not facing a certain and immediate risk of harm. The plaintiffs have not presented
any evidence that demonstrates that members of their organizations are likely to have their
ballots erroneously rejected under the current procedures.            And the plaintiffs have not
demonstrated that anyone whose ballot may be erroneously rejected will ultimately be unable to
cast a ballot, either absentee or by provisional ballot. Therefore, there is no evidence that
anyone’s constitutional rights are likely to be infringed by the Tennessee procedures. For that
reason, the plaintiffs cannot show that they will be irreparably harmed if they are not issued a
preliminary injunction.

                                                 III.

       For these reasons, we AFFIRM the district court’s denial of a preliminary injunction.
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.        Page 17


                                      _________________

                                            DISSENT
                                      _________________

       KAREN NELSON MOORE, Circuit Judge, dissenting. Make no mistake: today’s
majority opinion is yet another chapter in the concentrated effort to restrict the vote. See, e.g.,
Raysor v. DeSantis, 

(2020) (mem.); Republican Nat’l Comm. v. Democratic
Nat’l Comm., 

(2020); Democratic Nat’l Comm. v. Bostelmann, --- F.3d ----,

(7th Cir. 2020) (per curiam); New Ga. Project v. Raffensperger, --- F.3d ----,

(11th Cir. 2020); A. Philip Randolph Institute of Ohio v. LaRose, --- F. App’x
----, 

(6th Cir. 2020); see generally Richard L. Hasen, The 2016 U.S. Voting
Wars: From Bad to Worse, 26 Wm. & Mary Bill Rights J. 629 (2018). To be sure, it does not
cast itself as such—invoking instead the disinterested language of justiciability—but this only
makes today’s majority opinion more troubling. As a result of today’s decision, Tennessee is
free to—and will—disenfranchise hundreds, if not thousands of its citizens who cast their votes
absentee by mail. Masking today’s outcome in standing doctrine obscures that result, but that
makes it all the more disquieting.      I will not be a party to this passive sanctioning of
disenfranchisement. I dissent.

                                                I.

       To read the majority opinion, you would be forgiven for thinking that there was no
question that Plaintiffs likely lack standing to pursue their constitutional claims. The majority
crafts this illusion by misapplying inapposite authority, glossing over binding case law, and
torturing the evidentiary record. In doing so, it sets a dangerous precedent of its own that will
doubtless close the courthouse door to litigants, like Plaintiffs, seeking nothing more than to
ensure that their votes are counted. That result strikes at the core of our democratic system. As
the Supreme Court said in Reynolds v. Sims, 

, 561–62 (1964):

       Undoubtedly, the right of suffrage is a fundamental matter in a free and
       democratic society. Especially since the right to exercise the franchise in a free
       and unimpaired manner is preservative of other basic civil and political rights, any
       alleged infringement of the right of citizens to vote must be carefully and
       meticulously scrutinized.
 No. 20-6046            Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.         Page 18


The majority has abdicated this fundamental duty by conjuring up fictional barriers at the
threshold that allow them to turn a blind eye to the merits. Their justifications for doing so are
feeble.

          A plaintiff’s likelihood of success on the merits “‘necessarily includes a likelihood of the
court’s reaching the merits, which in turn depends on a likelihood that plaintiff has standing.’”
Waskul v. Washtenaw Cnty. Cmty. Mental Health, 

, 256 n.4 (6th Cir. 2018) (quoting
Nat’l Wildlife Fed’n v. Burford, 

, 328 (D.C. Cir. 1987) (Williams, J., concurring in
part) (original emphasis)).       Thus, for a preliminary injunction to issue, Plaintiffs must
demonstrate a likelihood of at least one of them establishing standing. See id.; Rumsfeld v.
Forum for Acad. and Institutional Rights, Inc., 

, 53 n.2 (2006) (“one party with
standing is sufficient to satisfy Article III’s case-or-controversy requirement”); Ne. Ohio Coal.
for the Homeless v. Husted, 

, 623 (6th Cir. 2016) (“When one party has standing to
bring a claim, the identical claims brought by other parties to the same lawsuit are justiciable.”).
Plaintiffs here have done so.

          The foundational elements of standing are well established: “The plaintiff must have
(1) suffered an injury in fact, (2) that is fairly traceable to the challenged conduct of the
defendant, and (3) that is likely to be redressed by a favorable judicial decision.” Spokeo, Inc. v.
Robins, 

, 1547 (2016), as revised (May 24, 2016) (citing Lujan v. Defs. of
Wildlife, 

, 560–61 (1992)). “An injury, for standing purposes, means the ‘invasion
of a legally protected interest which is (a) concrete and particularized, and (b) “actual or
imminent.”’” Daunt v. Benson, 

, 417 (6th Cir. 2020) (quoting 

). In the context of a claim for injunctive or declaratory relief, the plaintiff must demonstrate
that they are “subject to ‘actual present harm or a significant possibility of future harm.’”
Grendell v. Ohio S. Ct., 

, 833 (6th Cir. 2001) (quoting Nat’l Rifle Ass’n of Am. v.
Magaw, 

, 279 (6th Cir. 1997)).

          At its core, “[t]he purpose of the standing doctrine is to ensure that courts do not render
advisory opinions rather than resolve genuine controversies between adverse parties.” 

n.4 (Blackmun, J., dissenting); Chapman v. Tristar Prods., Inc., 

,
303 (6th Cir. 2019) (“The standing requirement prevents federal courts from issuing advisory
 No. 20-6046           Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.       Page 19


opinions.”). This limit on judicial power is said to ensure that the courts do not overstep the
separation of powers, see 

–04, and that the courts render decisions
with the benefit of argument that has been tested by the crucible of truly adversarial proceedings,
see United Food & Com. Workers Union Loc. 751 v. Brown Grp., Inc., 

, 555–56
(1996). See also City of Los Angeles v. Lyons, 

, 101 (1983).

         Decisional law recognizes that organizational plaintiffs are no less able to demonstrate
standing than individuals. Indeed, an organization has two avenues for establishing that they
have standing to sue: “(1) on its own behalf because it has suffered a palpable injury as a result
of the defendants’ actions; or (2) as the representative of its members.” MX Grp., Inc. v. City of
Covington, 

, 332–33 (6th Cir. 2002). Plaintiffs contend that they can demonstrate a
likelihood of establishing standing under either avenue. I agree.

                                                 A.

         To begin with, I believe that Plaintiffs have demonstrated a likelihood that they have
standing as representatives of their members. “An association has standing to bring suit on
behalf of its members when its members would otherwise have standing to sue in their own right,
the interests at stake are germane to the organization’s purpose, and neither the claim asserted
nor the relief requested requires the participation of individual members in the lawsuit.” Friends
of the Earth, Inc. v. Laidlaw Env’t Servs. (TOC), Inc., 

, 181 (2000). The only
question in dispute here is whether any of Plaintiffs’ members would likely have standing to sue
in their own right—Defendants do not contend that the issue falls outside of Plaintiffs’
organizational purposes or that the suit requires individual members’ participation (nor could
they).

         Defendants’ argument boils down to this: Plaintiffs do not have standing to sue on behalf
of their members because they have failed to—indeed, cannot—identify a specific member who
would have standing to sue in their own right. But, in these circumstances, Plaintiffs were
required to do no such thing. That is because we do not require organizational plaintiffs to
specify members who would themselves have standing where the injury in question could not be
“specifically identified in advance.” Sandusky Cnty. Democratic Party v. Blackwell, 387 F.3d
 No. 20-6046           Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.        Page 20


565, 574 (6th Cir. 2004) (per curiam); see also Block Commc’ns, Inc. v. F.C.C., 808 F. App’x
332, 336 (6th Cir. 2020). In other words, where an organizational plaintiff seeking to establish
standing to sue as a representative of its members could not identify the specific member that
will be harmed due to the nature of the injury, this court does not hold them to an impossible
standard—it is enough to demonstrate that at least one of their members will likely suffer the
injury.    See 

–74.        Here, Plaintiffs cannot know which of their
members’ absentee ballots will be rejected until election day, and thus they cannot identify the
injury that will occur with the level of specificity that Defendants demand and the majority
implicitly requires. But under this court’s precedent, that is not an insurmountable barrier.

          This doctrine is typified by Sandusky, a case to which the majority barely pays lip
service. In Sandusky, a panel of this court considered a challenge to Ohio’s provisional ballot
system brought by various political groups and labor 

–71. The challenge
arose under the Help America Vote Act (“HAVA”), which protected voters from being
mistakenly turned away from their polling place by “creating a system for provisional balloting,
that is, a system under which a ballot would be submitted on election day but counted if and only
if the person was later determined to have been entitled to vote.”

The crux of the
plaintiffs’ claim was that Ohio’s provisional ballot system failed to live up to the promises of
HAVA because it limited the availability of provisional ballots and “unduly limit[ed] the
circumstances in which a provisional ballot [would] be counted as a valid ballot.”

In
particular, the plaintiffs argued that under Ohio’s provisional ballot system, a poll worker could
withhold a provisional ballot from a would-be voter based on a mistaken, on-the-spot
determination that the voter was not a resident of the precinct, in violation of HAVA’s
protections.

n short, like Plaintiffs here, the plaintiffs in Sandusky brought suit to protect
against the risk of human error inherent in a state’s election laws.

          Acknowledging that the plaintiffs had not identified specific members “who will seek to
vote at a polling place that will be deemed wrong by election workers,” the Sandusky panel
nevertheless concluded the organizational plaintiffs had standing to pursue their claim.

The panel explained its reasoning thus:
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.         Page 21


       Appellees have not identified specific voters who will seek to vote at a polling
       place that will be deemed wrong by election workers, but this is understandable;
       by their nature, mistakes cannot be specifically identified in advance. Thus, a
       voter cannot know in advance that his or her name will be dropped from the rolls,
       or listed in an incorrect precinct, or listed correctly but subject to a human error
       by an election worker who mistakenly believes the voter is at the wrong polling
       place. It is inevitable, however, that there will be such mistakes. The issues [that
       the organizational plaintiffs] raise are not speculative or remote; they are real and
       imminent.

panel, then, the touchstone for representative organizational standing was
not the identification of a specific member who was sure to be denied a provisional ballot.
Instead, under Sandusky, in cases where the injury cannot be “specifically identified in advance,”
it is enough for a plaintiff organization to demonstrate that mistakes are certain to occur and that
at least some of the organization’s members will be affected. See

context of
suits involving elections, this rule makes good sense. Forcing plaintiffs to wait until they (or
their members) have been denied the opportunity to vote on election day would preclude any
meaningful relief because there is no casting a new ballot once the election is over. See League
of Women Voters of N. Carolina v. North Carolina, 

, 247 (4th Cir. 2014) (“[O]nce
the election occurs, there can be no do-over and no redress. The injury to these voters is real and
completely irreparable if nothing is done to enjoin this law.”). Yet that is the outcome the
majority effectively endorses today.

       Applying Sandusky to the matter at hand, Plaintiffs have established a likelihood that
some of their members will have their ballots erroneously rejected, and that is the most that this
court’s precedent requires. See 

. In particular, Plaintiffs have pointed
to evidence that the most recent reported rejection rate for absentee ballots due to perceived
signature invalidity is 0.09% and have offered unrefuted evidence that many of those rejections
are highly likely to be erroneous due to the inadequate training that election officials receive. R.
40-4 (Mohammed Decl. I at ¶¶ 21–23, 28–37) (Page ID #1535–45); R. 54-7 (Mohammed Decl.
II at ¶¶ 4–6) (Page ID #2312); see also R. 46-1 (Goins Decl. at ¶¶ 29–30) (Page ID #1832). This
rejection rate—to the extent it was accurate in the first place, see R. 43 (Mot. for Prelim. Inj. at
11 n.12) (Page ID #1669) (noting rejection rate was not reported for all Tennessee counties and
that not all absentee ballots were accounted for)—is likely to be significantly higher this year due
 No. 20-6046              Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.                       Page 22


to time pressures caused by the dramatic increase in the use of absentee ballots that Tennessee
expects. R. 54-7 (Mohammed Decl. II at ¶ 3) (Page ID #2312). And for Plaintiffs’ members,
many of whom fall within groups that are more likely to have benign signature variations due to
factors like their age, the rejection rate may be higher still. See R. 40-4 (Mohammed Decl. I at
¶¶ 21–23, 30–37) (Page ID #1535–45); R. 40-5 (Tennessee NAACP Decl. at ¶ 16) (Page ID
#1556). Setting that aside, even assuming a 0.09% rejection rate—which is generous—for the
2020 General Election, it is highly likely that some of Plaintiffs’ members will be among the
Tennessee voters who will suffer an erroneous rejection for perceived signature invalidity in
November if Tennessee’s signature verification procedure is left in place.1 That is all this court
requires of Plaintiffs in these circumstances. See 

.2


         1
           Plaintiffs count at least 30,000 members between them. See R. 40-6 (MCLC Decl. at ¶ 4) (Page ID #1568
(20,000 members); R. 40-5 (Tennessee NAACP Decl. at ¶ 11) (Page ID #1556) (10,000 members). Assuming one
third of those members (10,000) will vote absentee by mail, then it is reasonably likely somewhere around nine of
Plaintiffs’ members will have their ballots rejected for signature invalidity. At least some of those rejections are
likely to be erroneous according to the unrefuted opinion of Plaintiffs’ expert, who opined that untrained
comparisons are highly likely to result in erroneous rejections. R. 40-4 (Mohammed Decl. I at ¶¶ 21–23, 28–37)
(Page ID #1535–45); R. 54-7 (Mohammed Decl. II at ¶¶ 4–6) (Page ID #2312). I chose this rough one-third
estimate based on Tennessee’s own evidence that over 30% of the state’s registered voters are over the age of sixty,
thus qualify to vote absentee, Tenn. Code § 2-6-201(5)(A), and have a strong incentive to do so. See R. 40-2 (Doshi
Decl., Ex. 4 at 46) (Page ID #260). Of course, not all of those individuals will vote absentee by mail. But the 30%
number represents just one of the nine classes (not to mention further sub-classes) of Tennesseans who are
statutorily eligible to vote absentee by mail, including the significant-but-unknown number of Tennesseans who
qualify under the new interpretation rendering those at high risk for COVID-19 and their caretakers eligible to vote
absentee by mail. See Tenn. Code § 2-6-201; Fisher v. Hargett, 

, 385 (Tenn. 2020). Thus, I think
the one-third estimate is fair, especially considering that Plaintiffs have offered evidence suggesting that their
members qualify to vote absentee by mail at a higher rate than Tennessee generally. In any case, even if I cut my
estimate in half, and assumed that 5,000 of Plaintiffs’ members will vote absentee by mail, it would not change my
conclusion that Plaintiffs are likely to establish that at least some of their members will suffer an erroneous absentee
ballot rejection on account of perceived signature invalidity.
         2
           The Supreme Court has not since limited this principle, as Defendants assert. In Summers v. Earth Island
Institute, the case Defendants point to as abrogating Sandusky, the Supreme Court faulted an environmental
organization for failing to identify a specific member who would be injured by the application of the regulation in
question to particular national forests. See 

, 494–95 (2009). But there, the organization had a
straightforward way to identify specific members who would be injured by the regulation—offering evidence that
the member intended to visit the impacted forests. See

there is no dispute that a significant
proportion of Plaintiffs’ membership will vote, there is simply the unanswerable question as to which of them will
have their ballots erroneously rejected. Summers, accordingly, is no bar. See also Nat’l Council of La Raza v.
Cegavske, 

, 1041 (9th Cir. 2015) (“Where it is relatively clear, rather than merely speculative, that one
or more members have been or will be adversely affected by a defendant’s action, and where the defendant need not
know the identity of a particular member to understand and respond to an organization’s claim of injury, we see no
purpose to be served by requiring an organization to identify by name the member or members injured.”)
(distinguishing Summers).
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.          Page 23


       In order to avoid Sandusky, the majority resorts to a lame attack on the evidence
supporting Plaintiffs’ position and two inapplicable cases, namely, Lyons, 

, and
Shelby Advocates for Valid Elections v. Hargett, 

(6th Cir. 2020) (per curiam), cert.
denied, --- S. Ct. ----, 

(2020). On the evidentiary front, the majority decries
as speculative Plaintiffs’ conclusion that Tennessee election officials will erroneously reject at
least some of their members’ absentee ballots due to perceived signature invalidity.             The
majority even goes so far as to conclude that “it is far from inevitable that an absentee ballot will
be incorrectly rejected for inconsistent signature” at all. Maj. Op. at 10. What appears to be
motivating the majority’s reasoning is the absence of evidence of specific voters who have had
their absentee ballots erroneously rejected. Indeed, the majority bemoans Plaintiffs’ failure to
“cite any official data to support their theory that some of the absentee ballots will be incorrectly
rejected” or to point to a member who “has had an absentee ballot erroneously rejected in the
past.” Maj. Op. at 10. But the absence of evidence is not evidence of absence, and Plaintiffs
have offered a more than adequate alternative in the expert opinions of Dr. Linton A.
Mohammed, a certified Forensic Document Examiner (“FDE”) with impeccable credentials.
R. 40-4 (Mohammed Decl. I at ¶ 1) (Page ID #1530).

       Dr. Mohammed set forth a comprehensive analysis of the risk of error inherent in
signature verification, explaining that even highly trained FDEs will erroneously conclude that a
signature is non-genuine. See

Page ID #1550). Indeed, in one study referenced by Dr.
Mohammed, FDEs erroneously concluded that signatures were non-genuine at a rate of 7.05%.

(Page ID #1538–39). If FDEs—who typically receive two-to-three years of full time
training—will erroneously conclude a signature is non-genuine, then it does not require any
speculation whatsoever to conclude that Tennessee election officials—even with the benefit of a
45-minute video and the directives the majority recounts—will make the same error at a higher
rate. See

(Page ID #1541). Moreover, Dr. Mohammed identified numerous features
of Tennessee’s signature verification program that would increase the rate of error, such as
inadequate training, a lack of the requisite equipment, time pressure, a failure to screen of
election officials for “form blindness,” and limited comparison signatures upon which to base the
conclusion that a given absentee ballot held an invalid signature.

(Page ID
#1541, 1549). In light of this, Dr. Mohammed had no trouble opining that Tennessee election
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.          Page 24


officials are “highly likely” to erroneously reject absentee ballots due to perceived signature
invalidity

(Page ID #1535), going so far as to say that “Tennessee’s signature
matching procedures are all but guaranteed to result in the erroneous rejection of properly cast
ballots,”

(Page ID #1551–52). To take this evidence and conclude that errors may not
occur at all—as the majority does—is downright unreasonable.

       In an ineffective effort to undercut Dr. Mohammed’s testimony, the majority points out
that Defendants offered affidavits showing that Dr. Mohammed was unaware of certain extra-
statutory features of Tennessee’s signature verification process when he offered his opinions.
But none of Defendants’ evidence—detailing the (meager) signature verification training that
Tennessee election officials receive, including directives to employ a presumption favoring the
validity of signatures and vague assertions that a ballot will not be rejected unless three officials
deem the signature invalid—directly refuted Dr. Mohammed’s opinion that when signature
rejections occur, they highly likely to be erroneous. Indeed, Defendants produced no evidence
suggesting what the actual rate of error for signature verification rejections might be—assuming
Dr. Mohammed was incorrect—aside from a single, unverified case of suspected voter fraud.
R. 46-3 (Warren Decl. at ¶ 7) (Page ID #1854).          Moreover, Defendants’ evidence largely
affirmed that Tennessee’s signature verification process lacked the features that Dr. Mohammed
had originally opined would be required of a system that resulted in reliable signature
verification determinations. Thus, even after reviewing Defendants’ evidence, Dr. Mohammed
did not change his earlier opinions. R. 54-7 (Mohammed Decl. II at ¶¶ 1–6) (Page ID #2311–
12); see also R. 40-4 (Mohammed Decl. I at ¶ 36) (Page ID #1544) (opining that “Tennessee
election officials, even if put through a short training session, are unlikely to be able to
accurately account for these differences, particularly in an expedient time frame or when only
one or a few specimen signatures are available for comparison.”). In short, the majority’s effort
to downplay Dr. Mohammed’s expert opinions fails. Coupled with the certainty that ballots will
be rejected on the basis of a perceived signature invalidity and the historic rate for such
 No. 20-6046             Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.                  Page 25


rejections, Dr. Mohammed’s opinion renders it a certainty that erroneous rejections will happen
(and will very likely happen to Plaintiffs’ members).3

        The majority takes further issue with one of Dr. Mohammed’s other opinions: that the
rate of erroneous rejections due to perceived signature invalidity will increase along with the
number of absentee ballots Tennessee election officials must review. Maj. Op. at 10–11. But as
explained above, my conclusion that Plaintiffs are likely to establish standing does not depend
upon that particular opinion because Plaintiffs can establish that one or more of their members
will likely have their absentee ballots erroneously rejected due to perceived signature invalidity
even if the historic rate of rejection remains the same. In any case, the majority’s suggestion that
Dr. Mohammed is not qualified to render that particular opinion is ludicrous. Dr. Mohammed is
an accomplished researcher and academic in the field of document examination and signature
authenticity in particular. In addition to his 2019 Book, Forensic Examination of Signatures, he
has published sixteen peer-reviewed articles in the field, many focused specifically on “the
analysis of genuine, disguised, and forged signatures, and handwriting examination.” R. 40-4
(Mohammed Decl. I at ¶¶ 7, 8) (Page ID #1532–34). Moreover, he has trained investigators,
attorneys, and graduate students in document verification, and developed standards and practices
regarding the same.

(Page ID #1531–32). Even ignoring Dr. Mohammed’s other
accomplishments and experience, this would more than qualify him to opine that “In my
experience, the more signatures an election official has to review, the more likely they are to
make mistakes, particularly when they lack adequate time in which to conduct a review.” R. 54-
7 (Mohammed Decl. II at ¶ 3) (Page ID #2312). Contrary to the majority’s characterization, this
opinion is based not on Dr. Mohammed’s assessment of the “personal characteristics” of election
officials but on the conditions under which the election officials must perform their duties—and
in particular the insufficient time the election officials would have to examine the signatures in
question. Id.; R. 40-4 (Mohammed Decl. I at ¶ 25) (Page ID #1536–37) (opining that “because a
minimum of two hours is required to accurately compare signatures, election officials with

        3
           The majority suggests—citing no authority—that the only injury that would satisfy the standing inquiry
would be the state’s wholesale denial of a member’s ability to vote. See Maj. Op. at 11. Nonsense. The erroneous
rejection of a validly cast absentee ballot is plainly a concrete, particularized, and actual or imminent injury
sufficient to demonstrate standing.
 No. 20-6046             Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.                    Page 26


insufficient time to evaluate the signature on the ballot return envelope are likely to make
additional errors”); R. 46-1 (Goins Decl. at ¶ 9) (Page ID #1826) (anticipating delays in counting
absentee ballots due to significant increase in volume received). Dr. Mohammed was qualified
to offer that opinion, even though Plaintiffs’ standing argument does not depend on it.

        As for the two cases the majority relies on, they do not compel its strained conclusion,
which is contrary to this court’s binding precedent in Sandusky. First, the majority points to
Lyons, where the Supreme Court rejected, on standing grounds, a suit that sought to enjoin the
Los Angeles Police Department (“LAPD”) from employing the use of 

–98. Specifically, the Court concluded that the plaintiff’s allegation that the LAPD sometimes
used chokeholds in the course of policing (including in the past on plaintiff himself) did not
establish an injury that was sufficiently imminent.

In other words, the plaintiff failed
to demonstrate that he was likely to once again have a chokehold used upon him in the future as
would be required to seek injunctive relief.

an individual plaintiff, not
organizations like Plaintiffs, and this distinction is fatal to the majority’s reasoning. Although
evidence of the LAPD’s regular use of chokeholds would not establish a likelihood that it will
employ a chokehold on a given individual, that same evidence could establish that the LAPD
will employ a chokehold in an interaction with a plaintiff organization’s members, so long as the
membership was large enough relative to the rate at which the LAPD employed chokeholds. In
the language of Sandusky, such an organizational plaintiff could not “know in advance” the
member who would find themselves on the receiving end of an LAPD chokehold, but it could
demonstrate the inevitability of such an interaction taking place. 

.
Lyons is of no consequence here.4

        Second, the majority relies on Shelby Advocates. But that case serves only to underscore
the applicability of Sandusky and the sufficiency of Plaintiffs’ asserted injury.                      In Shelby

        4
          The majority faults Plaintiffs for failing to present evidence that one of their members has had their
absentee ballot erroneously rejected in the past, analogizing to the plaintiff in Lyons, who had presented that
evidence the LAPD had previously used a chokehold on him in the past. Maj. Op. at 11–12. The majority’s
suggestion is that a plaintiff cannot have standing to seeking forward-looking injunctive relief unless they have
suffered a past injury. This is absurd. Nothing in the law of standing—which allows suits seeking injunctive relief
upon a showing of an “imminent” injury and the possibility of “future harm”—supports the majority’s careless
suggestion. See 

.
 No. 20-6046            Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.       Page 27


Advocates, the plaintiffs, including at least one organization, contended that the past occurrence
of election administration issues—“system vulnerabilities, previous equipment malfunctions, and
past election mistakes”—created a future risk of vote denial that was sufficient to constitute an
injury in 

.       The panel rejected this theory as lacking the requisite
imminence.     See

–83. 
     But in doing so, the panel specifically took the time to
distinguish Sandusky:

       In Sandusky, the challenged policy . . . made it “inevitable” that the defendants
       would deny individuals their voting rights. Here, by contrast, plaintiffs allege
       only policies that add risk to the ever present possibility that an election worker
       will make a mistake. No injury may occur at all. Any analogy to Sandusky falls
       short.


Id. at 983. 
As explained at length above, Plaintiffs have established the inevitability of erroneous
absentee ballot rejections due to perceived signature issues, and they have further demonstrated a
significant likelihood that their members will be among the ones inevitably affected.
Accordingly, Shelby Advocates does not control the outcome of this case.

       In sum, the majority’s analysis of Plaintiffs’ standing to represent the interests of its
members is flawed in numerous respects—misconstruing the evidence and the law in a
misguided effort to affirm the denial of a preliminary injunction. I have no doubt that Plaintiffs
have a “‘personal stake in the outcome’” that “‘assure[s] that concrete adverseness which
sharpens the presentation of issues’ necessary for the proper resolution of constitutional
questions.” 
Lyons, 461 U.S. at 101 
(quoting Baker v. Carr, 
369 U.S. 186
, 204 (1962)). By
concluding that Plaintiffs likely have not demonstrated such an interest, the majority has assured
there will be no timely resolution of Plaintiffs’ constitutional questions, let alone a proper one.
Such a resolution would not have been a mere advisory opinion, undermining the separation of
powers. It would have been the resolution of a genuine constitutional controversy, central to this
court’s fundamental role—one with hundreds if not thousands of legitimate votes in the balance.

                                                  B.

       I would also conclude that Plaintiffs are likely to establish direct organizational standing
to pursue their constitutional claims. To demonstrate direct organizational standing, a plaintiff
organization must show that it suffered a “palpable injury as a result of the defendants’ actions.”
 No. 20-6046            Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.      Page 28


MX 
Group, 293 F.3d at 333
. To do so, Plaintiffs point to various changes made to their
operations, activities, and expenditures to account for and protect against Tennessee’s signature
verification process.

       Plaintiffs’ direct organizational standing argument turns on this court’s decision in
Coalition for the Homeless, 
837 F.3d 612
. In that case, the organizational plaintiffs—including
an advocacy organization that conducted voter outreach for persons without homes—brought
suit to enjoin two newly enacted Ohio laws that, among other things:

       (1) required county boards of elections to reject the ballots of absentee voters and
       provisional voters whose identification envelopes or affirmation forms,
       respectively, contain an address or birthdate that does not perfectly match voting
       records; [and] (2) reduced the number of post-election days for absentee voters to
       cure identification-envelope errors, and provisional voters to present valid
       identification, from ten to seven . . . .


Id. at 618, 624. 
The defendants argued that the organizational plaintiffs’ asserted injury, which
stemmed from actions undertaken in response to the changes in the law, was insufficient to
establish direct organizational standing.
 Id. at 624. 
The panel disagreed, reasoning that the
change in law triggered a comprehensive shift in the advocacy organization’s voter outreach
sufficient to establish an injury in fact; before the new laws the organization had focused its
outreach efforts towards encouraging homeless individuals to vote early in-person, whereas after
the shift in the law occurred the organization had focused its outreach efforts on encouraging
homeless individuals to vote by mail.
 Id. This represented “an 
overhaul of the get-out-the-vote
strategy of an organization that uses its limited resources helping homeless voters cast ballots”
requiring “more volunteers, time, and expenditures” that went beyond “the ‘effort and expense’
associated with advising voters how to ‘comport’ with the law,” which alone would have been
insufficient to establish an injury in fact.
 Id. (quoting Fair Elections 
Ohio v. Husted, 
770 F.3d
456
, 460 (6th Cir. 2014)).

       Here, organizational plaintiffs have offered evidence of a similar overhaul in their
outreach and advocacy activities sufficient to constitute an injury in fact for standing purposes.
Take, for example Plaintiff Equity Alliance, a “grassroots advocacy group that seeks to equip
citizens with tools and strategies to engage in the civic process and empower them to take action
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.        Page 29


on issues affecting their daily lives.” R. 40-8 (Equity Alliance Decl. at ¶ 5) (Page ID #1585).
This year, Equity Alliance has redirected significant time, effort, and resources towards absentee
voting advocacy in Tennessee, which previously was not a focus for the organization due to the
high rate of in-person voting in the state prior to the COVID-19 pandemic.
 Id. (¶ 23–25) (Page
ID #1588). This shift in focus includes plans to dedicate the organization’s limited resources to
voter outreach regarding Tennessee’s signature verification process.
 Id. (¶¶ 40–41) (Page 
ID
#1591).    The other organizational plaintiffs have demonstrated similar overhauls of their
advocacy and outreach activities in response to the expected surge in absentee voting due to the
COVID-19 pandemic, including efforts specifically targeting signature verification. See R. 40-7
(APRI Decl. at ¶¶ 27–30) (Page ID #1580) (prior to this year, efforts focused on encouraging in-
person voting, but have shifted to absentee voting, including informing voters about signature
verification issues); R. 40-6 (MCLC Decl. at ¶¶ 17–19, 27–29) (Page ID #1570–72) (diverted
resources from in-person voting initiatives towards absentee voter outreach including informing
voters about the signature verification process and how to avoid erroneous rejections); R. 40-5
(Tennessee NAACP Decl. at ¶¶ 41–46, 56–66) (Page ID #1561–64) (shifted outreach to focus on
absentee voting, including efforts directed towards signature verification); R. 40-9 (Free Hearts
Decl. at ¶¶ 14–15, 27–31) (Page ID #1597, 1599–1600) (to the same effect). These efforts have
gone beyond simply informing constituents about how to comply with absentee voter laws, to
include activities that encourage voters to practice their signatures in order to avoid erroneous
rejections and to proactively reach out to election officials to ensure their absentee ballot has
been accepted. See, e.g., R. 40-5 (Tennessee NAACP Decl. at ¶¶ 61–63) (Page ID #1564).
These injuries are fairly traceable to Tennessee’s signature verification procedures and would be
redressed by a favorable judicial outcome because, if Plaintiffs obtained their desired injunction,
they could redirect the resources currently being expended on signature verification efforts to
other initiatives. See, e.g.
, id. (¶ 66) (Page 
ID #1564).

       Defendants raise a distinction between Coalition for the Homeless and this case.
Specifically, they point out that unlike in Coalition for the Homeless, where the organizational
plaintiff’s shift in organizational strategy resulted from newly enacted laws, Tennessee’s
absentee voting laws are “hardly new.” Appellees’ Br. at 32. I acknowledge that the newly
enacted status of the Ohio laws at issue in Coalition for the Homeless was a factor considered by
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.          Page 30


that panel, but I do not believe that renders the case inapplicable here. That is because Coalition
for the Homeless does not stand for the proposition that challenges to newly enacted laws and
only challenges to newly elected laws can result in a cognizable injury for the purposes of direct
organizational standing in cases involving state-law challenges. The case does not purport to
make such an all-encompassing pronouncement that would hamstring countless legitimate
lawsuits, and indeed, if it did so purport, that would be dicta. At most, the case stands for the
proposition that some sort of change in circumstances must have led to an organizational plaintiff
redirecting its efforts in some fashion. That could result from a change in the law, but it could
also result from any number of other things. The recent change in circumstances resulting from
the COVID-19 pandemic identified by the organizational plaintiffs here suffices to meet such a
standard, to the extent Coalition for the Homeless calls for it.

       But I do not think that Coalition for the Homeless stands for even that much. The upshot
of that case is that an organization must establish that it has changed or will change its activities
in some fashion to address the challenged law—it cannot rely on insignificant changes to
activities that it was already conducting. I glean this from the context in which the panel in
Coalition for the Homeless invoked the newly enacted status of the Ohio laws at issue, which
was in distinguishing the case from an earlier decision of this court, Fair Elections Ohio,
770 F.3d 456
. See Coalition for the 
Homeless, 837 F.3d at 624
. Fair Elections Ohio, another
elections case, involved a significantly less robust claim of injury. In essence, the organizational
plaintiff there claimed that it suffered an injury in fact sufficient to challenge “years-old”
absentee ballot procedures based on the fact that it needed to make changes to the content of its
volunteer trainings to account for the laws. Fair Elections 
Ohio, 770 F.3d at 459
–60. But the
organizational plaintiff had conceded it would already be providing broadly similar trainings at
already-scheduled meetings.       See
 id.    Thus, there were 
no changes of any meaningful
significance to the organization’s activities upon which to base the finding of an injury in fact.

       In this context, the panel’s invocation in Coalition for the Homeless of the newly enacted
status of the Ohio laws at issue is best understood as a helpful explanation for why the
organizational plaintiffs in that case stood on different footing rather than the articulation of a
new limitation on direct organizational standing.          Whereas in Fair Elections Ohio the
 No. 20-6046              Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.                        Page 31


organization claimed an injury on the basis of substantially similar pre-existing activities, in
Coalition for the Homeless the organizations claimed standing on the basis of overhauled
operations due to a change in the law. Read together, these cases stand for the uncontroversial
position that an organization may not manufacture standing from its pre-existing work, but must
demonstrate a significant shift in their operations, activities, or strategies. See, e.g., Common
Cause Ind. v. Lawson, 
937 F.3d 944
, 955 (7th Cir. 2019) (“That is not to say that organizations
have standing based solely on the baseline work they are already doing. They ‘cannot convert [ ]
ordinary program costs into an injury in fact.’ The question is what additional or new burdens
are created by the law the organization is challenging. It must show that the disruption is real
and its response is warranted.”) (internal citations omitted) (original alterations). As explained
above, Plaintiffs have made an adequate showing of that here.

         The majority’s only retort to this argument is to repeat its assertion that there is no
imminent threat that Plaintiffs’ members’ absentee ballots will be erroneously rejected, which
would justify Plaintiffs’ overhauls and expenditures. For the reasons explained above, that false
premise cannot support the majority’s conclusion that Plaintiffs are unlikely to establish direct
organizational standing any more than it can support the majority’s conclusion that Plaintiffs are
unlikely to establish standing as representatives of their members.                        Indeed, Plaintiffs have
demonstrated a likelihood of establishing standing under either avenue. Accordingly, I would
resolve the question of whether Plaintiffs are likely to succeed on the merits of their
constitutional claims—a question the majority so eagerly has avoided.5

                                                          II.

         For the first time on appeal, Defendants argue that this court should apply the same
analytical framework to address both of Plaintiffs’ constitutional claims.                               Specifically,
Defendants contend that this court should eschew the typical procedural due process analysis—
characterized by a two-step inquiry and balancing test the Supreme Court articulated in Mathews


         5
          Commensurate with the brevity of Defendants’ argument that Plaintiffs lack prudential standing, I will
briefly address it here. Given that Plaintiffs are asserting the rights of their members, they have plainly established a
“close relationship” justifying their bringing suit. Cf. Fair Elections 
Ohio, 770 F.3d at 461
. This court routinely
allows organizations to proceed in cases such as these. See, e.g., 
Sandusky, 387 F.3d at 574
.
 No. 20-6046              Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.                        Page 32


v. Eldridge, 
424 U.S. 319 
(1976)—in favor of the standard applicable to allegations that a state
law unduly burdens various constitutional rights involving elections—another balancing test set
forth by the Supreme Court, but this time in Anderson v. Celebrezze, 
460 U.S. 780 
(1983), and
Burdick v. Takushi, 
504 U.S. 428 
(1992). According to Defendants, renouncing Mathews in
favor of Anderson-Burdick is appropriate because Anderson-Burdick is the framework for
analyzing every constitutional challenge to a state election law, regardless of whether that claim
sounds in procedural due process, equal protection, or the fundamental right to vote. I disagree.

         To begin, our precedent rejects Defendants’ position—a point the majority neglects to
mention in unhelpfully noting the existence of these two constitutional tests but nothing further.6
Take, for example, Miller v. Lorain County Board of Elections, where a panel of this court
addressed the constitutionality of an Ohio ballot access law, which required an independent
congressional candidate to “file a nominating petition that contains valid signatures of at least
one percent of qualified electors voting in the last gubernatorial election who reside within the
district . . . where the election is to be held” in order to be placed on the ballot. 
141 F.3d 252
,
254 (6th Cir. 1998). The panel applied Anderson-Burdick to conclude that the ballot access law
did not offend the First and Fourteenth Amendments’ guarantees of freedom of speech and
association and the Fourteenth Amendment’s equal protection guarantees
, id. at 256–59, 
but
applied the standard due process analysis to determine that the state had provided the plaintiff
constitutionally sound procedures for challenging the state’s determination that 386 of the
signatures he produced were invalid
, id. at 259–60. 
And in Schmitt v. LaRose, a panel of this
court applied Anderson-Burdick to determine whether Ohio’s ballot-initiative laws imposed an
impermissible prior restraint on the plaintiffs’ political speech in violation of the First and
Fourteenth Amendments, 
933 F.3d 628
, 639 (6th Cir. 2019), cert. denied, 
140 S. Ct. 2803
(2020), but applied the standard due process analysis to determine whether the state afforded

         6
           Noting that there was no need to carry on after their conclusion that Plaintiffs “failed to show a substantial
likelihood of establishing standing,” the majority nevertheless offers musings on a few merits issues under the guise
of “provid[ing] limited commentary to guide the district court in the ultimate resolution of various issues in this
case.” Maj. Op. at 13. The question of which standard to apply to Plaintiffs’ procedural due process claim is one of
those the majority deigned to pass upon, although its musings on this question are particularly lackluster. The
majority simply notes the existence of the two tests and informs the district court it will need to choose between
them. Maj. Op. at 13–15. This does nothing to “guide” the district court—it is the equivalent of directions that tell
you that there is a fork in the road but not which prong to take.
 No. 20-6046             Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.     Page 33


aggrieved proponents of a ballot initiative adequate procedures to challenge a decision to leave
their initiative off the ballot
, id. at 642.
          To be sure, as Defendants point out, our precedent also broadly states that Anderson-
Burdick is “a single standard for evaluating challenges to voting restrictions.” Obama for Am. v.
Husted, 
697 F.3d 423
, 430 (6th Cir. 2012). Indeed, we have applied Anderson-Burdick in
analyzing challenges alleging that a state’s deadline for requesting an absentee ballot
impermissibly infringes upon the right to vote protected by the First and Fourteenth
Amendments, Mays v. LaRose, 
951 F.3d 775
, 791–92 (6th Cir. 2020), ballot access laws like the
one at issue in Miller, and even equal protection claims alleging disparate treatment of voters,
Obama for 
Am., 697 F.3d at 428
–30; but see 
Mays, 951 F.3d at 783 
n.4 (dicta expressing doubt
as to whether Anderson-Burdick properly applies to all equal protection challenges involving
state election laws). But it does not follow from the application of Anderson-Burdick in those
contexts that it should also be applied to Plaintiffs’ procedural due process claim here. That is
because a procedural due process claim is different in kind from the other sorts of constitutional
claims that a plaintiff might bring to challenge a state election law, which are fundamentally
substantive in nature.

          The typical constitutional challenge to a state election law is an effort to vindicate a
substantive right. For example, as in Miller, an independent candidate for Congress might argue
that requiring that they obtain a certain number of signatures before they can appear on a ballot
violates their rights to freedom of speech and political association protected by the First and
Fourteenth Amendments. The gravamen of the claim is that the Constitution outright forbids the
challenged conduct of the state:          in the example, imposing the signature requirement as
constituted.     Put differently, the allegation is that the state’s conduct is constitutionally
“wrongful,” however it goes about undertaking it. See Zinermon v. Burch, 
494 U.S. 113
, 125
(1990).

          Procedural due process claims are different. The gravamen of a procedural due process
claim is not that the government’s conduct—in the procedural due process context, the
deprivation of liberty or property—is constitutionally prohibited of its own accord, but that the
state’s procedures do not adequately protect against mistaken, unjustified, or erroneous
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.           Page 34


deprivations. See Carey v. Piphus, 
435 U.S. 247
, 259 (1978) (“Procedural due process rules are
meant to protect persons not from the deprivation, but from the mistaken or unjustified
deprivation of life, liberty, or property.”) (emphasis added). To continue the example based on
Miller, the independent candidate might argue that even if the state may impose the signature
requirement at issue, the state failed to provide adequate procedures by which the candidate
could challenge the state’s determination that the signatures that the candidate provided were
invalid (say, because they came from outside the electoral district), in violation of their
procedural due process rights. Simply put, in the procedural due process context, the inquiry
turns on procedure, not substance. See 
Zinermon, 494 U.S. at 125 
(“In procedural due process
claims, the deprivation by state action of a constitutionally protected interest in ‘life, liberty, or
property’ is not in itself unconstitutional; what is unconstitutional is the deprivation of such an
interest without due process of law.”) (original emphasis).

       Miller and cases like it, which apply Anderson-Burdick to substantive constitutional
challenges to state election laws but apply the usual procedural due process analysis to related
procedural due process claims, appropriately preserve the conceptual distinction between these
two kinds of challenges. Doing so is important because those two tests are carefully calibrated to
vindicate substantive and procedural rights respectively. On the one hand, the Anderson-Burdick
standard is classically substantive, balancing the state’s justification for regulating an election in
a particular way against the burden that regulation places on a substantive right. Compare
Burdick, 504 U.S. at 434 
(“A court considering a challenge to a state election law must weigh
‘the character and magnitude of the asserted injury to the rights protected by the First and
Fourteenth Amendments that the plaintiff seeks to vindicate’ against ‘the precise interests put
forward by the State as justifications for the burden imposed by its rule,’ taking into
consideration ‘the extent to which those interests make it necessary to burden the plaintiff’s
rights.’”) (quoting 
Anderson, 460 U.S. at 789
), with Susan B. Anthony List v. Driehaus, 
814 F.3d
466
, 473 (6th Cir. 2016) (analyzing First Amendment challenge to speech regulation by applying
strict scrutiny, which asks whether the statute “serve[s] a compelling state interest”); Tyler v.
Hillsdale Cnty. Sheriff’s Dep’t, 
837 F.3d 678
, 693 (6th Cir. 2016) (analyzing Second
Amendment challenge to firearm restriction by applying intermediate scrutiny, which asks
whether the government’s objective is “‘significant, substantial, or important’”) (quoting United
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.        Page 35


States v. Chovan, 
735 F.3d 1127
, 1139 (9th Cir. 2013)).            See also Erwin Chemerinsky,
Substantive Due Process, 15 Touro L. Rev. 1501, 1501 (1999) (“Substantive due process looks
to whether there is a sufficient substantive justification, a good enough reason for such a
deprivation [of life, liberty, or property].”). On the other hand, procedural due process analysis
has an unsurprisingly procedural bent, turning on the risk of error inherent in the established
procedural regime and the burden that imposing additional procedures would place on the state,
among other things. 
Mathews, 424 U.S. at 335
.

       The standards are superficially similar—balancing factors and considering state and
private interests—but differ in crucial respects. Most obviously, Anderson-Burdick does not
expressly account for procedure.       That is not to say that Anderson-Burdick forbids the
consideration of procedure, but the standard relegates procedural inquires to the background,
even where they bear on the substantive question. Moreover, these two standards do not always
focus the inquiry on the same government interests. Under Anderson-Burdick, the inquiry
centers on the state interest that justifies burdening the substantive right, 
Burdick, 504 U.S. at
434
, but under Mathews, the inquiry turns, at least in part, on a consideration of the state’s
interest in avoiding more burdensome 
procedures, 424 U.S. at 335
. This is significant, because
the government’s interest in imposing a given election law may justify a substantive burden in
the abstract, but it does not follow that the same interest justifies imposing that burden without
procedures to safeguard against erroneous applications of the requirement. Admittedly, the line
between procedure and substance is blurry, see Erie R. Co. v. Tompkins, 
304 U.S. 64
, 92 (1938)
(“The line between procedural and substantive law is hazy . . . .”) (Reed, J., concurring in part),
but I would not be so quick to obliterate entirely that line in this context by applying a
substantive standard, Anderson-Burdick, to procedural challenges. Doing so risks negating the
Constitution’s procedural protections by subjugating them to a secondary role (at best) under
Anderson-Burdick.

       Defendants do not convince me otherwise.          They contend that applying Anderson-
Burdick to procedural due process challenges “respects the States’ broad constitutional authority
to regulate elections and its [sic] important interests in ‘assur[ing] that elections are operated
equitably and efficiently.” Appellees’ Br. at 36 (quoting 
Burdick, 504 U.S. at 433
). That might
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.           Page 36


be more persuasive if this case confronted us with a choice between Anderson-Burdick and the
unforgiving strict scrutiny standard.      Indeed, as Defendants themselves recognize, “[t]he
Supreme Court adopted the Anderson-Burdick framework as a more deferential alternative to
strict scrutiny.”
 Id. at 34 
(citing 
Burdick, 504 U.S. at 433
). But procedural due process analysis
is a far cry from strict scrutiny and, as illustrated in the discussion that follows, affords
Defendants a flexible standard capable of balancing their asserted interests against those asserted
by Plaintiffs. In this regard, the procedural due process analysis is similar to Anderson-Burdick,
but with the added benefit of affording procedural considerations their due.

       With this discussion behind us, I will now turn to the application of these two standards.
I will apply the usual procedural due process analysis to Plaintiffs’ procedural due process claim.
As for Plaintiffs’ claim that Tennessee’s signature verification requirement violates the
fundamental right to vote, I will apply Anderson-Burdick. Ultimately, I would conclude that
Plaintiffs are likely to succeed under either standard, rendering much of what I have said above
somewhat academic. Nevertheless, that may not always be the case for other plaintiffs, and so it
is important to lay out the reasons that support taking this path before I begin upon it.

                                                 III.

       Plaintiffs’ first challenge to Tennessee’s signature verification process seeks to vindicate
their members’ (and Tennessee voters’) procedural due process rights, which the Fourteenth
Amendment protects. “The Fourteenth Amendment prohibits states from depriving individuals
of life, liberty, or property without due process of law.” Johnson v. Morales, 
946 F.3d 911
, 921
(6th Cir. 2020) (citing U.S. Const. amend. XIV, § 1). Thus, the first step in analyzing a
procedural due process claim is to determine “whether a liberty or property interest exists that
has been interfered with by the state.” 
Miller, 141 F.3d at 259
. If so, at the second step of the
analysis, the court must determine whether “the procedures attendant upon that deprivation were
constitutionally sufficient”—i.e., whether the state has afforded due process.
 Id. The district
court 
concluded that Plaintiffs were unable to establish a likelihood of success on the merits at
the first step because their procedural due process claim failed to identify a protected interest that
would trigger due process requirements. R. 77 (Mem. Op. at 21–22) (Page ID #2478–79). I
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.          Page 37


disagree and, reaching the second step of the analysis, would conclude that Plaintiffs are likely to
succeed on the merits of their procedural due process claim.

                                                 A.

       “Protected liberty interests spring from two possible sources, the due process clause itself
and the laws of the state involved.” Codd v. Brown, 
949 F.2d 879
, 882 (6th Cir. 1991) (citing
Kentucky Dep’t of Corr. v. Thompson, 
490 U.S. 454
, 460 (1989)). Plaintiffs’ primary argument
is that Tennessee law establishes a liberty interest in exercising the right to vote absentee by mail
and to have that vote counted. Appellants’ Br. at 25. Because I agree, I would decline to rule on
Plaintiffs’ secondary argument: that the Constitution itself establishes a liberty interest in the
absentee voting context sufficient to trigger due process requirements.

       This court synthesized the standard for determining whether state law creates a protected
liberty interest in Tony L. By and Through Simpson v. Childers:

       State-created liberty interests arise when a state places “substantive limitations on
       official discretion.”     A state substantively limits official discretion “by
       establishing ‘substantive predicates’ to govern official decisionmaking . . . and,
       further, by mandating the outcome to be reached upon a finding that the relevant
       criteria have been met.” The state statutes or regulations in question also must
       use “explicitly mandatory language” requiring a particular outcome if the
       articulated substantive predicates are present. Finally, the statute or regulation
       must require a particular substantive outcome. State-created procedural rights
       that do not guarantee a particular substantive outcome are not protected by the
       Fourteenth Amendment, even where such procedural rights are mandatory.

71 F.3d 1182
, 1185 (6th Cir. 1995) (internal citations omitted) (alteration original). The absentee
ballot laws at issue here fall squarely within this framework.

       To begin, Tennessee law establishes substantive criteria that, if met, entitle a voter to vote
absentee by mail. Tenn. Code § 2-6-201. If these predicates are met, then election officials
“shall” mail the voter an absentee ballot upon receiving a proper request and “shall” deposit the
absentee ballot to be counted upon receiving the ballot and concluding that the voter is entitled to
vote.
 

(g). 
In other words, Tennessee law mandates that election officials
provide absentee ballots to voters who satisfy established criteria, and mandates that those ballots
be counted upon their return. See Tony 
L., 71 F.3d at 1185
. Election officials have no discretion
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.           Page 38


where the established criteria are met—the statute uses mandatory language (“shall”) to dictate
the voter’s right to obtain and vote with an absentee ballot in these circumstances. See
 id.

       Tennessee law thus 
goes beyond providing voters a mere “expectation of receiving a
certain process,” insufficient to create a liberty interest, and directs “a particular substantive
outcome” with regard to a voter’s right to vote absentee by mail. See
 id. at 1185–86. 
Tennessee
voters “have an expectation that a particular result”—having their absentee ballots counted—
“will follow from a particular, required action”—an official determining that they are entitled to
vote absentee—that is mandated where the voter meets established substantive criteria. See
 id. at
1186 
(concluding that no liberty interest is created by a statute that requires state official to take
certain steps in response to reports of child abuse, but that does not mandate a particular outcome
from those actions); see also Pusey v. City of Youngstown, 
11 F.3d 652
, 656 (6th Cir. 1993)
(“The Ohio victim impact law does not create a liberty interest here because it only provides that
the victim has the right to be notified. The statute does not specify how the victim’s statement
must affect the hearing nor does it require a particular outcome based on what the victim has
said.”). The analysis is no more complicated than that. In these circumstances, Tennessee has
created a liberty interest in voting absentee by mail sufficient to trigger due process protection.

       And yet, the district court concluded otherwise. On appeal, Plaintiffs have correctly
identified the two primary errors that led the district court astray. First, the district court placed
undue significance on this court’s decision in League of Women Voters of Ohio v. Brunner,
548 F.3d 463 
(6th Cir. 2008). Second, the district court relied on inapposite authority involving
the liberty interests of incarcerated prisoners to conclude erroneously that only a narrow range of
interests—involving freedom from restraint—qualify as state-created liberty interests for the
purposes of a procedural due process analysis.

       In Brunner, a panel of this court dealt with a complaint alleging that Ohio was “utiliz[ing]
‘non-uniform rules, standards, and procedures’ that result[ed] in ‘massive disenfranchisement
and unreasonable dilution of the 
vote.’” 548 F.3d at 478
. The plaintiffs in that case brought
equal protection, substantive due process, and procedural due process claims, seeking injunctive
and declaratory relief in the face of allegations of myriad issues saturating Ohio’s voting system.

Id. at 466. 
When the district court denied the defendants’ motion to dismiss each of these claims,
 No. 20-6046           Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.         Page 39


they sought and were granted interlocutory review.
 Id. at 473. 
On appeal, the panel affirmed the
district court’s ruling as to the plaintiffs’ equal protection and substantive due process claims but
reversed the district court as to their procedural due process claims.
 Id. at 479. 
In a single
paragraph—reflecting the “brevity of argument” supporting the procedural due process claim—
the panel remarked, without citing authority, that the plaintiffs’ procedural due process claim
“subsume[d] procedural due process into the substantive due process analysis” and that the
plaintiff had failed to “allege[] a constitutionally protected interest.”
 Id.

        Brunner does not 
control the outcome of this case for at least three reasons. First, as
Plaintiffs rightly point out, Brunner does not address the circumstances in which state law can
create a liberty interest, let alone a liberty interest in voting absentee by mail under Tennessee
law. Indeed, the Brunner plaintiffs’ asserted liberty interest was simply the constitutionally
protected right to vote, and they do not appear to have argued that Ohio law created a cognizable
liberty interest for the purposes of procedural due process. See
 id. Cases do not 
stand for
propositions they did not consider.

        Second, unlike the procedural due process claim in Brunner, Plaintiffs’ procedural due
process claim is wholly distinct from its claim that Tennessee’s absentee ballot laws burden the
fundamental right to vote. In Brunner, the plaintiffs alleged that a litany of state practices
amounted, in the aggregate, to a “fundamentally unfair” voting system in violation of their
substantive due process 
rights. 548 F.3d at 478
–79. The plaintiffs’ procedural due process claim
duplicated those allegations of widespread error, rather than challenging particular facets of
Ohio’s voting system and identifying the procedural safeguards due process requires, as would
be expected of a standard procedural due process claim.
 Id. Here, in contrast, 
Plaintiffs have
alleged and pursued a distinct procedural due process claim that targets a specific absentee
voting procedure and clearly identifies the procedural safeguards that they believe due process
requires.

        Third, and finally, the plaintiffs in Brunner failed to adequately brief their procedural due
process argument.
 Id. at 479. 
For this reason alone, Brunner is not controlling here, nor was it
controlling on the district court below.
 No. 20-6046           Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.              Page 40


       As for the prison litigation cases that the district court relied on, they simply do not stand
for the proposition that “the kinds of interests that will be deemed ‘liberty’ interests in [the
procedural due process] context is narrow, relating only to freedom from restraint.” R. 77 (Mem.
Op. at 16) (Page ID #2473). Sandin v. Conner, 
515 U.S. 472 
(1995), the case from which the
district court distilled this sweeping principle, does not support the district court’s conclusion. In
Sandin, the Court addressed the standard it would apply to determine whether prison regulations
created a cognizable liberty interest for the purpose of prisoners’ procedural due process claims.
Noting various “undesirable effects,”
 id. at 482, 
that stemmed from cases treating mandatory
language in prison regulations as establishing cognizable liberty interests for procedural due
process purposes, the Court “recognize[d] that States may under certain circumstances create
liberty interests which are protected by the Due Process Clause,”
 id. at 483–84, 
but concluded
that “these interests will be generally limited to freedom from restraint which, while not
exceeding the sentence in such an unexpected manner as to give rise to protection by the Due
Process Clause of its own force, nonetheless imposes atypical and significant hardship on the
inmate in relation to the ordinary incidents of prison life,”
 id. at 484 
(citations omitted).

       Thus, Sandin does represent a change in the legal framework for analyzing the existence
of state-law created liberty interests in the context of prison regulations, shifting the inquiry from
one focused on the language of the regulation (as is the case for the typical state-created interest
analysis) back to one focused on the “nature of the deprivation” relative to the strictures of prison
life. See
 id. at 479–82, 485–87. 
But Sandin did not purport to displace the established standard
for determining whether a state law establishes a liberty interest triggering due process
requirements outside of the context of prison regulations. Instead, the Court expressly limited its
inquiry to “the circumstances under which state prison regulations afford inmates a liberty
interest protected by the Due Process Clause.”
 Id. at 474. 
Indeed, the Court emphasized the
unique position of prison litigation, reiterating its view that in the context of prisoner litigation
“federal courts ought to afford appropriate deference and flexibility to state officials trying to
manage a volatile environment.”
 Id. at 482. 
Moreover, the considerations that motivated the
Court—a desire to avoid “disincentives for States to codify prison management procedures”
while affording state officials the flexibility to “fine-tun[e] . . . the ordinary incidents of prison
life”—have no bearing when considering procedural due process claims that do not involve
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.           Page 41


prison regulations or incarcerated prisoners.
 Id. at 482–83. 
The Court recognized as much,
remarking that a focus on mandatory language “may be entirely sensible in the ordinary task of
construing a statute defining rights and remedies available to the general public” but that “[i]t is a
good deal less sensible in the case of a prison regulation primarily designed to guide correctional
officials in the administration of a prison.”
 Id. at 481–82.
       Although this court has embraced Sandin in the context of prisoner litigation, it has done
so while reiterating the same unique concerns implicated by prisoner litigation that motivated the
Court in Sandin. See Bazzetta v. McGinnis, 
430 F.3d 795
, 801 (6th Cir. 2005) (“In evaluating a
claimed liberty interest by prison inmates, courts are mindful that imprisonment necessarily
‘carries with it the circumscription or loss of many significant rights.’”) (quoting Hudson v.
Palmer, 
468 U.S. 517
, 524 (1984)). Indeed, this court has tacitly rejected the applicability of
Sandin outside the prison litigation context, applying the usual state-law created interests
standard outside of that context. See Tony 
L., 71 F.3d at 1185 
(remarking that under Sandin, the
analysis for prison regulations “should focus on the nature of the deprivation rather than the
language of the regulation involved” as it would for other state laws); see also Jasinski v. Tyler,
729 F.3d 531
, 541 (6th Cir. 2013).

       The district court brushed aside these material distinctions, reading a sweeping holding
into a narrowly drawn Supreme Court decision. This was unwarranted. As explained above,
Tennessee law creates a protected liberty interest in voting absentee by mail. Lower courts
addressing substantively similar state laws have come to the same conclusion with near
uniformity. See also Self Advoc. Sols. N.D. v. Jaeger, --- F. Supp. 3d ---, 
2020 WL 2951012
, at
*8 (D.N.D. 2020); Frederick v. Lawson, --- F. Supp. 3d ---, 
2020 WL 4882696
, at *12 (S.D. Ind.
2020); Richardson v. Texas Sec. of State, --- F. Supp. 3d ---, 
2020 WL 5367216
, at *20–21
(W.D. Tex. 2020); Democracy N.C. v. N.C. State Bd. of Elections, --- F. Supp. 3d ---, 
2020 WL
4484063
, at *53 (M.D.N.C. 2020); 
Saucedo, 335 F. Supp. 3d at 217
; Martin v. Kemp, 341 F.
Supp. 3d 1326, 1338 (N.D. Ga. 2018), appeal dismissed sub nom. Martin v. Sec. of State of
Georgia, 18-14503-GG, 
2018 WL 7139247 
(11th Cir. 2018); Zessar v. Helander, 05 C 1917,
2006 WL 642646
, at *5 (N.D. Ill. Mar. 13, 2006). I find their straightforward application of
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.          Page 42


established principles more persuasive than the unsupported approach taken by the district court
here.

                                                 B.

        Having concluded that Plaintiffs are likely to prove that there is a constitutionally
protected, Tennessee-law created liberty interest in voting absentee by mail, I would accept
Plaintiffs’ invitation to address the second step of the procedural due process inquiry, which the
district court eschewed. “Once it is determined that due process applies, the question remains
what process is due.” Morrissey v. Brewer, 
408 U.S. 471
, 481 (1972). Courts answer that
question using the familiar balancing test from Mathews, which directs us to balance the private
interest at stake “against the government’s interest in avoiding additional or substitute process, in
light of ‘the risk of an erroneous deprivation’ of a [liberty] interest ‘and the probable value, if
any, of additional or substitute procedural safeguards.’” Hicks v. Comm’r of Soc. Sec., 
909 F.3d
786
, 799 (6th Cir. 2018) (quoting 
Mathews, 424 U.S. at 335
). These considerations uniformly
favor Plaintiffs.

        In cases like this one—where the state has failed to provide even the most rudimentary
ingredients of due process—it is appropriate to begin with a consideration of “‘the risk of
erroneous deprivation’” and the value inherent in additional procedures.            See
 id. (quoting
Mathews, 424 
U.S. at 333). That is because, despite all the case-specific flexibility of the
procedural due process inquiry, Leary v. Daeschner, 
228 F.3d 729
, 743 (6th Cir. 2000) (“Due
process is a flexible principle whose requirements depend on the facts of the individual case”),
“‘the government is never relieved of its duty to provide some notice and some opportunity to be
heard prior to a final deprivation of a [liberty] interest,’” 
Hicks, 909 F.3d at 799 
(original
emphasis) (quoting Propert v. District of Columbia, 
948 F.2d 1327
, 1332 (D.C. Cir. 1991)).
Thus, “[a]t some foundational level, this factor is dispositive.” 
Hicks, 909 F.3d at 800
. Where
the state has failed to provide the most basic requirements of due process—notice and a
meaningful opportunity to be heard—the private and government interests take a back seat. See
id.; see also Flaim v. Med. Coll. of Ohio, 
418 F.3d 629
, 635 (6th Cir. 2005) (“Notice and an
opportunity to be heard remain the most basic requirements of due process.”).
 No. 20-6046           Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.      Page 43


        Tennessee’s absentee voting law fails to provide these fundamental protections against
the risk of erroneous rejections of absentee ballots on account of perceived signature invalidity.
Presently, Tennessee law requires election officials to notify absentee voters if their ballot is
rejected, apparently including where there is a signature verification issue. Tenn. Code §§ 2-6-
202(g), 2-6-204(b). However, the state does not afford the voter an opportunity to cure the
signature issue before the rejection occurs. See
 id. Plaintiffs seek, primarily, 
a procedure that
would provide for pre-rejection notice, and an opportunity to cure any signature defect before
their absentee ballot is rejected.

        I begin with notice. The Tennessee statute does provide notice, but only after election
officials have rejected the ballot.      Post-deprivation notice is appropriate in only limited
circumstances that do not apply here—this is not an emergency situation requiring immediate
action, and Tennessee cannot effectively remedy an erroneously rejected absentee ballot once the
election is over. See 
Johnson, 946 F.3d at 921 
(identifying the exceptions to the “general rule
that due process ‘requires some kind of a hearing before the State deprives a person of liberty or
property.’”) (quoting 
Zinermon, 494 U.S. at 127
); Quinn v. Shirey, 
293 F.3d 315
, 319 (6th Cir.
2002) (“Except in exceptional circumstances, not applicable here, before a person is deprived of
either a liberty or property interest, he has a right to some kind of hearing.”). Thus, the notice
provided by Tennessee law is legally insufficient.

        As for an opportunity to be heard, Tennessee provides none. Even with adequate notice,
this would doom the state’s argument—“after all, ‘some form of hearing is required before an
individual is finally deprived of a [liberty] interest,’ no matter how small the interest or how
great the governmental burden.” 
Hicks, 909 F.3d at 800 
(quoting 
Mathews, 424 U.S. at 333
(emphasis added)). Plaintiffs are asking Defendants to provide the baseline requirements of due
process—their request for an opportunity to cure is best understood as one for a form of
hearing—but have been denied.

        I do not place any significance on Defendants’ (and now the majority’s) insistence that
some voters whose absentee ballots are rejected, may, if they receive their rejection notice in
time, be able to request a new absentee ballot or vote in person. This relies, as Defendants
acknowledge, on the assumption that the voter will receive the notice of signature invalidity
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.            Page 44


sufficiently ahead of the election, such that they could take advantage of these other avenues.
But aside from Defendants’ vague assertions that this is possible “if time allows,” Appellees’ Br.
at 10, it is far from clear that the notice would arrive in time. At the absolute least, it seems
exceedingly unlikely that notices of signature invalidity would arrive in time for those voters
who submitted the absentee ballot requests a week before election day, as Tennessee law allows
them to do. See Tenn. Code § 2-6-202(a). And even if the state were able to timely send these
notices in prior years, delay is inevitable given the anticipated surge in absentee ballots this year.
Furthermore, even where an absentee voter did receive notice of signature invalidity in time,
there is no reason to think that they would be able to avail themselves of the in-person voting
options, given that voting absentee by mail in Tennessee is generally limited to voters who
cannot go to the poll for significant reasons. See generally
 id. § 2-6-201. That 
being the case,
and given Plaintiffs’ unrefuted expert testimony regarding the significant flaws in Tennessee’s
signature verification training regimen, it is a certainty that a significant number of erroneous
rejections of ballots will occur, and that this will result in disenfranchisement. Especially given
the anticipated surge in voting absentee by mail in Tennessee—the state has printed four million
absentee ballots in preparation for the election, R. 40-2 (Doshi Decl., Ex. 2 at 7) (Page ID
#221)—this is a disgrace.

       In any case, the remaining factors also favor Plaintiffs. Whether derived from the
Constitution or Tennessee law, the private interest at stake is significant. Simply put, voters—
however they cast their ballot—have a profound interest in having that ballot counted. See
McCutcheon v. Fed. Election Comm’n, 
572 U.S. 185
, 191 (2014) (“There is no right more basic
in our democracy than the right to participate in electing our political leaders.”); Illinois State Bd.
of Elections v. Socialist Workers Party, 
440 U.S. 173
, 184 (1979) (“And for reasons too self-
evident to warrant amplification here, we have often reiterated that voting is of the most
fundamental significance under our constitutional structure.”); Yick Wo v. Hopkins, 
118 U.S.
356
, 370 (1886) (“voting . . . . is regarded as a fundamental political right, because preservative
of all rights”). The state does not reduce the significance of this interest by offering voters
alternatives to in-person voting on election day.
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.           Page 45


       As for the state’s interests, first, Defendants assert an interest in “‘counting only the votes
of eligible voters.’” Appellees’ Br. at 48 (quoting Crawford v. Marion Cnty. Election Bd.,
553 U.S. 181
, 196 (2008) (plurality opinion) (emphasis added)).             Fair enough.      But an
opportunity to cure signature defects is fully consistent with this interest because it ensures votes
will be counted more accurately than they would be otherwise.            See Frederick, 
2020 WL
4882696
, at *15 (“The State’s claimed interests here, to wit, in preventing voter fraud and
maintaining election integrity, are undeniably compelling interests. As Plaintiffs point out,
however, providing mail-in absentee voters notice and the opportunity to cure a perceived
signature mismatch by confirming their identity in fact promotes these important governmental
interests.”); 
Saucedo, 335 F. Supp. 3d at 220 
(“[I]f anything, additional procedures further the
State’s interest in preventing voter fraud while ensuring that qualified voters are not wrongly
disenfranchised.”). Indeed, the state itself, as much as the public, has an interest in “permitting
as many qualified voters to vote as possible.”          See Obama for 
Am., 697 F.3d at 436
.
Furthermore, the state’s interest in maintaining voter confidence, see 
Crawford, 553 U.S. at 194
–
96, favors an injunction—by affording voters an opportunity to cure perceived signature defects,
the state assures its citizens that proper votes will be counted, and improper votes will not. A
system that wrongly invalidates legitimate votes is at least as damaging to voter confidence as a
system that does not protect against the vague threat of voter fraud. See Purcell v. Gonzalez, 
549
U.S. 1
, 4 (2006) (“Voters who fear their legitimate votes will be outweighed by fraudulent ones
will feel disenfranchised.”).

       Second, Defendants assert an interest in orderly elections.             Again, however, an
opportunity to cure perceived signature defects furthers the government’s interest rather than
harming it. Defendants’ preferred alternative to an opportunity to cure is to have voters take
steps to ensure their vote is counted, before or after receiving a signature rejection notice. For
example, Defendants insist that a voter whose absentee ballot is rejected can request another
absentee ballot, cast a provisional ballot in person prior to the election, or vote in person on
election day. Appellees’ Br. at 10. They even suggest that a concerned voter could cast a
provisional ballot in person before receiving a rejection of their absentee ballot.
 Id. I 
fail to see
how these are more orderly alternatives to a streamlined opportunity to cure a perceived
signature defect that forecloses the possibility of confused voters clogging up polling places
 No. 20-6046            Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.         Page 46


despite the fact that they have already attempted to vote. And to the extent the state’s argument
targets Plaintiffs’ request that the opportunity to cure should extend beyond election day, I fail to
see how this disrupts the state’s electoral process where Tennessee law provides election
officials until the third Monday after the election to certify election results, Tenn. Code § 2-8-
101(a), Defendants themselves acknowledge that counting absentee ballots may end up taking
days, R. 46-1 (Goins Decl. at ¶ 9) (Page ID #1826), and Tennessee law already provides post-
election-day cure periods for failures to provide proper identification when submitting a
provisional ballot
, id. § 2-7-112(e)(5). Finally, 
Defendants’ suggestion that absentee voters who
are concerned that their mail ballots will be rejected should cast a preemptive provisional ballot
in person in addition to their absentee ballot is more that “passing strange,” Appellants’ Reply at
14 n.11; indeed, it is downright farcical. Asking voters to work under the assumption that the
state will fail to accurately employ its procedures is not a replacement for due process. Rather, it
is an acknowledgement that the state has no interest in counting these votes at all. The effect of
the majority is to tolerate this sham.

       Defendants do not seriously argue that Plaintiffs’ requested procedures would create a
significant administrative burden. Nor could they given their insistence that a “strikingly small”
number of ballots are likely to be rejected. See, e.g., Appellees’ Br. at 11. Indeed, some counties
in Tennessee already go beyond the statutory procedure and take informal steps to help voters
cure signature irregularities, suggesting it would not be burdensome to mandate an opportunity to
cure across the board. R. 46-4 (Farley Decl. at ¶ 6) (Page ID #1857); R. 46-3 (Warren Decl. at
¶ 5) (Page ID #1854).

       In sum, Tennessee’s absentee ballot signature verification procedures fail to provide even
the baseline protections required by due process. As a result, Plaintiffs would be likely to
succeed on the merits even if their liberty interest were minimal and the state’s interests were
significant. The opposite is true here—Plaintiffs’ interests are significant and the state’s interests
are not substantial—further demonstrating Plaintiffs’ likelihood of success on their procedural
due process claim. The district court erred in concluding otherwise, and the majority erred
further in evading the question.
 No. 20-6046         Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.          Page 47


                                               IV.

       Plaintiffs also allege that Tennessee’s absentee ballot signature process unconstitutionally
burdens their right to vote under the First and Fourteenth Amendments. This claim does not turn
on whether Plaintiffs have a constitutionally protected right to vote absentee by mail; rather, it
turns on the burden that Tennessee has placed on their right to vote, which their members
indisputably enjoy, and whether that burden exceeds that which the Constitution will endure.

       “The right to vote is a ‘precious’ and ‘fundamental’ right.” Obama for 
Am., 697 F.3d at
428 
(quoting Harper v. Va. State Bd. of Elections, 
383 U.S. 663
, 670 (1966)). Indeed, “[i]t is
beyond cavil that ‘voting is of the most fundamental significance under our constitutional
structure.’” 
Burdick, 504 U.S. at 433 
(quoting Illinois Bd. of 
Elections, 440 U.S. at 184
).
Recognizing, however, that “[e]lection laws will invariably impose some burden upon individual
voters,” the constitutional question is not whether the state law in question burdens the right to
vote, but to what extent, and whether it is justified.
 

n short, “while 
States can regulate
elections, they must be careful not to unduly burden the right to vote when doing so.” 
Mays,
951 F.3d at 783 
(emphasis added).

       As explained above, we assess whether a state has unduly burdened the right to vote by
applying the “flexible” Anderson-Burdick standard. 
Daunt, 956 F.3d at 406
. “Under this
standard, the rigorousness of our inquiry into the propriety of a state election law depends upon
the extent to which a challenged regulation burdens First and Fourteenth Amendment rights.”
Burdick, 504 U.S. at 434
. Thus, “[u]nder Anderson-Burdick, we first look to the burden the
State’s regulation imposes on the right to vote.” 
Mays, 951 F.3d at 784
. If the burden is
“severe,” then we apply strict scrutiny.
 Id. I
f the burden is reasonable, nondiscriminatory, and
otherwise minimal, then we apply rational basis review. See id.; Timmons v. Twin Cities Area
New Party, 
520 U.S. 351
, 358–59 (1997). Of course, this leaves cases that “fall between these
two extremes.” 
Mays, 951 F.3d at 784
.

       For these intermediate cases, where the burden on the right to vote is moderate,
       we must weigh that burden against “‘the precise interests put forward by the State
       as justifications for the burden imposed by its rule,’ taking into consideration ‘the
       extent to which those interests make it necessary to burden the plaintiffs’ rights.’”
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.           Page 48


Id. (quoting 
Burdick, 504 U.S. at 434
).         After all, especially in this intermediate zone,
“[n]o bright line separates permissible election-related regulation from unconstitutional
infringements on First Amendment freedoms.” 
Timmons, 520 U.S. at 359
.

       This case falls in such an intermediate zone, as most cases do. See Obama for 
Am.,
697 F.3d at 429 
(“Most cases fall in between these two extremes.”). On the one hand, where
Tennessee’s signature verification process results in an erroneous rejection, the burden is
significant, particularly so in those cases where due to the timing of the rejection the voter has no
opportunity to cast a ballot. On the other hand, it does seem as though at least some voters who
are timely notified of the rejection of their absentee ballots due to a perceived signature error will
be able to successfully vote (although I harbor significant doubts about how often that will be the
case), either by requesting and resubmitting a new absentee ballot by mail or voting in person, to
the extent they are able. Unlike the district court, I do not think Anderson-Burdick requires us to
artificially divorce the signature verification requirement from the procedures that render it
burdensome. See 
Crawford, 553 U.S. at 198
–99 (plurality opinion) (regarding voter ID law and
opportunities to vote without the ID requirements). But it is of no matter, because the district
court ultimately assumed, correctly, that Tennessee’s signature verification process fell within
the intermediate zone and proceeded accordingly. R. 77 (Mem. Op. at 35) (Page ID #2492).

       To resolve this case, I do not think it necessary further to delineate where exactly that
burden on the right to vote falls on the Anderson-Burdick scale. That is because, “[h]owever
slight that burden may appear, . . . it must be justified by relevant and legitimate state interests
‘sufficiently weighty to justify the limitation.’” 
Crawford, 553 U.S. at 191 
(quoting Norman v.
Reed, 
502 U.S. 279
, 288–89 (1992)). Here, Defendants’ asserted interests do not suffice for the
same reasons discussed above with regard to Plaintiffs’ procedural due process claim; those
interests favor offering an opportunity to cure signature verification rejections, even after
election day. In short, Defendants’ asserted interests harm their case rather than helping it.
Plaintiffs established a likelihood of success on the merits of their claim that Tennessee’s
signature verification process violates the First and Fourteenth Amendments.
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.              Page 49


                                                  V.

       “When a party seeks a preliminary injunction on the basis of a potential constitutional
violation, ‘the likelihood of success on the merits often will be the determinative factor.’”
Obama for 
Am., 697 F.3d at 436 
(quoting Jones v. Caruso, 
569 F.3d 258
, 265 (6th Cir. 2009)).
With that said, this election case involves uncommon circumstances with regard to the final three
“equitable factors” that inform a decision on whether to grant or deny a preliminary injunction:
whether the movant will suffer irreparable harm and the public and government interests. In
light of those considerations, I would conclude that the equities favor Plaintiffs and that a
preliminary injunction should have issued below.

       Beginning with irreparable harm, “[w]hen constitutional rights are threatened or
impaired, irreparable injury is presumed.” Obama for 
Am., 697 F.3d at 436 
(citing ACLU of Ky.
v. McCreary County, 
354 F.3d 438
, 445 (6th Cir. 2003)). Here, the primary harm that will be
suffered is the erroneous rejection of otherwise legitimate votes. There is no remedy for such a
harm once the election is over. See
 id. (“‘A plaintiff’s harm 
from the denial of a preliminary
injunction is irreparable if it is not fully compensable by monetary damages.’”) (quoting
Certified Restoration Dry Cleaning Network, L.L.C. v. Tenke Corp., 
511 F.3d 535
, 550 (6th Cir.
2007)). Thus, our precedent firmly establishes that “[a] restriction on the fundamental right to
vote therefore constitutes irreparable injury.”
 Id.

       The district court 
took a different approach, concluding that Plaintiffs failed to establish
an irreparable harm after concluding that there is an “exceeding[ly] low” rejection rate for
absentee ballots. R. 77 (Mem. Op. at 39) (Page ID #2496). It followed, according to the district
court, that even though rejections occur, and even though “presumably,” some of those rejections
are erroneous, and even though Plaintiffs had “tens of thousands of members,” Plaintiffs had not
shown a likelihood of irreparable harm without some more specific assertion from Plaintiffs as to
why they are likely to be harmed.
 Id. at 38–42 
(Page ID #2495–99). The majority beats a
similar track, repeating again its perception that Plaintiffs’ harm is entirely speculative.

       I think, to reverse the old adage, that the district court—and now the majority—missed
the trees for the forest. When we review a facial challenge like Plaintiffs have presented here,
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.          Page 50


our focus “is the group for whom the law is a restriction, not the group for whom the law is
irrelevant.” Planned Parenthood of Se. Pennsylvania v. Casey, 
505 U.S. 833
, 894 (1992); see
also 
Kemp, 918 F.3d at 1269
–70. Thus, we are concerned with the nature of the harm, not its
frequency. See Frank v. Walker, 
819 F.3d 384
, 386 (7th Cir. 2016) (“The right to vote is
personal and is not defeated by the fact that 99% of other people can secure the necessary
credentials easily.”). As Plaintiffs aptly put it, “The fact that others have their ballots counted
will be cold comfort to those disenfranchised. Even a single absentee ballot erroneously rejected
comprises an irreparable injury warranting this Court’s scrutiny.”            Appellants’ Br. at 48.
Tennessee’s low rate of rejection is a distraction from this inquiry, and it permeated the district
court’s reasoning, now captivating the majority as well. Plaintiffs substantiated the likelihood
that some of their members would have their absentee ballots erroneously rejected due to a
perceived signature invalidity. That was enough.

       Moving on, neither party spends much time on the public and government interests at
stake, which makes sense given that it was not a focus for the district court and would be
somewhat duplicative of the constitutional arguments above. Nor does the majority pass upon
these factors. Suffice it to say, “[t]he public interest . . . favors permitting as many qualified
voters to vote as possible.” Obama for 
Am., 697 F.3d at 436
–37. The state—I would think—
shares this interest. And, as explained above, the states’ interests are furthered by an opportunity
to cure signature invalidity issues, not hampered by it.        Plaintiffs’ requested relief is not
particularly onerous, as an administrative matter, especially compared to relief we have upheld in
similar contexts. See
 id. (three extra days 
of in-person and early voting).

       More fervently briefed is one final factor—timing. Indeed, this factor is high fashion in
election cases these days. See, e.g., Republican Nat’l 
Comm., 140 S. Ct. at 1207
; Bostelmann,
2020 WL 5951359
, at *1. Defendants argue both that Plaintiffs improperly delayed seeking
relief and that in any case, relief is inappropriate this close to the 2020 General Election.
Defendants’ first objection—Plaintiffs’ purported delay in bringing this suit—is unpersuasive.
Plaintiffs brought suit just a week after Tennessee issued its contingency plan for the 2020
elections, which anticipated an unprecedented surge in voting absentee by mail. Before that
time, which was at the beginning of the COVID-19 pandemic, Plaintiffs can be forgiven for not
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.          Page 51


focusing their efforts on Tennessee’s absentee voting laws given that historically only about
2.5% of Tennesseans made use of that option. I do not fault Plaintiffs for failing to anticipate an
abrupt, unexpected, and paradigm-altering occurrence like the COVID-19 pandemic and
responding to it promptly when it became apparent that it would have implications for the 2020
election that they wished to address.

       As for Defendants’ second timing argument, I do not think Supreme Court precedent
forecloses Plaintiffs’ relief due to the proximity of the November 3, 2020 General Election,
although I am cognizant that the Supreme Court “has repeatedly emphasized that lower federal
courts should ordinarily not alter the election rules on the eve of an election.” Republican Nat’l
Comm., 140 S. Ct. at 1207 
(citing Purcell, 
549 U.S. 1 
(2006) (per curiam)). Under this “Purcell
doctrine,” lower courts are required to weigh various factors specific to election cases, including
the possibility of creating voter confusion, and risking disincentivizing voters from going to the
polls. 
Purcell, 549 U.S. at 4
–5. I agree with Plaintiffs that Purcell is no obstacle here because
Plaintiffs’ preliminary injunction would minimize voter confusion and does nothing to
disincentivize voter turnout.

       Regarding voter confusion, Plaintiffs’ injunction would add a final step to Tennessee’s
signature verification process, requiring election officials to give an opportunity to cure signature
defects before a ballot is rejected.     This added procedural safeguard would prevent voter
confusion caused by erroneous signature rejections by offering voters a clear mechanism by
which to ensure their vote is counted. Without this added procedure, Defendants suggest voters
could ensure their vote is counted by voting early or the day of the election in person, even if they
are not sure whether their absentee ballot has been counted. This is certain to lead to confusion
for the voter themselves, for poll workers, and for others. Purcell does not mandate us to rubber-
stamp such a warped outcome. Moreover, even if this added mechanism somehow prompted
voter confusion, it would be far from widespread. As Defendants have tirelessly repeated,
absentee ballot rejections are relatively rare.     Defendants have not presented any cogent
argument for why these rare occurrences would lead to the sort of voter confusion that would
mandate leaving in place an otherwise constitutionally inadequate procedure. After all, there is
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.          Page 52


no reason to think most voters will even become aware of the absentee ballot curing system if
they are not one of the relative few who have their absentee ballots rejected.

       Similarly, I do not think Plaintiffs’ requested relief disincentivizes voter turnout. To the
contrary, I think it more likely that it will increase the number of proper votes counted.
Specifically, it seems to me that absentee voters in Tennessee who receive signature rejection
notices are unlikely to make use of the alternative voting methods that might be available to them
even if they were to receive notice in time that their ballots were rejected. Absentee voters in
Tennessee must qualify as such, in many cases because they are unable to vote in person, at least
without seriously risking their health and safety during the COVID-19 pandemic. And having
their absentee ballots rejected for perceived signature issues is likely to lead to frustration and a
lack of confidence in their state’s electoral process as much as anything else.

       To quote Plaintiffs, “[t]he Purcell doctrine . . . is not a magic wand that prevents voters
from obtaining relief in an election year.” Appellants’ Br. at 49. Rather, it is an equitable
doctrine that mandates that we consider concerns unique to election cases. Those considerations
would not warrant withholding Plaintiffs’ requested relief of an opportunity to timely cure
perceived signature issues. The COVID-19 pandemic has upended American life in countless
ways. In Tennessee, one of those ways is an expected surge in absentee voting, which has shone
the spotlight on heretofore unscrutinized aspects of Tennessee’s absentee voting system. Having
concluded that one of those aspects—the state’s signature verification procedure—is likely
constitutionally inadequate in failing to provide an opportunity to cure, I do not think Purcell
would have prevented us from acting. Indeed, the equities favor the issuance of a preliminary
injunction, and the district court erred in refusing it. The majority—in failing to even reach the
merits—irreparably compounded that mistake.

                                                VI.

       “While I am saddened, I am not surprised by today’s ruling.” Warshak v. United States,
532 F.3d 521
, 538 (6th Cir. 2008) (en banc) (Martin, J., dissenting). That is because many
federal courts—more specifically, many federal courts of review—have sanctioned a systematic
effort to suppress voter turnout and undermine the right to vote. Rarely does this have anything
 No. 20-6046          Memphis A. Philip Randolph Inst., et al. v. Hargett, et al.        Page 53


to do with the merits of the case. No, the effort has not been so bold as that. Most often, Purcell
provides the cover—a convenient court-made doctrine that provides plausible deniability
sounding in vague cries of “confidence in the electoral process.” See 
Purcell, 549 U.S. at 4
.
Today, however, standing is the shroud of choice. Whatever the disguise, the result is the same.
Hiding behind closed courthouse doors does not change the fact that ruling by ruling, many
courts are chipping away at votes that ought to be counted. It is a disgrace to the federal courts’
foundational role in ensuring democracy’s function, and a betrayal to the persons that wish to
participate in it fully. See Wesberry v. Sanders, 
376 U.S. 1
, 17 (1964) (“Other rights, even the
most basic, are illusory if the right to vote is undermined.”).

       On its own, today’s ruling may not—likely will not—change the course of this election.
But it is another drop in the bucket that is the degradation of the right to vote in this country.
See, e.g., 
Raysor, 140 S. Ct. at 2600 
(“This Court’s order prevents thousands of otherwise
eligible voters from participating in Florida’s primary election simply because they are poor.”)
(Sotomayor, J., dissenting); Republican Nat’l 
Comm., 140 S. Ct. at 1211 
(“The majority of this
Court declares that this case presents a ‘narrow, technical question.’ That is wrong. The question
here is whether tens of thousands of Wisconsin citizens can vote safely in the midst of a
pandemic.”) (internal citation omitted) (Ginsburg, J., dissenting); Bostelmann, 
2020 WL
5951359
, at *13 (“It is a virtual certainty that current conditions will result in many voters,
possibly tens of thousands, being disenfranchised absent changes to an election code designed
for in-person voting on election day.”) (Rovner, J., dissenting). I fear the day we come out from
behind the courthouse doors only to realize these drops have become a flood.

       I dissent.