Opinion
24-5141
10-24-2024
Michael L. Walden, Plaintiff-Appellant, v. General Electric International, Inc.; Communications Workers of America, AFL-CIO, CLC, Defendants-Appellees.
Austin P. Vowels, VOWELS LAW PLC, Henderson, Kentucky, for Appellant. Jay Inman, LITTLER MENDELSON, P.S.C., Lexington, Kentucky, for Appellee General Electric. Robert F. Holt, IUE-CWA, Dayton, Ohio, for Appellee Communication Workers of America.
Appeal from the United States District Court for the Western District of Kentucky at Owensboro. No. 4:19-cv-00159-David J. Hale, District Judge.
COUNSEL
ON BRIEF:
Austin P. Vowels, VOWELS LAW PLC, Henderson, Kentucky, for Appellant.
Jay Inman, LITTLER MENDELSON, P.S.C., Lexington, Kentucky, for Appellee General Electric. Robert F. Holt, IUE-CWA, Dayton, Ohio, for Appellee Communication Workers of America.
Before: McKEAGUE, KETHLEDGE, and NALBANDIAN, Circuit Judges.
OPINION
NALBANDIAN, Circuit Judge.
Multiple times, Michael Walden applied for a manufacturing job with General Electric (GE). Multiple times, he failed required tests. When he didn't get the job, he sued GE for age discrimination. Walden also sued his union, alleging unfair representation in his challenges to the company's hiring decisions. GE and the union both point to the failed tests to explain their dealings with him. The district court granted summary judgment for GE and the union. We affirm.
I.
Walden was born in 1955. For several decades he worked as a tool and die maker. Then, in September 2011, he took a job with GE as a "Flexible Machine Operator II." R. 113-5, Signed Offer, p. 1, PageID 1539. While at GE, Walden joined the union, an affiliated local with the Communications Workers of America. Eventually, he began applying for a promotion to toolmaker.
Under the Collective Bargaining Agreement (CBA) then in place, the company had to fill toolmaker jobs "on the basis of Total Plant Seniority to those employees that possess the minimum qualifications." R. 113-1, CBA, p. 3, PageID 1286. For some time, GE had evaluated candidates' qualifications and seniority from resumes alone. Then, in 2016, GE and the union agreed to implement a written test requirement, too. Several toolmakers in the union designed the test, which initially had twenty questions. The toolmakers agreed that "anyone that understood toolmaking should not miss more than three questions," or in other words, score lower than 85%. R. 113-6, Riggs Deposition, p. 17, PageID 1556. The graded tests of candidates GE hired confirm as much; successful candidates all passed with a score of 85% or better.
In December 2018, GE had two toolmaker openings. It published a job posting that read as follows:
Qualifications:
Graduate of GE approved Tool & Die Apprenticeship Program or 5 years of Tool and Die job experience.
Job Description:
Build and assemble various types of tools, fixtures, gages, dies, etc. of intricate and complicated designs requiring the sue [sic] of equipment to produce compound angles and complicated dimensional and functional relationships, mechanical movements, and multi-plane machining. Diagnose and repair original designs of complex and intricate tools and fixtures....
Note: The job's classifications and definitions are merely for purposes of identification and general description and do not purport to be all inclusive or exhaustive of the actual requirements of any job so classified or defined. To be considered, you must list your qualifications or attach a resume or record of experience.R. 113-2, 2018 Job Posting, p. 1, PageID 1313. Walden applied, as did two others-Brad Brown and Nathan Bryant. Both Brown and Bryant were younger than forty and were less senior than Walden. All three candidates had the required level of experience, so GE had them take the test. Based on shift availability, Brown and Bryant tested first. GE manager Tim Riggs administered the test.
But before Walden could also test, union steward Kevin Christian informed Riggs of a rumor that someone had passed around test questions and answers. Riggs testified that he and Christian took the matter to the skilled trades chairman (confusingly also named Brad Brown), and the three agreed to add five new questions to the test. They did not change the passing score. Shortly after, Walden took the now-twenty-five-question test while Brown (the candidate) and Bryant came back to answer the five new questions.
The passing score remained the same whether scores were based on missed answers or percentages. On both versions of the test, missing more than three questions would result in a score below 85%. Missing four out of twenty would result in an 80%, while missing four out of twenty-five would garner an 84%.
Riggs graded the tests. When Walden turned his test in, he had a brief conversation with Riggs about some questions, including one that he did not answer. Riggs considered Walden's concerns about the question and agreed not to count it against him. Walden also claims that Riggs made a vague comment that "it's not going to be an issue," though it's not obvious what this comment might have referred to. R. 113-4, Walden Deposition, p. 57, PageID 1371.
Brown and Bryant passed. Walden did not. According to Walden, Riggs informed him that he had missed six while Brown and Bryant had each missed four. But the graded tests show something else. Brown's test displays no missed answers, a perfect score. Bryant's test had two wrong answers, a 92%. Walden's test, on the other hand, shows five wrong answers, an 80%. With these results in hand, GE filled the openings with Brown and Bryant. Walden discussed losing the job opportunity with union officials but elected not to file a grievance after they advised him that he would have little chance of success.
The graded test shows only one answer (on Question 23) marked incorrect. GE and Riggs admit that Riggs made a mistake in grading Question 7, which also should have been marked incorrect. With two missed answers, Bryant still passed.
The next month, in January 2019, GE sought two more toolmakers. It published a job posting with the same description as before. But this time, the test changed. GE and the union agreed to a new test with a written component and a hands-on component, both administered by local community colleges. For the written component, GE and the union settled on a standardized test from the National Occupational Competency Testing Institute. They decided not to use a few parts of the test, considering them inapplicable to GE toolmaker responsibilities. GE and the union agreed that a candidate would need to score a combined 70% to pass.
Once again, Walden applied for the job. Once again, Walden failed the test. He scored an 86.89% on the written component but only a 16% on the skills component, for a combined score of about 51%. A younger candidate, Keith Crowley, scored a combined 69.6%, which the company rounded up to 70%. GE hired Crowley over Walden for the January opening.
After the January test, Walden asked the union to file a grievance. It did so. On the grievance form, Walden argued that despite the failed tests he had the minimum qualifications required for the job and that the CBA therefore required GE to pick him over candidates with less seniority. GE denied the grievance. The union appealed, but GE again denied it. The company wrote Walden to explain its decision and reiterated that testing had been required for years. It offered to let Walden test again for the next job posting.
In June, Walden filed a charge with the Equal Employment Opportunity Commission (EEOC) alleging that GE had discriminated against him based on age. The EEOC issued a right-to-sue notice in August, noting that it was "unable to conclude that the information [it] obtained establishe[d]" any violation of federal law. R. 108-30, EEOC Dismissal, PageID 1220. Later on, Walden also filed a charge with the National Labor Relations Board (NLRB) against the union over its handling of the grievance process. He claimed that the union breached its duty of fair representation by refusing to process his grievances and refusing to communicate with him. After an investigation, the regional NLRB office dismissed his charge. It noted that the union had filed and appealed the grievance as Walden had requested, and it found "insufficient evidence" to support his claims. R. 108-26, NLRB Dismissal, p. 1-2, PageID 1210-11.
Around this time, the union also considered taking Walden's grievance to arbitration. It ultimately decided against doing so, seeing dim prospects of success. So when Walden appealed the NLRB's decision to the agency's General Counsel, he raised the union's decision not to arbitrate. The General Counsel found that the union's actions-including its arbitration decision-did not violate the duty of fair representation, and he denied the appeal "substantially for the [same] reasons" as those identified in the regional office's decision. R. 108-28, NLRB Appellate Decision, p. 2, PageID 1216.
Walden sued GE and the union in November 2019. The operative complaint has two claims-age discrimination under the Age Discrimination in Employment Act (ADEA) and a violation of section 301 of the Labor Management Relations Act (LMRA). See 29 U.S.C. § 621 et seq.; 29 U.S.C. § 185. GE and the union sought summary judgment, which the district court granted. Walden appealed.
II.
We consider one preliminary matter before turning to summary judgment. In the district court, Walden asked to file a surreply to GE's reply brief. The district court denied the request, and Walden appeals that decision here. We review a decision to deny a surreply "under the deferential abuse-of-discretion standard." Mirando v. U.S. Dep't of Treas., 766 F.3d 540, 549 (6th Cir. 2014). Surreplies are "highly disfavored, as they usually are a strategic effort by the nonmoving party to have the last word on a matter." Cousins Smokehouse, LLC v. Louisville Processing & Cold Storage, Inc., 588 F.Supp.3d 753, 763 (W.D. Ky. 2022) (internal quotation marks omitted); see also Cornice & Rose Int'l, LLC v. Four Keys, LLC, 76 F.4th 1116, 1123 (8th Cir. 2023) (same).
GE included with its reply a new declaration from Riggs that, according to Walden, entitled Walden to a surreply. In his proposed surreply, Walden sought to attach a new affidavit responding to the claims in Riggs's declaration. Walden did not contend that Riggs's declaration contradicted his previous testimony, did not move to strike it, nor did he ask to reopen the deposition. At most, the declaration put a new spin on GE's position on the testing options it gave Walden. But in response, Walden's proposed surreply simply reiterated the same argument he'd already made-that GE treated others more favorably. Compare R. 113, Walden Response, p. 14, 19, PageID 1266, 1271, with R. 117-1, Proposed Surreply, p. 1-3, PageID 2115-17. The proposed surreply had a single paragraph of substantive argument and cited Walden's proposed affidavit only once but cited his previously filed response and its exhibits multiple times, leading the district court to conclude that Walden's request was "simply an attempt to have the last word." R. 117-1, Proposed Surreply, p. 1-2, PageID 2115-16; R. 128, Mem. and Order, p. 7, PageID 2155. At any rate, even if Walden had gotten his affidavit into the record, our summary judgment analysis would not change. The surreply wouldn't have added much, if anything. Under these circumstances, we cannot say that the district court abused its discretion.
We affirm the district court's denial of the motion to file a surreply.
III.
On to summary judgment. District courts must grant summary judgment when "there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law." Fed.R.Civ.P. 56(a). We review a grant of summary judgment de novo, and we draw all reasonable inferences for the nonmoving party. M.J. ex rel. S.J. v. Akron City Sch. Dist. Bd. Educ., 1 F.4th 436, 445 (6th Cir. 2021). But the nonmoving party must put forth more than a "scintilla" of evidence to support its claims. Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 252 (1986). "[C]onjecture" and "conclusory accusations" will not suffice. Arendale v. City of Memphis, 519 F.3d 587, 605 (6th Cir. 2008). To survive summary judgment, the nonmoving party "must present significant probative evidence" putting the material facts in doubt. Green Genie, Inc. v. City of Detroit, 63 F.4th 521, 526 (6th Cir. 2023) (internal quotation marks omitted).
IV.
A.
The ADEA prohibits employers from discriminating against people forty or older on the basis of age. 29 U.S.C. §§ 623(a)(1), 631(a). Plaintiffs can prove a violation of the ADEA through direct or (as relevant here) circumstantial evidence. Pelcha v. MW Bancorp, Inc., 988 F.3d 318, 324 (6th Cir. 2021). We evaluate claims based on circumstantial evidence using the burden-shifting analysis from McDonnell Douglas Corp. v. Green, 411 U.S. 792, 802-03 (1973).
McDonnell Douglas first requires the plaintiff to establish a prima facie case of age discrimination. A plaintiff does so by showing "(1) that he is a member of a protected class; (2) that he applied for, and did not receive, a job; (3) that he was qualified for the job; and (4) that a similarly-situated person who was not in the plaintiff's protected class received the job." Anthony v. BTR Auto. Sealing Sys., 339 F.3d 506, 515 (6th Cir. 2003). If the plaintiff succeeds, the burden then shifts to the employer to identify a "legitimate, nondiscriminatory reason" for its decision. Pelcha, 988 F.3d at 325. If the employer does so, the burden shifts back to the plaintiff to prove that the employer's reason "is a mere pretext" for unlawful age discrimination. Id.
There's no dispute that Walden clears prongs one and two of the prima facie test. This case turns on prongs three and four-whether Walden was qualified and whether the younger candidates GE hired were similarly situated.
GE argued that Walden's briefing "relies extensively on unpreserved arguments." Appellee Br. at 22. Despite stressing again and again that Walden "did not initially raise certain arguments," that his briefing "has many arguments [he] never raised," and that he "make[s] new arguments he never raised," GE never actually tells us which of Walden's appellate arguments it refers to. Id. at p. 22-23. GE offers us a single "see, e.g." citation that identifies only one "brand-new contract theory" Walden supposedly raised. Id. at p. 23. This sole example consists more of new authorities than new arguments. There "is no rule prohibiting a party from [citing new authorities]," and new citations to relevant authorities are "as welcome on appeal as in the court below." Reeser v. Henry Ford Hosp., 695 Fed.Appx. 876, 882 (6th Cir. 2017) (internal quotation marks omitted). As to the rest of Walden's supposedly forfeited arguments, GE's failure to point us to specifics forfeits its argument as underdeveloped. See United States v. Watkins, Nos. 23-3091/3467, 2024 WL 3218151, at *4 (6th Cir. June 27, 2024); cf. Murthy v. Missouri, 144 S.Ct. 1972, 1991 n.7 (2024) ("Judges are not like pigs, hunting for truffles buried in the record.") (cleaned up). But this matters little since we find Walden's arguments unpersuasive anyway.
1.
The toolmaker qualifications included either five years of relevant experience or an apprenticeship, as the job posting spelled out, as well as passing a written test. Riggs and multiple union officials testified that GE had required toolmaker candidates to pass a test for years before Walden applied. And the company did so in consultation with the union; indeed, Brown (the union official) described how union members, including himself, wrote the test. Walden presents no evidence contradicting this testimony.
Though Walden had more than five years' experience, he did not pass the December 2018 test. Riggs testified that a candidate needed 85% to pass, and Walden's test shows a score of only 80%. Nothing in the record contradicts that. So Walden cannot establish that he qualified for the job under prong three. See Anthony, 339 F.3d at 516. His prima facie case fails.
For a similar reason, Walden also cannot satisfy prong four. The record evidence shows that the younger and less senior candidates were not "similarly situated" to Walden because they did pass the test. And since they had the minimum years of experience, they met GE's standards. Had Walden passed the test, he would have been similarly situated to Brad Brown and Nathan Bryant, and GE would have had to award him the job under the CBA based on his seniority. But he didn't.
Walden makes several arguments in response. We find none persuasive. First, he asserts that taking the test was not actually a requirement because GE's job posting did not mention it. But the posting says that it's not exhaustive. After the posting lists certain minimum qualifications and a job description, a disclaimer states that any "classifications and definitions are merely for purposes of identification and general description and do not purport to be all inclusive or exhaustive of the actual requirements of any job so classified or defined." R. 113-2, 2018 Job Posting, p. 1, PageID 1313. Walden parses these terms finely, arguing that the nonexhaustive "classifications," "definitions," and "requirements" differ from "qualifications," and so we should not read the disclaimer to apply to the posting's "qualifications." This argument fails because the posting on its face does not use these divisions strictly. For example, it states outside the paragraph labeled "Qualifications" that candidates must also have "satisfactory performance on their present job." Id. That's clearly a minimum qualification. And in any event, we have noted that "employers are not rigidly bound by the language in a job description." Browning v. Dep't of Army, 436 F.3d 692, 696 (6th Cir. 2006). GE was free to implement a testing requirement, multiple witnesses described how it did so, and the company made Walden aware of that when he applied.
Second, Walden contends that our consideration of the unwritten testing requirement violates the contract-law rule against parol evidence. This argument is misplaced. Walden states that "[b]etween the job posting and the CBA, the contract is not ambiguous," so we cannot look to parol evidence. Appellant Br. at 32. Walden's phrasing seems to bootstrap the CBA's contractual nature onto the job posting. But though the CBA was a contract, the job posting was not. The CBA required GE to hire qualified candidates based on seniority, but it did not dictate which qualifications GE could set in the first place. Contract rules do not apply to discerning GE's intentions with the job posting.
Walden probably means to say that the posting was an offer, one that he "accepted" by applying with the most seniority. But the posting wasn't even that. At best, it was an invitation to be considered, or in contract-law terms, an invitation to offer, since nothing would have obliged GE to take any candidates. See Premier Land Dev. Co. v. Bedrock Contracting Inc., No. 22-5598, 2023 WL 1800272, at *1 (6th Cir. Feb. 7, 2023) (per curiam) (noting that a "bid is an offer, not a contract"); Guardian Angel Staffing Agency, Inc. v. Commonwealth, No. 2013000090, 2015 WL 3826343, at *4 (Ky. Ct. App. June 19, 2015) (noting that a request for applications "does not constitute an offer but merely a solicitation of offers, which does not impose any contractual obligations" (internal quotation marks omitted)). But here, the job posting probably wasn't even that. Walden's application was not the offer; GE's extension of a job offer (as the term suggests) would've been. At that point, Walden could have accepted or walked away.
Faced with testimony from GE and the union about its longstanding practice of testing toolmaker candidates, as well as the actual evidence of testing in the record, Walden has presented no evidence putting the test requirement in real dispute. See Scott v. Harris, 550 U.S. 372, 380 (2007) (directing courts at the summary judgment stage to ignore narratives "utterly discredited by the record").
Third, Walden maintains that he actually did pass the test. He opines that his 80% would be a passing score "[o]n most scales" and suggests that GE made up the 85% number to gerrymander him out. Appellant Br. at 33. But Walden's opinion is not what sets company policy. And he offers no evidence that the passing score was anything other than 85%, leaving that number uncontradicted. He points out that GE never put the passing score in writing. But he also cites no authority that would require GE to do so. And GE had enforced an 85% passing score for years, as shown by the graded tests of past toolmaker candidates in the record. Walden details how some of those passing tests display a few "correct" answers that do not correspond perfectly with the answer key, arguing that this puts the "real" passing score in some dispute. Yet he fails to acknowledge that GE graded the tests for substantive accuracy, not a verbatim match. Riggs stated that he used his "professional judgment based on decades of relevant employment at GE Aviation, to determine when answers were substantively correct, even if they did not exactly match the language set out in the answer key." R. 116-5, Riggs Declaration, p. 1, PageID 2092.
And the evidence shows that Riggs's professional judgment and leniency benefited Walden too. For example, for Question 4 the answer key reads "left side," but Walden's answer, which discussed the "Lowest point of wheell [sic]," did not use those words. Compare R. 11317, Answer Key, p. 1, PageID 1712, with R. 108-10, Walden Test 2018, p. 1, PageID 1177. Still, Riggs judged Walden's answer substantively correct and gave him credit. Had Riggs graded strictly, as Walden admonishes that he should have, Walden would have done worse. Walden's accusation that GE manufactured the passing score and should have evaluated some answers differently rests on nothing but speculation and his subjective opinion about grading-not enough to defeat summary judgment. See Smith v. City of Toledo, 13 F.4th 508, 515, 518 (6th Cir. 2021) (affirming summary judgment against a plaintiff who provided no evidence to support claims that "his testing and training conditions were materially different" than his peers' or that his employer "rigged" his multiple test failures).
Walden's final argument fares no better. Claiming that Bryant's test contains handwriting discrepancies, Walden questions whether Bryant completed the test on his own. He argues that the handwriting on parts of three answers (1) is written in a lighter, grey color while the rest of the answers are written in black, and (2) differs from other answers' handwriting. Walden also points out that the final five answers (which GE tested separately) are answered in a lighter color. This, he believes, raises doubts about the test's integrity and thus whether the test was a necessary job qualification rather than just a pretext to mask discrimination.
It's true that a few of Bryant's letters appear more grey while the rest appear black, but the same is true of the other tests in the record. When we look at those tests, the same slight differences in color gradation show up. Keep in mind that we can see only digital copies of the original documents on our electronic docket. To support some alternative, nefarious explanation that would account for all the tests, Walden would have to claim (and provide evidence) that Riggs rigged each one, not just Bryant's. This Walden has not done. That a bit of color was lost in electronic translation does not support an inference of discrimination.
The final five answers on Bryant's test similarly provide no evidentiary basis for questioning the test's integrity. Though the writing appears in grey, that alone would not permit a reasonable jury to infer that Bryant did not complete the test. Perhaps if GE claimed that Bryant wrote only with a black pen, and a few answers appeared in red ink, then a court could conclude from the document's face that a jury could reject GE's version of events. Cf. Moyer v. Gov't Emps. Ins., 114 F.4th 563, 569 (6th Cir. 2024) (finding it an open "factual question" whether a document with multiple redlines and electronic comments could be authenticated). Not so here. GE's position is not that Bryant couldn't have used different writing utensils, say, a pen on the first sitting and a pencil on the second. It's merely that Bryant completed the test. So nothing Walden gives us contradicts GE's position.
Walden's stab at a handwriting analysis likewise fails. Bryant testified that the handwriting appeared to be his after reviewing his test at his deposition. While he admitted that he couldn't confirm it with certainty, Bryant stated that the lighter colored answers "[l]ook[ed] the same" as his handwriting, that it didn't "really look any different," and that it "all look[ed] the same other than being lighter." R. 113-10, Bryant Deposition, p. 24-26, PageID 1665-67. He also stated that it was "possible" he had used two different writing utensils, though he could not recall what exactly he had used given the passage of several years. Id. at p. 24, PageID 1665.
Walden's argument, at bottom, is one of authentication under Federal Rule of Evidence 901. At trial, Walden would have to prove that the document is what he claims it is-the work, in relevant part, of someone other than Bryant. And since GE has put forth Bryant's testimony recognizing the handwriting, Walden has to convince us that he would have something in response. He doesn't. He gives us nothing concrete to work with, such as an expert analysis or another handwriting sample for comparison, that could contradict the sworn testimony of Bryant and Riggs. See Fed.R.Evid. 901(b)(3); e.g., Nelson v. Gowdy, No. 03-10116-BC, 2006 WL 2604679, at *6-7 (E.D. Mich. Sept. 11, 2006) (sending a handwriting issue to trial when the plaintiff pointed to likely similarities between an anonymous letter and the defendant's authenticated handwriting sample); United States v. Mallory, 902 F.3d 584, 593-94 (6th Cir. 2018) (expert using ninety-one authenticated handwriting samples for comparison); see also 31 Charles Alan Wright & Arthur R. Miller, Federal Practice and Procedure § 7108 n.19, Westlaw (database updated June 2024) (noting that "[t]here should be as many [comparators] as possible up to about 50-75," though admitting that "in some circumstances even one . . . will be enough" (internal quotation marks omitted)). Nor has Walden contended that he has any personal familiarity with Bryant's handwriting. See Fed.R.Evid. 901(b)(2); see, e.g., United States v. Harris, 786 F.3d 443 (6th Cir. 2015).
To be sure, Rule 901 provides a nonexhaustive list of authentication methods; experts and comparators aren't the only ways Walden could contest the facts. The Rule still requires something, however, and Walden hasn't come up with any creative alternative. Without an expert, another handwriting sample, or personal familiarity with Bryant's lettering, Walden seems left with nothing but his own amateur opinion on a subject where we usually permit only expert testimony. See Mallory, 902 F.3d at 593 ("[H]andwriting analysis [is] 'technical' or 'specialized' knowledge that, subject to thorough gatekeeping, is a proper area of expertise." (quoting Fed.R.Evid. 702(a) and discussing Kumho Tire Co. v. Carmichael, 526 U.S. 137 (1999)).
So the fact remains: a litigant has failed to put handwriting in dispute when he cannot offer the factfinder solid evidentiary ground on which to accept his claims. See 31 Wright & Miller § 7108 (calling it an "inadequate basis for authentication where [the evidence] provides insufficient points of comparison to permit a reasonable inference as to authenticity . . . or where the number of exemplars is insufficient to make a meaningful comparison"). Because Walden has not provided a future factfinder anything to set Bryant's test and sworn testimony against, he has failed to carry his evidentiary burden.
We are also skeptical that the handwriting is "material." Fed.R.Civ.P. 56(a). Even if the factfinder could reasonably find that a few answers were written differently (which we doubt), Walden would then need the factfinder to infer from there that Bryant didn't answer all the questions, then infer that Bryant wouldn't have passed under the "substantive accuracy" grading anyway, then infer that Riggs (or someone else) filled in the correct answers intending to deceive anyone who might second-guess the results years later in litigation, then infer that Riggs did so hoping to harm Walden (not merely to help Bryant), and finally infer that Riggs was motivated by ageist animus (not, hypothetically, because they merely didn't get along). This we cannot permit. See Kubala v. Smith, 984 F.3d 1132, 1141 (6th Cir. 2021) (rejecting plaintiff's theory as requiring "too many inferences"); cf. In re E.I. Du Pont de Nemours & Co. C-8 Pers. Inj. Litig., 54 F.4th 912, 935 (6th Cir. 2022) (withholding an issue from a jury when a party's theory would have required the jury to "draw inference upon inference upon inference" (internal quotation marks omitted)).
Summary judgment does not require a court to draw every possible inference, no matter how stretched, in the nonmovant's favor. It requires only "reasonable inferences." Kinlin v. Kline, 749 F.3d 573, 576 (6th Cir. 2014). At bottom, Walden's argument that someone monkeyed with Bryant's test rests not on "significant probative evidence," Green Genie, 63 F.4th at 526 (internal quotation marks omitted), but on "[c]onjecture" and "speculation," Arendale, 519 F.3d at 605 (internal quotation marks omitted); Milczak v. General Motors, LLC, 102 F.4th 772, 785 (6th Cir. 2024), which do not create a triable factual matter.
Walden missed out on a passing score-and thus the job-by only one or two questions. The Court sympathizes with such a close call. But employers commonly enforce mandatory job tests, especially in high-skill or dangerous occupations. See, e.g., Smith, 13 F.4th 508 (6th Cir. 2021) (affirming summary judgment against a firefighter candidate who failed a chainsaw test nine times and then sued for racial discrimination); Bonds v. Varian Med. Sys., 672 Fed.Appx. 562 (6th Cir. 2016) (affirming summary judgment against a job candidate who failed a radiation machine operating test and then sued for age discrimination). "If a company is allowed to require employees to pass an exam, . . . it must set a passing grade at some point." Bonds, 672 Fed.Appx. at 564. And "[o]nce a passing grade is set, it follows that some individuals will just meet it and some will fall just short of it." Id. It's not our job to second-guess those cutoffs.
2.
The next month, GE looked to hire again. The company asked for the same qualifications as before. So once again, Walden wouldn't have been eligible for the job unless he can show that he was qualified.
He can't. He failed the new test, scoring a 16% on the skills half with a total score of about 51%. Walden does not contest his grade this time around. Because he wasn't qualified, he can't clear prong three of the prima facie case. On this record, nothing would permit a reasonable jury to infer that GE passed Walden over because of his age, and not cold, hard qualifications.
Rather than show why he qualified for the job, Walden attacks the qualifications of Crowley, who did get the job. First, he questions whether Crowley had the minimum "5 years of Tool and Die job experience" that GE wanted. R. 113-13, January 2019 Job Posting, PageID 1702. But Riggs stated that he found Crowley's "experience and background, including his working supervisor role at Atlantis Plastics" enough to meet GE's standards. R. 116-5, Riggs Declaration, p. 2, PageID 2093. And Crowley's resume backs this up. For nearly two decades before GE hired him, Crowley worked as a "Tool Shop Manager" and "Tool Room Supervisor" with different companies. R. 113-33, Crowley Application, p. 3, PageID 1884. GE, not Walden, got to determine whether this experience counted. See Flowers v. WestRock Servs., Inc., 979 F.3d 1127, 1131 (6th Cir. 2020) ("[A]s the one who creates the position in question, the employer largely enjoys the right to decide the qualifications it prefers in one who holds the position and, it follows, whether an applicant lacks the necessary knowledge or experience.").
Second, Walden claims that GE "change[d] the scoring of the tests two or three times" to ensure Crowley would pass. Appellant Br. at 36. That gets things backwards. As the record evidence shows, GE and the union agreed to phase in new tests, to have the community college score it, not to count certain sections as unnecessary, and to average the written and hands-on components all before the testing took place. They'd made these decisions by the time Crowley-and Walden-tested. That Walden may not have known the precise details ahead of time does not provide an evidentiary basis from which one can infer age discrimination.
Walden also insists that the rounding of Crowley's 69.6% up to 70% shows that Riggs played favorites. Putting aside Riggs's statement in the record that GE "consistently rounds to the nearest whole number," rounding up is a familiar and benign practice. R. 116-5, Riggs Declaration, p. 2-3, PageID 2093-94. Indeed, in describing his own score on the same test, Walden himself rounds up, telling us that he scored an "87%" on the written test while the document itself says "86.89%." Compare Appellant Br. at 18, with R. 113-20, Walden January 2019 Written Test Score, p. 3, PageID 1729. Nothing here puts the reason for GE's hiring decision into genuine doubt.
Though it bears little on our decision, GE represented to the district court that Walden failed the test a third time in February 2023, scoring a 30% on the hand-on skills component. Reply Br. at 1 n.2.
The district court properly granted summary judgment against Walden on the ADEA claim.
B.
We turn now to Walden's hybrid section 301 claim. Section 301 of the LMRA gives federal courts jurisdiction to hear suits for violations of collective bargaining agreements. 29 U.S.C. § 185; Swanigan v. FCA U.S. LLC, 938 F.3d 779, 783-84 (6th Cir. 2019). An employee bringing a hybrid section 301 claim must prove "(1) that the employer breached the collective bargaining agreement and (2) that the union breached its duty of fair representation." Swanigan, 938 F.3d at 784 (internal quotation marks omitted). If the employee cannot prove both elements, he cannot succeed against any defendant. Id.
The district court held that issue preclusion bars Walden's claim because the NLRB already adjudicated Walden's unfair representation charge and dismissed it. We agree. Because Walden's unfair representation argument is precluded, his section 301 claim cannot succeed.
Issue preclusion can apply to determinations made by agencies like the NLRB. B &B Hardware Inc. v. Hargis Indus., Inc., 575 U.S. 138, 147-49 (2015); Elec. Workers Loc. 58 Pension Tr. Fund v. Gary's Elec. Serv. Co., 227 F.3d 646, 658-59 (6th Cir. 2000). Our "classic test" for issue preclusion in NLRB cases asks (1) whether the issue is identical to that decided by the NLRB, (2) whether the issue was necessary to the NLRB's judgment, and (3) whether the party against whom preclusion would operate had a full and fair opportunity to litigate the issue. Elec. Workers, 227 F.3d at 659. Litigants who had "every incentive to litigate [an issue] fully and vigorously" had a full and fair opportunity. Parklane Hosiery Co., v. Shore, 439 U.S. 322, 332 (1979).
All three prongs are satisfied here. First, the NLRB decided the same issue that Walden now seeks to relitigate-whether the union unfairly represented him. Second, the issue was "necessary" to the NLRB's ruling against him. In fact, it constituted the whole basis of the ruling against him. And third, Walden had a full and fair opportunity to litigate his case because he had every incentive to do so. Nothing in the record suggests that he had any reason for hesitation or that anything obstructed his path. So issue preclusion applies.
To stave off preclusion, Walden argues that the NLRB's decision considered narrower circumstances than he presented in federal court and thus did not address the same issue. He cites the union's decision not to take his case to arbitration and claims that the NLRB did not consider all events stretching back to December 2018. As to the former, the agency's appellate decision makes clear that the union's decision not to arbitrate was, in fact, considered. And as to the latter, the decision discusses how the regional office "considered all the evidence" from all relevant timeframes, including "going back to December 2018." R. 108-28, NLRB Appellate Decision, p. 2, PageID 1216. Walden fails to show that the NLRB decided something different from what he asks us to decide today.
Walden also asserts that because he appeared before the NLRB pro se, he did not have a chance to fully and fairly litigate the case. We disagree. As the district court pointed out, courts routinely apply issue preclusion against pro se parties. See, e.g., Otworth v. Fifth Third Bank, No. 20-1286, 2020 WL 9211025, at *3-4 (6th Cir. Oct. 27, 2020); McCord v. Kentucky Educ. Ass'n, No. 17-286, 2017 WL 5162817, at *3 (E.D. Ky. Nov. 7, 2017), aff'd sub nom., McCord v. Brooks, McComb and Fields, LLP, No. 17-6459, 2018 WL 8514480 (6th Cir. Dec. 12, 2018); see also Lansing v. Wells Fargo Bank, 894 F.3d 967, 972 (8th Cir. 2018) ("A plaintiff's pro se status . . . does not deprive the plaintiff of a full and fair opportunity to litigate. Parties proceeding pro se are not exempt from the doctrine of claim preclusion.").
All in all, Walden cannot relitigate his unfair representation claim against the union. As a result, he cannot establish a key element of his section 301 claim and it fails as a matter of law. The district court properly granted summary judgment against him.
V.
For these reasons, we affirm the district court's judgment.