FILED
United States Court of Appeals PUBLISH Tenth Circuit
UNITED STATES COURT OF APPEALS January 7, 2025 Christopher M. Wolpert
FOR THE TENTH CIRCUIT Clerk of Court
___________________________________________
DONETTA RAYMOND; FREDERICK HESTON; JILUN SHA; RANDY WILLIAMS; WILLIAM SCOTT DENNY; DEBRA HATCHER; BRIAN MARKS; RUSSELL BALLARD; GREGORY BUCCHIN; BRUCE ENSOR; FORREST FARIS; CHERYL RENEE GARDNER; CLARK T. HARBAUGH; CRAIG HOOBLER; BRIAN SCOTT JACKSON; WILLIAM KOCH; FRED LONGAN; DAVID B. MILLER; KENNETH L. POOLE, JR.; BAHRAM RAHBAR; RUSSELL SPRAGUE; CRAIG TOLSON; ROBERT TROILO; CURTIS J. VINES, on behalf of themselves and all others similarly situated,
Plaintiffs - Appellants,
v. No. 23-3126
SPIRIT AEROSYSTEMS HOLDINGS, INC.; SPIRIT AEROSYSTEMS, INC.,
Defendants - Appellees.
--------------------------------
EQUAL EMPLOYMENT OPPORTUNITY COMMISSION,
Amicus Curiae.
2
___________________________________________
APPEAL FROM THE UNITED STATES DISTRICT COURT FOR THE DISTRICT OF KANSAS (D.C. NO. 6:16-CV-01282-JWB)
___________________________________________
James Kaster of Nichols Kaster LLP, Minneapolis, Minnesota (Diane S. King and Marianna McLean of King Employment Law, Denver, Colorado, Robert Schug of Nichols Kaster LLP, Minneapolis, Minnesota, Daniel Kohrman and Lauren Naylor of AARP Foundation, Washington, DC, and Randall K. Rathbun of Depew Gillen Rathbun & McInteer, Wichita, Kansas with him on the briefs) for Plaintiffs-Appellants.
Steven W. Moore of Fox Rothschild LLP (Stacy D. Mueller and Renee J. Sheyko of Fox Rothschild LLP, Denver, Colorado, and James M. Armstrong, Jeff P. DeGraffenreid, Teresa L. Shulda, and Charles E. McClellan of Foulston Siefkin LLP, Wichita, Kansas, with him on the brief) for Defendants-Appellees.
Karla Gilbride, General Counsel, Jennifer S. Goldstein, Associate General Counsel, Anne Noel Occhialino, Assistant General Counsel, and Georgina C. Yeomans, Attorney, Equal Employment Opportunity Commission, Washington, D.C., filed an amicus curiae brief in support of Appellants.
___________________________________________
Before HARTZ, KELLY, and BACHARACH, Circuit Judges.
___________________________________________
BACHARACH, Circuit Judge.
___________________________________________
Spirit AeroSystems, Inc. adopted a reduction-in-force that led to 271 firings. Many of the fired employees brought suit, including a collective action against Spirit for age discrimination. To prevail on the collective action, the former employees needed to prove a pattern or practice of age discrimination.
3
Embracing this burden at the summary-judgment stage, the former employees relied on documentary evidence and testimony. In response, Spirit argued that it was trying to shed underperformers regardless of their ages. To resolve this disagreement, we must consider whether the evidence could reasonably support a finding of an unlawful pattern or practice of age discrimination. We answer no.
Background
I. Spirit tries to improve performance and cut costs.
By early 2011, Spirit was experiencing financial problems. So in late 2011, executives changed the way that the company assessed employee performance. This change resulted in lower performance ratings, but no meaningful reduction in the workforce.
Without a meaningful dent in labor costs, Spirit continued to experience financial problems. Spirit executives thus discussed other ways to cut costs, like reviewing management overhead, increasing firings for poor performance, changing the requirements for hiring, optimizing shifts, and offering a voluntary severance package to long-time employees.
Through these discussions, Spirit decided in 2012 to restructure the system for evaluating employee performance. Under the new system, managers would rate employees so that 15% would exceed expectations, 70% would meet expectations, and
4
15% would not match at least some of the expectations.
Spirit executives conducted training sessions to educate managers on the new system. During these training sessions, managers met with employees in the Human Resources Department to discuss the performance of various employees. The meetings spurred resistance. For example, one manager complained of pressure from Spirit executives to downgrade older or less healthy workers even when they were performing well. Another employee complained that Spirit was unfairly putting some employees on performance-improvement plans while exempting new employees. II. Spirit adopts a new health insurance plan to reduce costs.
Spirit executives not only tightened the system to evaluate performance, but also discussed funding of healthcare expenses for employees. In these discussions, a contractor told Spirit that its healthcare expenses would increase, largely because of the “aging factor.” Appellants’ App’x vol. 9, at 126. So Spirit decided to self-fund its health insurance.
III. Spirit opts for a reduction-in-force, and the ensuing litigation results in summary judgment.
After tightening the system for evaluating employee performance, Spirit planned a reduction-in-force that would trim the Wichita plant’s workforce by 10%. To carry out this plan, Spirit had to comply with a collective bargaining agreement. Under this agreement, Spirit could include unionized employees in a reduction-in-force only after conducting
5
a “retention exercise.” Appellants’ App’x vol. 5, at 241–44. So Spirit conducted a retention exercise in 2013.
In this exercise, Spirit considered an employee’s 2012 performance rating, 2013 performance, versatility, and criticality. 1 Based on these factors, managers would put 70% of the employees in the top category (A), 20% in the second category (B), and 10% in the bottom category (C). With few exceptions, the first employees to go in a reduction-in-force would come from the bottom category (C).
Spirit also softened the role of tenure in an employee’s performance rating. Until then, tenured employees would ordinarily enjoy an advantage. For example, Spirit’s agreement with the union generally allowed 20-year employees to advance one category in a retention exercise. So an employee with a C rating and 20 years’ experience would automatically advance to a B rating. But the union agreement authorized Spirit to exempt employees,
1 Spirit defined versatility based on an employee’s “critical thinking skills,” “flexibility/resiliency/adaptability/attitude,” and other related factors. The company assessed criticality based on whether the employee had the skills necessary to “best run the critical business functions” or to meet “future business requirements.” Appellees’ Supp. App’x vol. 11, at 3077.
6
removing this advantage. Spirit exercised this authority, exempting every tenured employee from this advantage in the retention exercise. Spirit also softened the role of tenure by exempting new employees in the retention exercise.
These actions spurred many former employees to assert a collective action under the Age Discrimination in Employment Act. But the district court granted summary judgment to Spirit based on the failure to prove an unlawful pattern or practice.
Discussion
I. We consider whether a reasonable factfinder could find a “pattern or practice” of age discrimination.
For the grant of partial summary judgment, we conduct de novo review, considering the evidence in the light most favorable to the former employees. Blehm v. Jacobs, 702 F.3d 1193, 1199 (10th Cir. 2012). 2 In considering the evidence in this light, we determine whether a “genuine dispute exists on any material fact” and whether Spirit is entitled to judgment as a matter of law.
2 The former employees argue that the district court appeared to
credit certain facts when certifying a collective action and
discredit these facts when ruling on the summary-judgment motion.
We need not address the alleged inconsistency because we conduct de novo review on the grant of summary judgment.
7
Fed. R. Civ. P. 56(a).
The viability of this collective claim turns on whether the reduction-
in-force entailed a pattern or practice of age discrimination. Thiessen v. Gen. Elec. Cap. Corp., 267 F.3d 1095, 1105–06 (10th Cir. 2001). For this claim, the former employees must show that “unlawful discrimination has been a regular procedure or policy followed by an employer.” Int’l Bhd. of Teamsters v. United States, 431 U.S. 324, 360 (1977). A regular procedure or policy exists if “discrimination was the company’s standard operating procedure . . . rather than the unusual practice.” Id. at 336.
II. The former employees have not shown that the reduction-in-force was motivated by ageism.
The former employees complain that the reduction-in-force was ageist, relying on lay testimony, documentary evidence, and expert opinions.
A. Lay testimony and documentary evidence The former employees characterize the reduction-in-force as age discrimination, relying on six categories of lay testimony and documentary evidence:
1. Slides addressing tenure and health-care costs 2. Planning documents 3. Policies 4. Statements by Spirit executives
8
5. Statements by Spirit employees 6. Training materials Some of this evidence might support a finding of age discrimination by some of Spirit’s managers or executives. But the combination of evidence doesn’t create a reasonable inference of a standard operating procedure to shed older employees.
1. Slides addressing tenure and health-care costs In 2012, Spirit executives periodically discussed ways to reduce costs, using slides addressing healthcare expenses and salaries. These slides arguably related to age, for older employees might generally incur greater healthcare expenses and earn higher salaries. But that relationship did not necessarily reflect age discrimination: an employer can use factors often associated with age as long as the employer doesn’t act on the basis of age. Hazen Paper Co. v. Biggins, 507 U.S. 604, 612–13 (1993).
The former employees focus on two slides. The first slide addressed healthcare costs, stating that Spirit could expect a 4% increase in healthcare costs for every digit increase in the workforce’s average age. The second slide
noted the existence of a “retirement risk” for employees over 62 and
recommended a voluntary retirement program to “[r]eplace long service/high cost employees with new hires.”
Appellants’ App’x vol. 8, at 39.
9
From the first slide, a factfinder could reasonably infer recognition of a connection between age and healthcare expense. But where is the evidence suggesting that Spirit acted because of this connection? The former employees answer with the second slide, which shows that Spirit was considering a voluntary retirement program to replace high-cost employees with new hires.
From the second slide, a factfinder might reasonably infer concern over upcoming retirements and the cost of employees with long tenures. But the Supreme Court has distinguished between tenure and age in Hazen Paper Co. v. Biggins, 507 U.S. 604 (1993). There the Court held that termination of an employee to prevent vesting of a tenure-based pension was not sufficient in itself to “constitute discriminatory treatment on the basis of age.” Id. at 612–13. Under Hazen, the former employees needed to show that Spirit had used tenure as a proxy for age. And there’s no evidence of that.
Moreover, the second slide was addressing a voluntary retirement program; and federal law generally permits such a program even when eligibility turns on age. 29 U.S.C. § 623(f)(2)(B)(ii); 3 see Fagan v. New
3 This law allows the employer to provide a voluntary retirement plan only if it’s consistent with the federal age-discrimination law. 29 U.S.C. § 623(f)(2)(B)(ii). But the federal law doesn’t prohibit an employer from offering compensation to encourage older workers to retire. See, e.g., Cipriano v. Bd. of Educ., 785 F.2d 51, 56 (2d Cir. 1986) (“There is nothing
10
York State Elec. & Gas Corp., 186 F.3d 127, 133 (2d Cir. 1999) (concluding that the implementation of a voluntary early retirement plan, which conforms to the purposes of the federal age-discrimination law, “cannot serve as evidence of unlawful age discrimination”). And the former employees haven’t presented any reason to infer ageism from Spirit’s consideration of a voluntary retirement program.
Granted, the result might be different if the former employees had coupled the slides with other probative evidence showing efforts to reduce healthcare costs by firing older workers. An example took place in Tramp v. Associated Underwriters, Inc., 768 F.3d 793 (8th Cir. 2014). There an employer asked an insurer to reduce the premium because the company had already done its “best” by firing the “oldest and sickest employees.” Id. at 802. If we were to follow Tramp, we might infer a discriminatory policy from evidence that Spirit had tried to reduce healthcare costs by firing older workers. But no such evidence exists.
2. Planning documents In 2012, Spirit also addressed the demographics of the workforce.
From this consideration of demographics, the former employees point to two documents:
inconsistent with the [federal age-discrimination law] in offering older employees compensation for leaving the workforce.”), abrogated on other grounds, Pub. Emps. Ret. Sys. of Ohio v. Betts, 492 U.S. 158, 171 (1989).
11
1. an email showing the average age of the workforce in 2011 and the anticipated average in 2017
2. a slide showing the evolving retirement risk, length of service, and average age of the workforce
These documents reflect only the aggregate ages of employees, which Spirit says that it used to project potential retirements. Spirit’s use of these planning documents is “innocuous,” particularly without evidence that managers were told about these documents. See Apsley v. Boeing Co., 691 F.3d 1184, 1205 (10th Cir. 2012). Absent that evidence, Spirit’s collection of demographic information wouldn’t suggest an ageist protocol in the way that managers evaluated performance.
3. Spirit’s Policies When conducting the retention exercise, Spirit had to decide which employees to fire. For this exercise, Spirit exempted new employees and softened the protections for senior employees.
The former employees regard the exemptions for new employees as evidence of age discrimination. Spirit disagrees, explaining that it protected new employees in order to preserve recruitment sources.
A factfinder could infer an ageist policy or practice only if the exemptions hadn’t resulted from an age-neutral justification. Fallis v. Kerr-McGee Corp., 944 F.2d 743, 745 (10th Cir. 1991). For example, we addressed a similar issue in Pippin v. Burlington Resource Oil & Gas Co., 440 F.3d 1186 (10th Cir. 2005). There too the plaintiff alleged age
12
discrimination in a reduction-in-force. Id. at 1201. The plaintiff claimed that the employer had fired a disproportionate number of older workers. Id. The employer explained that it had tried to hire several new employees, arguing that protection of those employees was necessary to preserve credibility with schools used for recruitment. Id.
We concluded that this purpose had prevented a reasonable inference of age-discrimination. Id.; see also Fallis, 944 F.2d at 745 (concluding that a retention exercise wasn’t discriminatory when an employer exempted first-year employees from a reduction-in-force because the company lacked a reliable basis to assess their performance). Under Pippin, the exemption of new hires wasn’t ageist if Spirit had been trying to preserve recruitment sources. And the former employees present no reason to question Spirit’s explanation.
The former employees also attack the mechanics of Spirit’s retention exercise. Under a contract with the union, Spirit needed to rate employees and fire those employees with the lowest ratings. Spirit did so, rating employees as A, B, and C. See Background–Part III, above. Employees with a C rating would generally go first in the reduction-in-force. See id.
Despite the risk of a C rating, the union contract provided a measure that could benefit employees who had stayed with Spirit for at least 20 years. See id. Under the contract, Spirit could elevate these employees to a higher category. See id. For example, senior employees with a C rating
13
could generally rise to a B rating, decreasing the risk that they would lose their jobs in a reduction-in-force.
But the contract contained an exception: Spirit could prevent designated employees with 20+ years from raising their ratings. Spirit invoked this exception, designating all C rated employees with at least 20 or more years of experience. See id. In this way, Spirit eliminated the protection ordinarily afforded to senior employees. The former employees criticize this decision, arguing that Spirit removed the protection for senior employees in order to fire more older workers.
In addressing this argument, we must consider two aspects of the former employees’ argument:
1. Can we consider tenure of 20 or more years as a proxy for age?
2. Does removal of this protection constitute age discrimination or an effort to focus solely on performance?
We’ve already answered the first question, noting that the Supreme Court has distinguished between tenure and age. See Discussion–Part II(A)(1), above. Given this distinction, an employer doesn’t necessarily discriminate based on age when focusing on tenure. See id.
Nor does an employer necessarily commit age discrimination when removing benefits to senior employees. Jones v. Unisys Corp., 54 F.3d 624, 630 n.6, 632 (10th Cir. 1995). For example, an employee can’t show discrimination based solely on an employer’s refusal to give a preference
14
to more senior employees in a reduction-in-force. Id. After all, “removing a feature that gave extra benefits to the old differs from discriminating against them. Replacing a plan that discriminates against the young with one that is age-neutral does not discriminate against the old.” Tomlinson v. El Paso Corp., 653 F.3d 1281, 1289 (10th Cir. 2011) (quoting Cooper v. IBM Pers. Pension Plan, 457 F.3d 636, 639 (7th Cir. 2006)).
We thus conclude that Spirit’s policies can’t reasonably suggest an ageist operating procedure.
4. Statements by Spirit executives The former employees also point to statements by three individuals (David Walker, Sam Marnick, and Cassie Caster), who were then serving as Spirit executives.
a. David Walker In 2012, Mr. David Walker served as Spirit’s Chief Technology Officer. In that capacity, he attended some of the meetings for Spirit managers. At one meeting, Mr. Walker allegedly said that young people were “key” and represented “the future of the company.” Appellants’ App’x vol. 7, at 74–75. The former employees point to these statements as evidence of favoritism toward younger workers.
But there’s no evidence of Mr. Walker’s participation in the decision to tighten performance standards in 2012 or to influence the ratings in the retention exercise.
15
We addressed similar statements in Cone v. Longmont United Hospital Ass’n, 14 F.3d 526 (10th Cir. 1994). There a chief executive officer said that the organization needed “some new young blood” because “long-term employees” provided “a diminishing return.” Id. at 531. We concluded that these statements hadn’t provided evidence of discrimination for two reasons:
1. We saw no reason to believe that the chief executive officer had played a role in the decision to fire the plaintiff.
2. We lacked any information about the context for the chief executive officer ’s statements.
Id.
Both rationales are equally applicable here. We lack any
evidence that Mr. Walker had played a role in tightening the performance standards or changing the employee ratings or
information about the context for Mr. Walker ’s reference to youthful workers as the key or future of Spirit.
We thus can’t rely on Mr. Walker’s purported statements as evidence of an ageist operating procedure.
b. Samantha Marnick The former employees also point to an email sent two months after the reduction-in-force. The author, Chief Administration Officer Samantha Marnick, addressed concerns that some managers might prevent workers with long tenures from taking early retirement. This possibility worried
16
Ms. Marnick because she viewed these workers as “blockers” in the “development” of employees with shorter tenures. The former employees argue that the district court erred by discounting this evidence that Spirit had wanted to fire older “blockers” of younger talent. We disagree based on the content and context of the email.
The email’s content didn’t suggest ageism. In fact, the email constituted only part of Ms. Marnick’s response to a question about the advantages of a voluntary retirement program. Ms. Marnick explained the advantages, pointing out that 345 employees had already applied. She added, however, that she faced a “challenge” in getting leadership to accept the applications. Appellants’ App’x vol. 8, at 8. The email doesn’t refer to age.
The timing of the email also prevented a reasonable inference of ageism in the reduction-in-force. Ms. Marnick made this statement after the reduction-in-force, and there’s no evidence tying these remarks to the firings. Without such evidence, a factfinder couldn’t reasonably infer ageism from Ms. Marnick’s statement. See Cone v. Longmont United Hosp. Ass’n, 14 F.3d 526, 531 (10th Cir. 1994) (stating that the plaintiff in an age-discrimination action must show “a nexus” between the discriminatory statement and the firing).
17
c. Cassie Caster The former employees also point to a statement by Ms. Cassie Caster, Vice-President of Human Resources. In June 2013, she collected data to create a presentation on cost reduction. Through this data, Ms. Caster learned that 76% of Spirit’s salaried workers in the United States were at least 40 years old. When Ms. Caster learned of this percentage, she
said “OMG” and asked if she had read the statistic correctly and
used this statistic when reporting on the potential profitability of a voluntary retirement program.
Appellants’ App’x vol. 8, at 23. The former employees point to Ms. Caster’s statements as evidence of age discrimination. For two reasons, however, Ms. Caster’s statements don’t suggest an ageist operating procedure.
First, the statements reflect only the aggregate percentage of older workers (rather than information bearing on whom to fire), and there is no indication that ranking managers had access to, nor acted on, the aggregate data. See Apsley v. Boeing Co., 691 F.3d 1184, 1205 (10th Cir. 2012).
Second, Ms. Caster made the statements in connection with a voluntary retirement program, which would ordinarily be lawful even if it had been designed to encourage older workers to retire. 29 U.S.C. § 623(f)(2)(B)(ii); see Fagan v. New York State Elec. & Gas Corp., 186 F.3d 127, 133 (2d Cir. 1999) (concluding that the implementation of a
18
voluntary early retirement plan, which conforms to the purposes of the federal age-discrimination law, “cannot serve as evidence of unlawful age discrimination”); see also Discussion–Part II(A)(1), above (discussing similar evidence in the context of an employer’s voluntary retirement program).
5. Statements by Spirit employees Hundreds of managers participated in the 2012 ratings of performance. The former employees focus on the process for these ratings, pointing to statements by Deborah Miller, Jeremy Fuller, and Tracy Gillespie.
a. Deborah Miller Ms. Miller said that
she had felt pressure from the Human Resources Department to give poor ratings to older and less healthy workers and
Spirit had lowered the ratings that she had given to some of the employees.
These statements could suggest discrimination against some employees. But how could these statements suggest a standard operating procedure of ageism?
Ms. Miller acknowledged that she didn’t know whether the pressure had come from upper-level managers or the Human Resources Department. And Ms. Miller didn’t say whether the pressure had extended to the hundreds of other managers participating in the evaluations. Without such
19
evidence, a factfinder might infer individual instances of age discrimination, but not a standard operating procedure to oust older workers through the change in performance standards in 2012 or the implementation of a retention exercise.
Ms. Miller also testified that colleagues had changed her assessments. The former employees rely on these changes, arguing that Spirit adjusted ratings without the managers’ consent. But Ms. Miller said that she had shared her evaluations with another manager and that manager had changed the ratings.
Without evidence that Spirit executives (rather than individual managers) had changed the ratings, a factfinder could not reasonably infer a policy of changing the assessments for older workers.
b. Jeremy Fuller Mr. Jeremy Fuller was another manager who participated in evaluating employees and implementing the retention exercise. Mr. Fuller stated that when Spirit implemented the retention exercise, someone had “probably” told him not to give C ratings to new employees because they were desirable when the workforce has aged. The former employees point to Mr. Fuller’s statements as evidence of favoritism toward younger employees.
20
But the former employees lack any evidence of Mr. Fuller’s involvement in the exemption of new hires. As noted above, an employer can exempt new hires in a retention exercise if the reasons are age-neutral. See Background–Part III, above (discussing the exemption of new hires in Spirit’s retention exercise). And without evidence of Mr. Fuller’s participation in policymaking, his statements wouldn’t suggest a standard operating procedure of ageism in the retention exercise.
c. Statements by Tracy Gillespie The former employees also rely on statements by Ms. Tracy Gillespie. In 2012, Ms. Gillespie worked in Spirit’s Human Resources Department. Ms. Gillespie testified that she had disagreed with the exemption for new hires. When Ms. Gillespie expressed that disagreement, someone explained that Spirit had spent a lot on recruitment and new employees had no chance to prove themselves.
The former employees argue that these statements reflect ageism.
But the explanation to Ms. Gillespie reflects two age-neutral reasons for exempting new hires: (1) the recruitment cost and (2) an inadequate period for assessment. In fact, we’ve upheld an exemption for first-year employees when the employer reasons that the employees haven’t had enough time to make a reliable assessment. Fallis v. Kerr-McGee Corp., 944 F.2d 743, 745 (10th Cir. 1991). Given Ms. Gillespie’s age-neutral
21
explanation and the former employees’ lack of contrary evidence, a factfinder couldn’t reasonably characterize the exemption of new employees as an ageist operating procedure.
6. Training Materials The former employees also rely on materials that Spirit used to train managers on how to rate employee performance. In one exercise, Spirit asked managers to rank hypothetical employees. Two of these employees were named Gus and John. Gus had been with Spirit longer; but he was hostile, resisted change, and did work that was subpar and late. John was new, but he was eager to learn and generated high-quality work. See id. In the training sessions, Spirit encouraged managers to rank John higher than Gus despite the difference in tenure. See id.
The former employees argue that this guidance relied on stereotypes of older workers as lazy, unfriendly, and inflexible. But the guidance referred to tenure, not age; and the Supreme Court has distinguished between tenure and age. Hazen Paper Co. v. Biggins, 507 U.S. 604, 612–13 (1993); see Discussion–Part II(A)(1), above. So this example of Gus and John does not support an inference of an ageist operating procedure.
* * *
The former employees insist that they presented evidence of ageism.
But the issue is whether Spirit had a standard operating procedure to get
22
rid of older employees. Individually or collectively, the former employees’ evidence doesn’t support an inference of an ageist operating procedure.
One manager complained of ageist pressure, but she didn’t know who had created the pressure. And some employees disagreed with the exemption of new employees. But there’s no evidence linking this exemption to a bias against older workers.
Spirit also removed a benefit ordinarily afforded to senior workers.
But the union contract entitled Spirit to remove that benefit, and the move put all employees on an equal playing field.
We thus conclude that the lay testimony and documentary evidence didn’t suggest an ageist operating procedure in the reduction-in-force.
B. Expert opinions The former employees also rely on the expert opinions of Dr. Stanley Kaufman and Dr. Kevin Cahill.
1. Dr. Kaufman’s statistical comparisons Dr. Kaufman presented opinions about statistical comparisons among Spirit’s workers.
a. Comparison of older workers with low ratings in 2010 and 2013
The first set of opinions involved the retention exercises that Spirit had conducted in 2010 and 2013. Dr. Kaufman examined these exercises and compared the results, finding that older workers had been more likely
23
to get a poor rating (C) in the 2013 retention exercise than the one in 2010. The former employees use this comparison to argue that Spirit decided between 2010 and 2013 to target older workers.
But Dr. Kaufman wasn’t comparing apples and apples. In the 2010 retention exercise, Spirit generally upgraded employees with 20+ years of experience from a C to a B. In the 2013 retention exercise, Spirit removed that advantage for all employees with 20+ years of experience. See Background–Part III, above. We’ve regarded the removal of similar advantages as age-neutral. Jones v. Unisys Corp., 54 F.3d 624, 630 n.6, 632 (10th Cir. 1995); see Discussion–Part II(A)(3), above. Given Dr. Kaufman’s failure to consider this age-neutral explanation, we can’t rely on his statistical comparison to infer an ageist operating procedure from the change between 2010 and 2013. See Rea v. Martin Marietta Corp., 29 F.3d 1450, 1456 (10th Cir. 1994) (stating that for an inference of discrimination, the statistician must eliminate nondiscriminatory reasons for the disparity).
b. Comparison of older workers’ performance scores in 2010 and 2012
Dr. Kaufman also opined that older workers had been more likely to see their assessments drop between 2010 and 2012. Dr. Kaufman blamed this drop on ageist changes in the way Spirit evaluated performance in 2012. But Spirit also made changes in 2011 in the evaluation of
24
performance. These changes could provide an age-neutral explanation for the disparity in scores. Given Dr. Kaufman’s failure to account for this possibility, his comparison did not support an ageist operating procedure. See id.
c. Disproportionate firings of older workers Dr. Kaufman also opined that in the reduction-in-force, older workers had been far more likely to lose their jobs than younger workers. To reach this opinion, Dr. Kaufman examined the firings, controlling for age, discipline, performance rating, and job code.
During the retention exercise, managers had relied on four factors when rating an employee:
1. 2012 performance, 2. 2013 performance-to-date, 3. versatility, and 4. criticality.
See Background–Part III, above. The former employees don’t question Dr. Kaufman’s failure to control for 2013 performance. But they argue that Dr. Kaufman did control for versatility and criticality by assuming their connection to job codes.
In assuming this connection, Dr. Kaufman lacked evidentiary support. Spirit pointed out that it had compared the criticality and versatility of some employees within the same job codes. And the
25
criticality and versatility of employees could differ even within the same job code. So Spirit argued that Dr. Kaufman had failed to account for differences among these employees. Without any response from the former employees, the district court agreed.
On appeal, the former employees argue that the district court failed to consider Dr. Kaufman’s explanation for how he had used job codes to control for criticality and versatility. But when the former employees responded to Spirit’s motion for partial summary judgment, they didn’t say anything about Dr. Kaufman’s consideration of job codes in his regression analysis.
Even if the former employees had done so, they didn’t explain how versatility and criticality could refer to a job code when Spirit sometimes compared employees within the same job code. And the former employees present no evidence that versatility or criticality are proxies for age or that Dr. Kaufman controlled for 2013 performance.
See, e.g., City of Richmond v. J.A. Croson Co., 488 U.S. 469, 501 (1989) (stating that when special qualifications are necessary, statistical comparisons must be confined to the individuals qualified for the particular task). Given Dr. Kaufman’s failure to consider these age-neutral justifications for the statistical disparity, his opinion lacks any probative
26
value. 4 See Rea v. Martin Marietta Corp., 29 F.3d 1450, 1456 (10th Cir. 1994). 5
4 A statistical disparity might be enough to create a reasonable inference of a discriminatory motive. To determine whether a disparity is significant enough, social scientists sometimes compare the actual composition of a given sample to the possibility of chance. Castaneda v. Partida, 430 U.S. 482, 496 n.17 (1977). To quantify these comparisons, social scientists have devised the concept of standard deviations. Id. A high standard deviation suggests that the makeup of a given category resulted from an intentional act or decision rather than chance. See Carpenter v. Boeing Co., 456 F.3d 1183, 1202 (10th Cir. 2006) (“what the large number of standard deviations means is that the departure from equality, whatever its magnitude, is highly unlikely to be random”).
The significance of the standard deviation, however, is case-specific.
Watson v. Fort Worth Bank & Tr., 487 U.S. 977, 995 n.3 (1988) (plurality op.). For example, if the sample is big, a standard deviation exceeding two or three would generally arouse suspicion among social scientists. Castaneda, 430 U.S. at 496 n.17. So a standard deviation exceeding two or three might “undercut” a hypothesis of chance. Hazelwood Sch. Dist. v. United States, 433 U.S. 299, 311 n.17 (1977). On the other hand, some courts use caution in drawing conclusions when the standard deviation falls in the range of one to three. Kilgo v. Bowman Transp., Inc., 789 F.2d 859, 872 (11th Cir. 1986); Gay v. Waiters’ & Dairy Lunchmen’s Union, Local No. 30, 694 F.2d 531, 551 (9th Cir. 1982); EEOC v. Amer. Nat. Bank, 652 F.2d 1176, 1192 (4th Cir. 1981).
But we haven’t adopted to a litmus test based on the number of standard deviations. See Watson, 487 U.S. at 995 n.3. After all, the usefulness of statistics turns “on all of the surrounding facts and circumstances.” Int’l Bhd. of Teamsters v. United States, 431 U.S. 324, 340 (1977). Here the surrounding circumstances involve Dr. Kaufman’s failure to account for some of the factors bearing on the retention ratings. 5 Dr. Kaufman also opined that age had influenced firings among low performers. For that opinion, Dr. Kaufman did not purport to control for 2013 performance, versatility, or criticality. See Rea, 29 F.3d at 1456.
27
2. Dr. Cahill’s expert opinions The former employees used two declarations by Kevin Cahill. He signed one in March 2022 and the other in November 2022. The former employees relied on both declarations in opposing Spirit’s motion for partial summary judgment. But the district court excluded both declarations. As a result, the court declined to consider either declaration when ruling on Spirit’s motion for partial summary judgment. Given the effect of admissibility on the summary-judgment ruling, we must decide whether the district court erred in excluding the two declarations. 6 a. The March declaration The March declaration contained two sections. In the first section, Dr. Cahill predicted Spirit’s conduct based on possibilities of a new business strategy, an effort to fire underperforming workers, or an effort to fire older or more expensive workers.
In the second section, Dr. Cahill applied the evidence of Spirit’s conduct to the three models.
6 The former employees addressed the admissibility of Dr. Cahill’s opinions as a stand-alone argument. But the appeal is interlocutory, and the district court certified only the grant of summary judgment. See Fed. R. Civ. P. 54(b). Though our jurisdiction extends only to the summary- judgment ruling, the former employees urged a fact-issue on a pattern or practice based in part on Dr. Cahill’s declarations. Given this use of the declarations, we consider their admissibility.
28
From the two sections, Dr. Cahill opined that Spirit’s layoffs hadn’t suggested a new business strategy or an effort to fire underperforming workers. For example, Dr. Cahill questioned the existence of a new business strategy because public filings had said nothing about a new strategy and training new workers could be expensive.
Similarly, Dr. Cahill questioned Spirit’s explanation involving an effort to eliminate underperforming employees, pointing out that Spirit hadn’t said anything in public filings about workplace inefficiencies.
Though Dr. Cahill discounted the first two possibilities, he regarded Spirit’s actions as consistent with a strategy to rid the workforce of older or more expensive employees. For this opinion, Dr. Cahill relied on Spirit’s short-term incentive to fire more expensive employees,
failure to say anything in public filings about a change in strategy or performance, and
internal references to expensive employees.
Spirit moved to exclude the March declaration. T he district court granted Spirit’s motion to exclude, reasoning that Dr. Cahill
had lacked expertise on the significance of omissions in public filings and
hadn’t clearly identified the empirical evidence supporting his considerations.
29
The former employees argue that the district court failed to apply adequate oversight. In overseeing opinion testimony, the district court must act as a gatekeeper, determining whether
the expert’s specialized knowledge will help the factfinder to understand the evidence or determine a disputed fact,
the expert’s opinion is based on sufficient facts or data,
the expert’s opinion is based on reliable principles and methods, and
the expert’s opinion reflects a reliable application of the facts.
Fed. R. Evid. 702. We determine de novo whether “the district court applied the proper standard and actually performed its gatekeeper role.” Dodge v. Cotter Corp., 328 F.3d 1212, 1223 (10th Cir. 2003). To meet its gatekeeping obligation, the district court must address the challenges to the expert testimony, furnishing a record that allows the reviewing court to ensure application of the law. Storagecraft Tech. Corp. v. Kirby, 744 F.3d 1183, 1190 (10th Cir. 2014). The former employees invoked this requirement, arguing that the district court had erred as a gatekeeper.
The district court appeared to reject this opinion testimony based on Dr. Cahill’s qualifications, regarding them as inadequate to support reliance on public filings. The former employees argue that this reasoning didn’t address Dr. Cahill’s ultimate opinion. But Dr. Cahill based that opinion testimony on omissions in the public filings. So the district court’s explanation addressed the basis for Dr. Cahill’s opinion. See Mitchell v.
30
Gencorp Inc., 165 F.3d 778, 782 (10th Cir. 1999) (stating that expert testimony is rendered inadmissible when it’s based on a step that renders the analysis unreliable).
The former employees insist that the district court ignored some of the evidence, other courts have accepted similar methodologies, and a jury might need Dr. Cahill’s help to reconstruct his model.
Even if the former employees were right, however, the district court would have satisfied its gatekeeping role by addressing Spirit’s challenge and providing a meaningful appellate record. See p. 29, above.
b. The November declaration In the November declaration, Dr. Cahill provided further support for his opinions, presenting new evidence that union members probably had health insurance,
Spirit’s older workers with health insurance were more likely than younger workers to lose their jobs in the reduction-in-
force, and
older workers in the same categories for tenure and job were more likely than younger workers to go in the reduction-in-
force.
In urging an unlawful pattern or practice, the former employees rely
in part on Dr. Cahill’s declaration. But the district court excluded this declaration because it had been late.
31
Exclusion was appropriate if the delay hadn’t been justified or harmless. Fed. R. Civ. P. 37(c)(1). Justification and harmlessness involved four considerations:
1. Did the delay result in surprise or prejudice to Spirit?
2. Was Spirit able to cure the potential prejudice?
3. Would introduction of the declaration disrupt the trial?
4. Did the former employees act in bad faith or with willfulness?
Woodworker ’s Supply, Inc. v. Principal Mut. Life Ins., 170 F.3d 985, 993 (10th Cir. 1999). We generally expect the district court to consider these factors when deciding whether to exclude a late declaration. HCG Platinum, LLC v. Preferred Prod. Placement Corp., 873 F.3d 1191, 1201, 1203 (10th Cir. 2017). The district court did that here. 7
7 The former employees argue that
the district court failed to address the four factors and three of the factors support admissibility.
We’ve not said whether a district court needs to discuss each of the four factors. Instead, we have reversed only when the district court overlooked “the lion’s share” of the four factors or misapplied two of the factors. HCG Platinum, LLC v. Preferred Prod. Placement Corp., 873 F.3d 1191, 1201– 03 (10th Cir. 2017) (ignored “the lion’s share”); Jacobsen v. Deseret Book Co., 287 F.3d 936, 953 (10th Cir. 2002) (misapplied two factors). But we need not decide here whether the court must discuss all the factors because the court did that here.
32
The former employees don’t question whether the court considered prejudice. The court apparently addressed prejudice by pointing out that Spirit had relied on the scope of Dr. Cahill’s initial declaration by declining to file a motion to exclude his testimony or to depose him.
The court also addressed Spirit’s ability to cure potential prejudice, reasoning that a deposition would be impractical because the trial was only a month away.
The district court also considered disruption of the trial, reasoning that a deposition would require a continuance. 8 The final factor is bad faith or willfulness. The court apparently considered this factor by chiding the former employees for trying “to insert new expert analyses” even though the data had been available earlier. Appellants’ App’x vol. 7, at 255 n.9.
The court thus considered all of the pertinent factors and reasonably concluded that they supported exclusion. As a result, the court didn’t err in excluding the November declaration.
3. Dr. Bardwell’s expert opinion The parties also discuss Dr. Robert Bardwell’s opinion testimony, but the former employees don’t appear to rely on it. In seeking summary
8 The former employees argue that if we were to remand for another reason, this factor would no longer support exclusion. But we’re not remanding for another reason.
33
judgment, Spirit presented evidence that the average age of the Wichita workforce had remained relatively constant before and after the reduction- in-force, undercutting any potential inference of ageism. In response, the former employees argued that Spirit had disregarded the flow of incoming and outgoing workers in the reduction-in-force. For this argument, the former employees relied on Dr. Bardwell’s expert report.
On appeal, Spirit argues that Dr. Bardwell compared dissimilar categories in two respects:
1. Dr. Bardwell mistakenly included older workers who had quit as well as those who were fired and
who had received an annual salary as well as those who had earned an hourly wage.
2. Dr. Bardwell erroneously compared
older workers regardless of whether they had been union members or earned a salary and
younger workers only if they had obtained salaries and union memberships.
In their reply brief, the former employees don’t address these criticisms of Dr. Bardwell’s report.
But the former employees apparently relied on this part of Dr. Bardwell’s report only to rebut Spirit’s reliance on the relative consistency of the average age of Spirit’s workforce before and after the reduction-in-force. We haven’t relied on Spirit’s evidence involving the
34
average age of the workforce. So we need not address this opinion by Dr. Bardwell.
* * *
For the reduction-in-force, we conclude that the former employees didn’t show a triable fact-issue on the existence of an unlawful pattern or practice.
III. The evidence didn’t create a triable fact-issue on the refusal to rehire older workers.
The former employees also complain that Spirit refused to rehire older workers after the reduction-in-force. The district court again concluded that the former employees hadn’t shown a standard operating procedure involving ageism, and we agree.
In part, the former employees argue that Spirit lulled the former employees into believing that they were eligible for rehire. The former employees were indeed eligible for rehire. But suppose for the sake of argument that Spirit had deceived the former employees. How would that deception suggest ageism?
In alleging deception, the former employees complain that Spirit said at a job fair that former employees would need to apply online. But everyone at the job fair was told the same thing. How could a factfinder infer ageism from a uniform instruction to apply online?
35
Similarly, the former employees complain that when they reapplied, Spirit put holds on their applications. Spirit did that for all former employees to determine whether they were eligible for rehire. But the former employees complain that the holds prevented any realistic chance of rehire.
With this complaint, the former employees imply that Spirit didn’t want to rehire them. A factfinder might reasonably agree. After all, Spirit had just fired these employees based on their poor performance, inferior versatility, and lack of criticality. See Background–Part III, above. So a factfinder could reasonably infer that Spirit wouldn’t want to rehire these individuals.
The former employees point to Spirit’s explanation, noting that Spirit was concerned about credibility with the union because Spirit had fired these employees based on poor performance. This explanation has nothing to do with age. So even if Spirit had resisted reemployment of these workers, the factfinder couldn’t reasonably infer an ageist operating procedure. IV. Conclusion The district court was right to grant Spirit’s motion for partial summary judgment on the claims involving an ageist pattern or practice. 9
9 The parties also addressed two other issues:
36
The former employees presented an array of evidence involving lay testimony, documentary evidence, and expert opinions. In combination, the evidence might reasonably support an inference that Spirit was concerned about its aging workforce. But when that evidence is viewed individually or collectively, a factfinder couldn’t reasonably infer that Spirit had an ageist pattern or practice in firing older workers or in refusing to rehire them. We thus affirm the grant of partial summary judgment to Spirit on the claim of an unlawful pattern or practice.
the enforceability of waivers by some of the former employees and
the characterization of a one-time reduction-in-force as a pattern or practice.
First, Spirit argued that many of the former employees had waived their claims. The former employees responded by challenging the enforceability of the waivers. Addressing these arguments, the district court ruled that some of the waivers were enforceable; and the former employees appeal that ruling.
Second, Spirit argued that a one-time reduction-in-force can’t constitute an unlawful pattern or practice. The district court didn’t decide this issue, but Spirit raises it as an alternative basis to affirm.
We need not address either issue because we conclude that the former employees failed to show an unlawful pattern or practice.