BACKGROUND

In 2005, the Corps sought to fill a GS-14 Supervisory Pro- curement Analyst position in the Contracting Division at its Kansas City District. The position was also known as Chief of Contracting. The Corps pursued a two-step hiring process, in which it advertised an opening for a 120-day temporary 146 SHELLEY v. GEREN position, and then announced a formal process to hire a per- manent Chief of Contracting.

An email announcement of the 120-day position was made on October 3, 2005, and courtesy copies were sent to Major Kelly Butler and Regional Contracting Chief Joseph Scanlan. The announcement explained that recruitment for the perma- nent position would begin in the near future. Applicants were instructed to email or fax their current resume/application, copies of their two most recent performance appraisals, proof that they had completed the educational requirements, and proof of eligibility for a Critical Acquisition Position.

Shelley applied for the 120-day position. At the time, he had been serving as Assistant Chief of the Contracting Divi- sion for the Walla Walla District, a GS-13 position, for over a year. In that position, Shelley supervised, coordinated, and managed the work of Team Leaders. Shelley was supervised by Connie Oberle, the Chief of Contracting at Walla Walla. He held a master’s degree in business administration, had twenty-nine years of experience in contracting (twenty-six of which were with the Corps), and had received numerous awards for his work. In 2005, Shelley had received a “Special Act Award” for “major acquisition accomplishments and acquisition research [and] policy.” In 2003, he had received a Bronze Star Medal for exceptionally meritorious acquisition service as a Contracting Officer with the Corps while deployed to the Afghanistan Area Office. He was born in 1951 and was fifty-four years old at the time of his applica- tion for the 120-day position.

Butler served as the selecting official for the 120-day posi- tion. She received about nine resumes and rated them accord- ing to the criteria from the position announcement, which she summarized as “[b]asically their experience.” No other offi- cials reviewed the resumes for the 120-day position. Butler also spoke with the applicants’ references. SHELLEY v. GEREN 147 Butler testified that Oberle gave a negative reference for Shelley. She stated, “I called Connie for a reference for [Shel- ley]. And, really, Connie’s reference is why we did not choose [Shelley].” Butler explained, “When I get a big No feeling from the supervisor, that sends a red flag.” Later, But- ler changed her testimony, stating that Oberle’s reference was “one of the reasons we did not choose [Shelley].” Oberle, however, denied ever having spoken with Butler regarding Shelley’s qualifications. Oberle testified that she spoke with Scanlan and informed him that Shelley was interested in the 120-day position, and that the job would be a “wonderful opportunity for him.”

Butler consulted about the applicants with Colonel Michael Rossi, Commander of the Kansas City District, and Steve Iverson, Deputy District Engineer for Project Management for the same district. They agreed that Vince Marsh should be hired. Marsh was serving as a Supervisory Procurement Ana- lyst and Chief of the Business Management Division in Huntsville, Alabama, a GS-14 position, and had been serving in the position for more than a year. He supervised approxi- mately fifteen employees and served as Director of Contract- ing for the Business Management Division “as requested.” In his prior position, as Business Operations Manager at the United States Army Contracting Command in Europe, Marsh had also served as Director of Contracting “as requested,” supervising seventy-five contract specialists on those occa- sions. He was forty-two years old at that time, born in 1963. He had twenty years of experience in contracting (fourteen in contracting positions and six in procurement positions). He had been with the Corps for less than two years. The most recent award listed on Marsh’s resume was a “Sustained Superior Performance Award” he had received in 2002, before joining the Corps.

Butler interviewed Marsh for the position. There is no evi- dence that she interviewed other candidates. 148 SHELLEY v. GEREN On November 2, 2005, Kevin Brice, Business Management Division Chief, sent an email seeking approval to hire Marsh for the 120-day position. Brice stated that he and Rossi rec- ommended Marsh for the position, that Scanlan had partici- pated in the selection process, and that Butler believed Marsh was Scanlan’s top pick. Marsh’s selection was approved.

Scanlan knew Shelley and was aware that Shelley was in his fifties. Scanlan had served as a superior to Shelley, and had worked with him for about six years. He was familiar with Shelley’s credentials and experience working for the Corps. He knew that, as Assistant Chief of Contracting for the Walla Walla District, Shelley had, at various times, served as Acting Chief of Contracting when Oberle was absent. Scanlan expressed confidence in Shelley’s performance of his duties as Acting Chief of Contracting and testified that both techni- cally and professionally Shelley was a good contracting offi- cer.

At the time Scanlan supported Marsh for the 120-day posi- tion, he knew Marsh only by reputation. They had met at a social event for the Army Contracting Command in Germany. Scanlan’s belief that Marsh was the best candidate for the 120-day position was not based on any personal experience working with him. Scanlan, however, told Butler, Brice, Iver- son, Rossi and Kevin Bond, District Counsel Chief, who later joined the selection panel for the permanent position, that he had worked with Marsh in Germany, and that he believed Marsh would do very well in the 120-day position.

Shelley learned that he was not selected for the 120-day position on or about November 4, 2005.

Meanwhile, on October 24, 2005, the permanent position and job description had been announced, and the Corps began accepting applications. The selection plan called for a panel of five members to review applications. The panel members were Scanlan, Brice, Rossi, Bond, and Mary Parks, Chief SHELLEY v. GEREN 149 Contracting Specialist. Scanlan, Brice and Rossi had all par- ticipated in the hiring decision for the 120-day position. Rossi was assigned to chair the panel.

The selection plan identified four criteria on which to screen applicants for interviews: technical competency, man- agement skills, leadership and teamwork. On each criterion, the applicants were to be evaluated as “outstanding,” “fully successful,” or “minimally acceptable.” Possession of a grad- uate level degree was a factor in ranking a candidate as out- standing for technical competency. A factor to be considered with regard to management skills was supervision of over thirty employees.

Oberle testified that, around the time the hiring process was taking place, Scanlan and Brice requested from the contract- ing chiefs information about projected retirement dates for employees in their districts and divisions. Scanlan did not recall asking his chiefs for information on retirement eligibil- ity. He admitted, though, that in 2004 or 2005 he had requested, from the districts, certain data which, at that time, was provided in a spreadsheet entitled Capable Workforce Matrix. Although the matrix did not include the names of the employees, it included information such as job titles, grade levels, number of employees in a particular position in a divi- sion, as well as their anticipated retirement dates. The exam- ple in the record of this matrix for the Walla Walla Contracting Division is dated March 21, 2006, but apparently the same format was used in 2004 and 2005. It is clear from the 2006 version of the matrix for the Walla Walla Contract- ing Division that it would be a simple matter to deduce the names of the incumbents from the position titles within the division.

Thirty-three individuals applied for the permanent position, including Shelley, Marsh, and Oberle. The panel members independently evaluated the applicants as outstanding, fully successful, or minimally acceptable, on each of the four selec- 150 SHELLEY v. GEREN tion criteria, based on their resumes. On December 19, 2005, the panel members convened by teleconference to select can- didates for interviews. Scanlan testified that he did not share any age-related information about Shelley at the teleconfer- ence. Brice testified that age was not a consideration in evalu- ating the applicants, although information on the resumes could allow panelists to estimate applicants’ ages.

During the teleconference, each panelist placed the candi- dates in either the top third, middle third, or bottom third of the applicant pool. The spreadsheet summarizing these scores does not identify the panelists by name, but it shows that two candidates received a top score from each of the five panel- ists. Marsh and another candidate received four top scores and a mid score. A fifth candidate received four top scores and a bottom score. An applicant named Robert received three top scores and two mid scores. These were the six candidates selected for interviews. Shelley was given a top score by three of the five panelists. He was initially given a mid score by two panelists. This ranking would have been equal to that which had earned Robert an interview. But one panelist— whose identity is not disclosed in the record—changed Shel- ley’s mid score to a bottom score. Shelley was not given an interview.

Marsh was, at forty-two years old, the youngest inter- viewee. The oldest interviewee was fifty-five years old, one year older than Shelley. The other interviewees were forty-six (two of them), fifty, and fifty-three years old.

On January 20, 2006, the panel recommended Marsh for the permanent position. On or about February 17, 2006, Shel- ley learned that he had not been afforded an opportunity to interview for the permanent position. On April 16, 2006, Marsh was reassigned to the permanent Chief of Contracting position.

On March 6, 2006, seventeen days after Shelley learned that he had been denied an interview for the permanent posi- SHELLEY v. GEREN 151 tion, he made initial contact with the Corps’ Equal Employ- ment Opportunity (EEO) officer. On May 12, 2006, after receiving notice of his right to file a formal complaint of dis- crimination, Shelley did so, alleging that he had been discrim- inated against between November 2005 and January 2006 due to his age, in that he was “not afforded the anticipated inter- view opportunity . . . thereby eliminating his promotion opportunity for the Kansas City District GS-14, Chief, Con- tracting Division position.” After the EEO office denied his claim in its Final Agency Action on June 27, 2008, Shelley filed suit in federal district court on July 28, 2008.

The district court granted summary judgment in favor of the Corps. The court assumed, without deciding, that Shelley timely exhausted his administrative remedies as to both the 120-day position and the permanent position. The court declined to analyze the motion in accordance with McDonnell Douglas Corporation v. Green, 411 U.S. 792 (1973), finding it inapplicable to ADEA cases after the Supreme Court’s deci- sion in Gross v. FBL Financial Services, Inc., 129 S. Ct. 2343 (2009). Relying on Gross, the district court held that Shelley put forth insufficient facts that his age was the “but-for” cause of his non-selection for the 120-day position and for an inter- view for and promotion to the permanent position.

Shelley appeals.

STANDARD OF REVIEW

We review de novo a district court’s grant of summary judgment. Bagdadi v. Nazar, 84 F.3d 1194, 1197 (9th Cir. 1994). We must determine, viewing the evidence in the light most favorable to the non-moving party, whether there are any genuine issues of material fact, and whether the district court applied the relevant substantive law. Id. “Whether a plaintiff has exhausted administrative remedies as required before filing suit is a question of law, which we review de novo.” Bankston v. White, 345 F.3d 768, 770 (9th Cir. 2003). 152 SHELLEY v. GEREN

DISCUSSION

I. Administrative Remedies

Preliminarily, the Corps argues that we may not consider Shelley’s complaint of non-selection for the 120-day position because he failed to seek administrative remedies for that decision in a timely manner. The Corps argues that Shelley failed to contact the EEO office within forty-five days of learning that he was not selected for the 120-day position and failed to complain about his non-selection for the 120-day position in his formal complaint of discrimination in the EEO administrative process.

Federal employees who believe they have been discrimi- nated against on the basis of age have “the option of pursuing administrative remedies, either through the agency’s EEO procedures, or through the Merit Systems Protection Board.” Bankston, 345 F.3d at 770 (internal citations omitted). Equal Employment Opportunity Commission (EEOC) regulations provide that an aggrieved federal employee who pursues the EEO avenue must consult an EEO counselor within forty-five days of the effective date of the contested personnel action, prior to filing a complaint alleging age discrimination. 29 C.F.R. §§ 1614.103, 1614.105(a)(1).1 We have stated that

although the regulatory pre-filing exhaustion requirement at § 1614.105 “does not carry the full weight of statutory authority” and is not a jurisdic- tional prerequisite for suit in federal court, we have consistently held that, absent waiver, estoppel, or 1 As an alternative to filing an administrative complaint, a federal employee may file a civil action in a United States district court under the ADEA after giving the EEOC not less than thirty days’ notice of intent to sue. 29 U.S.C. § 633a(d); 29 C.F.R. § 1614.201(a); see Bankston, 345 F.3d at 770. There is no evidence that Shelley provided the EEOC with notice of his intent to sue the Corps, and he does not appear to rely on this alter- native avenue. SHELLEY v. GEREN 153 equitable tolling, “failure to comply with this regula- tion [is] . . . fatal to a federal employee’s discrimina- tion claim” in federal court.

Kraus v. Presidio Trust Facilities Div./Residential Mgmt. Branch, 572 F.3d 1039, 1043 (9th Cir. 2009) (alterations in the original) (quoting Lyons v. England, 307 F.3d 1092, 1105 (9th Cir. 2002)).

Shelley took the position that he timely initiated the EEO process on March 6, 2006, after he learned on or about Febru- ary 17, 2006, that he had been denied an opportunity to inter- view for the permanent Chief of Contracting position. In his EEO complaint, Shelley asserted that he was discriminated against, between November 2005 and January 2006, based on his age. Shelley asserted that because of his age he was not given an interview, thereby eliminating his promotion oppor- tunity. Shelley learned that he was not selected for the 120- day position on or about November 4, 2005. Thus, the time period Shelley specified in his EEO complaint encompasses the hiring process for both the 120-day and the permanent Chief of Contracting positions. Reading the EEO complaint liberally, as we must, see Greenlaw v. Garrett, 59 F.3d 994, 999 (9th Cir. 1995), it is apparent that Shelley complained about both hiring decisions.

Further, the decisions were not discrete employment actions, but were part of a single, two-step, hiring process. The Corps sought to fill the position first on a temporary basis, followed by a permanent appointment after 120 days. It is obvious that the person selected for the temporary posi- tion would have a significant competitive advantage over the other applicants for the permanent position, and therefore that the temporary appointment could be seen as a step towards the permanent appointment. Also, the limited nature of the hiring process for the temporary position, and the fact that the recruitment for the permanent position started less than a 154 SHELLEY v. GEREN month after the temporary position was announced, could have led an applicant to view the processes as a continuum.

The interrelatedness of the two positions and of the hiring processes for them persuades us that the process for filling the Chief of Contracting position was a single course of conduct that began in 2005 with the selection of Marsh for the 120- day position and ended on April 16, 2006, when Marsh was confirmed as the new Chief of Contracting. Because Shelley filed his EEO complaint on March 6, seventeen days after he learned that he had not been selected to interview for the per- manent position, he met the 45-day requirement of 29 C.F.R. § 1614.105(a).

Even if we assume that the two promotions were discrete employment actions, Shelley’s complaint was still timely. “Incidents of discrimination not included in an EEOC charge may not be considered by a federal court unless the new claims are like or reasonably related to the allegations con- tained in the EEOC charge.” Green v. Los Angeles County Superintendent of Schools, 883 F.2d 1472, 1476 (9th Cir. 1989) (internal quotation marks omitted). In determining whether a new claim is like or reasonably related to allega- tions contained in the previous charge, the court inquires into “whether the original EEOC investigation would have encom- passed the additional charges.” Id. The same is true of a com- plaint of discrimination submitted to a federal agency’s EEO office. See Greenlaw, 59 F.3d at 1000 (citing Sosa v. Hiraoka, 920 F.2d 1451, 1456-57 & n.2 (9th Cir. 1990)).

Here, the crux of Shelley’s complaint is that he was bypassed for promotion to the permanent Chief of Contracting position because of his age. Because of the close relationship between the two positions and the temporally-overlapping hir- ing processes for them, an EEO investigation into the hiring process for the permanent position would necessarily have led to the investigation of the hiring process for the temporary position. The case is therefore distinguishable from Williams SHELLEY v. GEREN 155 v. Little Rock Municipal Water Works, 21 F.3d 218 (8th Cir. 1994), upon which the Corps relies. There, the Eighth Circuit affirmed partial summary judgment in favor of the defendant on the plaintiff’s racial discrimination claim, finding that the plaintiff’s EEOC complaint for retaliation included no men- tion of racial discrimination, and her allegations of racial dis- crimination submitted to the EEOC years earlier were not deemed reasonably related to her current claim for retaliation. Id. at 222-23.

[1] In sum, Shelley’s initial contact with the Corps’ EEO officer seventeen days after he learned that he had been denied the opportunity to interview for the permanent Chief of Contracting position timely initiated his administrative claim based on being denied interviews and selection for the 120-day and the permanent positions, all of which occurred as part of the same course of conduct by the Corps. Shelley timely exhausted available administrative remedies.

II. Summary Judgment Disposition of Age Discrimination

Claim

[2] Shelley’s failure-to-promote claim is a claim of dispa- rate treatment under the ADEA. The ADEA makes it unlaw- ful for an employer to discriminate “because of [an] individual’s age.” 29 U.S.C. § 623(a)(1). The prohibition is “limited to individuals who are at least 40 years of age.” 29 U.S.C. § 631(a). The ADEA applies to protect federal employees and applicants for federal employment. 29 U.S.C. § 633a(a). To prevail on a claim for age discrimination under the ADEA, a plaintiff must prove at trial that age was the “but-for” cause of the employer’s adverse action. Gross, 129 S. Ct. at 2350. “Unlike Title VII, the ADEA’s text does not provide that a plaintiff may establish discrimination by show- ing that age was simply a motivating factor.” Id. at 2349.

[3] This case, however, was resolved on summary judg- ment, and not on the merits. Prior to Gross, our circuit applied 156 SHELLEY v. GEREN the burden-shifting evidentiary framework of McDonnell Douglas, 411 U.S. at 802, to motions for summary judgment on ADEA claims. See, e.g., Coleman v. Quaker Oats Co., 232 F.3d 1271, 1281 (9th Cir. 2000); Wallis v. J.R. Simplot Co., 26 F.3d 885, 889 (9th Cir. 1994); Rose v. Wells Fargo & Co., 902 F.2d 1417, 1420 (9th Cir. 1990); Steckl v. Motorola, Inc., 703 F.2d 392, 393 (9th Cir. 1983). The district court declined to apply this framework, believing that Gross rejected it. The Corps argues to the same effect.

[4] We disagree. In Gross, the Court grappled with whether a mixed-motives instruction may be given to the jury in an ADEA case.2 129 S. Ct. at 2348. Relying on the text of 29 U.S.C. § 623(a)(1) and case law allocating the burden of persuasion, the Court held that a plaintiff retains at all times the burden of persuasion to establish that age was the “but- for” cause of an employer’s adverse action. Id. at 2352. Because Gross involved a case that had already progressed to trial, it did not address the evidentiary framework applicable to a motion for summary judgment. The Court, in fact, explic- itly noted that it “has not definitively decided whether the evi- dentiary framework of McDonnell Douglas utilized in Title VII cases is appropriate in the ADEA context.” Id. at 2349 n.2.

[5] Since the decision in Gross, several sister circuits have continued to utilize the McDonnell Douglas framework to decide motions for summary judgment in ADEA cases. See, e.g., Leibowitz v. Cornell Univ., 584 F.3d 487, 498 (2d Cir. 2 Mixed-motives jury instructions are used in Title VII cases where an employee alleges that he or she suffered an adverse employment action because of both permissible and impermissible considerations, i.e., a “mixed-motives” case. Gross, 129 S. Ct. at 2347 (citing Price Waterhouse v. Hopkins, 490 U.S. 228 (1989)). Under such instructions, if a Title VII plaintiff shows that discrimination was a “motivating” or a “substantial” factor in the employer’s action, the burden of persuasion would shift to the employer to show that it would have taken the same action regardless of that impermissible consideration. Id. SHELLEY v. GEREN 157 2009); Velez v. Thermo King Day P.R., Inc., 585 F.3d 441, 446-47 (1st Cir. 2009); Connolly v. Pepsi Bottling Grp., LLC, 2009 WL 3154445 at *2-3 (3d Cir. 2009) (unpublished).3 We join them and hold that nothing in Gross overruled our cases utilizing this framework to decide summary judgment motions in ADEA cases. The McDonnell Douglas test is used on summary judgment, not at trial. Costa v. Desert Palace, Inc., 299 F.3d 838, 855 (9th Cir. 2002) (“This legal proof structure is a tool to assist plaintiffs at the summary judgment stage so that they may reach trial. . . [I]t is not normally appropriate to introduce the McDonnell Douglas burden- shifting framework to the jury.”). The McDonnell Douglas test shifts only the burden of production, after the plaintiff makes a prima facie case. See, e.g., Tusing v. Des Moines Cmty. Sch. Dist., 639 F.3d 507, 515 n.3 (8th Cir. 2011) (“The McDonnell Douglas analysis is likely still an appropriate way to analyze ADEA ‘pretext’ claims, however, because McDon- nell Douglas only shifts the burden of production.”); Smith v. City of Allentown, 589 F.3d 684, 691 (3d Cir. 2009) (“Gross stands for the proposition that it is improper to shift the bur- den of persuasion to the defendant in an age discrimination case. McDonnell Douglas, however, imposes no shift in that particular burden.”). “If plaintiffs establish a prima facie case, ‘[t]he burden of production, but not persuasion, then shifts to the employer to articulate some legitimate, nondiscriminatory reason for the challenged actions.’ ” Hawn v. Exec. Jet Mgmt., Inc., 615 F.3d 1151, 1155 (9th Cir. 2010) (quoting Chuang v. Univ. of Cal. Davis, Bd. of Trs., 225 F.3d 1115, 1123-24 (9th Cir. 2000)). “If defendant meets this burden, plaintiffs must then raise a triable issue of material fact as to whether the defendant’s proffered reasons for their terminations are mere pretext for unlawful discrimination.” Id. Because the contin- ued use of McDonnell Douglas in summary judgment motions 3 We have done the same in non-precedential, unpublished decisions. Russell v. Mountain Park Health Ctr. Props., LLC, 403 Fed. App’x. 195, 196 (9th Cir. 2010); EEOC v. Banner Health, 402 Fed. App’x. 289, 2010 WL 4324409, at *1 (9th Cir. 2010). 158 SHELLEY v. GEREN on ADEA claims is not inconsistent with Gross, we cannot overrule our prior precedent because of Gross.

[6] Thus, to survive summary judgment on his claim for a violation of the ADEA under the disparate treatment theory of liability, Shelley must first establish a prima facie case of age discrimination. Coleman, 232 F.3d at 1280-81. If he is suc- cessful, the burden of production shifts to the Corps to articu- late a legitimate non-discriminatory reason for its adverse employment action. Id. at 1281. It is then Shelley’s task to demonstrate that there is a material genuine issue of fact as to whether the employer’s purported reason is pretext for age discrimination. Id. At trial, he must carry the burden to prove that age was the “but-for” cause of his non-selection.

A. Prima Facie Case

A “prima facie case requires evidence adequate to create an inference that an employment decision was based on a[n] [illegal] discriminatory criterion.” O’Connor v. Consol. Coin Caterers Corp., 517 U.S. 308, 312 (1996) (internal quotation marks omitted) (alterations in original).

[7] In a failure-to-promote case, a plaintiff may establish a prima facie case of discrimination in violation of the ADEA by producing evidence that he or she was (1) at least forty years old, (2) qualified for the position for which an applica- tion was submitted, (3) denied the position, and (4) the pro- motion was given to a substantially younger person. See Steckl, 703 F.2d at 393 (holding that the plaintiff established a prima facie case for age discrimination under the McDon- nell Douglas framework because he “was clearly within the protected class, had applied for an available position for which he was qualified, and was denied a promotion which was given to a younger person”); see also O’Connor, 517 U.S. at 313 (“Because the ADEA prohibits discrimination on the basis of age and not class membership, the fact that a replacement is substantially younger than the plaintiff is a far SHELLEY v. GEREN 159 more reliable indicator of age discrimination than is the fact that the plaintiff was replaced by someone outside the pro- tected class.”); Nidds v. Schindler Elevator Corp., 113 F.3d 912, 917 (9th Cir. 1996) (finding that the requirements for a prima facie case of age discrimination were satisfied by evi- dence that the fifty-four year old plaintiff was discharged, he was performing his job satisfactorily, and his duties continued to be performed by a substantially younger employee, and holding that the district court “erred in concluding that to establish a prima facie case, [the plaintiff] was required to show that he was at least as qualified as his replacement”).4

[8] It is undisputed that Shelley was fifty-four at the rele- vant time, he was qualified for both the temporary and the permanent positions, he was denied both positions, and both went to a substantially younger candidate. Accordingly, Shel- ley has established a prima facie case of age discrimination.

B. Facial Legitimacy of the Corps’ Explanation

[9] The burden now shifts to the Corps to provide a non- discriminatory explanation for its hiring decisions. Coleman, 232 F.3d at 1281. The Corps did so here. In its brief on appeal, the Corps proffers as its non-discriminatory explana- tion that Marsh was already employed as a GS-14 Supervi- sory Procurement Analyst, and hiring him caused a lateral move, whereas the position would have been a promotion for Shelley, who was serving at the GS-13 level. This is a facially legitimate explanation. 4 It thus appears that, as part of the prima facie case, a plaintiff does not have to show that he was discriminated against in favor of a substantially younger employee with equal or inferior qualifications. If Shelley were required to show this, he has done so, as discussed below in connection with his showing of pretext. See Lowe v. City of Monrovia, 775 F.2d 998, 1005 (9th Cir. 1985), as amended, 784 F.2d 1407 (1986) (holding that in order to show pretext, a plaintiff may rely on the same evidence that he offered to establish a prima facie case). 160 SHELLEY v. GEREN C. Pretext

[10] The Corps’ articulation of a legitimate non- discriminatory reason shifts the burden back to Shelley to raise a genuine factual question as to whether the proffered reason is pretextual. The plaintiff can prove pretext “(1) indi- rectly, by showing that the employer’s proffered explanation is ‘unworthy of credence’ because it is internally inconsistent or otherwise not believable, or (2) directly, by showing that unlawful discrimination more likely motivated the employer.” Chuang, 225 F.3d at 1124. All of the evidence—whether direct or indirect—is to be considered cumulatively. Id. We conclude that Shelley offered both direct and indirect evi- dence of pretext.

i. Direct Evidence

[11] Shelley presented direct evidence of age discrimina- tion to rebut the Corps’ purported non-discriminatory reason. Oberle testified that Scanlan and Brice inquired about the pro- jected retirement dates for employees in the contracting divi- sions during the hiring period for the 120-day and permanent positions. A fact-finder could infer from this that they consid- ered age and projected retirement relevant to the hiring deci- sion. Despite the absence of names in the Capable Workforce Matrix, the format of the matrix permitted the identification of specific employees. The matrix contradicts Scanlan’s testi- mony that individual age information was not provided in that format.5

[12] The Corps contends that the matrix, at best, estab- lishes that Scanlan and Brice knew Shelley’s prospective 5 Although the Walla Walla Capable Workforce Matrix in the record is dated March 21, 2006, after Shelley was denied an opportunity to inter- view, it displays the same information that Scanlan and Brice had requested earlier, and shows that one could deduce the names of individual employees. SHELLEY v. GEREN 161 retirement date, which, standing alone, would not support an inference of age discrimination. But the fact that Scanlan and Brice sought out the retirement dates at the time of their par- ticipation in the hiring process for the temporary and perma- nent positions shows more than that the decision-makers may have known of the candidates’ ages. It raises an inference that they considered this information relevant to their decisions. Although Scanlan and Brice did not make the hiring decisions alone, evidence of their inquiry and of their influence over the process supports an inference that the Corps’ proffered expla- nation for hiring Marsh was a pretext for age discrimination. See Staub v. Proctor Hospital, 131 S. Ct. 1186, 1194 (2011) (under Uniformed Services Employment and Reemployment Rights Act, discriminatory motive imputed to employer where a supervisor performs an act motivated by discriminatory ani- mus that is intended by the supervisor to cause an adverse employment action, and that act is a proximate cause of the ultimate employment action).

ii. Indirect Evidence

[13] Evidence of a plaintiff’s superior qualifications, standing alone, may be sufficient to prove pretext. Raad v. Fairbanks North Star Borough School Dist., 323 F.3d 1185, 1194 (9th Cir. 2003) (citing Odima v. Westin Tucson Hotel, 53 F.3d 1484, 1492 (9th Cir. 1995)). A comparison of Shel- ley’s and Marsh’s resumes gives rise to a factual dispute as to whether Shelley was better qualified for the position than Marsh. Compared to Marsh, Shelley had significantly more years of work experience related to contracting, and more experience employed in the Corps. As of October 2005, Shel- ley had twenty-nine years of experience in contracting, whereas Marsh had twenty years. Unlike Marsh, Shelley spent most of his career in the Corps. Shelley had been an employee of the Corps for over nineteen years, Marsh for five and a half years.

Furthermore, Shelley was already employed in a Contract- ing Division, while Marsh was a Supervisory Procurement 162 SHELLEY v. GEREN Analyst in a Business Management Division. Shelley had been employed as Assistant Chief in the Contracting Division under Oberle in Walla Walla for over a year, and before that for four years as a Team Leader in Walla Walla under Oberle. Contrary to the Corps’ contention, Marsh served as Director of Contracting only “as requested.” Shelley, too, served as Acting Chief of Contracting in his supervisor’s absence. Although this fact was not included in his resume, Scanlan was aware of it, and testified that he had confidence in Shel- ley’s performance as Acting Chief of Contracting. Shelley’s resume identified more impressive and recent awards for on- the-job accomplishments than Marsh’s. Shelley’s educational qualifications were superior to Marsh’s. Shelley held an M.B.A., while Marsh had no gradu- ate level degree. The selection criteria for the permanent posi- tion indicate that possession of a graduate level degree is a factor in ranking a candidate as outstanding for technical competency.

The selection criteria for the permanent position listed supervision of at least thirty employees as a factor to be con- sidered in evaluating management skills. The Corps incor- rectly asserts that Shelley’s resume failed to indicate that he acted as a supervisor. While Shelley’s resume did not specify the number of employees he supervised, it did disclose that, as Assistant Chief of Contracting, he supervised, coordinated and managed the work of subordinate Team Leaders, who presumably led teams. Marsh relayed that he supervised fif- teen employees in the position that he held at the time of his application. Although Marsh had supervised seventy-five employees on occasion when he served temporarily as Direc- tor of Contracting in Europe, it was not part of his regular job duties. Neither candidate clearly demonstrated that he met this criterion.

[14] The Corps argues that, while Shelley’s qualifications as described after the fact were extensive, he did not include SHELLEY v. GEREN 163 all of his work experience and skills in his resume. On appeal, however, Shelley relies exclusively on information that was listed in his resume to argue that his qualifications for the position were superior. Even absent the additional information (e.g., the value of construction contracts he successfully shep- herded), Shelley’s resume demonstrated sufficient qualifica- tions that a reasonable jury could find that he was substantially better qualified than Marsh. In addition, Scanlan testified that he had worked with Shelley for six years, and was familiar with Shelley’s experience and credentials.

Further, that Shelley was a GS-13 employee, while Marsh was a GS-14, was relevant only as to the 120-day position. Had Shelley been given that position, he would have become a GS-14 and his move to the permanent position would have been lateral, like Marsh’s.

None of the officials whose support for Marsh was cited in the email seeking approval for his hire for the 120-day posi- tion had reviewed the applicants’ resumes. Only Butler reviewed the resumes. Viewed in the light most favorable to Shelley, Butler’s testimony, read in conjunction with Oberle’s testimony, could be understood to indicate that Butler initially favored Shelley for the position, but that she used an alleged negative reference from Oberle (which Oberle denies she ever gave) as a pretext for hiring Marsh after learning that Scanlan favored Marsh for the position. Scanlan, it bears repeating, represented to Butler, Rossi, Brice, Iverson and Bond that he had worked with Marsh in Germany. He testified, however, that he recommended Marsh based only on his reputation. Scanlan had met him at a social event, but had never worked with him. As noted, Scanlan had sought out information about employees’ retirement eligibility at the time of the hiring pro- cess.

[15] Accordingly, Shelley’s rejection for the 120-day posi- tion could be found to be based on age discrimination. If it was, then the inference that the decision-makers were biased 164 SHELLEY v. GEREN would carry over to their decision-making for the permanent position. Further, the denial of the temporary position was clearly a causative factor in the denial of the interview and selection for the permanent position. If the first decision was caused by discrimination, a strong inference is raised that the subsequent decisions were as well.

The Corps argues that age bias cannot be inferred in the selection for the permanent position, because other applicants close in age to Shelley were interviewed for that position and Shelley was not. Those applicants, however, were not selected for the position, and instead Marsh, the significantly younger applicant, was hired. Stacking the interview pool with older candidates does not immunize the decision to hire a younger one. Of the five panelists who selected the interviewees, Scanlan, Brice, and Rossi had all participated in the hiring decision for the 120-day position. Shelley received a top score from three of the five panelists. It was only the alteration of one unidentified panelist’s score for Shelley from mid to bot- tom that cost Shelley an interview and disqualified him for the permanent position. The evidence of Scanlan’s discriminatory animus discussed above supports an inference that Scanlan was biased against Shelley and in favor of Marsh based on age. The evidence of the workings of the hiring process sup- ports an inference that Scanlan was able to influence the inter- view and selection decisions. Neither Shelley’s non-selection for an interview for the permanent position, nor the interviews of other older applicants who were not selected for promotion, disproves Shelley’s evidence supporting a prima facie case and pretext.

[16] The evidence, viewed in the light most favorable to Shelley, is sufficient to allow a reasonable jury to find that the Corps’ reliance on Marsh’s GS-14 level, as compared to Shel- ley’s GS-13 level, was pretextual in the light of Shelley’s oth- erwise superior experience, education and recognition.

In sum, Shelley produced sufficient evidence to establish a prima facie case of age discrimination, and has responded to SHELLEY v. GEREN 165 the Corps’ alleged non-discriminatory reason for refusing to promote him by identifying evidence, both direct and indirect, showing that the Corps’ explanation is pretextual.

CONCLUSION

[17] Because Shelley initiated a timely administrative complaint and produced sufficient evidence in support of his ADEA claim, the district court’s grant of summary judgment in the Corps’ favor is reversed. The case is remanded to the district court for further proceedings in accordance with this opinion.

REVERSED and REMANDED.

BYBEE, Circuit Judge, concurring in part and dissenting in part:

This should be a straightforward case. Plaintiff Devon Scott Shelley (“Shelley”) claims he became a victim of age discrim- ination when the Army Corps of Engineers (“the Corps”) denied him an opportunity to interview for a GS-14 position as Chief of the Contracting Division that ultimately went to a younger candidate. An unfiltered look at the facts reveals that of the six finalists the Corps interviewed, one candidate was older than Shelley, and two others were close to Shelley in age. The candidate the Corps ultimately hired, Vince Marsh, was already a GS-14 Supervisory Procurement Ana- lyst, while Shelley was a GS-13 Assistant Chief.

On this record, there is no way Shelley can show that the Corps passed him over for an equally or less qualified candi- date on account of his age. Rather, the record shows that 32 individuals applied for the position. A five-member selection committee, chaired by an army colonel and advised by an EEO officer, independently ranked the candidates on the basis 166 SHELLEY v. GEREN of their resumes. Based on their individual evaluations of the candidates, the committee held a telephone conference and produced a list of the top six candidates. Shelley was not ran- ked among the top six, but was included in the second-tier group of nine candidates. The six finalists included three men and three women (one of whom was Shelley’s own supervi- sor, Connie Oberle), and were born in 1950, 1952, 1955, 1959 (2), and 1963; Shelley was born in 1951. The committee then jointly interviewed the finalists and unanimously recom- mended hiring Marsh. In its report, the committee found that Marsh had the “strongest interview. . . . [It] demonstrated [his] technical competency, professionalism, leadership and strategic thinking.” The committee also found that Marsh had “the highest overall positive references.” In the Corps’s own investigation of Shelley’s complaint, every member of the committee denied that age played any role in the committee’s decision.

To this overwhelming evidence that age was not the reason the committee decided to hire Marsh and not Shelley, the majority simply points to two facts: (1) some members of the committee likely knew how old Shelley was, Maj. Op. at 161, and (2) Shelley believed he had more experience and, there- fore, was better qualified than Marsh, Maj. Op. at 161-163. This doesn’t come close to proving that the Corps “refuse[d] to hire [Shelley] . . . because of such individual’s age. 29 U.S.C. § 623(a)(1) (emphasis added). And because the Age Discrimination in Employment Act (“ADEA”) does not per- mit a mixed-motive theory, Gross v. FBL Fin. Servs., Inc., 129 S. Ct. 2343, 2350 (2009), Shelley has no case at all. I would affirm the district court’s award of summary judgment, and I respectfully dissent.1 1 Because I agree that Shelley exhausted his administrative remedies, I concur in Part I of the majority opinion. I also agree that the district court erred in holding that the McDonnell Douglas framework does not apply to ADEA claims. Accordingly, I concur in Parts I through II.A of the majority opinion. SHELLEY v. GEREN 167 I

In Gross, the Supreme Court held that “under the plain lan- guage of the ADEA . . . a plaintiff must prove that age was the ‘but-for’ cause of the employer’s adverse decision.” Id. at 2350. This “but-for” test applies no matter whether the case was resolved on summary judgment or after a jury trial, and it does not permit the plaintiff to rely on a mixed-motive the- ory. See id.

Because Shelley has no direct evidence of discrimination based on age, he must rely on the burden-shifting approach articulated in McDonnell Douglas Corp. v. Green, 411 U.S. 792 ( 1973). The majority holds that Gross did not overrule our prior cases holding that the McDonnell Douglas frame- work applies to a disparate-treatment ADEA claim. Maj. Op. at 156-158. I concur in that part of the opinion. Although I think the McDonnell Douglas framework is going to be diffi- cult to apply to an ADEA claim after Gross—and the Court was coy about whether McDonnell Douglas is compatible with Gross, 129 S. Ct. at 2349 n.2—I agree that Gross does not clearly overrule our prior precedents. See Smith v. City of Allentown, 589 F.3d 684, 691 (3d Cir. 2009) (“Gross does not conflict with our continued application of the McDonnell Douglas paradigm in age discrimination cases.”).

Even if we continue to apply the McDonnell Douglas framework to the ADEA, Shelley still shoulders the ultimate burden of showing that the Corps’s explanation—that Shelley was not as qualified as Marsh—was pretextual and that the necessary reason Shelley was not hired was because of his age. “To establish a disparate-treatment claim under the plain language of the ADEA . . . a plaintiff must prove that age was the ‘but-for’ cause of the employer’s adverse decision,” Gross, 129 S. Ct. at 2350, because “the burden of persuasion [n]ever shifts to the party defending an alleged mix-motives discrimination claim brought under the ADEA,” id. at 2348. 168 SHELLEY v. GEREN That is a heavy burden for Shelley to carry. And he doesn’t come close on this record. The processes used by the Corps rule out impermissible bias, and the evidence relied on by the majority is insufficient to raise an issue of material fact.

II

The majority finds that “Shelley’s rejection for the 120-day position could be found to be based on age discrimination.” Maj. Op. at 163. It claims that Major Kelly Butler’s and Con- nie Oberle’s testimonies could show that “Butler initially favored Shelley for the position, but that she used an alleged negative reference from Oberle (which Oberle denies she ever gave) as a pretext for hiring Marsh after learning that Scanlan favored Marsh for the position.” Id. at 163. I think the 120- day position is not really an issue.2 In any event, this is simply not a reasonable inference. Although there is some dispute over whether or how Oberle conveyed a reference to Butler, whether she did offer a negative reference is beside the point (though she was going to be a candidate for the permanent position herself). It is not reasonable to infer that Butler would have chosen Shelley but for an alleged reference from Oberle. See Maj. Op. at 163. Butler did state in the Depart- 2 The majority mischaracterizes the Corps’s explanation for its decisions as being that Marsh was already a GS-14 whereas Shelley was only a GS- 13. Maj. Op. at 159. As I discuss in Part IV, Marsh’s prior permanent GS- 14 position is of course relevant and demonstrates his superior qualifica- tions, though it is hardly the only way in which he was a better applicant than Shelley. The Corps has contended throughout that the six candidates selected for an interview were each generally better qualified than Shelley and that the selection process worked correctly; even Shelley acknowl- edged this as the explanation offered by the Corps. Because of this mis- characterization, the majority can claim that the Corps’s explanation was irrelevant for the permanent hiring decision because if Shelley had received the temporary position, his receiving the permanent position would have been a lateral move, too. See Maj. Op. at 163. It also paves the way for the majority’s erroneous conclusion that a “reasonable jury [could] find that the Corps’ reliance on Marsh’s GS-14 level, as compared to Shelley’s GS-13 level, was pretextual.” Maj. Op. at 164. SHELLEY v. GEREN 169 ment of Defense Fact-Finding Conference of March 14, 2007 that “really, Connie’s reference is why we did not choose Scott [Shelley].” Yet the Fact-Finding Conference at which she testified was not convened to consider Shelley’s com- plaint, and, in context, her statement is not particularly proba- tive. Moreover, in her deposition for this case she clarified her testimony, saying: “I didn’t mean to indicate that I was ready to hire him and Connie Oberle said no. That’s not the case. . . . Connie’s reference is one of the reasons we did not choose Scott. . . . I did not mean to infer that she was the only reason that . . . he wasn’t chosen.” She stated that she based her hir- ing decision on Marsh’s being “the most qualified out of the nine” who applied.

Furthermore, it is unreasonable to infer that Butler changed her mind after input from Scanlan. She testified that she was the only one who reviewed the resumes for that position, and she made the decision in consultation with Colonel Michael Rossi, Commander for the Kansas City District, and Steve Iverson, Deputy District Engineer for Project Management for the Kansas City District. Although Scanlan may have been “a part of the process” and let Butler know that Marsh was his top pick, there is no indication that he was involved in the substance of the decision, and Shelley produced no evidence to contradict Butler’s account of how the decision was made. A reasonable jury could not reasonably conclude from the evidence that discrimination was the but-for cause of Shelley failing to receive the 120-day position.

With regard to the permanent position, Shelley’s biggest problem is that before he can make out a case that he did not get the position because of his age, Shelley has to show that he was passed over for an interview because of his age. This he cannot do. Of the six candidates who were actually selected for a final interview (from a list of 32), one was older than Shelley, one was only a year younger, and another was four years younger. The fact that the Corps considered quali- fied candidates older than (or about the same age as) Shelley 170 SHELLEY v. GEREN without offering Shelley an interview is fatal to his claim. It is true that the fourth traditional element for establishing a prima facie case in failure to promote cases under the McDonnell-Douglas framework—that the employee be replaced by someone substantially younger, O’Connor v. Consol. Coin Caterers Corp., 517 U.S. 308, 313 (1996)—is not a strict requirement. A plaintiff can produce more proba- tive evidence of discrimination, or he can show that the deci- sion to hire another member of the protected class (or in ADEA cases, a person of similar age) was a pretext to hide the discriminatory decision. See Diaz v. Am. Tel. & Tel., 752 F.2d 1356, 1359-62 (9th Cir. 1985) (Title VII); see also Diaz v. Eagle Produce Ltd. P’ship, 521 F.3d 1201, 1207-08 & n.2 (9th Cir. 2008). While I deal with the former type of evidence below, there is no evidence of the latter grand conspiracy, that candidates of the same age as Shelley were given interviews to hide the already-made decision that Marsh would be given the position. Because age must be the but-for cause, see Gross, 129 S. Ct. at 2350, the fact that the committee ranked other similarly-aged applicants higher than Shelley suggests that even without any alleged age bias, he still would have not received the job because another similarly-aged applicant would have taken it instead.

Since the Corps interviewed similarly situated candidates, but not Shelley, the only plausible conclusion from this set of facts is that some reason other than age caused the selection committee to decide not to interview Shelley. And if the com- mittee had some reason other than age—indeed, if it had any other reason—then Shelley cannot satisfy Gross’s “but-for” test.

III

The majority finds that “Shelley presented direct evidence of age discrimination to rebut the Corps’ purported non- discriminatory reason.” Maj. Op. at 160. The majority is just wrong on all accounts. SHELLEY v. GEREN 171 First, the majority (as did Shelley) framed its theory of the case in terms of the McDonnell Douglas burden-shifting anal- ysis. See Maj. Op. at 157. But that test is used when the plain- tiff has to rely on inferences that are not based on direct evidence. As we have explained, “[w]hen a plaintiff alleges disparate treatment based on direct evidence in an ADEA claim, we do not apply the burden-shifting analysis set forth in [McDonnell Douglas].” Enlow v. Salem-Keizer Yellow Cab Co., 389 F.3d 802, 812 (9th Cir. 2004). In Enlow, we were simply following the Supreme Court’s instruction that “the McDonnell Douglas test is inapplicable where the plaintiff presents direct evidence of discrimination.” Trans World Air- lines, Inc. v. Thurston, 469 U.S. 111, 121 (1985).

Second, even if we look at the majority’s “direct evidence,” it is anything but direct. As we explained in Coghlan v. Amer- ican Seafoods Co.: “Direct evidence is evidence ‘which, if believed, proves the fact [of discriminatory animus] without inference or presumption.’ ” 413 F.3d 1090, 1095 (9th Cir. 2005) (alteration in original) (quoting Godwin v. Hunt Wes- son, Inc., 150 F.3d 1217, 1221 (9th Cir. 1998)). It “typically consists of clearly sexist, racist, or similarly discriminatory statements or actions by the employer.” Id.; see also Enlow, 389 F.3d at 812 (“Direct evidence, in the context of an ADEA claim, is defined as ‘evidence of conduct or statements by persons involved in the decision-making process that may be viewed as directly reflecting the alleged discriminatory atti- tude . . . sufficient to permit the fact finder to infer that that attitude was more likely than not a motivating factor in the employer’s decision.’ ” (quoting Walton v. McDonnell Doug- las Corp., 167 F.3d 423, 426 (8th Cir. 1999) (internal quota- tion marks omitted)).

The majority’s “direct evidence” is a single fact—that two members of the permanent position selection committee (Regional Contracting Chief Joseph Scanlan and Business Management Division Chief Kevin Brice) obtained a docu- ment called the “Capable Workforce Matrix” (the “matrix”), 172 SHELLEY v. GEREN which lists projected vacancies in what appears to be the Walla Walla Contracting Division.3 Maj. Op. at 160. The matrix lists each of the positions in the division and, among other information, when the positions were expected to become vacant due to planned departures or retirements. The matrix does not mention the name of any member of the divi- sion, but one who is familiar with the division could deduce names based on the titles listed. For instance, the matrix indi- cates that the division currently employs one Assistant Chief of Contracting, and that the incumbent is scheduled to either leave or retire in fiscal year 2006. One who is familiar with the division would know that the Assistant Chief of Contract- ing is Shelley and would know, based on the matrix, that he is scheduled to either leave or retire from his position in the Walla Walla division in 2006. From this fact, the majority deduces “an inference that they considered this information relevant to their decision.” Maj. Op. at 161 (emphasis added). But an inference, we have said quite clearly, is not direct evi- dence. Coghlan, 413 F.3d at 1095.

Moreover, even taking this evidence as circumstantial evi- dence, it doesn’t amount to a hill of beans. There is no evi- dence in the record that Scanlan or Brice actually looked at the list, that either of them tried to link up Shelley’s name to the list, or (even assuming that they did) that either took into account Shelley’s age. In fact, the only evidence in the record is to the contrary. Scanlan testified that although he probably received the matrix, he did not recall seeing it; both testified that age was not a consideration. Nothing indicates that this request was anything out of the ordinary. Scanlan and Brice were entitled to these documents by virtue of their positions. Scanlan testified that “the matrix is a working document used for regional workforce planning purposes[, a]nd it is updated periodically for workforce planning purposes”; all of the dis- 3 For the reasons set out above, this piece of evidence is irrelevant to the 120-day position and only potentially probative for discrimination in the selection for the permanent position. SHELLEY v. GEREN 173 tricts, not just Walla Walla, were periodically to provide that information for use in projecting future requirements for dif- ferent types of positions. So even if Scanlan or Brice had determined Shelley’s retirement date from the matrix, that fact, without more, is irrelevant. And because Shelley’s super- visor, Oberle, was on the same sheet and as readily identifi- able as Shelley, there is no reason to think that they would not also deduce her age and discriminate against her—which they did not, because she received an interview.

Furthermore, as the majority concedes, all of the applicants had submitted resumes from which their ages could have been estimated, Maj. Op. at 149-150, and Scanlan had worked with Shelley for a number of years, so it would not be surprising if he knew Shelley’s age. Indeed, the majority holds that Shel- ley has proved a prima facie case under McDonnell Douglas, which in the context of an ADEA claim means that Shelley has shown that “the promotion was given to a substantially younger person.” Maj. Op. at 158. Thus, the majority began from the premise that everyone knew that Marsh was younger than Shelley. But aside from the bare fact of knowing Shel- ley’s age, there are no statements by Scanlan or Brice, no emails, and no off-hand remarks to Shelley or others about Shelley’s age. Yet “[w]ithout more, . . . the fact that [Marsh] was younger than [Shelley] does not create a triable issue of pretext.” Pottenger v. Potlatch Corp., 329 F.3d 740, 748 (9th Cir. 2003). If mere awareness of an applicant’s age is direct evidence of discrimination sufficient to show pretext, no disappointed-applicant-turned-plaintiff need ever worry about summary judgment again. Indeed, even in cases in which employers have not only noticed but also commented in potentially negative ways about a plaintiff’s age, we have found that this weak evidence is insufficient to justify a trial. See id. at 747 (compiling cases).

IV

The majority works very hard to come up with indirect evi- dence of pretext. Shelley may establish pretext “indirectly, by 174 SHELLEY v. GEREN showing that the [Corps]’s proffered explanation is ‘unworthy of credence’ because it is internally inconsistent or otherwise not believable, or [ ] directly, by showing that unlawful dis- crimination more likely motivated the employer.” Chuang v. Univ. of Cal. Davis, Bd. of Trs., 225 F.3d 1115, 1127 (9th Cir. 2000) (quoting Godwin, 150 F.3d at 1220-22). Because Shel- ley has no direct evidence, the circumstantial evidence upon which he relies to refute the Corps’s proffered explanation must be “ ‘specific’ and ‘substantial’ to create a genuine issue of material fact.” Cornwell v. Electra Cent. Credit Union, 439 F.3d 1018, 1029 (9th Cir. 2006) (quoting Godwin, 150 F.3d at 1222). What the majority comes up with is strict scrutiny of Shelley’s and Marsh’s resumes and a conclusion that “Shelley’s resume demonstrated sufficient qualifications that a reasonable jury could find that he was substantially better qualified than Marsh.” Maj. Op. at 163.

The majority did not articulate the correct legal standard. While we have held that a “district court’s finding that a Title VII plaintiff’s qualifications were clearly superior to the qual- ifications of the applicant selected is a proper basis for a find- ing of discrimination,” Odima v. Westin Tucson Hotel, 53 F.3d 1484, 1492 (9th Cir. 1995), we have yet to articulate a precise standard in the summary judgment context. In Raad v. Fairbanks N. Star Borough School Dist., 323 F.3d 1185 (9th Cir. 2003), we declined to establish the high hurdle that the Fifth Circuit adopted in Odom v. Frank, 3 F.3d 839, 847 (5th Cir.1993) (requiring that the “disparities [be] so apparent as virtually to jump off the page and slap us in the face”). 323 F.3d at 1194. The Supreme Court also thought the standard in Odom was “unhelpful and imprecise,” though it did not choose to give its own articulation of the correct standard. Ash v. Tyson Foods, Inc., 546 U.S. 454, 457 (2006). Thus, we are left with the notion in Raad that a “pronounced difference” in qualifications can be enough, 323 F.3d at 1194, but without more guidance on what lesser quantum would also be suffi- cient. SHELLEY v. GEREN 175 Yet, surely situations in which the qualifications are so sim- ilar that they could easily be thought to be equal cannot justify a trial. More importantly, it cannot be that the standard is that a reasonable jury could find that one applicant is more qualified—however slightly—than another; we cannot ask the jurors which of two candidates they prefer. Rather, it must be that a reasonable jury could think that there is such a disparity in their qualifications that the choosing of Marsh over Shelley is only explainable because of the differences in their age. This is a higher threshold than the majority’s new formula- tion. See Ash, 546 U.S. at 457 (“Under this Court’s decisions, qualifications evidence may suffice, at least in some circum- stances, to show pretext.”); Cornwell, 439 F.3d at 1033 (“[A] reasonable jury might also view the disparity between Corn- well’s management experience and Hall’s as proof that Defendants’ explanation for Cornwell’s demotion was a pre- text for race discrimination.”); Raad, 323 F.3d at 1197 (“[T]he fact that an employer hired a far less qualified person than the plaintiff naturally gives rise to an inference that the non-discriminatory explanation offered by the employer is pre- textual.”).4

Shelley cannot show that his qualifications were so clearly superior to Marsh’s that a jury could reasonably find that Marsh was promoted over Shelley on account of age. In fact, it is far from clear that Shelley was even marginally better qualified than Marsh. Shelley relied on no evidence other than his own declaration and his resume. He offered no expert wit- nesses who could evaluate the very technical language of con- tracting and procurement hierarchy, he proffered no colleagues who thought that he was better qualified, and he couldn’t point to any irregularities in the selection process. 4 I note that in these cases the standard of proof is not as demanding as in this ADEA case. Because these cases were brought under Title VII and similar causes of action, demonstrating mixed motives would have been sufficient in each; the plaintiffs did not need to prove but-for causation. See, e.g., Desert Palace, Inc. v. Costa, 539 U.S. 90, 94 (2003). 176 SHELLEY v. GEREN All we have is Shelley’s own opinion of his relative qualifica- tions. And because he has no other evidence to support his claim of age discrimination, Shelley rests his entire case on the theory that his qualifications are so vastly superior to Marsh’s that there is no other explanation for the selection committee’s decision other than the disparity in their ages. With all due respect, we have no business substituting our judgment for the selection committee. And our judgment does not improve by inviting jurors to decide which of the two they would have hired.

According to the selection criteria, the selection committee was supposed to evaluate resumes according to four criteria: (1) technical competency, (2) management, (3) leadership, and (4) teamwork. With regard to technical competency, the criteria required the committee to consider four factors: dem- onstrated knowledge of federal contracting regulations, expe- rience in overseeing multimillion dollar contracts, knowledge and experience in applying federal regulations, and experi- ence in developing contract strategies to support large acquisi- tion programs. The selection criteria add that an outstanding candidate should possess a graduate level degree.

Shelley had two things going for him relative to Marsh. He had more experience—Shelley had 29 years of contracting experience, while Marsh had 21 years of experience—and Shelley possessed an MBA from Northwest Nazarene Univer- sity. The MBA degree was a plus for Shelley, but the years of experience tell us nothing. Both candidates had at least 20 years of experience. We have no basis for deciding that Shel- ley’s additional years made him a superior candidate; if so, then every employer must promote on the basis of years of experience alone. But every employer knows that mere years in service is not a perfect proxy for competence. And indeed, if one were to rely on “experience” alone, Oberle would be an obvious choice over Shelley because she already had the very same position in Walla Walla as was being offered in Kansas City. SHELLEY v. GEREN 177 More importantly, the technical competency criterion cal- led for the candidates to have “[e]xperience in overseeing multi-million dollar contracts.” While Shelley’s resume lists “supervis[ing]” and “coordinat[ing]” contracts as among his responsibilities, he failed to list any specifics or examples of the contracts that he managed. In his declaration prepared for this lawsuit, Shelley claimed that he had supervised “multi- million dollar military construction contracts” in Afghanistan, a $100,000,000 power house at Minidoka Dam for the Bureau of Reclamation, multi-million dollar dam projects in the Pacific Northwest, a $200,000,000 modernization project at McNary Dam, and other contracts exceeding $10,000,000 on the Columbia and Snake rivers. All of that is very impressive. Unfortunately, none of it was on Shelley’s resume. For exam- ple, Shelley testified as to the projects he worked on in Afghanistan, for which he was awarded the Bronze Star.

Q. Are those—are those examples listed in your resume?

A. They’re referenced in my Bronze Star award. Not the contracts specifically.

Q. But if I look at the passage on the Bronze Star award will I see any reference to those specific projects?

A. I don’t think so.

Q. Should the panel have considered those projects when they were looking at your resume?

A. Yes.

Q. And they have done that by finding it out from something other than your resume?

A. Yes. 178 SHELLEY v. GEREN He didn’t fare any better on the other projects he claimed in his declaration. Here is the testimony on the Minidoka Dam project:

Q. Is that example listed anywhere in your resume?

....

A. No.

He was asked about his experience with the dams in the Pacific Northwest:

Q. Where in your resume does it mention . . . the complex earth dams throughout the Pacific Northwest? Or the Grand Coulee Dam?

A. I don’t list them specifically.

His Walla Walla experience didn’t show up either:

Q. Where is that in your resume?

A. It’s not specifically stated, but it’s implied with my unlimited warrant in my resume.

By contrast, Marsh’s resume specifies that he “[s]erved as the Contract Administrator for the $1.3 billion Health Care Delivery and Administrative Support Services contract.” The majority’s answer for Shelley’s failings in this regard is to explain that Scanlan, who served on the selection committee, was already personally “familiar with Shelley’s experience and credentials.” Maj. Op. at 163. But we have no idea whether Scanlan knew all the details Shelley omitted, and suggesting that one member of the selection committee “knows your record” is not the same thing as submitting a complete application to the committee. The decision not to interview Shelley was not up to Scanlan alone—it was up to SHELLEY v. GEREN 179 a five-member selection committee, of which Scanlan was just one member. As Colonel Rossi, who chaired the selection committee, explained, each member of the selection commit- tee independently reviewed resumes from dozens of appli- cants before convening via teleconference to decide on who to interview. Given the structure of the application process, Shelley should not have expected Scanlan to fill in the blanks in his resume if he failed to submit a complete resume in the first place; it is unreasonable for him to expect the remaining members of the committee to rank him based on information he didn’t supply. In sum, though Shelley claimed to have had experience managing large contracts, there is no way to tell based on his resume, which offers only general descriptions of his past experience.

With respect to management and leadership, the selection criteria emphasize experience serving as branch or section supervisor with over 30 employees, managing “large, multi- disciplined” organizations and overseeing the execution of multi-million dollar contracts. Here, the differences between the candidates’ resumes really show. At the time he applied, Shelley was Assistant Chief in the Contracting Division, a GS-13 position. He previously had positions as a Team Leader, Supervisory Contract Specialist, and Contract Spe- cialist. Although the criteria specifically mentioned that the candidates should be “branch or section supervisor . . . with over 30 employees,” Shelley did not list the number of employees he supervised in any of his positions. By contrast, Marsh was a Supervisory Procurement Analyst, which was a GS-14 position, and he stated that he supervised 15 employ- ees. In his previous position as Supervisory Contract Special- ist, he had supervised 75 contract specialists, including 14 contracting officers. And prior to that he had supervised 13 German and American contract specialists and three contract- ing officers in Germany.

The majority recognizes that Shelley didn’t respond directly to the criteria, but props him up anyway: “While 180 SHELLEY v. GEREN Shelley’s resume did not specify the number of employees he supervised, it did disclose that, as Assistant Chief of Contract- ing, he supervised, coordinated and managed the work of sub- ordinate Team Leaders, who presumably led teams.” Maj. Op. at 179-180 (emphasis added). Unfortunately, it is a big pre- sumption. Shelley stated that, around 1985, he was temporar- ily made the Chief of the Walla Walla Division. Here is his testimony:

A. How long have you served as a Division Chief, how many days?

Q. I can’t even number them on my head . . . . I was constantly made the Division Chief.

....

Q. Is that noted on your resume?

....

A. I don’t see that I covered that in there. . . . They wouldn’t have seen it from my resume.

....

Q. So is there anywhere in your resume where it reflects that you acted as a Division Chief for any significant period of time?

A. I don’t see it in there.

To counteract the failings in Shelley’s resume, the majority then disparages Marsh’s resume because his supervising 75 employees was not part of his regular job duties. See Maj. Op. at 162. From all of this, the majority calls the round a tie because “[n]either candidate clearly demonstrated that he met this criterion.” Id. at 162. SHELLEY v. GEREN 181 The majority is wrong, of course. The most notable distinc- tion between the two resumes was that Shelley was not the head of his office and Marsh was. That inconvenient fact is also reflected in one other critical fact: Marsh’s demonstrated competence had been rewarded with a GS-14 position, while Shelley was still a GS-13. The majority’s answer to this is incomprehensible. It says that this was “relevant only as to the 120-day position” because, if Shelley had been given the 120- day position, he too would have been a GS-14. Id. at 163. The majority has missed the whole point: Marsh was already a GS-14 when the 120-day position opened; even though he was younger than Shelley and had fewer years with the Corps, he held a higher position, at least as measured by his supervi- sory responsibilities and his pay grade. See Pottenger, 329 F.3d at 748 (“Nor does the fact that the company moved a younger employee ahead of Pottenger on the CEO successor list suggest that [the company] acted with any discriminatory motive, for that employee had held a higher position in the company than Pottenger.”).

With regard to the final criterion, teamwork, the criteria emphasize the ability to work with customers and other departments, offices, and teams in a multi-disciplinary setting. Marsh’s resume lists relevant experience such as serving as his directorate’s point of contact with Congress and the Army Audit Agency, in addition to maintaining working relation- ships with counterparts throughout the Department of Defense, U.S. Small Business Administration, and other fed- eral agencies. Shelley’s resume, on the other hand, fails to list any comparable experience. The majority opinion is just silent on this criterion.

From all of this, the majority deduces that “a reasonable jury could find that [Shelley] was substantially better quali- fied than Marsh.” Maj. Op. at 163 (emphasis added). The majority thus sides with Shelley, who bitterly claims that “Marsh should have received a ‘minimally acceptable’ evalu- ation.” But repeating it does not make it so. At the very least, 182 SHELLEY v. GEREN Marsh had a resume equal to or better than Shelley’s. And these were not the only two candidates. The selection commit- tee was responsible for ranking 32 applicants, among whom were a number of qualified individuals, including Shelley’s supervisor. Based on the relevant resume screening criteria, along with the fact that Marsh was already a Supervisory Pro- curement Analyst, there is no evidence to show that Shelley was “substantially better qualified” than Marsh. And without that evidence, Shelley has nothing to show that the Corps’s explanation—that the selection committee thought there were six candidates better qualified than Shelley—was pretextual.5

Indeed, the majority overlooks a straightforward conclusion based on evidence that it discusses. In the panel members’ ini- tial rankings—which were done individually and prior to any discussion with other panel members—Marsh received four top rankings and one mid ranking; Shelley received three top rankings and two mid rankings. Even assuming that Scanlan was the one who gave Marsh a top ranking and Shelley a mid ranking, that means that the other panel members indepen- dently came to the conclusion that Marsh was at least as qual- ified as Shelley. Furthermore, a different panel member changed his score for Shelley (downgrading him from mid to bottom). Even if this change came after the panel members’ discussion—in which the evidence only shows that age was never discussed and neither was Shelley in particular—that means that at least two panel members concluded that Marsh was a better applicant than Shelley. Thus, even if Scanlan did have animosity toward Shelley based on his age (for which there is no evidence whatsoever), Marsh would still have been strongly preferred over Shelley by the committee as a whole. 5 It should also be noted that, for the same reasons that Marsh demon- strated better qualifications for the permanent position, he also demon- strated better qualifications for the 120-day position. In any case, Marsh already held the same job title as the 120-day position. Thus, for neither position can Shelley rebut the Corps’s legitimate, nondiscriminatory justi- fication. SHELLEY v. GEREN 183 There is more than sufficient evidence to affirm the district court’s grant of summary judgment, even without considering the Gross “but-for” test. See Coleman v. Quaker Oats Co., 232 F.3d 1271, 1285 (9th Cir. 2000) (holding that comments related to age, the plaintiff’s own evaluations of his qualifica- tions, and the use of subjective evaluations for promotion were still insufficient to raise an issue of fact concerning dis- criminatory motive). Once we factor Gross into the mix, it is apparent that Shelley cannot show that the Corps’s decision is unexplainable on any basis other than age discrimination.

*****

There is not only no evidence of age discrimination against Shelley, there is no evidence of age discrimination in favor of Marsh. On this record, Shelley has failed to satisfy even his minimal burden of showing that his age “actually played a role in [the Corps’s decisionmaking] process and had a deter- minative influence on the outcome,” Reeves v. Sanderson Plumbing Prods., Inc., 530 U.S. 133, 141 (2000) (quoting Hazen Paper Co. v. Biggins, 507 U.S. 604, 610 (1993)), much less advance any evidence of “but for” causation for his ADEA claim. Accordingly, I would affirm the grant of sum- mary judgment.