Reset A A Font size: Print

United States Court of Appeals, First Circuit.

Darry Mason HENDERSON, Plaintiff, Appellant, v. MASSACHUSETTS BAY TRANSPORTATION AUTHORITY, Defendant, Appellee.

No. 19-1720

Decided: September 30, 2020

Before Torruella, Lynch, and Barron, Circuit Judges. James R. Tewhey, Sudbury, MA, with whom Michelle Carnevale, Salem, MA, was on brief, for appellant. David J. Santeusanio, Boston, MA, with whom Andrew E. Silvia and Holland & Knight LLP were on brief, for appellee.

Plaintiff Darry Mason Henderson, a black male and Massachusetts Bay Transportation Authority (“MBTA”) foreman, brought claims of racial discrimination, unlawful retaliation, and negligent infliction of emotional distress against the MBTA. The racial discrimination claim stems from Henderson's unsuccessful application for promotion to two MBTA supervisor positions in September 2012. The positions went to two white persons who received higher interview rankings than Henderson. In fact, Henderson's rankings placed him nineteenth out of the twenty candidates interviewed. Henderson claims that the MBTA did not select him because of his race. Henderson also contends that the MBTA retaliated against him by no longer assigning him podium duty, a type of work assignment, because he complained of racially motivated verbal abuse by a supervisor. The district court granted summary judgment to the MBTA on all three claims.

On appeal, Henderson challenges the grant of summary judgment on the racial discrimination and retaliation claims. Both challenges are meritless. We affirm.


A. Facts

Henderson began working at the MBTA as a construction laborer in 1991. In 1995, he was promoted to the position of laborer foreperson, in which he led a two- to four-person crew that identified and fixed maintenance issues. He also served as a temporary-change supervisor (“TC supervisor”) from 2000 to 2005. A TC supervisor fills in for a permanent supervisor when a supervisor is absent (typically when a supervisor is on leave) or when a supervisor position is vacant and has not yet been permanently filled. When he was a TC supervisor, Henderson supervised carpenters, roofers, laborers, and cement finishers. He supervised up to thirty of these employees at any given time. During this time, Henderson unsuccessfully applied to two permanent supervisor positions.1 In 2005, he returned to his former role as a laborer foreperson. By the time he applied for the permanent supervisor position in 2012, he had not performed supervisory duties in seven years. Further, he had applied for two permanent supervisory positions in 2005, and others were found to be better qualified for both positions.

1. The Hiring Process

In November 2011, the MBTA posted openings for two permanent supervisor positions of Building and Station Maintenance. The “minimum entrance requirements” (“MERs”) for the positions included: a high school diploma or GED; at minimum five years' work history in building and equipment maintenance; supervisory experience; the ability to use Word, Excel, Database, PeopleSoft, or Mainframe applications; effective organizational skills; and passing a Criminal Offender Record Information (“CORI”) check, background check, and medical screening. According to the longstanding practice of the MBTA, which it uniformly applies, the MERs fell into two categories: those that an application and resume must demonstrate for an applicant to receive an interview, and those that are not required for an interview but that an applicant must have to receive the job. An applicant need not demonstrate the ability to use the computer applications listed in the computer skills MER to receive an interview, as it fell into the latter category. Two of the listed computer applications were proprietary to the MBTA and outside applicants were not expected to know those programs. A temporary hiring freeze at the MBTA delayed the hiring process for the permanent supervisor positions.

Before restarting the hiring process the following summer, Steven Emde, an MBTA human resources (“HR”) staffing manager, met with current employees to recruit internal applicants.2 Both Henderson and Gilcoine were at this meeting. After Gilcoine told Henderson she would be on the positions' selection committee, Henderson expressed to Emde his concern that Gilcoine's longtime friendship with one of the other applicants, painter foreperson Bernadette Higgins, would make Gilcoine's participation unfair. Emde notified his superior about Henderson's concern and they removed Gilcoine from the committee. Emde did not notify any other members of the committee or anyone else involved in the hiring process about Henderson's concern.

On September 5, 2012, the permanent supervisor positions were reposted. The application form for the positions asked roughly twenty questions, including about an applicant's personal information, education and skills, work history, past or current work at the MBTA, and professional references. The application form's “Availability & Eligibility” section asked specific yes-or-no questions about whether the applicant could work all days and all shifts, was at least eighteen years old, and was “legally eligible” to work in the United States. The “Education & Skills” section included space for the applicant to list any schools the applicant had attended, and stated “Please list any additional education or training relevant to this position” and “Please list other skills, including computer or language skills, that are relevant to this position.”

To decide whom to invite to interview, Emde considered four MERs: high school diploma/GED, relevant work history, supervisory experience, and, for internal applicants, a satisfactory work record for the past two years. At this point, Emde did not consider the effective organizational skills, computer skills, CORI check, background check, or medical requirements MERs. Emde reviewed the applications for the four screening MERs to decide whom to invite for an interview. For certain MERs, like the computer skills MER, no answer was required by the form. The applicants were free to leave certain sections unrelated to the four screening MERs unanswered, like the other relevant skills “including computer or language skills” section. The application did not ask applicants to answer whether they met the computer skills MER or which of the relevant computer applications they had the ability to use.

The applications and resumes of the selected applicants were made available to the selection committee before the interviews. The committee was given a scoring sheet with the questions to be asked. This was longstanding MBTA practice and was uniformly applied. Each member of the committee used the same form, which required each member to post numeric scores for the candidates' answers to each question. Emde also provided the selection committee with an “answer key” to the interview questions, which expressly focused on many of the MERs, including the effective organizational skills and computer skills MERs. The answer key provided the interviewers with the same important elements to look for in the candidates' answers, which allowed the committee to take a more objective approach in its scoring.

One hundred nineteen people applied for the two positions, including Henderson and Higgins. Higgins listed Gilcoine as a reference. William Melchionda, a white male who had not worked for the MBTA but had twenty years of construction and building maintenance experience, also applied.3 Melchionda did not fill in the section of his application regarding “other skills, including computer or language skills, that are relevant to this position.”

Twenty-three applicants were invited to interview, of whom twenty accepted. The selection committee interviewed these twenty applicants over the course of six days from November 28, 2012, to December 18, 2012. The committee was composed of Emde; John Martin, a supervisor in the electrical power department; and Andrew Baker, an engineering and maintenance director.4 All three are white men. Emde has served on a couple hundred selection committees. Over the course of his career at the MBTA, Martin has served on the selection committee in twenty to thirty interview processes. At that time, Martin had about fourteen years of supervisory experience at the MBTA. In consequence, Martin understood the duties and necessary qualifications of an MBTA supervisor. It was his understanding that the hiring/promotion decision would be based solely on the candidates' interview scores. This committee interviewed Henderson and Higgins on November 28, Melchionda on December 4, and the seventeen other candidates during the interview period. Of these seventeen other candidates, two were black, one was Asian, and the rest were white.

MBTA policy states that the committee members ask each applicant a set of uniform questions and each committee member, using guidelines developed by HR, assigns a numerical score to the applicant's response to each question. During each of the candidate's responses, the interviewers take notes. The interviewers typically record the scores for the candidate's responses at the end of each interview. Generally, the members of the committee do not discuss the applicants or their answers with each other before recording scores. Under its policy and longstanding, uniform practice, the MBTA hires the applicant(s) with the highest scores, unless an applicant fails a background check or medical screening.

Before the final hiring decision is made, HR sends a summary of the hiring process and the resumes of the recommended hires to the MBTA's Office of Diversity and Civil Rights (“ODCR”), which must approve any new hire. The ODCR reviews information on the demographics of the applicants, the application process, and the resumes of the recommended candidates. It will stop the application process if it is not comfortable with the way the process is proceeding. There is no evidence that the MBTA did not follow its uniform policy and practice, and evidence that it did.5

Henderson received a total cumulative score of 117, with a sixty-one from Martin and a fifty-six from Emde, and so had the second-lowest combined score of the entire group of applicants. At least one black candidate received higher scores than Henderson did. Eighteen of the nineteen other candidates received higher scores. Higgins, who is white, had the highest score: Martin gave her a ninety-four and Emde gave her a ninety-three.6

Higgins had supervised far larger groups of MBTA employees than Henderson had and she had done so more recently as a TC supervisor. Emde considered Higgins's responses to be “the best among those interviewed and substantially better than Henderson's responses.”

For example, interview question two asked the candidate to “tell us about your experience managing large maintenance and/or construction projects” and “what your responsibilities were during these projects.” In response, Higgins stated that she had managed multiple MBTA paint shop shifts for the past decade and had personally managed the MBTA's Revive & Guide program and its fifty-four employees while ensuring it was completed on time.

As another example, question nine asked: “Have you ever had to discipline an employee? Please tell us about your experience in handling employee discipline and include in your answer if the discipline was defined by a union agreement.” Higgins spoke about her investigation of an employee's attendance policy violation in which she followed union procedures and issued the employee a written warning.

Melchionda had the next highest score, with scores of 92 from both Martin and Emde. Melchionda had supervised a significantly larger group of employees than Henderson had: sixty, rather than thirty. His supervisory experience was not only more significant but more recent. Melchionda had independently managed complex construction projects and had experience with managing a unionized crew.

In response to question two, Melchionda gave a detailed explanation of his supervision of sixty employees in a unionized production and manufacturing facility, his management of a $1.5 million roofing project, and his experience leading a thirteen- to twenty-eight-person crew that maintained twenty-one properties. As to question nine, Melchionda spoke about his experience supervising unionized labor with the progressive discipline outlined in the union contract, and his conflict resolution preferences.

Question four of the interview asked:

Please tell us about your computer skills? [sic] Have you ever used and [sic] Asset Management system? What would be the value of having a computerized system that tracks the status of needed repairs?

This question was, in effect, composed of two questions: The first question addressed a candidate's computer skills and experience with an Asset Management system. The second question addressed a candidate's views on and understanding of a computerized repair-tracking system. Only the first question was covered by a MER. Martin and Emde each gave Henderson a score of four and Melchionda a score of seven for question four. The committee notes show that Henderson's response focused on his experience with word processing software and “Blue Zone” software and he mentioned that a computerized system would “overcome gaps to get trades together.” The notes also imply that he repeatedly stated that such a system “would be great.”

The notes show that Melchionda stated he had “minimal [computer] use [at] this time,” but had “operational exp[erience]” using computer systems with maintenance logs and work orders. The notes also state that Melchionda used a computer at home and “everyday.” As to the second computer question, the notes reflect that Melchionda stated that a computerized system would “save[ ] time,” be “more efficient,” and “keep records.”

After scoring the applicants, HR contacted David Benson, one of Melchionda's references. Benson said: “[Melchionda's] been gone for a long time and was my foreman. He moved on to better/greener pastures. Got to go now.”

In January 2013, the selection committee recommended Higgins and Melchionda for the positions, as they had the highest scores. Its recommendation was referred to the ODCR, which approved of hiring Melchionda and Higgins. In February, Higgins and Melchionda were informed that they had received the positions. On February 28, Henderson was told that he was not selected. He has continued in his same role at the MBTA and Melchionda is now his supervisor. After Melchionda was hired, Henderson taught him about the MBTA and its policies, procedures, and internal proprietary computer system. Melchionda later stated at deposition that the MBTA-specific computer applications Henderson taught him were “all very basic” and not hard to learn.

2. The Retaliation Claim

On October 12, 2012, the MBTA did not assign Henderson to “podium duty,” and instead chose another laborer foreperson. Those assigned to podium duty help set up the podium and speakers for government officials giving speeches. The duty often results in overtime pay. The union steward typically decides who receives such duties, but Gilcoine had some influence over the assignment. Henderson had worked podium duty for the five years before October 2012 but has not been assigned it since. He has, however, been assigned other overtime duties at the same rate as podium duty since October 2012. Podium duty also does not always result in overtime pay.

Henderson claims that Gilcoine yelled at him for using an MBTA computer in Charlestown to fill out paperwork on January 31, 2013. Specifically, he states that Gilcoine yelled: “Get out of the office, Darry. I don't want you in there.” Other employees often used this office as well. Henderson claims that he discussed this incident with Sayten Patel,7 a deputy director in their department, about a week later. Henderson mentioned that he thought it was due to his race. He claims that Patel said he would speak to Gilcoine about the incident. Gilcoine testified that this incident never occurred and Patel testified that he and Henderson never spoke about the incident.

3. The MCAD Complaint

On September 20, 2013, Henderson filed a charge of discrimination and retaliation with the Massachusetts Commission Against Discrimination (“MCAD”). The Equal Employment Opportunity Commission sent Henderson a right to sue letter on August 4, 2017.

B. Procedural History

On October 26, 2017, Henderson sued the MBTA in the U.S. District Court for the District of Massachusetts. He alleged racial discrimination in violation of Title VII, retaliation in violation of Title VII, and negligent infliction of emotional distress. After discovery concluded, the MBTA moved on February 8, 2019, for summary judgment on all counts. Henderson opposed summary judgment on the discrimination and retaliation claims, but conceded that the claim for negligent infliction of emotional distress was meritless.

Henderson argued that the MBTA hired both Higgins and Melchionda instead of him because of his race and that the MBTA's stated reliance on the interview scores to make the hiring decision was pretextual. Henderson also argued that, in response to his complaint to Patel, Gilcoine retaliated against him by no longer assigning him podium duty. The district court concluded both arguments were meritless. The district court reasoned that, although it read the record as showing that Henderson may have had better computer skills and that there was some subjectivity in the interview process, the MBTA's use of the highest scores after the interviews was not shown to be pretextual. The court also concluded that Henderson did not offer any evidence that discrimination motivated the hiring decision.

The district court concluded that the retaliation claim was meritless because Henderson could not prove causation: Henderson was taken off podium duty five months before he allegedly complained of the computer incident with Gilcoine. Although Henderson argued that his continued denial of podium duty was retaliatory, he provided no evidence that either the union steward or Gilcoine knew of his alleged complaint to Patel.

On July 16, 2019, Henderson appealed.


A. Standard of Review

We review the grant of summary judgment de novo. Theidon v. Harvard Univ., 948 F.3d 477, 494 (1st Cir. 2020). “We view the record in the light most favorable to the nonmoving party and make all reasonable inferences in that party's favor.” Johnson v. Univ. of P.R., 714 F.3d 48, 52 (1st Cir. 2013) (citations omitted). In opposing summary judgment, the plaintiff bears the burden of producing evidence sufficient to rebut the defendant's arguments but cannot rely on “conclusory allegations, improbable inferences, ․ or rank speculation.” Theidon, 948 F.3d at 494 (quoting Ahern v. Shinseki, 629 F.3d 49, 54 (1st Cir. 2010)). Similarly, the plaintiff's “subjective belief of discrimination is not sufficient to withstand summary judgment.” Tyree v. Foxx, 835 F.3d 35, 42 (1st Cir. 2016).

B. Henderson Has Not Shown Pretext and Discriminatory Motivation as to the Hiring/Promotion Decision

Henderson argues he has produced sufficient evidence to get to a jury on his claim that he was denied a promotion based on his race. Henderson provides no evidence to support his assertion that the MBTA deviated from its policy in that he would not have been given an interview had he left blank the computer skills portion of the application form. He argues an inference of discrimination can be drawn from the decision to interview Melchionda despite Melchionda leaving blank the question as to the computer skills on the application form, and from the higher scores given to Melchionda on that particular question after Melchionda explained his computer abilities at the interview. He acknowledges that he has no direct evidence that any member of the interviewing team or the reviewing team was motivated by racial animus against him. His argument, in essence, is that because he was a long time MBTA employee who had five years of experience as a TC supervisor ending some six to seven years before, the MBTA's decision not to promote him must have been racially motivated. He supports this contention with evidence not specific to him or to the promotion process used here, but to what he characterizes as a sorry history of racism at the MBTA. He does not respond to the MBTA's argument that even if Henderson had been given a higher score on that question, Henderson still would not have received the promotion given his being next to last on the scores.

Henderson's claim relies on indirect evidence, and so we apply the McDonnell Douglas burden-shifting test. McDonnell Douglas Corp. v. Green, 411 U.S. 792, 802-04, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973); see also Thompson v. Coca-Cola Co., 522 F.3d 168, 176 (1st Cir. 2008). Henderson first bears the burden of establishing a prima facie case. To do so, he must show that (1) he is a member of a protected class, (2) he was qualified for the position to which he applied, (3) he was not hired, and (4) an applicant with similar qualifications received the position. See Goncalves v. Plymouth Cty. Sheriff's Dep't, 659 F.3d 101, 105 (1st Cir. 2011).

The burden of production then shifts to the MBTA, which must “present a legitimate, non-discriminatory reason” for choosing Melchionda over Henderson. See Thompson, 522 F.3d at 176 (quoting Quinones v. Houser Buick, 436 F.3d 284, 289 (1st Cir. 2006)).

If the MBTA satisfies this burden, Henderson must “show by a preponderance of the evidence that the [MBTA's] proffered reason is pretextual and that the actual reason for [not promoting him] is discriminatory.” Johnson, 714 F.3d at 54; see also Pearson v. Mass. Bay Transp. Auth., 723 F.3d 36, 40 (1st Cir. 2013) (Souter, J.) (“To defeat summary judgment, [the plaintiff] must offer ‘some minimally sufficient evidence, direct or indirect, both of pretext and of [the MBTA's] discriminatory animus.’ ” (second alteration in original) (second emphasis added) (quoting Acevedo-Parrilla v. Novartis Ex-Lax, Inc., 696 F.3d 128, 140 (1st Cir. 2012))). To show pretext, Henderson must do more than just allege he was more qualified: “[I]n the absence of strong objective evidence (e.g., test scores), proof of competing qualifications will seldom, in and of itself, be sufficient to create a triable issue of pretext.” Rathbun v. Autozone, Inc., 361 F.3d 62, 74 (1st Cir. 2004). He has not met this burden.

Henderson independently has not met his burden to produce “evidence ․ permit[ting] a factfinder reasonably to infer that unlawful discrimination was a determinative factor in the [MBTA]'s decision.” Feliciano de la Cruz v. El Conquistador Resort & Country Club, 218 F.3d 1, 8 (1st Cir. 2000). Henderson must show that the MBTA's motivation was “unlawful,” not merely “inappropriate.” Ahmed v. Johnson, 752 F.3d 490, 498 (1st Cir. 2014).

As we often do, “[w]e ․ ‘bypass the prima facie case issue,’ ” and get to the question of whether Henderson's evidence of pretext is sufficient. Luceus v. Rhode Island, 923 F.3d 255, 258 (1st Cir. 2019) (first alteration in original) (quoting Cham v. Station Operators, Inc., 685 F.3d 87, 95 (1st Cir. 2012)); accord Espinal v. Nat'l Grid NE Holdings 2, LLC, 693 F.3d 31, 35 (1st Cir. 2012).

The MBTA's stated reason for not hiring Henderson was that he “performed poorly on the interview -- so poorly that his scores placed him 19th of the 20 candidates to receive an interview.” The MBTA said that it followed its “hiring practices” and selected Melchionda and Higgins because they “received the top two scores of all interviewees.” Henderson concedes that his interview score was ranked nineteenth of twenty and that Higgins's and Melchionda's scores were the two highest. The MBTA also stated that “Melchionda's and Higgins's high interview scores reflect their better qualifications and that they more effectively communicated their experience and credentials.” The record supports this legitimate, nondiscriminatory reason, which satisfies the MBTA's burden of production. See Hicks v. Johnson, 755 F.3d 738, 744, 746-47 (1st Cir. 2014). Henderson has not produced sufficient evidence that this nondiscriminatory reason was pretextual to avoid summary judgment.

1. The MBTA Did Not Apply a Racially Differential Criteria as to the Computer Skills MER

Henderson first argues that the MBTA applied racially discriminatory standards to Melchionda's benefit throughout the hiring/promotion process.8 He contends that Melchionda's application did not satisfy the computer skills MER so Melchionda should not have received an interview. Henderson also argues that Melchionda's interview responses did not show he had the requisite computer skills, that he did not satisfy the computer skills MER, and so should not have been eligible to be hired. Henderson asserts that the MBTA applied different standards as to the computer skills MER and did so because Melchionda is white. For several reasons, these arguments lack merit.

The MBTA did not apply a differential standard by inviting Melchionda to interview. Henderson contends that an applicant must meet all of the MERs to receive an interview. The record refutes this assertion.

An applicant only needs to meet some of the MERs to receive an interview. The MBTA had a consistent policy and practice of using some MERs for one reason and others for a different reason to be satisfied only before final hiring. Some MERs were listed to provide notice to applicants that certain skills were requirements of the job and that, to carry out the job, a successful candidate might need to use the skills listed. These MERs informed applicants that, even if they received an interview, they would still need to meet those MERs to be hired. Henderson also admits that the interview process was as the MBTA described it, in that some MERs would be assessed after the decision to extend an interview offer. The questions in which an applicant might choose to address certain MERs, including the computer skills MER, were not required to receive an interview. Melchionda receiving an interview without answering the question related to relevant computer skills does not evidence pretext.9

Further, because Emde did not screen any of the applicants for the computer skills MER, Henderson's contention that the MBTA applied a different screening standard to white applicants is entirely unsupported. There is no evidence that the MBTA followed a different policy about not considering this MER as to any other candidate and certainly no evidence of racial differentiation.10

The MBTA also did not apply a differential criteria by hiring Melchionda, as the committee reasonably found that his interview responses satisfied the computer skills MER. The selection committee described Melchionda as stating he had “minimal [computer] use at this time” but had (1) “operational exp[erience] on [a] P.C.” filling out “work orders” and “logs,” (2) used a computer at home, and (3) used a computer “everyday.” (Emphasis added.) There is no evidence in the record that the committee could not reasonably find these three statements to satisfy the computer skills MER or that the committee's acceptance of these statements as satisfying the MER was racially motivated. The committee reasonably inferred that Melchionda had the ability to use Word, Excel, Database, PeopleSoft, or Mainframe applications.

There is no evidence of a differential criteria applied to Melchionda in determining whether he satisfied the MERs necessary to progress in the hiring/promotion process. Further, the record contains no evidence of a differential criteria that suggests any racial considerations.

2. The Interview Process Does Not Evidence Pretext for Racial Discrimination

Henderson next attacks the interview process on which the hiring/promotion decision was solely based. The MBTA, however, may rely on interview scores in its hiring/promotion decisions.11 In Martinez-Burgos v. Guayama Corp., 656 F.3d 7 (1st Cir. 2011), we concluded that an employer may hire solely based on candidates' interview scores and that “better interview scores” may reflect an individual's superior qualifications.12 Id. at 13; see also Goncalves, 659 F.3d at 107 (concluding that a candidate's lower interview scores reflected her inferior qualifications); Prescott v. Higgins, 538 F.3d 32, 40-41 (1st Cir. 2008) (holding that an employer can judge a candidate's qualification through interview questions alone). The record shows that the interview scores reflected the candidates' qualifications, as the two highest scoring candidates were more qualified than Henderson.

Like the defendant in Hicks v. Johnson, the MBTA “took pains to standardize the interview process as well as record and quantify the candidates' performance on a uniform scale.” 755 F.3d at 747. The MBTA asked the “same ․ questions” of the candidates, worded the questions “broadly ․ as to provide [a candidate] with ample running room to tout her qualifications and experience,” and provided the selection committee with an answer key. Id. Henderson admits that the committee asked him the same questions as the other candidates and that the interview was thirty to forty minutes long (ample time for Henderson to describe his qualifications and experience).13 The ODCR review also provided another check on the interviewers' decision. “In essence, the [MBTA] made the subjective part of the [hiring/]promotion process as objective as possible ․ [and] on this record, these measures do preclude any reasonable inference that the interview process was evidence of pretext.” Id.

Relying solely on interview scores also does not, as Henderson contends, violate MBTA Policy and Procedure 2.2, which defines the MBTA's hiring policy and the procedures to be used in the MBTA's “Hiring and Selection Process.” Nothing in this policy forbids the procedure the committee used and Henderson admits this. His argument is that it should.

3. Henderson Alleges a Scoring Inconsistency that Does Not Exist and Which Still Would Not Support an Inference of Pretext for Racial Discrimination

Henderson alleges a single inconsistency between his and Melchionda's interview scores and the committee notes which, he argues, “impl[ies] that the proffered reason offered by the [MBTA] was pretext.” The alleged inconsistency is that, although each interviewer scored Melchionda three points higher than Henderson on the computer skills question, according to the district court, the committee's notes show that Melchionda “described his computer skills as ‘minimal’ ” and that Henderson “had more computer experience.” For several reasons, we reject this argument.

Henderson did not raise this argument before the district court and does not sufficiently develop this argument. In consequence, he has waived it. Arrieta-Gimenez, 859 F.2d at 1037; Zannino, 895 F.2d at 17.14

Not only did Henderson waive this argument, but it is without merit because there is no inconsistency in the interview scoring. The district court's reading of the record, in which it stated it found an inconsistency, was erroneous. The district court described the committee notes as stating that Melchionda “described his computer skills as ‘minimal,’ ” when, in fact, they show that Melchionda stated he had “minimal [computer] use [at] this time.” (Emphasis added.) There is no indication in the interview notes or elsewhere in the record that Melchionda lacked the computer skills required by the MER. The district court's conclusion that “Henderson had more computer experience than Melchionda” is similarly unsupported. The notes detail Melchionda's “operational exp[erience]” with computer systems and his use of a computer “everyday” at home. According to the notes, Henderson's response, in contrast, focused more on word processing skills and having taken a word processing class but does not show any particular expertise in word processing or other applications. The notes do not show that Henderson's computer skills were superior to Melchionda's.15

Second, the committee's scores also reflect Henderson's and Melchionda's responses to the second computer question: “What would be the value of having a computerized system that tracks the status of needed repairs?” According to the committee notes, Henderson repeated that such a system “would be great” and mentioned that it would “overcome gaps to get trades together,” while Melchionda stated that such a system would “save[ ] time,” be “more efficient,” and “keep records.” Henderson testified that, although he believed he said more in response to this second question, he could not recall what more he might have said. The notes of each of the interviewers do not in any sense purport to be verbatim. They are simply notes, not purporting to be a complete recounting of everything said in the interview. Further, the notes as to each of the candidates are consistent in merely being notes.16

There is no evidence that the committee's assignment to Melchionda of a higher score as to computer skills was either error or, if error, that the error was racially motivated. The interviewers reasonably found Melchionda's operational experience with computers to outweigh Henderson's word processing skills. Further, Melchionda's response to the second, “computerized system” question was more detailed. Weighing these experiences and responses, the interviewers reasonably gave Melchionda a higher score. See Hicks, 755 F.3d at 746 (“Weighing the value of [two interviewees' different types of] experience required the interview panel to make a judgment that it was entitled to make.”). We are left with Henderson's bald assertion that he was more qualified than Melchionda, which cannot support an inference of racial discrimination.17 Tyree, 835 F.3d at 42. Tellingly, the district court also found no evidence of racial discrimination, despite its reservations about the three-point score difference between Melchionda and Henderson. Further, even assuming dubitante that a reasonable factfinder could read the committee notes as warranting a higher computer score for Henderson, such a reading would not create a triable issue of material fact.

First, even under Henderson's proposed reading of the record, the committee's notes and computer skills scores “are not so inconsistent as to be unworthy of credence, which is the test.”18 Echevarría v. AstraZeneca Pharm. LP, 856 F.3d 119, 136 (1st Cir. 2017) (internal quotation marks omitted) (quoting Collazo-Rosado v. Univ. of P.R., 765 F.3d 86, 94 (1st Cir. 2014)).

Second, Feliciano de la Cruz v. El Conquistador Resort and Country Club, 218 F.3d 1 (1st Cir. 2000), requires our holding. In Feliciano de la Cruz, the court held that a plaintiff's offer of “thin” evidence of pretext by itself cannot defeat summary judgment if that evidence, and any other offered evidence, does not support a reasonable inference of discrimination. Id. at 8, 10. The court concluded that the plaintiff had not met this burden, explaining that she had not offered evidence of a pattern of racially disproportionate hiring or firing, evidence of discriminatory corporate policies, evidence of discriminatory statements by her company's management, or evidence that her employer's “evaluation of her performance was infected by stereotyped thinking or other types of unconscious ․ bias.” Id. at 8-9. Similarly here, Henderson conceded that “[t]here was nothing discriminatory or retaliatory about the questions the Selection Committee asked [him] in the interview” and that he did “not have any reason to believe that the individuals on the Selection Committee would discriminate or retaliate against him on the basis of his race ․ other than ․ that all three Selection Committee members were white.”19 The evidence Henderson contends shows the MBTA's discriminatory patterns, practices, and statements does not permit a reasonable inference of discrimination and so does not distinguish this case from Feliciano de la Cruz.20 See id. at 9-10. Henderson's remaining general assertions that the MBTA is a racist employer, based on his view of historical evidence, do not meet his burden to survive summary judgment. Just as the court concluded in Feliciano de la Cruz, “if we remanded for trial, the jury ‘would be left to guess at the reasons behind the pretext.’ ” Id. at 9 (quoting Medina-Muñoz v. R.J. Reynolds Tobacco Co., 896 F.2d 5, 10 (1st Cir. 1990)).

4. Henderson's Newly Raised Cronyism Argument is Waived and Wholly Unsupported

Henderson argues for the first time on appeal that cronyism affected the interview scores, Melchionda's hiring resulted from cronyism, and this cronyism was a “smokescreen” for racial discrimination. He also did not raise this argument in his appellate brief, but rather for the first time at oral argument. We reject this argument as both doubly waived and unsupported by the record. See Pignons S.A. de Mecanique, 701 F.2d at 3; Arrieta-Gimenez, 859 F.2d at 1037.

Henderson raises this new argument to try to take advantage of the district court's speculation that there was cronyism in the hiring/promotion decision. On de novo review, we find no support in the record for any such speculation and speculation is inappropriate in any event. The only evidence which Henderson has ever argued supported an inference of cronyism was that Melchionda listed Perez, the MBTA's head of HR, and Perez's brother as references. But Emde's and Martin's uncontradicted testimonies that they did not speak to Perez about Melchionda rebut such an inference.21

Moreover, the record shows that the MBTA actively combatted potential cronyism during this hiring process: First, the MBTA removed Gilcoine from the selection committee in part because of her close relationship with Higgins. Second, upon suspicion of favoritism, the MBTA excluded all of Baker's scores.

5. Henderson Has Not Shown that He Would Have Received the Position Regardless

Finally, Henderson's claim fails at a basic causation level. Henderson has not shown that, absent the alleged double-standard applied to Melchionda, he would have gotten the job. See Bostock v. Clayton County, ––– U.S. ––––, 140 S. Ct. 1731, 1739, 207 L.Ed.2d 218 (2020) (“Title VII's ‘because of’ test incorporates the ‘simple’ and ‘traditional’ standard of but-for causation ․ [which] is established whenever a particular outcome would not have happened ‘but for’ the purported cause.” (citations omitted));22 Chadwick v. WellPoint, Inc., 561 F.3d 38, 48 (1st Cir. 2009) (determining whether “a reasonable jury could conclude that the promotion denial was more probably than not caused by discrimination”). That is because there were sixteen candidates other than Higgins and Melchionda with higher scores, at least one of whom was black. Henderson scored seven points below the next two higher-scoring candidates. So even if he had scored the same on the computer skills questions as Melchionda, which would have increased his total score by six, Henderson would not have risen in the rankings. Further, Henderson scored sixty-six and fifty-four points lower than the next two highest-scoring candidates after Melchionda. Even if Henderson had received the maximum scores of ten from both interviewers on the computer skills question, and Melchionda and Higgins received the minimum score of zero, Henderson would still not have had a top-two interview score or a higher score than Melchionda or Higgins. Such a change would only have moved Henderson into sixteenth place, out of twenty interview candidates. Given his low interview ranking, Henderson cannot independently show, and has not tried to show, that the others ranked above him were chosen for discriminatory reasons and he would have gotten the job.

To the extent that Henderson argues that the alleged inconsistency in the computer skills scoring would allow a reasonable jury to infer that racial discrimination affected his scores for the other, more subjective interview questions, Henderson has doubly waived this argument by failing to raise it before the district court or to sufficiently develop it. Arrieta-Gimenez, 859 F.2d at 1037; Zannino, 895 F.2d at 17.

Henderson has not shown that, absent the alleged racially discriminatory lowering of his scores for the “subjective” questions, his total score would be higher than Melchionda's and those of the sixteen candidates who outscored Henderson but also did not receive the position.

6. The Dissent is Without Merit and Contrary to Clear Precedent

Because our holding follows directly from our holding in Feliciano de la Cruz, the dissent necessarily relies on the argument that Reeves v. Sanderson Plumbing Products, Inc., 530 U.S. 133, 120 S.Ct. 2097, 147 L.Ed.2d 105 (2000), limited that decision to its facts. But this Court has rejected that argument both in denying rehearing in Feliciano de la Cruz and in subsequent opinions. After Reeves we have continued to rely on Feliciano de la Cruz in both published and unpublished decisions. See Meléndez v. Autogermana, Inc., 622 F.3d 46, 53 (1st Cir. 2010) (citing Feliciano de la Cruz, 218 F.3d at 8) (thin evidence of pretext, without more, did not create a triable issue as to the discriminatory purpose); see also Chouinard v. N.H. Dep't of Corr., 157 F. App'x 322, 325 (1st Cir. 2005) (unpublished opinion) (“Even if these hiring decisions were irrational or unfair, that would not be the same as discrimination.” (citing Feliciano de la Cruz, 218 F.3d at 8) (other citations omitted)); Céspedes Rodriguez v. Rivera Hernandes, 135 F. App'x 441, 443 (1st Cir. 2005) (unpublished opinion) (“[Plaintiff] also highlights the evidence that his termination was based on trumped-up charges, but even if we accept for the sake of argument that there is a trialworthy issue here, [plaintiff] still must show that the pretext masked unlawful ․ discrimination.” (citing Reeves, 530 U.S. at 146-49, 120 S.Ct. 2097; Feliciano de la Cruz, 218 F.3d at 8)). Indeed, in Ronda-Perez v. Banco Bilbao Vizcaya Argentaria--Puerto Rico we observed:

Shortly after Reeves ․ we had occasion to reconsider a ruling we made in [Feliciano de la Cruz]. In an order denying panel rehearing, we held that our analysis was consistent with Reeves, and reiterated that the thinness of the plaintiff's showing of pretext ․ failed to shed any light on what the true reason [for the adverse employment action] was.

404 F.3d 42, 44 (1st Cir. 2005). The dissent's argument that Feliciano de la Cruz is “a fact-dependent ruling ․ readily distinguished from this one” is thus contrary to our established precedent. The dissent's claim that “Feliciano [cannot be] understood to state a generalizable rule” is inconsistent with our holdings in each of the cases discussed above. Nor is it compelled by the decision in Reeves. There, the Supreme Court made clear that “a number of factors” weigh on whether a showing of pretext, combined with a prima facie case of discrimination, are enough to raise a jury question about discriminatory intent. Reeves, 530 U.S. at 148-49, 120 S.Ct. 2097. Feliciano de la Cruz identifies one such set of circumstances where weak evidence of pretext is insufficient to raise a jury question as to discriminatory intent. Nor is the rule announced in Feliciano de la Cruz as amorphous as the dissent claims. It directs only that the plaintiff must present some evidence of discriminatory intent beyond a barebones allegation that the employer's stated reason for an employment decision was pretextual.

Moreover, the dissent misunderstands our reasons for finding plaintiff has not produced evidence of a racially discriminatory motive. We do not rely only on the “asserted weakness of Henderson's pretext showing.” Henderson failed to show any form of differential treatment. Nor has he come close to showing that the committee's stated reasons were not the true reasons. Nor has he shown the true reason was race discrimination.

The dissent also misunderstands the district court's misgivings as to the computer skills interview question score. Nowhere does the district court claim that the MBTA's interview scoring was “indefensible,” or “inexplicable,”23 as the dissent claims. Instead, the district court stated the three-point difference between Melchionda's and Henderson's scores was inconsistent. Indeed, it is ambiguous from the summary judgment opinion whether the district was concerned by the fact that Melchionda received a higher score than Henderson, or by the size of the difference in scores. The latter view is consistent with the district court's emphasis on the fact that Henderson received a “three points lower” score. Further, the district court concluded, as we do, that the committee's scoring of the computer skills interview question was not evidence of racial discrimination.

The district court raised cronyism as one possible non-racial explanation for the MBTA's decision to interview Melchionda. And it did so only because of a misreading of the record. The district court was concerned that Melchionda received an interview even though he did not complete the computer skills portion of the application. It is clear, however, that the MBTA did not screen applications based on that question.

Nor is there support in the record for the dissent's conclusion that Henderson objectively should have scored higher than Melchionda on the computer skills question. It is true Henderson was more familiar with the MBTA's proprietary software, and also highlighted his word processing experience before the committee. But there is no reason to conclude Melchionda did not inform the committee he was also familiar with Microsoft Office and Excel. At minimum, he told the committee he had “operational exp[erience]” with work orders and maintenance logs. Additionally, the interview notes and deposition testimony indicate Melchionda's answer to the second part of the computer skills question was more thorough than Henderson's. Melchionda listed specific advantages to a computerized system, while Henderson noted such a system would be “great” and would “get trades together.” At deposition, Henderson claimed to have said more, but was unable to recall any part of his answer not recorded in the interview notes. In these circumstances, the MBTA could reasonably conclude that Melchionda's familiarity with computers and understanding of the advantages of a “computerized system that tracks repairs” outweighed Henderson's experience with the MBTA's internal software.

Finally, the dissent addresses the fact that sixteen other unsuccessful candidates received higher scores than Henderson in a single footnote. There is no evidence in the record that Henderson would have received a higher score, and there is absolutely no evidence that he would have received a higher score than the sixteen other unsuccessful applicants, including at least one other African-American candidate. The dissent's arguments do not overcome Henderson's basic causation issue. Even if race discrimination played a role in the difference in interview scores between Henderson and Melchionda, Henderson cannot explain why the job would not have gone at the very least to another unsuccessful non-white applicant with a higher interview score than Henderson.

C. Henderson Has Not Established a Prima Facie Case of Retaliation

Henderson argues that the MBTA retaliated against him for complaining of Gilcoine's conduct by denying him podium duty assignments. This argument is meritless.24

To establish a prima facie case of retaliation, Henderson must show that “(1) [ ]he engaged in protected conduct; (2) [ ]he suffered an adverse employment action; and (3) that a causal nexus exists between the protected [conduct] and the adverse action.” Carlson v. Univ. of New Eng., 899 F.3d 36, 43 (1st Cir. 2018) (third alteration in original) (internal quotation marks omitted) (quoting Garayalde-Rijos v. Municipality of Carolina, 747 F.3d 15, 24 (1st Cir. 2014)).

Henderson did not present a prima facie case. Although his complaint to Patel constituted protected conduct,25 Henderson has not shown that not assigning him podium duty was an adverse employment action or that it was causally related to his complaint.26

The MBTA correctly argues that not assigning Henderson podium duty did not materially change his employment and so was not an adverse employment action. “An adverse employment action ‘typically involves discrete changes in the terms of employment, such as ․ reassignment with significantly different responsibilities, or a decision causing significant change in benefits.’ ” Garmon v. Nat'l R.R. Passenger Corp., 844 F.3d 307, 314 (1st Cir. 2016) (quoting Cham, 685 F.3d at 94). Denial of overtime opportunities can be a materially adverse action in certain contexts, but “must be more disruptive than a mere ․ alteration of job responsibilities.” Id. (quoting Morales-Vallellanes v. Potter, 605 F.3d 27, 35 (1st Cir. 2010)).

Henderson has not been assigned podium duty since October 12, 2012. But he concedes that he has received other overtime opportunities and does not claim he was denied any other overtime opportunities. He testified that he still received overtime opportunities and pay in 2013 and in 2018 (as of the time of his testimony). Podium duty also does not always result in overtime pay. Henderson has not shown “an actual decrease in ․ overtime opportunities” that might constitute a “materially adverse change” (instead of a mere “alteration of job responsibilities”). Id. (quoting Morales-Vallellanes, 605 F.3d at 35).

Henderson also argues that the “temporal proximity” of his complaint to Patel and the MBTA not assigning him podium duty “provide[d] the required inference of causation for a prima facie case of retaliation.” This argument also lacks merit.

“Causation moves forward, not backwards, and no protected conduct after an adverse employment action can serve as the predicate for a retaliation claim.” Pearson, 723 F.3d at 42. Temporal proximity only supports an inference of causation when the record shows “that the decisionmaker knew of the ․ protected conduct when he or she decided to take the adverse employment action.” Planadeball v. Wyndham Vacation Resorts, Inc., 793 F.3d 169, 177 (1st Cir. 2015) (quoting Pomales v. Celulares Telefónica, Inc., 447 F.3d 79, 85 (1st Cir. 2006)).

Henderson complained to Patel in February 2013, four months after he argues he was no longer assigned podium duty. This podium duty assignment cannot serve as the predicate for Henderson's retaliation claim, as it occurred before Henderson's protected conduct. See Pearson, 723 F.3d at 42.

To the extent Henderson argues that the MBTA retaliated against him by continuing to deny him podium duty after he complained to Patel, the argument also fails. Henderson does not point to any evidence that Gilcoine or the union stewards responsible for assigning podium duty knew of his complaint to Patel. See Planadeball, 793 F.3d at 177.



Darry Henderson has worked at the Massachusetts Bay Transportation Authority (“MBTA”) for more than two decades. During most of that time, he has served as a laborer foreperson, managing small crews of workers as they go about doing repairs and other jobs for the MBTA's maintenance department. And, it appears, he has done that work well. In fact, starting in the early 2000s, the MBTA asked him to serve as a temporary supervisor of building and station maintenance, and he went on to serve in that post for about five years, supervising much larger crews of laborers, carpenters, roofers, and others.

Eventually, though, Henderson, who is African-American, became discouraged about his chances of getting hired for the supervisory position on a permanent basis, having twice lost out to other candidates for such a post. He thus decided not to apply to continue as a temporary supervisor. Instead, he returned to his work as a laborer foreperson.

In 2012, however, Henderson decided to try his luck once again for the bigger supervisory job, after he heard about two openings for permanent supervisor positions of the kind that he previously had filled in a temporary capacity. The MBTA interviewed Henderson for these positions, but the all-white panel of reviewers gave him low scores for his answers to the interview questions. The MBTA filled both positions with candidates who had scored higher in their interviews, each of whom was white.

One of these new hires was William Melchionda, who had never previously performed the MBTA supervisor job in any capacity, let alone performed it well in a temporary status for as many years as Henderson had. In fact, Melchionda had never worked at the MBTA at all, and, after Melchionda began in his new role there, Henderson was forced to help train him due to his lack of experience at the MBTA and notwithstanding his considerable supervisory experience in the private sector.

Rather than simply accepting this outcome as if it had been based on interview performance as the MBTA claimed, Henderson chose to file suit for employment discrimination under Title VII. He alleged, among other things, that the MBTA had discriminated against him “because of ․ race” by declining to hire him for the supervisory posts and choosing Melchionda instead. 42 U.S.C. § 2000e–2(a)(1).

From this brief review of the facts, all of which were supported by the summary judgment record, a reasonable juror could find that Henderson had done what he needed to do to make out a prima facie case that the MBTA was liable for violating Title VII. After all, he had all but done the job at issue for years, while Melchionda had not, and so a juror easily could find that they were similarly qualified even though Melchionda had received much higher interview scores than Henderson. Consistent with that conclusion, neither the District Court nor the majority suggests that Henderson did not put forward enough evidence to permit a juror to so find, as neither the District Court nor the majority suggests that Henderson's interview scores alone prevented him from showing that he and Melchionda were similarly qualified for the supervisor positions on offer.27

The result is that, as is usually the case under Title VII, Henderson's effort to defeat his employer's motion for summary judgment turns on whether the employer's claimed reason for its hiring choice was pretextual and whether, insofar as it was, the employer's actual motive for making that choice was because of race. As I will explain, those questions are ordinarily ones of fact that a jury should be permitted to resolve when a plaintiff has made out a prima facie case of an employer's Title VII liability. But here, the majority holds that Henderson's showing as to pretext was too weak to entitle him to have a jury assess it, notwithstanding that a prima facie case was supportably in place. And the majority further holds that, despite the record support for the prima facie case, a jury could not reasonably infer from the evidence of the MBTA's false account of its reason for hiring Melchionda, even if it were strong enough to support a finding of pretext, that it discriminated against Henderson because of race. The majority thus affirms the District Court's grant of summary judgment to the MBTA on each of these independent grounds.

By doing so, however, the majority prevents a long-term African-American employee of the MBTA from having a jury decide whether it was “because of ․ race” that he was passed over for a promotion to a supervisory position that he had successfully held on a temporary basis for years in favor of a white candidate who had not worked at the MBTA for even a single day. And, the majority does so even though the MBTA purported to base that hiring decision solely on the higher scores that the white candidate received for the answers that he gave during his interview, when, as we will see, the District Court itself raised the concern that an objective review of the content of that candidate's interview answers could not support the higher scores that the all-white panel of reviewers gave them.

In consequence, the majority's summary judgment ruling necessarily rests in my view on an unduly limited conception of the jury's proper role in resolving the difficult questions of pretext and motive on which Title VII claims so often turn, given the reality that a “smoking gun” that effectively announces the employer's unlawful discrimination is a rarity. In fact, the District Court's own analysis of the record suggests the reason to send a case like this to the jury.

The District Court expressly found that the MBTA's decision even to interview Melchionda was “hard to understand,” and it also speculated that “cronyism” may have driven the decision to give him the job. Given the thin line between a decision based on cronyism and one made “because of ․ race” in this context, the District Court's puzzlement over and speculation about the MBTA's true motive only serves to underscore to me that -- in light of the evidence that Henderson put forth as to why the interview scoring lacked integrity -- a jury should have been permitted to find for itself what that true motive was.28


The framework that we must use to evaluate the MBTA's motion for summary judgment in this case -- and that the majority relies upon -- is easy enough to describe. It unfolds in three stages, as Henderson has not put forward any direct evidence of race discrimination.

Henderson first must put forth enough evidence to permit a reasonable juror to find that he has made out a prima facie case of employment discrimination based on race. See Ahmed v. Johnson, 752 F.3d 490, 495-96 (1st Cir. 2014) (discussing McDonnell Douglas Corp. v. Green, 411 U.S. 792, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973)). To do so, he need only make a credible case that he was qualified for the position for which he applied and that he was passed over for it in favor of a candidate of a different race with similar qualifications. See Kosereis v. Rhode Island, 331 F.3d 207, 212-213 (1st Cir. 2003).

At the next stage, the burden shifts to the MBTA, as the employer. It must give a legitimate, nondiscriminatory reason for having made the hiring choice that it did. See Paul v. Murphy, 948 F.3d 42, 49 (1st Cir. 2020). This, too, is not an onerous requirement, as the employer bears only a burden of production, see Medina-Munoz v. R.J. Reynolds Tobacco Co., 896 F.2d 5, 9 (1st Cir. 1990), which it can meet even by giving an explanation based on a mistaken understanding of the plaintiff's qualifications, see Paul, 948 F.3d at 51-52.

At the third and final stage, the burden shifts back to Henderson, who, as the plaintiff, must establish that a reasonable juror could find that the MBTA's assertedly legitimate, nondiscriminatory reason for its hiring choice was in fact a pretext for race discrimination. See id. at 49-50. Here, things get somewhat tougher for the plaintiff.

Henderson must not only put forth enough evidence to permit a juror reasonably to find that the MBTA's assertedly nondiscriminatory reason for its hiring choice was not its real reason -- or, otherwise put, that its claimed reason was pretextual. See id. He also must show that a juror could reasonably infer from the pretextual nature of the employer's asserted reason that the employer actually discriminated against him in making that decision “because of ․ race.” 42 U.S.C.A. § 2000e-2(a)(1); see also Reeves v. Sanderson Plumbing Prods., Inc., 530 U.S. 133, 146-47, 120 S.Ct. 2097, 147 L.Ed.2d 105 (2000).

The framework's realism is its virtue. It spares the plaintiff from having to produce what it will often be impossible to produce -- direct evidence of the employer's racially discriminatory motive. See Vélez v. Thermo King de P.R., Inc., 585 F.3d 441, 446-47 (1st Cir. 2009). It contemplates instead that the plaintiff usually may succeed in defeating a motion for summary judgment simply by putting forth enough evidence to permit a juror reasonably to find both that he has made a prima facie case that the employer discriminated against him because of race and that the employer's stated, nondiscriminatory reason for its hiring choice was pretextual. See Reeves, 530 U.S. at 147-48, 120 S.Ct. 2097.

As the United States Supreme Court has explained, “it is permissible for the trier of fact to infer the ultimate fact of discrimination from the falsity of the employer's explanation.” Id. at 147, 120 S.Ct. 2097 (emphasis omitted). That is because, as the Court has also noted, “when all legitimate reasons for rejecting an applicant have been eliminated ․, it is more likely than not the employer, who we generally assume acts only with some reason, based his decision on an impermissible consideration such as race.” Furnco Constr. Corp. v. Waters, 438 U.S. 567, 577, 98 S.Ct. 2943, 57 L.Ed.2d 957 (1978), quoted in Reeves, 530 U.S. at 147-48, 120 S.Ct. 2097; see also Reeves, 530 U.S. at 147, 120 S.Ct. 2097 (noting that “the employer is in the best position to put forth the actual reason for its decision,” so if its “justification has been eliminated, discrimination may well be the most likely alternative explanation”). This understanding, the Court has further explained, comports with “the general principle of evidence law that the factfinder is entitled to consider a party's dishonesty about a material fact as ‘affirmative evidence of guilt.’ ” Reeves, 530 U.S. at 147, 120 S.Ct. 2097 (quoting Wright v. West, 505 U.S. 277, 296, 112 S.Ct. 2482, 120 L.Ed.2d 225 (1992)).

It may be, of course, that in an unusual case, the evidence of pretext will suffice to permit a juror to find it but not to permit a juror to infer from it that the employer discriminated against the plaintiff because of race. See id. at 146-47, 120 S.Ct. 2097 (citing St. Mary's Honor Ctr. v. Hicks, 509 U.S. 502, 511, 519, 524, 113 S.Ct. 2742, 125 L.Ed.2d 407 (1993)). In the main, though, the unmasking of the employer's claimed reason for its hiring choice suffices to create a permissible inference of a racially discriminatory motive when the prima facie case for Title VII liability has been made. See id. at 147, 120 S.Ct. 2097 (citing St. Mary's, 509 U.S. at 511, 113 S.Ct. 2742).

Thus, judges must be careful not to make the test for determining whether there is a genuine dispute of material fact as to pretext unduly stringent. Otherwise, they will cut off at the pass a victim of race discrimination's most realistic means of ensuring that the institution best suited to assess individualized motive -- the jury -- will decide the ultimate question of whether the employer discriminated because of race. See, e.g., Harrington v. Aggregate Indus. Ne. Region, Inc., 668 F.3d 25, 33 (1st Cir. 2012) (“Courts should be especially cautious before granting summary judgment when pretext ․ [is] at issue.”); EEOC v. Unión Independiente de la Autoridad de Acueductos y Alcantarillados, 279 F.3d 49, 56 (1st Cir. 2002) (“Credibility issues ․ ordinarily should be reserved ‘for the factfinder at trial, not for the court at summary judgment.’ ” (quoting Simas v. First Citizens' Fed. Credit Union, 170 F.3d 37, 49 (1st Cir. 1999))).

Similarly, judges must be careful, absent unusual circumstances, not to prevent the jury from deciding for itself whether race discrimination best explains the employer's hiring choice when the evidence suffices to show that the employer's claimed reason for making that choice was pretextual. See, e.g., Theidon v. Harvard Univ., 948 F.3d 477, 496 (1st Cir. 2020) (“We proceed with caution and restraint when considering summary judgment motions where ․ issues of motive and intent must be resolved.”); Soto-Feliciano v. Villa Cofresí Hotels, Inc., 779 F.3d 19, 25 (1st Cir. 2015) (“[W]here ․ the issue [is] whether the employer's stated nondiscriminatory reason is a pretext for discrimination, courts must be particularly cautious about granting the employer's motion for summary judgment.” (internal quotation marks omitted) (quoting Hodgens v. Gen. Dynamics Corp., 144 F.3d 151, 167 (1st Cir. 1998))). Otherwise, the burden-shifting framework will function less as the tool for protecting against subtle forms of race discrimination that it was crafted to be and more as a means of insulating employers from scrutiny. The result then will be that Title VII will offer most plaintiffs merely a nominal -- rather than a meaningful -- day in court.


This understanding of the burden-shifting framework convinces me that Henderson should be permitted to present his case to the jury. To explain my thinking, I first address why the question of pretext was fit for the jury to decide in his case. I then address why the question of whether the evidence of pretext supported an inference of race discrimination was too.


As I have noted, the MBTA asserts that it chose Melchionda over Henderson because of his superior performance during the interview process. It then rightly points out that we have credited claims of reliance on interview performance in granting summary judgment to employers in prior Title VII cases. See, e.g., Goncalves v. Plymouth Cnty. Sheriff's Dep't, 659 F.3d 101, 105-06 (1st Cir. 2011). Thus, the MBTA contends that we have no basis for doubting its claim that it picked Melchionda over Henderson because Melchionda performed so much better in his interview and that, in consequence, there is nothing for the jury to resolve when it comes to the question of pretext.29

I can see the reason to be wary of a claim of pretext that challenges an employer's claimed reliance on an assessment of how candidates answered a question in an interview about, say, what makes one a good employee. Different employers can reasonably disagree about whether one answer to such a question was stronger than another, given the subjective assessment of the quality of the answer that a question of that kind invites. See Hicks v. Johnson, 755 F.3d 738, 746 (1st Cir. 2014). For that reason, the employer's inability in that context to explain convincingly why it found one candidate's answer strong and another's weak ordinarily will fail -- in and of itself -- to create a triable issue of fact as to pretext. See Hidalgo v. Overseas Condado Ins. Agencies, Inc., 120 F.3d 328, 337 (1st Cir. 1997) (“Courts may not sit as super personnel departments, assessing the merits -- or even the rationality -- of employers' nondiscriminatory business decisions.” (quoting Mesnick v. Gen. Elec. Co., 950 F.2d 816, 825 (1st Cir. 1991))).

Moreover, nothing about the interview process that the MBTA used here seems suspect on its face. The MBTA's human resources staff worked with a “Selection Committee” composed of three MBTA employees to create a standardized list of interview questions. The group assembled an “answer key,” which “lists specific points the Selection Committee seeks in an answer to each question.” The Selection Committee then interviewed each candidate in turn, asking each one the same agreed-upon questions in the same order and taking notes on the candidates' responses.

In addition, after each interview ended, each participating member of the Selection Committee assigned the candidate's response to each question a score from zero to ten, supposedly based on both the answer key developed by the Selection Committee and a score sheet. The scores assigned then were summed together,30 and, as was the MBTA's standard practice, the MBTA hired the two candidates with the highest combined scores.

As it turned out, moreover, Melchionda and one other white candidate, Bernadette Higgins, who was also white, received combined interview scores of 184 and 187, respectively, and each was hired for one of the open positions. Henderson, by contrast, received a combined score of 117 and was not chosen.

If our inquiry ended there, a juror would not appear to have any reason to disbelieve the MBTA's assertion that it hired Melchionda over Henderson because of his superior interview performance or to conclude that this was not a legitimate, nondiscriminatory reason for making that choice. But, as I will explain, our inquiry cannot end there.

I am aware of no precedent that precludes us from concluding that a juror reasonably could find that the MBTA's reliance on the differential scoring of the interviews was not the real reason for the hiring decision if that juror reasonably could find that the scoring of the interview answers was not on the level. Nor do I see how such an argument could be successfully advanced.

We have repeatedly recognized that “[o]ne way to show pretext” is to identify “weaknesses, implausibilities, inconsistencies, incoherencies, or contradictions in the employer's proffered legitimate reasons for its action.” Billings v. Town of Grafton, 515 F.3d 39, 55 (1st Cir. 2008) (second quotation quoting Hodgens, 144 F.3d at 168). Such holes in the employer's official story, we also have repeatedly said, may establish that the account offered by the employer is “unworthy of credence” and thus that “the employer did not act for the asserted non-discriminatory reasons.” Id. (quoting Hodgens, 144 F.3d at 168).

To be sure, as I have noted, some interview questions do not invite answers that lend themselves to objective assessment. That can make it difficult for a plaintiff to prove that the scoring was suspicious, even when there is room for debate about the strength of the answers given. I have no quarrel with that reality, given the important role that qualitative judgments about candidates based on interview performance can play in hiring.

But, some interview questions seek answers that are readily reviewed objectively. For example, if an interview question sought information about whether a candidate had post-secondary education, and the answer key indicated that the employer wanted to hire someone with at least a college education, it would be concerning if the candidate who answered that he had gone to college scored worse than the one who answered that he did not.

As a result, there is no reason that I can see to bar a juror from declining to credit the employer's claim to have relied on an assessment of an answer that invites objective assessment when the record supportably shows that the employer's actual assessment of that answer is, objectively, indefensible. Furthermore, in my view, if a juror reasonably could conclude that the employer fudged the numbers in scoring that one answer -- at least if the “objective” question that prompted it appears to be a critical one -- then that same juror reasonably could draw the additional inference that the scoring of answers even to more “subjective” questions was also suspect. See Hicks, 755 F.3d at 746 (acknowledging that “the subjectivity necessarily introduced by the interview process can mask discrimination”); United States v. González-Martínez, 825 F.3d 51, 56 (1st Cir. 2016) (“Jurors ․ are not expected to resist commonsense inferences on the realities of human experience.” (quoting United States v. Saccoccia, 58 F.3d 754, 782 (1st Cir. 1995))).31 Nor have we had occasion in any of our cases concerning interview performance to hold that it is necessarily unreasonable for a juror to infer that the evaluative process was tainted generally when the evidence of the employer's problematic evaluation of a specific, important question is clear enough.

It is thus significant to me that there are inexplicable 32 discrepancies regarding the scoring of Henderson's and Melchionda's answers to the question that each was asked at the interview regarding computer skills, which was undeniably an important one. The computer skills question was as follows: “Please tell us about your computer skills[.] Have you ever used an Asset Management system? What would be the value of having a computerized system that tracks the status of needed repairs?”

On its face, this question hardly seeks an answer that defies objective assessment. Indeed, the “answer key” reveals that the Selection Committee was looking for three specific and objectively verifiable things in the candidates' response: (1) “MS [Microsoft] Office exp[erience],” (2) “Familiarity,” and (3) “Use of Asset management system.”

It is therefore troubling that the answers that Henderson and Melchionda each gave to this question were scored so differently. For, as we will see, the answers themselves, so far as the record reveals, do not appear to warrant such disparate assessments in Melchionda's favor.

The interview notes show that Henderson claimed to have “exp[erience] w[ith] Microsoft,” including “outlining work” and “editing work.” They also reveal that Henderson stated that he had experience with “Blue Zone,” a software program used at the MBTA, and that he had gone to “school for word processing.” As to knowledge of Microsoft Office and “familiarity,” in other words, Henderson seemed to answer in just the way that the answer key suggested a strong candidate would.

With regard to the Committee's question about a computerized repair tracking system, Henderson's answer was not as strong. He apparently stated that such a system would be “great,” as it would avoid “overlap[ ]” and would help “overcome gaps to get trades together.”

The interview notes show that when Melchionda responded to the question about computer skills, he did not claim to have any experience with Microsoft programs, as Henderson did and as the answer key indicated that the Selection Committee desired a candidate would. Nor do they show that he claimed to have any experience with Blue Zone or any other program used at the MBTA.33

Like the notes taken for Henderson's response, moreover, the notes for Melchionda's do not shed much light on his response to the computerized repair system question and certainly do not show it to have been very strong. They indicate that he said such a system would “save[ ] time,” be “more efficient,” and “keep records,” without additional detail.

The notes that one of the interviewers took do indicate that Melchionda claimed to “use [a] computer every[ ]day,” and the other's indicate that he had “operational exp[erience]” on a P.C. Melchionda did also mention that he “print[ed] out purchase orders, request forms, [and] maintenance logs” and “ke[pt] up the logs to respond to tenants.” But, even still, both interviewers noted that Melchionda conceded that his “use” of computers at the time of the interview was “minimal.”

Objectively, then, one would be surprised to learn that Melchionda was given nearly twice as good a score for his answer to the computer skills question as Henderson.34 Yet, Henderson received two “fours” for his response to the question, indicating that the interviewers thought his response was a “Fair” one that “[m]issed important item(s),” while Melchionda's response earned him two “sevens.” According to the score sheet, that score implies Melchionda's response fell somewhere between a “Good” answer that made “the most salient points” and a “Very Good” one that “[a]nswered most of the question.”

I am hardly alone in thinking that the scoring of the candidates' answers to this question was fishy. The District Court itself was puzzled: “Henderson had more computer experience than Melchionda ․ yet Henderson received three points lower” for his response to the question.35

On appeal, the MBTA takes issue with the District Court's characterization -- repeated by Henderson -- of the record as showing that Melchionda claimed to have “minimal” computer skills. But, it fails to identify any evidence that addresses the key point that bothered the District Court: what in Melchionda's response to the question would validate the decision to award him nearly twice as many points as Henderson for his response to it?

The majority suggests that an interviewer could have preferred aspects of Melchionda's response, such as his mention of “operational exp[erience]” or his response to the Selection Committee's question about a hypothetical computerized repair system,36 to Henderson's. Maj. Op. –––– – ––––, ––––. But, nothing in the record indicates that the Selection Committee in fact had those preferences. Nor does the MBTA itself assert as much on appeal. Indeed, the answer key reveals that the Selection Committee was concerned, at least in large part, with Microsoft Office experience, a qualification that Henderson alone mentioned. I do not see how we can rely on our own speculation about the Selection Committee's unspoken preferences to justify taking the pretext issue from the jury when the MBTA stresses to us the critical role that the answer key plays in ensuring that the scoring process is a fair one that is not infected by bias.

Of course, Henderson's three-point-lower score as to this one question regarding computer skills could not itself have been decisive. The scoring gap between the two candidates was much larger. But, this question was by no means a trivial one. The showing that Henderson has made about the unexplained way that the answer to it was scored thus provides a basis from which a reasonable juror could surmise that the way that the interviewers scored the answers to the other eleven questions -- including those of a more subjective bent -- lacked integrity.37

That is so because the record provides support for just that finding, once the basis for questioning the scoring of the computer skills question is considered. The District Court found that the decision to award Melchionda three more points than Henderson on one question, which asked the candidates to explain their reasons for applying for the position, had no “objective basis” in the answers that the candidates gave. The MBTA does not even attempt to identify such a basis on appeal. Additionally, three questions asked the candidates to describe their experience with certain job-related tasks. Melchionda scored a combined twenty points higher on them, even though Henderson alone had five years of direct experience performing the supervisor job.

For these reasons, Henderson has shown, in my view, that there is a genuine issue of material fact as to whether the evidence supportably shows that the interview scores were so lacking in integrity that the MBTA's claimed reliance on them to justify its hiring of Melchionda may be deemed pretextual. Accordingly, I disagree with the majority's decision to grant summary judgment to the MBTA based on an inadequate showing of pretext, as I think that ruling risks making the interview process an easy means by which an employer may insulate a hiring choice from the kind of scrutiny that Title VII contemplates.38


That still leaves the question of whether a juror reasonably could find on this record that the MBTA's stated reason for choosing Melchionda over Henderson was not only a pretext for its true motive but also a pretext for race-based discrimination. See Paul, 948 F.3d at 49-50. Unless Henderson can make that showing, after all, he cannot fend off the MBTA's motion for summary judgment.

Given that the majority holds Henderson's showing was too weak to permit a finding of pretext at all, it need not reach this additional issue about what inference might be drawn from a finding of pretext. But, in an independent holding, the majority does so nonetheless and concludes that Henderson's showing on this score is also too weak to get his case to a jury. Here, too, though, I disagree with the majority's decision to cut the jury out of the process.

Neither the MBTA nor the majority contends that some unarticulated but nevertheless nondiscriminatory reason other than the interview scores drove the hiring decision. See Reeves, 530 U.S. at 148, 120 S.Ct. 2097 (“[A]n employer would be entitled to judgment as a matter of law if the record conclusively revealed some other, nondiscriminatory reason for the employer's decision ․”). Nor does either the MBTA or the majority contend that any evidence affirmatively shows that race did not influence that decision. See id. (noting that judgment as a matter of law would be proper if there was “only a weak issue of fact” regarding pretext “and there was abundant and uncontroverted independent evidence that no discrimination had occurred”). Thus, neither the MBTA nor the majority disputes that, insofar as the evidence permits a finding of pretext, the MBTA fails to offer any reason at all for having made the hiring choice that it did.

The majority relies instead on the asserted weakness of Henderson's pretext showing, even though it assumes in this part of its analysis that the record suffices to support a pretext finding. To justify doing so, the majority leans on our decision in Feliciano de la Cruz v. El Conquistador Resort and Country Club, 218 F.3d 1 (1st Cir. 2000). Maj. Op. –––– – ––––.

There, we held that the plaintiff, a credit manager fired from her job at a hotel, was able to make a “thin” but supportable showing that the reasons given for her termination -- poor job performance, supposedly -- were pretextual, largely due to her receipt of commendations and a pay raise from her employer combined with evidence that the hotel's financial woes were not her fault. Feliciano, 218 F.3d at 7-8. Nevertheless, we affirmed the grant of summary judgment to the employer because, we concluded, the evidence was insufficient for a juror to infer that, as the plaintiff had claimed, she was fired due to her Puerto Rican origin. Id. at 8-9.

Feliciano was decided just days before the Supreme Court's decision in Reeves, which purported to disagree with what it understood to be our Circuit's precedent that a jury may not infer discrimination based only on the combination of a prima facie case of discrimination and evidence of pretext. See Reeves, 530 U.S. at 140-41, 148, 120 S.Ct. 2097. The Feliciano panel in a brief order did later deny a post-Reeves petition for rehearing, in which the plaintiff had argued that Reeves was at odds with the panel's analysis. 218 F.3d at 10. In doing so, the panel stated that the “conclusion that Feliciano failed to adduce sufficient evidence to survive summary judgment” was “based on the particular weakness of her case.” Id.

That order makes clear that Feliciano is a binding ruling that we must follow, notwithstanding Reeves. But, by its own terms, Feliciano is best read to be a necessarily fact-dependent ruling. As such, it is readily distinguished from this case, given the different nature of the pretext showings in each.

Henderson premises his pretext showing on evidence that casts doubt on the integrity of the scoring of the candidates' interview answers. Yet, far from presenting a “weak” or “thin” case in that regard, the District Court itself expressed concern based on Henderson's showing about the scoring. So, whatever one makes of the strength of the pretext showing in Feliciano, I do not see how Henderson's showing is comparably bare bones.

Of course, Reeves recognizes that some rare showings of pretext -- even if more than strong enough to support a finding of pretext -- in their nature have little “probative value” in showing race discrimination. 530 U.S. at 148-49, 120 S.Ct. 2097. For example, a plaintiff may prove that an employer's claimed reliance on the winning candidate's superior prior experience is pretextual if he can show that the winning candidate bribed the employer. That plaintiff, however, will have done little to advance his case for proving race discrimination. Indeed, he arguably would have all but disproved it by demonstrating the true but non-race-based ground for the employer's hiring decision.

But, even if Feliciano could be read to have been relying on the exception that Reeves carves out for rare cases of that kind, that exception could not be understood to encompass Henderson's case. He has supportably shown that his white competitor received higher scores for objectively worse answers than he gave to the exact same question and that this white competitor's answers to other more subjective questions were not evidently better than his own, even though they, too, were scored as if they were.

When the showing of pretext rests on evidence that the Title VII plaintiff was not hired for doing exactly what the candidate of a different race who was hired did, see Thomas v. Eastman Kodak Co., 183 F.3d 38, 62 (1st Cir. 1999), the pretext showing is inherently one that involves evidence that is at least suggestive of race-based disparate treatment. How, then, can we say that the evidence of pretext in such a case is of no “probative value,” Reeves, 530 U.S. at 149, 120 S.Ct. 2097, to the ultimate question of whether race discrimination was the real basis for the decision?39

Thus, I do not see how Henderson's case qualifies as the unusual one in which a jury should be denied the chance to draw an inference of race discrimination once it finds that the employer's stated reason for its hiring choice was not the real one. And that is so even if we do not add into the mix the letter from the Federal Transit Administration and the Federal Highway Administration that alleged that, over a three-year period around the time the MBTA hired for the two supervisor positions, the agency received 750 Equal Employment Opportunity complaints, paid out more than $4 million in settlements and legal fees, and yet failed adequately to document the allegations.40

That the District Court was moved to posit that cronyism may have best explained the basis for the MBTA's choice serves to reinforce this conclusion. That explanation -- which is the only one other than the MBTA's merit-based explanation for its hiring choice that the District Court referenced -- was one that was itself a close cousin of the kind of race-based discrimination that Title VII forbids. That speculation about the MBTA's true but unacknowledged motive thus highlights to me that -- insofar as Henderson's pretext showing sufficed -- the question of motive here necessarily turned on the kind of inferential assessments, based on common experience, that lay jurors are best qualified to make.


No doubt, a future case that implicates the proper application of the final stage of the Title VII burden-shifting framework will have enough factual differences from this one to permit it to be distinguished, just as I think this one can be distinguished on the facts from Feliciano. The ruling there was one of law -- given that it was affirming a grant of summary judgment to the defendant. But, it is expressed in terms that necessarily limit its reach to that case's specific facts. See Feliciano, 218 F.3d at 7 (emphasizing that, “[i]n evaluating whether summary judgment was proper ․ ‘everything depends on individual facts’ ” (quoting Thomas, 183 F.3d at 57)).

Indeed, were Feliciano understood to state a generalizable rule, as the majority suggests that it purported to do, Maj. Op. –––– – ––––, then it is hard to know what that rule would be other than that a showing of pretext is not enough when, given the particular facts involved, that showing is too “thin” or, as the majority now terms it, “barebones,” id. at 37. Certainly, the rule cannot be that a plaintiff must do what Henderson has not, which is to provide affirmative evidence of discriminatory motive. Reeves makes clear that is not required, see 530 U.S. at 147, 120 S.Ct. 2097, and Feliciano is not to the contrary, see 218 F.3d at 6, 10. I thus must assume that the majority's extensive recounting of the various types of such affirmative evidence that Henderson does not put forth is not intended to suggest that his failure to produce it weighs against -- rather than merely fails to aid -- his effort to have a jury decide the motive issue. Maj. Op. –––– – –––– & nn. 19-20.

I am nonetheless concerned that, by now extending Feliciano to this case, we may be stumbling -- through an accretion of fact-dependent rulings -- into what will turn out to be the application of an “I know it when I see it” test. But, if we were to adopt such a test de facto, then it seems to me that we necessarily would be adopting an approach to applying the burden-shifting framework that also could not be squared with Reeves. Such an application would bring about a shift in the approach to the respective roles of judge and jury in deciding questions of individual motive under Title VII that Reeves does not permit.

There, the Supreme Court expressly identified just two types of cases in which a jury should be barred under the burden-shifting framework from finding from the plaintiff's supportable showing of pretext that the employer made its hiring decision because of race. They are: (1) where “conclusive[ ]” evidence of “some other[ ] nondiscriminatory reason for the employer's decision” exists, and (2) where “weak” evidence of pretext is undercut by “abundant and uncontroverted independent evidence that no discrimination had occurred.” 530 U.S. at 148, 120 S.Ct. 2097.

These illustrative examples show that the rare exceptions Reeves has in mind are ones in which some feature of the record directly undercuts the normally permissible inference that the jury may draw from an otherwise adequate showing of a prima facie case plus pretext. 530 U.S. at 148, 120 S.Ct. 2097; see also id. at 154, 120 S.Ct. 2097 (Ginsburg, J., concurring) (“anticipat[ing]” that “circumstances in which plaintiffs will be required to submit evidence beyond” what is necessary to show a prima facie case and pretext “in order to survive a motion for judgment as a matter of law ․ will be uncommon”). Consistent with this understanding, other circuits have read Reeves both to presume that once a finding of pretext is permitted and a prima facie case has been made the jury gets to make the ultimate call and that the evidentiary burden for overcoming that presumption is a substantial one. See Blow v. City of San Antonio, 236 F.3d 293, 298 (5th Cir. 2001) (finding a “prima facie case,” a “material issue of disputed fact as to whether the employer's explanation was false,” and “no unusual circumstances that would prevent a rational fact-finder from concluding that the employer's reasons ․ were discriminatory” to be enough to defeat summary judgment under “a straightforward application of Reeves”); Rowe v. Marley Co., 233 F.3d 825, 830 (4th Cir. 2000) (holding that “absent” “evidence that precludes a finding of discrimination,” “courts may not require a plaintiff who proves both a prima facie case and pretext to produce additional proof of discrimination in order to survive a defendant's motion for summary judgment”).

Here, however, there is no such directly undermining evidence in the record, nor is there anything else in it that could suffice to dispel the concerning inference created by Henderson's pretext showing (assuming, as the majority does at points, that it was supportably made). In fact, as I have explained, the pretext showing in this case itself sounds in the disparate treatment of candidates of different races, which was not true in Feliciano. Yet, the majority grants summary judgment to the MBTA nonetheless.

In justifying the decision to take the ultimate question of motive based on pretext out of the hands of the jury, the majority treats the logical chain of reasoning that Henderson asks us to recognize as if it necessarily falls outside the realm of permissible inferential reasoning because it can only be understood to require the fact-finder to engage in impermissible speculation. Maj. Op. –––– – ––––. But, while the summary judgment standard requires us to draw the line between inference-making and speculative guessing, it is not a Platonic one that may be limned without reference to the source of law that sparks the need for a motive determination. It is necessarily a context-dependent one that must be assessed with respect to the specific pretext and motive issues that are at play under the specific statute that grounds the plaintiff's claim.

Under the Supreme Court's Title VII jurisprudence, the berth given to juries to draw on their own experiences and lay understandings to make inferences about employer motive based on findings of pretext is a wide one. Otherwise, the Court has made clear, the promise of a discrimination-free hiring process that Title VII was enacted to secure would be thwarted. And that is surely so. Indeed, were jurors not given that leeway, Title VII would undermine itself, as the very legal bar to race-based hiring that it imposes incentivizes employers to cover up a racially discriminatory motive that in an earlier era they could have more freely risked making known without fear of violating federal law.

The concern, then, is that the majority's reluctance to permit the jury to engage in “speculation” reflects a misunderstanding about the kind of permissible inferential reasoning that, as a matter of law, Title VII's burden-shifting framework inherently contemplates. Per Reeves, as I understand it, this framework aims to ensure that juries may decide difficult questions of motive under Title VII even though conclusive or even affirmative evidence of employer motive often will prove elusive. There will, then, usually be only judgments to be made from the fact that the plaintiff has shown that he lost out to a similarly qualified person of a different race for the job, that the employer falsely explained its reason for making that choice, and that it has offered no non-discriminatory reason to explain what in consequence has been left unexplained. To deem impermissibly speculative in such a case a jury's judgment that racial discrimination best explains the inexplicable -- especially when the plaintiff is so qualified that he was forced to train the one to whom he lost out and the pretext showing relies on evidence of disparate treatment -- is to undermine the burden-shifting framework itself.

For these reasons, I am concerned that, through a series of individualized, seemingly fact-dependent rulings, Reeves is at risk of suffering death by a thousand cuts. Insofar as that is so, the time to stop the bleeding, in my view, is now.


For these reasons, I respectfully dissent.


1.   During the same period, Debra Gilcoine was promoted to one of these permanent supervisor positions. She was promoted to superintendent three to four years later.

2.   Emde began working at the MBTA in 1998 as an HR representative. Before that, he had worked as both an adjunct professor at Bridgewater State College and at MVP Sports from 1995 to 1997. In 2007, Emde was promoted to senior HR generalist. In 2009, he was promoted to manager of staffing. As the manager of staffing, he oversaw the work of HR generalists. Throughout his time at the MBTA, Emde screened resumes, made job postings, and conducted interviews.

3.   Melchionda listed Bill Perez, the head of the MBTA's HR department, and Perez's brother as references. Perez never spoke with anyone involved in the selection process and the selection committee did not discuss Perez's inclusion as a reference.

4.   Emde testified that a selection committee typically comprises, at minimum, one HR representative and one or more members of the department of which the position is a part.

5.   Emde and Martin testified that the MBTA followed this uniform policy and practice. Sayten Patel described the interview process and scoring just as the MBTA has described it was used. There is no contrary evidence.

6.   Baker gave all but one applicant very low scores and, after HR discovered that Baker's subordinate had recommended that applicant, HR removed Baker's scores from the hiring calculations. The removal of Baker's scores did not affect Henderson's ranking with respect to the other candidates.

7.   Patel knew Henderson wanted to be a supervisor, and testified that he was “trying to help ․ Henderson develop his skills so that he would be a viable candidate for the [supervisor] position.”

8.   Henderson's complaint alleges that he was more qualified than both Higgins and Melchionda and they were afforded disparate treatment because of their race. But on appeal, Henderson argues only that he was more qualified than Melchionda and the MBTA applied different rules to Melchionda due to his race. Henderson's claim based on Higgins's selection has therefore been dropped on appeal. Nonetheless the process used in the decision to promote her is relevant to and supports the MBTA's position, and refutes Henderson's claim of differential treatment based on race.

9.   Our review is de novo and, save the district court's conclusion that Henderson did not show pretext, we do not credit or agree with the district court's characterizations of the evidence or find its conclusions supported by the record.There is no basis to question, as the district court seemingly did, the MBTA's explanation that only some of the MERs determine whether an applicant receives an interview, and the computer skills MER is not one of those. The district court stated: “It is hard to understand why someone who left an answer blank on a minimum required job skill was given an interview ․” The record does not support the court's language.The district court also treated all interview scoring together as a subjective evaluation, and that was an error. That some of the interview criteria were more subjective than others does not alter the fact that Melchionda objectively had more experience than Henderson and met the MERs. Having met the objective criteria at issue, the committee could reasonably conclude that Melchionda was qualified to get the job.

10.   Henderson argued for the first time at oral argument that a provision of the application required that Melchionda be disqualified for not signing the “Notification & Agreement” section of the application. Henderson then contended that this showed racially differential treatment of Melchionda's job application. But Henderson did not argue this point to the district court or in his initial brief to this court, and so has doubly waived it. Arrieta-Gimenez v. Arrieta-Negron, 859 F.2d 1033, 1037 (1st Cir. 1988); Pignons S.A. de Mecanique v. Polaroid Corp., 701 F.2d 1, 3 (1st Cir. 1983).Moreover, Martin testified that, in his experience, it was “not unusual” for this portion of the application to be unsigned. Emde testified that there was no requirement that an applicant sign this section. Henderson admitted that both of these statements were true and has not introduced any evidence that rebuts these statements.

11.   This reliance on interview scores rebuts Henderson's waived argument that David Benson's neutral recommendation for Melchionda shows pretext, as the recommendation could not impact the interview scores and, as Emde testified, the MBTA mainly used references to “verify dates of employment.” See United States v. Zannino, 895 F.2d 1, 17 (1st Cir. 1990).

12.   Henderson never argued in his brief to this court that his nonselection despite his supervisory experience is evidence of pretext and so he has waived such an argument. Pignons S.A. de Mecanique, 701 F.2d at 3.

13.   Henderson's only dispute with the answers listed for him in the interview notes is that he contends they are incomplete, as he asserts he spoke for two to three minutes for each question and so said more in each response. He cannot, however, identify any specific inaccuracy or omission in the answers listed for him in the notes. This unsupported contention does not lessen the objectivity of the interview process.

14.   The dissent now argues that Henderson preserved this challenge when his counsel stated in oral argument before the district court that there was “a different set of standards as to how the interview answers are ‘scored’ between African-American and white candidates.” Counsel's conclusory remarks at oral argument are not sufficient to preserve the issue. An argument “not fully developed below” is waived. Ryan v. Royal Ins. Co., 916 F.2d 731, 734 (1st Cir. 1990); see also In re Olympic Mills Corp., 477 F.3d 1, 17 (1st Cir. 2007).

15.   Henderson asserted for the first time at oral argument that he “had a number of certificates from the MBTA relative to his computer training.” Henderson's brief does not mention these certificates nor are they in the committee notes or the record.

16.   The notes do not reveal whether Melchionda informed the committee he had skills in Microsoft Office and Excel, although he did in deposition taken by plaintiff's counsel state he had those skills.

17.   The MBTA need not consider other evidence of qualifications and we do not “sit as super personnel departments, assessing the merits -- or even the rationality -- of [the MBTA's] nondiscriminatory business decision[ ]” to use such a process. Goncalves, 659 F.3d at 107 (quoting Mesnick v. Gen. Elec. Co., 950 F.2d 816, 825 (1st Cir. 1991)).

18.   We note that, in the past, the inconsistencies we have concluded to show pretext have been far more significant than the score of a single interview question. See, e.g., Billings v. Town of Grafton, 515 F.3d 39, 56 (1st Cir. 2008) (inconsistent “accounts about who made the decision to transfer [the plaintiff] and, more importantly, how it was made”); Santiago-Ramos v. Centennial P.R. Wireless Corp., 217 F.3d 46, 56 (1st Cir. 2000) (inconsistent justifications for firing the plaintiff).

19.   Although Henderson argues that studies on the effect of race in interviews show pretext for racial discrimination, Henderson did not actually introduce any such studies and waived this argument by failing to develop it. Zannino, 895 F.2d at 17. Moreover, that “the decision makers were [not of Henderson's protected class] does not alone ․ create an inference [of] ․ discrimination.” Rivas Rosado v. Radio Shack, Inc., 312 F.3d 532, 534 (1st Cir. 2002).

20.   Henderson argues that his, and other minority MBTA workers', assignment to remove snow from the Orange Line, various MBTA hiring statistics, and an August 2013 Federal Transit Administration (“FTA”) letter to the MBTA support an inference of discriminatory hiring at the MBTA. All three arguments lack merit, and Henderson waived the first two by failing to develop them. Zannino, 895 F.2d at 17.Henderson alleges that the Orange Line work assignment is more difficult than others (despite testifying it required the same number of hours as other assignments and “[s]ome might say” that the other lines required more work). He also contends that many MBTA employees call the Orange Line the “Soul Train Line” because of the neighborhoods through which it runs. But Henderson never identified anyone who said the phrase “Soul Train Line” and he testified that no supervisor ever said it. None of the hiring/promotion decision makers decided which subway line Henderson worked. See Thompson, 522 F.3d at 178. Moreover, Henderson's two Orange Line subordinates were white. The record does not support the contention that race affected this work assignment. The record does not support the contention that this work assignment had anything to do with the hiring/promotion decision.Henderson's statistics do not support an inference of pretext for racial discrimination here. “ ‘[S]tatistical evidence of a company's general hiring patterns ․’ is only helpful ‘if it tends to prove the discriminatory intent of the decision makers involved.’ ” Ray v. Ropes & Gray, LLP, 799 F.3d 99, 116 (1st Cir. 2015) (first quoting LeBlanc v. Great Am. Ins. Co., 6 F.3d 836, 848 (1st Cir. 1993); and then quoting Hillstrom v. Best W. TLC Hotel, 354 F.3d 27, 32 (1st Cir. 2003)). Henderson's hiring statistics ended four years before the hiring/promotion decision, lack detail, and lack any “meaningful connection” with the decision not to hire him. Id.Finally, Henderson argues that an August 2013 FTA letter to the MBTA evidences hiring discrimination. But this letter merely stated that the FTA deemed the MBTA as non-compliant because the MBTA had not provided enough information to the FTA for it to investigate the allegations of discrimination and the MBTA had received 750 unspecified Equal Employment Opportunity complaints over the previous three years. After the MBTA provided more information on October 10, 2014, the FTA, on November 25, 2014, deemed the MBTA in compliance with federal law and stated that it had “amended [its] finding of probable non-compliance to probable compliance.”

21.   Further, Martin testified that he did not recall anyone ever mentioning Perez's name during the interview process and, at the time, did not know Melchionda had listed Perez as a reference.

22.   Henderson's complaint did not allege under 42 U.S.C. § 2000e-2(m) that race was a “motivating factor” in the MBTA's decision, which would allow Henderson to succeed without showing but-for causation. See Bostock, 140 S. Ct. at 1739-40; see also Chadwick v. WellPoint, Inc., 561 F.3d 38, 48 (1st Cir. 2009) (viewing separately the “mixed motives” approach and traditional approach under McDonnell Douglas). Henderson's complaint alleges that the hiring decision was made “not ․ for any legitimate business reason but because of his race.” But even had Henderson made a mixed-motives argument, it would similarly fail, because he has not provided evidence that the hiring decision was “attributable even in part to a forbidden bias.” Burton v. Town of Littleton, 426 F.3d 9, 20 (1st Cir. 2005).

23.   The district court did state it was “hard to understand why someone who left an answer blank on a minimum required job skill was given an interview.” But the record clearly shows that applicants were not screened for interviews on the basis of the computer skills MER.

24.   Henderson also includes this retaliation claim as evidence of racial discrimination. To the extent this could show an atmosphere of racial discrimination and Henderson has not waived this argument by failing to develop it, see Zannino, 895 F.2d at 17, it fails for the same reasons.

25.   Although the parties dispute whether Gilcoine yelled at Henderson and whether Henderson complained to Patel, under our standard of review, we assume that both events occurred. See Johnson, 714 F.3d at 52. Henderson alleges Gilcoine's outburst was motivated by racism, so his complaint to Patel was protected conduct. See Fantini v. Salem State Coll., 557 F.3d 22, 32 (1st Cir. 2009) (stating that protected conduct can be the opposition to “any practice made an unlawful employment practice by Title VII” or that the plaintiff reasonably and in good faith believes violated Title VII (quoting Long v. Eastfield Coll., 88 F.3d 300, 304 (5th Cir. 1996))).

26.   There is also no evidence that the decision to give podium duty to another employee was discriminatory.

27.   The MBTA does argue that Henderson failed to supportably make out a prima facie case because of his poor performance at the interview for the job compared to Melchionda's. As I will explain, however, that contention does not hold up because of what the evidence shows about the reasons to doubt the interview scoring. For that reason, I agree with the majority in following the District Court in not resting its ruling as to summary judgment on this ground.

28.   I agree fully, however, with the majority's grant of summary judgment as to Henderson's retaliation claim.

29.   Some of our prior cases involving a Title VII plaintiff's relatively poor interview performance have held that no triable issue of material fact existed as to whether a prima facie case had been established because that poor interview performance itself precluded the plaintiff from establishing either adequate qualifications for the position or similar qualifications to the person whom the employer ended up hiring. See, e.g., Goncalves, 659 F.3d at 105-06. But, while the MBTA contends those precedents support that same result here, the majority, like the District Court, does not so hold. I agree with that approach, given how evidently qualified for the post Henderson was based on his past experience as an MBTA supervisor and the fact that Melchionda had not worked at the MBTA at all. Thus, here, the interview performance is relevant to the pretext issue, if at all. In any event, for the reasons I set forth below, the lower scores are themselves sufficiently suspect that, for the same reasons they cannot suffice to preclude Henderson from having the jury decide the pretext question, they also cannot suffice to preclude him from having the jury decide the prima facie case issue.

30.   While the Selection Committee was originally composed of three members, only two of the interviewers' scores ultimately were used for the hiring decision. The MBTA contends that the scores awarded by one of the original members of the Selection Committee were removed due to a concern about improper favoritism.

31.   This inference is bolstered by Henderson's testimony that the Selection Committee's notes fail to reflect the level of detail he provided in each of his responses to the interview questions.

32.   It is true the District Court did not use this word to characterize the MBTA's scoring of this answer, nor did it deem it “indefensible.” After all, it refused to let the jury resolve the pretext question. But, I do not see how scoring of answers that are “inconsistent,” Maj. Op. ––––, can explain the decision by an employer to credit them or make them defensible.

33.   The parties do not discuss whether “Blue Zone” is an asset management system of the type that the answer key indicates that the MBTA valued experience in using.

34.   The majority -- but, notably, not the defendants -- conclude that Henderson has waived this line of argumentation by failing to raise it below or develop it here. Maj. Op. ––––. But, at a hearing on the MBTA's summary judgment motion below, Henderson's attorney argued that there was “a different set of standards as to how the interview answers are scored” between African-American and white candidates. Later on, the District Court pressed counsel for the MBTA about what the record showed about Melchionda's responses to the computer skills question in his interview and stated it would “look ․ up” the answer to its question. The District Court also emphasized in this respect that, at least as it understood things, its point about the content of the candidates' responses to the computer skills question was the precise “point [Henderson's] raising.” And, as noted, the District Court relied on these responses to assess whether summary judgment was properly granted to the MBTA.On appeal, Henderson argues that “the scoring sheets themselves give a clear indication that numerical values assigned to the answers could be and were arbitrary.” In particular, he notes that the computer skills question was arbitrary, as Melchionda, “who described himself as having minimal computer skills, nevertheless[ ] scored higher on this question than Henderson[,] who had significant computer skills.” Furthermore, Henderson contends, the “arbitra[ry] nature of the test scores” “imply that the ․ reason offered by the employer was pretext.” Thus, I do not agree with the majority that this claim has not been properly preserved.

35.   I note that the District Court was also puzzled by the fact that Melchionda had been granted an interview when his application left blank the line that asked for him to describe his computer skills. As the District Court put it, “[i]t is hard to understand why someone who left an answer blank on a minimum required job skill was given an interview.”

36.   To the extent that the majority contends that Melchionda's response to this question is objectively “more detailed,” Maj. Op. ––––, that characterization is not supported by the record. As described above, there is no objective basis on this record for finding one response to be preferable to the other, and the MBTA provides no explanation for why Melchionda's answer better matched its desired response.

37.   To make up the scoring gap, the score given for each question by each interviewer would need to shift by only a little more than a point up for Henderson and down for Melchionda.

38.   I should add that the fact that many other applicants who were not hired scored better than Henderson is of no significance to the pretext inquiry, given that the evidence supportably shows that the answers given by the even higher scoring candidate that the MBTA did hire -- Melchionda -- were in an objective sense no better (and, in some respects, even worse) than those given by a candidate who scored as low as Henderson. Once that evidence of the suspicious nature of the scoring was in place, a juror could reasonably find that the MBTA must not have relied on the interview performance of the candidates in hiring Melchionda, notwithstanding its representation to the contrary.

39.   That Henderson has not shown that each of the applicants who scored higher than he did gave answers no better than his poses no problem. If he can show that a juror could find that the MBTA was telling a falsehood in asserting that it picked Melchionda because of their divergent interview performances and that he was similarly qualified for the job but of a different race, then, for the reasons explained above, nothing in this record would prevent a juror, per Reeves, from inferring from that falsehood that race was driving the hiring choice in this round of hiring. Henderson need show no more to permit a jury to find that he suffered an adverse employment consequence “because of ․ race” and thus to permit it to find for him on his Title VII claim.The majority does separately assert that, because Henderson scored lower on the interview than sixteen other applicants who were not themselves selected for the open positions, he could succeed on a Title VII claim, if at all, only pursuant to the theory that race was at most a “motivating factor” for the MBTA's hiring decision. Maj. Op. –––– n.22. It then proceeds to contend that he cannot rely on that theory because he failed to plead it expressly in his complaint. Id. But, we are at the summary judgment stage, see Ríos-Campbell v. U.S. Dep't of Com., 927 F.3d 21, 25-26 (1st Cir. 2019), and, in any event, no authority indicates that Henderson's general allegation of a Title VII violation due to his having been “subjected to race discrimination” is not itself a claim that race was at least a motivating factor in the MBTA's decision.Moreover, while the MBTA in theory could have contended that the higher scores that these other applicants received show that it would not have hired Henderson even if race had been a motivating factor for its actual decision, such a defense would not itself have defeated his claim of status-based discrimination; it would have merely limited his remedies. See 42 U.S.C. § 2000e-5(g)(2)(B); see also Univ. of Tex. Sw. Med. Ctr. v. Nassar, 570 U.S. 338, 355, 133 S.Ct. 2517, 186 L.Ed.2d 503 (2013). Not surprisingly, therefore, the MBTA argues on appeal only that a fair application of the burden-shifting framework under McDonnell Douglas compels the conclusion that no reasonable juror could find that race played any role at all -- motivating factor or otherwise -- in its hiring decision. As a result, I, like the District Court, focus on that contention.

40.   The MBTA points to a subsequent letter from the federal agencies that it claims shows that it “was in compliance with federal requirements regarding equal-opportunity employment.” But, the subsequent letter did not address the “extremely troubling and disappointing” allegations of widespread complaints and large settlements and, in any case, only expressed approval of changes that the MBTA undertook to address deficiencies in its equal-opportunity program after January of 2013, when Melchionda and Higgins were selected.

LYNCH, Circuit Judge.