dissenting in part, and concurring in part.
I disagree with the Court’s conclusion that Torgerson and Mundell “created the requisite inference of unlawful discrimination under a McDonnell Douglas analysis, including sufficient evidence of pretext.” Ante, at 595. In my view, Torgerson and Mundell do not produce evidence from which a reasonable jury could find that Rochester’s stated reason for not hiring them — that they scored lower in the hiring process than candidates hired — is pretext for discrimination.
There are at least two ways a plaintiff can establish a material question of fact regarding pretext. Wallace v. DTG Operations, Inc., 442 F.3d 1112, 1120 (8th Cir. 2006). A plaintiff may show the employer’s explanation is “unworthy of credence, because it has no basis in fact.” Id. (quotations and citations omitted). A plaintiff may also show pretext directly, by “persuading the court that a [prohibited] reason more likely motivated the employer.” Id.
The Court concludes that the following evidence creates a material question as to pretext: (1) the “hiring of some white candidates with similar or lower qualifications than the appellants,” (2) the subjective nature of parts of the hiring process, and (3) *601Fire Chief Kapler’s reference to Torgerson and Mundell as “unfit.” Ante, at 597-600.
1. Comparison of qualifications
“Where ... the employer contends that the selected candidate was more qualified for the position than the plaintiff, a comparative analysis of the qualifications is relevant to determine whether there is reason to disbelieve the employer’s proffered reason for its employment decision.” Chock v. Northwest Airlines, Inc., 113 F.3d 861, 864 (8th Cir.1997). “[A] comparison that reveals that the plaintiff was only similarly qualified or not as qualified as the selected candidate would not raise an inference of ... discrimination.” Id. .
After quoting Chock, the Court states that “Torgerson and Mundell possessed similar qualifications as the hired candidates.” Ante, at 598. “Similar qualifications” do “not raise an inference of ... discrimination.” Chock, 113 F.3d at 864. See also Lidge-Myrtil v. Deere & Co., 49 F.3d 1308, 1311 (8th Cir.1995) (“Although [plaintiff] does possess the experience and some of the other qualities essential for success in the position, this does not suffice to raise an inference that [the employer’s] stated rationale for giving the position to another is pretextual”); Pierce v. Marsh, 859 F.2d 601, 603 (8th Cir.1988) (“The mere existence of comparable qualifications between two applicants ... alone does not raise an inference of ... discrimination.”).
In Pierce, the Army’s personnel office prepared a list of job candidates in rank order. The Army hired Stokes (a black male) and Webb (a white female), who ranked ahead of Pierce (a black male). Pierce sued for race discrimination; the Army explained that Webb was more qualified, as reflected by her higher rank on the list. The district court granted summary judgment to the Army.
On appeal, Pierce stressed an exhibit indicating “that his qualifications far outweigh Webb’s qualifications, thus demonstrating that the Secretary’s legitimate nondiscriminatory reason for the decision to hire Webb serves as a mere pretext. ...” Id. at 603. Reviewing the exhibit, this court concluded that Pierce and Webb had “relatively similar qualifications,” and “neither candidate appears better qualified for the foreman position.” Id. “Relatively similar qualifications” were not enough for Pierce to survive summary judgment. See id. at 604 (“... Pierce failed to provide any evidence from which a rational trier of fact could infer that the selecting committee’s articulated nondiscriminatory reason for hiring Webb over Pierce was overcome by any evidence establishing that reason as pretextual.”).
Here, like Pierce, Torgerson and Mun-dell appeared on an eligibility list, ranked below the hired candidates. At best, they have “relatively similar qualifications” to some hired candidates. As in Pierce, “relatively similar qualifications” do not create a material issue of fact as to pretext.
2. Subjectivity of hiring process
“Although subjective [hiring] procedures are susceptible of discriminatory abuse and require close scrutiny, subjectivity alone does not render an employment decision infirm.” Brooks v. Ameren UE, 345 F.3d 986, 988 (8th Cir.2003) (internal citation omitted).
Torgerson and Mundell emphasize that the panel interviews account for 40 percent of applicants’ scores, and stress the wide range of scores in these interviews. But they fail to provide any evidence that the interviews were discriminatory. Panels consisted of one interviewer from Rochester’s human resources department, one from the Fire Civil Service Commission, *602and one from the Fire Department. The interviewers received a list of (human-resources-prepared) questions to ask of all candidates and were given scoring criteria to evaluate responses. Interviewees received, in advance, the list of potential questions. Even Torgerson and Mundell agree that the questions asked were what they anticipated based on the list of possible questions, and that none of the questions was inappropriate.
After stating that the interviews “warrant [close] scrutiny,” the Court concludes that “deciding the efficacy of Rochester’s steps to ensure nondiscriminatory evaluation is, on the facts in this record, better reserved for the jury,” and “the material fact of potential discriminatory abuse of the subjective interview process remains.” Ante, at 598. The Court does not refer to any evidence that discrimination took place11 — only to the “potential” for discrimination in the interviews. Plaintiffs cannot survive summary judgment by identifying a part of the hiring process where there is only “potential” for discrimination. Again, “subjectivity alone does not render an employment decision infirm.” Brooks, 345 F.3d at 988.
In Pierce, as in this case, hiring officials ranked qualified candidates on a list based on a combination of factors. This court said:
The selecting officials testified by affidavit that subjective factors, i.e., their evaluation based on personal observation of the candidates, also entered into the selection decision. Even if these subjective reasons could be rejected on credibility grounds, such a rejection of that evidence would not add anything to the lack of a showing of pretext by Pierce.
Pierce, 859 F.2d at 603-04. Here, the fact that parts of the hiring process were more subjective does not add anything to the lack of showing of pretext by Torgerson and Mundell.
The Court notes that “neither this court nor the district court can determine from the record whether Torgerson and Mun-dell received lower scores than other candidates who gave similar answers [in the panel interview],” and concludes the “reasons for their absence likely involve fact questions best resolved” at trial. Ante, at 598. The burden, however, is on Torgerson and Mundell to provide evidence of pretext. See, e.g., Ramlet v. E.F. Johnson Co., 507 F.3d 1149, 1153 (8th Cir.2007) (If the defendant provides a non-discriminatory reasons for its decision, “the presumption [of discrimination] disappears, and the burden shifts back to the plaintiff to show that the proffered reason was pretext for ... discrimination”); Pope v. ESA Serv., Inc., 406 F.3d 1001, 1007 (8th Cir.2005) (“[T]he ultimate burden falls on [plaintiffs] to produce evidence sufficient to create a genuine issue of material fact regarding whether [the employer’s] proffered nondiscriminatory justifications are mere pretext for intentional discrimination.”). Here, a reasonable jury could not infer pretext from the absence of this evidence. See Pineda v. UPS, 360 F.3d 483, 487 (5th Cir.2004) (“To satisfy this burden, the plaintiff must offer some evidence that permits the jury to infer that the proffered *603explanation was a pretext for discrimination. The trier of fact may not simply choose to disbelieve the employer’s explanation in the absence of any evidence showing why it should do so.”) (internal quotations and alterations omitted).
Torgerson and Mundell’s objections to the process creating the eligibility list are unpersuasive. Rochester’s explanation of its hiring decisions has a “basis in fact.” See Wallace, 442 F.3d at 1120.
3. Fire Chief Kapler’s description of Torgerson and Mundell as “unfit”
Torgerson and Mundell may also show pretext by “persuading the court that a [prohibited] reason more likely motivated the employer.” Id.
City Councilmember Patrick Carr testified that Fire Chief Kapler said he did not hire Torgerson and Mundell because he “found them unfit.” At his deposition, Fire Chief Kapler explained his remark, distinguishing “fit” from “qualified.” The Court states that “the district court credited Kapler’s explanation of his use of the word ‘unfit,’ ” which “is not the only reasonable inference from the facts.” 12 Ante, at 599.
Nothing in the record suggests Fire Chief Kapler’s use of “unfit” was based on national origin or gender. Minnesota law requires “public competitive examinations to test the relative fitness of applicants.” Minn.Stat. § 420.07(2) (emphasis added). “While we are required to make all reasonable inferences in favor of the nonmoving party in considering summary judgment, we do so without resort to speculation.” Twymon v. Wells Fargo & Co., 462 F.3d 925, 934 (8th Cir.2006). “Facially race-neutral [or gender-neutral] statements, without more, do not demonstrate [discriminatory] animus on the part of the speaker.” Id.See also Hannoon v. Fawn Eng’g Corp., 324 F.3d 1041, 1047 (8th Cir. 2003) (“Because the comments regarding body odor did not suggest any reference to race or national origin, we are unwilling to hold such comments reasonably capable of supporting an inference of discriminatory intent.”).
4. Other evidence
Although the Court does not specifically rely on it, other evidence the Court references does not support an inference of discrimination, when read in context. See Twymon, 462 F.3d at 934. Councilmember Carr testified that Fire Commissioner John Withers — who was on the panel that interviewed Torgerson — said he recommended a convicted felon for a firefighting position because he is “a big, strong guy.” As Carr makes clear, this statement came in the context of a conversation about a particular male candidate, not about other applicants.
Councilmember Carr also testified about an exchange with Fire Commissioner Roger Field:
I said — the first question I asked [Field] was are you aware of all the terms and conditions of the SAFER grant. And then he said, “What do you mean?” And I said, “Well, they stipulated you hire women and minorities.” And he said, “I knew nothing of that.” He said, “Had I known, I would have recommended that the City not take the grant.” He said the City should never have taken the grant if that was the stipulation.
*604As the district court noted, Carr’s description “of the SAFER grant failed to mention the qualifying language ‘to the extent possible’ .... Testimony that Field recommended against taking a grant that ‘stipulated’ the City hire women and minorities, regardless of relative qualifications, is not evidence of discriminatory animus.” As the district court concluded, “Field’s statement is unrelated to any challenged step in the decisional process.... ”
Torgerson and Mundell stress that when Fire Chief Kapler conducted the final interviews of candidates from the top of the eligibility list, he looked only for “red flags,” but when interviewing those at the bottom (them), he looked for “something about [them] that would elevate them to the level of being better than the candidates who were at the top of the list.” This cannot reasonably be construed as discriminatory because, despite moving on to the fire-chief interview, Torgerson and Mundell retained their ranks on the eligibility list. As the Court recognizes, “although all candidates on the eligibility roster meet the minimum qualifications for the firefighter position, those at the top of the list are recognized as more qualified for the position than those at the bottom of the list.” Ante, at 589.
* * *
Torgerson and Mundell fail at step three of the McDonnell Douglas analysis: they do not offer evidence from which a reasonable jury could conclude Rochester’s reason for not hiring them was pretextual.
I would affirm the judgment of the district court in all respects, and I therefore dissent in part.
. The Court notes that Torgerson scored better than one, and equal to another, hired candidate on the written and agility tests (and Mundell almost scored equal to the lowest-scoring hired candidate). Ante, at 595. But the fact remains: "At the end of the objective written and agility phases ... Torgerson was ranked 41st and Mundell was ranked 46th out of 48 candidates.” Ante, at 590. Thus, the interview scores significantly affected the ranking of other candidates. The interview scores did not significantly affect the rankings of Torgerson or Mundell, who ranked low before and after the interviews.
. The Court notes statistical evidence in support of its conclusion. Ante, at 599 n. 10. '‘[SJtatistical evidence will rarely suffice to rebut an employer's legitimate, nondiscriminatory reasons for a particular adverse employment action.” Bogren v. Minnesota, 236 F.3d 399, 406 (8th Cir.2000).