This appeal is from a summary judgment for the defendant on a claim of race-based employment discrimination. The key question is whether the plaintiff presented sufficient evidence that the defendant’s stated reasons for denying him a promotion did not actually motivate its decisions. In our view, the plaintiff did not present such evidence. For that reason, and because the statistical records have not been shown to support a finding of “disparate impact” discrimination, we shall affirm the judgment entered by the district court.

I

The plaintiff, Mario Heath, began working for the Ohio Turnpike Commission as a part-time toll collector in 1987. Because part-time toll collectors are paid less and receive fewer benefits than full-time toll collectors, he applied on several occasions to be reclassified as a full-time collector. It is undisputed that Mr. Heath was qualified for a full-time position.

The Commission administers mathematics tests to part-time toll collectors who want to be reclassified as full-time collectors. It also conducts interviews before deciding which part-time collectors to promote. Mr. Heath first took the math test in February of 1997. receiving a score of 275.1 Of the 61 other part-time toll collectors who took the same test that year,2 only 14 failed to achieve a higher score. Mr. Heath was not interviewed in 1997 and was not selected for reclassification. Ten part-time collectors, none of whom scored lower than Heath on the math test, were promoted in 1997. Each of those collectors is white: Mr. Heath is black.

In March of 1998, Mr. Heath took a different math test in a renewed effort to become a full-time toll collector. This time he scored 100 percent, better than all but 12 of the other 60 part-time collectors who sought reclassification that year. Heath *496was then interviewed by Sharon Isaac, the Commission’s Director of Toll Operations.3

Ms. Isaac perceived Mr. Heath as “very subdued” during the interview. As she later testified, “[h]e didn’t express ... any enthusiasm about the possibility of becoming a full-time employee....” Ms. Isaac’s notes reflect that Mr. Heath did not face her during the interview, but rather sat with his “body turned sideways.” He did not make eye contact with her. It is unlikely that Ms. Isaac’s race had anything to do with Mr. Heath’s peculiar way of presenting himself, because Ms. Isaac too is black.

Ms. Isaac testified that the interview caused her to be concerned about Mr. Heath’s ability to interact with customers. She said it was the interview, coupled with some flaws in his attendance record, that led her not to select Heath for reclassification.

Seven part-time toll collectors, all of them white, were reclassified in 1998. Six of the seven scored lower than Mr. Heath on the 1998 math test (with scores of 77, 89, 89, 94, 94, and 94 percent). Twelve white part-time collectors achieved perfect scores on the math test, but only one of these twelve white applicants was promoted.

Mr. Heath stopped working for the Commission in the summer of 1999. In August of that year. Heath filed an administrative charge of race and sex discrimination. Three months after receiving a right-to-sue letter. Heath sued the Commission in federal district court. In an amended complaint filed in February of 2001, Heath alleged that the Commission’s employment practices, including the use of written mathematics tests to evaluate candidates for full-time positions, discriminated against blacks because of their race.

After a period of discovery, the Commission moved for summary judgment. It argued that Mr. Heath could not prevail under either a “disparate impact” or a “disparate treatment” theory of liability. The district court granted the motion. Relying on an expert witness analysis of statistical data, the court held that the Commission’s selection procedures had not been shown to have a disparate impact on black part-time toll collectors who sought reclassification to full-time status. The court held further that, although Mr. Heath had made out a prima facie case of disparate treatment, the Commission had articulated legitimate, nondiscriminatory reasons for its actions and there was no evidence sufficient to prove that those reasons were pretexts for racial discrimination. Following the entry of final judgment, Mr. Heath filed this timely appeal.

II

Mr. Heath’s primary argument is that he presented evidence sufficient to support a finding of intentional race-based discrimination. He argues secondarily that statistical evidence supports his claim of “disparate impact” discrimination. We review the district court’s grant of summary judgment de novo. See, e.g., Cleveland Branch, N.A.A.C.P. v. City of Parma, 263 F.3d 513, 523 (6th Cir. 2001), cert. denied, 535 U.S. 971, 122 S.Ct. 1438, 152 L.Ed.2d 382 (2002).

A

The Commission conceded in the district court that Mr. Heath had made out a prima facie case of intentional discrimi*497nation under McDonnell Douglas Corp. v. Green, 411 U.S. 792, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973). Heath did so by showing (1) he is a member of a protected class. (2) he is qualified for the job of full-time toll collector, (3) he was not given such a job, although there were several openings, and (4) the jobs went instead to persons outside the protected class. See McDonnell Douglas, 411 U.S. at 802; Nguyen v. City of Cleveland, 229 F.3d 559, 562-63 (6th Cir. 2000). The Commission was therefore required to produce evidence of legitimate, nondiscrimimtory reasons for its refusal to promote Mr. Heath. See McDonnell Douglas, 411 U.S. at 802. This it did by presenting evidence of Heath’s score on the 1997 math test and his performance in the 1998 interview.

In these circumstances, and in the absence of direct evidence that race motivated the Commission’s decisions, Mr. Heath was required to produce evidence sufficient to show that the explanation offered by the Commission is “unworthy of credence.” Texas Dep’t of Community Affairs v. Burdine, 450 U.S. 248, 256, 101 S.Ct. 1089, 67 L.Ed.2d 207 (1981); see Manzer v. Diamond Shamrock Chemicals Co., 29 F.3d 1078, 1083 (1994). Sufficient evidence that an employer’s stated reasons are false, when combined with a prima facie case of discrimination, will in most cases “permit the trier of fact to conclude that the employer unlawfully discriminated.” Reeves v. Sanderson Plumbing Products, Inc., 530 U.S. 133, 148, 120 S.Ct. 2097, 147 L.Ed.2d 105 (2000); see St. Mary’s Honor Center v. Hicks, 509 U.S. 502, 511, 113 S.Ct. 2742, 125 L.Ed.2d 407 (1993). Absent such evidence, on the other hand, a jury will have no adequate basis on which to infer intentional discrimination. See Manzer, 29 F.3d at 1083.

The question before us here is whether Mr. Heath presented evidence that would allow a rational fact-finder to reject the Commission’s stated reasons for denying him a promotion. “To make a submissible case” on the issue of pretext, Manzer teaches, “the plaintiff is required to show ... either (1) that the proffered reasons had no basis in fact, (2) that the proffered reasons did not actually motivate [the adverse action], or (3) that they were insufficient to motivate [the action].” Id. at 1084 (internal quotation marks omitted). We are not persuaded that Mr. Heath has carried this burden.

First, Mr. Heath has not shown that the bases on which the Commission says it acted “never happened.” Id. Heath disputes neither his 1997 test score nor Ms. Isaac’s characterization of his 1998 interview.

Second. Mr. Heath has not demonstrated that his test score and interview performance were insufficient to motivate the Commission’s decisions. A showing that an employer’s stated reasons are insufficient to motivate its actions ordinarily “consists of evidence that other employees, particularly employees not in the protected class” were not adversely treated despite “substantially identical conduct.” Id. It is undisputed that no part-time toll collector with a test score as low as, or lower than, Heath’s was reclassified as a full-time collector in 1997.4 Similarly, Heath has not contended that any part-time collector was promoted despite an interview performance comparable to his own.

Finally, we do not think Mr. Heath has shown that his test score and interview performance did not actually motivate the Commission’s decisions. Unlike the other two methods of establishing pretext (dem*498onstrating that the defendant’s reasons are factually false and demonstrating that the defendant’s reasons are insufficient to justify the adverse action), this method generally involves an indirect attack on the credibility of the proffered reasons. See id. As the Manzer court explained.

“the plaintiff attempts to indict the credibility of his employer’s explanation by showing circumstances which tend to prove that an illegal motivation was more likely than that offered by the defendant. In other words, the plaintiff argues that the sheer weight of the circumstantial evidence of discrimination makes it ‘more likely than not that the employer’s explanation is a pretext.” Id. (emphasis omitted).

Heath’s prima facie case is not sufficient to show that his test score and interview performance did not actually motivate the Commission. If the prima facie case alone allowed a jury to infer pretext, then “[n]o case could ever be culled out after the prima facie stage.” Id.; see Cicero v. Borg-Warner Automotive, Inc., 280 F.3d 579, 589 (6th Cir. 2002). Heath has presented some evidence in addition to his prima facie case, but, in our view, his evidence would not justify a jury in rejecting the Commission’s explanation for its actions.

Mr. Heath points to the fact that, after being passed over in 1997 on the stated basis of a low math score, he was passed over again in 1998 despite a math score of 100 percent. But an acceptable math score was not the sole criterion for promotion, and the record, as we read it, does not show that the Commission was being inconsistent. It is true that the Commission required an acceptable test score as a necessary condition for promotion; neither Heath nor any of the 20 white collectors who scored below 295 on the 1997 math test was reclassified that year.5 It is also true, however, that the Commission never regarded a high test score, standing alone, as a confirmed ticket to promotion. In 1997, five part-time toll collectors (four whites and one black) scored above 400 on the Commission’s math test, but only one of these high-scoring collectors was promoted. In 1998, 13 part-time collectors (12 whites and one black) scored 100 percent on the math test, but, again, only one was promoted. In the light of this evidence, we do not think a rational fact-finder could infer from the Commission’s treatment of Heath’s test scores that the stated reasons for not promoting Heath either in 1997 or 1998 were pretexts for racial discrimination.

Mr. Heath has shown that he had more to recommend him than his score on the 1998 math test. His evidence includes a written performance evaluation for calendar year 1997. wherein Heath was rated an “above average” collector with “excellent” work habits, knowledge, behavior, and availability. It also includes deposition testimony in which Heath’s toll plaza supervisor characterized him as an “excellent collector.” This evidence tends to suggest that the concerns Ms. Isaac developed during and after Heath’s interview — concerns, specifically, about Heath’s interaction with customers and his availability — might not have proved to be well-founded. But it does not suggest, in our view, that Isaac’s concerns were not genuinely held.

For one thing, Ms. Isaac’s concerns had an undisputed factual basis. Mr. Heath’s interview was “unusual” in that he did not make eye contact, did not express enthusi*499asm, and did not “seem to be concerned about what was happening” in the interview room. Notwithstanding the positive evaluations of Heath’s past performance, Isaac was entitled to form and rely upon first-hand impressions of Heath’s demean- or.6

It is significant, in this connection, that Ms. Isaac was obliged to reject a substantial majority of the part-time toll collectors seeking full-time positions in 1997 and 1998. In 1997, 77 part-time collectors applied for only 10 full-time positions. In 1998, 61 collectors applied for seven positions. When only one of every eight or nine applicants can be promoted, it is likely that many qualified — even highly qualified — candidates will be passed over. In this setting, we believe, the rejection of a deserving black applicant does not in itself suggest that racial discrimination motivated his employer. This is all the more true when the applicant has, in the eyes of the decision-maker, shown little interest in the position.

Mr. Heath contends that the evidence reveals a pattern of rejecting black part-time toll collectors who applied for full-time status. The contention is based on records showing that there were no blacks among the 26 part-time collectors who were reclassified between 1997 and March of 2000.

We are not persuaded that these records are probative of pretext. Only three of the 77 applicants for reclassification in 1997 — 3.9 percent — were black.7 Therefore, all things being equal, the probabilities were against a black applicant receiving one of the 10 available positions. The odds were only slightly better in 1998, when four of the 61 applicants — 6.6 percent — were black.8 but just seven positions were available. In 1999, the statistical chance of a black applicant being reclassified was the lowest of the three years: only 3.7 percent of the applicants (two of 53) were black, and nine positions were available. Statistically, black applicants as a group could expect to receive less than one half of a position in each of the three years. In light of that fact, we cannot say that the selection of zero black applicants each year supports an inference that the Commission’s stated reasons for not selecting Mr. Heath were pretexts for intentional discrimination.

Aggregating the data from 1997 through March of 2000 does little to aid Mr. Heath’s case. White part-time toll collectors who applied for reclassification at least once during that time period outnumbered their black counterparts by about 24 to one (122 to 5). If applicants were selected randomly, therefore, it is likely that only one of the 26 available positions would be filled by a black applicant. Common sense, we believe, suggests that Ms. Isaac’s selection of zero black applicants, when a random draw would most likely have selected just one, is insufficient to prove pretext.

*500The testimony of Robert Gitter, Ph.D., Mr. Heath’s expert economist, supports that intuition. Dr. Gitter was asked to assume that four out of 100 applicants were black and that 27 whites and zero blacks were reclassified. Given these facts — which mirror the actual ratio of white applicants to black applicants and include one more position than was actually available — Dr. Gitter preliminarily concluded that the difference between the number of blacks that one would expect to be reclassified and the number that actually were reclassified was not statistically significant.9

Finally, Mr. Heath points to testimony concerning racial epithets used by supervisors and other toll collectors at the toll plaza where he worked. Ms. Isaac, however, was the sole decision-maker with respect to reclassification of part-time toll collectors. Nothing links Ms. Isaac to the epithets in question, so this evidence is not probative of the Commission’s motivation.

In sum, we see nothing in the record that would entitle a fact-finder to reject the Commission’s explanation and find that the likelier explanation was an intent to discriminate against Mr. Heath because he is black. Summary judgment was therefore appropriate on Mr. Heath’s claim of disparate treatment.

B

To make out a prima facie case of “disparate impact” discrimination, Mr. Heath was required to identify a “specific employment practice” that allegedly had an adverse effect on black part-time toll collectors and then to offer statistical evidence sufficient to show that the challenged practice caused the adverse effect. See, e.g., Kovacevich v. Kent State University, 224 F.3d 806, 830 (6th Cir. 2000). We agree with the district court’s conclusion that Heath failed to carry his burden in this regard.

In his amended complaint, Mr. Heath pointed to what he claimed were two specific practices that affected black part-time toll collectors adversely: the use of mathematics tests in evaluating candidates for reclassification and a lack of emphasis on “successful job performance” and length of service when considering candidates for reclassification. Heath did not offer any statistical evidence, however, tending to show that these specific practices (assuming that all of them were in fact “practices”) disadvantaged black part-time toll collectors.

In opposing the motion for summary judgment. Mr. Heath argued more generally that the Commission’s “testing and interview processes” disproportionately blocked black part-time toll collectors from reclassification as full-time collectors. We are far from sure that this satisfied Heath’s burden to “isolat[e] and identify] the specific employment practices that are allegedly responsible for any observed statistical disparities.” Watson v. Fort Worth Bank & Trust, 487 U.S. 977, 994, 108 S.Ct. 2777, 101 L.Ed.2d 827 (1988).

We need not resolve that issue, however, because the statistical disparity noted by Mr. Heath — the reclassification of 26 white and zero black part-time toll collectors between 1997 and early 2000 — has not been shown “sufficiently substantial” to raise “an inference of causation.” Id. at 995. As Heath’s own expert witness concluded (albeit preliminarily), given the racial composition of the applicant pool, the number of black applicants selected for reclassification did not differ significantly from what would be expected absent improper dis*501crimination. No contrary opinion was offered.

Because Mr. Heath did not make out a prima facie “disparate impact” case, summary judgment was appropriate on this theory as well.

AFFIRMED.