ORDER
In his action, plaintiff William Kidd, III, who is African-American, alleges that the Illinois State Police (“ISP”) unlawfully discharged him on the basis of his race, in violation of Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e-2(a). The case was tried to the bench and resulted in a finding for the ISP. Kidd v. Illinois State Police, 1997 WL 361140 (N.D.Ill. June 20, 1997). Kidd appealed. In an opinion dated January 12, 1999, we remanded this case to the district court for further consideration. Kidd v. Illinois State Police, 167 F.3d 1084 (7th Cir.1999). We asked Judge Rosemond, who tried the case, to address certain evidence in the record that stood in apparent conflict with the testimony of a key defense witness, whom Judge Rosemond had found to be credible. See id. at 1097-1101. We also asked Judge Rosemond to give further consideration to statistical evidence indicating that African-American probationary troopers (like Kidd) had been terminated at much higher rates than their white counterparts. Id. at 1101-02. We retained jurisdiction over Kidd’s appeal while awaiting the district court’s supplemental findings.
After taking additional submissions from the parties, Judge Rosemond issued an order addressing the evidence that we asked him to consider on remand. Kidd v. Illinois State Police, 138 F.Supp.2d 1047 (N.D.Ill.2001). We in turn invited the parties to file supplemental briefs addressing that order.
I.
Our previous opinion contains a complete recitation of the relevant facts. Our factual and procedural summary here shall be more brief.
The Illinois State Police (“ISP”) discharged Kidd shortly before he completed *460a year-long period of training to become a trooper due to Kidd’s sub-par performance in report-writing and radio communications. Kidd had been exhibiting difficulty with his report-writing from the beginning of his training; his difficulty with radio communications (which primarily involved transposing license plate numbers) did not become apparent until he began supervised patrols later on in the training process. Several months prior to this discharge, the ISP sent Kidd for professional evaluation and ascertained that Kidd had a learning disability, or something close to it; and approximately three weeks prior to his discharge, Kidd had entered into an agreement committing himself to the remedial training he needed to compensate for this apparent disability. But because Kidd’s initial, probationary period of employment was nearly at an end by this time and Kidd had not yet demonstrated an ability to engage in accurate written and radio communications, the ISP was reluctant to retain him in its employ and decided instead to discharge him.
Kidd subsequently filed this Title VII action alleging that his discharge was discriminatory. Among other contentions, Kidd asserted that the ISP had intervened much earlier in the training process and much more constructively on behalf of another trainee in Kidd’s class, Robert Tucker, who also had evident writing problems from the outset of training (later attributed to a learning disability) but who was white. Tucker, like Kidd, was terminated at the end of his probationary year because he had not yet managed to correct that problem. By the time of Tucker’s termination, however, a special “Field Training Program” comprised of both tutoring and a variety of special workplace accommodations for Tucker had been in place for more than six months. Six months, coincidentally, was what Kidd’s evaluators and supervisors thought he would need in order to correct his problems with writing and radio usage. But, as we have noted, Kidd’s apparent learning disability was not identified until late in his training, and a remedial program for him was not put into place until weeks before the decision was made to discharge him. Consequently, Kidd did not have the opportunity that Tucker did to attempt to overcome his apparent disability. Kidd contends that the marked difference in treatment was due to the difference in their races.
After conducting a bench trial, Judge Rosemond found that the ISP had not engaged in racial discrimination. Although the evidence left no doubt that the ISP had endeavored to identify and address the source of Tucker’s problems with report-writing much sooner than it did in Kidd’s case, Judge Rosemond found that the ISP had intervened earlier on Tucker’s behalf because it had perceived his writing deficiencies to be more pronounced, and more in need of special remediation, than it did Kidd’s. In making this finding, Judge Rosemond relied in the main upon the testimony of ISP Sergeant Frank DeBerry. DeBerry was the statewide supervisor of Field Training Officers (who in turn provide field training to probationary officers like Kidd), and he also taught report writing to both Kidd and Tucker during the initial phase of their training at the ISP academy. DeBerry was also the individual who decided that Tucker ought to be evaluated for potential learning deficiencies and who, after Tucker’s learning disability was identified, helped to develop the special Field Training Officer program for Tucker. DeBerry was, in short, “the catalyst for the divergent paths taken by Kidd and Tucker[.]” Kidd I, 1997 WL 361140, at *4. DeBerry testified at trial that although both Kidd and Tucker had writing problems, he believed that Tucker’s problems were worse than Kidd’s, *461thus necessitating more specialized intervention.
I thought that with Trooper Tucker, there was very little if any hope of him ever writing a police report that would not be embarrassing to the State Police. As regard[s] to Kidd, it was my belief that he needed work but that our field training officers, upon graduation [from first phase of training at the ISP academy], could help him get to an acceptable performance level in that area.
Tr. 174. Judge Rosemond believed DeBerry, and DeBerry’s explanation for the earlier intervention on Tucker’s behalf supplied the foundation for Judge Rosemond’s finding that the ISP had not delayed seeking assistance for Kidd on account of Kidd’s race. See Kidd I, 1997 WL 361140, at *4.
Kidd argued on appeal that DeBerry’s testimony stood in conflict with other evidence indicating that Tucker and Kidd had equally poor writing skills, and after reviewing the record, we agreed that a remand was required so that Judge Rosemond might address the conflict. Specifically, we asked Judge Rosemond to consider three pieces of evidence. First, Tucker and Kidd had been given comparable marks in their writing skills during weeks three through twenty of their classroom training at the ISP academy. At times, in fact, Tucker’s writing skills were rated a bit higher than Kidd’s were. Second, both Kidd and Tucker failed report-writing at the academy, even after they were given remedial instruction in that area. Indeed, Kidd and Tucker were the only cadets in their class of eighty-two people to fail in that subject. Third, copies of an essay Kidd and Tucker had each written during their first week at the ISP academy (entitled Why I Want To Be A State Police Officer) revealed that both men had marked problems with spelling, grammar, and sentence structure. Each of these three bits of evidence supplied a fairly objective measure of the men’s writing abilities, and collectively the evidence suggested that Tucker and Kidd had similarly poor writing skills — a notion that stood in obvious conflict with DeBerry’s assessment that Tucker was significantly worse off as a writer than Kidd. We believed it was necessary for Judge Rosemond to explicitly address this evidence, and its seeming conflict with DeBerry’s testimony, before we could in turn decide whether Judge Rosemond’s decision to credit DeBerry was clearly erroneous. 167 F.3d at 1097-1101.
We also asked Judge Rosemond to consider certain statistical proof that Kidd had offered in support of his claim of racial discrimination and which the judge had disregarded in his initial opinion. Kidd had presented data indicating that whereas African-Americans comprised seventeen percent of the nine cadet classes that entered the ISP academy from June 1986 through November 1990, African-Americans accounted for sixty-one percent of the individuals discharged from those same classes. Noting that statistics can be probative, albeit not necessarily determinative, of whether discrimination was at work in an individual case of discrimination, 167 F.3d at 1101 n. 16, we asked the court to reconsider the statistics and to “take them for what they are worth,” id. at 1102.
II.
On remand, Judge Rosemond revisited the case with an eye to the evidence we flagged in our opinion. The judge accepted supplemental evidence from Kidd concerning the statistical data and supplemental briefing from both parties as to the import of that data and the other evidence we had asked the judge to consider on remand. Upon reconsideration, Judge Rosemond again concluded that DeBerry’s *462testimony regarding the differences he perceived between Kidd and Tucker was credible, and that the delay in offering remedial training and support to Kidd was not due to Kidd’s race.
A.
The judge first addressed the Why I Want To Be A State Police Officer essays that Kidd and Tucker submitted during their first week of instruction at the ISP academy. Judge Rosemond found that Kidd’s essay reflected more structure: it had an introduction, a body, and a conclusion, and its ideas were addressed in discrete paragraphs. Tucker’s essay, by contrast, amounted to one unending paragraph that rambled among ideas in a stream-of-consciousness style. Further, Kidd’s essay addressed more sophisticated themes (for example, making a difference in the world, helping people to respect the constitutional rights of others, and police misconduct) and reflected a logical progression (from Kidd’s childhood to his eventual decision to become a trooper). Kidd’s choice of words and employment of a simile also reflected greater sophistication. Finally, although both essays contained many spelling errors, Kidd’s misspellings were not consistent. For example, within the same essay, Kidd sometimes spelled the word “enforcement” correctly and sometimes he did not. This suggested that Kidd had simply failed to proofread his work. 138 F.Supp.2d at 1049-50.
The record also contained the multiple revisions of the “Why I Want to Be a State Police Officer” essay that both Kidd and Tucker had submitted after they were corrected by their supervisors. In Judge Rosemond’s view, the revisions also reflected significant differences in the men’s abilities. Although Tucker’s successive drafts reflected improvement, they continued to reflect basic difficulties with sentence completion, word choice, and spelling. Tucker’s final draft, for example, contained at least eleven spelling errors. Errors also persisted in Kidd’s initial revisions, but his later efforts showed marked improvements. Kidd’s final draft contained but one error (in apostrophe placement) that the reviewer had not even noticed. None of Tucker’s revisions came close to being error-free. Finally, although Tucker and Kidd both failed to correct certain errors even after they were flagged, in Kidd’s case, the failure was in some instances more prolonged. The original draft of Kidd’s essay, for example, contained an incomplete sentence that was flagged no fewer than six times before Kidd managed to correct it. From this the judge inferred that Kidd was either stubborn or reluctant to correct errors. Id, at 1050-53.
Several other circumstances were, in Judge Rosemond’s view, consistent with the notion that the deficiencies in Kidd’s writing were due as much to his lack of motivation as to an inability to write well. First, despite an invitation to ask his instructor for assistance, Kidd apparently had never asked for help in correcting the errors in his essay. Second, there was some evidence that Kidd was resistant to criticism. The judge pointed out that when asked at trial about a report that he occasionally got excited during traffic stops and relayed erroneous information, Kidd was reluctant to acknowledge the criticism as valid. Third, although Kidd continued to seek employment in law enforcement after his discharge, he made no independent effort to get to the bottom of the deficiencies that the ISP had cited as the reason for his discharge. Id. at 1053.
Regarding the significance of the Why I Want To Be A State Police Officer essay vis á vis DeBerry’s assessment of the relative writing abilities of Kidd and Tucker, Judge Rosemond made two additional ob*463servations. Having taught report-writing to both cadets over the course of twenty weeks, DeBerry had seen many other examples of their written work and also had the advantage of having witnessed their progress, attitudes, and receptivity to criticism. His judgment as to their relative capabilities thus carried much more weight, in the judge’s view, than a single essay that the cadets had written during their first week of training. Id. at 1054. Moreover, to the extent those essays seemed to be inconsistent with DeBerry’s assessment, he was never asked while on the witness stand to review the essays and comment on the inconsistency. Id.
B.
Judge Rosemond next considered the marks that Tucker and Kidd had been given for their writing skills during their training at the ISP Academy.1 The judge acknowledged that the low ratings the two men had received in this area showed that both were doing poorly. However, in the judge’s view, the comparable marks did not demonstrate that both men were performing to the extent of their abilities. Kidd had demonstrated that he was capable of improving his written work, but apparently he did not always apply himself appropriately. Tucker, on the other hand, was nearly illiterate. Thus, it was DeBerry’s impression that Kidd could do better than he was (and better than Tucker was doing) but that he was not applying himself. DeBerry knew, for example, that Kidd had on occasion failed to comply with requests that he revise and re-submit various writing assignments. Accordingly, it was reasonable for DeBerry (who again had observed both cadets first-hand over the course of their training at the academy) to perceive that Tucker had a more evident need for special assistance than Kidd. Id. at 1057.
This was [DeBerry’s] judgment call and his impression. There’s no indication that it was made with any racial bias towards Kidd. It was his impression that Tucker was doing the best he could and could not do too much better.
Id.
C.
In the course of discussing the cadets’ written work and marks, Judge Rosemond cited a number of other circumstances that distinguished Kidd from Tucker. In the judge’s view, these circumstances lent further support to the notion that DeBerry genuinely perceived a need to intervene on Tucker’s behalf at an earlier date in the men’s training and that neither DeBerry nor the ISP purposely withheld such assistance from Kidd on account of Kidd’s race.
First, after Tucker (like Kidd) had failed his remedial coursework in report-writing, DeBerry had recommended Tucker’s termination&emdash;a fact that suggested DeBerry truly did view Tucker as being less able to bring his writing up to snuff. (DeBerry was overruled by his colleagues, however.) Faced with a candidate he believed could not improve in the normal course of troop- er training), DeBerry concluded that out- side intervention was warranted. On the other hand, DeBerry remained convinced that Kidd could still get up to speed with help from the Field Training Officers who would oversee the next phase of his train- ing. In Judge Rosemond’s view, this was a plausible assessment. Whereas Tucker had only a high school education, Kidd had as he *464a college degree. Furthermore, during the second phase of training, trooper candidates are taken out of the classroom and paired with Field Training Officers for hands-on experience in the field. DeBerry believed that once he was exposed to actual law enforcement duties, Kidd would come to realize how important his writing skills were to the job of a state police officer and apply himself accordingly. As an experienced instructor, DeBerry also knew that some candidates who perform below par in the classroom excel in the field under the more nurturing tutelage of their Field Training Officers. DeBerry thus could reasonably have believed that Kidd’s writing would improve to an acceptable level during Phase II without tutoring or other specialized intervention. Id. at 1054-56.
Second, the facts suggested to Judge Rosemond that Kidd suffered from a lack of motivation. Again, Kidd had a college degree; and as Kidd’s final version of the Why I Want to be a State Police Officer essay demonstrated, he was capable of writing an (almost) error-free paper. Yet, as the persistent errors in the earlier versions established, Kidd was either unwilling to follow directions and/or unwilling to ask for help. DeBerry testified that he also knew of instances in which Kidd had failed to complete revisions of his written assignments as requested. Id. at 1055-56. “Thus, from Sergeant DeBerry’s perspective, as well as that of the other instructors, Kidd had the ability to do better, but simply was not motivated to do so.” Id. at 1056.
Third, Judge Rosemond pointed out that the problem Kidd had with transposing letters and numbers, and the left-to-right confusion that a professional evaluator later noted, did not become evident until Phase II of Kidd’s training. DeBerry thus did not have the benefit of those clues at the conclusion of Phase I, when he made the decision to seek outside help for Tucker but not for Kidd. Id. at 1056-57.
D.
Finally, Judge Rosemond considered the statistical data indicating that probationary troopers who were African-American were discharged at a much higher rate than other candidates, but found that the data did not suggest that discrimination had occurred in Kidd’s ease. The judge acknowledged that the disparity between the two termination rates was “striking.” Id. at 1059. He also credited the view propounded by Kidd’s expert that the difference was not due to chance. Id. at 1060. However, the record told the court nothing about the African-American candidates who had been discharged — for example, their test scores, field performance ratings, previous education and experience, or the articulated reasons for their discharge. Consequently, it was impossible for the court to determine whether the higher termination rate for African-American candidates might be attributable to factors other than race. Id. Moreover, Judge Rosemond was convinced that even if the ISP were discriminating against African-Americans generally, Kidd’s discharge clearly was not discriminatory. At the time of his termination, Kidd was not qualified for employment as a trooper given his difficulties completing written reports and conveying accurate information by radio. Kidd had been given numerous opportunities prior to his termination to address the shortcomings in his written work, which had been noted at the outset of his training. Although the ISP did not attempt to identify and address the source of Kidd’s persistent writing problems as soon as it did in Tucker’s case, that was due to DeBerry’s assessment that Kidd was not so much in need of help as Tucker was, and Judge Rosemond reiterated that there was no reason to believe that DeBer*465ry's assessment, even if mistaken, was motivated by racial animus. To the extent that Kidd's tendency to transpose letters and numbers suggested a need for professional intervention, Judge Rosemond reiterated that this tendency had not emerged until much later in Kidd's training, when he began patrols in the field. Id. at 1060-61.
III.
Our task at this juncture is a limited one. We remanded this case with the request that the district court address (1) objective evidence that appeared to conflict with Sergeant DeBerry's testimony that the ISP sought out help for Tucker sooner than it did for Kidd because Tucker's written work demonstrated a more pronounced need for professional intervention than Kidd's did; and (2) the statistical data reflecting a markedly higher termination rate for African-Americans than for other cadets. Judge Rosemond's order on remand reflects a thorough consideration of both categories of evidence. The inferences that he drew from the evidence, and the findings that he made-including in particular his decision to credit DeBerryare all factual in nature. See generally Anderson v. Bessemer City, N.C., 470 U.S. 564, 573-75, 105 S.Ct. 1504, 1511-12, 84 L.Ed.2d 518 (1985). Accordingly, our review is for clear error. FED. R. CIV. P. 52(a); see Anderson, 470 U.S. at 573-75, 105 S.Ct. at 1511-12. A factual finding is clearly erroneous "when, although there is evidence to support it, the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed." United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 542, 92 L.Ed. 746 (1948). So long as the finding amounts to a permissible construction of the evidence, we will not deem the finding erroneous, even if another view of the evidence is possible. See Anderson, 470 U.S. at 573-74, 105 S.Ct. at 1511.
Our decision to remand the case was not a signal that we believed any of Judge Rosemond's findings to be erroneous. Rather, because certain objective evidence stood in seeming conflict with his decision to credit a key witness-DeBerry-we believed it necessary for the judge to expressly address that evidence and to reconcile it with DeBerry's testimony. Kidd, 167 F.3d at 1101 (citing Griffin v. City of Omaha, 785 F.2d 620, 628 (8th Cir.1986)). "Only with the benefit of Judge Rosemond's thinking in this regard can we exercise meaningful review of his critical factual findings." Id. Similarly, with respect to the statistical data, we did not compel the judge to give the figures any particular weight. We simply asked him to consider the statistics and to "take them for what they are worth." Id. at 1102.
A.
We turn first to the objective evidence. As summarized above, Judge Rosemond considered on remand the evidence that we believed cast doubt upon DeBerry's testimony as to the perceived need to intervene sooner on Tucker's behalf than on JUdd's. Even in light of this evidence, Judge Rosemond found no reason to believe that DeBerry did not genuinely perceive that JUdd could correct his writing deficiencies without professional intervention. After reviewing the record for a second time, we cannot say that Judge Rosemond's decision to credit Tucker was clearly erroneous.
Fir.s4 as Judge Rosemond emphasized, DeBerry had a unique perspective on the relative writing abifities of Kidd and Tucker. DeBerry taught report-writing to both cadets in the first phase of their training and, as a result, arguably was better situated than any other witness to assess not *466only the shortcomings in each man’s writing, but the ability of each cadet to correct those deficiencies. Thus, to the extent that other witnesses or other evidence suggested, in seeming conflict with DeBerry’s testimony, that Kidd and Tucker were equally poor writers, DeBerry’s assessment was not necessarily incredible.
Second, although Tucker and Kidd were given roughly equivalent scores on their written work at the academy and although ultimately they both failed report-writing, their comparable performance did not necessarily mean that Kidd and Tucker wrote poorly for the same reasons or that they needed the same type of assistance in order to improve. The essence of DeBerry’s judgment as to the distinction between the two cadets was not necessarily that Kidd was a better writer than Tucker, but that Kidd might be able to improve his writing with the one-on-one supervision and inspiration that would be available to him in the second, hands-on phase of trooper training, while Tucker appeared to need more specialized assistance. Individual writing problems can indeed have different causes and different solutions. We can find nothing in the record that precluded DeBerry from believing, at the conclusion of Kidd’s and Tucker’s classroom training, that Kidd’s writing might improve with help from his Field Training Officer, while Tucker required professional intervention.
Third, although the Why I Want To Be A State Police Officer essays that Kidd and Tucker submitted during the first week of training reveal that both men had serious problems with spelling, grammar, and sentence completion, there are, as Judge Rosemond pointed out, some differences evident in both the original drafts and the revisions. Kidd’s original essay had a paragraph structure and a greater degree of focus that made it much easier to read than Tucker’s essay. Despite persistent spelling errors, Kidd’s use of language, and some of the ideas he was expressing (e.g., protecting the constitutional rights of others and working for a non-corrupt agency) were arguably a bit more sophisticated than Tucker’s, although the difference was by no means dramatic. More significant, perhaps, is the fact that although both cadets submitted many rewrites of their essays, Kidd ultimately was able (with a great deal of help) to submit a version nearly free of errors, whereas even Tucker’s final draft contained a number of errors.
Fourth, there is some indication that Kidd’s errors were partially due to carelessness. As Judge Rosemond noted, Kidd’s spelling errors were not always consistent. The fact that he spelled certain words correctly at some points but not others within a single piece of writing arguably bespeaks a failure to proofread rather than an inability to spell.
Fifth, Kidd had earned a college degree, while Tucker had not progressed beyond high school in his formal schooling. The fact that Kidd had managed to graduate from college suggests not only that Kidd was better educated than his classmate, but also that his prospects for improving his writing were better than Tucker’s. In other words, although both men had failed report writing at the academy, Kidd obviously had sufficient learning skills to graduate from college. One might reasonably conclude that if Kidd had the ability to obtain a college degree, he had the innate ability to bring his spelling, grammar, and syntax to an acceptable level without special tutoring.
As Kidd points out, some of the judge’s observations on remand may not have the full support of the record. For example, based on Kidd’s failure on five different occasions to correct an incomplete sentence in his Why I Want To Be A State Police Officer essay, Judge Rosemond con-*467eluded that Kidd’s persistent errors were due to stubbornness or a reluctance to acknowledge his errors and follow instructions. See 138 F.Supp.2d at 1053 (“the [various] rewrites are not reflective of five different endeavors to correct an incomplete sentence[;][r]ather, they are reflective of no effort having been made on five different occasions”) (emphasis in original); id. (“Apparently, Kidd viewed for five times that the incomplete sentence was merely his instructor’s personal opinion on the matter and simply refused to change it.”). Yet, the record tells us nothing about the reasons for Kidd’s failure to correct the sentence that the judge cited. In fact, Kidd actually did attempt, albeit unsuccessfully, to revise the sentence on several occasions; those attempts are facially inconsistent with the notion that Kidd was unwilling to acknowledge his mistake or reluctant to try to correct it. Kidd may simply not have understood the precise nature of his error or the proper way in which to correct it. Inferring obstinance from the prolonged inability to correct one writing error may therefore be somewhat of a stretch. On the other hand, Kidd’s failure to seek assistance in correcting such a persistent error, although one of his instructors had extended an offer of help to him, may indeed reflect a lack of initiative, as Judge Rosemond also thought. See id. at 1053,1056.
The judge also indicated that Kidd’s instructors perceived him to be in need of inspiration rather than tutoring. Id. at 1056, 1057. This may be putting words in their mouths. DeBerry did testify that, at the time he made the decision to seek outside help for Tucker but not Kidd in November 1988, he knew of certain instances in which Kidd failed to comply with certain requests to submit revisions of his written work. Tr. 180-81. We readily agree that DeBerry’s testimony would support the inference that DeBerry, at least, perceived Kidd not to be making every effort possible to improve his writing. Yet, we can find no support in the record for the notion that there was shared perception among Kidd’s other instructors by this point in time that Kidd suffered from a lack of motivation.2 But this may be quibbling. If DeBerry knew from his own experience and/or the reports of other instructors (his testimony is not exactly clear on this point) that Kidd occasionally had not completed revisions as requested, then DeBerry — who on behalf of the ISP decided to seek professional intervention for Tucker but not Kidd— could have reasonably believed Kidd was in greater need of motivation than he was of a tutor.
In any case, these asserted discrepancies in the district court’s observations do not fatally undermine the decision to credit *468DeBerry’s testimony. Notwithstanding the evidence suggesting that Kidd and Tucker were performing at equivalent levels in writing when DeBerry made the decision to seek professional assistance for Tucker but not Kidd, the record does provide some support for the notion that Kidd and Tucker were not equally situated in terms of their prospects for becoming better writers. Kidd had a college education, his writing was better in some respects than Tucker’s although the two cadets made many of the same types of errors, and Kidd demonstrated a superior ability to (eventually) correct most of those errors. These differences, and DeBerry’s unique familiarity with the relative writing abilities of Kidd and Tucker, rendered the distinction he drew between the two cadets plausible, if not compelling. Moreover, as the district court pointed out, even if DeBerry’s assessment was wrong, nothing indicates that the error was deliberate or otherwise colored by racial bias. 138 F.Supp.2d at 1057; see generally Russell v. Acme-Evans Co., 51 F.3d 64, 68 (7th Cir.1995) (distinguishing pretext from mistake). Consequently, we must defer to the district court’s decision to credit DeBerry’s testimony.
B.
We also find no reason to question the district court’s findings in light of the statistical evidence relating to the comparative rates of discharge for African-American and other cadets. This evidence does reveal a dramatically higher discharge rate for African-American candidates, as we noted in our previous opinion and as the district court acknowledged on remand. The record does not tell us anything, however, about the reasons given for the discharge of the African-American candidates, nor does it shed any light on whether the discharged African-American candidates were situated similarly with their classmates in such key respects as their education, previous experience, classroom performance, performance in the field, and so on. Without such evidence, it is impossible to assess whether, and to what extent, the higher discharge rate for African-American candidates may have been due to factors other than their race. See, e.g., Chavez v. Illinois State Police, 251 F.3d 612, 638 (7th Cir.2001) (“The statistics proffered must address the crucial question of whether one class is being treated differently from another class that is otherwise similarly situated.”). At the same time, the statistical proof, although it reflects a striking disparity in the discharge rate for African-American candidates, does not necessarily reveal whether discrimination was at work in the discharge of Kidd. The district court entertained the possibility that the ISP was discriminating against African-American cadets and probationary troopers generally, but nonetheless concluded that ISP had delayed offering Kidd the type of specialized assistance it had offered Tucker for non-discriminatory reasons. 138 F.Supp.2d at 1060-61. Because the ISP did not realize that Kidd required such assistance until late in his training, and because at the time of his discharge Kidd’s written and radio communications were not adequate, the court again concluded that the ISP had not discriminated against him. We find no clear error in that assessment.
IV.
Without question, the ISP intervened much earlier on Tucker’s behalf than it did on Kidd’s. Had the ISP fashioned a remedial program for Kidd in November 1988 (when it did so for Tucker) rather than waiting until February 1989 to begin investigating the source of his difficulties (and later still to assemble a remedial program), Kidd might have brought his written work and radio communications up to an accept*469able level by the time the ISP made its final employment decisions in May and June of 1989. However, DeBerry testified at the time that the ISP created the special program for Tucker, he believed that Kidd was capable of coming up to par without the tutoring and other specialized assistance that was put in place for Tucker. DeBerry articulated his reasons for viewing Tucker and Kidd differently, and the district court credited DeBerry’s testimony on this point. Judge Rosemond has fully considered the evidence that stands in apparent tension with DeBerry’s assessment and has concluded that it was nonetheless plausible for DeBerry to believe that Tucker’s need for intervention surfaced at an earlier date than did Kidd’s need. We cannot say that the judge’s decision to credit DeBerry, or his resulting conclusion that the delay in offering Kidd remedial assistance was not due to Kidd’s race, was clearly erroneous.
We note, as Judge Rosemond has, that when the ISP finally did intervene on Kidd’s behalf, it appears to have done so in a less constructive manner than it did in Tucker’s case. Kidd I, 1997 WL 361140, at *9. Yet, having reviewed the record for a second time, we find no basis to disturb Judge Rosemond’s finding that the differential treatment was not due to Kidd’s race.
For the foregoing reasons, we AFFIRM the district court’s judgment.
. The judge did not separately consider the fact that both Kidd and Tucker failed report-writing at the ISP academy. See 167 F.3d at 1098. However, it is clear from his supplemental findings that he had this fact in mind addressed the plausibility of DeBerry's testimony that Tucker was in greater need of remedial intervention than Kidd. See 138 F.Supp.2d at 1054. at 1054.
. We should note that later in Kidd's training, Master Sergeant Carter came to the conclusion that Kidd was unwilling to acknowledge his shortcomings. See Tr. 263 ("At times when I had talked to him about the problem with changing numbers and other problems like that, he would refuse to admit that he had a problem.... It was never his fault.”) This might be construed as additional evidence that Kidd was not sufficiently motivated to improve his performance. But Carter's observation post-dates DeBerry's decision to create a special remedial program for Tucker; consequently, it does not bear on DeBerry's assessment at that particular point in time. It should also be noted that during Phases II and III, other persons at the ISP had very positive impressions of Kidd’s motivation. See, e.g., Tr. 137 (Field Training Supervisor Art Guedel) ("[W]e all said basically the same thing. Trooper Kidd had such a good attitude and really wanted to be a state trooper, we wanted to keep him. He had some areas of deficiency that he was working very hard on. Okay? And to us, that showed desire on his part. And the hardest thing to get is desire, and he already had that.”).