Erlinda Regner, a Chicago librarian, brought suit under Title VII against the *535Chicago Public Library System. She alleged that promotion practices used at the City’s main library had a disparate impact on minorities and offered evidence showing that the percentage of minorities promoted at the main building was substantially less than the percentage of minorities in the total library work force. The City of Chicago, the operator of the library, countered that one could not look solely at promotion statistics for one library location when the system as a whole showed minorities receiving a fair percentage of promotions. The district court agreed and granted the City’s summary judgment motion. 601 F.Supp. 830 (N.D.Ill.1985). While this case is perhaps unnecessarily complicated by the way statistics were offered and legal theories presented, we find summary judgment inappropriate and therefore reverse.
Regner, a Filipino woman, has been employed by the Chicago Public Library since 1977. Between 1979 and the time of filing of this suit, she had worked as a Librarian I in the Business/Science and Technology Department of the main library of the Chicago Public Library system. Plaintiff has a master’s degree in library science and considers business, science and technology her specialties. Prior to filing suit, plaintiff applied for four promotions in the Business/Science and Technology department of the main library. Two of these were to Librarian II positions and two were for Librarian I positions that were to be upgraded to Librarian II. She was denied all these promotions and they were given instead to Caucasians.
Employees throughout the Chicago Public Library system may apply for promotion at any branch. To be a Librarian II, an individual must at least have a master’s degree and one year of experience. To be a Librarian III, an applicant needs a master’s degree and at least two years of experience.
Somewhat subjective criteria are used for evaluating candidates for promotion at the public library. A panel of reviewers— usually including the unit head and the unit head’s supervisor — prepares a job description and an evaluation of the abilities required for the job. The reviewers compose a list of interview questions and, as a panel, interview the candidates. The interviewers are free to ask as many questions as they like and to weight the questions and answers as they deem best. Following the interview, the panel fills out rating sheets and then agrees on a candidate.
Library procedures call for the Director of Library Personnel to review the rating sheets and interview questions and to refer all questions concerning problems with the interview process to the Librarian II and III Committee. In fact, with respect to the four promotions for which Regner applied, the Director of Library Personnel did not review the interview questions prior to their use.
Plaintiff believed that the main library’s promotion practices had a disparate impact on minorities. She therefore filed suit in the District Court for the Northern District of Illinois under Title VII, 42 U.S.C. § 2000e et seq. She introduced evidence showing that of twenty-five promotions and upgrades from Librarian I to Librarians II and III at the main library between 1981 and 1983 twenty-four of those promoted were white while one was black. The percentage of minorities promoted at the main library, plaintiff argued, was greatly disproportionate to the percentage of minorities qualified for promotion throughout the Chicago library system. The case for discrimination against minorities became more apparent, plaintiff contended, when one examined time-in-grade statistics for the library system. Plaintiff presented statistics showing that looking at the last one hundred promotions within the system it took whites an average of 3.9 years to rise from Librarian I to Librarian II. Blacks waited an average of 5.13 years, Hispanic librarians 6.08 years and Asian librarians, such as Regner, 7.08 years before being promoted.
The City of Chicago responded that since any librarian could apply for promotion anywhere in the system and since movement among the branches was fluid, the *536relevant labor market was the library system as a whole. Thus looking at the sixty-one promotions that took place in the library system between 1981 and 1983, minorities were promoted in proportion to their representation in the work force. In fact, the percentage of minorities receiving promotions within the system was slightly higher than their representation in the work force.
At the district judge’s urging, both sides filed motions for summary judgment. Upon reviewing defendant’s motion, plaintiff reached the conclusion that genuine issues of fact remained in controversy. Regner’s Reply Memorandum in Support of Her Motion for Summary Judgment and In Opposition to Defendant’s Motion notes several factual issues that remain in dispute.
For example, the sides did not agree on the relevant labor market. Plaintiff argued that promotions at the main library were more prized plums that made them qualitatively different from promotions at branch libraries. The main library offered the opportunity to specialize, more experienced staffs, larger collections and bigger budgets, even though the pay for Librarian II was uniform throughout the system. Similarly, while defendant noted that minorities received a fair proportion of promotions, plaintiff intimated that this was achieved through impermissibly segregated work forces, with minorities being promoted in the branches and whites concentrated in the main library. Also, defendants did not rebut plaintiff’s expert testimony that the main library’s selection criteria were wrongfully subjective and inefficacious at measuring potential job performance.
Despite plaintiff’s protests, the district court denied plaintiff’s motion for summary judgment and granted defendant’s. The judge concluded that plaintiffs tried to “gerrymander the statistics” by pointing to the entire library system to show that a large pool of minority applicants was available for promotion and then looking only to the main library to show that minorities were denied promotion. The court also rejected plaintiff’s time-in-grade statistics, in part because they covered a period longer than the period during which discrimination was alleged.
Plaintiff’s motion to reconsider the grant of summary judgment was denied and this appeal followed. Because we consider this an inapt case for summary judgment we vacate and remand.
I.
After both sides moved for summary judgment, the district court granted the City’s motion. However, “the fact that both sides have moved for summary judgment does not mean that the court must grant judgment as a matter of law for one side or the other.” Schwabenbauer v. Board of Education, 667 F.2d 305, 313 (2d Cir.1981). Summary judgment is only appropriate when there is no material fact in dispute. See Liberies v. County of Cook, 709 F.2d 1122, 1129 (7th Cir.1983). In determining whether any material fact remains in dispute, all inferences must be viewed in the light most favorable to the party against whom summary judgment would be granted. See Fitzsimmons v. Best, 528 F.2d 692 (7th Cir.1976).
Because of the relative ease with which a prima facie case of employment discrimination may be made and the factual complexities inherent in Title VII litigation, summary judgment, although certainly available, is generally not favored in Title VII cases. See Logan v. General Fireproofing, 521 F.2d 881, 883 (4th Cir.1971) (“Ordinarily, summary disposition of Title VII cases is not favored, especially on the basis of a potentially inadequate factual presentation.”); Jones v. Western Geophysical Co., 669 F.2d 280, 283 (5th Cir.1982) (summary judgment should be “used cautiously” in Title VII cases). But cf. Aguilera v. Cook County Police and Corrections Merit Board, 760 F.2d 844, 849 (7th Cir.1985) (Summary judgment appropriate when plaintiff puts in no evidence to oppose a “massively supported” motion.).
In reviewing the district court’s disposition of the case, then, we must examine the *537standards for establishing a prima facie case under Title VII in light of the facts presented here to see whether summary judgment is appropriate.
II.
Title VII forbids an employer “to discriminate against any individual with respect to his compensation, terms, conditions, or privileges of employment because of such individual’s race” or to use race “to limit, segregate, or classify his employees ... in any way which would deprive or tend to deprive any individual of employment opportunities.” 42 U.S.C. § 2000e-2(a)(l), (2). Plaintiffs may pursue Title VII claims under either a disparate treatment or a disparate impact analysis. In general, either theory may be applicable to a particular set of facts. See International Brotherhood of Teamsters v. United States, 431 U.S. 324, 335 n. 15, 97 S.Ct. 1843, 1854-55 n. 15, 52 L.Ed.2d 396 (1977); Rowe v. Cleveland Pneumatic Co., 690 F.2d 88, 92 (6th Cir.1982).
Unlike a plaintiff in a disparate treatment case, a plaintiff in a disparate impact case need not show an intent to discriminate. “To establish a prima facie case of discrimination, a plaintiff need only show that the facially neutral standards in question select applicants for hire in a significantly discriminatory pattern.” Dothard v. Rawlinson, 433 U.S. 321, 329, 97 S.Ct. 2720, 2726-28, 53 L.Ed.2d 786 (1977); Liberies v. County of Cook, 709 F.2d at 1132. In some respects, however, the standards for success under a disparate impact theory are more rigorous than those under a disparate treatment analysis. To ultimately prevail in a disparate impact case, “a plaintiff must not merely prove circumstances raising an inference of discriminatory impact [but] must prove the discriminatory impact at issue.” Lowe v. Monrovia, 775 F.2d 998, 1004 (9th Cir.1985).
Once a plaintiff has made out a prima facie case of discriminatory impact, the burden shifts to the defendant to rebut plaintiff’s statistics or to show that the allegedly discriminatory practice is mandated by business necessity, Griggs v. Duke Power Co., 401 U.S. 424, 431, 91 S.Ct. 849, 853, 28 L.Ed.2d 158 (1971), or that it has a “manifest relation to the employment in question” in order to avoid a finding of discrimination. Connecticut v. Teal, 457 U.S. 440, 446, 102 S.Ct. 2525, 2530, 73 L.Ed.2d 130 (1982). Even if the employer can demonstrate a business necessity, a plaintiff may still prevail by showing that the employer was using' the practice in question as a mere pretext for discrimination or that other selection devices without a significant discriminatory effect would also “serve the employer’s legitimate interest in efficient and trustworthy workmanship.” Albermarle Paper Co. v. Moody, 422 U.S. 405, 425, 95 S.Ct. 2362, 2375, 45 L.Ed.2d 280 (1975).
Thus in Dothard v. Rawlinson, supra, a woman applied to be a prison guard in the Alabama state prison system. Her application was rejected because she did not meet the minimum weight requirement. The prison required all applicants to be at least 5’2" in height and to weigh at least 120 lbs. Plaintiff presented evidence that these requirements combined to eliminate 40% of women applicants but only 1% of men. Work force statistics bore out this fact. Of 435 prison guards in the Alabama system, only 56, or 12%, were female. The court found the requirements to have a disparate impact in violation of Title VII.
In Connecticut v. Teal, suit was brought by four black employees of a state agency. These employees had desired to become permanent supervisors at the agency. The state required that employees pass a promotion test before becoming permanent supervisors. The rate of blacks passing the test was only 68% of the passing rate of whites. The court found the test to have a disproportionate impact that made the selection process invalid. The court rejected the state’s “bottom line” defense that the state supervisory rolls had an appropriate number of blacks even if individual blacks were harmed by the promotion test.
*538III.
Here, the plaintiff claimed that the library’s promotion policies had a disparate impact on minorities. She noted that the main library generally made its promotion selections for Librarians II and III from the eligible librarians within the Chicago system. Minorities represented over 20% of this pool yet only one minority was promoted at the main library out of 25 promotions over a two and a half year period. The district court concluded as a matter of law that Regner could not limit her work force statistics to the main library. First, the court held, “the free flow of promotions within the entire CPL system does not justify limiting the number of promotions in this case to a specific library branch or a specific division of the library.” Second, the court found that “plaintiff’s reliance on the disparate impact theory as the sole basis for her discrimination claim prevents her from arguing for promotion statistics limited to the main library” because to compare the percentage of minorities in the general library with the percentage promoted at only one branch would “gerrymander the statistics.”
We think the district court erred in refusing as a matter of law to consider work force data from one library branch. While minorities may ultimately get their fair share of promotions within the library system as a whole, plaintiff’s data suggest that the promotions awarded to minorities are clustered in the branch libraries while the main library primarily promotes whites. Even though there is no indication in the record before us that this clustering is intentional, it may well amount to de facto segregation in contravention of Title VII. See Williams v. Colorado Springs School District, 641 F.2d 835 (10th Cir.1981) (Black teachers impermissibly congregated at schools with predominantly black enrollments); Sears v. Bennett, 645 F.2d 1365, 1370-74 (10th Cir.1981). This is especially true if, as Regner contends, the main library offers a unique working environment and career advantages. We are not suggesting that the City could not rebut plaintiff’s statistics or could not offer a legitimate reason for the disparities. Rather, we think only that the basis offered by the district court for granting summary judgment was insufficient and that the defendant has not shown an absence of issues of material fact with respect to liability.
We also believe that the district court erred in not considering plaintiff’s time-in-grade statistics. Plaintiff presented evidence showing that minorities wait much longer than whites before reaching higher librarian grades. The district court barred plaintiff’s statistics because they included promotions that took place before the alleged period of discrimination. But the court apparently did not allow plaintiff to revise her statistical sample. Plaintiff suggests that, if one looks at time-in-grade statistics for the period at issue here, the results are similar to those originally presented and show significant disparities. While a district court need not allow a plaintiff to correct every error in the presentation of evidence, here, at the pretrial stage of a heavily fact-dependent case, we think the court should have been less restrictive in excluding potentially probative evidence.
While plaintiff’s work force and time-in-grade statistics may establish a prima facie case for discrimination under Title VII, there is an additional hurdle to be surmounted in finding plaintiff’s claim appropriate under a disparate impact theory. A disparate impact claim generally concerns the impact of facially neutral criteria. See, e.g., Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849; Cedillo v. International Association of Bridge & Structural Iron Workers, 603 F.2d 7, 11 (7th Cir.1979). Facial neutrality is most easily demonstrated in the context of defined, concrete requirements such as requiring a minimum height or weight or a minimum score on an objective test. Raising a disparate impact claim is more problematic when subjective factors are present in the decisionmaking process. As the Ninth Circuit has noted, “there is some question as to whether [disparate impact] analysis may be applied to *539all to subjective employment decisionmak-ing.” Moore v. Hughes Helicopters, 708 F.2d 475, 481 (9th Cir.1983).
Here, the review panels could weight questions as they saw fit. There was no “correct” answer to the questions asked in the interviews or to the essays written by applicants. The City argues that this subjectivity makes considerations of disparate impact inappropriate. Some circuits have taken this position. See Pegues v. Mississippi State Employment Service, 699 F.2d 760, 765 (5th Cir.), cert. denied, 464 U.S. 991, 104 S.Ct. 482, 78 L.Ed.2d 679 (1983) (“Because the classification and referral practices complained of effectively turn on discretionary decisions, they do not fall within the category of facially neutral procedures to which the disparate impact model is traditionally applied,”); Harris v. Ford Motor Co., 651 F.2d 609, 611 (8th Cir. 1981) (affirming trial decision denying disparate impact claim because “a subjective decisionmaking system is not the type of practice outlawed under Griggs and cannot alone form the foundation of a discriminatory impact case”).
Many courts considering the issue, however, have allowed a disparate impact claim to proceed even in reference to a subjective employment practice. In Griffin v. Carlin, 755 F.2d 1516, 1524 (11th Cir.1985), black employees complained of a disparate impact in promotions in the Jacksonville, Florida postal system. Promotions were, in part, based on a written exam that was subjectively graded. The court allowed a disparate impact claim despite the existence of certain subjective elements. It noted that Griggs “did not differentiate between subjective and objective employment barriers” and that to rule otherwise would encourage employers to use subjective evaluation criteria rather than validated objective tests. See also Bauer v. Bailar, 647 F.2d 1037, 1043 (10th Cir. 1981) (applying disparate impact claim to subjective practices and concluding, “[sjubjective hiring and promotion decisions ... may not go unexplained if there is a significant disproportionate non-selection of members of a group over an extended period.”); Segar v. Smith, 738 F.2d 1249, 1288 n. 34 (D.C.Cir. 1984) (“Though these practices arguably encompass some subjective judgments as to agents’ performance, we find that disparate impact appropriately applies to them.”); Rowe v. Cleveland Pneumatic Co., 690 F.2d 88, 93 n. 10 (6th Cir.1982) (“Many Title VII cases involving subjective evaluation procedures have been analyzed under the disparate impact doctrine.”)
We find the Griffin and Bauer analysis appropriate for claims considered at the early stages of the trial process.
In view of the requirements for a prima facie case under Title VII, we believe that the plaintiff met this initial burden.1 The City may certainly be able to rebut this case but the district court should have put the onus on it to do so. We think that summary judgment was inappropriate at this time, where, in effect, possible work force segregation is suggested and where potentially invalid promotion criteria were used. We therefore reverse and remand for further proceedings not inconsistent with this opinion.