joined by all other Members of the Court, concurring in part.
I
A
The purpose of North Carolina’s agricultural extension program, administered through the North Carolina Agricultural Extension Service (Extension Service), is to aid in the dissemination of “useful and practical information on sub*389jects relating to agriculture and home economics.” App. to Pet. for Cert, in No. 85-93, p. 7a (hereinafter Pet. App.). The Extension Service is a division of the School of Agriculture and Life Sciences at North Carolina State University (NCSU). It is headed by a Director who exercises authority over District Extension Chairmen responsible for administering all Extension Service programs within the State’s six Extension Service districts. The District Extension Chairmen, in turn, supervise the 100 County Extension Chairmen who are responsible for developing and coordinating all Extension Service activities within their respective counties. The County Extension Chairmen also report to their respective Board of County Commissioners (Board), a unit of local government, on extension programs and on matters relating to budgeting and personnel.
The Extension Service operates in four major areas: home economics, agriculture, 4-H and youth, and community resource development. In both the home economics and 4-H areas, one of the Extension Service’s methods entails the establishment of clubs to educate the club members in home economics and other useful and practical skills. The agricultural program educates and encourages farmers to adopt scientific methods and to adjust to changing economic circumstances. The community resource development program emphasizes group action through citizen groups and organizations. Each of these programs is implemented by local agents who are selected for employment jointly by the Extension Service and the county Boards. Agents are divided into three ranks: full agent, associate agent, and assistant agent. “While the three ranks of agents perform essentially the same types of tasks, when an agent is promoted his responsibilities increase and a higher level of performance is expected of him.” Id., at 17a.
The salaries of all workers are determined jointly by the Extension Service and the Boards. Id., at 33a; CA App. *390223; DX 78, CA App. 1684.1 The federal, state, and county governments all contribute to these salaries. The Boards and the Extension Service determine jointly the proportionate share of salaries to be paid by the State and by the county. Moreover, all county extension hirings and firings are decided “‘jointly between the North Carolina Agricultural Extension Service and the Board of County Commissioners.’” Pet. App. 24a (quoting Memorandum of Understanding, DX 78).
The Extension Service has overall responsibility for establishing qualifications for employment in the Service and for screening applicants before recommending qualified applicants to the county commissioners for appointment to vacant or new positions. The Extension Service also prepares and submits an annual budget request to the Board for the county’s share of funds for salaries.
Each Board reviews the budget requests from the Extension Service each year and confers with and advises the District and County Extension Chairman concerning Extension Service programs. The Board furnishes the county’s share of salaries for extension personnel. In addition, it provides office space and equipment, utilities, telephone, demonstration materials, etc.
Prior to August 1,1965, the Extension Service was divided into two branches: a white branch and a “Negro branch.” Only the “Negro branch” had a formal racial designation. The “Negro branch” was composed entirely of black personnel and served only black farmers, homemakers, and youth. The white branch employed no blacks, but did on occasion serve blacks. On August 1, 1965, in response to the Civil Rights Act of 1964, the State merged the two branches of the *391Extension Service into a single organization. However, as the District Court subsequently found, “[the] unification and integration of the Extension Service did not result immediately in the elimination of some disparities which had existed between the salaries of white personnel and black personnel . . . Id., at 31a.
B
The private petitioners include employees of the Extension Service, recipients of its services, members of Extension Homemaker Clubs, and parents of 4-H Club youths. Complaint 12. They brought this action in 1971 alleging racial discrimination in employment and in the provision of services on the part of the Extension Service in violation of the First, Fifth, and Fourteenth Amendments to the Constitution, 42 U. S. C. §§ 1981, 1983 and 2000d, and 7 U. S. C. §341 et seq. The defendants, respondents here, were William C. Friday, President of NCSU, and various officials associated with the University and its School of Agriculture. In addition, County Commissioners from Alamance, Edgecomb, and Mecklenburg Counties were also named as defendants.
On April 7, 1972, the United States intervened under § 902 of Title IX and §§601 and 602 of Title VI of the Civil Rights Act of 1964, 42 U. S. C. §§2000h-2, 2000d, and 2000d-l. The United States subsequently amended its complaint in intervention to include allegations that defendants had also violated §§703 and 706 of Title VII of the Civil Rights Act of 1964, as amended, 42 U. S. C. §§2000e-2 and 2000e-5. The United States’ complaint essentially tracked the claims made by the private petitioners. The private petitioners were permitted on the eve of trial to amend their complaint to add a claim under Title VII as well.
On two occasions prior to trial the District Court was asked, but declined, to certify the action as a class action. *392Near the close of trial the plaintiffs again requested the court to certify four classes of plaintiffs and one class of defendants.2 However, the District Court once again declined to do so, and this decision was subsequently upheld by the Court of Appeals. On the merits, the trial court explored allegations of racial discrimination in virtually every aspect of the Extension Service’s employment practices and provision of services.3 The District Court ruled in favor of respondents in all respects. On most issues it concluded that peti*393tioners had failed to carry their burden of proof. As a general proposition, the District Court was of the view that the Extension Service had conducted itself in a nondiscriminatory manner since it became subject to Title VII and since the merger of the black and white branches in 1965. Both the private petitioners and the United States limited their appeals to the claims that the District Court erred in considering the evidence before it regarding salaries and promotions to County Chairmen, and in concluding that the Extension Service had not discriminated against blacks with respect to salaries and promotions to County Chairmen. The United States also claimed that the system used to determine merit pay increases violated Title VII. The private petitioners also appealed the rejection of their claim that respondents were unlawfully providing services and materials to segregated 4-H and Extension Homemaker Clubs, and the District Court’s refusal to certify the case as a class action. The Court of Appeals affirmed the District Court in all respects. 751 F. 2d 662 (CA4 1984). We granted certiorari, 474 U. S. 978 (1985).4
*394II
The first issue we must decide is whether the Court of Appeals erred in upholding the District Court’s finding that petitioners had not proved by a preponderance of the evidence that respondents had discriminated against black Extension Service employees in violation of Title VII by paying them less than whites employed in the same positions. The Court of Appeals reasoned that the Extension Service was under no obligation to eliminate any salary disparity between blacks and whites that had its origin prior to 1972 when Title VII became applicable to public employers such as the Extension Service.5 It also reasoned that factors, other than those included in petitioners’ multiple regression analyses, affected salary, and that therefore those regression analyses were incapable of sustaining a finding in favor of petitioners.
A
Both the Court of Appeals and the District Court found that before the black and white Extension Service branches were merged in 1965, the Extension Service maintained two separate, racially segregated branches and paid black employees less than white employees. Pet. App. 120a; 751 F. 2d, at 666. The Court of Appeals also acknowledged that after the merger of the Extension Service, “[s]ome preexisting salary disparities continued to linger on,” and that these disparities continued after Title VII became applicable to the Extension Service in March 1972 and after this suit was filed. Ibid. Indeed, the Court of Appeals noted that “the Extension Service admits that, while it had made some adjustments to try to get rid of the salary disparity resulting *395on account of pre-Act discrimination, it has not made all the adjustments necessary to get rid of all such disparity.” Id., at 672. See also Brief for Respondents 32 (“[EJfforts were made to reduce the average differences but due to the county by county salary differences and finding [sic] structure 1971 [sic], the averages were not eliminated”). The court interpreted petitioners’ claim on appeal to be that “the pre-Act discriminatory difference in salaries should have been affirmatively eliminated but has not.” 751 F. 2d, at 670. Relying on our cases in Hazelwood School District v. United States, 433 U. S. 299 (1977), and United Air Lines, Inc. v. Evans, 431 U. S. 553 (1977), it concluded, “[w]e do not think this is the law.” 751 F. 2d, at 670.
The error of the Court of Appeals with respect to salary disparities created prior to 1972 and perpetuated thereafter is too obvious to warrant extended discussion: that the Extension Service discriminated with respect to salaries prior to the time it was covered by Title VII does not excuse perpetuating that discrimination after the Extension Service became covered by Title VII. To hold otherwise would have the effect of exempting from liability those employers who were historically the greatest offenders of the rights of blacks. A pattern or practice that would have constituted a violation of Title VII, but for the fact that the statute had not yet become effective, became a violation upon Title VII’s effective date, and to the extent an employer continued to engage in that act or practice, it is liable under that statute. While recovery may not be permitted for pre-1972 acts of discrimination, to the extent that this discrimination was perpetuated after 1972, liability may be imposed.
Each week’s paycheck that delivers less to a black than to a similarly situated white is a wrong actionable under Title VII, regardless of the fact that this pattern was begun prior *396to the effective date of Title VII. The Court of Appeals plainly erred in holding that the pre-Act discriminatory difference in salaries did not have to be eliminated.6
*397The Court of Appeals’ conclusion that pre-Act salary discrimination did not have to be eliminated undermines the rest of its analysis of the District Court opinion. Having rejected the effect of pre-Act discrimination, the court considered solely whether the Extension Service discriminated with respect to the application of quartile rankings which, according to the Court of Appeals, were “the only aspect of salary computation in which the Extension Service exercised any discretion.” 751 F. 2d, at 674.7 Because, as we have explained, the Extension Service was under an obligation to eradicate salary disparities based on race that began prior to the effective date of Title VII,8 the Court of Appeals erred in concentrating its analysis solely on the issue whether there was racial discrimination in the ranking system.
B
We now turn to the issue whether the Court of Appeals erred in upholding the District Court’s refusal to accept the petitioners’ expert statistical evidence as proof of discrimina*398tion by a preponderance of the evidence. In a case alleging that a defendant has engaged in a pattern and practice of discrimination under § 707(a) of the Civil Rights Act of 1964, 42 U. S. C. § 2000e-6(a), plaintiffs must “establish by a preponderance of the evidence that racial discrimination was the company’s standard operating procedure — the regular rather than the unusual practice.” Teamsters v. United States, 431 U. S. 324, 336 (1977). Further, our decision in United States Postal Service Board of Governors v. Aikens, 460 U. S. 711 (1983), although not decided in the context of a pattern-and-practice case, makes clear that if the defendants have not succeeded in having a case dismissed on the ground that plaintiffs have failed to establish a prima facie case, and have responded to the plaintiffs’ proof by offering evidence of their own, the factfinder then must decide whether the plaintiffs have demonstrated a pattern or practice of discrimination by a preponderance of the evidence. This is because the only issue to be decided at that point is whether the plaintiffs have actually proved discrimination. Id., at 715. This determination is subject to the clearly-erroneous standard on appellate review. See Anderson v. Bessemer City, 470 U. S. 564 (1985); Pullman-Standard v. Swint, 456 U. S. 273 (1982).
At trial, petitioners relied heavily on multiple regression analyses designed to demonstrate that blacks were paid less than similarly situated whites. The United States’ expert prepared multiple regression analyses relating to salaries for the years 1974, 1975, and 1981. Certain of these regressions used four independent variables — race, education, tenure, and job title. Petitioners selected these variables based on discovery testimony by an Extension Service official that four factors were determinative of salary: education, tenure, job title, and job performance. GX 159, pp. 90, 96. In addition, regressions done by the Extension Service itself for 1971 included the variables race, sex, education, and experience; and another in 1974 used the variables race, education, *399and tenure to check for disparities between the salaries of blacks and whites. GX 214; Tr. 3915-3918; CA App. 1681; Tr. 3920.
The regressions purported to demonstrate that in 1974 the average black employee earned $331 less per year than a white employee with the same job title, education, and tenure, GX 123; CA App. 1601; Tr. 364-365, and that in 1975 the disparity was $395, GX 123; CA App. 1589; Tr. 377.9 The regression for 1981 showed a smaller disparity which lacked statistical significance.
The Court of Appeals stated:
“[The] district court refused to accept plaintiffs’ expert testimony as proof of discrimination by a preponderance of the evidence because the plaintiffs’ expert had not included a number of variable factors the court considered relevant, among them being the across the board and percentage pay increases which varied from county to county. The district court was, of course, correct in this analysis.” 751 F. 2d, at 672.
The Court of Appeals thought the District Court correct for essentially two reasons: First, the Court of Appeals rejected petitioners’ regression analysis because it “contained salary figures which reflect the effect of pre-Act discrimination, a consideration not actionable under Title VII . . . .” Ibid, (footnote omitted). Second, the court believed that “[a]n appropriate regression analysis of salary should ... include all measurable variables thought to have an effect on salary level.” Ibid. In particular, the court found that the failure to consider county-to-county differences in salary increases was significant. It concluded, noting: “[B]oth experts omitted from their respective analysis variables which ought to be reasonably viewed as determinants of salary. As a result, the regression analysis presented here must be *400considered unacceptable as evidence of discrimination.” Ibid. The Court of Appeals’ treatment of the statistical evidence in this case was erroneous in important respects.
1
The Court of Appeals erred in stating that petitioners’ regression analyses were “unacceptable as evidence of discrimination,” because they did not include “all measurable variables thought to have an effect on salary level.” The court’s view of the evidentiary value of the regression analyses was plainly incorrect. While the omission of variables from a regression analysis may render the analysis less probative than it otherwise might be, it can hardly be said, absent some other infirmity, that an analysis which accounts for the major factors “must be considered unacceptable as evidence of discrimination.” Ibid. Normally, failure to include variables will affect the analysis’ probativeness, not its admissibility.10
Importantly, it is clear that a regression analysis that includes less than “all measurable variables” may serve to prove a plaintiff’s case. A plaintiff in a Title VII suit need not prove discrimination with scientific certainty; rather, his or her burden is to prove discrimination by a preponderance of the evidence. Texas Dept. of Community Affairs v. Burdine, 450 U. S. 248, 252 (1981). Whether, in fact, such a regression analysis does carry the plaintiffs’ ultimate burden will depend in a given case on the factual context of each case in light of all the evidence presented by both the plaintiff and the defendant. However, as long as the court may fairly conclude, in light of all the evidence, that it is more likely *401than not that impermissible discrimination exists, the plaintiff is entitled to prevail.
2
In this case the Court of Appeals failed utterly to examine the regression analyses in light of all the evidence in the record. Looked at in its entirety, petitioners offered an impressive array of evidence to support their contention that the Extension Service engaged in a pattern or practice of discrimination with respect to salaries. In addition to their own regression analyses described above, petitioners offered regressions done by the Extension Service for 1971 and 1974 that showed results similar to those revealed by petitioners’ regressions. Tr. 3917; CA App. 1681. Petitioners also claim support from multiple regressions presented by respondents at trial for the year 1975. Using the same model that petitioners had used, and similar variables, respondents’ expert obtained substantially the same result for 1975, a statistically significant racial effect of $384. CA App. 1716. Indeed, respondents also included in their analysis, “quartile rank” as an independent variable, and this increased the racial effect to $475.11
Petitioners also presented evidence of pre-Act salary discrimination, and of respondents’ ineffectual attempts to eradicate it. For example, petitioners submitted evidence, and the District Court found, that blacks were paid less than whites in comparable positions prior to the merger of the black and white services in 1965. Pet. App. 120a. Moreover, in 1971, respondents acknowledged that substantial sal*402ary differences between blacks and whites existed.12 In addition, evidence was offered to show that the efforts by the Extension Service to equalize those salaries in 1971 were insufficient to accomplish the goal. Tr. 242-246; GX 98. As we made clear in Hazelwood School District v. United States, 433 U. S., at 309-310, n. 15 “[p]roof that an employer engaged in racial discrimination prior to the effective date of Title VII might in some circumstances support the inference that such discrimination continued, particularly where relevant aspects of the decisionmaking process had undergone little change.”13
Further, petitioners presented evidence to rebut respondents’ contention that county-to-county variations in contributions to salary explain the established disparity between black and white salaries. The United States presented evidence, which it claims respondents did not rebut, establishing that black employees were not located disproportionately in the counties that contributed only a small amount to Extension Service salaries. GX 216; see also CA App. 189. Absent a disproportionate concentration of blacks in such counties, it is difficult, if not impossible, to understand how the fact that some counties contribute less to salaries than others could explain disparities between black and white salaries. *403In addition, the United States presented an exhibit based on 1973 data for 23 counties showing 29 black employees who were earning less than whites in the same county who had comparable or lower positions and tenure. GX 102.
Finally, and there was some overlap here with evidence used to discredit the county-to-county variation theory, petitioners presented evidence consisting of individual comparisons between salaries of blacks and whites similarly situated. GX 102, DX 48. Witness testimony, claimed by petitioners to be unrebutted, also confirmed the continued existence of such disparities. CA App. 190; Tr. 2010-2012, 2685, 2825-2826.
Setting out the range of persuasive evidence offered by petitioners demonstrates the error of the Court of Appeals in focusing solely on the characteristics of the regression analysis. Although we think that consideration of the evidence makes a strong case for finding the District Court’s conclusion clearly erroneous,14 we leave that task to the Court of *404Appeals on remand which must make such a determination based on the “entire evidence” in the record. United States v. United States Gypsum Co., 333 U. S. 364 (1948).15
I — I I — I HH
The private petitioners complain that the District Court and Court of Appeals erred in failing to certify this case as a class action. They seek the certification of three distinct classes: (1) all black employees of the Extension Service on or after November 18, 1971; (2) all current black members and potential black members of the 4-H and Extension Home*405maker Clubs on or after November 18, 1971;16 and as a defendant (3) all County Commissioners in North Carolina who held that position on or after November 18, 1971. 751 F. 2d, at 667. The Court of Appeals upheld the District Court’s denial of class certification.
A
With respect to the class of black employees, the Court of Appeals held that due to the fact that salaries are made up of money from several distinct sources, the Federal Government, the State, and the counties, the “claim of a potential plaintiff against one county will not be typical of the claim of another potential plaintiff against a different county.” Id., at 668.17 It applied the same reasoning to the employees’ charge of discrimination in the hiring of County Chairmen. Ibid. Yet the claims here were not asserted solely against the counties; they were asserted also against the Extension Service. And, as against the Extension Service, at least, it is clear that the claims of the named plaintiffs were “typical” *406of other black employees who may have been paid less or denied promotion to chairman. Although it seems likely that the other requirements of Federal Rule of Civil Procedure 23 were met by this class, neither court below expressly considered the issue and we therefore leave that determination to the Court of Appeals on remand.18
B
The Court of Appeals also upheld the District Court’s decision not to certify a class of County Commissioner defendants because there “was simply no evidence of any standardized practice among the one hundred separate counties in the state to deprive anyone of any rights solely because of race.” Pet. App. 47a-48a. The Court of Appeals was of the view that “to have a proper class of defendants in a case such as this there must be either a statewide rule or practice so that relief is available if the rule or practice is invalid, or the adjudication with respect to a member of a defendant class must as a practical matter be dispositive of the interests of the other members of the class as provided in FRCP 23(b)(1)(B).” 751 F. 2d, at 670. We agree with the Court of Appeals that certification of a defendant class under Rule 23(b)(1)(B) in this case would have been improper. Whether an individual county acted intentionally with the Extension Service in setting salaries or in selecting County Chairmen in a discriminatory manner, is an issue that once decided with respect to a *407particular county could not “be dispositive of the interests of the other members of the class.” The private petitioners have suggested no theory to support any different result.
In this opinion, and in my opinion dissenting in part, post, p. 309, the following designations are used to refer to the record. GX, exhibit of Federal Government; DX, defendant’s exhibit; Tr., trial transcript; CA App., Appendix in the Court of Appeals.
The classes considered by the District Court in its August 20, 1982, memorandum were:
“(1) All Black and Indian employees and potential employees of the [Extension Service] since November 18, 1971, and thereafter;
“(2) All Black and Indian persons who were recipients or potential recipients of service from the [Extension Service] on November 18, 1971, and thereafter;
“(3) All Black and Indian members or potential members of the [Extension Service’s] 4-H Clubs on November 18, 1971, and thereafter;
“(4) All Black and Indian persons who were members or potential members of the [Extension Service’s] Homemaker Clubs on November 18, 1971, and thereafter, and
“(5) [A defendant class consisting of a]ll County Commissioners in North Carolina, in their official capacities, on November 18, 1971, and thereafter.” Pet. App. 37a.
The claims relating to American Indians were dismissed by the District Court because petitioners at trial presented no evidence to support these claims. Id., at 49a, n. 11. No issue with respect to these claims is before us.
Petitioners sought to prove that respondents had continued to assign black employees only to counties that had black employees prior to 1965; had failed to recruit, hire, and assign blacks on an equal basis with whites; had denied blacks the same compensation, terms, conditions, and privileges as were provided to whites; had segregated blacks in work assignments; had failed to establish selection standards sufficiently objective to prevent discrimination in hiring and promotion; had failed to correct the present effects of past discrimination; had failed to provide minorities with services equal to the services provided white persons; and had failed to maintain nonracially segregated 4-H Clubs and Extension Homemaker Clubs. Id., at 50a-51a.
The question presented in the Federal Government’s petition is whether black state employees establish a claim under § 703(a) of the 1964 Civil Rights Act, 42 U. S. C. 2000e-2(a), by identifying current salary disparities between themselves and white employees holding the same jobs, when such disparities result from a state policy before 1965 of paying blacks lower salaries than whites.
The private petitioners presented the same question as that presented by the Federal Government, and four additional questions:
(1) May a regression analysis be treated as probative evidence of discrimination where the analysis does not incorporate every conceivable relevant variable?
(2) May North Carolina satisfy its obligation to desegregate the de jure system of 4-H Clubs and Extension Homemaker Clubs by adopting a freedom of choice plan that fails?
(3) May an employer immunize itself from liability for illegal discrimination by delegating its hiring decisions to a third party?
(4) Did the Fourth Circuit err in denying class certification in this case?
As originally enacted, Title VII of the Civil Rights Act of 1964 applied only to private employers. The Act was expanded to include public employees by the Equal Employment Opportunity Act of 1972, 86 Stat. 103, whose effective date was March 24,1972. See 42 U. S. C. §§ 2000e(a), (b), (f), (h).
Neither Hazelwood nor Evans suggests any different rule, and indeed those cases support the result here. In Evans, respondent, a female flight attendant, was forced to resign in 1968 from her position due to her employer’s policy forbidding female flight attendants to marry. Respondent there never brought an action with respect to this forced resignation. In 1972 she was rehired by the airline as a new hire, and given seniority only from that date. Although her claim with respect to the 1968 resignation was time barred, respondent filed suit claiming that the airline was guilty of a present, continuing violation of Title VII because the seniority system treated her less favorably than males who were hired after her termination in 1968 and prior to her reemployment. Further, she claimed that the seniority system gave present effect to the past, illegal forced retirement and thereby perpetuated the consequences of forbidden discrimination. Respondent had made no allegation that the seniority system itself was intentionally designed to discriminate. Because a lawsuit on the forced resignation was time barred, however, it was to be treated as an act occurring before the statute was passed, and therefore it had “no present legal consequences,” 431 U. S., at 558, even though “[i]t may constitute relevant background evidence in a proceeding in which the status of a current practice is at issue.” Ibid. The “critical question,” the Court declared, “is whether any present violation exists.” Ibid, (emphasis added). Because the employer was not engaged in discriminatory practices at the time the respondent in Evans brought suit, there simply was no violation of Title VII.
In Hazelwood, the Attorney General brought suit against the Hazelwood School District and various of its officials claiming that they were engaged in a pattern or practice of discriminatory hiring in violation of Title VII. We vacated the decision of the Court of Appeals that directed judgment for the Government, because that decision did not take into account the possibility that the prima facie statistical proof in the record “might at the trial court level be rebutted by statistics dealing with Hazelwood’s hiring after it became subject to Title VII.” 433 U. S., at 309. We explained that “[a] public employer who from [1972] forward made all its employment decisions in a wholly nondiscriminatory way would not violate Title VII even if it had formerly maintained an all-white work force by purposefully excluding Negroes.” Ibid.
Here, however, petitioners are alleging that in continuing to pay blacks less than similarly situated whites, respondents have not from the date of *397the Act forward “made all [their] employment decisions in a wholly nondiscriminatory way.” Ibid. Our holding in no sense gives legal effect to the pre-1972 actions, but, consistent with Evans and Hazelwood, focuses on the present salary structure, which is illegal if it is a mere continuation of the pre-1965 discriminatory pay structure.
Quartile ranking refers to the practice of the Extension Service of placing each agent in the first, second, third, or fourth quartile, according to his or her performance for the previous period. These rankings influence salary decisions.
This lawsuit involves two distinct types of salary claims: those of employees subject to the premerger discriminatory pay structure and those hired after the merger of the black and white branches. If the acknowledged pre-1965 disparities continued for employees employed prior to 1965, then respondents violated the law. But for employees covered by this suit who were never employed under the dual system, it is meaningless to say that the pre-1965 disparity “continued” past 1972, absent (1) evidence that new disparities were created or begun after the merger that continued past 1972, or (2) evidence that new disparities were created after 1972. See Brief for Plaintiffs-Appellants Bazemore et al. in Nos. 82-1873(L), 82-1881, 82-1927, 82-2065 (CA4), pp. 24-41.
Petitioners’ expert testified that both of these disparities were statistically significant. Tr. 364-365, 377.
There may, of course, be some regressions so incomplete as to be inadmissible as irrelevant; but such was clearly not the case here.
With respect to the increased disparity when quartile rank was added to the regression analysis, respondents’ expert stated that 20% of the data was missing when quartile rankings were added, and he was unable to explain the effect of the increase in the disparity when those rankings were added. Tr. 6242.
Dr. T. Carlton Blalock, then Assistant Director for Administration in the Extension Service, pointed out that black county professionals were earning an average of $800 to $1,100 per year less than whites in 1970. Tr. 3905, 3911. The Blalock memorandum stated:
“Believe you’d agree our salaries for women & non-white men on average are lower — Our figures verify — Due to several factors-
-The competitive market — This is not acceptable as a reason though.
-Tradition — not just in Ext.
-Less county support for non-white positions.” GX 157, App. 129.
On appeal petitioners specifically complained that the District Court had not given any weight to the pre-Act discrimination in its analysis. However, to the extent that proof is required to establish discrimination with respect to salary disparities created after 1972, see supra, at 394-397, evidence of pre-Act discrimination is quite probative.
There was very little evidence to show that there was in fact no disparity in salaries between blacks and whites, or to demonstrate that any disparities that existed were the product of chance. The District Court did point to eases of individual differences that it found to be successfully rebutted by respondents. In addition, the District Court alluded to evidence presented by defendants relating to salaries for 1976 and thereafter, Pet. App. 146a-147a, CA App. 2227-2231, but, putting aside whether that evidence actually contradicted petitioners’, it is simply not very probative of whether there existed a pattern or practice of discrimination prior to 1976. The District Court also pointed to “scattergrams” or graphs based on the data in respondents’ regressions, concluding that these graphs displayed the salaries of blacks and whites “in a completely random distribution.” Pet. App. 148a. Yet, as pointed out by the United States in its brief below, the very purpose of a regression analysis is to organize and explain data that may appear to be random. See Fisher, Multiple Regression in Legal Proceedings, 80 Colum. L. Rev. 702, 705-707 (1980). Thus, it is simply wrong to give weight to a scattergram while ignoring the underlying regression analysis. Respondents’ strategy at trial was to declare simply that many factors go into making up an individual employee’s salary; they made no attempt that we are aware of — statistical or otherwise — to demonstrate that when these factors were properly organized and *404accounted for there was no significant disparity between the salaries of blacks and whites.
We do note, however, that certain conclusions of the District Court are inexplicable in light of the record. First, the District Court, in referring to petitioners’ expert’s analyses stated that the regressions on which petitioners principally relied did not include job title. Pet. App. 119a. Yet the District Court expressly noted that in other regressions in which petitioners did include job title, a statistically significant disparity was noted. Second, the District Court stated that “the single most important factor in determining salaries for the Extension Service professional staff is job performance.” Id., at 134a. Yet the District Court failed even to note that respondents’ regression analysis for 1975 which included a performance variable showed an even greater disparity in salary than did petitioners’.
Third, the District Court complained about the inclusion of the County Chairmen in petitioners’ regression analysis, fearing that the fact that they were disproportionately white would skew the salary statistics to show whites earning more than blacks. Yet, because the regressions controlled for job title, adding County Chairmen as a variable in the regression would simply mean that the salaries of white County Chairmen would be compared with those of nonwhite County Chairmen. In any event, respondents’ own regression at trial excluded County Chairmen and revealed a differential between black and white salaries.
Finally, the District Court listed nine variables that it believed were not accounted for in petitioners’ regressions. See id., at 133a. It did not, however, determine whether these variables were included in the evidence in other respects. For example, several of the “missing” variables relate to eounty-to-county variations, while others relate to performance, a variable expressly included in respondents’ own regression.
Given the Court’s disposition on the merits of the claims relating to the 4-H Clubs and the Extension Homemaker Clubs, we agree that the issue whether the District Court erred in refusing to certify a class of club members is now moot.
The Court of Appeals analogized the present case to its decision in Stastny v. Southern Bell Tel. & Tel. Co., 628 F. 2d 267 (CA4 1980), “in which we held that promotion and pay decisions subject to almost complete local autonomy in the various offices of Southern Bell throughout North Carolina would not support the typicality requirement under FRCP 23(a)(3) for a statewide class of employees.” 751 F. 2d, at 668. The findings of the District Court flatly contradict the Court of Appeals’ conclusion that salaries are reached in any “autonomous” fashion or are arrived at by the counties; rather the District Court explained that “[t]he salaries are determined jointly by the Service and the county board of commissioners.” App. 77. This finding is supported by the Memorandum of Understanding between the Extension Service and the Boards of Commissioners for the relevant time period. See id., at 162. Of course, that this case may be one in which it is proper to certify a class, is distinct from the question whether county variations serve as a basis for the demonstrated disparities between the salaries of black and white employees.
The District Court believed that a “most importan[t]” reason for not certifying this case as a class action was that “it is now settled law that class action certification is inappropriate and unnecessary in pattern and practice suits brought by the EEOC and the government pursuant to Title VII.” Pet. App. 44a-46a. It cited our opinion in General Telephone Co. v. EEOC, 446 U. S. 318 (1980), for this proposition. The District Court misread that opinion. In General Telephone Co., we held that in a pattern-and-practice ease the Government need not be certified as a representative of the class of alleged victims. That case does not stand for the erroneous proposition that once the Government intervenes in a case brought by private plaintiffs, those plaintiffs lose their right to proceed as a class.