Cornelius COOPER, Michael Edwards, et al., Plaintiffs-Appellants, v. SOUTHERN COMPANY, Georgia Power Company, et al., Defendants-Appellees, 390 F.3d 695

MARCUS, Circuit Judge: Cornelius Cooper and six other plaintiffs appeal from the district court’s orders denying class certification in their employment discrimination suit and entering final

summary judgment in favor of the defendants, the Southern Company, Georgia Power Company, Southern Company Services, Inc., and Southern Company Energy Solutions, Inc. Seven current or former employees of the various defendants brought this putative class action, alleging violations of Title VII of the Civil Rights Act of 1964 (“Title VII”), 42 U.S.C. §§ 2000e, et seq., and 42 U.S.C. § 1981 (“Section 1981”). The plaintiffs claimed discrimination in promotions and compensation, and sought declaratory and injunc-tive relief, back pay, and compensatory and punitive damages. After denying class certification, the district court entered seven separate orders of

summary judgment in favor of the defendants. After painstaking review of the record, as well as careful consideration of the briefs and oral argument, we conclude that the district court did not abuse its discretion in denying class certification, because the plaintiffs demonstrated neither that they had satisfied the commonality and typicality requirements of Federal Rule of Civil Procedure 23(a), nor that damages were incidental to equitable and declaratory relief or that common questions of law or fact predominated, as required under Federal Rule of Civil Procedure 23(b)(2) or (b)(3). In addition, we are persuaded by none of the plaintiffs’ arguments regarding the individual

summary judgment orders and, accordingly, affirm in all respects the judgments of the district court. I. The complex facts and procedural history underlying this appeal are these. Seven African-American past or present employees of the defendant companies filed suit on July 27, 2000, alleging that the Southern Company and several of its subsidiaries — Georgia Power Company, Southern Company Services, Inc. and Southern Company Energy Solutions, Inc. — unlawfully discriminated against their employees on account of race.

The plaintiffs alleged that the defendants discriminated against them in connection with promotion opportunities and performance evaluations, as well as in terms of compensation, and claimed that the defendants tolerated a racially hostile working environment.

Notably, the plaintiffs sought to represent a Rule 23 class that they described in the following terms: All African-American persons employed by Southern Company’s Corporate Office, Georgia Power Company, Southern Company Services, Inc. or Southern Company Energy Solutions, Inc., in the United States at any time from July, 1998 to the present, who are subject to the Defendants’ employment, personnel and human resources policies and practices and who have been, continue to be, or may in the future be adversely affected by the Defendants’ racially diserimi-natory employment policies and practices (“the Class”).

Cooper v. Southern Co., 205 F.R.D. 596, 598-99 (N.D.Ga.2001). At the time suit was filed, the proposed class included approximately 2,400 individuals residing in Georgia, Alabama, Florida, and Mississippi, and working at locations dispersed throughout the four states. The proposed class consisted of entry-level manual laborers and skilled professionals, among others, and encompassed exempt, non-exempt, unionized and non-unionized workers.

In essence, the plaintiffs maintain that the various defendant companies share a common system for personnel decision-making, which constitutes a “common policy or practice” that is appropriately challenged on a class-wide basis. Specifically, plaintiffs say that there are “common promotion and compensation policies and practices, which give managers unfettered discretion to make subjective, non-job-related decisions.” Appellants’ Brief at 2.

According to plaintiffs, these “common policies and practices ... foster a pattern or practice of race discrimination or have a disparate impact on black employees,” id., and plaintiffs sought to redress the claimed wrongdoing in the form of a class action. A. The defendants include four companies that provide a wide range of energy-related products and services throughout the southeastern United States. The Southern Company (“TSC”) is a holding company that itself has no employees, but which owns stock in the other defendant companies, all Southern Company subsidiaries.

TSC is a Delaware corporation that is domesticated under the laws of Georgia and maintains its corporate headquarters in Atlanta, Georgia. In addition, TSC is the corporate parent of various other utility companies not directly involved in the lawsuit. Georgia Power Company (“GPC”) is the Southern Company’s largest subsidiary, and provides electricity to approximately 1.9 million customers in 153 of Georgia’s 159 counties.

GPC employs approximately 9,000 employees in the state of Georgia, some 44% of whom are covered by a collective bargaining agreement (“CBA”) with the International Brotherhood of Electrical Workers (“IBEW”), Local No. 84. GPC’s operations are extremely diversified, and its facilities, workforce, and management are dispersed across the entire state of Georgia. Certain functions within GPC are divided among different regions in the state, and widely divergent operating and management structures exist within different regions.

Thus, for example, GPC’s Customer Operations function across 16 regions that have very different workforces. The expansive and rural Southeast Region has very different characteristics with respect to technology, customer service, and distribution capacity than does the metropolitan Atlanta region, a dense and highly-populated area that includes numerous commercial clients. Southern Company Energy Solutions (“SCES”), in turn, is a non-regulated, non-utility subsidiary of the Southern Company that develops for sale various energy-related products and services.

SCES consults with commercial and industrial clients on energy efficiency, provides energy efficiency and environmental programs and services, and offers other services to commercial and residential customers. Most of the fewer than 300 employees employed by SCES are exempt employees whose compensation is determined by sales commissions; none of SCES’s employees is a member of a union. Finally, Southern Company Services (“SCS”) provides the other defendants with human resources services and administers common compensation and promotion policies.

SCS also provides various public relations and employee relations services to the other defendant companies. SCS employs approximately 3,500 non-unionized employees who work primarily in Georgia and Alabama, although they also coordinate services in Mississippi, Florida, New York, and Washington, D.C. The differences in management structures, working environments, and criteria for employment decisions vary substantially among the defendant companies.

The promotion and compensation decisions affecting laborers involved in electrical power transmission and distribution, for example, take into account very different criteria than do decisions involving the professional and managerial ranks of the companies. Similarly, the more than 200 locations in which employees of the different defendants work are spread throughout a widely dispersed geographical area and encompass an extremely diverse range of working environments. While the operations and workforces of the defendant companies are substantially different, they are linked in several ways.

All the defendants maintain a common job, salary, and pay-grade system that dictates salary ranges for both exempt and nonexempt employees. Non-exempt employees are classified from grade NE1 through NE9; exempt employees are classified at grades El through E15. Salary ranges for each grade vary significantly. While the same pay-grade system is used within all of TSC’s subsidiaries, the workforces within the different companies are strikingly different.

Not every position exists in every company, and in some subsidiaries the job responsibilities for a given position differ greatly from the responsibilities for the same position in another subsidiary. Job positions may also correspond to different pay-grade ranges within different subsidiaries. The terms of compensation for unionized workers are strictly governed by the terms of the CBA, giving managers precious little discretion over salary and similar issues.

For non-unionized workers, hiring and promotion decisions are made by managers of various rank based in geographically dispersed facilities within each of the defendant companies. Different managers assign various weights to different qualifications in making their employment decisions, and managers have the discretion to utilize a wide range of processes and procedures when making compensation and promotion decisions. Several different types of promotions are available to non-unionized employees.

Most competitive job openings — jobs open to any qualified employee — are posted on a computer intranet system known as “Job-Net,” which is accessible to employees within each TSC subsidiary. SCS recommends standardized processes for filling JobNet positions, but managers within the various subsidiaries retain discretion in choosing whether to post job vacancies. Other promotions occur outside the context of the competitive promotions available on JobNet.

For example, employees can advance within their own “job family.” These “progressive promotions” occur when employees meet pre-determined performance goals or demonstrate satisfactory work over a given period. Under these circumstances, an employee may be promoted, for example, from an Operator II position to an Operator I position. These promotions are not competitive since they do not involve unfilled positions, but rather are available only to employees progressing individually within their own job family.

SCS also provides numerous programs and procedures designed to assist managers within the other defendant companies in evaluating employees and making compensation and promotion decisions. These procedures are designed to ensure that employees are evaluated, compensated, and promoted based on job-related factors, in accordance with the companies’ strict prohibitions on discrimination on account of race. Thus, for instance, SCS provides the defendant companies with resources to assist managers in promotion decisions, including structured interview guidelines designed to evaluate job-related skills.

SCS also recommends that employees receive regular written and oral evaluations, and develops specific evaluation forms for this purpose. SCS and the other defendants have all developed affirmative action plans to address workforce diversity, and they also provide professional development and professional mentoring programs to facilitate employees’ professional growth. SCS’s Equal Employment Opportunity (“EEO”) Department has primary responsibility for ensuring compliance with anti-discrimination and equal opportunity laws.

The EEO Department’s policy is to investigate discrimination complaints, and if any discrimination or harassment is revealed, to take remedial action to eliminate the discrimination or harassment. SCS’s EEO Department also acts in conjunction with Compliance Officers within each defendant company to conduct annual compliance surveys. Every individual employee within the defendant companies receives an annual compliance questionnaire, which asks whether the employee is aware of any incidents of discrimination or harassment.

The EEO Department also conducts quarterly Equal Employment Opportunity reviews designed to ensure that all the defendant companies comply not only with the applicable anti-discrimination laws, but also with their own internal policies. In these reviews, the EEO Department evaluates the hiring, promotion, and compensation practices of randomly selected departments within the various defendant companies, with the goal of remedying any inconsistencies that may exist between a department’s practice and the defendant’s EEO policies.

The defendant companies have also implemented various training programs to promote compliance with all EEO policies and anti-discrimination laws. Many of the defendants’ managers are required to attend “FEDLAW,” a one-day seminar that provides an overview of the federal anti-discrimination laws applicable to the defendants’ workplaces. The defendants also encourage their employees and managers to attend another one-day training session, the “Civil Treatment” seminar. In the “Civil Treatment” training, employees are reminded of the defendants’ anti-discrimination policies and encouraged to report any violations.

Finally, all employees of SCS and SCES are entitled to raise EEO concerns with SCS’s Employee Concerns department, and unionized employees at GPC may raise discrimination complaints with their union stewards, pursuant to successive agreements between GPC and Local 84 of the IBEW. GPC employees may also bring complaints to the Corporate Concerns Department or the Corporate Security Department. B.

The seven named plaintiffs are African Americans who work (or formerly worked) in a wide assortment of union and nonunion positions within GPC, SCS and SCES, and who have been subjected to compensation and promotion decisions made by different managers utilizing different procedures. We analyze in detail the individual claims of each plaintiff in

Section III, infra, which addresses each of the

summary judgment orders, but offer here a brief

summary of each plaintiffs work history in order to measure his or her claims against the standards embodied in Rule 23 of the Federal Rules of Civil Procedure. Cornelius Cooper (“Cooper”) works as a Lineman for GPC in Atlanta, Georgia, where he has been employed since 1973. At the time the suit was filed, Cooper had been employed by GPC for some 28 years. He has been a union member during his employment at GPC, and thus, his compensation and promotions within the bargaining unit have been governed by the terms of the CBA.

Cooper received his current Lineman position in 1981 and has not been promoted since, in spite of positive work evaluations; he claims that less qualified white employees have been promoted while he could not advance for nearly 20 years. Cooper has not taken the First Line Supervisor test, nor has he applied for a Foreman position since at least 1990, because no Foreman positions have been open in that time. He has not applied to be a crew leader. However, Cooper maintains that he suffered discrimination when he was denied two promotions in 1998.

The first position that Cooper was denied was as a Lighting Coordinator, an opening that was posted on JobNet in July 1998. Later that year, Cooper applied for and was denied a non-union Methods & Training Specialist (“Trainer”) position at the Klondike Training Center. Of several Trainer jobs open at the time, at least one was awarded to an African-American employee, Charlie Johnson. Michael Edwards (“Edwards”) had worked at GPC for 13 years when this suit was commenced. Like Cooper, Edwards has held a variety of union positions in which his compensation and promotions were subject to the terms of the CBA.

Over the course of his employment with GPC, Edwards worked in several different departments of the company, based in different cities throughout Georgia. He maintains that he has been denied various promotions to union positions, especially in the two years immediately before the suit was filed, and was frequently denied even an opportunity to interview for these positions. In April 2000, Edwards did interview for a Cable Locator position, but was denied the job.

Edwards says that he was fully qualified for the position, having worked for 5 months as a Cable Locator in a temporary capacity, and having worked as a Lineman, which involved more extensive and difficult responsibilities than a Cable Locator. In spite of his qualifications, Edwards did not receive the job, which was instead given to a white co-worker whom Edwards argues was less qualified.

Edwards was denied the Cable Locator position after his interviewers found his demeanor “too relaxed.” During his tenure at GPC, Edwards suffered injuries to his back, and as a result he has had several extended absences and light duty assignments. Between May and November, 1998, Edwards took a medical leave from GPC, returning to take a temporary position which he held until its elimination in April 2000. Since that time he has been out of work on disability leave.

Edwards filed a complaint with the EEOC in 1999, alleging a violation of the Americans with Disabilities Act (“ADA”), but he did not allege any race-based discrimination at that time. Less than a week prior to the filing of this suit, however, he lodged a complaint with the EEOC that alleged individual and class-wide racial discrimination. On appeal, Edwards raises three other promotion claims that were not included in the Complaint. Charcella Green (“Green”) had been employed at GPC for 17 years when this suit was filed.

For the last 11 of those years she worked as an Education Services Coordinator, after having been employed in various capacities since initially joining GPC as a temporary intern in 1988. While Green advanced through various positions at GPC, she acquired several educational credentials. Prior to joining GPC, Green received a B.S. degree in Human Services Administration from Mercer University, and in 1983 she was awarded a Master of Social Work degree. While working at GPC, she continued her education, eventually receiving a Ph.D. from Clark Atlanta University.

From 1996 until her department was eliminated in 2002, Green worked at GPC’s headquarters in Atlanta in a range of exempt positions in the Educational Services department. She complains that she was wrongfully denied an Educational Services Manager position in 1996 and a Community Advisor position in 1997. On July 11, 1998, Green was again denied a Manager of Educational Services position that was awarded to a white employee. Green also says that she has been compensated at a less favorable rate than white employees. Patricia Harris (“P.

Harris”) voluntarily resigned from SCS in August 1999, prior to the commencement of this case. P. Harris began working at GPC in 1990, and became an employee of SCS in 1997, when she began working as a Market Research Analyst in Atlanta. P. Harris did not apply for any promotions during the relevant period and did not file any complaints with the EEOC. However, she maintains that she was only granted her Market Research Analyst position after repeatedly complaining — including to a GPC Vice President — about the difficulty of advancing. P.

Harris contends that her white counterparts were promoted much more easily, and says that once she began working as a Market Research Analyst, she was provided with less training and support than her white colleagues. P. Harris also claims that she suffered discrimination in compensation on account of race, observing that she was paid at the low end of her salary grade, while white colleagues were more generously compensated, notwithstanding their being less qualified. Indeed, P. Harris points out that she earned a Master of Business Administration degree, while her white coworkers lacked that credential.

Sarah Jean Hams (“S.J. .Harris”) was employed by GPC from 1979 until 2000, working in various non-exempt, non-union customer services positions in Lawrence-ville, Duluth, and Gwinnet, Georgia. Hired in 1979 as a General Clerk, she had been promoted to Operating Assistant by 1989. In 1994, S.J. Harris was assigned to work in Duluth as a Secretary; her job title was changed to Region Support Representative in 1997. Working as a Region Support Representative, S.J. Harris was the only African-American employee in an 18-person department.

She alleges that she was discriminated against in her employee evaluations, claiming that she received lower evaluation scores than she deserved, while white employees received inappropriately positive evaluations. She also observes that she received only “minuscule” salary increases during her tenure, and has therefore received lower compensation than white employees. S.J. Harris did not apply for any promotions during the relevant period and has brought no promotion claims. However, in addition to raising her compensation discrimination claims, S.J.

Harris contends that she was wrongfully fired in retaliation for her involvement in this lawsuit. GPC denies any retaliatory motive and contends that she was fired for disciplinary reasons, after having received several negative evaluations and having been placed in a disciplinary program. While on “last-chance” probation, S.J. Harris was again disciplined, and then was terminated. Irene McCullers (“McCullers”) was hired by GPC in 1978, and worked in several clerical positions until 1997, when she was transferred to SCS.

Since 1997, McCullers has worked at SCS as a nonexempt Processing Operator I in the Information Management Services Department. Altogether, McCullers had spent 23 years working for the defendant companies at the time suit was filed. McCullers alleges that she was discriminated against in terms of her compensation, promotions, and performance evaluations. Specifically, McCullers contends that within her department there are six employees working as Processing Operators: two African-American and two white employees hold Processing Operator I titles, and two white employees hold Processing Operator Senior positions.

McCullers alleges that she and her African-American colleague are victims of racial discrimination in that they are the lowest paid employees. McCullers adds that she is paid approximately 10% less than her white colleagues, even though she has an associates degree in business administration, while the white employees have no more than a high school education. McCullers also maintains that she has been discriminated against in promotions.

She avers that at the time her department was reorganized in 1997, she was discrimi-natorily denied a progressive promotion from Processing Operator I to Processing Operator Senior, while at least one white counterpart was promoted to the senior position. In the two years before this litigation commenced, McCullers was again denied a promotion to Processing Operator Senior, while another white colleague was promoted. Finally, Carolyn Wilson (“Wilson”), the seventh of the named plaintiffs, began working for the defendants in 1985, when she was hired by GPC as an employee in its Customer Service Operations.

In 1997, she was hired as a Project Analyst in the Finance Center at SCES. Wilson maintains that she experienced racial discrimination in promotion and compensation decisions, as well as with respect to her performance evaluations. Wilson says that she uniformly performed her job well and received positive performance evaluations, and yet was denied numerous promotions for which she was qualified. She adds that while she applied for approximately 15 promotions within the defendant companies, she was granted only 3 interviews, and was denied several promotions because of her race.

In November 1997, Wilson applied for and was granted a Project Analyst Senior position at SCES, which she learned of through the JobNet system, but her position was re-classified as a Project Analyst III, a lower-level position than she had sought. The following year, Wilson applied for and was offered a Customer Service Representative position, but turned it down. At approximately the same time this lawsuit was commenced, Wilson filed a charge with the EEOC, but she did not identify the specific event or events that constituted the discrimination she faced, instead alleging discrimination more generally.

She maintains that she is compensated less favorably than at least four white colleagues who perform the same job as she does, even though they are classified at the same, or lower, salary grades. To summarize, of the seven named plaintiffs, three no longer worked for any of the defendants at the time this case was commenced (Edwards, P. Harris, and S.J. Harris). Of the four remaining employees, one held a position which was covered by the CBA (Cooper), two held exempt positions outside senior management (Green and Wilson), and one held a non-exempt position (McCullers).

Although at various times during their employment with the defendants, some of the plaintiffs worked in Georgia locations outside metropolitan Atlanta, most of them were employed in Atlanta for most or all of their tenures. Several of the named plaintiffs worked in office environments characteristic of the central administrative operations that- took place in Atlanta, and even Edwards and Cooper — who labored in connection with the direct delivery of electricity — worked in more developed urban areas of Georgia.

None of the named plaintiffs worked in the less densely populated, rural areas of the state, and although the plaintiffs sought to represent a broad class including employees based throughout Florida, Mississippi, and Alabama, none of them worked within these states.

All of the plaintiffs complained of discrimination in either compensation or promotion, and while none of the plaintiffs produced direct evidence that the personnel decisions in their individual cases were motivated by race, each plaintiff alleged that he or she was victimized by a common policy of subjective decision-making that allowed discrimination to persist in the defendant companies.

Specifically, the plaintiffs’ Complaint asserted that there was a class-wide “continuing pattern and practice of racial discrimination,” that they and the putative class members were subjected to discriminatory treatment, and that the defendants’ policies and practices “had an ongoing disparate impact.” Two major issues are presented in this appeal. First, plaintiffs argue that the district court abused its discretion when it denied their motion for class certification under Rule 23 of the Federal Rules of Civil Procedure. Second, they say that the district court erred when it granted

summary judgment in favor of the defendants on each of the seven named plaintiffs’ individual claims. We address each of these issues in turn. II. We begin with the plaintiffs’ basic argument that the district court abused its discretion when it denied their motion for class certification. In denying the plaintiffs’ motion, the district court first considered the plaintiffs’ anecdotal and statistical evidence, before concluding that the plaintiffs could not satisfy the basic requirements of Rule 23(

a) of the Federal Rules of Civil Procedure. Alternatively, the district court held that individual questions of law and fact plainly outweighed any common questions, and therefore that this suit was not appropriate for certification under either subsection (2) or (3) of Rule 23(

b) of the Federal Rules of Civil Procedure. The plaintiffs maintain that the district court’s denial of class certification results from “fundamental errors of law and a fundamental misconception of the role of the court in ruling on a motion for class certification.” Appellants’ Brief at 8. We remain unpersuaded. A. Questions concerning class certification are left to the sound discretion of the district court. Armstrong v. Martin Marietta Corp., 138 F.3d 1374, 1386 (11th Cir.1998) (en banc); Freeman v.

Motor Convoy, Inc., 700 F.2d 1339, 1347 (11th Cir.1983). “A district court’s decision whether or not to certify a class under Rule 23 of the FRCP is reviewed for abuse of discretion. As long as the district court’s reasoning stays within the parameters of Rule 23’s requirements for the certification of a class, the district court decision will not be disturbed.” Hines v. Widnall, 334 F.3d 1253, 1255 (11th Cir.2003) (citations omitted). Even if we would have certified a class, that does not mean the district court abused its discretion in declining to do so. Id.; see also Shroder v.

Suburban Coastal Corp., 729 F.2d 1371, 1374 (11th Cir.1984). Indeed, the distinguishing hallmark of abuse-of-discretion review is that it “presupposes a zone of choice within which the trial courts may go either way.” Kern v. TXO Prod. Corp., 738 F.2d 968, 971 (8th Cir.1984). By definition ... under the abuse of discretion standard of review there will be occasions in which we affirm the district court even though we would have gone the other way had it been our call. That is how an abuse of discretion standard differs from a de novo standard of review.

As we have stated previously, the abuse of discretion standard allows “a range of choice for the district court, so long as that choice does not constitute a clear error of judgment.” Rasbury v. I.R.S., 24 F.3d 159, 168 (11th Cir.1994) (quoting United States v. Kelly, 888 F.2d 732, 745 (11th Cir.1989) (citation omitted)). First, the plaintiffs arg-ue that the district court erred .by “reaching the merits” of the plaintiffs’ claim at the preliminary stage of class certification.

Thus, we are required to determine whether the trial court properly assessed the plaintiffs’ evidence only to the extent necessary to decide the issue of class certification, or whether the trial judge impermissibly usurped the role of the jury by ruling on the merits of the plaintiffs’ statistical and expert evidence. After thorough review of this record, we are satisfied that the district court did not improperly invade the jury’s province when it conducted (as it was required to do in this case) a rigorous analysis of the evidence proffered by the parties at the class certification stage.

The plaintiffs rely on the Supreme Court’s often-cited admonition in Eisen v. Carlisle & Jacquelin, 417 U.S. 156, 94 S.Ct. 2140, 40 L.Ed.2d 732 (1974), that “nothing in either the language or history of Rule 23 ... gives a court any authority to conduct a preliminary inquiry into the merits of a suit in order to determine whether it may be maintained as a class action.” Id. at 177, 94 S.Ct. at 2152.

However, we have noted that “[w]hile it is true that a trial court may not properly reach the merits of a claim when determining whether class certification is warranted, this principle should not be talismani-cally invoked to artificially limit a trial court’s examination of the factors necessary to a reasoned determination of whether a plaintiff has met her burden of establishing each of the Rule 23 class action requirements.” Love v. Turlington, 733 F.2d 1562, 1564 (11th Cir.1984) (citation omitted).

Indeed, both the Supreme Court and this Court have noted since Eisen that evidence pertaining to the requirements embodied in Rule 23 is often intertwined with the merits, making it impossible to meaningfully address the Rule 23 criteria without at least touching on the “merits” of the litigation. See Coopers & Lybrand v. Livesay, 437 U.S. 463, 469 n. 12, 98 S.Ct. 2454, 2458 n. 12, 57 L.Ed.2d 351 (1978) (“Evaluation of many of the questions entering into determination of class action questions is intimately involved with the merits of the claims.” (citation omitted)); Nelson v. U.S.

Steel Corp., 709 F.2d 675, 679 (11th Cir.1983) (explaining that Rule 23(

a) evidence “is often intertwined with the merits”). As the Supreme Court has noted, the bare allegation that racial discrimination has occurred “neither determines whether a class action may be maintained ... nor defines the class that may be certified.” Gen. Tel. Co. of the Southwest v. Falcon, 457 U.S. 147, 157, 102 S.Ct. 2364, 2370, 72 L.Ed.2d 740 (1982). Rather, before a district court determines the efficacy of class certification, it may be required to make an informed assessment of the parties’ evidence.

That a trial court does so does not mean that it has erroneously “reached the merits” of the litigation. See Kirkpatrick v. J.C. Bradford & Co., 827 F.2d 718, 722 (11th Cir.1987). In this case, the district court was obliged to make some preliminary assessment of the plaintiffs’ evidence to determine, at the very least, whether the named plaintiffs were claiming discrimination that was common to the members of the putative class. Indeed, the district court would have erred if it had certified a class without first determining that the named plaintiffs had claims common to those of the unnamed class members. B.

We turn, then, to whether the court’s substantive determinations at the class certification stage amounted to an abuse of discretion. The plaintiffs say that the district court erred by concluding that they had failed to establish the commonality, typicality, and adequacy requirements of Rule 23(a). They claim that there were questions of law and fact common to the entire class and that the claims of the named plaintiffs were in fact typical of all the class members. The four basic requirements of Rule 23(

a) are “designed to effectively limit class claims to those ‘fairly encompassed’ by the named plaintiffs’ individual claims.” Prado-Steiman, 221 F.3d at 1278 (quoting Falcon, 457 U.S. at 156, 102 S.Ct. at 2370) (internal quotation marks omitted). As the district court put it, Rule 23(

a) is designed to ensure that “the common bond between the class representatives’ claims and those of the class is strong enough so that it is fair for the fortunes of the class members to rise or fall with the fortunes of the class representatives.” Cooper, 205 F.R.D. at 608-09. The typicality and commonality requirements are distinct but interrelated, as the Supreme Court has made clear: The commonality and typicality requirements of Rule 23(

a) tend to merge. Both serve as guideposts for determining whether under the particular circumstances maintenance of a class action is economical and whether the named plaintiffs claim and the class claims are so interrelated that the interests of the class members will be fairly and adequately protected in their absence. Falcon, 457 U.S. at 157 n. 13, 102 S.Ct. at 2370 n. 13. We have similarly explained that “the commonality and typicality requirements of Rule 23(

a) overlap. Both requirements focus on whether a sufficient nexus exists between the legal claims of the named class representatives and those of individual class members to warrant class certification.” Prado-Steiman, 221 F.3d at 1278. “A class representative must possess the same interest and suffer the same injury as the class members in order to be typical under Rule 23(a)(3).” Murray v. Auslander, 244 F.3d 807, 811 (11th Cir.2001). “[Tjypicality measures whether a sufficient nexus exists between the claims of the named representatives and those of the class at large.” Prado-Steiman, 221 F.3d at 1279.

While not entirely dissimilar to the typicality requirement, Rule 23(a)’s commonality requirement measures the extent to which all members of a putative class have similar claims. “Traditionally, commonality refers to the group characteristics of the class as a whole [while] typicality refers to the individual characteristics of the named plaintiff in relation to the class.” Id. Under the commonality requirement, “a class action must involve issues that are susceptible to class-wide proof.” Murray, 244 F.3d at 811.

Neither the typicality nor the commonality requirement “mandates that all putative class members share identical claims, and ... factual differences among the claims of the putative class members do not defeat certification.” Baby Neal v. Casey, 43 F.3d 48, 56 (3d Cir.1994) (citation omitted); see also Prado-Steiman, 221 F.3d at 1279 n. 14; Appleyard v. Wallace, 754 F.2d 955, 958 (11th Cir.1985). Nevertheless, the named plaintiffs’ claims must still share “the same essential characteristics as the claims of the class at large.” Appleyard, 754 F.2d at 958 (citation and internal quotation marks omitted).

As for the typicality requirement, the district court concluded that while some of the named plaintiffs had claims that could be typical of subgroups within the larger class, collectively the named plaintiffs did not have claims typical of the entire class. Thus, for example, plaintiff Edwards’s claims were uniquely tied to issues relating to his disability; plaintiff S.J.

Haras’s claims could not be divorced from the disciplinary issues relating to her case; and although Cooper’s claims were typical of union members denied promotion to non-union jobs, they were not typical of those union members who might have complaints relating to the denial of union positions. Similarly, none of the named plaintiffs had claims typical of members of the putative class who were denied senior management positions, nor did any of the named plaintiffs have compensation claims typical of upper-management employees.

And while the plaintiffs’ claims may have been typical of some subclasses of the defendants’ workforce, it is not certain that their claims would be typical of other subclasses. Thus, for example, the specific job-related criteria GPC uses to make employment decisions affecting its rural power-delivery operations may be very different from the criteria underlying decisions about employees in their administrative offices. As a result, the named plaintiffs did not present claims typical of the full range of employees in their putative class.

The plaintiffs argue, nevertheless, that the district court erred by focusing on the idiosyncratic features of the different named plaintiffs’ claims instead of acknowledging that the claims of the plaintiffs and the members of the larger class were based on the same legal theories. That is, the plaintiffs say that their individual injuries “arise out of the same pattern and practice [of discrimination] and same [discriminatory] policies that give rise to the injuries of the class members.” Appellants’ Brief at 18.

We reiterate that our review is only for abuse of discretion, and that whether we would have certified the class in the first instance — or even some grouping of subclasses — is not dispositive. Hines, 334 F.3d at 1257; see also Shroder, 729 F.2d at 1374. Here, the district court’s analysis as to typicality fell within the range of choice permissible on abuse-of-discretion review and did not represent a clear error in judgment. See Kelly, 888 F.2d at 745.

First, the compensation and promotion decisions affecting each of the named plaintiffs were made by individual managers in disparate locations, based on the individual plaintiffs’ characteristics, including their educational backgrounds, experiences, work achievements, and performance in interviews, among other factors. In Love v. Turlington, we held that the district court had not abused its discretion in denying class certification based on the plaintiffs’ failure to meet the typicality and commonality requirements. 733 F.2d at 1564.

That case involved a challenge to Florida’s SSAT-I, a test designed to identify high school students who had failed to master one of the state’s academic performance standards. Students failing the test were identified for remedial assistance, but a failure did not automatically render a student ineligible for a high school diploma. Rather, decisions on each individual student’s eligibility for a diploma were “made by individual teachers on the basis of [the student’s] individual achievements.

Each student’s situation differs, and the diploma is denied each student for reasons which are unique to his situation, and which do not necessarily correspond to his performance on the SSAT-I.” Id. Similarly, in the instant case the compensation and promotion decisions affecting each of the named plaintiffs were based on factors that, by and large, appear unique to each of the plaintiffs in question.

In addition, the plaintiffs sought to represent a very broad class that purported to represent all African-American employees of the defendants, at all levels of the corporate hierarchy of all the defendant companies. Because the plaintiffs asserted broad claims on behalf of a broad class, they were required to identify representative plaintiffs who shared those broad claims.

However, while the different named plaintiffs may have had claims that were typical of some conceivable subgroups of the overall class, the seven named plaintiffs, collectively, did not have claims that would have been typical of the entire class. See Falcon, 457 U.S. at 156, 102 S.Ct. at 2370; Hines, 334 F.3d at 1257-58. As a result, we are constrained to conclude that the district court’s conclusions regarding typicality did not constitute a clear error of judgment, nor were they otherwise outside the range of choices the district court was allowed to make.

As for the commonality requirement, the district court concluded that it was impossible to say that the employment histories of the named plaintiffs and the ways in which they claimed to have experienced discrimination could be “fairly compared with the history or individual experiences of absent class members,” or that the claims of class members shared such common features that rulings could be fashioned to fairly adjudicate the claims as a group. Cooper, 205 F.R.D. at 610-611.

Where, as here, class certification was sought by employees working in widely diverse job types, spread throughout different facilities and geographic locations, courts have frequently declined to certify classes. See, e.g., Zachery v. Texaco Exploration & Prod., Inc., 185 F.R.D. 230, 239 (W.D.Tex.1999) (denying certification where proposed class spread across fifteen states and involved seventeen business units); Troupe v. Randall’s Food & Drug, Inc., No. CIV.A. 3:98-CV-2462, 1999 WL 552727, at *5 (N.D.Tex.

July 28, 1999) (declining to certify proposed class that covered at least “fifty separate stores spread over two large cities and their outlying suburbs” with “management practices varying] widely according to stores across the division”). Since the hiring, compensation, and promotion decisions at issue were made by different managers in different companies implementing different policies, even if the named plaintiffs established that they were, individually, subjected to intentional discrimination, they would not necessarily have established that other class members suffered from the same discrimination.

Commonality in the claims of the broad class the plaintiffs sought to certify would have to be established by showing that the discrimination sustained was either part of an overarching pattern and practice of intentional discrimination on the part of the defendants or the result of the discriminatory disparate impact of a facially neutral employment policy. We explain the elements required to sustain a successful pattern and practice or disparate impact claim more fully in

Section III.B., infra, but a brief overview suffices for our purposes here. To establish a “pattern or practice” of disparate treatment, the plaintiff must show that intentional discrimination was the employer’s “standard operating procedure.” Joe’s Stone Crab, 220 F.3d at 1274 (explaining different types of discrimination claims in context of gender discrimination). The plaintiff can prove that discrimination was the standard operating procedure “through a combination of statistics and anecdotes.” Id.

To prove disparate impact, a plaintiff must establish the existence of a statistically significant disparity among members of different groups affected by a type of employment decision; a specific, facially neutral employment practice involved in the decision; and a causal nexus between the facially neutral employment practice and the statistically significant disparity. Id. Thus, it was plainly necessary for the district court to evaluate the statistical evidence the plaintiffs submitted in order to determine whether it established the discrimination of which plaintiffs complained.

The district court concluded the statistical “analysis ha[d] some limitations which undermine[d] its usefulness in measuring whether Defendants’ employment practices [we]re racially neutral.” Cooper, 205 F.R.D. at 613. Because of substantial analytical flaws in the statistical evidence, the plaintiffs had made an “inadequate showing ... to raise a presumption of discrimination arising from application of the collective whole of Defendants’ compensation and promotion policies. Thus, disparate impact analysis produce[d] no evidence common to the claims of all class members.

Also, the [statistical evidence] fail[ed] to establish evidence of a pattern and practice of discrimination.” Id. at 615 (emphasis added). The plaintiffs’ primary statistical evidence was found in reports by Dr. Janice Madden, who analyzed employment data provided by the defendants. Dr. Madden’s reports concluded that African Americans received fewer promotions than would be expected given their numbers in the workforce. However, after analyzing Madden’s methodology and conclusions, the district court identified several basic infirmities that undermined the reports’ substantive value.

The district court did not exclude Dr. Madden’s reports because she was unqualified or because the reports were based on a wholly unreliable methodology; rather, the court accepted the reports’ conclusions but determined that they still failed to establish that the named plaintiffs had claims in common with other class members under either a pattern and practice or disparate impact theory.

In the first place, the district court noted that Madden’s reports did not incorporate variables that would allow for the comparison of individuals who were similarly situated with respect to managerial decision makers, job types, locations, departments, and the specific criteria relevant for the jobs in question.

Madden did not tailor her analysis to the specific positions, job locations, or departmental and organizational structures in question; however, the wide-ranging and highly diversified nature of the defendants’ operations requires that employee comparisons take these distinctions into account in order to ensure that the black and white employees being compared are similarly situated. See Wards Cove Packing Co. v.

Atonio, 490 U.S. 642, 656-57, 109 S.Ct. 2115, 2124-25, 104 L.Ed.2d 733 (1989) (“[A] Title VII plaintiff does not make a case of disparate impact simply by showing that, ‘at the bottom line,’ there is racial imbalance in the work force.”); see also Brown v. Am. Honda Motor Co., 939 F.2d 946, 952 (11th Cir.1991) (“Statistics ... without a [proper] analytic foundation[ ] are virtually meaningless.”); Balderston v.

Fairbanks Morse Engine Div. of Coltec Indus., 328 F.3d 309, 319 (7th Cir.2003) (“The Supreme Court has emphasized the importance of looking to the proper base ‘group’ when making statistical comparisons and examining all of the surrounding facts and circumstances which create the statistics themselves.”). As the district court observed, Madden’s reports did not take into account the type or level of acquired skills of job applicants.

Madden’s “education” component failed to take into account the field of study, the relevance of that field to any position in question, or the quality of the educational institutions involved, factors which may be important in some managers’ employment decisions. Moreover, Madden calculated “experience” only by tabulating the amount of time that had passed since an individual finished his or her formal education and the amount of time that individual had spent on the defendants’ payroll.

That two employees completed their education in the same year does not ensure that they have similar levels of job-related experience, since what those employees have done in the intervening years may be extraordinarily different. While Madden’s “experience” variable may have measured the passage of time, it did not in any way factor in the quality, type, or relevance of an employee’s experience. And the reports did not factor in any employee’s actual job performance, a consideration that is undeniably important in decisions relating to compensation and promotion.

Indeed, without taking into account job performance at all, the fact that two employees began their employment with the defendants at the same time and have not been promoted at the same rate does not establish that discrimination is at work. In essence, the district court concluded that these methodological deficiencies rendered it “impossible to determine what the [salary and promotions] gaps [we]re, whether they [we]re statistically significant, or whether factors other than race [we]re involved.” Cooper, 205 F.R.D. at 615.

Thus, the district court observed that “there ha[d] been an inadequate showing by Plaintiffs to raise a presumption of discrimination arising from application of the collective whole of Defendants’ compensation and promotion policies.” Id.; see also Maddox v. Claytor, 764 F.2d 1539, 1552 (11th Cir.1985) (“[Multiple regression analysis] measures the probability that the calculated disparity could occur randomly — but the analysis in no way validates the calculation of the disparity itself.

If the tested disparity is based on erroneous assumptions or suffers from flaws in the underlying data, then standard deviation analysis is foredoomed to yield an equally faulty result.”); Eastland v. Tenn. Valley Auth., 704 F.2d 613, 621-24 (11th Cir.1983). The district court also observed that Madden’s analysis of promotion rates was, similarly, grounded in insufficiently tailored data.

Significantly, Madden failed to distinguish between competitive promotions — -those open to any qualified candidate — and progressive promotions — • those available only to individual employees progressing in a “job family” — a distinction that is crucial when analyzing promotions. See Forehand v. Fla. State Hosp., 89 F.3d 1562, 1572-73 (11th Cir.1996).

As for progressive promotions, because Madden recognized promotions only when an employee received an increase in salary grade, moved from a non-exempt to an exempt position, or moved from a union job into a higher-paying non-union job, Madden’s calculations did not capture those that occurred within the same salary grade. As for competitive promotions, Madden did not take into account how many and which employees actually applied for them, also a significant factor in making any meaningful generalizations about allegedly discriminatory promotion practices.

As part of her analysis, Madden divided the employees into different “pools,” or subsets, of employees, and her analysis revealed statistical deficiencies in only some of the pools. Of the 148 pools of employees analyzed in 1998-1999, Madden’s analysis revealed a representational deficiency of African-American employees in only three pools: two union salary-grade pools and one pool for non-exempt salary-grade-5 employees.

The district court concluded that this actually could be evidence that the promotional deficiencies for African-Americans may have existed only in some segments of the workforce, and therefore not throughout the entire class. Finally, the district court observed that Madden’s reports did not establish the existence of any specific promotion or compensation policy or practice, let alone trace the alleged racial disparities to such a policy or practice.

In addition, the Madden reports did not make any reference to any of the specific named plaintiffs or their specific similarly-situated comparators, making it altogether unclear how the reports could establish commonality among these named plaintiffs’ claims and the overall claims of the affected class. The district court concluded that because of these flaws in Madden’s models, it was impossible to determine how wide a real gap, if any, existed in salaries and promotion rates among white and black employees, or whether any such gap was attributable to factors other than race.

We are persuaded that the district court’s basic conclusion — that plaintiffs’ statistical evidence was insufficient to establish a presumption of discrimination common to the claims of all members of the putative class — did not constitute a clear error of judgment, nor was the conclusion otherwise outside the acceptable range of choices. The plaintiffs also proffered, in addition to the statistical evidence, a range of anecdotal evidence that, they claimed, shed additional light on the defendants’ allegedly discriminatory policies and practices.

Among the evidence presented were reports of several ugly incidents involving nooses displayed at some locations within some of the defendants’ facilities, reports of racial jokes and slurs, and the belief expressed by various employees that there was a “glass ceiling” for African-American employees. Finally, plaintiffs filed a series of affidavits of prospective class members who offered anecdotal evidence that repeated many of the claims plaintiffs had made in their pleadings.

The district court determined that the anecdotal evidence was “inadequate to establish ... a pattern and practice of discrimination.” Cooper, 205 F.R.D. at 619. Again, we conclude that the district court acted within its discretion in determining that, given the sheer size and geographically dispersed nature of the defendants’ workforce, the anecdotal evidence — disturbing as some of it may have been — was inadequate to establish discrimination class-wide.

After thorough review of the entire record, we conclude that the district court did not abuse its discretion in determining that the statistical and anecdotal evidence submitted by the plaintiffs did not establish pattern and practice discrimination common to the class or a common disparate impact affecting the defendants’ African-American employees class-wide.

The powerful differences between the named plaintiffs’ claims and the claims of the overall class members meant that procedural fairness for all members of the putative class would not be ensured if the claims of the entire class were allowed to rise or fall on the fortunes of the named plaintiffs’ claims. C. Even if we concluded, however, that the district court abused its discretion in determining that the plaintiffs did not meet two of the requirements of Rule 23(a), certification would still be inappropriate unless the plaintiffs also could satisfy at least one of the requirements of Rule 23(b).

See Jackson v. Motel 6 Multipurpose, Inc., 130 F.3d 999, 1005 (11th Cir.1997). As another ground for its decision, the district court concluded that certification was inappropriate under either Rule 23(b)(2) or Rule 23(b)(3). On appeal the plaintiffs say that the district court erred in refusing to certify a class under Rule 23(b)(2) for injunctive relief and back pay or to certify a hybrid class under Rule 23(b)(2) for injunctive relief, while severing the damages phase of the proceedings by allowing opt-outs for damages.

They also argue that the court should have certified a class under Rule 23(b)(3) for injunctive relief, back pay, and damages. We remain unconvinced and, accordingly, hold that the district court did not abuse its discretion in concluding that the requirements of Rule 23(

b) were not satisfied either. Under Rule 23(b)(2), class certification is appropriate when the requirements of Rule 23(

a) are met and the defendant “has acted or refused to act on grounds generally applicable to the class, thereby making appropriate final injunctive relief or corresponding declaratory relief with respect to the class as a whole.” Fed.R.Civ.P. 23(b)(2). Back pay is considered equitable relief and can therefore be awarded in a case certified under Rule 23(b)(2). Pettway v. Am. Cast Iron Pipe Co., 494 F.2d 211, 257 (5th Cir.1974).

Here, however, the plaintiffs also sought compensatory and punitive damages, legal relief that can only be awarded in a Rule 23(b)(2) class action when the damages sought are “incidental” to the claims for equitable and declaratory relief. Murray, 244 F.3d at 812.

Damage claims in a Rule 23(b)(2) class action must be incidental to the equitable and declaratory relief because the basic premise of such a class action' — -that class members suffer a common injury properly addressed by class-wide equitable relief— “begins to break down when the class seeks to recover ... monetary relief to be allocated based on individual injuries.” Allison v. Citgo Petroleum Corp., 151 F.3d 402, 413 (5th Cir.1998) (quoting Eubanks v. Billington, 110 F.3d 87, 95 (D.C.Cir.1997)) (internal quotation marks omitted).

We explained in Murray those criteria used to evaluate whether money damages should be considered “incidental” to equitable claims in a 23(b)(2) case. In Murray, we adopted the position of the Fifth Circuit in Allison v. Citgo Petroleum Corp.: By incidental, we mean damages that flow directly from liability to the class as a whole on the claims forming the basis of the injunctive or declaratory relief.... Liability for incidental damages should not ... entail complex individualized determinations.

Thus, incidental damages will, by definition, be more in the nature of a group remedy, consistent with the forms of relief intended for (b)(2) class actions. Murray, 244 F.3d at 812 (quoting Allison, 151 F.3d at 415). In keeping with the requirement that monetary damages be incidental to equitable relief in Rule 28(b)(2) class actions, the advisory committee’s note to Rule 23 explains that certification under

section (b)(2) is not proper in “cases in which the appropriate final relief relates exclusively or predominantly to money damages.” Fed.R.Civ.P. 23, advisory committee’s note (emphasis added). The plaintiffs argue, nevertheless, that the district court could have certified a class under Rule 23(b)(2) only as to the injunctive and declaratory prayer for relief, thereby excluding altogether the damages issues from class certification.

However, to the extent the named plaintiffs were willing to forego class certification on damages in order to pursue injunctive relief that consisted of an admonition to follow general principles of settled law, it is far from clear that the named plaintiffs would adequately represent the interests of the other putative class members.

Indeed, to many of the class members (and especially to those who no longer work for the defendants), the monetary damages requested might be of far greater significance than injunctive relief, stated at a high order of abstraction, that simply directs the defendants not to discriminate. Moreover, determining the level of damages to which each class member was entitled plainly would require detailed, ease-by-case fact finding, carefully calibrated for each individual employee.

The “complex, individualized determinations” necessary to fix the appropriate level of individual damage awards in this case are exactly the type that Murray and Allison make clear should not be considered “incidental” to the claims for injunctive and declaratory relief. Finally, as the district court noted, since the plaintiffs demanded a jury trial in this case, the parties were entitled under the Seventh Amendment to have all matters at law determined by a single jury before having decisions concerning equitable relief made by the trial court. See Cooper, 205 F.R.D. at 629 (citing Ross v.

Bernhard, 396 U.S. 531, 538-39, 90 S.Ct. 733, 738, 24 L.Ed.2d 729 (1970)). Under the circumstances of this case, even assuming that the district court could conduct an initial bench trial on the merits of the equitable claims, and that the court actually found in favor of the plaintiffs, it would still be necessary for a single jury to hear and rule on more than 2,000 individual claims for compensatory damages. Again, on this record we can discern no abuse of discretion in the district court’s refusal to proceed under the class action umbrella.

The plaintiffs also say that even if certification under Rule 23(b)(2) was not appropriate, the district court still should have certified the class for equitable relief, back pay, and damages under Rule 23(b)(3). But, for class certification under Rule 23(b)(3) to be appropriate, common questions still must “predominate over any questions affecting only individual members” and the class action mechanism must be “superior ... for the fair and efficient adjudication of the controversy.” Fed. R.Civ.P. 23(b)(3).

We have explained that in addition to the requirements embodied in Rule 23(a), Rule 23(b)(3) “imposes two additional requirements, and increased efficiency is only one of them. Predominance is the other.... ” Jackson, 130 F.3d at 1006. Therefore, “the issues in the class action that are subject to generalized proof and thus applicable to the class as a whole, must predominate over those issues that are subject only to individualized proof.” Kerr v. City of W. Palm Beach, 875 F.2d 1546, 1558 (11th Cir.1989) (citation and internal quotation marks omitted).

Common issues will not predominate over individual questions if, “as a practical matter, the resolution of ... [an] overarching common issue breaks down into an unmanageable variety of individual legal and factual issues.” Andrews v. Am. Tel. & Tel. Co., 95 F.3d 1014, 1023 (11th Cir.1996).

Our case law plainly establishes that a class action should not proceed under Rule 23(b)(3) when it appears that “most, if not all, of the plaintiffs’ claims will stand or fall, not on the question whether [the defendant] has a practice or policy of [race] discrimination, but on the resolution of ... highly case-specific factual issues.” Rutstein, 211 F.3d at 1235 (quoting Jackson, 130 F.3d at 1006) (internal quotation marks omitted). Indeed, the predominance requirement under Rule 23(b)(3) is “far more demanding” than Rule 23(a)’s commonality requirement.

Jackson, 130 F.3d at 1005 (citation and internal quotation marks omitted). Here, the plaintiffs attempted to use statistical proof to establish the existence of a generalized policy of discrimination that provided the common thread linking together plaintiffs’ otherwise disparate and individualized claims.

However, we agree with the district court that substantial limitations in the plaintiffs’ statistical evidence rendered the form of proof wholly insufficient to show that any pattern or practice of discrimination disparately affected the plaintiffs’ class, or, for that matter, that the defendants had a general policy of discrimination. See Cooper, 205 F.R.D. at 630. Moreover, the individual determinations on liability and damages necessary for the individual plaintiffs to succeed would require highly fact-specific inquiries concerning each plaintiff.

Thus, we can discern no abuse of discretion in the district court’s determination that the issues subject to individualized proof predominated over those that could be established with class-wide proof, and therefore that the Rule 23(b)(3) class action procedure would not be superior for the fair and efficient adjudication of the controversy. III. A. Finally, we turn to the district court’s orders granting

summary judgment in favor of the defendants on the claims of each of the seven individual plaintiffs. We review the district court’s orders granting

summary judgment de novo, applying the same legal standards as the district court did and viewing all of the facts in the light most favorable to the non-moving party. See, e.g., Johnson v. Booker. T. Washington Broad. Serv., Inc., 234 F.3d 501, 507 (11th Cir.2000).

Summary judgment is appropriate when “the pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any, show that there is no genuine issue as to any material fact and that the moving party is entitled to a judgment as a matter of law.” Fed.R.Civ.P. 56(c). ‘Where the record taken as a whole could not lead a rational trier of fact to find for the non-moving party, there is no ‘genuine issue for trial.’ ” Matsushita Elec. Indus. Co. v. Zenith Radio Corp., 475 U.S. 574, 587, 106 S.Ct. 1348, 1356, 89 L.Ed.2d 538 (1986) (quoting First Nat’l Bank of Ariz. v.

Cities Serv. Co., 391 U.S. 253, 289, 88 S.Ct. 1575, 1592, 20 L.Ed.2d 569 (1968)). “[A]ll reasonable doubts about the facts should be resolved in favor of the non-movant.” Burton v. City of Belle Glade, 178 F.3d 1175, 1187 (11th Cir.1999) (quoting Clemons v. Dougherty County, 684 F.2d 1365, 1368-69 (11th Cir.1982)). B. As we have observed, the plaintiffs claimed racial discrimination under three distinct Title VII theories: pattern and practice discrimination, disparate treatment discrimination, and disparate impact discrimination. See Joe’s Stone Crab, 220 F.3d at 1273.

The first two theories require the plaintiffs to prove discriminatory intent; the third does not. Id. The plaintiffs also assert claims under

Section 1981, which, unlike Title VII, only provides a cause of action for claims involving intentional discrimination. Gen. Bldg. Contractors Ass’n v. Pennsylvania, 458 U.S. 375, 391, 102 S.Ct. 3141, 3150, 73 L.Ed.2d 835 (1982); Brown, 939 F.2d at 949. The basic legal framework governing these theories is by now well-established. In a disparate treatment case, the plaintiff bears the burden of proving that the employer intentionally discriminated against him because of his race. See Holifield v. Reno, 115 F.3d 1555, 1565 (11th Cir.1997).

The plaintiff can establish discriminatory intent through either direct or indirect evidence; to establish a prima facie case of intentional discrimination using circumstantial evidence, plaintiffs may use the framework established in McDonnell Douglas Corp. v. Green, 411 U.S. 792, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973).

Again, in a “pattern and practice” disparate treatment case, “the plaintiff must prove, normally through a combination of statistics and anecdotes, that discrimination is the company’s ‘standard operating procedure.’ ” Joe’s Stone Crab, 220 F.3d at 1274 (citation omitted). “To meet this burden of proof, a plaintiff must ‘prove more than the mere occurrence of isolated or accidental or sporadic discriminatory acts.

It ha[s] to establish by a preponderance of the evidence that ... discrimination [is] ... the regular rather than unusual practice.’ ” Id. at 1286-87 (quoting Teamsters, 431 U.S. at 336, 97 S.Ct. at 1843). To state a claim under a disparate impact theory, in contrast, a plaintiff need not establish that he suffered intentional discrimination. Rather, “disparate impact theory prohibits neutral employment practices which, while non-discriminatory on their face, visit an adverse, disproportionate impact on a statutorily-protected group.” Id. at 1274.

We have noted that the disparate impact theory is “a doctrinal surrogate for eliminating unprovable acts of intentional discrimination hidden innocuously behind facially-neutral policies or practices.” Id. Thus, to state a prima facie case of disparate impact discrimination, a plaintiff must establish that (1) there is a significant statistical disparity among members of different racial groups; (2) there is a specific, facially-neutral employment policy or practice; and (3) there is a causal nexus between the specific policy or practice and the statistical disparity. Id.

Moreover, in order to satisfy the third, critical requirement, a plaintiff “must offer statistical evidence of a kind and degree sufficient to show that the practice in question has caused the exclusion of applicants for jobs or promotions because of their membership in a protected group.” Id. at 1274-75 (citation and internal quotation marks omitted). In general terms, disparate impact cases are more amenable than other types of discrimination claims to class treatment, since they arise out of specific policies or practices that have a disproportionate impact on an entire class of employees.

In contrast, when disparate treatment is the basis of a class action, it is generally more likely that individual issues will predominate and make proceeding in a class action format inappropriate. See, e.g., Reyes v.

Walt Disney World Co., 176 F.R.D. 654, 658 (M.D.Fla.1998) (explaining that disparate treatment claims are inherently “highly individualized” and where employees worked in different divisions with different structures and decision makers, each plaintiff was subject to his “own set of unique circumstances surrounding the adverse employment action about which they now attempt to collectively complain”). In a discrimination action brought under either Title VII or

Section 1981, a plaintiff must first establish a pri-ma facie case of discrimination, which the defendant can rebut by offering a legitimate, non-discriminatory reason for the allegedly discriminatory act. See Reeves v. Sanderson Plumbing Prods., Inc., 530 U.S. 133, 142, 120 S.Ct. 2097, 2106, 147 L.Ed.2d 105 (2000). “In other words, the defendant must produc[e] evidence that the plaintiff was rejected, or someone else was preferred, for a legitimate, nondiscriminatory reason.” U.S. Postal Serv. Bd. of Governors v.

Aikens, 460 U.S. 711, 714, 103 S.Ct. 1478, 1481, 75 L.Ed.2d 403 (1983) (citation and internal quotation marks omitted). “It is important to bear in mind ... that the defendant’s burden of rebuttal is exceedingly light.... At this stage of the inquiry, the defendant need not persuade the court that its proffered reasons are legitimate; the defendant’s burden is merely one of production, not proof.” Perryman v. Johnson Prods. Co., 698 F.2d 1138, 1142 (11th Cir.1983) (citation and internal quotation marks omitted).

If the defendant successfully rebuts the plaintiffs prima facie case, “the presumption of discrimination is eliminated.” Chapman v. AI Transp., 229 F.3d 1012, 1024 (11th Cir.2000) (en banc). To survive

summary judgment, the plaintiff must then “come forward with evidence, including the previously produced evidence establishing the prima facie case, sufficient to permit a reasonable factfinder to conclude that the reasons given by the employer were not the real reasons for the adverse employment decision.” Id. (quoting Combs v. Plantation Patterns, 106 F.3d 1519, 1528 (11th Cir.1997)); see also Reeves, 530 U.S. at 143, 120 S.Ct. at 2106; Perryman, 698 F.2d at 1142.

To show that the employer’s reasons were pretextual, the plaintiff must demonstrate “such weaknesses, implausibilities, inconsistencies, incoherencies, or contradictions in the employer’s proffered legitimate reasons for its action that a reasonable factfinder could find them unworthy of credence.” Combs, 106 F.3d at 1538. “If the plaintiff does not proffer sufficient evidence to create a genuine issue of material fact regarding whether each of the defendant employer’s articulated reasons is pretextual, the employer is entitled to

summary judgment on the plaintiffs claim.” Chapman, 229 F.3d at 1024-25; see also Combs, 106 F.3d at 1529 (explaining that the plaintiff must present “sufficient evidence to demonstrate the existence of a genuine issue of fact as to the truth of each of the employer’s proffered reasons for its challenged action”). The plaintiffs argue that they have presented sufficient statistical evidence in support of their pattern and practice and disparate impact claims to raise genuine issues of material fact, and thus to defeat

summary judgment. In contrast to the individual claims of disparate treatment, which we address infra, the pattern and practice and disparate impact claims depend on the same body of statistical evidence. Accordingly, these claims stand or fall in concert. As we have observed at some length, the statistical evidence consists primarily of two expert reports authored by Dr. Janice Madden. However, Dr. Madden’s reports were not sufficiently tailored to compare similarly situated employees.

They failed to effectively measure job-related skills, education, experience, and job performance, using only broad, imprecise measurements as proxies for work experience. Nor did the reports sufficiently narrow the classes of employees who were compared, instead using company-wide data that did not reliably identify discrimination within designated segments of the workforce.

And, as we have observed, the reports did not distinguish between competitive and non-competitive promotions, did not capture in-grade progressive promotions and did not measure outcomes for employees who actually applied for promotions as opposed to those who did not. After reviewing this body of statistical evidence, we agree with the district court that basic analytical deficiencies in the reports render them insufficient to support a conclusion that intentional discrimination was the defendants’ standard operating procedure.

Even if we accept all the calculations underlying Madden’s reports as correct, the reports’ conclusions are of very limited value because the data on which those calculations were based was not meaningfully tailored. That is, even crediting as true the conclusions, we do not believe that a jury reasonably could infer, based on the evidence, that the plaintiffs established a pattern and practice claim.

Summary judgment was therefore properly entered for the defendants as to the plaintiffs’ discriminatory pattern and practice claims. Moreover, the district court properly entered

summary judgment for the defendants on the disparate impact claims. The plaintiffs say they identified a statistical disparity between white and black employees, along with a facially neutral employment practice — the use of subjective hiring criteria by the defendants — that has caused this disparity. First, however, a substantial question exists as to whether the plaintiffs’ statistical evidence is sufficiently tailored to establish statistical disparities within the specific segments of the defendants’ workforce in which the individual plaintiffs worked. See, e.g., Maddox, 764 F.2d at 1549-50.

But even assuming that the plaintiffs had met this requirement, they have failed to demonstrate a causal nexus between any statistical disparities in the defendants’ workforce and the practice of using partially subjective hiring criteria by some managers in some of the defendants’ facilities. See, e.g., Wards Cove Packing, 490 U.S. at 651-55, 109 S.Ct. at 2121-24. Accordingly, the district court did not err in entering

summary judgment for the defendants on the plaintiffs’ disparate impact claims. We turn, then, to the arguments mounted by each plaintiff concerning his or her disparate treatment claims. C. Cornelius Cooper Cooper asserts that the district court erred by entering final

summary judgment in favor of the defendants on his individual disparate treatment claims. Specifically, Cooper argues that the district court erred in entering

summary judgment against him as to the denial of two promotions in 1998, one to a Lighting Coordinator position and the other to a Trainer position. Cooper also argues that the district court erred in excluding an expert report which, he claims, would have established that it was highly unlikely that the selection process used to deny him his position was “race blind.” Cooper first says that he established pretext as to the denial of a Lighting Coordinator job within GPC’s Lighting Services Business Unit (“LSBU”).

Lighting Coordinators at GPC are involved in selling outdoor lighting systems and overseeing the installation and maintenance of these systems by outside contractors. Cooper was one of more than fifty applicants for two Lighting Coordinator positions posted on the defendants’ JobNet system in 1998, but was denied an interview for the position. He established a prima facie case that the denial of the promotion was discriminatory, but the district court concluded he had failed to show that the legitimate, non-discriminatory reasons for GPC’s denying him the promotion were pretextual.

We dispense first with Cooper’s suggestion that statistical evidence found in the declaration of expert John Del Rocc-ili raised a genuine issue of material fact concerning whether the interview selection process used by GPC was race blind. The district court excluded Del Roccili’s declaration after finding that the plaintiffs had not properly identified Del Roccili as an expert or provided a statement of his intended testimony, as required by Rule 26 of the Federal Rules of Civil Procedure.

The district court also held that, to the extent the plaintiffs sought to use Del Roccili as a fact witness, his declaration was inadmissible because he lacked personal knowledge of the data about which he opined. We review the district court’s decision to exclude Del Roccili’s declaration for abuse of discretion, see Benson v. Tocco, Inc., 113 F.3d 1203, 1208 (11th Cir.1997), and conclude that the district court clearly acted within its discretion here. Under Federal Rule of Civil Procedure 26, all parties must disclose expert opinion reports when directed by the court, or at least 90 days before trial.

These reports must contain “a complete statement of all opinions to be expressed and the basis and reasons therefor.” Fed.R.Civ.P. 26(a)(2)(B). Under Rule 37(c)(1) of the Federal Rules of Civil Procedure, a party “that without substantial justification fails to disclose information required by Rule 26(

a) or 26(e)(1) ... is not, unless such failure is harmless, permitted to use as evidence at a trial ... any witness or information not so disclosed.” Fed.R.Civ.P. 37(c)(1). In this case, plaintiffs originally represented that Del Roccili was a fact witness, but later identified him as an expert witness in response to defendants’ discovery requests. However, the plaintiffs never provided defendants with an expert report containing Del Roccili’s opinions or, for that matter, the bases for his opinions, as required under Rule 26.

Because the expert witness discovery rules are designed to allow both sides in a case to prepare their cases adequately and to prevent surprise, see Sherrod v. Lingle, 223 F.3d 605, 613 (7th Cir.2000), compliance with the requirements of Rule 26 is not merely aspirational. Plaintiffs have identified Del Roccili as “an expert in extracting and analyzing data from human resources databases,” R282 at 2, and in his declaration, Del Roccili extracted substantial data from the defendants’ databases, performed statistical probability calculations, and expressed opinions about the significance of these calculations.

While we do not evaluate the reliability of DelRoccili’s opinions, the techniques utilized and opinions presented plainly require the kind of technical or specialized knowledge that is the hallmark of expert evidence. See Fed.R.Evid. 702. We therefore conclude that Del Roccili’s declaration was properly construed as an expert report and, thus, because the plaintiffs failed to comply with the explicit requirements of Rule 26, it was within the sound discretion of the trial judge to sanction plaintiffs for their failure to disclose by enforcing the unambiguous terms of Rule 37(c).

See Firefighters’ Inst. for Racial Equality ex rel. Anderson v. City of St. Louis, 220 F.3d 898, 902 (8th Cir.2000). Cooper claims, however, that he has demonstrated pretext even in the absence of Del Roccili’s declaration. We remain unpersuaded. It is undisputed that LSBU Sales Manager Mike Clay was responsible for choosing for interviews applicants who had applied for the Lighting Coordinator positions. After receiving more than 50 applications for the openings, Clay reviewed a computer-generated list of applicants, including a candidate profile listing each applicant’s work history.

Clay reviewed the applicant profiles and selected candidates to interview based on the following specific job-related criteria: applicants needed at least two years of experience in power delivery; demonstrated communication, leadership and teamwork skills; experience in contract administration and contractor supe

[…]