Appendix — Phillips v. Gates (No. 09-151)

APPENDIX A

NOT RECOMMENDED FOR

FULL-TEXT PUBLICATION

File Name: 09a0319n.06

UNITED STATES COURT OF APPEALS

FOR THE SIXTH CIRCUIT

Case No. 07-3515

[Filed May 5, 2009]

WILLIE S. PHILLIPS, et. al.,

Plaintiffs-Appellants,

v.

ROBERT GATES, SECRETARY OF THE

DEPARTMENT OF DEFENSE,

Defendant-Appellee.

Nee eee eee eee eee ee”

ON APPEAL FROM

THE UNITED STATES DISTRICT COURT

FOR THE SOUTHERN DISTRICT OF OHIO

BEFORE: MARTIN, BATCHELDER, and

DAUGHTREY, Circuit Judges.

2a

ALICE M. BATCHELDER, Circuit Judge.

Appellants are African-American current and former

employees of the Defense Finance and Accounting

Service in Columbus, Ohio (“DFAS-CO”), an

independent agency within the federal Department of

Defense (“DOD”). They claim that DFAS-CO’s

promotion policies disparately impacted them on

account of their race. Following a bench trial, the

magistrate judge ruled in favor of the Secretary of

Defense (“Secretary”). For the reasons that follow, we

affirm.

DFAS-CO’s main function is to pay the DOD’s

vendors and contractors; it also issues disbursements

and provides accounting services for smaller agencies

within the DOD. DFAS-CO was created in 1991 to

consolidate the work of several] regional offices. The

DOD sought to recruit employees from these offices to

come to the new center in Columbus, but fewer than

10% of the employees agreed to transfer. A significant

percentage of the employees who did transfer were

African-American.

From 1991 to 1995 the DFAS-CO workforce

expanded from 400-500 employees to 3,500 employees.

This increase featured rapid hiring (an average of 70

new hires per month) and numerous opportunities for

promotions. Because only a small number of employees

transferred from other offices, most of the hires were

“off the street” — of people from the Columbus area.

Some African-American DFAS-CO_ employees

perceived that many of the newly-hired white

employees were being promoted at a faster rate than

they were, despite the African-American employees’

seniority. These African-American employees also

perceived that they had been subject to more

3a

disciplinary actions and had received fewer awards

than white employees, that promotions were made on

the basis of personal friendships and connections, and

that supervisors engaged in preselection — all of

which, they felt, limited African-American employees’

opportunities for promotions. g

In response to employees’ complaints about the

promotions process, the directors of DFAS-CO

organized a Process Action Team (“PAT”) to

investigate the allegations. This team, which included

DFAS-CO employees and two outside consultants, was

to address specific issues: (1) whether only

non-minorities were advancing to the GS-11 pay grade

and above; (2) whether preselection of jobs existed;

(3) whether management selection patterns were

discriminatory; and (4) whether personal relationships

were impacting selections and promotions. In its

investigation, the team interviewed 25% of the

DFAS-CO workforce and analyzed promotions and

Equal Employment Opportunity (“EEO”) statistical

data for the years 1995 and 1996. The interviewed

employees included both randomly selected workers

and volunteers.

In March 1997, the team released a report (“the

PAT report”) finding that: (1) minorities were

promoted at lower rates than non-minorities,

especially at higher grade levels; (2) severai groups,

especially African-American men, were

under-represented at higher grade levels;

(3) evaluation of application content was subjective;

(4) there existed a_ strong perception that

discrimination, whether race-based or not, existed at

DFAS-CO; (5) management may have committed a

prohibited personnel practice in the case of a

4a

particular white employee; (6) white women received

72% of the exceptional performance ratings in 1995;

(7) supervisors rated employees within their own race

or national origin more highly than employees from

other groups; and (8) selecting officials who were white

men showed strong, ethnic preferences in awarding

promotions.

Following the release of the PAT report, Appellant

Willie Phillips filed a discrimination complaint with

DFAS-CO’s EEO office. After an investigation, the

EEO office issued its own report in’ which it found,

among other things, that African-American employees

received about 22% of all promotions in 1995-1996,

which reflected the approximate percentage ‘of

African-American employees in the DFAS-CO

workforce during that period. An administrative judge

subsequently dismissed the complaint.

The Appellants then filed suit, alleging that the

Secretary's practices and procedures regarding

employee promotions had a disparate impact on

African-American employees in violation of Title VII.

The district court granted summary judgment in favor

of the Secretary on the ground that Appellants had

failed to file a timely administrative complaint. We

reversed, finding that equitable tolling applied.

Phillips v. Cohen, 3 F. App’x 212 (6th Cir. 2001)

(“Phillips I’).

On remand, the district court referred the case toa

magistrate judge, who entered summary judgment for

the Secretary — this time on the grounds that

Appellants had not produced any evidence of a

disparate impact or traced the alleged impact to any of

the Secretary’s policies. We again held in favor of

5a

Appellants, finding that there remained for trial a

material issue of fact regarding the existence of a

disparate impact. Phillips v. Cohen, 400 F.3d 388 (6th

Cir. 2005) (“Phillips II”). We also held that the

magistrate judge had failed both to assess the quality

of evidence that had been lost by Appellee, or, in

violation of a court order, routinely destroyed pursuant

to the agency’s internal regulations, and to determine

appropriate sanctions.

On further remand, the magistrate judge conducted

a bench trial and found that Appellants had failed to

prove a disparate impact by a preponderance of the

evidence. The magistrate judge also evaluated the

quality of the lost evidence and determined that the

loss did not negatively impact Appellants’ case; the

judge did, however, award Appellants attorney’s fees

and costs incurred in connection with their motion for

sanctions. This appeal followed.

Il.

Appellants raise several arguments on appeal.

First, they contend that the trial court erred in finding

that they had failed to prove a disparate impact by a

preponderance of the evidence. Second, they argue that

the trial court did not sanction the Secretary harshly

enough for failing to preserve evidence. Finally, they

maintain that the court erroneously dismissed

Plaintiff-Appellant Todd Brooks.

A.

“This Court’s standard of review in a Title VII

discrimination case is ‘narrow.” Dunlap v. TVA, 519

F.3d 626, 629 (6th Cir. 2008) (quoting /sabel v. City of

6a

Memphis, 404 F.3d 404, 411 (6th Cir. 2005)). In a

disparate impact case, a trial court’s findings of fact

should stand unless clearly erroneous. Jd. “The issue

is not whether the [trial] court reached the best

conclusion, but whether the evidence before the court

supported the [trial] court’s findings.” Jd. (citation

omitted). “Also, the [trial] court’s findings based on the

credibility of the witnesses before it are entitled to

great deference on appeal.” Jd. (citation omitted).

To succeed with a disparate-impact claim, a

plaintiff must first “establish a prima facie case of

discrimination — i.e., the plaintiff must establish that

an adverse impact has occurred.” Jd. “If he succeeds,

the employer must show that the protocol in question

has ‘a manifest relationship to the employment’ — the

so-called ‘business necessity’ justification.” /d. (quoting

Griggs v. Duke Power, 401 U.S. 424, 431 (1971)). “The

plaintiff must then show that other tests or selection

protocols would serve the employer’s interest without

creating the undesirable discriminatory effect.” Id.

(citing Albemarle Paper Co. v. Moody, 422 U.S. 405,

425, 432 (1975)).

To establish a prima facie case, the plaintiff must

challenge a specific employment practice and prove,

through relevant statistical analysis, that the

challenged practice has an adverse impact on a

protected group. Id. (citing Johnson v. U.S. Dep’t of

Health and Human Servs., 30 F.3d 45, 48 (6th Cir.

1994)). The prima facie case is sometimes said to

comprise three elements — identification, disparate

impact, and causation. See, e.g., EEOC v. Steamship

Clerks Union, Local 1066, 48 F.3d 594, 601 (1st Cir.

1995). If, however, “the employee challenges the

employer’s promotion process as a whole ... then the

Ta

disparate impact and causation elements merge.”

Phillips I], 400 F.3d at 397-98 n.8. Accepting arguendo

Appellants’ argument that the DFAS-CO promotion

practices are incapable of separation,’ we consider the

disparate-impact and causation elements to address

the same question: “whether the evidence in the record

supported a finding that African-American employees

were promoted at a lower rate than white employees.”

Id.

“In cases involving promotion policies, the relevant

inquiry is comparing the number of protected group

members benefitting from promotions with the number

seeking them; this figure is then contrasted with the

corresponding ratio for the non-protected group.” Jd. at

399 (citing Connecticut v. Teal, 457 U.S. 440, 448

(1982)). “[S]tatistics based on an applicant pool

containing individuals lacking minimal qualifications

for the job would be of little probative value,” however.

Watson v. Ft. Worth Bank & Trust, 487 U.S. 977, 997

(1988) (citations omitted). And while “sufficiently

substantial” statistical disparities raise an inference of

disparate impact, the relevant analysis should not be

“framed in terms of any rigid mathematical formula.”

Id. at 994-95.

Ideally, to determine if African-American

employees were promoted at a lower rate than their

"In the second appeal in this litigation, we noted that Appellee did

not challenge the magistrate judge’s finding that DFAS-CO’s

promotion process “should be considered as a whole” or

Appellant’s argument that the process was “incapable of

separation.” Phillips 11, 400 F.3d at 397-98. The Secretary does

not argue before us that the promotion process should be analyzed

on a piecemeal basis.

8a

white counterparts, one would compare the respective

ratios of promoted African-American and _ white

employees to the number of qualified

African-American and white applicants. The parties

could not make this comparison here, however,

because under the DFAS-CO document retention

policy, applicant flow data was destroyed every two

years, on a rolling basis. By the time Appellants filed

their complaint in October 1998, only data for 1997

and part of 1996 existed.

In an attempt to reconstruct the relevant ratios,

Appellants and the Secretary presented statistical

experts who employed different methods. Appellants’

expert, Dr. Ramona Paetzold, compared the racial

composition of all promoted empleyees to the racial

composition of the entire DFAS-CO workforce. This

method assumed that every DFAS-CO employee

applied for every job opening, irrespective of whether

the employee would, in fact, be qualified, unqualified,

or overqualified for the position. Dr. Paetzold thus

based her statistical analysis on an applicant pool that

necessarily included “individuals lacking minimal

qualifications.” See Watson, 487 U.S. at 997. Dr.

Paetzold determined that African-American employees

were promoted at a significantly lower rate than were

white employees.

The Secretary's expert, Dr. John Claudy, created

“constructed pools” of applicants for each promotion.

Dr. Claudy examined employee grades in pairs of

successive years to determine when promotions had

occurred. For example, if an employee was listed as a

grade GS-7 in 1995 but as a grade GS-8 in 1996, then

Dr. Claudy assumed that a promotion had taken place.

For each promotion, Dr. Claudy constructed a pool of

9a

employees in the same (pre-promotion) grade level and

job series as the successful applicant. This method had

the advantage of using an applicant pool that was

more likely to approximate the group of employees who

were qualified for and actually applied for a given

promotion. It was nonetheless flawed in that it failed

to account for the fact that sometimes employees could

qualify for and would apply for promotions outside of

their job series. Dr. Claudy determined that: in some

grade levels African-Americans were promoted at a

rate higher than would be expected, in other grade

levels African-Americans were promoted at a rate

lower than would be expected; and in only one grade

level (Grade 5) were African-Americans promoted at a

rate lower than two standard deviations from what

would be expected.

The magistrate judge noted the flaws with both

approaches, found that the experts’ conclusions stood

in equipoise, and determined that Appellants had

failed to show a disparate impact by a preponderance

of the evidence. The magistrate judge further found

that the PAT report did not prove a disparity in

promotion rates between African-Americans and

whites because the report compared the percentage of

African-American employees receiving promotions at

DFAS-CO to the percentage of African-American

employees in the Columbus labor force. The PAT

report did not account for minimal qualifications for

the promotions, nor did it purport to analyze the

alleged disparity in terms of standard deviations. As

for the EEO report, the magistrate judge determined

that its findings were skewed because its survey

sample was not entirely random but included a

disproportionate number of African-American

respondents. Finally, the magistrate judge found that

10a

Appellants’ testimony did not establish a disparate

impact because the individual witnesses testified only

that they were denied promotions for which they

applied and for which they believed themselves to be

qualified. They did not establish that they were denied

the promotions on account of their race, and their

anecdotal evidence did not prove a _ significant

difference in the promotion rates between whites and

African-Americans at DFAS-CO.

The magistrate judge’s findings are not clearly

erroneous. Although Appellants argue that “(t]he trial

court erred in finding Dr. Claudy’s expert opinion as

credible and trustworthy as Dr. Paetzold’s,” it might be

better said that the magistrate judge found both expert

opinions equally problematic. He therefore accepted

neither expert’s analysis, and concluded that

Appellants had not carried their burden. Although the

PAT report, the EEO report, and Appellants’ testimony

provided some evidence of a perceived race-based

disparity at DFAS-CO, the evidence was not sufficient

to require a finding that Appellants had demonstrated

the type of statistical dispaiity necessary to establish

a Title VII prima face case.

B.

“We review a district court’s imposition of sanctions

for abuse of discretion.” Fieger v. Cox, 524 F.3d 770,

774 (6th Cir. 2008) (quoting Frank v. D’Ambrosi, 4

F.3d 1378, 1387 (6th Cir. 1993) (citing in turn Cooter

& Gell v. Hartmarx Corp., 496 U.S. 384, 399 (1990))).

On December 18, 1998, the district court issued a

retention order that required DFAS-CO to preserve

existing records related to this litigation. By that time,

lla

however, DFAS-CO already had destroyed almost all

of the applicant flow data for the years 1996 and

earlier because its document retention policy called for

such materials to be destroyed on a two-year, rolling

basis. Thomas Gary, head of Togar and Associates, the

outside consultant that assisted DFAS-CO with the

PAT report, testified that he took with him two boxes

of materials used in the creation of the PAT report.

These materials would have included only data for the

years 1995-1996 and likely did not include all

applicant flow data even from that period. Gary

testified «hat in late 1997 or early 1998 he sent the

boxes to DFAS-CO headquarters in Arlington,

Virginia. The magistrate judge held that if DFAS-CO

had taken steps to look for those m-terials

immediately after the district court’s order, DFAS-CO

might have been able to locate the PAT documents.

The magistrate judge granted Appellants attorney’s

fees and costs incurred in pursuing their motion for

sanctions.

Appellants claim that this sanction was not enough

and that the magistrate judge should have “lowered

the bar” on differences in the promotion rate or

allowed an inference of adverse impact pursuant to 29

C.F.R. § 1607.4(D). That regulation provides in part:

Where the user has not maintained data on

adverse impact as _ required by _ the

documentation section of applicable guidelines,

the Federal enforcement agencies may draw an

inference of adverse impact of the selection

process from the failure of the user to maintain

such data, if the user has an underutilization of

a group in the job category, as compared to the

group’s representation in the relevant labor

12a

market or, in the case of jobs filled from within,

the applicable work force.

Nothing in this regulation requires federal judges to

impose any particular sanction on an entity that has

failed to preserve records that may bear on adverse

employment impacts. It permits, but does not require,

federal enforcement agencies to draw inferences. It

does not abrogate the discretion trial judges have in

imposing appropriate sanctions. Moreover, both

statistical experts at trial testified that the data from

the PAT investigation would not have substantially

altered their analyses because it would only have

included a sample of information from a two-year

period and would not have helped to establish

promotion rates for the total period in question. We

find no abuse of discretion here.

C.

We review for abuse of discretion a trial court’s

decision to dismiss a party. Sutherland v. Michigan

Dep't of Treasury, 344 F.3d 603, 612 (6th Cir. 2003);

see FED. R. CIV. PRO. 21 (“On motion or on its own, the

court may at any time, on just terms, add or drop a

party.”). We therefore “affirm the dismissal of a party

for misjoinder unless this court is left with a definite

and firm conviction that the trial court committed a

clear error of judgment.” Sutherland, 344 F.3d at 612

(citation and internal quotation marks omitted).

The magistrate judge dismissed Plaintiff-Appellant

Todd Brooks because he was unable to be physically

present at trial. Appellants argue that Brooks’s

individual testimony was unnecessary because it

would have been “cumulative and redundant”: he, like

13a

all the other plaintiffs, would have testified that he

was denied promotions for which he was qualified and

that he believed those denials were racially-based. His

testimony, Appellants insist, would have been required

only if the trial had proceeded to the damages phase.

This argument is without merit. Although Brooks

most likely would have testified to being denied

promotions in much the same way as the other

plaintiffs allegedly were, his testimony would not for

that reason have been “cumulative” or “redundant.” As

a plaintiff in this action, Brooks was required to prove

that DFAS-CO’s allegedly discriminatory promotions

policy injured him in some way. Whether Brooks was

denied promotions for which he was qualified was an

issue of fact for determination in the trial’s lability

phase.

Appellants further argue that instead of dismissing

Brooks, the trial court should have admitted his

affidavit into evidence. Quoting McIntyre v. Reynolds

Metals Co., 468 F.2d 1092, 1094 (5th Cir. 1972),

Appellants maintain that the affidavit would not have

been inadmissible hearsay because “[w]hen a witness

... affirms the truth of a prior statement, the earlier

statement is to be considered not only as bearing on

the credibility of the witness but also as affirmative

evidence.” McIntyre, however, addressed a situation

where a deposition witness explicitly affirmed a prior

statement, and opposing counsel had the opportunity

at the deposition to cross-examine the witness

regarding the statement. Here, the Secretary never

had the opportunity to cross-examine Brooks about the

statements in his affidavit. Appellants argue that the

Secretary could have deposed Brooks if he had wished,

but this argument misses the mark — a defendant

l4a

does not waive a hearsay objection by waiting to

confront a plaintiff at trial. The magistrate judge did

not abuse his discretion in dismissing Brooks.

CONCLUSION

Accordingly, this Court AFFIRMS the judgment of

the district court.

15a

APPENDIX B

IN THE UNITED STATES DISTRICT COURT

FOR THE SOUTHERN DISTRICT OF OHIO

EASTERN DIVISION

Civil Action 2:98-cv-799

[Filed January 5, 2007]

Willie S. Phillips, et al.,

Plaintiffs

v.

Robert Gates, Secretary,

Department of Defense,

Defendant

le ee ee ee

Magistrate Judge Abe!

Opinion and Order

Plaintiffs bring this employment discrimination

action under Title VII of the Civil Rights Act of 1964,

42 U.S.C. §2000e-16. Plaintiffs are employees of the

Defense Finance and Accounting Service, Columbus

Center (““DFAS-CO”), an entity within the Department

of Defense under Secretary Robert Gates, the

defendant. Plaintiffs allege that between 1990 and

16a

1996 DFAS-CO discriminated against African-

American employees by using policies and practices

that had a disparate impact on their ability to obtain

promotions. This case was tried to the Court beginning

January 23, 2006.

I. Allegations in the Amended Complaint

Plaintiffs Lerdon Woodfolk, Jean Washington,

Marcia Spence, Lisa Prater, Mary Phillips, Linda R.

Pettes, Carilsa Miles, Susan McNeal, Jackie McGrady,

Kevin Gray, Denise Goudy, Richard W. Goudy, Emma

Fields, Yulander Edwards, Lori Dillard, Todd Brooks,

Curtissene Anderson and Willie S. Phillips filed an

amended complaint on October 15, 2002 (doc. 37).'

The Defense Finance Accounting Service was

created in 1991 as an independent agency within the

United States Department of Defense (“DOD”). DFAS’s

headquarters are located in Arlingten, Virginia. There

are five main centers, one of which is located in

Columbus, Ohio. DFAS pays Department of Defense

contractors and vendors, provides accounting services,

and disburses morey for smaller DOD agencies. Am.

Compl. 4¥ 12-13.

’ Jean Washington, Curtissene Anderson, Linda R. Pettes, and

Richard W. Goudy were dismissed because they lacked standing.

See Phillips v. Cohen, 400 F.3d 388, 397 (6 Cir. 2005). Jean

Washington was reinstated as a party plaintiff. See doc. 103. At

trial, the Court granted defendant’s motion for a directed verdict

at trial with respect to plaintiff Todd Brooks. Mr. Brooks did not

attend the trial, and no evidence was presented with respect to his

claim. See 1V-96:16-102:8. Jean Washington, Emma Fields, and

Lori Dillard also failed to appear for trial.

17a

When DFAS-CO was created in 1991, there was not

a workforce already established in Columbus. Jd. at

J 14. DFAS-CO began with a workforce of 400-500

employees, many of whom were transferred federal

employees from Atlanta, Boston, Los Angeles and

elsewhere. Jd. at 17 16 & 19. Plaintiffs state that this

initial workforce of transferred employees was racially

diverse. Id. at { 19. The workforce reached 3,500

employees in 1995. Id. at J 16. Between 1991 and

1995, DFAS-CO hired approximately 3,000 new

employees and offered many _ promotional

opportunities. DFAS-CO hired an average of 70

employees per month between 1991 and 1995. Id. at

q 20.

African-American employees of DFAS-CO believed

that they were not getting the same promotional

opportunities as white employees. Id. at J 23.

Plaintiffs believed that white employees who were

hired with no prior experience were promoted more

rapidly and progressed at a faster rate than African

American employees. Plaintiffs believe that African

American employees trained white employees who

then became their supervisors. Plaintiffs also maintain

that there was favoritism and nepotism in the hiring

and promotion process. Id. They allege that there were

abuses of the detail process. Detailing is a policy that

permits a manager to fill a position without posting a

position for competitive bidding. Plaintiffs allege that

once a position was filled by detailing, that person was

in a more favorable position to get the ultimate

promotion. Jd. at J 35. Asa result of these employment

practices, the complaint alleges, “white employees at

the GS-6 through GS-12 Grades’ were

disproportionately chosen over qualified

African-Americans for promotions.” Jd. at 32.

18a

Further, “African-Americans are under-represented at

the GS-13 through GS-15 Grades due to unlawful

discrimination.” Jd. at 731.

On December 6, 1996, Charles Coffee, Director of

DFAS-CO, set up a Process Action Team (“PAT”) to

review and evaluate DFAS-CO’s merit promotion

policy and procedures. Jd. at J 26. Among the matters

it investigated were whether only non-minorities were

advancing to GS-11 level and above and whether there

were any patterns of discrimination in management

selection. On March 3, 1997, PAT issued its report. Jd.

at 7 50.

On March 5, 1998, each plaintiff contacted

DFAS-CO Equal Employment Opportunity officials

with an informal complaint. After receiving informal

counseling, plaintiffs filed a formal individual and

group complaint with the EEOC. Jd. at J 51. On

August 7, 1998, plaintiffs filed this action. Plaintiffs

maintain that DFAS-CO’s discriminatory employment

policies and practices have routinely resulted in the

selection of whites over qualified African Americans,

and these specific employment practices have had and

continue to have a disparate impact on African

American employees. Id. at | 52-53.

Il. Arguments of the Parties

A. Plaintiffs

To show that defendant’s promotion practices had

a disparate impact on African Americans, plaintiffs

rely on the conclusions of the PAT report, statements

made by Ida Faye Groves in a presentation to DFAS

headquarters, the EEO report, and a statistical report

19a

prepared by their expert, Dr. Ramona Paetzold.

Plaintiffs argue that Dr. Paetzold presented statistical

proof that African Americans were promoted at a

statistically significantly lower rate when compared to

white employees. The defendant’s approach, on the

other hand, was less reliable and criticized by the

Sixth Circuit.

Plaintiffs contend that this Court should determine

whether defendant violated Title VII based on

evidence showing that African American employees

were promoted less frequently than white employees;

that factors in the promotions process, including

awards and discipline, disadvantaged African

Americans because of racial differences in application;

and that employees perceived favoritism and a lack of

transparency as pervading the promotion process.

Plaintiffs ask the Court to admit the testimony of

Delana Taylor to rebut defendants’ assertions that

PAT documents were inadvertently lost and that the

applicant flow data was destroyed pursuant to

DFAS-CO’s customary document destruction policy.

Taylor testified that she read a letter indicating that

the PAT report placed DFAS-CO in a “bad light” and

should be “sanitized.” Plaintiffs maintain that Taylor's

testimony is admissible pursuant to Rules 803(1) & (2),

and 807 of the Federal Rules of Evidence.

Finally, plaintiffs argue that the Court should

sanction defendant for destroying evidence by

awarding plaintiffs their expert and attorney fees, or

in the alternative, finding liability on behalf of

defendant. Plaintiffs maintain that it was clear that

they were prejudiced as a result of the destruction of

the documents. The prohibition against destroying

20a

relevant evidence once litigation has begun is clear.

Plaintiffs also rely on sections 1607.1 and 1607.14 of

the Code of Federal Regulations to show that

defendants were obligated to preserve promotional

records.

B. Defendant

Defendants argue that plaintiffs have failed to

establish all of the elements of disparate impact by a

preponderance of the evidence. Defendants maintain

that plaintiffs have not demonstrated that the

elements of the promotion decisionmaking process at

DFAS-CO are incapable of separation for analysis, nor

have they established that specific DFAS-CO

employment practices had a disparate impact on

whether African Americans were promoted as required

by Watson v. Fort Worth Bank & Trust, 487 U.S. 877,

994 (1988).

Defendants also argue that plaintiffs have failed to

establish the existence of a disparate impact as a

result of the facially neutral employment practices.

Defendants contend that the evidence in the record,

both statistical and non-statistical evidence, fails to

establish any disparate impact on African Americans

at DFAS-CO from 1992 to 1996.

Defendant maintains that the statistical analysis

performed by Dr. Ramona Paetzold, plaintiff's expert,

was flawed, and that defendant’s expert provided a

more credible analysis. Defendant argues that the

single pool approach employed by Dr. Paetzold failed

to account for the importance of grade and series.

Under her analysis, all employees at DFAS-CO were

equally likely to apply for all positions. Defendant

2la

maintains that DFAS-CO employees simply were not

all equally likely to apply for all positions, and Dr.

Claudy’s approach better approximated who comprised

the actual applicant pool.

Defendant maintains that Dr. Claudy made the

functional equivalent of the type of analysis the Sixth

Circuit stated was required in a disparate impact

statistical analysis, and plaintiffs expert did not.

Instead, Dr. Paetzold said that the prescribed analysis

was impossible under the circumstances and compared

recipients of promotions to the total number of

employees at DFAS-CO. Defendant argues that

promotional policies cannot have a disparate impact on

employees who have never applied for a promotion, so

the appropriate analysis compares the ratios for

African Americans and non-African Americans of the

number of each promoted to the number of each

seeking them. Finally, defendant maintains that the

constructed pool approach used by Dr. Claudy does not

warrant suspicion under Mozee v. Am. Commercial

Marine Serv. Co., 940 F. 2d 1036, 1045 (7™ Cir. 1991).

Defendant also argues that the non-statistical

evidence in the record, including the PAT report and

anecdotal evidence offered by plaintiffs, does not

provide credible evidence of disparate impact.

According to defendant, the PAT team failed to employ

any commonly accepted analytical procedures such as

surveying a truly random sample of employees;

examining applicant flow data from randomly selected

promotions; and conducting statistical analysis of the

available data in terms of standard deviation or

employing a statistician to do so. Consequently,

defendant maintains that the accuracy and reliability

of the PAT report’s results are questionable at best.

22a

Defendant further argues that the anecdotal evidence

of plaintiffs’ individual experiences applying for

promotions is irrelevant to a disparate impact

analysis.

With respect to sanctions, defendant maintains that

the evidence presented at trial clearly established that

the evidence in existence at the time of the Court’s

preservation order would have been useless for a

disparate impact analysis. Consequently, defendant

argues that sanctions are not warranted.

lil. Plaintiffs’ Motion for Sanctions

This suit was filed August 7, 1998. Plaintiffs served

interrogatories and requests for production of

documents on September 2, 1998. Defendant did not

timely respond to them. On December 7, 1998,

plaintiffs filed a request for status conference

regarding defendant’s failure to respond to the

outstanding discovery requests (doc. 6). On December

15, 1998, defendant filed a motion for protective order

staying discovery until the Court had ruled on its

November 2, 1998 motion to dismiss or, in the

alternative, for summary judgment (doc. 7). On

December 18, 1998, I issued a Discovery and

Scheduling Conference Order that granted defendant’s

motion for protective order “on the condition that

defendant’s counse! locate and gather all documents

responsive to the discovery requests and produce

documents responsive to it within eight (8) business

days of the Court issuing any decision denying

the pending motion to dismiss/for summary

judgment” (doc. 8).

On April 24, 20038, plaintiff filed a motion for

sanctions because defendants failed to preserve the

23a

following evidence:

(1)

(2)

(3)

(4)

(5)

(6)

The Job Opportunity Announcements of

the Columbus Center for Fiscal Years

1995 & 1996; reviewed by the PAT

members listed in the PAT Report

bibliography;

The Certificate of Referrals of the

Columbus Center for Fiscal Years 1995 &

1996 reviewed by the PAT members

listed in the PAT Report bibliography;

Any of the “Other data, such as

discipline, awards, referrals, selections,

and performance ratings, that required

manual compilation and stratification”

mentioned in the PAT Report;

The “data readily available through

current databases, such as DBMS”

mentioned as reviewed by the PAT

members;

Any of the four hundred and twenty

seven (427) completed survey forms of the

DFAS Columbus employees;

Any documents regarding the one

hundred and eighteen (118) individual

interviews with employees;

24a

(7) Any of the documents provided by the

employees that were received by PAT

members Togar and Associates;

(8) The spread sheets created by Togar and

Associates for the PAT Report; and

(9) DFAS Columbus merit promotion records

prior to October 1998 which were

apparently destroyed after two years

pursuant to “DFAS’ internal Regulation

1402.1.”

In their April 24, 2003 motion, plaintiffs argued that

defendant has failed to respond to their September 2,

1998 discovery request and to comply with the Court’s

December 18, 1998 Order. Plaintiffs maintained that

defendant had lost or destroyed documents that were

responsive to the discovery requests. Plaintiffs

requested that the Court enter judgment in their favor

or, in the alternative, strike the report of defendant’s

expert and admit the report of plaintiffs’ expert as a

true statement of the statistical evidence in this

action.

On July 2, 2003, I granted plaintiffs’ April 24, 2003

motion for sanctions but deferred imposing sanctions

until trial because I could not determine to what

degree the loss of the documents would prejudice

plaintiffs. See doc. 78. On August 3, 2003, defendant’s

motion for summary judgment was granted, and as a

result, I did not revisit what sanctions were

appropriate in light of defendant’s failure to comply

with the December 18, 1998 Discovery and Scheduling

Conference Order. The Sixth Circuit, however, found

that I erred by not making an assessment of the

25a

quality of the evidence lost and determining the

appropriate sanctions te be imposed on defendant for

failing to comply with tire Court’s Order.

Trial Testimony. Items 1 through 8 on p.aintiffs’

list of allegedly lost or destroyed documents relate to

materials gathered by the consulting firm Togar and

Associates (“Togar”) in the course of the preparation of

the January 24, 1997 PAT report that was reieased

March 3, 1997.* Plaintiff initially alleged that

defendant lost these materials, although plaintiffs now

assert that these documents were purposefully

gathered in an effort to “sanitize” the PAT report.

These documents were at some point in the hands of

Togar and Associates. Thomas Gary, who founded and

is employed by Togar and Associates, testified at trial

that he had the survey sheets placed in two corrugated

boxes and shipped to his home. Tr. Vol. I at 168:13-17.

The employees who responded to the survey were

promised that their response would be kept

confidential. Mr. Gary believed that if the surveys

were left at DFAS-CO, the employees might be

subjected to retaliation against for the comments they

made in their written responses to the questionnaire.

Tr. Vol. I at 168:19-25 and 169:1-17. He kept the

decuments at his home for some unknown period of

time until he received a letter from Mr. Coffee

instructing him to forward the documents to DFAS

headquarters in Virginia. Tr. Vol. I at 1701-7. Mr.

Gary testified that he sent the documents as requested

? On March 3, 1997, PAT released a 121-page report. It indicated

that considerable data was examined during the investigation.

Survey forms were completed by 427 DFAS-CO employees. PAT

also interviewed 118 employees. About 30 employees provided

documents to PAT.

26a

and could not explain why the documents were never

located. Tr. Vol I at 170:7-18.

Both experts indicated that even if the survey

documents were accessible, these documents would not

assist them in determining whether or not the

promotion process had a disparate impact on

African-Americans at DFAS. They would have

preferred to have applicant flow data, the Job

Opportunity Announcement (which includes job

qualifications), all applicants, those eligible for the

position, those qualified for it, the certificate of referral

(the top 15 applicants plus ties), and the applicant

selected. Tr. Vol. IV at 113:17—114:15.

Dr. Claudy testified that the PAT report did not

provide information that indicated whether or not

management selection patterns were discriminatory:

Q. Is there any information on these pages, that

is 37 through 38, covering the issue of, are

management selections patterns

discriminatory, to suggest that the polices

and/or practices of DFAS have an adverse

impact on African Americans?

A. There is really no information here that

allows you to makes those inferences about

adverse impact. There is no statistical

information here, there is no inferential

statistics here.

Q. Can the conclusions of the PAT report be

compared with your own? Are we talking

about the sare thing?

27a

A. My report focused on answering a fairly

specific question, whether during the time

period from 1992 through ’96, fewer African

Americans were selected for promotion than

would have been suggested by chance, given

there represeiitation in the work force. The

PAT report looks at all sorts of different

allegations that were made, but it does it

using purely descriptive statistics. It looks at

only 94 — or ’95 and ’96, I’m sorry, rather

than the entire time period, and it uses a

single pool rather than breaking them out.

The result — this really doesn’t — this

report was not designed to answer the

allegations in the subject that we are here

htigating today.

Tr. Vol V at 60:7-61:9.

Dr. Claudy further testified that even if he had

access to the approximately 427 completed surveys

from interviews with DFAS employees in 1996 about

their com plaints or concerns, these documents would

not have been relevant to a statistical analysis of

promotions. Tr. Vol. V at 62:1-5. He also would not

have uscd data on discipline, awards, and performance

ratings of DFAS employees had it been available. Tr.

Vol. V at 61:22-25.

The PAT team also viewed documents relating to

approximately thirty specific promotions as a result of

individual complaints about the selection process. Dr.

Claudy testified that these thirty Job Opportunity

28a

Announcements (“JOAs”)® would not have been useful

in determining whether there was disparate impact in

the promotion practices for African Americans from

1992 to 1996:

Q: How about approximately 30 files from 30

promotion selections, taking place between

1995 and 1996 and pulled for the principal

reason that the non-selectee had complained.

These selected files would include all of the

applicants, all of the eligibles, all the

referred and the selectee for just those 30

promotion selections?

A. If those same files had been available for the

other 3000-plus vacancies, yes. But under

the circumstances, those 30 by themselves,

they represent a 1 percent sample that was

not selected randomly but were only selected

from two of the years, no. They would not

have provided any useful information for an

applicant flow analysis of the process from

"92 to 96.

Tr. Vol V at 62:10-22. Dr. Claudy also testified that the

certificates of referrals would not have improved his

analysis. With respect to the applicant flow for the last

four or five months of 1996, Dr. Cleudv testified:

A. Well, again, if we had all of those folders for

all of the years, I would want them. But

what was going on in the last four or five

* The terms JOAs and applicant flow data were used

interchangeably throughout trial. See Tr. Vol V at 62:13-15.

29a

months of a 92 to 96, that’s a five-year time

period inclusive, four or five months does not

represent an accurate picture of what was

going on during that entire time period. So,

no.

Q. Would it even give you an accurate

perception of what was going on in 1996?

A. No. It would only - - it might give you some

perception of what was going on in those four

months or five months but not for the year as

a whole, no, and certainly not for the whole

time period.

‘l'r. Vol V at 63:24-25-64:1-10.

Dr. Paetzold agreed that data on discipline, awards,

and performance ratings for DFAS employees for 1995

and 1996 by themselves would not have been helpful.

Tr. Vol. VI at 41:9-13. She also stated that the

approximately 427 completed survey forms or the

notes taken when they were filled out would not have

been useful to her analysis. Tr. Vol. VI at 41:14-20. Dr.

Paetzold acknowledged that the thirty JOA files

viewed by the PAT ieam would not have provided

reliable conclusions with respect to whether there was

a disparate impact in 1995 or 1996.

Q. How about approximately 30 files from 30

prumotion selections taking place between

1995 and 1996 and pulled for the principal

reason that the _ non-selectee' had

complained? These selected files would

include all of the eligibles, all of the referred

and the selectee for 30 promotion selections?

30a

A. So, it had the actual outcome of who was

selected for the promotions?

Q. For 30 files based on complaints from

non-selectees for, you know, randomly for

two years, 95 and ’96?

A. And the data for those who weren’t selected

would have been available, also?

Q. Yes, ma’am.

A. I could have isolated those and analyzed

those separately, yes.

@. Could you have to come to a reliable result

for 1995 with that data?

A. No.

Q. How about 1996? Im sorry, I missed your

answer.

A. No. lam assuming that that’s not a complete

set of data. But that would have given me

some very specific separate information, but

I don’t know that I could have made a full

determination for 1995 or 1996.

Tr. Vol. VI at 41:21-42:18.

Derek Saunders and Carol Scott testified that the

applicant flow data was routinely destroyed two years

following the close of the application period for the

position in question. Both Saunders and Scott also

testified that the PAT did not review or access the

3la

complete applicant flow data for the years 1995 and

1996, even though the bibliography of the PAT report

references “Job Opportunity Announcements of the

Columbus Center for Fiscal Years 1995 & 1996.” Pls.

Ex. 4

Plaintiffs offered the testimony of Delana Taylor to

rebut defendant’s evidence that the underlying PAT

data was destroyed through routine DFAS data

retention and destruction practices. Ms. Taylor

testified as follows. As Director Tom O’Brien’s

secretary, she handled all of the incoming mail to the

director’s office. When opening the mail, she came

across a letter from DFAS headquarters which

aiscussed the PAT report. The letter indicated that the

PAT report had put DFAS in a bad light. March 17,

2006 transcript at 10. The writer indicated that the

PAT report should be “sanitized” and the records

should be collected. Jd. at 11. Ms. Taylor could not

recollect who wrote the letter or to whom it was

addressed, although she surmised that it was directed

to either Mr. Coffee or Ida Fay Groves.

Thomas O’Brien testified that he never received the

letter described by Ms. Taylor. He said that he would

have remembered such a letter because it would have

been “totally inappropriate.” Jd. at 24:21-24.

Discussion. Defendants argue that the Court

should not admit the testimony of Delana Taylor.

Plaintiffs offered her testimony to rebut defendant’s

evidence that the underlying PAT data was destreyed

according to regular agency regulations, but was,

instead, destroyed in an effort to minimize the impact

of the PAT report and sanitize its contents. Defendant

argues that the issue of whether the PAT data was

32a

destroyed pursuant to agency regulations or an alleged

directive to gather and destroy the documents has

been contested since 1998. Ms. Taylor informed

plaintiff Miles of the alleged letter no later than 2000.

Ms. Taylor left DFAS-(O in 2000, and therefore could

not have any fear of retaliation in the event that she

testified. Plaintiffs, however, failed to disclose Ms.

Taylor or the existence of the alleged letter until the

week of trial. Defendant maintains that plaintiffs were

obligated to disclose Ms. Taylor as a potential witness.

Defendant also argues that Ms. 'l'aylor’s testimony

constitutes inadmissible hearsay. Plaintiffs maintains

that Ms. Taylor’s testimony is admissible pursuant to

Rule 803(1), (2), and 807 of the Federal Rules of

Evidence. Plaintiffs argues that Ms. Taylor specifically

recalled the use of the word “sanitized,” which falls

within the hearsay exception for a present sense

impression under Rule 803(1). Defendant maintains

that Rule 803(1) is not applicable because Ms. Taylor

is not the declarant, and the author of the alleged

letter was not describing or explaining an event or

condition while the author was perceiving the event or

condition. Consequently, Ms. Taylor’s impression upon

reading the letter is irrelevant to the hearsay analysis.

Plaintiffs also argue that Ms. Taylor’s testimony is

admissible as an excited utterance under Rule 803(2).

As noted above, Ms. Taylor is not the declarant.

Instead, she is testifying about someone else’s

statement. Rule 803(3) would be implicated if the

author had written the letter shortly after a startling

event.

Finally, plaintiffs argue that Ms. Taylor’s testimony

is admissible under Rule 807 because there are

33a

“equivalent circumstantial guarantees of

trustworthiness.” Fed. R. Evid. 807. Rule 807 states:

[A] statement may not be admitted under this

exception unless the proponent of it makes

known to the adverse party sufficiently in

advance of trial or hearing to provide the

adverse party with a fair opportunity to prepare

to meet it, the proponent’s intention to offer the

statement and the particulars of it, including

the name and address of the declarant.

Having considered the parties’ arguments, I

conclude that Ms. Taylor’s testimony was not rebuttal

evidence. Plaintiffs had the duty to disclose her as a

potential witness and failed to do so. Alternatively, Ms.

Taylor’s testimony is not admissible based on any

exception to the hearsay rules argued by plaintiffs.‘

Her testimony lacked corroborating detail. Ms. Taylor

knew neither the sender or the recipient of the letter.

She read just a portion of it. Her recollection of its

details is sketchy. While I believe Ms. Taylor testified

truthfully, her testimony falls well short of

establishing that DFAS intentionally destroyed data

collected by the PAT team.

* Tlowever I do note that if the letter had directed the recipients

to collect and destroy the documents gathered by the PAT team,

the statement would not be offered for the truth of the matter

stated. Rather, it would have been offered to prove that the

direction to destroy the documents had been given. But Ms.

Taylor’s testimony falls well short of proof by a preponderance

that the letter directed high level supervisors at DFAS-CO to

destroy PAT report data.

34a

Having determined that plaintiffs have failed to

prove that defendant intentionally destroyed the PAT

data, I turn to the question of what sanction is

appropriate for defendant’s failure to take action to

preserve the documents. It is undisputed that the best

data to assess whether DFAS promotion practices had

a disparate impact would have been all the applicant

flow data from 1992 through 1996. Plaintiffs argue

that DFAS-CO should have known that the applicant

flow data might be relevant to future litigation. First,

Director Coffee’s December 6, 1996 memo regarded an

investigation to determine whether management

selection practices were discriminatory. Second,

plaintiffs’ March 5, 1997 EEO complaint, Pls. Ex. 1,

put DFAS-CO on notice that litigation over whether

those practices were discriminatory was possible. Pls.

Ex. 9. However, assuming that the government’s

record retention and destruction policy was followed,

at the time the PAT team began gathering information

the applicant flow data for 1992 through 1994 would

already have been destroyed. Tr. Vol. IV at 126:16-20.

When the EEO complaint was filed in mid-1997,

applicant flow data from only part of 1995 and all of

1996 would have been available. Tr. Vol. IV at

126:3-20. In August 1998, when this suit was filed, all

applicant flow data through the first three quarters of

1996 should have been destroyed. Tr. Vol IV at

127:12-14 and 22-25. By December 1998, when the

Court ordered defendant to preserve documents

responsive to plaintiffs’ September 2, 1998 written

discovery requests, all applicant flow data from

1992-1996 should have been destroyed. Tr. Vol. IV at

127:2-4.

Based on the experts’ testimony that they needed

the applicant flow data from substantially the entire

35a

period 1992-1996 to make a statistical analysis of

competitive promotions and the destruction of most of

that data in accordance with DFAS’s records retention

policy before the PAT team _ began collecting

information, I conclude that the evidence does not

support plaintiffs’ position that the PAT documents

would have assisted the finder of fact with

determining whether or not the DFAS promotion

practices had a disparate impact on African Americans

from 1992 through 1996. The PAT team looked at no

applicant flow data from 1992-1994. Despite the PAT

report bibliography’s reference to “Job Opportunity

Announcements of the Columbus Center for Fiscal

Years 1995 & 1996,” the testimony of Thomas Gary,

Derek Saunders, and Carol Scott at trial establishes by

a preponderance of the evidence that the PAT team did

not review the entire applicant flow data for 1995 and

1996. Mr. Gary indicated that the two boxes of

documents were primarily the completed surveys and

corresponding notes. Had Mr. Gary retained copies of

all the applicant flow data for 1995 and 1996 in

addition to the 427 completed surveys, all these

documents could not have been stored in two boxes.

Furthermore, Mr. Gary testified that the reason he

took the surveys was his concern with protecting the

confidentiality of information provided by the survey

respondents. The applicant flow data was accessible by

DFAS management and maintained by the Human

Resources department. Mr. Gary had no motive to take

the applicant flow data to protect DFAS employees

from retaliation from management.

Both experts agreed that neither the approximately

30 Job Opportunity Announcement files reviewed by

PAT nor the completed surveys and corresponding

notes would have assisted them in analyzing the

36a

promotion rates of African Americans at DFAS from

1992 through 1996. Even if the applicant flow data

that was extant at the time PAT convened was

available, both experts agreed that they would have

had insufficient data to extrapolate findings regarding

the five years of promotions at issue in this litigation.

It is undisputed that some documents responsive to

plaintiffs’ September 2, 1998 first request for

production of documents and first set of interrogatories

were lost and that defendant failed to take steps in

December 1998 to preserve the documents and other

information needed to respond to those discovery

requests. Had defendant’s counsel complied with the

Court’s December 18, 1998 Discovery and Scheduling

Conference Order, it seems probable that the

documents would have been located and preserved

since plaintiffs served their document requests on

September 2, 1998 and Gary had them in his

possession until at least late 1997 or eariy 1998. Had

a search been instituted in December 1998 as my

Order, the Rules of Civil Procedure, and the case law

regarding preservation of evidence require, it could

have been quickly determined that Gary took

possession of tre PAT surveys and, perhaps, some

other documents related to his investigation in or

about January 1997. Had he been contacted in

December 1998, his memory about the whereabouts of

those documents would have }2en fresh. If, as he

testified, he had received a letter directing him to send

the documents to DFAS headquarters in late 1997 or

early 1998, he would likely have remembered the

details and the records could well have been located. If,

contrary to his testimony, the records were still in

storage at Togar and Associates, he could have turned

them over to aefendant’s counsel.

37a

My July 2, 2003 Order concluded that the “fault lies

with defendant for the loss of the survey forms,

documents from DFAS-CO employees, spread sheets,

documents provided by DFAS to Togar and Associates,

and other documents developed by Togar and

Associates.” (Doc. 78, at p. 13.) It does not expressly

address plaintiffs request for sanctions based on the

destruction of the applicant flow data for 1992-1996.

Plaintiffs, who were represented by counsel during the

proceedings on the EEO complaint, did not then ask

for applicant flow data, and they did not ask that

applicant flow data for 1992 through 1996 be

preserved. None of the discovery requests plaintifis

had served on defendant on or before December 18,

1998 sought applicant flow data. They did seek the

personnel files of the individual plaintiffs, but there is

no evidence that applicant flow data would have been

in those files. Although, the July 2, 2003 Order asserts

that “plaintiffs’ discovery requests encompassed

promotion records.” That statement is accurate insofar

as records relating to the fact that an individual

plaintiff was promoted would be contained in that

employee’s personnel file, but it is inaccurate if it

taken to mean that plaintiffs’ discovery requests

sought the applicant flow data that the experts needed

to make a statistical analysis of the promotions made

during the period 1992-1996.

Even though plaintiffs had not sought applicant

flow data in their first wave of written discovery,

defendant may have had an independent legal duty to

preserve it. Generally, a party has the duty to preserve

documents once it knows that they are relevant to

resolution of the clatms and defenses raised in a

lawsuit. See Broccoli v. Echostar Communications

Corp., 229 F.R.D. 506, 510 (D. Md. 2005). Here the

38a

expert testimony and the case law, both of which base

a disparate impact analysis on a comparison of the

promotion rates of members of the protected class with

that of employees who are not a member of the

protected class, strongly support the existence of a.

duty to preserve the applicant flow data. Defendant’s

counsel should have recognized that the applicant flow

data was essential to a factual determination of

plaintiffs’ disparate impact claim.

Plaintiffs argue that under 29 C.F.R. § 1607.1 et

seq. and § 1607.4(A) as interpreted in Wards Cove, 490

U.S. at 657-58, DFAS-CO had an obligation to retain

applicant flow data. Plaintiffs alleged disparate impact

in their EEO charge and the complaint initiating this

lawsuit. The best evidence to prove-or

disprove—disparate impact is a statistical analysis of

the ratio of the number of minorities promoted to the

number of minorities seeking promotion compared

with the ratio of the number of non-minorities

promoted to the number of non-minorities seeking

promotion. Applicant flow data is the best evidence to

use for that analysis.

At this stage of the proceedings, plaintiffs argue

that the Court should sanction defendant for its

destruction of evidence by awarding plaintiffs their

expert and attorney fees, or in the alternative, finding

liability on behalf of defendant.°

Rule 37(b)(2) of the Federal Rules of Civil

Procedure provides in pertinent part:

° In my July 2, 2003 Order, I determined that an entry of

judgment against defendant was not warranted.

39a

(2) Sanctions by Court in Which Action is

Pending. If a party . . . fails to obey an order to

provide or permit discovery, including an order

made under subdivision (a) of this rule or Rule

35, . .. the court in which the action is pending

may make such orders in regard to the failure as

are just, and among others the following:

(A) An order that the matters regarding which

the order was made or any other designated

facts shall be taken to be established for the

purposes of the action in accordance with the

claim of the party obtaining the order;

(B) An order refusing to allow the disobedient

party to support or oppose designated claims or

defenses, or prohibiting that party from

introducing designated matters in evidence;

(C) An order striking out pleadings or parts

thereof, or staying further proceedings until the

order is obeyed, or dismissing the action or

proceeding or any part thereof, or rendering a

judgment by default against the disobedient

party;

(D) In lieu of any of the foregoing orders or in

addition thereto, an order treating as a

contempt of court the failure to obey any orders

except an order to submit to a physical or

mental examination;

In heu of any of the foregoing orders or in

addition thereto, the court shall require the

40a

party failing to obey the order or the attorney

advising that party or both to pay the

reasonable expenses, including attorney’s fees,

caused by the failure, unless the court finds that

the failure was substantially justified or that

other circumstances make an award of expenses

unjust.

When the PAT team convened in December 1996,

Mr. Saunders testified that only applicant flow data

from 1996 and part of 1995 was available. Tr. Vol. [IV

at 126:3-20. And, on March 5, 1997, when plaintiffs

filed the class complaint with the EEO counselor,

DFAS would only have had JOAs from 1996 and part

of 1995. Of the thousands of JOAs for the five year

period 1992-1996, there were only something over 100

JOAs in fiscal year 1995 and just 57 in fiscal year

1996.° Tr. Vol. IV at 122:17-19. As previously noted,

both experts testified that this data would not have

provided sufficient information from which one could

draw conclusions regarding the entire period that is

the subject of this litigation. Even so, defendant should

have known that this data would play a central role in

this case, and consequently, as I previously ruled,

sanctions are warranted.

Mr. Saunders also testified that the PAT report did

not rely on all of the applicant flow data that was still

in existence. He testified that he was the sole person

who retrieved applicant flow data from the Human

Resources ofiice for the PAT team. He further testified

* The PAT report would reverse these numbers. It states that

there were 57 Job Vacancy Announcements in FY 95 and 134 in

FY 96. Pls. Ex. 4 at 15..

4la

that he only pulled approximately 30 JOAs for review

by the PAT team. These 30 JOAs were selected

because the PAT team received employee complaints

of favoritism or other partiality in the selection process

for these promotions. Tr. Vol. IV at 118-120. Togar and

Associates then selected the portions of the JOAs they

wanted to copy.

Although the record establishes that the PAT team

reviewed partial data provided by Mr. Saunders

regarding approximately 30 specific promotion

decisions, there 1s no evidence that the PAT team

reviewed any other applicant flow data. Although I do

not believe that the PAT report’s supporting

documentation or the JOAs still in existence at the

tinie this action was initiated would have provided the

fact finder with the necessary data to assess whether

promotional practices had a disparate impact on

African Americans at DFAS, I do find that sanctions

are warranted under the circumstances given defense

counsel’s failure to comply with the Court’s December

18, 1998 Order. Had counsel ivucated and gathered

documents as she had been ordered to do, the Court

would have had the benefit of viewing the supporting

documentation without the need for extensive

testimony and briefing with regard to its import.

As a sanction for this failure to comply with a court

order, defendant is ORDERED to pay for all of

plaintiffs’ attorney fees and costs expended in

litigating and briefing this issue, including plaintiffs

expenses incurred with respect to offering rebuttal

testimony by way of Ms. Taylor. However, I do not find

that plaintiffs should be awarded their expert fees or

that hability should be imposed as a sanction because

the un-controverted expert testimony is that the

.

a

42a

missing documents would not have provided the

information needed to make a finding of liability. The

applicant flow data for 1992 through 1996 would have

served as the best evidence for determining whether or

not DFAS promotion practices disparately impacted

African Americans, and this data was not contained in

the missing documents. Without the applicant flow

data for 1992 through 1996, plaintiffs had little choice

but to retain an expert to attempt to recreate what she

believed that applicant flow data would have shown.

Had plaintiffs obtained the applicant flow data from

1996 and a part of 1995, this expert analysis in all

likelihood would still have veen necessary for the years

1992-1994 and part of 1995.

As a further sanction, the Court accepts the

statistical information contained in the PAT report as

accurate. The Court also admits for the truth of the

matter asserted therein the PAT report summaries of

the employee statements complaining about

employment practices that gave supervisors discretion

which they believed had a disparate impact on them as

African Americans.

Plaintiff is ORDERED to submit an itemized

statement of such costs and fees and supporting

memorandum within ten (10) days of the date of

this Order. Thereafter, defendant may file a brief in

opposition within seven (7) days of the date of service

of plaintiff's statement and supporting memorandum.

43a

IV. Findings of Fact

A. Overview

DFAS-CO periorms centralized finance and

accounting services for smaller agencies within the

Department of Defense. It was created when DOD

consolidated field offices and transferred their

functions to DFAS-CO. Employees of the consolidated

filed offices were offered the opportunity to transfer to

Columbus, but fewer than 10 percent did. A large

segment of those who transferred were African

American.

During the period 1991-95, DFAS-CO hired around

3,000 employees, an average of 70 a month. There

were many promotional opportunities. Tr. Vol. IV at

112:7-8. It was common to have upwards of 200

applicants for a vacancy. Tr. Vol. IV at 113:3-11,

171:19-172:11. Plaintiffs’ perception was that new

hires were from “off the streets” of Columbus. They

were mostly Caucasian and had no prior government

experience; yet plaintiffs believed they were promoted

at a faster rate than African Americans who had

transferred to DFAS-CO from other DOD facilities.

Many plaintiffs testified that they trained whites who

later became their supervisors. They believed that

favoritism and other promotional practices limited

their opportunities for promotion to GS-7 through

GS-11 positions.

B. Plaintiffs’ Testimony

Kevin Gray. Before coming to DFAS-CO, Gray was

a Financial Accounts Clerk GS 4/5/6. Gray had a B.A.

in business administration management and an M.S.

44a

in information resource management. He was

employed at DFAS-CO from 1990 to 2000. In 1996,

Gray was sclected for a 510 Accounting Technician

position in another directorate. He ended his

employment at DFAS-CO as a GS-11 in the Accounting

Directorate. Gray applied for financial specialist

positions with promotional potential to GS-11 in the

501 professional series. Tr. Vol. 1 at 22. During his 10

years at DFAS-CO, Gray applied for approximately 10

positions. Tr. Vol. I at 26. He testified that limited

areas of consideration prevented him from applying for

other positions. Tr. Vol. I at 43; Pls. Ex. 39.

Gray testified that persons holding an accounting

technician position were rarely’ selected for

professional positions. The 501 Financial Specialist

series was similar to an accountant position. However,

the 510 series was designated Accountants.

In 1994, Gray was a GS-7 Accounting Technician.

After not being selected several times, Gray started

gathering information about selection for jobs that he

did not get. He found that the selected candidates

often had a higher grade than that required to be

qualified for the position. Gray applied for several

positions where the Job Opportunity Announcement

stated that a lower grade qualified for the position, but

a higher grade employee was selected. Gray saw

employees detailed into supervisory positions over

him. He believed that selecting officials had a lot of

discretion.

On one occasion, Gray applied for promotion as an

accountant. Human Resources said he was not

qualified, but after further review he was determined

to be qualified. However, the selection had already

45a

been made. The selected employee told Gray that he

was a friend of the selecting official.

On another occasion, the employees selected had

very limited experience, but the selecting official — a

white female — said that the selected employee showed

favorably from experience in another area unrelated to

the job.

Gray was given priority consideration for the next

promotion because of an EEO complaint, but a white

female selecting official selected a white female with a

higher grade for Vacancy No. 96-113 CJ. Pls. Ex. 39 at

77.

Gray believed that there was a pattern of selection

that hindered minorities from receiving promotions to

GS-7 through GS-11 positions. Consequently, few

minorities were qualified to apply for GS-12 and

higher up positions. Because of this belief, he and some

30 other African American employees met in 1996.

They presented their concerns to Mr. Coffee, who then

appointed the Process Action ‘eam committee.

Yulander Edwards. Edwards began her federal

employment in 1984 at a Department of Defense

facility in St. Louis as a GS-3 file clerk. She has two

years of junior college credits. When she transferred to

Columbus in 1992, she was placed in a GS-7, 503

series position. She was then transferred from that

lead position to a front pay monitor position. In late

1993, Edwards transferred back to a GS-7, 503 series

lead position. In January 1995, she was reassigned to

a GS-7, 525 Accounting Technician position, the same

position she had held at the end of her employment in

46a

S.. Louis. In 2001, she was promoted to a GS-9, 509

position. Tr. Vol. Il at 7.

From 1992 to 1995, Edwards applied for at least 10

promotions, but she/received none. Tr. Vol II at 10.

After that, promotional opportunities diminished. She

applied several times for GS-7- target-11 501 series

financial specialist positions and made the certification

list. Tr. Vol. I] at 13. Once the selecting official chose

Michelle Kandrack, a white female, who had only two

years of federal service. Edwards then had four years

of qualifying experience. She applied for several 501

series jobs, made the certification list, but was told

that they were not going to fill the position with that

certification list.

Edwards applied for a 501 Financial Specialist

target 11 position and was told that she was not

qualified. Tr. Vol. Il at 14. She applied for several

other 501 Financial Specialist target 11 jobs and made

the certification list. She applied for a GS-8, 525 series

supervisory position and was told she was not

qualified even though she had held a GS-7, 525 series

position in St. Louis. Tr. Vol. II at 16. She applied for

a personnel specialist position GS-5 target 11 and was

told she was not qualified. Edwards believed that her

being assigned to the 503 series before coming to

Columbus prevented her from advancing. Tr. Vol II at

18.

Kdwards was once detailed into a_ research

supervisor position for 90 days, but she was removed

after thirty days. Tr. Vol. II at 17. A white female with

less experience was promoted to the financial

specialist research supervisor position. She also

testified that there were positions she could not apply

4ila

for because they were restricted to a limited area of

consideration. Tr. Vol. II at 44.

Lisa Prater. Prater has a B.S. in_ business

management. She began working at DFAS-CO in

January 1991 as a GS-4, 503 series cash clerk in the

Dispensing Directorate. After one year, she was

promoted to a GS-5-target-6 financial accounts clerk

and later to the GS-6. Tr. Vol Il at 51; Pls. Ex. 45.

Between 1990 and 1996, Prater applied for no less

than 10 jobs. Generally, Prater was looking for a 501

series, GS-5-target-11 or 7-target-9 position. When she

applied and did not get a job, Prater would often make

an appointment with the selecting official to ask

whether she had problems with the interview or other

problems that might cause her to lose & promotion.

Prater was concerned that less well qualified

candidates were selected for promotion because they

were friends of the clique, were favored, or were

preselected. She believed that Ida Faye Groves, Patsy

Frisch, and Maura Shaffer, all white females,

determined who was promoted in the West Directorate

where she worked.

In 1996, Prater had problems with a white female

supervisor who wanted to rate her “Fully,” rather than

“Exceptional.” She also ranked high on _ the

certification list for a 501 series 7-target-9 position

(Job Announcement No. 96-018-AS) that was given to

an intern who was a GS-4 or GS-5. Tr. Vol. II at

63:22—65:19. In May 1996, Prater ranked high on the

certification list, but the promotion went to Michelle

48a

Little, a white female, who ranked much lower.’ In the

same month, B. J. Johns, a white female, was a

selected for a position on the same Job Opportunity

Announcement as Ms. Little, even though she was

lower on the list than Prater. Pls. Ex. 45 at ] 7.

Jarlisa_ Miles. In July 1990, Miles started at

DFAS-CO as a GS-5, 525 series Accounting

Technician. Tr. Vol. II at 83; Plfs. Ex. 42. She has a

B.S. in accounting and management. In 1991, Miles

was promoted to a GS-6, 525 series Accounting

Technician position. Later she applied for a GS-7/9/1i

position, but she believes that the supervisor had

already preselected Debbi Temesvary for the position.

Tr. Vol II at 95. Miles was later promoted to a GS-7,

525 series Accounting Technician position in Stock

Fund. Tr. Vol. II at 84. In 1994, she moved to another

agency. She was absorbed back into DFAS-CO in

November 1994 as a GS-7/9/11 accountant. Tr. Vol. II

at 84. After that, she made a normal progression to

GS-9, then GS-11. Then she was promoted to a GS-12

supervisor position. In June 2000, Miles left DFAS-CO.

She is now a GS-13 Director for the Department of

Navy, supervising 50 people who are graded GS-9/11.

Miles testified that during the period 1990-1996 she

applied for hundreds of positions. In 1991, she applied

’ Several plaintiffs testified that they learned that they ranked

higher than applicants on a certificate of referral or higher than

the applicant selected. While I do not doubt that someone may

have told them this, I do not believe that fact is evidence of

disparate impact. Although Human Resources did rank the

candidates, the top 15 and ties were listed on the certificate of

referral alphabetically. They were not ranked on the certificate,

and the hiring supervisor was not told their rank.

49a

for a GS-7/9/11, but a supervisor told her she was

wasting her time. The supervisor said that a friend of

the white male selecting officer was going to get the

job. A white female who did not have a college degree

was selected. Miles believed that her promotional

opportunities were diminished when a supervisor gave

her a “Fully” rating instead of a _ deserved

“Highly’rating. Rather than adjust the rating, the

supervisor gave Miles a cash award for duties she did

not perform, which she refused. Pls. Ex. 42, 4] 10.

Miles testified that when she worked under Ida

Faye Groves in the Stock Fund Directorate, she was

forced to train a lead, Judy Carnes, who was not

qualified when placed in the position.

Miles testified that she once observed Tom Dorner,

a white male, ask Bridget Watkins to look at a

certificate list and tell him who she recognized as a

Caucasian. Miles believed that supervisors promoted

whomever they wanted to.

Jackie McGrady. McGrady transferred to DFAS-CO

in September 1990 as a GS-6-target-11. Pls. Ex. 38.

She had a bachelors degree in social work, 20 hours of

course work in statistics, budgeting and accounting,

plus years of accounting and budgeting experience.

When she transferred to Columbus, she continued in

the Grow Your Own Accounting program, which she

had begun while working for the Department of

Defense in Chicago. In 1992, McGrady dropped out of

the program because she could not complete the course

work while adopting a child. Tr. Vol. II at 159. She

reverted back to the GS-6, 525 series position. In 1998,

McGrady, still a GS-6, left DFAS-CO.

50a

While at DFAS-CO, McGrady applied for several

promotions. Tr. Vol. I] at 161. Sometimes she was

erroneously not placed on the certification list. Other

times, another employee lateraled in. McGrady did not

get superior and excellence performance appraisals as

she had done in Chicago. She believed that

non-minorities who were doing less work were given

higher evaluations. McGrady testified that she and

other African Americans were doing eight or nine

times the work of those getting awards and better

performance evaluations. She said that very few

African Americans got awards. Their supervisors were

people they had trained.

McGrady began creating her own database. She

testified that more than 50-60 employees told her

about their qualifications, but she could only name

Stacy Tolson. McGrady further testified that she once

analyzed the work production of employees in the unit

where she worked. During a period of perhaps two

years, she analyzed the work loads of about 60

employees.

McGrady testified that from 1992 to 1997, she

made numerous applications for account technician

and financial assistance positions. She applied for a

supervisor position, but it was filled by a white female.

In contrast, she testified that Amy Thomas, a white

female, was detailed to a position, then she became a

supervisor and moved up the chain of command.

McGrady once got a temporary promotion to GS-7

because she was doing a difficult audit.

McGrady testified that Stacey Tolson got her job by

favoritism. Either her father or her father-in-law was

5la

the Director of Payroll; and she was close to Ms.

Frisch.

Susan McNeal. McNeal’s first federal employment

was in 1985. Between then and June 1992, when she

transferred to DFAS-CO, McNeal held a number of a

clerical positions. Pls. Ex. 41. She has an associate

degree in business administration. McNeal testified

that she was told prior to moving to Columbu: that she

was eligible for positions as a 540 series voucher

examiner, a 525 series accounting technician, and her

then current 503 series position. She came in as a

GS-4/5/6, in a 503 series position, as a G S-5. She

became a GS-6 after nine months. She remained a

GS-6 for 14 years. Tr. Vol. II at 236.

From 1992 to 1996, McNeal applied for 29

promotions. Tr. Vol. II at 215. She applied for positions

as a contract input lead, management analyst,

financial management central management intern,

and financial specialist. Tr. Vol. II at 217 and 224. She

testified, “At one point I was applying for everything.”

Tr. Vol. U at 251.

McNeal testified that although she was not

promoted, people from off the streets were employed at

DFAS-CO. She said that employees promoted to

contract input technician positions had no relevant

government experience before coming to DFAS-CO.

For example, Robin Terry, a white female, advanced

from a GS-4 to a GS-5 in six months.

Mary L. Phillips. Phillips transferred to DFAS-CO

in August 1990. She has a high school education with

some college and vocational training. Pls. Ex. 43.

Phillips came in as a GS-5 in the 525 series. Tr. Vol.

52a

IiI at 49. In October 1992, she became a GS-7/8

supervisor. Tr. Vol. HI at 10,12-13. In 1999, she was

promoted to GS-9. Tr. Vol. III at 49.

Phillips applied for promotions to positions as a 525

series accounting technician and a 501 series financial

specialist. Tr. Vol. III at 9. She also applied for a recon

technician position and a GS-9/11, 501 series position.

Tr. Vol. III at 13. Phillips testified that in 1996 she

applied for Vacancy Announcement No. 96-103R, a

GS-5/7/9/11 financial specialist position in the 501

series. She was interviewed, but B. J. Johns, a white

female, was selected. Phillips testified that she scored

more points on the certification list than Johns. She

said that she had more experience, interviewed well,

and reaked higher than Johns. On cross-examination,

Phillips testified that she had no complaints about her

progression from a GS-5, to a GS-c, then to a GS-7.

Marcia Spence. Spence came to DFAS-CO as a

GS-7, series 525, accounting technician. Tr. Vol. II" at

65. When she was coming to Columbus, she appliec for

a GS-7/9/11, series 501, financial specialist position.

Tr. Vol. III at 65. In May 1993, Spence was promoted

to GS-9/11, 501 series position. Tr. Vol. III at 77.

Spence testified that she applied for several 501

series promotions In November 1995, she applied for

several GS-ll-target-12 financial prevalidation

speciaiist positions. Three people were selected for the

positions, including Chris Carter, “employee X” in the

PAT report. In December 1995-January 1996, Human

Resources determined that Carter did not have the

time-in-grade at GS-11, 501 series. Management then

elected not to fill the position. In March 1996, Spence

applied again for the same financial prevalidation

53a

specialist position. She was interviewed, but Carter

again was selected for the job. In November 2000,

Spence became a GS-12, 501 series, Financial Systems

Specialist. Tr. Vol. III at 76.

On cross-examination, Spence testified that the list

of favored employees included maybe 15 names.

Denise Goudy. In August 1989, Goudy came to

DFAS-CO as an input technician, GS-5, 503 series.

She is a high school graduate with course work in

business law and accounting principles. Pls. Ex. 38. In

June 1990, Goudy received a competitive promotion to

GS-6, 503 series. In late 1991, she became a lead input

technician. In late 1992, she was promoted to contro]

supervisor, GS-7, 503 series. In early1993, Goudy

became a contract input supervisor, GS-8, 503 series.

In 1996, Goudy applied for a Financial Specialist

position, GS-7/9/11, 525 series. Vacancy

Announcement No. 96-103. The person selected scored

a 47, which was well-below the JOA cutoff of 75.

Goudy’s own experience as a supervisor was that no

one exerted any pressure to select black or white job

promotion applicants. Her division chief did remind

her to give every qualified applicant fair consideration.

Lerdon Woodfolk. Woodfolk came to DFAS-CO in

December 1990 in a GS-11, 501 series position. He had

wanted to enter as a GS-11-target-12, but because of a

delay in the completion of his security clearance the

position was no longer available when he arrived in

Columbus. Woodfolk is a high school graduate with

three years of college and a number of Defense

Acquisitions University courses. Pls. Ex. 48. In

54a

February 1992, Woodfolk was promoted to a GS-12

Financial Specialist position.

Woodfolk testified that when he worked in North

Fiscal Budget he instructed Tim Mock, a white male,

about the budgeting process that he had helped set up.

Mock then gained a temporary assignment to the

position Woodfolk trained him for. Mock became a

GS-13. Mock was selected for promotion by Nancy

Drew, a white female, who had been an intern when

Woodfolk worked at DCASR-Cleveland.

Woodfolk applied for a Supervisory Management

and Program Analyst position, GS-13, 343 series. JOA

94-051-AS, Pls. Ex. 48, at 8. In February 1994-April

1994, he had been detailed to the position to assist the

unit during its transition. Patsy Frisch, a white

female, selected Jim Hunter, a white male, for the

position.

Reba Rivera, who had no directly related

experience, was selected over Woodfolk for a

Supervisory Management Analyst position in the

Stock Fund Directorate. Pls. Ex. 48, at 8. Woodfolk

identified other job opportunities set out in Plaintiffs’

exhibit 48 that he was denied. He further testified that

Lou Ricks, Lori Hunter, Jeff Groslosc, Paul Huey and

Greg Allen were examples of white employees who

made rapid advancement from GS-5 through GS-9 to

GS-18 and GS-15.

Woodfolk testified that he felt betrayed because he

was a “good soldier.” He described himself as a

self-starter and an independent worker wno got the job

done. He believed his advancement should have been

better.

55a

Willie Phillips. Phillips transferred to DFAS-CO in

Sevtember 1989 as an accounting technician, GS-7,

525 series. He is a high school graduate with three

years of college as well as job-related education. In the

fall of 1990, Phillips was promoted to Voucher

Examiner Superv:sor, GS-8, 540 series. In 1992,

Phillips had a disagreement about supervision

management practices and resigned the position. He

reverted to his original accounting technician GS-7,

525 series position. Later that year, he was again

promoted to a GS-8, 525 series position. Sometime

during 1993-94, Phillips was reclassified to become a

GS-9. In February 1996, Phillips became a financial

prevalidation specialist, GS-9, 501 series.

Phillips applied for jobs in both the 501 and 525

series. He also applied for some jobs at a lower grade

level because the target grade was higher.

Phillips testified that promotions at DFAS-CO were

based more on who you knew, and who you hung out

with, rather than job experience and qualifications. He

believed that some favored employees were promoted

through unearned awards, experience gained when

detailed, and transfers between directorates.

Phillips also testified that promotional

opportunities were restricted because some were

available only to employees working in the directorate

where the vacancy occurred. He believed awards have

a large impact on promotions because the points given

for awards boosted the applicant’s rating. Discipline

also impacts one’s performance appraisal and the

applicant’s saleability to some extent.

56a

Phillips’ perception was that as new employees

rapidly came into DFAS-CO, African Americans were

not promoted as rapidly to middle manager positions.

Phillips himself made the top three for several

positions he applied for, only to have the select-ing

official, Sue Jennings, a white female, recall the

certification list. He also lost the opportunity to apply

for promotion during a six-month period in 1996

because he was rated only “Met expectations.” During

the preceding three years he had been rated

“Exceptional” twice and “Highly” once.

Phillips testified that when he worked for Pat Rose,

she hired Recon Supervisors and clerks who had no

experience. Philip showed them how to do the job, but

he received a very bad rating. Phillips asked Steve

Frisch for a transfer so he could apply for the same job

in another division, but Frisch said that he was not in

the habit of transferring. Yet a white male was

permitted to transfer under similar circumstances.

C. Defendant’s Witnesses

Derek Saunders. Saunders, an African American

who worked for 11 years at DFAS-CO as a Human

Resources Specialist, testified for the government

about promotional practices and the operation of the

PAT team. When there was a Job Vacancy

Announcement, applicants would fill out a SF-171. It

does not ask for the applicant’s race. At no time is the

applicant ever asked about his or her race. Human

resources refers the top 15 applicants and ties to the

selecting official by a Certificate of Referral.

Applicants are listed alphabetically. They are not

ranked by score. Tr. Vol. IV at 116:1-14. During the

period 1990 through 1996, there were 50 Human

57a

Resources Specialists who looked through applications

and developed certificates of referral.

Plaintiffs testified that when a job vacancy occurs,

an announcement is posted stating the minimum

qualifications needed and the application deadline. See

Tr. Vol. | at 27:2-28:21. AHuman Resources Specialist

reviews the applications, applies selection criteria, and

ranks the top 15 candidates. See Tr. Vol. I at

28:22-131:20. This hst goes through the Human

Resources office and equal employment office and is

given to the supervisor in charge of filling the vacancy.

The supervisor has the discretion to interview, or not

interview, anyone on the list. He may request that

more persons be added to the list. He may choose

anyone on the list to fill the vacancy. See Tr. Vol. I at

31:21-32:8; 27:2-132:8; 33:11-37:18; 42:13-52:2.

Saunders testified that each certificate of referral

could include fifteen or more names. Tr. Vol. IV at

11:14-20. The top fifteen candidates plus ties were

referred to the selecting official. Jd.

DFAS-CO started with several hundred employees.

It grew to 3600 employees in 1996. There could be

hundreds of applicants for a specific Job Opportunity

Announcement. It was not unusual to have over 300

applicants for a position.

Saunders was the Human Resources Specialist

member of the PAT team. He was the only member of

the PAT team who could get the JOA files. So all the

JOA material came through Saunders to Togar and

Associates. JOA files included the JOA Vacancy

Announcement, the applications, those eligible for the

positions, those qualified for it, the Certificate of

58a

Referral, and the applicant selected from either 1995

or 1996. Tr. Vol. IV at 120:10-18; 130:5-21.

Around 30 issues were identified by Togar, so

Saunders pulled about 30 JOAs. He testified that all

the issues regarding the 30 JOA files are documented

in the PAT report. He further testified that Togar did

not say there was discrimination as to any of the 30

JOA files that the PAT Team looked at.

Saunders testified that when Togar asked for a

JOA file, he got it. Saunders stayed in the room while

Togar reviewed the file. He copied whatever Togar

asked for, then returned the file. Togar kept the copies.

At the time Saunders was pulling the files, Human

Resources would have had all of the JOA files for FY96

and 9 months of FY95.° There were no JOA files for

1990 through 1994 available. There were

approximately 100° JOA files for FY95 and 57 for

FY96. See Tr. Vol IV at 120:19-121:24.

Section 41 of DFAS’s Records Management

Administrative Procedures provides that JOA files are

destroyed two calendar years after the JOA vacancy

announcement closed. Assuming that policy was

followed, in August 1998 only the JOA files from the

last quarter of 1996 would have been available. During

the entire year 1996, there were only 57 JOA files. In

September 1998, ali of the JOA files from 1992

through 1996 would have been destroyed.

* The Fiscal Year begins October 1. In December 1996, it was

FY97.

59a

The standards for job series are set by OPM. The

501, 503, 510, and 511 series are two-grade interval

series. The 501 and 503 series are a combination of two

or more different positions and duties in one position.

The 510 series is for accountants, and the 511 series is

for auditors. The 525 series is for accounting

technicians, and the 540 series is for voucher

examiners. Line employees in the 525 and 540 series

go up to GS-7. Supervisors are GS-8s or GS-9s, but

rarely GS-10s. Generally, an employee must have one

year in grade or specialized experience to be eligible

for promotion to the next grade. Because of this

requirement, it is rare that an applicant would be

eligible for a two-grade promotion.

Saunders testified that when EEO notified Human

Resources that an EEO charge had been filed

regarding a specific JOA file, that file would be

preserved.

Carol Scott. Scott, who was an EEO Specialist in

1996 and a member of the PAT team, testified for the

government. She said that the EEO MAS database had

no information from the JOA files. EEO recorded race,

nationality, gender and age information for the top 15

applicants. It had no information about other

applicants eligible for promotion who did not make the

Certification of Referral. The EEO office did not store

JOA files. It would check out JOA files, if needed, for

review in connection with an investigation.

D. The PAT Report

African Americans were concerned that few African

Americans were being promoted to higher GS levels.

Pls. Exs. 8-11. Director Charles Coffee formed the PAT

60a

team to investigate allegations of favoritism, racism,

and pre-selection. Groves, p. 47; Pls. Ex. 5. Ethnic

preferences and the race of the supervisor played a

role in selection. Pls. Ex. 4 at 37-38. 12/6/96 Coffee

Memorandum.

In 1996, amid employee concerns regarding the

fairness of the DFAS-CO promotion process, Director

Charles Coffee and Principal Deputy Director Ida Faye

Groves directed a team of DFAS employees and two

outside consultants to investigate allegations of

favoritism, racism, and pre-selection. The Process

Action Team (“PAT team”) was asked to examine the

following issues:

) Fi Limited upward mobility opportunities

for employees.

2. Limited internal mobility.

3. Consistency in rating, ranking, and

referral of applicants.

4. Are only non-minorities advancing to

GS-11 and above?

5. Is there disparity in assessing candidates’

qualifications?

6. Is there pre-selection for jobs?

7 Are personal relationships impacting

selection and promotions?

8. Does nepotism exist - family members

(spouses/children) in the same

organizations or in supervisory chains?

9. Are management selection patterns

discriminatory?

10. Is the = distribution of awards,

performance appraisals, and disciplinary

actions equitable?

6la

11. Do employees fear reprisal from

DFAS-HQ and Columbus Center EEO

personnel?

12. Does the EEO staff only comply with

management’s decision?

Pls. Ex. 4 at 0007-09. The PAT team interviewed 25%

of the DFAS-CC workforce and analyzed the merit

promotion and EEO statistical data for 1995 and 1996.

Employees who were not randomly sclected to be

interviewed were invited to speak with Togar

representatives about any concerns that they had with

respect to the promotion processes. There were 427

surveys completed. Some of the survey respondents

were randomly selected, but anyone else who wanted

to was permitted to complete a survey.

The PAT report made the following conclusions:

1. Minorities continue to be promoted at

lower rates than non-minority employees

especially at the higher levels;

2. Several groups, especially Black males,

continue to be underrepresented at the

higher grade levels at the Columbus

Center;

3. Evaluation of application content was

subjective;

4. There is a_ strong perception§ that

discrimination, whether race based or

not, exists at the Columbus Center. In

the case of employee X, management may

62a

have committed a prohibited personnel

practice;

5. White females received 72 percent of the

exceptional rating in 1995. This may have

impacted their selection during the Merit

Promotion process. This research has

shown that supervisors rate employees of

their race or national origin group higher

than they rate employees from other

groups.

6. The data shows that white male selecting

officials show strong ethnic preferences in

awarding promotions.

The DFAS EEO report provides data on several

employment practices that meet the four-fifths rule as

noted by the Sixth Circuit. Pls. Ex. 1. Plaintiffs argue

that the PAT report findings are admissions by

DFAS-CO. Plaintiffs also argue that EEO officer

findings are also admissions. But defendant maintains

that the PAT report was merely a perception report.

Tr. Vol. I at 177:16-25, 178:1-5. The data was not

designed to measure disparate impact. See Tr. Vol. I at

176:1-9. The PAT team did not randomly select

applicant flow data. The survey was not random.

There was no statistician on the PAT team. Tr. Vol. IV

at 132:17-18; 190: 7-8. Standard statistical procedures

were not used. Of the 427 survey respondents, some

were randomly selected, but a significant number

volunteered to complete the survey. Tr. Vol. I at

189:4-8. The PAT report states that 188 employees

who were not randomly selected were interviewed. Pls.

Ex. 4 at 57. During FY 95 and FY 96, 77% of the

awards went to non-minorities, and only 23% went to

63a

minorities. Pls. Ex. 4 at 1-42. With respect to

performance ratings, whites received 72% of the

exceptional ratings, although they comprised only 46%

of DFAS-CO. Pls. Ex. 4 at 43. Black males received

only 4.5% of the highly and exceptional ratings,

although they made up 6.47% of DEFAS-CO.

Although 21% of DFAS-CO’s employees were

African American, they received approximutely 49% of

the disciplinary actions taken during FY 95 and FY 96.

Id. at 39. They also received more suspensions. /d.

During FY 95 there were a total of 21 reprimands, 60

suspensions of up to 14 days, and 15 suspensions of 15

days/removals. /d. at 106. During FY 96 there were a

total of 14 reprimands, 33 suspensions of up to 14

days, and 17 suspensions of 15 days/removals. Jd. at

106.”

The PAT report also contained some data about the

distribution of jobs at DFAS-CO. It states that there

were 3076 positions. It listed the total number of

positions at each GS level:

GS-3 66 GS-7 671 GS-11 344

GS-4 248 15-8 54 GS-12 278

GS-5 = 493 GS-9 —s-:125 GS-13 77

GS-6 661 GS-10 2 GS-14 332

GS-15 = 12

Id. at 15. In FY 96, 134 vacancy announcements were

issued. In FY 95, there were 57. /d. In FY 96, 51 of the

vacancies were limited to employees within the

directorate were the vacancy was. In FY 95, just two of

” Some employees received multiple disciplinary actions

64a

the vacancy announcements were so limited.*® Jd. at

16. During FY 95 and FY 96, 475 positions were filled.

The referred candidate pool for those jobs was 5,264.

Id. at 23.

The PAT report stated that “minorities continue to

be promoted at a lower rates than non-minority

employees especially at higher levels.” Jd. at 33. It

include an analysis of the number of white males,

white females, African American males and African

American females who are employed in and promoted

to grades 5-8, 9-12 and 13-15. Id. at 25-33. However,

the report includes only percentages of each group in

or promoted to a particular grade range. It does not

include the number of qualified African Americans

applying for and the number promoted to positions

within a grade range. It does give the number of each

group as a percentage of the total number of DFAS-CO

employees and as a percentage of those employed in

each grade range. It gives the percentages of each

group included in the referred candidate pool for

vacancies in each grade range, but it contains no

information about the number of African American

and non-minority qualified candidates who applied for

position vacancies in the grade range.

'° Defendant offered testimony that once the number of employee

positions at DFAS-CO began to contract job vacancy

announcements were often limited to the directorate where the

vacancy occurred because funds were not available for the position

if it was filled by someone from another directorate. Tr. Vo}. I at

219:13-20.

65a

V. Conclusions of Law

Title VII makes it unlawful for an employer “to fail

or refuse to hire or to discharge any individual, or

otherwise to discriminate against any individual with

respect to his compensation, terms, conditions, or

privileges of employment, because of such individual’s

race, color, religion, sex, or national origin; or (2) to

limit, segregate, or classify his employees or applicants

for employment in any way which would deprive or

tend to deprive any individual of employment

opportunities or otherwise adversely affect his status

as an employee, because of such individual’s race,

color, religion, sex, or national origin.” 42 U.S.C.

§ 2000e-2. Section 2000e-16 of title 42 of the United

States Code prohibits certain federal employers from

engaging in the same discriminatory conduct.

Title VII of the Civil Rights Act of 1964, 42 U.S.C.

§2000e et seq., prohibits both overt discrimination and

“practices that are fair in form but discriminatory in

operation.” Griggs v. Duke Power Co., 401i U.S. 424,

431 (1971); Alexander v. Local 496, Laborers’ Intern.

Union of N. Am., 177 F.3d 394, 405 (6th Cir. 1999).

Title VII precludes “employment practices that are

facially neutral in their treatment of different groups

but that in fact fall more harshly on one group than

another and cannot be justified by business necessity.”

Int'l Bhd. of Teamsters v. United States, 431 U.S. 324,

335 n. 15 (1977); see also Watson v. Fort Worth Bank

and Trust, 487 U.S. 977, 986-7 (1988).

“Disparate impact occurs when a_ neutral

employment practice has an adverse impact on a

protected class.” Ramos v. Baxter Healthcare Corp. of

Puerto Rico, Inc., 256 F.Supp.2d 127, 145(D.P.R. 2003)

66a

(citing Hazen Paper Co. v. Biggins, 507 U.S. 604, 609

(1993)). Proof of intentional discrimination is not

required. See Mullin v. Raytheon Co., 164 F.3d 696,

699-700 (1st Cir. 1999). Rather, a disparate impact

claim rests on discriminatory results, i.e., “on proof

that the employer utilizes employment practices that

are facially neutral in their treatment of different

groups but ... in fact fall more harshly on one group

than another and cannot be justified by business

necessity.” Id. at "00 (citations and internal quotation

marks omitted).

Typically, a Title VII disparate impact claim has

three elements: identification, disparate impact, and

causation. Hemmings v. Tidyman’s Inc., 285 F.3d

1174, 1190 (9"*Cir. 2002). Here plaintiffs ask the Court

to consider defendant’s promotion process as a whole,

rather than a specific hiring or promotion practice. If

the Court concludes “that the elements of a

respondent’s decisionmaking process are not capable

of separation for analysis, the decisionmaking process

may be analyzed as one employment practice.” 42

U.S.C. § 2000e-2(k)(1)(B)(). Phillips, 400 F.3d at 398

n.8. To satisfy the merged disparate impact/causation

element, plaintiffs must prove by a preponderance of

the evidence that African American employees were

promoted at a lower rate than white employees. Id.

Plaintiffs’ burden in a Title VII disparate impact case

is to identify the facially neutral employment practices

challenged and establish that the facially neutral

employment practice caused a disparate impact on the

basis of race. Phillips v. Cohen, 400 F.3d 388, 397-98

(6" Cir. 2005).

For disparate impact claims involving promotional

policies, “the relevant statistical inquiry involves

67a

comparing the number of protected group members

benefitting from promotions with the number seeking

them; this figure is then contrasted with the

corresponding ratio for the non-protected group.”

Phillips, 400 F.3d at 399. In determining whether

disparate impact has occurred, “expert statistical

evidence... is not to be considered in a vacuum. . . ; it

must be considered ‘in light of all the evidence in the

record.” Jd. at 401 (quoting Bazemore v. Friday. 478

U.S. 385, 401 (1986)).

Plaintiffs can satisfy their burden “by

demonstrating disparities through statistical evidence

of a kind and degree sufficient to show that the

practice in question has caused the exclusion of

applicants for jobs cr promotions because of their

membership in a protected group.” Bacon, 205 F.R.D.

at 473 (citing Watson, 487 U.S. at 986-87; E.E.0.C. v.

Joe’s Stone Crab, Inc., 220 F.3d 1263, 1274-75 (11th

Cir. 2000)).

Typically, a plaintiff must establish that each

challenged employment practice causes a disparate

impact, but when “the complaining party can

demonstrate to the court that the elements of a

respondent’s decisionmaking process are not capable

of separation for analysis, the decisionmaking process

may be analyzed as one employment practice.” 42

U.S.C. §2000e-2(k)(1)(B)U). “In offering statistical

evidence, the plaintiff must eliminate the most

common nondiscriminatory reasons for the disparity.”

Bacon, 205 F.R.D. at 473 (citing Mozee v. American

Commercial Marine Service Co., 940 F.2d 1036, 1045

(7th Cir. 1991)). Plaintiffs are not required to take into

account all measurable variables, but they must

include the major factors potentially responsible for

68a

any disparity. Jd. “Normally, failure to include

variables will affect the analysis’ probativeness, not its

admissibility.” Bazemore v. Friday, 478 U.S. 385, 400

(1986).

Once plaintiff establishes a prima facie case, the

burden shifts to the employer to produce evidence that

the challenged practice is a business necessity.

Alexander, 177 F.3d at 406 (citing Wards Cove, 490

U.S. at 659). Defendant may challenge any of the

elements of the prima facie case or “show . . . that the

challenged practice is job-related and consistent with

business necessity.” £.E.O.C. v. Steamship, 48 F.3d

594, 602 (1“ Cir. 1995).

If the defendant fails in its efforts to counter the

plaintiffs prima facie case, then the factfinder is

entitled--though not necessarily compelled--to

enter judgment for the plaintiff. On the other

hand, even if the defendant stalemates the

prima facie case by elucidating a legitimate,

nondiscriminatory rationale for utilizing the

challenged practice, the plaintiff may still

prevail if [he] is able to establish that the

professed rationale is pretextual. The plaintiff

might demonstrate, for example, that some

other practice, without a similarly undesirable

side effect, was available and would have served

the defendant’s legitimate interest equally well.

E.E.O.C. v. Steamship, 48 F.3d at 602 (citations and

internal quotation marks omitted).

Defendant maintains that plaintiffs have not

proven that the promotional process at DFAS-CO is

incapable of separation for analysis. Plaintiffs have

69a

failed to respond to this argument either in their April

7, 2006 brief (doc. 20) or in their April 28, 2006 reply

to defendant’s merits brief (doc. 123).

The statute states plaintiffs’ burden of proving

disparate impact:

(k) Burden of proof in disparate impact cases

(1)(A) An unlawful employment practice based

on disparate impact is established under this

subchapter only if--

(i) a complaining party demonstrates that a

respondent uses a particular employment

practice that causes a disparate impact on the

basis of race, color, religion, sex, or national

origin and the respondent fails to demonstrate

that the challenged practice is job related for the

position in question and consistent with

business necessity; . . . .

(B)G) With respect to demonstrating that a

particular employment practice causes a

disparate impact as described in subparagraph

(A)(i), the complaining party shall demonstrate

that each particular challenged employment

practice causes a disparate impact, except that

if the complaining party can demonstrate to the

court that the elements of a respondent’s

decisionmaking process are not capable of

separation for analysis, the decisionmaking

process may be analyzed as one employment

practice.

42 U.S.C. § 2000e-2 (k). While I believe that defendant

is correct that most of the practices identified by

70a

plaintiffs could be separately analyzed," plaintiffs

maintain that the identified employment practices, in

the aggregate, gave supervisors a great deal of

discretion, which was exercised to the disadvantage of

African Americans. Although plaintiffs’ briefs discuss

a number of selection practices, they do not expressly

state which practices they rely on to prove disparate

impact. Plaintiffs’ briefs discuss the following

employment practices: detailing; changing the series

number of a position; withdrawing a certificate of

referral and later issuing a new JOA; requesting

additional names; limiting applications to a particular

directorate; performance ratings; performance awards;

discipline; favoritism, nepotism and preselection. Each

of these practices could be analyzed separately for

whether they have a disparate impact on African

Americans.

All the promotion selection practices identified by

plaintiffs tend to give supervisors more leeway to

consider subjective selection criteria. Subjective

employment criteria may be challenged under the

disparate impact approach. Watson v. Fort Worth Bank

& Trust, 487 U.S. 977, 990 (1988). Plaintiffs here do

not challenge any of the objective selection criteria,

'' For instance, the disparate impact of detailing could be

determined by identifying each detailed employee and

determining whether the detailed employees were selected for the

positions, and determining the races of each person detailed, and

determining the race of each person selected for the positions. A

similar analysis could be performed for every position allegedly

filled by nepotism or favoritism. The PAT report identified

employees who allegedly received prometion through nepotism

and favoritism. Their races could be determined

Tla

such as time in grade or positive education

requirements.

Each supervisor had a range of selection practices

from which to choose. There is no evidence quantifying

how often a particular selection practice was used. For

example, there is no evidence about the number of

positions that were filled temporarily by detailing or

the races of all the employces detailed to a position.

There is evidence that not all person detailed into a

position were eventually promoted into it. Although

there is evidence that African Americans received

proportionately more discipline than Caucasians in

1995-1996, there is no evidence about the number of

disciplined African American employees who applied

for promotion or the number who were promoted.

Further, based on the number of disciplines identified

in the PAT report, only around 3% of employees were

disciplined each year during FY 95 and FY 96. There

is no evidence of the number of job vacancy

announcements that were withdrawn after the human

resources department had forwarded the names of the

top 15 applicants and ties to the selecting official. Nor

is there any evidence about how many withdrawn job

vacancy announcements were followed up by a new

announcement, a new list, and the selection of an

applicant not on the first list for the position.

The Court finds that plaintiffs have sufficiently

identified the practices under attack and demonstrated

by a preponderance of the evidence that the

promotional process at DFAS-CO is incapable of

separation for analysis. The competitive promotion

process at DFAS-CO was complex and consisted of

over 18 separate steps, which permitted the selecting

official to exercise a great deal of discretion when

72a

selecting a candidate. The supervisor could determine

the timing of an opening, limit the area of

consideration, detail an employee into a position, and

withdraw a vacancy when the favored promotion

candidate was not on the certificate and later seek a

new certificate.

Plaintiff Kevin Gray testified that practices such as

returning the certificate (Tr. Vol I at 34:6-35-21);

selecting a grade not announced in the vacancy

announcement (Tr. Vol. I at 35:21 -36:23); limiting the

area of consideration for a given vacancy (Tr. Vol. I at

42:16 - 43:13); and preselecting and detailing

candidates (Tr. Vol. I at 43:14 - 46:19) negatively

impacted African Americans. Plaintiff Willie Phillips

testified that the discipline process was administered

unfairly on the basis of race. Tr. Vol. IV at 33.

Additionally, plaintiffs rely on the conclusions of the

PAT report, which inaicated that African Americans

received approximately 49 percent of all disciplinary

actions in 1995 and 1996 although they comprised only

twenty-one percent of the DFAS-CO population. Mr.

Phillips testified that whether or not an individual

received any awards had a significant impact on

obtaining a promotion. Tr. Vol. IV at 22. He further

testified that he believed that awards were

disproportionately granted to people who he did not

believe should have received them. Tr. Vol. IV at 25.

Performance evaluations were also an important

consideration in determining who was promoted

because points were assigned on the basis of one’s

evaluation. Tr. Vol. [V at 25. Mr. Phillips testified that

he observed white employees being giving better

performance evaluations than their African American

counterparts. Tr. Vol. IV at 27.

73a

Saunders testified that a higher performance

appraisal generated more points in terms of the

promotion evaluation criteria. Tr. Vol. I at

216:17-25-217:1-7. If an individual received an

exceptional rating, he would be more likely to rank

higher on the referral list as a result.

Defendant argues that plaintiffs have a

fundamental misunderstanding of the promotion

system at DFAS-CO. Defendant maintains that

plaintiffs believe that they were discriminated against

if they were referred but not promoted on two or more

occasions. Because each certificate of referral could

include 15 or more names, it is likely that employees

were referred more than once without being promoted.

Defendant also points to Saunders’ testimony to show

that the selecting official was not privy to the “rank” of

the candidates of the certificate of referral. The

certificate of referral listed the names of the most

qualified applicants in alphabetical order.

Consequently, selecting officials sometimes selected

lower-ranked employees over higher-ranked

employees. Finally, defendant maintains that the

certificate of referral does not contain any in-formation

regarding the race of an applicant. Tr. Vol. IV at

114:21-115:11. Unless the selecting official knew the

applicant personally, the selecting official had not way

of knowing the race of a particular individual.

Plaintiffs testified that all or nearly all of the

selecting officials whom they knew were white. On

cross-examination, plaintiffs acknowledged that they

only knew ten to twenty selecting officials out of the

hundreds. As a result defendant argues that plaintiffs’

experiences are not indicative of DFAS-CO as a whole.

Furthermore, defendant maintains that African

74a

Americans occupied 26.23% of the supervisory position

although they comprised only 21.89% of the DFAS-CO

population. See Pls. Ex. 1 at 8.

Statistical Evidence.

The Sixth Circuit stated that “the relevant inquiry

is comparing the number of protected group members

benefitting from promotions with the number seeking

them; this figure is then contrasted with the

corresponding ratio for the non-protected group.”

Phillips, 400 F.3d at 399. To perform this analysis, the

following data is needed: (1) the number of promotions

received by protected group members; (2) the number

of protected group members applying for promotions;

(3) the number of promotions received by

non-protected group members; and, (4) the number of

non-protected group members applying for promotions.

As previously discussed, this data is no longer

available, and therefore the parties were forced to

attempt to approximate what that data would have

revealed if were still available.

Defendant argues that his expert conducted the

analysis prescribed by the Sixth Circuit, and plaintiffs’

expert did not. Both experts were able to determine

the number of promotions received by protected group

members and the number of promotions received by

non-protected group members by examining “end of

year snapshots.” If an employee was at a higher grade

in year 2 than he was in year 1, both experts assumed

that the employee was promoted.” See Def. Ex. H at

'’ It is not clear, however, whether the person was the recipient of

a competitive promotion or whether it was the result of an

75a

{1 14. However, each expert took a very different

approach to determine whether promotion policies had

a disparate impact on African Americans. Dr. Claudy

created “constructed pools” of applicants based on

grade, job series, and where possible, time-in-grade.

The constructed pools were an attempt to approximate

the number of protected and non-protected group

members who were likely to apply for promotions. Dr.

Pactzold, on the other hand, compared the total

number of African American employees to the number

of African Americans receiving promotions. She

performed the same calculation for white employees.

For both categories, she divided the number of

employees promoted by the total number of employees,

then compared these “two promotion rates.” The Sixth

Circuit’s remand requires a comparison of the ratio of

the number of protected employees applying for

promotion to the number promoted with the number of

non-protected employees applying for promotion to the

number promoted.

Thatis the protected employce promoted

the protected employees applying

is compared with

the non-protected employee promoted

the non-protected employees applying

automatic grade increase. Several plaintiffs testified that they

either held or sought positions that were, for example, a 7 target

11 position. In this type of a position, the person would

automatically advance from a 7 to a 9 to an 11 as long as the

employee satisfactorily performed their job duties

76a

The problem faced by Dr. Paetzold and Dr. Claudy

is that none of these numbers is available. They

attempted to approximate the numbers. However,

approximating the unavailable data is a very difficult,

if not impossible, task. The critical information

unavailable is staggering. Although a good fix on the

total number of promotions each year can be obtained

from the end of year snapshots by identifying the

employees who were at a higher grade than the

previous year, this case involves only competitive

promotions. An unknown number of employees who

were at a grade level higher than the previous year

received noncompetitive promotions. They were in a

job with a grade progression, e.g., GS 7/9/11, and they

had completed their time in grade at a lower level with

satisfactory job performance and evaluations and had

been “automatically” promoted to the next grade level

in the sequence. Neither plaintiffs nor defendant

offered any evidence on the number of positions with

non-competitive grade progressions or about the

number of such noncompetitive promotions during

1992-1996.

For the competitive promotions, the minimum

qualifications for applying for the promotion are

unknown because the Job Opportunity Announcement

for the positions are unavailable. Further, the number

of employees applying for each competitive vacancy is

unknown as is the race of each applicant.

Dr. Paetzold’s approach makes no attempt to

identify who applied for a promotion. And _ it

assumes—without a supporting explanation—that both

minorities and non-minorities would apply for a given

position at the same percentages as they are

represented in the workforce. In contrast, Dr. Claudy

T7Va

does attempt to identify the members of the protected

class and the mernbers of the non-protected class who

applied for promotion. The flaw in his approach is that

he assumes—without supporting explanation—that al]

persons eligible for promotion—however minimally

qualified—applied for promotion. He also considered an

employee eligible for promotion—after a year in

grade—only if the employee was in the same series as

the job vacancy. But depending on the JOA

qualifications for the position, employees in other

series may have been qualified.

Dr. Paetzold argued that Dr. Claudy did not

directly compare African Americans promoted and

Whites promoted. Tr. Vol. IV at 152-52. Ultimately,

neither approach is particularly helpful in determining

whether African Americans suffered a disparate

impact with respect to promotional opportunities at

DFAS-CO.

Dr. Claudy performed the Multiple Events Exact

Probability Test and analyzed the pools separately and

then aggregated the results. Dr. Claudy testified:

(The Multiple Events Exact Probability Test]

combines the best aspects of both [methods]: the

power of including many selection events in a

single statistical test so that even small

amounts of disparity are detectable, and the

precision of treating pools of similarly situated

employees separately, so that differences in the

selection rates for protected group members and

other employees are not confounded with

differences in selection rates across occupations

and grade levels.

78a

Def’s Exh. H at 30-31. Although the Sixth Circuit

noted that courts have exhibited skepticism toward

constructed pools, defendant maintains that Dr.

Claudy’s use of grade and series to construct applicant

pools does not warrant suspicion and is

distinguishable from Mozee v. Am. Commerce Marine

Ser. Co., 940 F. 2d 1036, 1045 (7" Cir. 1991). See

Phillips, 400 F.3d at 400. In Mozee, the Seventh

Circuit criticized the post-hoc construction of

qualifications that were neither obvious nor far from

uniformly applied. According to defendant, the use of

grade, job series, and to the extent possible, time-in-

grade are obvious, and uniformly applied. They are

critical components of the federal personnel system.

Dr. Claudy used the grade and job series to determine

who might have applied rather than, as the expert

criticized in Mozee did, to justify who was selected. I

agree that grade, job series, and time-in-grade are

reasonable criteria for attempting to reconstruct who

applied for job vacancies. Indeed, given’ the

unavailability of the applicant flow data, they are the

best available indicators of who may have applied for

a job vacancy. Dr. Claudy’s Fisher Exact Test results

are set out in Def.’s Exs. H and I. He did compare

African Americans with non-African Americans, Tr.

Vol. V at 29:17-30, 22, and found no statistical support

for disparate impact Tr. Vol. V at 42:4-7. Defendant

notes that Paetzold recommends constructed pool

analysis in her book.

Dr. Claudy’s constructed pools, however, are both

under- and over-inclusive. This approach does not

account for applicants who applied for positions in a

different series. It also does not differentiate between

competitive promotions and automatic promotions, as

in the case of a 7 target 11 position. It also fails to take

79a

into account that some promotional opportunities had

a limited area of consideration. Some _ potential

applicants were foreclosed from pursuing a particular

promotion, although they may have been within the

grade and series that Dr. Claudy believed to be the

likely pool of candidates, because it was only available

to employees of a particular directorate.

Although defendant maintains that grade and

senies were a significant indicator of who likely

comprised the applicant pool, several plaintiffs

testified that they applied for positions outside of their

series. Some plaintiffs were ultimately successful in

obtaining positions in a different series. Yulander

Edwards testified that she considered _ series

significant. She had been assigned to the 503 series

prior to her arrival in Columbus, and she believed that

being in this series prevented her from advancing. Tr.

Vol. II at 18. Before coming to DFAS, Ms. Edwards

had experience in the 525 and 540 series. She was

eventually promoted to GS 9/501 position in 2001.

Lisa Prater testified that she began working at

DFAS as a GS 4/503. 11-51. She was promoted to a

position in the 510 series. Tr. Vol. IJ at51. Susan

McNeal testified that she was told prior to moving to

Columbus that she was eligible for positions in 540,

525, and 503 series. Mary Phillips testified that as a

GS 5/525, she apphed for positions as a 525 accounting

technician and 501 financial specialist. ‘I'r. Vol. III at

9. Aithough she was in the 503 series, Denise Goudy

testified that she applied for a position in the 501

senies. Tr. Vol. Ill at 111. Willie Phillips testified that

he began at DFAS in the 525 series and now works in

the 501 series. Tr. Vol. IV at 4-5. Mr. Phillips also

worked in the 540 series, Tr. Vol. IV at 12, and he

80a

testified that he applied for at least one position in the

503 series. Tr. Vol. IV at 84. Several plaintiffs also

testified that om occasion they may have applied for a

position a grade lower than their current level given

that some positions had automatic grade increases

that would ultimately lead to grade levels higher than

their current position.

Given these lmitations, the Court finds that

although defendant’s expert's analysis is some

evidence on the issue of disparate impact, it is not

determinative of whether DFAS-CO’s promotional!

practices had a disparate impact on African Americans

from 1992 to 1996.

Plaintiffs argue that Dr. Paetzold presented

statistical proof, within a_ reasonable scientific

certainty, that African American promotion rates were

significantly lower than white promotion rates. Dr.

Paetzold looked at the aggregate promotion rates. She

calculated the “promotion rates” by dividing the

number of African American employees promoted by

the total number of African American employees at

DFAS-CO. She made the same calculation for white

employees. Then she compared the promotion rates for

African American and white employees by using a

binomial test.

Next, Dr. Paetzold evaluated whether any

differences in promotion rates were statistically

significant. Tr. Vol. VI at 14. Dr. Paetzoid considered

a difference to be statistically significant if it was two

or more standard deviations. She found that the

aggregate promotion rate for all African American

employees at DFAS-CO from 1992 to 1996 was 42.23%,

while the rate was 45.87% for Caucasian employees.

8la

Pls. Ex. 24, App. This resulted in a p-value of .013. Tr.

Vol. VI at 16. As a result, Dr. Paetzold concluded that

African American employees were promoted at a

statistically significant lower rate compared to white

employees because the difference in the promotional

rates for African Americans and whites were greater

than two standard deviations.

Defendant argues that Dr. Paetzold had _ to

separately analyze employees by gender to find a

statistical disparity. Paetzold dep. p. 43:4-9. According

to defendant, the aggregate analysis is of no value

because it fails to approximate the number of

candidates who actually sought promotion.

Although Dr. Paetzold conducted one large analysis

that spanned the years from 1992 to 1996, she also

examined the data year by year and found a

statistically significant difference in promotion rates

for African American men and Caucasian men who

were employees at DFAS-CO in 1993.'° The promotion

13

When asked to explain why she examined both race and sex,

Dr. Paetzold stated:

Sometimes it is helpful to look at sex in addition, since sex

was presented in the data file. The difference in the

promotion rates for the two racial groups could be due to

either men or women within the two racial groups, and so

I looked at both African American and white women and

African American and white men. | did not find any

difference for the men during that 1992 to 1996 period, but

I was able to identify that the difference in promotion

rates between African Americans and whites appear to be

as aresult of adifference in promotion rates, in particular,

between African American women and white women.

VI-19. However, Dr. Pactzold’s expert report indicates that she

82a

rate for African American men was 30.77%, while the

rate for Caucasian men was 37.79%. Oct. 22, 2002

Report of Dr. Paetzold, App. 5. According to Dr.

Paetzold, this resulted in a p-value of .023. The next

greatest difference that Dr. Paetzold found was in the

promotion rates of African American women and

Caucasian women in 1994. The promotion rate for

African American women was 26.5%, while the rate for

Caucasian women was 29.6%. Id. This resulted in a

p-value of .06, which Dr. Pactzold termed to be of

“marginal significance” and 1.5 standard deviations

beyond what would be expected in the absence of

discrimination. /d.

Defendant argues that Dr. Paetzold’s analysis is

contrary to the law of disparate impact and common

sense because it mistakenly relies on the assumption

that all employees sought each promotion. Dr.

Paetzold did not calculate the ratio of African

American employees promoted to African American

employees applying for positions and comparing it

with the ratio of non-minorities promoted to the

number of non-minorities applying for positions as

required by the case law. Phillips, 400 F.3d at 399.

Defendant maintains that this assumption is

particularly problematic in the context of the federal

personnel system, which has time-in-grade and series

requirements. Defendant argues that promotional

policies cannot have a disparate impact on employees

who never sought a promotion; therefore, the

appropriate analysis compares the number of

found statistically significant differences for African American

men, rather for African American women

83a

employees receiving promotions to the number of

employees seeking promotions.

Dr. Claudy testified that the single pool approach

employed by Dr. Paetzold is inherently flawed because

it treats all employees as being entirely fungible.

Defendant argues that one disproportionate applicant

pool can skew the entire result, and the single pool

approach results in a correct calculation only where

protected group members make up exactly the same

proportion of each applicant poo! or exactly the same

proportion is selected from every pool. Furthermore,

Dr. Paetzold’s analysis fails to eliminate the most

significant, obvious, and non-discriminatory reasons

for the disparity, job series and grade level.

Defendant criticizes Dr. Paetzold’s methodology.

Defendant argues that Dr. Paetzold failed to explain

her methodology and failed to report the results of

each analysis she conducted. In fact, Dr. Paetzold

could not even recall which analyses she conducted. Tr.

Vol. VI at 32-33. Dr. Paetzold only recalled those

analyses that were included in her report—those

yielding results beneficial to plaintiffs’ case. Dr.

Paetzold only reported those results that were

statistically significant, altnough she did not report all

statistically significant results. As a result of failing to

fully explain her methodology and report the complete

results of her analysis, the Court’s ability to evaluate

her conclusions is severely hampered.

A disparate impact analysis compares “the number

of protected group members benefitting from

promotions with the number seeking them; this figure

is contrasted with the corresponding ratio for the

non-protected group.” Phillips, 400 F.3d at 399. Dr.

84a

Paetzold treated all employees as equally eligible for a

promotion, regardless of their job series or grade level,

so she assumed that all African Americans and all

whites applied for each promotion. That assumption

has no factual basis in the record. Further, it ignores

the basic premise of a prima facie case, that if

qualified African American applicants receive

substantially fewer promotions than the ratio of

qualified African American applicants to the total

qualified applicant pool, there is an inference that race

was a factor in the selection. By failing to account for

job series and grade level, Dr. Paetzold failed to

eliminate the most common nondiscriminatory reasons

for the disparity, such as whether the applicant was

qualified. See Bacon v. Honda of Am. Mfg., Inc., 205

F.R.D. 466, 473 (S.D. Ohio 2001) (“In offering

statistical evidence, the plaintiff must eliminate the

most common nondiscriminatory reasons for the

disparity.”). As defendant argues, even when racial

discrimination is not a factor, a Caucasian employee at

grade 12 is more likely to receive a promotion to grade

13 than an African American employee at grade 4. Dr.

Paetzold’s analysis treats the grade 4 employee as if he

is equally likely as the grade 12 employee to receive

the promotion. Further, she treats a grade 12

employee as equally likely as a grade 4 employee to

receive a job opening at grade 5, even though it would

be a demotion.

Non-Statistical Evidence. In addition to relying on

Dr. Paetzold’s statistical analysis, plaintiffs rely on the

conclusions of the PAT report, statements made by Ida

Faye Groves in a presentation to DFAS headquarters,

the EEO report, and the testimony of plaintiffs to show

that defendant’s promotion practices had a disparate

impact on African Americans. Although the Sixth

85a

Circuit concluded that the PAT report and anecdotal!

evidence provided some evidence of disparate impact,

and, as a result, summary judgment was not

appropriate, defendant maintains that plaintiffs

cannot demonstrate disparate impact by a

preponderance of the evidence based on either

anecdotal evidence or the PAT report.

Plaintiffs maintain that the PAT report concluded

that defendant’s promotional practices created a

disparate impact on African American employees.

Defendant argues that the PAT Report is inherently

unreliable because it was a perception report that was

not designed to collect statistically meaningful data.

Tr. Vol. I at 177:16-25-178:1-5. The PAT team did not

conduct a statistical analysis of the collected data, and

the data was not collected from a random pool of

employees. Instead, the PAT Report was based on the

perceptions of a non-random, potentially biased, pool

of employees.

The PAT Report is not sufficient to support

plaintiffs’ allegations of disparate impact. The PAT

analyzed merit promotion and EEO statistical data for

fiscal years 1995 and 1996, as well as interviewed

employees concerning their perceptions ofemployment

practices at DFAS-CO. One of the issues examined by

the PAT was: “Are only non-minorities advancing to

the GS-11 level and above?” They based their analysis

on the Bureau of Labor’s 1990 Census of the civilian

labor force of the Columbus Metropolitan Area, which

reported that African Americans make up 10.7% of the

86a

civilian labor force in the Columbus Metropolitan

Area.“

The PAT found that, in comparison to the civilian

labor force of the Columbus Metropolitan Area, African

American employees at DFAS-CO were over

-represented in grades 5-8 (22.1% of the DFAS-CO

workforce)), grades 9-12 (20.0%), and grades 13-15

(13.2%). See PAT Report, at 28-31, 114. The PAT found

that, for fiscal years 1995 and 1996, African American

employees at DFAS-CO were selected for promotion

21.6% of the time 1n grades 5-8, 17.8% of the time in

grades 9-12, and 6.6% of the time in grades 13-15. See

PAT Report, at 28-31.

The difficulty with the PAT’s promotions data is

that it exists in a vacuum. The PAT did not perform a

statistical comparison of the promotions data with an

expected rate of promotion for African Americans at

DFAS-CO. The expected rate that the PAT references

is the percentage of African Americans in the civilian

labor force in the Columbus Metropolitan Area. This is

too broad, for it rests on the assumption that the

percentage of African Americans qualified for grades

13 through 15 is the same as the overall percentage of

African Americans in the civilian labor force in

Columbus--an assumption which plaintiffs have offered

no evidence to substantiate. See Wards Cove Packing

Company, Inc. v. Atonio, et al., 490 U.S. 642, 650-51

(1989) (“It is such a comparison—between the racial

‘* The PAT report analyzed African American males and African

American females separately, a distinction not relevant to the

case at hand. Therefore, this Opinion and Order uses the PAT’s

combined results for all African Americans

Siva

composition of the qualified persons in the labor

market and the persons holding at-issue jobs—that

generally forms the proper basis for the initial inquiry

in a disparte-impact case.”); Moore v. Hughes

Helicopters, Inc., 708 F.2d 475, 483-84 (9th Cir. 1983)

(“{Plaintiff] has failed to establish that there were any

black female employees qualified for the positions in

Labor Grades 15-20.”). The actual number of African

American employees at DFAS-CO qualified for a given

promotion could well have been greater than, or less

than, the number of African Americans in the

Columbus civilian labor force. Like Dr. Paetzold, the

PAT did not compare the actual promotion rates of

African Americans with an expected rate that

accounted for minimal qualifications. Further, the PAT

did not conduct a statistical analysis to determine how

great the disparity was in terms of standard

deviations. Simply put, the PAT Report does not

contain reliable evidence that African Americans at

DFAS-CO suffered a disparate impact in promotions

With respect to the PAT report, the Court concludes

that this type of survey of employee perceptions and

experiences, standing alone, cannot show disparate

impact. This survey was neither random nor of a large

enough scale to demonstrate that its findings were

reliable and valid. Many of the plaintiffs acknowledge

their participation in the survey, either through direct

or indirect communications with representatives of

Togar. The EEO report indicates that African

Americans were disproportionately represented in the

survey performed by the PAT team. Pls. Ex. 1, at 6.

Despite the Court’s belief that plaintiffs’ participation

was based on their sincere belief that DFAS-CO’s

promotional policies resulted in qualified African

SSa

Americans being denied promotion, this manner of

collecting data is susceptible to skewed results.

Plaintiffs also argue that talking points from Ida

Faye Groves’s presentation at DFAS headquarters in

Arlington, Virginia on February 24, 1997 is an

admission that minorities continued to be promoted at

lower rates than non-minorities, especially at the

higher grade levels. Plaintiffs further argue that the

July 24, 1997 EEO report presents compelling

evidence of ethnic preferences among’ white

supervisors by promoting white employees. Plaintiffs

cannot meet their burden of proof by relying on the

non-statistical evidence of record alone. Plaintiffs

mischaracterize Ms. Groves’s presentation to DFAS

headquarters. There is nothing in the records which

suggest that Ms. Groves was doing anything other

than communicating the conclusions of the PAT tearm

Finally, plaintiffs argue that their trial testimony

demonstrates that promotion policies had a disparate

impact on African Americans. This testimony, without

the benefit of supporting statistical analysis, does not

show that African Americans as a group were

negatively impacted. In Wards Cove Packing Co., Inc

v. Atonio, 490 U.S. 642 (1989), the United States

Supreme Court held that “the proper basis for the

initial inquiry in a disparate-impact case” is a

comparison “between the racial composition of the

qualified persons in the labor market and the persons

holding at-issue jobs.” 490 U.S. at 650-51 (emphasis

added). Plaintiff must demonstrate a_ statistical

disparity based on “the racial composition of [the

at-issue jobs] and the racial composition of the

qualified .. . population in the relevant labor market.”

Id. at 650 (quoting Hazelwood School Dist. v. United

89a

States, 433 U.S. 299, 308 (1977)). Unfortunately,

plaintiffs’ testimony does not provide this Court with

the evidence necessary to support a finding olf

disparate impact

Plaintiffs were all hard working DFAS-CO

employees who reasonably might have expected more

rapid promotion during 1992-1996 than they got. They

also reasonably believed that a few promotions singled

out in their testimony suggested favoritism or

preselection. They also had reason to be concerned

about whether detailing, changing the series of a

position, limiting applications to one directorate,

performance awards, and performance ratings might

have adversely affected their opportunities for

promotion.'” However, there is no evidence from which

the Court could conclude by a preponderance that any

of the plaintiffs were denied promotion to a particular

position because of their race. As to each of the job

vacancy announcements about which the individual

plaintiffs testified, there is no evidence of the position

description and job qualifications, the race of the top

fifteen applicants and ties, or their qualifications. This

is a disparate impact case, so the issue is whether

plaintiffs have presented statistical evidence which,

when weighed with their testimony and the relevant

facts in the PAT report, prove by a preponderance that

the employment process they have identified had a

disparate impact on African Americans

On balance, I find Dr. Claudy’s expert opinion that

there is not. statistical evidence that African

There is no evidence that they personally were held back by

unwarranted disciplins

O0a

Americans who applied for promotion were selected

less often than chance at least as credible and

trustworthy as Dr. Paetzold’s opinion to the contrary.

Both experts’ opinions are entitled to little weight

because both were unable to compare the ratios of the

number of African American applicants promoted to

the number of African American applicants with the

ratio of the number of non-minority applicants. Dr.

Claudy—unlike Dr. Paetzold—at least made an attempt

to re-construct the applicant pools, but he made a

number of assumptions about who would likely have

applied that are not supported by the _ record.

Moreover, he assumed every promotion was a

competitive promotion, while the record demonstrates

that anumber of promotions were non-competitive. Dr.

Pactzold’s opinion is unreliable because she assumed

every employee applied for every promotion. She also

assumed every promotion was a competitive one.

Further, she could not tell the Court all of the analysis

of the data she performed, and she reported only some

of the statistically significant results she found

Focusing solely on the statistical evidence, I find the

evidence to be evenly balanced, so that plaintiffs have

failed to meet their burden of proving that the ratio of

African Americans selected for promotions to those

applying statistically significantly smaller than the

ratio for whites. Weighing all the evidence admitted

into the record, I find that it is no better than evenly

balanced or the issue of whether the employment

practices identified by plaintiffs had a disparate

impact on qualified African Americans who applied for

promotion. Consequently, I find that plaintiffs have

failed to prove by a preponderance of the evidence that

African Americans were disparately impacted by the

employment practices they identified.

Gla

Accordingly, the Clerk of Court is DIRECTED to

enter JUDGMENT for defendant

s/ Mark R. Abel

United States Magistrate Judg