Harlem, New York City, 1964-1965, undated - 1 of 3

    BRIEF FOR FEDERAL CROSS/APPELLEES
REPLY BRIEF FOR FEDERAL APPELLANTS

UNITED STATES COURT OF APPEALS 
FOR THE DISTRICT OF COLUMBIA CIRCUIT

No.
No.

83-1678
83-1679

DENNIS HARRISON, et al. , Appellees-Cross/Appellants
V.

ELIZABETH DOLE, et al. , Appellants-Cross/Appellees

ON CROSS-APPEALS FROM THE UNITED STATES DISTRICT COURT 
FOR THE DISTRICT OF COLUMBIA

STANLEY S. HARRIS, 
United States Attorney.

ROYCE C. LAMBERTH,
R. CRAIG LAWRENCE,
JOHN H.E. BAYLY, JR.,
STUART H. NEWBERGER,
Assistant United States Attorneys.

OF COUNSEL:
TIMOTHY SHEA
Office of the Chief Counsel
Maritime Administration
U.S. Department of Transportation

C. A. No. 79-1816

Date: November 18, 1983



I N D E X

Page

INTRODUCTION ...........................................  1

Plaintiff Class' Cross-Appeal ....................  1
MarAd's Appeal ..........................  . . . .  2

FACTUAL BACKGROUND ..................................... 3
A. Plaintiff Spencer's Administrative Claim . . .  3
B. The "Destruction of Records" Claim ..........  4
C. Plaintiffs' Competitive Selection Analysis . . 5
D. MarAd's "Unvalidated Selection" System . . . .  5
E. The Anecdotal Evidence ......................  8
F. MarAd's Statistical Analyses ................  8
G. Plaintiffs' Statistical Analyses ............  13

ARGUMENT...............................................  15
I. Plaintiffs' Cross-Appeal .....................  15

A. The District Court Properly Held That
MarAd Has Not Discriminated Against The 
Certified, Compound Class ..............  15

B. The District Court Properly Held That 
MarAd Has Not Discriminated Against
Women Employees On the Basis of Sex . . .  17

C. The District Court Properly Refused to
Review Plaintiffs' Post-Trial Statistical 
A n a l y s i s ............................... 18

D. The District Court Properly Held That 
the Affirmative Action Provisions of 
42 U.S.C. §2000e-16 Do Not Create A
Private Right Of Action ................  20

E. The District Court Properly Limited 
The Class Relief to the Extent
Challenged by Plaintiffs' ..............  -23



II. MarAd's Appeal.................................  26
A. The District Court Erred In Certifying The

Compound, Across-the-Board Class . . . .  26
B. The District Court Erred In Finding

Partial Class (Race) Liability ...........  30
C. The District Court's Relief Order,

Requiring A Validation Study, was
Erroneous . . . .. .......................  32

CONCLUSION...............................................  34

TABLE OF CASES
Baca v. Butz, 394 F. Supp. 888 (D.N.M. 1975)..........  22
Borrell v. U.S. Intern. Communications Agency.
1T5TT.2d W L  (D.C. Cir. 1982)— . . .  ..........  20

Brown v. GSA, 425 U.S. 820 (1977) ....................  31
Bush v. Lucas, U.S. , 103 S. Ct.

2104 ( T O T  TT". . . T T ~ ........................... 22
Cannon v. University of Chicago, 441 U.S. 677

OT79) .— r ...........................................  21
Council of the Blind of Del. City Valley v. Regan,

709 F.2d 1521 (D.C. Cir. 1983) (en banc) ............  20
Court v. Ash, 422 U.S. 66 (1975)........................  20
Crown, Cork & Seal Co. v. Parker, U.S. ,

S. Ct. , 1T> L. Ed. 2d 628~(T283) .~TT . . . .  23
East Texas Motor Freight System, Inc. v. Rodriguez,

431 U.S. 295 (1978) . ............... . . . . . .  26
*EEOC v. Federal Reserve Bank of Richmond, 698 F.2d 633 

(4th Cir. 1983), cert, granted, sub, nom., Cooper v.
Federal Reserve Bank of Richmond, 52 U.S.L.W. 3342 
(October 31, 1983)(No. 83-185)....................  25, 28

^General Telephone Co. of the Southwest v. Falcon,
457 U.S. 148 (1982) ! ! ! ! ! ! ! ! ................  Passim

Lawler v. Alexander, 698 F.2d 439 (11th Cir. 1983) . . .  28

ZJ Cases chiefly relied upon.

Page

-ii-



McKenzie v. Sawyer, 684 F.2d 62 (D.C. Cir. 1982) . . . .  31
Mervin v. F.T.C., 591 F.2d 821 (D.C. Cir. 1978) . . . .  25
Payne v. Travenol Laboratories, Inc., 673 F.2d

798 (5th Cir. 1982), cert, denied, U.S. ,
103 S. Ct. 451 (1982)............................... 30

*Pouncy v. Prudential Ins. Co. of America, 668 F.2d
795 (5th Cir. 1982) ................................. 31

Shivers v. Landrieu 674 F.2d 906 (D.C. Cir. 1981) . . .  21
Talev v. Reinhardt, 662 F.2d 888 (D.C. Cir. 1981) . . .  31
Toney v. Block, 705 F.2d 1364 (D.C. Cir. 1983) ........  29
*Trout v. Lehman, 702 F.2d 1094 (D.C. Cir. 1983) cert.
pet. pending, 52 U.S.L.W. 3387 (No. 83-706) ........  Passim

*Valentino v. U.S. Postal Service, 674 F.2d 56
(D.C. Cir. 1982)..................................... Passim

Zipes v. Trans World Airlines, Inc., 455 U.S. 385
(1982)...............................................  23

OTHER AUTHORITIES
42 U.S.C. § 2000e—16 ................................... Passim
5 C.F.R. § 713.604 (1977) ............................. 29
5 C.F.R. § 1613.214(a) ................................. 25
5 C.F.R. § 1613.602 ................................. 25
Administrative Procedure Act, 5 U.S.C. § 701 et seq. . . 21
Rule 23, F.R. Civ. P.................................... Passim
OPM X-118 Standards ................................... Passim

Page

-iii-



BRIEF FOR FEDERAL CROSS/APPELLEES
REPLY BRIEF FOR FEDERAL APPELLANTS

UNITED STATES COURT OF APPEALS 
FOR THE DISTRICT OF COLUMBIA CIRCUIT

No. 83-1678 
No. 83-1679

DENNIS HARRISON, et al. , Appellees-Cross/Appellants
v.

ELIZABETH DOLE, et al., Appellants-Cross/Appellees

ON CROSS-APPEALS FROM THE UNITED STATES DISTRICT COURT 
FOR THE DISTRICT OF COLUMBIA

INTRODUCTION

Federal cross-appellees (federal appellants) —  ̂ hereby 
respond to the issues raised on the cross-appeal and, in addition, 
reply to appellees' (cross-appellants) — brief filed in the 
present case.

The Plaintiff Class' Cross-Appeal 
In the pending cross-appeal, plaintiffs below challenge 

several adverse holdings of the District Court. More particularly,

T7 As with their opening brief, federal cross-appellees will 
refer to themselves as "MarAd."
2/ "Plaintiffs" or "Class members".



the issues on cross-appeal include: (1) whether the District
Court properly found that MarAd had not discriminated in its 
employment practices on the basis of sex or race either against
(a) a compound, across-the-board class of all non-white males, or
(b) that portion of the compound class consisting of women; (2) 
whether the District Court properly refused to consider a statisti­
cal report prepared by plaintiffs' experts after the trial on the 
merits had been completed; (3) whether the District Court properly 
held that the affirmative action provisions of 42 U.S.C. § 2000e-16 
do not provide federal employees with an independent, private 
cause of action against the employing agency; and (4) whether the 
District Court properly denied certain class relief. As dis­
cussed below, the District Court correctly ruled on all the 
issues raised in the cross-appeal and its judgment should be 
affirmed to that extent.

MarAd's Appeal
As noted in MarAd's opening brief, the District Court 

incorrectly ruled on several critical issues. These issues 
include: (1) the District Court's error in certifying an across-
the-board, compound class in contravention of the strict require­
ments of Rule 23, Federal Rules of Civil Procedure and the 
Supreme Court's ruling in General Telephone Co. of the Southwest 
v. Falcon, 457 U.S. 148 (1982) (Falcon); (2) the District Court's 
error in bifurcating the compound class,after trial, notwith­
standing its finding on the merits that MarAd had not discrimi­
nated against the certified, compound class; (3) the District 
Court's error in ordering agency-wide validation of MarAd.'s

2



promotion criteria (a) where plaintiffs had concededly failed to 
prove intentional, disparate treatment of black employees in 
promotions, and (b) where MarAd (like all federal agencies) was 
required to follow the promotion criteria set down by the Office 
of Personnel Management, (a non-party) in the OPM X-118 Standards; 
and (4) the District Court's overbroad inclusion of black class 
members in its order providing individual class member relief.
To that extent, the District Court's findings should be reversed 
and remanded.

FACTUAL BACKGROUND
These cross-appeals arise out of District Judge Louis B. 

Oberdorfer's findings in a Title VII class action wherein all 
black and women applicants and employees of the Maritime Admini­
stration challenged that agency's employment practices and 
policies. After certifying an "across-the-board" class, conducting 
extensive pretrial proceedings, and holding a two-week trial, the 
District Court found that MarAd had not engaged in discriminatory 
practices against women but had against blacks. Thereafter, 
various types of relief issued. MarAd's opening brief extensively 
discussed the factual background of the present appeals, (Brief 
pp. 9-42). Consequently, MarAds responds to plaintiffs' brief's 
numerous factual statements only as they require a response.

A. Plaintiff Spencer's Administrative Claim
Plaintiffs' assertion that Ms. Spencer was promoted after 

the class claim was filed apparently suggests that it was only 
because of the claim that she was promoted. On the contrary,

3



when Ms. Spencer (and other employees) complained about a posi­
tion selection in March 1977 MarAd agreed, in writing, to re­
advertise the position and designate a new selecting official. 
This position was readvertised on April 13, 1977 and closed in 
May 1977. Thus, Ms. Spencer's selection in August 1977 was the 
result of the April 1977 compromise wherein the position was 
readvertised. (DX 98, 110, and 132 Is 357-364). Because this 
position was not filled until Ms. Spencer was selected, she could 
not have any individual backpay claim relating to the present case.

In response to Ms. Spencer's "separate claim" regarding 
failure to re-classify her position in early 1977, MarAd recom­
mended that she be promoted by reclassification but, under 
established agency rules, had to refer the recommendation to the 
Department of Commerce. The Department denied the request 
shortly before Ms. Spencer was selected for the readvertised 
position in August 1977. DX 130, Is 152-154.

B. The "Destruction of Records" Claim 
Plaintiffs suggest that MarAd was engaged in a careless and 

broad effort to dispose of relevant employee records. However, 
the only relevant records that were unavailable were the SF-171 
application forms which MarAd continued to return to rejected 
applicants. This practice had little or no impact on the reliabi­
lity of either party's statistical analysis. —

Plaintiffs' claim on this "issue" is, therefore, unsupported 
by the record.

J7 Plaintiffs speculate that, if only they had available more 
information on the race of applicants, the results of their
(FOOTNOTE CONTINUED ON NEXT PAGE)

4



C. Plaintiffs' Competitive Selection 
Analysis_______________

Plaintiffs' treatment of competitive selections confuses the 
process. Their analysis describes MarAd's decisions on assignment 
of job series as involving substantial employer discretion and 
judgment. However, assigning the job series is the process of 
associating, i.e.; an accountant position with the appropriate 
federal job series, i.e. , series 510. Although there was 
evidence bearing on the degree of discretion involved in applying 
the OPM X-118 standards (see DX-129; 197 and 193; Op. at 7, 8 and 
35) (JA 204, 205 and 235), there was no testimony suggesting that 
the assignment of the series involved any substantial discretion. 
To the contrary, there was substantial evidence that these 
decisions are subject to detailed, government-wide guidance from 
OPM. See DX-130; Is 95-105.

3/ (FOOTNONTE CONTINUED FROM PREVIOUS PAGE)
statistical analyses might change to their advantage. (Plain­
tiffs' Brief p. 16) Yet, there was no testimony that the avail­
able data was insufficient to permit reliable conclusions. 
MarAd's expert, Dr. Michelson, specifically undertook analyses to 
measure the sensitivity of the data and fairly adopted conserva­
tive methodologies. (DX 120). Plaintiffs seek to blame MarAd's 
temporary continuance of its practice of returning Form SF 171's 
to applicants for this information loss. However, plaintiffs 
make no attempt to quantify the significance of the lost infor­
mation. Of course this would be quite a task, given that the SF 
171 Forms do not list race identification. Because all of the 
returned SF-171's would be from non-MarAd applicants, race 
information would have been obtainable from the only other 
available source, OPM (DX 120 p. 10 and 18) . However, cross­
checking applicants of unknown race with OPM records yielded 
racial identification for only a small fraction of applicants. Id.

5



In addition, plaintiffs assert that the number of candidates 
who may be referred to a selecting official is nearly unlimited. 
However, this is clearly not the case. Certificates usually 
include three to five candidates, with ten the maximum number 
that may be referred when meaningful distinctions cannot be made. 
(See DX 104, MAO 730-335 section 10 (1979); MAO 730-733 section
11 (1969)). A larger number of candidates may be referred when
multiple selections are made from the same certificate. (See, 
e.g., PX 180) .

D. MarAd1s "Unvalidated" Selection System
Plaintiffs' suggest that MarAd somehow was required to 

undertake a validation of its selection system prior to trial. 
This claim is utterly without support. Up until January 1980, 
collection of adverse impact data from applicants was prohibited 
by 0PM. Once the prohibition was removed, MarAd (along with the 
Department of Commerce), proceeded to develop a data collection 
and storage system. This system was developed and put into 
operation by late 1980. (DX-129, 2s 231-233).

In the same vein, MarAd offered extensive testimony about 
the job relatedness of its selection systems and its conformity 
with 0PM requirements. This evidence showed that, at the initial 
stage, the basic (or "minimum") qualifications were taken from 
the OPM X-118 and, since 1977, MarAd had added basic qualifica­
tions beyond the X-118 on only one occasion. (DX 129, 2 193).
At the next stages, (rating and ranking), the primary criteria 
were always training, education and prior experience. Prior to 
1979, awards and appraisals were given some consideration but

6



afterwards, not at all. DX-129 5s 198-200. These criteria are 
job related by definition. Because the data collection system 
was the first step in any formal evaluation, that process had not 
advanced to a validation by the time of trial (early 1982).

In order to bolster their unsupported suspicion that the 
MarAd personnel system was subject to supervisory manipulation, 
plaintiffs exaggerate the number and significance of positions 
they describe as "unique" (Brief p. 12) The District Court, 
however, did not find that there were a substantial number of 
positions at MarAd that were unique to it but found, instead, 
that there were "a number of unique or one of a kind jobs." (Op. 
7) (JA 204) Neither plaintiffs' nor the Court ever identify 
these positions. MarAd identified them as those classified as 
examiners in the 301 job series (DX 130 1 139) . Inspection of 
PX-1 indicates that this is not a substantial number of positions.

Plaintiffs then confuse the differing nature of the two sets 
of standards that are applied to these "unique" positions. 
Classification standards for these jobs are constructed by a 
cross-comparison of closely related functions. (DX 130, 5 141). 
On the other hand, qualification standards for the 301 series 
have requirements that call for general and specialized experi­
ence. (DX 129, 5 192).

Contrary to plaintiffs' assertion (p. 12), MarAd has never
contended that career-ladder positions must be "professional". 
However, at the GS levels upon which plaintiffs' statistical 
analysis focused (GS 7-15) , all career-ladder positions were

7



professional. The District Court found that "[wjhile nonpro­
fessional jobs at MarAd could possibly be designated career 
ladder, there is no evidence that failure to do so is in any way 
discriminatory." (Op. 29 and 30) (JA 229-230).

Finally, defendants note that plaintiffs' claim of inade­
quate training (p. 14) fails to recognize that OPM requires
training to be job related. DX-131, fs 276-278. -

E. The Anecdotal Evidence
Plaintiffs' discussion of the anecdotal evidence overstates 

the record. (pp. 14-15). The only testimony offered by plaintiffs
relating to individual discrimination was proffered by nine class

5 /members (other than the three named plaintiffs). —

4/ Plaintiffs devote undue attention to awards data with little 
explanation for its prominence. They assert that MarAd's awards 
data indicates a significant disproportion in favor of white 
males. This is in error. In all cash awards, no significant 
disproportion was observed. (DX-120, Appendix D). MarAd admitted 
that a disproportion of honorary awards (bronze medals) went to 
white males but explained that such awards carried no consequences 
in compensation. Because honorary awards are made only for 
significant contributions to agency programs over a long period 
of time, the type of position held bears on the award distribu­
tion. (DX-104, MAO 740-451, section b). Plaintiffs seek to 
argue that their awards totals should be adopted. Defendants 
objected to receipt of plaintiffs' awards totals because it was 
unsponsored by a competent witness. Conversely, MarAd's awards 
totals, sponsored by Maxine Anderson (the personnel specialist 
who maintained the data), is more reliable and must be accepted. 
See Tr. 664-676.
5/ These witnesses were John Blackburn; Gerald Brown; Mary 
Duckett; Joan Forman; Vontell Frost; Sharon Howard; Freddie 
Johnson; Rona LaPrade & James White. Ms. Forman and Ms. Frost 
did not testify —  facts relating to their claims were admitted 
by MarAd. Plaintiffs had offered testimony from Mary Arter, 
Jessie Fernanders and Joyce Campbell but these persons did not 
assert that they were victims of discrimination. See Tr. 267- 
404, 480-543. See also Plaintiffs' Substituted Pre-Trial Brief, 
pp. 6, 9, 12, 72, 76 and 116-118. (R 151).

8



For its part, MarAd vigorously opposed each of these asser­
tions of discrimination. —^

F. MarAd1s Statistical Analyses 
MarAd offered two statistical studies that were recognized 

by the District Court to be "more reliable" than plaintiffs'. 
(Op. 28) (JA 228). One of its studies was a survival analysis of 
non-competitive promotions that separately compared time-in-grade 
(tenure) by race, sex and white males against others for career 
ladder and non-career ladder jobs. In addition, a strict joint 
test was undertaken that sought to combine all the positions 
while preserving the differences. (DX-119).

For competitive selections MarAd's expert, Dr. Michelson, 
offered a multiple pools analysis of over 400 vacancy announce­
ments over the period from 1977 to early 1981. (DX-120 p.
17). — The advantage of the multiple pools analysis was that it 
correctly identified the rejected applicants for each vacancy as

6/ See Defendants' Post-Trial Brief P. 11-63-75) (R 205); Tr. pp. 631-770, 776-817, 881-894.
l_l While MarAd's expert did recognize that a multiple pools 
analysis might not be appropriate for candidates competing 
against a standard (as is the case at the qualification stage) , 
this does not compel the concludion that plaintiffs aggregation 
of all qualification decisions was appropriate. After all, 
applicants were competing against scores of standards. See DX-3. 
Plaintiffs, of course, made no attempt to allocate qualification 
success rates either by series or job type. In any event, where 
applicants are competing against each other at the ranking and 
certification stage, the use of the multiple pools analysis is 
wholly appropriate. (Op. 22 and 28) (JA 222 and 228)

9



only those who applied for the vacancy. Statistical techniques, 
such as plaintiffs, that lump together applicants for different 
vacancy announcements result in treating disappointed applicants 
from all announcements as "rejects"-- where the racial or sexual 
mix of the applicants varies from announcement to announcement, 
the difference in the results are substantial. (DX-120 Appendix 
B). The multiple pools analysis included breakdowns for clerical, 
non-clerical, low-level (GS-12 and below), career-ladder, and 
upper-level positions. Dr. Michelson's analysis separately 
compared the "success" rates of applicants by race, sex and white 
males against others, analyzing: (1) applicants vs. selectees;
(2) qualifieds vs. selectees; (3) certifieds v. selectees. 
After all of these permutations, the only area where any results 
remotely favorable to plaintiffs emerged was the disproportionate 
failure of black clericals to meet minimum qualifications. (DX 
120 pp. 60-63).

Moreover, the survival analysis of career ladder promotions 
(which are non-competitive) indicated no statistically signifi­
cant differences for any race/sex subgroup. The data indicated 
that, with only one exception, the differences in median time to 
promotion ranged from 12.6 to 13.3 months for grades 7, 9 and 11 
regarding any race/sex subgroup. In a white males v. others, the 
median times to promotion differed by only .2 months for any 
grades. (DX 119 pp. 13-20).

To meet plaintiffs' claim that MarAd unduly disaggregated 
the data, Dr. Wyant also performed a joint test where the pro­
fessional (e.g., career-ladder) and non-professional tenures were

10



all included and adjusted for their being in different groups. 
This analysis treated all positions within the two respective 
groups (professional v. non-professional) as equally promotable. 
Because of its complexity and its conservative assumptions, this 
strict analysis established, in effect, the outer perimeter for 
statistical speculation. The results of this analysis showed an 
unlikely, but still not statistically significant, .05 level. 
(DX. 119 p. 33-36). ^

Dr. Michelson's analysis of the competitive promotions 
strongly rebutted plaintiffs' claims. The analysis of the 
success rate of applicants by sex for grades GS-13 through 18 
showed that the probability of selecting the observed number of 
females (or fewer) from among those determined to be eligible —  ̂
was .99 -- indicating a statistically significant overselection 
of women. For GS-12s and below, Dr. Michelson likewise found 
that the probability of selecting the observed number of females 
(or fewer) was again .99. Similarly, Dr. Michelson found that 
the success rate of eligible females competing for entry into 
career ladder positions was .90. (DX-120 pp. 33-37).

17 The probabilities are expressed here in hundredths. Low 
probabilities indicate underselection. The .50 figure would 
represent the "expected" value. Figures over .50 indicate 
overselection. Plaintiffs, who are reduced to arguing that this 
analysis "comes within a hair" of showing statistical significance, ignore the severity of the test.
9/ Dr. Michelson testified that the relevant comparison was 
between qualified applicants vs. selectees because this compared 
applicants who were interested and minimally qualified. See 
generally Tr. 896-993. However, he also displayed data comparing 
applicants and selectees as well. Id.

11



This data firmly refuted any notion of sex discrimination. The
District Court agreed, noting that

[a]s the analysis of Dr. Michelson shows 
women have, if anything, been selected in 
disproportionately greater numbers in competi­
tive appointments at MarAd. Though some of 
this overselection of women disappears if the 
apparent bias in favor of applicants from 
within MarAd is factored in [footnote omit­
ted) , there remains a slight, statistically 
insignificant preference for women.

(emphasis in original) Op. 33. (JA 223)
Dr. Michelson's analysis of the success rate of black

eligibles at GS-13 and above arrived at a .31 level, a figure
that is fully consistent with an unbiased selection process. The
success rate of black eligibles at GS-12 and below was determined
to be effectively zero owing entirely to underselection of blacks
in clerical positions. Thus, the probability of selecting the
observed number (or fewer) of black clericals was .01. However,
for blacks competing for career ladder positions the figure was
.57 and for all low level announcements (excluding clericals) it
was .22. Because of the number of clerical selections, its
results predominated the low level announcement results. Dr.
Michelson's review of the clerical selections of blacks indicated
that they were underselected by 12 positions to the benefit of
white females and non-black minorities. (DX-120, pp. 40-55).
Thus, only one isolated, statistically significant outcome was
observed —  blacks competing for clerical positions.

Finally, Dr. Michelson also performed three white males
against others analyses which yielded results quite favorable to
MarAd. The first analysis yielded probabilities of .19 for GS-1

12



through 12 and .97 for grades GS 13 through 18 for the success 
rates of non-white male eligibles using only applicants whose 
race was known. The second analysis assigned race to males of 
unknown race in proportion to the race composition of the rejected 
applicants. This yielded probabilities of .68 for GS-1 through 
12 and .96 for GS-13 and above. (DX-120 pp. 58-71).

Finally, Dr. Michelson conducted a third white males against 
others analysis wherein he "increased the representation of 
blacks among males of unknown race by 10%." Even this indicated 
no underselection of non-white males. (See pp. 15-17 of Attach­
ment 1 to Defendants Post-Trial Brief) (R 205). — ^

G. Plaintiffs' Statistical Analysis
Plaintiffs' statistical data was greatly inferior to MarAd's 

for one basic reason: they carelessly lumped all MarAd Head­
quarters employees together without making any distinction 
between job classifications. It is undisputed that MarAd has 
many different types of professional and non-professional posi­
tions and that basic parameters of the job types vary greatly. 
Specifically, clericals and administrative positions usually 
range from GS-2 entry levels to GS -7 full performance levels, 
whereas professionals may enter at the GS-5 level and advance to

10/ Plaintiffs argue that the District Court was sympathetic to 
their argument that Dr. Michelson's multiple pools analysis was 
weakened because applicant pools with only one racial group of 
applicants would not influence the results. The District Court 
recognized that Dr. Michelson's additional analysis (imputing 
race to the race-unknown applicants based on and thereby increas­
ing the numbers of racially mixed pools), improved the results 
for MarAd. Op. 23 (JA 223) (DX-120 pp. 51 and 53). Thus, use of 
race known data was demonstrated to be more conservative and reliable.

13



the full performance level at GS-12. (DX-130 2 128) . The
District Court's observation that "it would be irrational to 
assume equal promotability" (Op. 29) (JA 229) among the posi­
tions examined by plaintiffs is plainly correct.

It is notable that when plaintiffs' studies examined more 
homogeneous positions the results were favorable to MarAd. For 
example, career ladder promotions are available only up to GS-12. 
Plaintiffs' analysis of tenure and frequencies of promotions 
(competitive and noncompetitive together) to GS-13 and 14 show that 
non-white males advanced faster and more frequently than white males 
though not to a statistically significant extent. (DX-4 p. 10).

The only analysis presented by plaintiffs that purported to 
address qualifications were its regressions analyses. —  ̂ Plain­
tiffs' regression analyses, like the rest of their analytical 
data, never factored in occupational classifications. As MarAd 
pointed out and the District Court found, regression analyses 
such as the ones proffered by plaintiffs "show[ ] little about 
how the agency became the way it is" and concluded that they

11/ These regressions used information from MarAd personnel 
computer tapes, supplied pursuant to defendants' Answers to 
Plaintiffs' Fourth Interrogatories in June 1980 (later updated), 
to attempt to identify the impact of race/sex characteristics on 
salary. Despite plaintiffs' ability to prepare regression 
analyses early in the case their first did not appear until one 
month before trial. (See PX 4 (printouts); DX-120 Appendix C, p. 
20; Deposition of J. Van Ryzin (R 202)).

14



"tell[ ] little about whether this situation came about as a 
result of actionable discrimination." (Op. 30) (JA 230); (DX- 
Appendix C, p. 11-14). — ^

ARGUMENT
I. Plaintiffs' Cross-Appeal
A. The District Court Properly Held That 

MarAd Had Not Discriminated Against 
the Certified, Compound Class________

As more fully discussed in MarAd's opening brief (pp. 34-37, 
68-71, 74-75, 78-82), the District Court rejected the class members' 
allegations of compound discrimination -- the cold numbers pur­
portedly reflecting the "white-maleness" of MarAd's hiring and 
promotional system were simply not found to have any discrimina­
tory basis. In challenging this finding, the plaintiff class 
again recites the same statistical analyses —  reviewed and 
rejected by the District Court —  which purport to prove that 
MarAd discriminated in favor of white males. (Plaintiffs' brief, 
at pp. 17-33). However, the District Court's findings on this 
issue are unassailable.

12/ The only analysis which purported to account for job clas­
sifications was plaintiffs' untimely (and rejected) post-trial 
submission. However, this last analysis merely identified all 
positions that plaintiffs' expert identified as requiring only 
"specific (rather than general) educational degrees or courses." 
Affidavit of J. Van Ryzin dated April 20, 1982, at p. 4 (R 334). 
It was further rife with fatal defects —  job series with specific 
educational requirements (e.q., accountant-series 510; operational 
research analyst-series 1515) were ommitted while jobs without 
such requirements (e.g., engineering techician-series 802; 
statistical assistant-series 1531) were misidentified as requiring specific education. Id.; DX-3.

15



First, the statistical evidence introduced at trial clearly 
demonstrated that a large portion of the compound class (e.g., 
women) were overselected for promotion. Op. at 33. (JA 233). 
Consequently, any notion of compound class discrimination is absurd.

Second, the statistical evidence proffered by the class at
trial —  in an effort to prove compound class discrimination —
was found to be less reliable than that presented by MarAd. Op.
at 28 (JA 228) . For instance, plaintiffs' statistical experts
failed to separately examine statistics regarding career-ladder
against non-career-ladder positions. The District Court termed
this failure "irrational", finding that such separate analyses
were an absolute requirement given the unequal promotability
between these two types of positions. Id. at 29 (JA 229). In
addition, the class members' regression analyses were accorded
"relatively little weight" by the District Court, which noted
that such a study, "repeating a static view of the agency's
distribution of salary and similar benefits, shows little about
how the agency became the way it is." Id. at 30-31 (JA 230-31).
Indeed, the District Court opined:

It is not unusual that at an established 
agency like MarAd white males dominate the 
higher level, higher paying jobs. Many cases 
have dealt with just such a situation and 
concluded that it tells little about whether 
this situation came about as a result of 
presently actionable discrimination.

Id. (See also MarAd's brief at pp. 34-37). — / For the same

_1_3/ The class members' expert presented three regression 
analyses at trial. See Plaintiffs' brief at 28, n. 26. None of 
these analyses addressed or considered how promotions requiring 
specialized skills, education or training- would affect the (FOOTNOTE CONTINUED ON NEXT PAGE)

16



reasons, the class members' "descriptive" statistics —  providing 
"snapshots" of MarAd's workforce on an annual basis —  are even 
less reliable than the rejected regression analyses. (See Plain­
tiffs' brief at pp. 17-22) . Simply counting the number of 
"groups at issue" (e.g,, white-males, black-females, etc.) in 
particular GS slots, in a particular time period, tells nothing 
about the underlying claim —  whether any significant statistical 
disparity (e.g., more white males in higher positions) was caused 
by a MarAd discriminatory policy or practice during the relevant 
time frame. Moreover, the class members' survival analysis, 
tracking the relative promotion rates of the respective sex/race 
groups (and lumping together all types of promotions), again 
failed to address the minimum objective qualifications neces­
sarily at issue in various promotions. See Valentino, and Trout, 
supra.

Therefore, there is no evidence in the record on which to 
set aside the District Court's refusal to grant relief to the 
compound class.

13/ (FOOTNOTE CONTINUED FROM PREVIOUS PAGE)
statistical promotabilities. Indeed, the only analysis attempting 
to address these latter requirements was prepared after trial and 
post-trial briefing, although even that study failed to adequately 
account for objective minimum qualifications. Indeed, this last 
regression analysis failed to factor in the position type at 
issue, even though this information was available in the data 
base. Instead, plaintiffs assumed that all positions could be 
lumped together for comparison sake. Id. The District Court 
properly refused to consider this untimely submission. See 
generally Valentino v. U.S. Postal Service, 674 F.2d 56 (D.C.
Cir. 1982), and Trout v. Lehman, 702 F.2d 1094 (D.C. Cir. 1983), 
cert pet, pending, 52 U.S.L.W. 3387 (No. 83-706).

17



B. The District Court Properly Held That 
MarAd Had Not Discriminated Against 
Women Employees On The Basis of Sex

As noted, the District Court held that plaintiffs had "not
proven either disparate treatment or a disparate impact on the
basis of sex from the hiring and promotional practices at MarAd."
Op. at 33 (JA 233). In relying on the more probative statistical
study submitted by MarAd, the Court stated:

...women have, if anything, been selected in 
disproportionately greater numbers in com­
petitive appointments at MarAd. Though some 
of this overselection of women disappears if 
the apparent bias in favor of applicants from 
within MarAd is factored in **_/ there remains 
a slight, statically insignificant preference 
for women. Such a result is obviously 
inconsistent with a pattern and practice of 
discrimination against women. And [MarAd's 
expert] found only a small, statistically 
insignificant difference in the time to 
promotion of women in career ladder positions 
and in other noncompetitive promotions....
Taken as- a whole, the evidence does not 
indicate the significant difference between 
the promotion and hiring rates of men and 
women required either to create an inference 
of discriminatory intent or a duty to validate 
the selection practices of the agency . . .

J7 By use of the term "bias," the Court does 
not imply that there is anything improper in 
the fact that MarAd employees have a better 
chance than others of being hired for higher 
level jobs. In fact, one would expect that 
persons already employed at MarAd would be 
more likely than others to have the qualifica­
tions for other positions at MarAd.

Op. at 33-34 (JA 233-34). This finding was clearly based on the 
more reliable evidence proffered by MarAd. In support of its 
appeal from this ruling, the plaintiff class again relies on its

18



"white-maleness" statistics discussed above. Of course that 
evidence -- whether viewed by itself or in conjunction with 
MarAd's more reliable evidence -- simply failed to meet the 
prerequisites for proving class-wide discrimination.

C. The District Court Properly Refused To
Review Plaintiffs' Post-Trial Statistical 
Analysis________________________________

The class assigns as error the District Court's refusal to 
consider an untimely statistical report which they attempted to 
"introduce" after trial and submission of post-trial briefs. 
(Plaintiffs' brief at p. 65, n. 59). This report, they assert, 
was prepared in light of this Court's decision in Valentino, 
supra, which, they further claim, "clearly defined the legal 
standards governing [class statistical analyses] in this circuit 
for the first time . . . "  ^d. This self-serving statement
fails to explain why the class failed to file this untimely 
evidence during the trial -- Valentino merely held that statis­
tical proof alleging class-wide discrimination in promotions and 
hiring must necessarily account for the specific objective 
minimum qualifications of the jobs at issue in order to meet 
plaintiff's prima facie burden under Title VII. That burden, of 
course, was on plaintiffs' shoulders long before this Court's 
decision in Valentino. — ^

14/ Indeed this Court Tn Trout severely criticized this very 
sort of post-trial effort to reopen the evidence through 
introduction of yet another statistical analysis. Trout, surpa, 
702 F. 2d at 1106-1107. ----- ---

19



Significantly, the "post-Valentino11 analysis was critically 
flawed and still failed to account for the specific minimum 
objective qualifications at issue in the promotion scheme. See 
p. 14, n. 12, supra. As such, even its post-trial consideration 
of this study could not have altered the District Court's adverse 
sex discrimination finding. In any event, plaintiffs fail to 
show that the District Court abused its discretion in refusing to 
credit these belated statistics.

D. The District Court Properly Held That 
The Affirmative Action Provisions Of 
42 U.S.C. §2000e-16 Do Not Create A 
Private Right of Action______________

The class members challenge the District Court's holding 
that the affirmative action provisions of 42 U.S.C. § 2000e-16(b) 
do not give rise to a private right of action. Op. at 3 7 (JA 
237) (Plaintiffs' brief, pp. 66-68). However, this challenge is 
only supported by the generalized legislative history cited at 
length in their brief and the absence of any relevant judicial 
authority.

The traditional test for creation of private rights of 
action was articulated in Court v. Ash, 422 U.S. 66 (1975) and 
was restated recently in this Circuit in Council of the Blind of 
Del. Cty. Valley v. Regan, 709 F.2d 1521-26 (D.C. Cir. 1983) (en 
banc):

First, is the plaintiff "one of the class for whose 
special benefit the statute was enacted," . . . 
that is, does the statute create a federal 
right in favor of the plaintiff? Second, is 
there any indication of legislative intent, 
explicit or implicit, either to create such a 
remedy or to deny one? . . . Third, is it 
consistent with the underlying purposes of

20



the legislative scheme to imply such a remedy 
for the plaintiff? . . . And finally, is the 
cause of action one traditionally relegated 
to state law, in an area basically the 
concern of the States, so that it would be 
inappropriate to infer a cause of action based solely on federal law?

Although analyses of this issue typically address all four 
issues, "[legislative intent has proved to be the preeminent 
test‘" Sorrell v. U.S. Intern. Communications Agency, 682 F.2d 
981, 986 (D.C. Cir. 1982).

Generally, it is the burden of the plaintiff to demonstrate 
the presence of the requisite factors -- especially the legisla­
tive intent. See Shivers v. Landrieu, 674 F.2d 906, 912 (D.C.
Cir. 1981) ("without some palpable indication of legislative 
intent we would be most ill-advised to discover an implied cause 
of action"). Similarly in Council for the Blind of Del. Cty. 
Valley, the government prevailed by demonstrating that in enact­
ing the legislation at issue Congress "chose to authorize private 
suits principally against the state or local government . . . rather 
than the [federal agency]." _Id. at 1530-31. Plaintiffs point to 
no legislative intent to create such a cause of action.

The District Court flatly rejected the notion that §2000e-
16(b) creates a private cause of action, holding

...there is no evidence that Congress intended 
that this [affirmative action] obligation, as 
opposed to the general obligation to be free 
from discrimination, was to be privately 
enforceable. Nor is it clear how such a duty 
could be measured for the purpose of, e.g., awarding back pay.

21



Op. at 37 (JA). — / In rejecting plaintiffs' reliance on Cannon
v. University of Chicago, 441 U.S. 677 (1979), the District Court
noted that a judicially-created private cause of action, such as
sought here, would raise serious questions of reviewability:

While it is difficult to determine which 
individuals would have received employment 
had a given workforce been free of discrimi­
nation, it is impossible to determine who 
would have benefited had a proper amount of 
affirmative action been instituted.

Op. at 37 (JA 237). Indeed, the only other court to have addressed
this issue also refused to find a private cause of action. See
Baca v * Butz, 394 F. Supp. 888, 894 n. 11 (D.N.M. 1975).

Cannon v. University of Chicago, supra, addressed an entirely 
aspect of federal civil rights law —  governmental 

distribution of federal funds to educational programs —  under
Title IX of the 1972 amendments to the Civil Rights Act of 1964. 
Here, however, it is the government's own role as employer that 
is at issue. Pursuant to the various administrative and judicial 
procedures expressly provided in Title VII, an aggrieved federal 
employee, (or class of employees) has wholly adequate remedies

_15/ In making its findings on this issue, the District Court 
also found that judicial review of MarAd's affirmative action 
obligations was available under the narrow and limited provisions 
of the Administrative Procedure Act, 5 U.S.’C. § 701 et seq. Op. 
at 37-39 (JA 237-39). However, the Court deferred ruling on the 
merits of APA review and, subsequently, incorporated certain 
aspects of the affirmative action issues into its relief orders. 
It did specifically find that plaintiffs had failed to prove that 
MarAd's affirmative action plan was "arbitrary and capricious." 
Sea 559 F. Supp. at 951; see also Inj. at 7 (JA 260).

22



where he or she has suffered discriminatory treatment. Where 
Congress has expressly provided for such relief in the federal 
sector, the Courts should refuse to create additional remedies.
See e.g., Bush v. Lucas, ____ U.S. ____, 103 S. Ct. 2104 (1983)
(courts will not create direct, constitutional cause of action 
for aggrieved federal employees where Congress established 
specific scheme for review under the Civil Service Reform Act of 
1978).

The District Court's ruling on this issue was therefore 
correct and should be affirmed.

E. The District Court Properly Limited 
The Class Relief To The Extent 
Challenged By Plaintiffs___________

The plaintiff class challenges three aspects of the District 
Court's relief order. (Plaintiffs' brief pp. 69 and 70) These 
challenges center on (1) the tolling of the statutory limitations 
period for non-prevailing class members to file individual 
discrimination complaints in the future; (2) the preclusive 
effect of the District Court's adverse class rulings on non­
prevailing class members who might later raise individual dis­
crimination claims; and (3) the scope of prevailing class relief. 
These claims are without merit.

First, plaintiffs apparently argue that members of the non­
prevailing class are not time-barred from filing individual 
complaints of their own where, in relying on the pendency of the 
class action, they failed to previously file such claims. In 
effect, they assert that the pendency of the class action tolled

23



the limitation period set forth in 42 U.S.C. § 2000e-5. See
Crown, Cork & Seal Co. v. Parker, ___ U.S. ___, ___ S. Ct. ___,
76 L.Ed.2d 628 (1983); —  ̂ see also Zipes v. Trans World Airlines, 
Inc., 455 U.S. 385 (1982) (filing requirements of Title VII are 
not jurisdictional but are statutorily limited).

In order to protect these as yet unfiled claims, plaintiffs 
unsuccessfully moved the District Court to amend its relief order 
(pursuant to Rule 59(e), F.R. Civ. P.) so that nonprevailing 
class members would receive notice of their right to pursue 
individual complaints of discrimination independently of the 
class action. This question, of course, is entirely speculative 
and premature where such unidentified individual claims have yet 
to make their way to the District Court. The question of whether 
such claims are time-barred, (or were grandfathered in because of 
the pending class action), can only be resolved if and when such 
persons actually file claims. — ^

Second, plaintiffs apparently seek a determination that the 
adverse class findings are in no way binding on any non-prevailing 
individual class member's claims of discrimination. They sought 
to establish this declaratory ruling through their unsuccessful 
motion to alter or amend the District Court's judgment. Again,

16/ Of course, Crown, Cork addressed class decertification.
17/ The basis for plaintiffs' claim is itself confusing. The 
District Court did provide in its relief that plaintiffs' counsel 
could, if they so chose, send out a notice to non-prevailing 
women class members. The only condition was that MarAd was not 
obligated to either provide the notice or pay the expense. Inj. 
at 6 (JA 259). While not part of the record, such a notice was sent out by plaintiffs.

24



however, this claim is highly speculative given the lack of 
identifiable, non-prevailing class members who have filed com­
plaints. In any event, to the extent that individual complaints 
are later filed by non-prevailing class members and rely on the 
same claims and/or evidence raised during the class proceedings, 
such persons are not entitled to any relief based on the doctrine 
of res .judicata. See EEOC v. Federal Reserve Bank of Richmond.
698 F .2d 633, 674-75 (4th Cir. 1983), cert, granted sub. nom., 
Cooper v. Federal Reserve Bank of Richmond. 52 U.S.L.W. 3342 
(October 31, 1983) (No. 83-185). —  ̂ See also, Mervin v. F.T.C.,
591 F.2d 821, 830 (D.C. Cir. 1978).

Third, plaintiffs assert that, in establishing the scope of 
prevailing class relief, the District Court (1) erred in apparent­
ly relying upon the individual complaint regulations (5 C.F.R. 
§1613.214 (a)) -- providing a 30-day throw-back period for filing 
individual, prevailing class claims -- rather than the class com­
plaint regulations (5 C.F.R. § 1613.602), which set forth a 135-day 
throw-back period, and (2) was correct in permitting prevailing in­
dividual class claims to go back two years prior to the filing of the 
administrative complaint. (Plaintiffs' brief at 70). With regard to 
the first claim, MarAd has already noted in its brief (pp. 83-84) 
that the class complaint regulations are, in fact, the appropriate 
provisions for any relief here. However, as also set forth in 
MarAd' s earlier brief, (id.) the District Court erred in providing 
relief for claims arising prior to March 21, 1977 -- 135 days prior 
to the filing of the administrative class complaint in this case.

T57 The Supreme Court has recently granted certiorari in that 
case in order to address this very issue.

25



II. MarAd's Appeal
A. The District Court Erred In Certifying 

The Compound, Across-the-Board Class

In its opening brief, MarAd extensively examined the District 
Court's error in certifying (and subsequently failing to modify) 
the compound, across-the-board class. (Brief at pp. 45-63). 
Several arguments raised by plaintiffs on this issue require a 
response.

First, plaintiffs claim that they satisfied the concededly 
strict requisites of Rule 23, F.R.Civ. P. at the original class 
certification hearing. Impliedly, they claim that any subsequent 
modification of the certified class was improper because "the 
propriety of a class certification [is not judged] by hindsight." 
Plaintiffs' brief at p. 44, quoting Falcon, 457 U.S. at 160. 
However, this argument totally misses the point of MarAd's 
appeal. Notwithstanding the impropriety of the District Court's 
original certification, —  its subsequent refusal to modify the 
class in any manner —  in light of MarAd's twice demonstrating 
the inherent conflicts, untypical and uncommon claims raised in 
the class —  flies in the face of the Supreme Court's ruling in 
Falcon. It is not hindsight upon which MarAd raises this appeal 
but, instead, the District Court's failure to (1) specifically

19/ At the original certification hearing, plaintiffs merely 
introduced unsponsored raw data lacking any statistical analysis or significance. (R. 333).

26



address and resolve the strict requirements of Rule 23 at the 
original certification stage, and (2) modify or otherwise limit 
the plaintiff class before trial so that MarAd could properly 
address the issues at trial. In the former situation, the 
District Court's class certification order is devoid of any such 
analysis, and merely sets forth the compound class' definition. 
(JA 152) In the latter, the District Court's failure deprived 
MarAd of an opportunity to meet the issues squarely at trial. 
Additionally, the District Court again failed to set forth the 
reasons for its failure to limit the class, despite extensive 
documentation supporting the need to do so.

Second, contrary to plaintiffs' claim, (Brief pp. 44-46) the 
Supreme Court in Falcon effectively ruled that across-the board 
classes —  at least those raising inconsistent and conflicting 
claims such as present here —  are legally deficient given Rule 
23's strict requirements. Id. 457 U.S. at 155-56. See also East 
Texas Motor Freight System, Inc, v. Rodriguez, 431 U.S. 295
(1978) . If anything, Falcon and East Texas Motor Frieght were 
based on class certifications presenting fewer and simpler 
inconsistencies and conflicts than are present here. —  ̂ Indeed,

20/ Plaintiffs also claim that, unlike Falcon and East Texas 
Motor Freight, the District Court here held an evidentiary 
hearing on certification, thereby legitamizing the process. 
However, that claim holds little water when the record is 
examined more closely —  the hearing in this case centered almost 
exclusively on MarAd's claim that plaintiff Lawrence was not 
suitable'to represent the class. (JA 26-150). There is little, 
if any, indication that the District Judge even considered the 
evidence submitted by plaintiff in certifying the class. Moreover, 
the District Court's failure to even generally analyze or discuss 
the requirements of.Rule 23 in its certification order only 
further underline the problems inherent in this case even at the 
threshhold point of certification. (JA 152)

27



plaintiffs' half-heartedly rationalize that the issues in Falcon 
which were fatal to certification -- inherent conflicts between 
promotion claims and hiring claims -- are somehow not relevant 
here. Such an argument, however, ignores a much more critical 
(and fatal) factor: challenges to MarAd's promotion and hiring
systems not only raised inconsistent claims themselves, but also 
raised conflicting interests within each system -- competitive v. 
non-competitive, professional v. clerical, etc. Merely investi­
gating "all the steps in the procedure" (Plaintiffs' brief at p. 
46) begs the question where there are several very different

O 1 /procedures being challenged. —
Third, plaintiffs' discussion of Rule 23' s typicality 

requirement ignores the fact that the named plaintiffs -- pro­
fessional, high-ranking employees-- presented claims wholly 
untypical from the main portion of the class. Instead, this lack 
of typicality is justified based on their allegation that all 
applicants and employees -- regardless of the concededly diverse 
and specialized positions at issue -- were subject to the "same 
discriminatory policies." Plaintiffs' brief at pp. 47-49. As in 
Valentino, the three named plaintiffs here could not present

21/ The fatal flaw to this argument is that the District Court 
Held that a statistical analysis would necessarily have to 
examine separately competitive v. non-competitive promotion 
procedures. Lumping the two together only further confuses the 
fset-finder's inquiry. See EEOC v. Federal Reserve Bank of 
Richmond, supra. But see Lawler v. Alexander, 698 F.2d ST5 Ml i-hClrV 1983). -------

28



claims typical of the far-ranging, diverse and specialized 
professional and administrative occupations concededly present at 
MarAd. *

Fourth, plaintiffs' brief analysis of Rule 23's commonality 
requirement again generally refers to an "overall pattern of 
discriminatory practices" and "pervasive discriminatory employ­
ment policy." Brief at 50-51. Such a perfunctory response to 
MarAd's argument does not create or justify a class consisting of 
"common" claims where the many uncommon claims are evident.

Finally, plaintiffs' analysis of Rule 23' s adequacy of 
representation requirement again fails specifically to address 
the inherent conflicts present in the compound class, already 
discussed at length in MarAd's opening brief. (pp. 58-63).

For all the these reasons, the class certification was 
improper and should be reversed. — ^

22/ Plaintiffs' further rationalize that the typicality requirement 
was met because "all three named plaintiffs made out a prima facie 
case of individual discrimination on the basis of evidence presented 
at trial." Brief at p. 49. Notwithstanding the undisputed fact that 
none of the three individually named plaintiffs were successful on 
the merits (or even established a prima facie case of disparate 
treatment), the proof of such claims was wholly different from the 
sort of evidence necessary to establish a class-wide, disparate impact 
case. See Trout, 702 F.2d at 1112. The named plaintiffs -- who.only 
presented disparate treatment claims at trial -- have never indicated 
that they were alleging disparate impact claims as part of their case.
23/ The significance which plaintiffs attach to the acceptance of 
the class complaint by the Department of Commerce Office of Civil 
Rights is greatly overstated. First, under the applicable regula­
tions (5 C.F.R. t 713.604) (1977) that decision was made without 
the benefit of any argument by or on behalf of MarAd. Second, 
the District Court never purported to attach any significance to 
that action -- there is little doubt that it had an independent 
obligation under Rule 23 to review the applicable legal require­
ments. Third, the final decisions of agency civil rights officers 
may be supplemented or rebutted by agencies in subsequent litiga­
tion. See Toney v. Block, 705 F.2d 1364, 1366 (D.C. Cir. 1983).

29



B. The District Court Erred In Finding 
Partial Class (Race) Liability_____

Plaintiffs criticism (Brief pp. 56-58, 62-64) of MarAd's
claim that the District Court erred in finding partial class 
liability —  race discrimination at least below the GS-13 level 
—  is based on the same flawed analysis relied upon by the trial 
court. While these issues have been discussed at length in 
MarAd's opening brief (pp. 68-82), certain arguments raised by 
plaintiffs require a brief response.

First, plaintiffs assert that the District Court's class 
race discrimination finding was correct because it relied on both 
the alleged subjective selection standards ( e.q., disparate 
treatment) and the "facially neutral" (e.g., disparate impact) 
selection criteria used by MarAd. (Brief at 56-57) In support 
thereof, plaintiffs cite Payne v. Travenol Laboratories, Inc.
673 F. 2d 798 (5th Cir. 1982), cert, denied, ___ U.S. ___, 103 S.
Ct. 451 (1982) . —  However unlike Payne, the District Court
here specifically and exclusively relied upon a disparate impact 
theory to find partial (race) class liability and never made a 
finding of intent. 559 F. Supp. at 948-950. (JA 245-47). Conse­
quently, because there has been no finding that MarAd disparately 
treated its black employees in the selection process, the "same 
facts" which purportedly made out the disparate impact analysis are 
irrelevant when examining rejected claims of disparate treatment.

24/ Contrary to plaintiffs' assertion, the Court in Payne noted 
that after the disparate impact issue was resolved adversely to 
plaintiffs, the case was "dramatically" transformed because proof 
of intent was "crucial." Id., 673 F.2d at 817.

30



Second, the District Court's use of a disparate impact 
analysis here was clearly in error. Indeed, Pouncy v. Prudential 
Ins. Co. of America, 668 F.2d 795 (5th Cir. 1982) is directly on 
point where plaintiffs -- in mounting an across-the-board attack 
against MarAd's entire personnel system -- sought relief even 
broader and more unfocused than that sought in Pouncy. —  ̂ The 
fact that MarAd is a federal employer rather than a private 
sector employer is a distinction without a difference. The 
burden of proof requirements under Title VII are concededly the 
same regardless of the defendant's identity. See generally Brown 
v. GSA, 425 U.S. 820 (1977). In any event, neither the affirma­
tive action provisions of the statute (42 U.S.C. § 2000e-16(b)) 
nor the various regulations and executive orders promulgated 
thereunder supplant plaintiffs' burden to prove their case on the 
merits.

Third, even if an impact analysis were somehow appropriate, 
plaintiffs' statistical evidence -- purporting to show that 
MarAd's "subjective" promotion decisions had as their goal the 
exclusion of blacks -- is not supported by the evidence. No 
signficant disparity in the number of blacks ultimately selected 
from eligibles was demonstrated. (MarAd's Brief at pp. 80-81).

25/ Plaintiffs' insistence that Pouncy not be followed in this 
Circuit, based on two previous cases before this Court, is 
totally misplaced. In Talev v. Reinhardt, 662 F.2d 888 (D.C. 
Cir. 1981), an individual plaintiff unsuccessfully challenged his 
failure to receive a promotion. In McKenzie v. Sawyer, 684 F.2d 
62 (D.C. Cir. 1982), the plaintiff class was limited to a very 
specific section of the defendant agency. Contrary to plain­
tiffs' claim here, those cases do not remotely suggest that this 
Court has condoned the sort of relief proscribed in Pouncy.

31



Moreover, plaintiffs' reliance on statistics focusing on the 
applicant to eligible stage -- where plaintiffs see significant 
racial disparity -- are legally deficient because that evidence 
failed to account for the minimum objective qualifications 
necessary to be considered for selection for many of the con- 
cededly specialized jobs at issue. See Valentino. 674 F.2d at 
67-68 and Trout, 702 F.2d at 1102.

Fourth, with regard to non-competitive (e.g., career ladder) 
promotions, plaintifffs suggest that some disparity was demon­
strated. However, plaintiffs never produced any analysis of 
carreer ladder promotions. MarAd's analysis clearly demonstrated 
that there was no statistically significant differences among the 
race/sex subgroups. (DX 119, pp. 17-20). Moreover, MarAd's 
competitive promotion analysis proved that blacks and women 
(either together or separately) were slightly over-selected in 
competing for entry into career ladders.

C. The District Court's Relief Order,
Requiring A Validation Study, was 
Erroneous_________________________

In response to MarAd's claim that it could not be held
accountable for binding 0PM position selection/promotion criteria
(Brief at p. 82), plaintiffs assert that the validation study
ordered by the Court is appropriate class relief. (Brief at pp.
59-62). They base this assertion on their view that:

The starting point for the carrying out of a 
competitive selection at MarAd is the drawing 
up by MarAd supervisors of a position descrip­
tion that describes the duties of the job in 
question. This process is one involving a 
degree of judgment and subjectivity and is

32



carried out at MarAd by a virtually all-white 
supervisory workforce. [footnote omitted].
From the position description all else flows.

Brief at 59. Notwithstanding the District Court's refusal to 
condemn the "whiteness" of the higher level positions plaintiffs' 
argument actually supports MarAd's claim. First, to the extent 
that MarAd subjectively treats black applicants and employees 
seeking promotions in a disparate manner (as plaintiffs allege), 
the District Court's liability finding -- based on a disparate 
impact model -- is wholly inconsistent with this sort of relief. 
Indeed, the only "facially neutral" selection standards at issue 
are those found in the 0PM X-118, which were concededly not 
challenged in this case. See Plaintiffs' Brief at 62, n. 56. 
Where the District Court held MarAd liable under a disparate 
impact theory, relief addressing the alleged subjective, inten­
tionally discriminatory employment practices of MarAd -- a 
disparate treatment theory -- is simply misplaced.

Second, to the extent the District Court ordered such relief 
based on the binding effect of the 0PM X-118 standards (the 
"facially neutral" criteria) on its selection process -- for 
which there is ample evidence in the record -- MarAd simply can 
not be held accountable for such standards. See Trout, 702 F.2d 
at 1105-14.

Therefore, the validation study is inconsistent with the 
District Court's findings and should be reversed.

33



CONCLUSION

For all the reasons stated above and in MarAd's opening 
brief, the District Court's findings should be affirmed in part 
and reversed in part.

STANLEY S. HARRIS, 
United States Attorney.

ROYCE C. LAMBERTH,
R. CRAIG LAWRENCE,
JOHN H.E. BAYLY, JR., 
STUART H. NEWBERGER, 
Assistant United States

OF COUNSEL:
TIMOTHY SHEA
Office of the Chief Counsel
Maritime Administration
U.S. Department of Transportation

Attorneys.

I

t

34



■v

%