Waisome v. Port Authority of New York and New Jersey Brief for Defendants-Appellees

    91-7213
llniteii States Court of Appeals

fo r  the

&edml» Circuit

F E L IX  W A ISO M E , F R E D D IE  M cM IL LA N , R IC H A R D  
B. K E IT H , R O B E R T  L . B E T H E A , E L L S W O R T H  
C O R U M , JR .,  H ILA RY K IN G , R O D E R IC K  W . UPSHUR 

on behalf of themselves and all those similarly situated,

Plaintiffs-Appellants,
— against —

T H E  P O R T  A U T H O R IT Y  O F  N EW  YO RK  AND N E W  
J E R S E Y ; T H E  B O A R D  O F  C O M M IS S IO N E R S ; 
ST E P H E N  B E R G E R ; H EN RY I. D E G E N E S T E ; and 
T H E  P O R T  A U T H O R IT Y  P O L IC E  B E N E V O L E N T  

A SSO C IA TIO N , IN C .,

Defendants-Appellees,

ON APPEAL FROM THE UNITED STATES DISTRICT COURT 
FOR THE SOUTHERN DISTRICT OF NEW YORK

BRIEF FOR THE PORT AUTHORITY 
DEFENDANTS-APPELLEES

On the Brief:
Arthur P. Berg 
Philip A. Maurer 
James Begley 
C arlene V . McIntyre

Milton H. Pachter 
Attorney for Defendants- 

Appellees The Port Authority 
of New York and New Jersey 

One World Trade Center - 66N 
New York, New York 10048 
(212) 435-6835



1

T A B L E  O F C O N TEN TS
Page

Table of Authorities.................................    iii

Statement Of The Issue Presented For Review v

Statement Of The Case....................   1

A. Nature of the Case and Course of
Proceedings............................................  1

B. Statement of the Facts ...............................  3

1. The Sergeant’s Promotion
Examination Process..........................   3

2. EEO C Administrative Proceedings . . 6

3. The District Court’s Decision . . . . . . .  7

Summary Of The Argument ............................. . 9

Argument

The Selection Process For Police Sergeant
Had No Disparate Impact On Black
Candidates ........................................................   12

A. Standard of Review and Proceedings
Under Review...............................   12

B. Applicable Legal Standards...............   13

C. The Multi-Component Examination 
Process Had No Disparate Impact on
Black Candidates.................................  20

1. The Written Exam ination....................  20

2. The Oral Examination and the
Performance Appraisal..........................  27

3. The Eligibility L is t ......................... 28

Conclusion ................................................................. 37





Ill

TABLE OF AUTHORITIES
Cases Page

Bilingual Bicultural Coalition on Mass 
M edia, Inc. v. FC C , 595 F.2d 621 (D.C.
Cir. 1978) ............................. ............... ............. 17, 21

Bridgeport Guardians, Inc. v. M em bers o f  
Bridgeport Civil Sew . C om m ’n, 482 F.2d 
1333 (2d Cir. 1973) ......................................... 23

Bushey v. New York State Civil Service 
C om m ’n, 733 F.2d 220 (2d Cir. 1984) . . .  14, 18, 22

Castaneda v. Partida, 430 U.S. 482 (1977) . 15

Celotex Corp. v. Catrett, A ll  U.S. 317 
(1986) ........................................    12

Connecticut v. Teal, 457 U.S. 440 (1982) . .  10, 19, 20,
21, 24, 25

G ilbert v. City o f  L ittle R ock, A rk., 722 
F.2d 1390 (8th Cir. 1983).................. 34, 35

Guardians Assn v. Civil Service C om m ’n,
630 F.2d 79 (2d Cir. 1980)...........................  15 n.8, 18,

19, 23

H azelw ood School District v. United States,
433 U.S. 299 (1977)............. .......................... 15

Jackson v. Nassau County Civil Service 
C om m ’n, 424 F.Supp. 1162 (E.D.N.Y.
1976) ............................................................... .... 22, 23

K irkland v. New York State Dept, o f  
C orrectional Services, 374 F.Supp. 1361 
(S.D.N.Y. 1974), a f f ’d, 520 F.2d 420 (2d 
Cir. 1975), cert, denied , 429 U.S. 823 
(1976) ................ .......................... ........................ 23, 24



IV

Cases Page
K irkland v. New York State D ept, o f  

C orrectional Services, 711 F.2d 1117 (2d 
Cir. 1983) ..................................................... .. 29, 30, 34

Ottaviani v. State U. o f  New York at New
Paltz, 875 F.2d 365 (2d Cir. 1 9 8 9 )...........  8, 16,

16 n.9-10, 
17, 26, 30

Pullman-Standard v. Swint, 456 U.S. 273
(1982)...................................................................  12

Teal v. Connecticut, 645 F.2d 133 (2d Cir.
1981) ...................................................................  18

U nderwood v. State o f  New York, 28 
F .E .P . 922 (S.D.N.Y 1980)............. ............. 35 n.16

W ade v. New York Tel. C o., 500 F.Supp.
1170 (S.D.N.Y. 1980) .................................... 33

W ards C ove Packing C o., Inc. v. Frank
Antonio, 109 S.Ct. 2115 (1 9 8 9 ).................. 13, 14 n.7,

16 n.9
Watson v. Fort W orth Bank and Trust, 487

U.S. 977 (1988).................................................  13-14 ,15 ,
16 n.9-10, 

18, 21
Statutes:
Civil Rights Act of 1866, 42 U.S.C. §1981 

(West 1981 & Supp. 1 9 9 1 ) ........................... 2
Civil Rights Act of 1866, 42 U.S.C. §1983 

(West 1981 & Supp. 1 9 9 1 ) ........................... 2
Civil Rights Act of 1964, Tit. V II, 42 

U .S.C. §2000d (West 1981 & Supp. 1991) 2
Civil Rights Act of 1964, Tit. V II, 42 

U .S.C. §2000e (West 1981 & Supp. 1991) 2
Other Authorities:
29 C .F .R . §1607,4D (1 9 8 8 ).............................  29, 31



Rniteii States (Court of Appeals
fo r  the

i>ec0ni> (Circuit

FE LIX  WAISOME, FRED D IE McMILLAN, RICHARD 
B. KEITH, ROBERT L. BETHEA, ELLSW ORTH 
CORUM, JR ., HILARY KING, RODERICK W. UPSHUR 

on behalf of themselves and all those similarly situated,

Plaintiffs-Appellants,

— against —

THE PORT AUTHORITY OF NEW YORK AND NEW 
JE R S E Y ; TH E BOARD O F CO M M ISSIO N ERS; 
STEPHEN BERG ER; HENRY I. D EGEN ESTE; and 
THE PORT AUTHORITY POLICE BENEVOLENT 

ASSOCIATION, INC.,

Defendants-Appellees.

ON APPEAL FROM THE UNITED STATES DISTRICT COURT 
FOR THE SOUTHERN DISTRICT O F NEW  YORK

BRIEF FOR THE PORT AUTHORITY 
DEFEND ANTS-APPELLEES

STATEMENT OF THE CASE

A. Nature o f  the Case and Course o f  Proceedings

In this case, plaintiffs-appellants, individual Black police 
officers who are employed by the Port Authority and a class 
of similarly situated officers whom they represent, challenge,



2

as discriminatory, a selection process used by the Port 
Authority to make promotions to the rank of Sergeant. 
Plaintiffs are currently employed as police officers by the 
Port Authority and now represent a class consisting of all 
Black police officers who took the promotion examination. 
They claimed below that defendants-appellees1 have 
violated Tide VII of the Civil Rights Act of 1964, 42 U.S.C. 
§2000d, §2000e, e tseq . (West 1981 & Supp. 1991), as well 
as the Civil Rights Act of 1866, 42 U.S.C. §1981 and §1983. 
Plaintiffs’ district court papers sought declaratory relief 
based on the defendants’ alleged maintenance of an 
unlawful employment practice and the administration of 
unlawful selection criteria. Specifically, in their original 
complaint, they alleged that the Port Authority violated 
plaintiffs’ rights by utilizing, as a prerequisite for promo­
tion to the rank of Sergeant, a selection procedure which 
had an adverse impact on Black applicants and which had 
not been demonstrated to be job-related. In their amended 
complaint, plaintiffs added a claim that the selection pro­
cess had intentionally discriminated against Black ap­
plicants (J.A, 20).2 Before filing their amended complaint, 
plaintiffs also served and filed a motion seeking preliminary 
injunctive relief.

By endorsement dated October 14, 1988, the district 
court denied plaintiffs’ motion for preliminary injunctive 
relief, ruling that plaintiffs had failed to demonstrate

1 Plaintiffs have also named as a defendant the Port Authority 
Police Benevolent Association, Inc. (hereinafter “PBA”), the union 
which is the collective bargaining agent for Port Authority police 
officers and which is represented in these proceedings by its own 
counsel.

2 “J.A .” refers to the joint appendix filed by plaintiffs in this Court. 
“Br.” refers to the opening brief filed by plaintiffs in this Court.



3

irreparable harm. On November 21, 1989, the court ap­
proved a stipulation between the parties whereby the plain­
tiffs agreed to discontinue their claims for intentional 
discrimination and the defendants agreed that if the ex­
amination was found by the court to have an adverse im­
pact on Black applicants, they would not seek to litigate 
the validity of the examination (J.A. 80-81).3 The parties 
cross-moved for summary judgment agreeing that, despite 
slight differences in methodology and computation, the 
underlying facts were essentially undisputed. Therefore, 
the parties filed statements consisting of the material facts 
not in dispute in accordance with the rules of the court 
(J.A. 82).

B. Statem ent o f  the Facts

1. The Sergeant’s Prom otion Exam ination Process

On July 11, 1986, the Personnel Department of the Port 
Authority announced the commencement of an examina­
tion process for the purpose of establishing a vertical list 
of Port Authority police officers eligible for promotion to 
the rank of Sergeant (J.A. 83; Promotion Examination An­
nouncement, No. 86-31, J.A. 126). In order to be eligible 
to participate, candidates for promotion were required to 
have at least two years in grade (including Academy train­
ing) as a Port Authority police officer and were also re­
quired to be actually employed as a police officer as of the 
first date of the written test (J.A. 126).

The selection process for placement on the Eligible List 
consisted of three basic components. The first component 
consisted of a written test, to “measure knowledge of law,

3 On November 2, 1988, the district court granted the motion to in­
tervene as defendants filed by Port Authority police officers Joseph 
Leather, et al.



4

police supervision and social and psychological problems 
in police work.” The second component was an individual 
oral test to “measure judgment and personal qualifica­
tions.” Finally, the third component was a performance 
appraisal consisting of two parts — a supervisory perform­
ance rating and a score based on the candidate’s attend­
ance record (J.A. 126, 127).

The written examination for police officers was ad­
ministered on September 6, 1986, and a make-up test was 
administered on September 20, 1986 (J.A. 85). In mid- 
November of 1986, candidates were notified of their scores 
on the written component. Candidates were given until 
December 19, 1986 to appeal the results of the written ex­
amination and, by January 8, 1987, all the appeals taken 
were completed (J.A. 85). The individual oral examina­
tions were administered between January 26, 1987 and 
February 13, 1987. The performance appraisal process 
began on March 2, 1987 and was completed by March 20, 
1987 (J.A. 85). The performance appraisal ratings were 
factored into the total test score.4 The eligibility list was 
issued on March 30, 1987 (J.A. 86).5

The passing score for the written examination was 66 % 
and the passing score for the oral was 69.5% (J.A. 86, 90). 
A passing score on the written was needed to proceed to 
the oral examination and a passing score on the oral was 
needed to proceed to the performance appraisal (J.A. 86,

4 There was an identical selection process commenced for detectives 
who desired to be promoted to the rank of sergeant (J.A. 134). That 
process also consisted of a written examination, oral examination and 
performance appraisal component. Successful detective candidates 
were placed on the same Eligible List as successful police officers.

5 A revised eligibility list was issued December 4, 1987 after an Ap­
peals Board resolved seventeen appeals from candidates on the per­
formance evaluation process (J.A. 97).



5

90). The weights accorded to the three components of the 
selection process were 55 % for the written examination, 
35% for the oral examination and 10% for the perfor­
mance appraisal process (J.A. 84-85).

A total of 617 police officers (including detectives) par­
ticipated in the selection process. Of the total number of 
participants, 508 were White and 64 were Black (J.A. 86). 
A total of 539 participants passed the written examination. 
Of those who passed the written, 455 were White and 50 
were Black (J.A. 86, 188).

Of the 539 successful candidates on the written examina­
tion, 531 participants took the oral examination. Of those 
who either decided not to or could not proceed further, 
7 were White and 1 was Black (J.A. 89-90, 188). On the 
written examination, the pass rate for Blacks (78.13 %) was 
87.2% of the pass rate for Whites (89.57 %) (J.A. 124,197). 
Of the 531 participants who took the oral examination, 
448 participants were White and 49 participants were 
Black. Of the 531 participants, 310 passed the oral exam. 
Of those who passed, 258 were White and 33 were Black 
(J.A. 90, 188). On the oral examination, the Black pass 
rate (67.35%) was 116.97% of the White pass rate (57.58) 
(J.A. 90, 188).

Of the White candidates who participated in both the 
written and the oral examinations, 51.7% passed (258 
divided by 499) and of those Black candidates who par­
ticipated in both  oral and written examinations, 51.56% 
(33 divided by 64) passed (J.A. 125, 188).

Of the 617 participants who began the selection process 
(508 Whites and 64 Black), 310 participants completed the 
process and were included on the Eligible List (J.A. 91, 
97-101). Of the 310 candidates included on the Eligible 
List, 258 were White and 33 were Black (J.A. 92, 125).



6

When the list expired on March 30, 1990, 79 promo­
tions had been made from the list and the 85th candidate 
on the list had been reached.6 The promoted officers in­
cluded 70 Whites, 5 Blacks, 2 others and two “grand­
fathered” candidates who were White (J.A. 124, 189).

2. EEO C  Administrative Proceedings

The Eligible List was issued on March 30, 1987. All 
named plaintiffs filed charges of discrimination with the 
Equal Employment Opportunity Commission (EEOC). In 
each case, a determination was made by the EEOC that 
each charging party had not been the subject of discrimina­
tion in violation of Title V II. The letters to all plaintiffs 
were dated December 3, 1987.

Mr. Felix Waisome filed a charge of discrimination on 
June 8, 1987. He passed the written test with a score of 
90, but failed the oral by one-half of a point (J.A. 95).

Mr. Freddie McMillan filed his charge on June 10, 1987. 
He passed both the written exam (score of 80) and the oral 
(score of 77.5). However, he complained that his combined 
scores placed him near the bottom of the Eligible List (No, 
195) (J.A. 95).

Mr. Richard B. Keith filed his charge on July 20, 1987. 
He passed the written test (score of 88) and failed the oral 
(score of 64.5) (J.A. 95).

Mr. Robert L. Bethea filed his charge on July 14, 1987. 
He passed the written (score of 68) and the oral (score of 
94). His claim of discrimination was based on his position 
on the Eligible List (No. 221) (J.A. 95).

6 Two candidates whose rankings were high enough to entitle them 
to promotions retired and four other candidates, all White, refused 
promotions and remained on the force (J.A. 91-92).



7

Mr. Ellsworth Corum, Jr. filed his charge on July 27, 
1987. He passed the written examination (score of 86) but 
failed the oral (score of 68) (J.A. 95).

Mr. Hilary A. King filed his charge on July 30, 1987. 
He alleged that he passed both the written and oral. He 
based his claim of discrimination on his position near the 
bottom of the Eligible List (No. 180) (J.A. 95).

Mr. Roderick W. Upshur filed his charge on September 
30, 1987. He failed the written test (score of 58). He is the 
only named plaintiff that failed the written component of 
the selection process (J.A. 96).

3. The District C ourt’s Decision

After outlining the essential facts, the district court (Duffy, 
J.) initially determined that it would be appropriate to cer­
tify a class, a determination which defendants-appellees do 
not contest on appeal. On the merits, the court determined 
that plaintiffs had not met their burden of demonstrating 
that the examination process had an adverse impact on 
Black candidates. The court then separately analyzed 
plaintiffs’ claims that the written component of the selec­
tion process as well as the overall process had an adverse 
effect on Black candidates. The court noted that, in order 
to prevail, plaintiffs were bound to demonstrate that “the 
disputed component denied minorities, to a dispropor­
tionate degree, the opportunity to be promoted” (J.A. 193). 
Comparing the pass rates on the written component, the 
court noted that it was undisputed that, using standard 
deviation analysis as a unit to measure the statistical 
significance of those differences, the differences were 
statistically significant at 2.68 standard deviations. 
However, the court also noted that the Second Circuit had 
“rejected the efficacy of a ‘minimum threshold level of



8

statistical significance’ for determining whether the plain­
tiff has established a prim a facie case of discrimination. 
Ottaviani v. State Univ. o f  New York at New Paltz, 875 
F.2d at 373.” (J.A. 195). Following Ottaviani, which had 
drawn from recent Supreme Court pronouncements on this 
issue, the court below noted that statistical analysis was 
not the only method adopted by the Court in order to 
determine whether such differences are legally significant. 
The court observed that “[i]n order to determine the im­
portance or magnitude of the differences, the statistical 
data must be analyzed in the context of the situation in 
practical terms” (J.A. 196).

The district court used two additional measures to deter­
mine whether the difference in pass rates on the written 
exam was practically significant. First, the court referred to 
the EEOC guidelines or the “four-fifths rule” (80 % rule con­
tained in those guidelines). Under that rule, a selection rate 
for any protected group which is less than four-fifths the 
selection rate for the group with the highest rate is general­
ly determined by the federal rules to be evidence of adverse 
impact. Applying this rule, the court noted that the pass 
rate for Blacks on the written component of the examina­
tion was 87.2 % of the pass rate for White candidates. Ad­
ditionally, the court observed that while a test for statistical 
significance yielded a result of 2.68 standard deviations, 
had two additional Black candidates passed, the difference 
would no longer be statistically significant (J.A. 198).

Further, the court below addressed plaintiffs’ claim that 
mathematically a score of 76 was needed on the written 
component to qualify an individual for promotion, and 
a comparison between the proportion of Blacks and Whites 
achieving such a score was statistically significant at 2.68 
standard deviations. The court observed that while some 
Blacks performed well on the written component, as a 
group Blacks scored lower than Whites. However, the



9

court held that this evidence, standing alone, did not in­
dicate adverse impact because a number of Blacks per­
formed better than Whites on the balance of the examina­
tion, and a comparison of overall pass rates on the entire 
examination revealed no statistical disparity (J.A. 199-200).

Finally, applying a test of statistical significance, a test 
which the court below held was appropriate due to the 
small number of promotions involved, the district court 
rejected plaintiffs’ claim that the overall promotion rates, 
as measured by the overall ranks of all candidates on the 
eligibility list, demonstrated adverse impact. The court 
observed that an application of standard deviation analysis 
revealed a difference in selection rates of 1.34 standard 
deviations, a number well below the level of 2 or 3 stan­
dard deviations which constitutes statistical significance 
(J.A. 200). Based on its determination that plaintiffs had 
failed to present evidence of a “sufficiently substantial dif­
ference” between W hite and Black selection or 
performance rates, the court granted judgment in favor 
of the defendants and this appeal followed.

SUM M ARY O F T H E  A RGUM ENT

The Supreme Court has indicated that Title VII covers 
intentional discrimination as well as facially neutral 
employment practices which have a significant adverse im­
pact on protected groups and are, therefore, regarded as 
the functional equivalents of intentional discrimination. 
Under this latter theory, the sole issue presented to this 
Court for review is whether the examination process used 
by the Port Authority, until March of 1990, for the pro­
motion of police officers to the rank of Sergeant had an 
adverse impact on Black candidates for promotion. The 
district court, in disposing of cross-motions for summary 
judgment, determined that the plaintiffs failed to



10

demonstrate a “sufficiently substantial difference” in the 
White and Black selection rates or performance rates which 
would warrant an inference of discrimination under Title 
V II. We submit that, on the material facts of this case, 
which all parties agree are not in dispute, the district court 
was plainly correct.

First, while the Supreme Court, in Connecticut v. Teal, 
457 U.S. 440 (1982), indicated that an examination which 
operates as a pass/fail barrier and which, therefore, ex­
cludes protected candidates from further participation in 
the examination process must be separately analyzed, ap­
plication of Teal to the facts of this case would indicate 
that no such barrier was in operation. In the instant case, 
unlike the situation in Teal, a comparison of pass/fail rates 
for the written test component reveals that the Black pass 
rate is 87.2% of the White pass rate. Therefore, here, a 
comparison of pass rates for the written test component 
satisfies the four-fifths rule, set forth in the EEO C guide­
lines, which is used as a benchmark by courts, including 
the Supreme Court in Teal, to measure the significance 
of such differences.

Plaintiffs, however, contend that a higher score than the 
established score on the written examination was needed 
in order to assure a position on the eligibility list within 
striking range for promotion and that the difference in 
Black and White performance at the higher passing score 
reveals adverse impact. Plaintiffs’ analysis is flawed. Com­
mon sense dictates that, when, as here, a multi-component 
examination is involved, any legally significant disparity 
in performance by protected groups on one component will 
be reflected in the aggregate results achieved by those in­
dividuals on the entire examination. Here, while Whites 
scored better than Blacks on the written component, Blacks 
scored better than Whites on the oral examination. Most 
importantly, a significant number of Blacks performed well



11

enough on the entire process that an analysis of the overall 
promotion rates reveals no statistical disparity between the 
number of Whites and the number of Blacks promoted.

Second, the selection process resulted in the promotion 
of only 79 individuals (5 Blacks, 70 Whites, 2 others and 
2 grandfathered candidates). As the district court noted, 
because of the small numbers involved, it is appropriate 
in this case to look at the statistical significance of the dif­
ference in the overall promotion rates in order to deter­
mine whether the selection process had an adverse impact. 
Applying standard deviation analysis, however, the dif­
ference in selection rates is 1.34 standard deviations, a 
measure well below the 2 or 3 standard deviations men­
tioned most often in court decisions as constituting 
statistical significance.

Moreover, while plaintiffs argue that the court below 
committed legal error in failing to compare the relative 
ranks of Black and White candidates on the resulting 
eligibility list, we maintain that plaintiffs’ argument on 
these facts is misplaced. Contrary to plaintiffs’ emphasis 
on relative rank, Title VII does not require that employers 
equalize scores or equalize the probabilities that all in­
dividuals in a protected class will be promoted. Rather, 
it requires that members of protected groups enjoy relative­
ly equal opportunities to be promoted. Here, there were 
proportionally equal numbers of Blacks and Whites cer­
tified as eligible for promotion and some Blacks significant­
ly outperformed others. Moreover, there were relatively 
few positions to be filled. Therefore, the district court cor­
rectly determined that since a significant number of Black 
candidates completed the entire process and qualified for 
these promotions, no adverse impact was shown against 
Black candidates.



12

ARGUMENT

THE SELECTION PROCESS FOR POLICE 
SERGEANT HAD NO DISPARATE IMPACT ON 
BLACK CANDIDATES

A. Standard o f  Review and Proceedings Under Review

In the court below, the parties stipulated that the only 
issue on the merits for the court to resolve was the issue 
of whether the examination process under challenge had 
an adverse impact on Black candidates (J.A. 81). The 
defendants agreed that if the district court found that the 
examination had an adverse impact on Blacks, they would 
not seek to litigate the validity of the examination. Based 
on the undisputed material facts, the parties submitted 
cross-motions for summary judgment. We submit that the 
district court correctly determined that application of the 
governing legal standards to the undisputed facts of this 
case leads to but one conclusion — the examination pro­
cess had no significant adverse impact on Black candidates. 
Thus, under the established standards for the grant of sum­
mary judgment, the district court properly entered judg­
ment in favor of the defendants. Celotex Corp. v. Catrett, 
477 U.S. 317, 325 (1986). Since this Court is now called 
upon to review an alleged misapplication (Br. at 13) of 
the governing legal standards to the undisputed facts in 
this case, the applicable standard of review is de novo. 
Pullman-Standard v. Swint, 456 U.S. 273, 289 n.19 (1982).

Plaintiffs contend that “at stake in the instant appeal 
are both  retrospective relief for blacks denied promotions 
as a result of the 1987 Eligibility List, and prospective relief 
regarding the continued use of a selection process similar 
to that utilized in preparing the 1987 List” (Br. at 10, em­
phasis supplied). While the meaning of this statement is 
less than clear, it is plain that there is only one selection



13

process at issue in this appeal. As the district court stated, 
“[t]he underlying claim concerns the legality, under Title 
VII, of a promotional examination given in 1986-1987, the 
results of which were to be used to make all promotions 
to the position of Sergeant for a three-year period from 
1987 through 1990” (J.A. 189-190). Thus, as plaintiffs have 
otherwise acknowledged, the only examination process 
which formed the basis of their underlying claim (Amend. 
Complaint - J.A. 7, 8, 16, 21), and on which the district 
court was asked to rule, was the process which resulted 
in an eligibility list that expired in March of 1990. Although 
a new examination was subsequently administered and has 
resulted in a new eligibility list, the validity of that ex­
amination was not challenged in the proceeding below. 
In fact, plaintiffs, who have informed this Court that ad­
ministrative proceedings are pending before the EEOC on 
claims concerning that separate examination, did not, in 
the court below, attempt to amend their complaint to raise 
allegations arising out of that exam (J.A. 207). In short, 
the only issue presented for review in this case is the district 
court’s rejection of plaintiffs’ claim that the examination 
process and resulting eligibility list, which expired on 
March 30, 1990, had a disparate impact on Black 
candidates.

B. A pplicable L egal Standards

The Supreme Court has now clarified the standards 
which apply to cases such as the one at bar which rely on 
the disparate impact theory to prove discrimination. That 
theory presumes that “some employment practices, 
adopted without a deliberately discriminatory motive, may 
in operation be functionally equivalent to intentional 
discrimination.” Watson v. Fort W orth Bank and Trust, 
487 U.S. 977 (1988). See also, Wards Cove Packing C o., 
Inc. v. Frank Antonio, 109 S.Ct. 2115 (1989). In disparate 
impact cases, “facially neutral employment practices that



14

have significant adverse effects on protected groups have 
been held to violate the Act without proof that the 
employer adopted those practices with a discriminatory 
intent.” Watson v. Fort W orth Bank and Trust, 487 U.S. 
at 977 (emphasis in original).7

The law in this Circuit has long established that in the 
appropriate case, a showing of disparate impact may be 
based on “ ‘statistical evidence showing that an employment 
practice has the effect of denying the members of one race 
equal access to employment opportunities.” ’ Bushey v. New 
York State Civil Service C om m ’n, 733 F.2d 220, 225 (2d 
Cir. 1984) (collecting cases), quoting New York City Transit 
Authority v. Beazer, 440 U.S. 568 (1979). The Supreme 
Court in Watson v. Fort W orth Bank and Trust provided 
generalized guidelines concerning the kind and quantity of 
statistical evidence that is sufficient to raise an inference 
of discriminatory impact. Eschewing “any rigid 
mathematical formula”, the Court explained that “statistical 
disparities must be sufficiently substantial [in order to] raise 
such an inference.” Watson v. Fort Worth Bank and Trust, 
487 U.S. at 995 (emphasis supplied). The Court emphasized 
that no consensus has developed around any one 
mathematical standard for demonstrating disparities in 
selection rates and “courts appear generally to have judged 
the ‘significance’ or ‘substantiality’ of numerical disparities 
on a case-by-case basis.” 487 U.S. at 995 n.3.

While no consensus has developed around any one 
theory, two approaches to the issue of whether a disparity

7 Once a prima facie case of discrimination based on a showing of 
disparate impact is established, then and only then will an employer 
carry the burden of producing evidence of a business justification for 
its employment practice. The courts have made clear, however, that 
while the employer may have the burden of producing evidence of 
business justification, the burden of persuasion always rests with the 
plaintiff. Wards Cove Packing Co., supra.



15

in selection rates is “sufficiently substantial”, both men­
tioned in Watson as providing a point of reference for the 
courts, have emerged. After outlining these approaches, 
we will show that the district court correctly determined 
that their application to the relevant aspects of the selec­
tion process in this case reveals either no disparity between 
the Black and White selection rates at all or, if a slight 
disparity does exist, that the disparity is not statistically 
and/or practically significant, and, therefore, is not “suf­
ficiently substantial.”

The Supreme Court (Castaneda v. Partida, 430 U.S. 
482, 496 n .17 (1977); H azelw ood School District v. United 
States, 433 U.S. 299 (1977)), as well as this Court,8 have 
referred to the standard deviation analysis as a unit of 
measurement to assess whether differences in selection rates

8 In Guardians Ass’n v. Civil Service Comrn ’n, 630 F.2d 79, 86-87 n,4 
(2d Cir. 1980), this Court defined the concept of standard deviation 
as follows:

“The standard deviation for a particular set of data pro­
vides a measure of how much the particular results of that 
data differ from the expected results. In essence, the stan­
dard deviation is a measure of the average variance of the 
sample, that is, the amount by which each item differs from 
the mean. The number of standard deviations by which 
the actual results differ from the expected results can be 
compared to the normal distribution curve, yielding the 
likelihood that this difference would have been the result 
of chance. The likelihood that the actual results will fall 
more than one standard deviation beyond the expected 
results is about 32% . For more than two standard devia­
tions, it is about 4.6%  and for more than three standard 
deviations, it is about .03% . On this basis, the Supreme 
Court concluded in Castaneda that when actual results fell 
more than three standard deviations from the expected 
result (that is, a race-neutral selection), the deviation could 
be regarded as caused by some factor other than chance.”



16

are statistically significant. “The greater the number of 
standard deviations, the less likely it is that chance is the 
cause of any difference between the expected and observed 
results” (Ottaviani v State Univ. o f  New York at New Paltz, 
875 F.2d 365, 372 (2d Cir. 1989) (citation omitted)),9 or, 
in other words, the more likely the difference is statistically 
significant. However, in Watson, the Supreme Court made 
clear that it had “not suggested that any particular number 
of ‘standard deviations’ can determine whether a plaintiff 
has made out a prima facie case in the complex area of 
employment discrimination.” 487 U.S. at 995 n.3 (cita­
tion omitted). This Court, following “recent Supreme 
Court pronouncements”, has soundly rejected the efficacy 
of a “minimum threshold level of statistical significance” 
for determining whether the plaintiff has established a 
prim a fa c ie  case of discrimination.10 Ottaviani, 875 F .2d

9 That portion of Justice O’Connor’s decision in Watson v. Fort Worth 
Bank and Trust, 487 U.S. at 993, dealing with causation requirements 
in disparate impact cases (IID ), commanded a plurality of the Court. 
However, in Wards Cove Packing Co., Inc. v. Frank Antonio, 109 
S.Ct. at 2124, a majority of the Court indicated that the law with 
respect to causation in disparate impact cases was “correctly stated” 
by Justice O’Connor in her opinion in Watson. Moreover, even prior 
to Wards Cove, this Court in Ottaviani essentially embraced the Wat­
son plurality’s analysis when, citing Watson, the Court noted that, 
“recent Supreme Court pronouncements instruct that there simply is 
no minimum threshold level of statistical significance which mandates 
a finding that Title VII plaintiffs have made out a prima facie case.” 
875 F.2d at 373.

10 While Ottaviani dealt with discriminatory treatment and not 
discriminatory impact, statistical analysis is used in a similar fashion 
in both theories. “[Pjlaintiffs in a disparate treatment case frequently 
rely on statistical evidence to establish that there is a disparity between 
the predicted and actual treatment of employees who are members 
of a disadvantaged group. . . .” Ottaviani, 875 F.2d at 370-371. In­
deed, after Watson, the distinction between the two theories is blurred 
by the fact that the Court has held that subjective employment prac­
tices may be judged under the disparate impact model.



17

at 372. In rejecting plaintiffs’ “argument that a finding of 
two standard deviations should be equated with a prima 
facie case of discrimination under Title V II . . .  ”, this 
Court explained:

“It is certainly true that a finding of two to three 
standard deviations can be highly probative of 
discriminatory treatment. * * * As tempting as 
it might be to announce a black letter rule of law, 
however, recent Supreme Court pronouncements 
instruct that there simply is no minimum 
threshold level of statistical significance which 
mandates a finding that Title VII plaintiffs have 
made out a prima facie case. See, e .g ., Watson 
v. Fort W orth Bank b  Trust, * * * * .”

Ottaviani, 875 F.2d at 372-73.

Standard deviation analysis, however, is not the only 
approach the courts have endorsed. Indeed, as will be ex­
plained in detail, a test for statistical significance measured 
by the standard deviation for a particular set of data may 
simply be insufficient, standing alone, to measure the legal 
significance or practical impact of any disparities in the 
selection process in this case.

In fact, as the district court recognized, it is obvious that 
“[w]hen a case of disparate impact rises or falls on a com­
parison of the numbers selected, the point at which there 
is a disparity in selection rates between the races must grow 
so large that it may be said that the numbers alone establish 
a ‘sufficiently substantial’ difference in selection rates to 
warrant an inference of discriminatory impact” (J.A. 196). 
“[Statistical significance is not the same as practical 
significance because in isolation it tells nothing about the 
importance or magnitude of the differences.” Bilingual 
Bicultural Coalition o f  Mass M edia, Inc. v. FCC, 595 F.2d



18

621, 641 n. 57 (D.C. Cir. 1978), quotingH. Blalock, Social 
Statistics at 163 (2d ed. 1972). In other words, in order 
to determine whether a disparity in selection rates is, as 
required by the Supreme Court, “sufficiently substantial”, 
or stated differently, in order to determine the importance 
or magnitude of the differences, courts must determine not 
only the bare statistical significance but also the 
significance of those differences in practical terms.

As the court below correctly held, to aid in making the 
determination of practical significance, guidelines are 
already in place. The Equal Employment Opportunity 
Commission has adopted guidelines on employee selection 
procedures. While the courts are not bound by the 
guidelines, and while the Supreme Court has stated that 
the “four-fifths rule” set forth in those guidelines provides 
no more than a “rule of thumb for the court” (W atson, 
487 U.S. at 995 n.3), this Court has repeatedly recognized 
that the guidelines are useful in determining the substan­
tiality of disparities in selection rates, and has, therefore, 
accorded the guidelines “great deference”. Teal v. Con­
necticut, 645 F.2d 133, 137 n.6 (2d Cir. 1981), q f f ’d, 457 
U.S. 440 (1982); Bushey v. New York State Civil Service 
C om m ’n, 733 F.2d 220, 225 (2d Cir. 1984), cert, denied, 
469 U.S. 1117 (1985); Guardians Ass’n v. Civil Service 
C om m ’n, 630 F.2d 79, 88 (2d Cir. 1980).

As is particularly pertinent here, the guidelines, in­
cluding the “four-fifths rule” contained in the guidelines, 
assess the difference in proportions of selection from a 
statistical and practical standpoint. The “four-fifths rule” 
or “80% rule” provides in pertinent part:

“D. Adverse im pact and the four-fifths rule.

A selection rate for any race, sex, or ethnic group
which is less than four-fifths (4/5) (or eighty



19

percent) of the rate for the group with the highest 
rate will generally be regarded by the Federal en­
forcement agencies as evidence of adverse impact, 
while a greater than four-fifths rate will generally 
not be regarded by Federal enforcement agen­
cies as evidence of adverse impact. Smaller dif­
ferences in selection rate may nevertheless con­
stitute adverse impact, where they are significant 
in both statistical and practical terms or where 
a user’s actions have discouraged applicants 
disproportionately on grounds of race, sex, or 
ethnic group. Greater differences in selection rate 
may not constitute adverse impact where the dif­
ferences are based on small numbers and are not 
statistically significant, or where special recruit­
ing or other programs cause the pool of minority 
or female candidates to be atypical of the normal 
pool of applicants from that group. 29 C .F.R . 
1607.4D (Exh. A to Plaintiffs’ Brief at iii).”

It should be noted that this Court has used both  stan­
dard deviation analysis and the four-fifths rule to deter­
mine the impact of a selection process which is alleged to 
have a discriminatory effect. See, e.g., Guardians Assn  
v. Civil Service C om m ’n, 630 F.2d at 88. Further, courts 
have recognized that, although the difference in selection 
rates between two groups may be statistically significant, 
because of the numbers involved, the magnitude of the dif­
ference between the groups may be so small that it is ap­
propriate to conclude that the difference is not practical­
ly significant, or, in other words, is, in real terms, simply 
an insignificant basis on which to find discriminatory 
impact.

As we will show in the following section, when the ex­
amination process under challenge here is scrutinized with 
an eye toward statistical, as well as practical, significance,



20

there are no disparities between the Black and White selec­
tion rates which warrant a finding of discriminatory 
impact.

C . The M ulti-Component Examination Process H ad No
D isparate Im pact on B lack Candidates

1. The W ritten Exam ination

To assess whether the selection process had an adverse 
impact on Black candidates, it is necessary to first conduct 
an analysis of the pass rates of Blacks and Whites on the 
written component of the examination. As the Port 
Authority’s expert, Dr. Abrams, explained in her affidavit, 
“in analyzing adverse impact of multi component selec­
tion procedures, it is important to analyze any separate 
components in the process which serve to exclude par­
ticipants from subsequent components of the selection pro­
cess” (J.A. 37) (emphasis supplied). This is the case because 
the Supreme Court has determined, in affirming a deci­
sion of this court, that a protected employee who 
demonstrates that a “pass/fail barrier” in a multi- 
component selection process had a discriminatory impact 
on his class makes out a prim a fa c ie  case of discrimina­
tion under Title VII despite the fact that the “bottom-line” 
of the selection process had no adverse impact against the 
class as a whole. Connecticut v. Teal, 457 U.S. 440 (1982). 
The plaintiffs in Teal had all failed the first step in a selec­
tion process for permanent supervisory promotions — a 
written examination. Plaintiffs, who were able to show 
that Blacks passed the written test at a rate less than 80 % 
of the White rate, were determined to have a cause of ac­
tion under Title VII even though the overall promotion 
rate for Blacks was higher than for Whites.

In the instant case, the written component of the ex­
amination was taken by 508 Whites and 64 Blacks. Of



21

those, 455 Whites and 50 Blacks passed the written ex­
amination (J.A. 86, 188). Significantly, the percentage of 
Whites passing was 89.57 % and the percentage of Blacks 
passing was 78.13% (J.A. 86). As the district court 
recognized, the passing rate for Blacks is 87.2%  
(78.13/89.57 = .872) of the passing rate for Whites, a rate 
which plainly is greater than the 80% rule (J.A. 86, 
197-98). Thus, the instant case is clearly distinguishable 
from Teal where the passing rate for Blacks on the writ­
ten exam was less than 80% of the White rate.

Further, to complete the analysis of pass rates, Dr. 
Abrams notes that the Uniform Guidelines state that 
“ ‘small differences in selection rates may nevertheless con­
stitute adverse impact, where they are significant in both 
statistical and practical terms. . . (J.A. 38). Here, a
test for the statistical significance between the difference 
in pass rates yields 2.68 standard deviations (J.A. 124, 145). 
However, as the district court noted, “had two more Black 
candidates passed the examination, the difference in pass 
rates would no longer be statistically significant” (J.A. 198). 
Obviously, therefore, the district court was correct in 
holding that “[a] proper analysis of a written examination 
with results which are as close as these requires considera­
tion of not only bare statistical significance, but also prac­
tical significance” (J.A. 198). As we have noted, one must 
often look at practical significance because bare statistical 
significance “tells nothing about the importance or 
magnitude of the differences” (Bilingual Bicultural Coali­
tion o f  Mass M edia, Inc. v. FC C , 595 F.2d at 642 n.57), 
and the Supreme Court has made clear that a disparity 
must be “sufficiently substantial” (W atson , supra) to war­
rant an inference that the disparity is attributable to race.

Here, since the difference in pass rates on the written 
examination meets the four-fifths rule, and no statistical 
significance would be shown if only two other Black



22

candidates had passed the examination, the court below 
appropriately concluded that the difference in pass rates 
is simply not practically significant or “sufficiently 
substantial”.11

Instructive in this regard is Jackson  v. Nassau County 
Civil Service C om m ’n, 424 F.Supp. 1162, 1168 (E.D.N.Y. 
1976). In that case, the court, obviously sensitive to prac­
tical significance, refused to find disparate impact after 
evaluating the results of an examination where a signifi­
cant alteration in the passing percentages would occur by 
the simple shift of several minority candidates from the 
failing to the passing column. In this case also, the failure 
of two more Black candidates to pass the written examina­
tion simply does not rise to the level of a significant 
disproportion that cries out for judicial remedy.

Additionally, it is interesting to compare the instant case 
to prior decisions of this Circuit which have found that 
the differences in percentages of pass rates on written ex­
aminations did have an adverse impact. Significantly, those 
cases involved differences in passing rates between Whites 
and minority candidates which were markedly greater 
than the differences in the case at bar.

Thus, in Bushey v. New York State Civil Service 
C om m ’n , 733 F.2d at 225-226 (emphasis supplied), this

11 Plaintiffs would fault the court below for using the four-fifths rule 
as a measure of practical significance. They assert that “[fjailure to meet 
the Guidelines’ four-fifths threshold is precisely what triggers the 
Guidelines’ additional inquiry into statistical and practical significance” 
(Br. at 32). Under plaintiffs’ analysis of the EEOC guidelines, however, 
the court’s alleged error on this point would be of no consequence on 
these facts since it is undisputed that the pass rates on the written com­
ponent meet the four-fifths rule and, therefore, there would be no need 
to further consider statistical or practical significance.



23

court, using the 80 per cent rule, noted that “the passing 
rate of minority candidates [for promotion to Correction 
Captain] was approximately fifty  percent low er  than the 
passing rate of the nonminority candidates.” In Kirkland  
v. New York State D ept, o f  C orrectional Services, 374 
F.Supp. 1361 (S.D.N.Y. 1974), a ff ’d, 520 F.2d420 (2d Cir. 
1975), cert, denied , 429 U.S. 823 (1976), 30.8%  of the 
Whites and only 7.7 % of the Blacks passed the exam for 
promotion to Correction Sergeant. Further, in Guardians 
Assn  v. Civil Service C om m ’n, 630 F.2d at 87, 88, on 
evidence that the minority pass rate for an exam ad­
ministered to determine selection for the position of City 
police officer was about two-fifths the passing rate for 
Whites (and 89 standard deviations), it was determined 
that “[b]y any reasonable measure, including the standard 
deviation . . .  or the four-fifths rule of the EEOC 
Guidelines”, the exam had a disparate impact on Blacks. 
And finally, in Bridgeport Guardians, Inc. v. M em bers o f  
Bridgeport Civil Serv. C om m ’n, 482 F.2d 1333, 1335 (2d 
Cir. 1973), the passing rate for White candidates for the 
position of police officer was three and one-half times the 
passing rate for minorities.

Aware that the actual pass rate for Blacks compared to 
the pass rate for Whites on the written component survives 
the four-fifths rule, plaintiffs argue instead that the Court 
should consider the fact that since the Port Authority 
reached only rank number 85 on the list, a higher score on 
the written test than the actual passing score was needed 
to reach a rank on the Eligible List that would qualify a 
candidate for promotion (Br. at 26-29). Plaintiffs argue that 
“[f]rom a purely mathematical perspective, a score of at 
least 76 was needed for an applicant to be eligible for pro­
motion” (Br. at 27). They assert that the difference in pro­
portion of Blacks scoring 76 or higher demonstrates adverse 
impact under the four-fifths rule and that the statistical 
significance for such proportional difference yields a result



24

of 2.68 standard deviations — also showing adverse im­
pact (Br. at 29). Plaintiffs attempt to support this argu­
ment for the significance of a higher passing score from 
a reading of the Supreme Court’s decision in Connecticut 
v. Teal, which they argue stands for the proposition that 
the test for adverse impact is whether one component of 
an exam process denied minorities to a disproportionate 
degree the opportunity to be promoted (Br. at 26). There 
are, however, several fallacies in plaintiffs’ reasoning, and, 
despite plaintiffs’ attempt to squeeze the very different facts 
of this case into the principle announced in Teal, that case 
is readily distinguishable.

Significantly, in Teal, all plaintiffs fa iled  the written ex­
amination component and were thus barred from proceed­
ing to the next step in the examination process. Here, in 
sharp contrast to that situation, when a separate analysis 
of the pass/fail rates on the written exam is undertaken, 
no disparate impact on Black candidates is shown. Once 
this pass/fail analysis is performed, Teal does not require, 
and it is not necessary for, the Court to separately analyze 
any disparity7 in the subscores or in the performance on sub­
tests of the examination because the actual selection of those 
to be promoted was based on the overall ranking on the 
list which, in turn, was based on a combination of scores 
on all components (J.A. 84, 199). Indeed, for this reason, 
as this Court has noted, “[wjhere all of the candidates par­
ticipate in the entire selection process, and the overall results 
reveal no significant disparity of impact, scrutinizing in­
dividual questions or individual sub-tests would, indeed, 
‘conflict with the dictates of common sense.’ ” Teal, 645 
F.2d at 138, quoting Kirkland  v. New York State Dept, 
o f Correctional Services, 374 F.Supp. 1361, 1370 (S.D.N.Y. 
1974), a f f ’d, 520 F.2d 420 (2d Cir. 1975).

At bottom, plaintiffs would apparently translate the 
language in Teal, which speaks of equality of opportunities



25

for individuals, into a requirement that a multi-component 
exam be constituted so that individuals from protected and 
unprotected groups score as well on each component of the 
examination, even though it is a composite score from all 
components which results in the final ranking and ultimate 
selection of candidates. But plaintiffs misconstrue Teal. 
Contrary to plaintiffs’ analysis, Teal does not dictate an 
inquiry into individual performance on subtests or, follow­
ing plaintiffs’ logic, individual performance on questions 
in a subtest. See, Connecticut v. Teal, 457 U.S. 440, 464 
n.8 (Powell, J . ,  dissenting) (“Another possibility [flowing 
from the majority decision] is that employers may integrate 
consideration of test results into one overall hiring deci­
sion based on that ‘factor’ and  additional factors. Such a 
process would not, even under the Court’s reasoning, result 
in a finding of discrimination on the basis of disparate im­
pact unless the actual hiring decisions had a disparate im­
pact on the minority group.” (emphasis in original).

In fact, “common sense” dictates that since there is a 
multi-component exam involved, any legally significant 
disparity in scores achieved by a protected class on one 
component of the examination will be reflected in the ag­
gregate results for that class and, thus, in the overall pro­
motion rates. Here, while a comparison of pass/fail rates 
for Blacks and Whites on the written exam satisfies the 80 
percent rule, it is undisputed that Whites did score better 
than Blacks on this exam component. However, it is also 
undisputed that Blacks outperformed Whites both in pass 
rates and scores on the oral examination component, and 
that the scores from all components were aggregated before 
candidates were placed on the eligibility list.

Most importantly, as the court below recognized, an 
analysis of the overall results of the process — the actual 
promotion rates — reveals no statistical disparity between 
the numbers of Blacks and the numbers of Whites



26

promoted (J.A. 200). Further, the facts reveal that the dif­
ference in overall average scores achieved by Blacks and 
Whites on the total test process is so slight that, even if 
those scores were determinative of adverse impact, the dif­
ference amounts to only 2.2 written exam questions (J.A. 
150).

Arguing further, plaintiffs state that a comparison of 
performance at the established pass rate for the written 
component (66% and above) “does not meet the initial 
four-fifths threshold” of the guidelines (Br. at 30). Actually, 
since the pass rate for Blacks was 87.2% of the pass rate 
for Whites, the rate is greater than 80 % and, therefore, 
the court below correctly determined that a comparison 
of pass rates does not violate the four-fifths or 80 % rule 
(J.A. 197-98). Plaintiffs, however, apparently fault the 
district court for failing to find, as legally determinative 
of adverse impact, the fact that this difference in pass rates 
was, nonetheless, statistically significant at 2.68 standard 
deviations (J.A. 88) (Br. at 32). Actually, the EEOC 
guidelines do not, by their terms (supra, at 18, 19), man­
date an inquiry into statistical significance if, as here, the 
four-fifths rule of thumb is not violated. Further, “there 
simply is no minimum threshold level of statistical 
significance which mandates a finding that Title VII plain­
tiffs have made out a prima facie case” (Ottaviani, 875 
F.2d at 372-73). Also, while the guidelines indicate that 
small differences in selection rates (rates that exceed 80 %) 
may constitute adverse impact if they are statistically and 
practically significant, as we have noted, the statistical dif­
ferences present here would disappear if only two addi­
tional Blacks had passed the examination.

Plaintiffs note that an additional two blacks scoring 
higher than the established pass rate of 66 % would not 
equalize the pass rates between Blacks and Whites (Br. at 
34). However, as we have noted, those pass rates already



27

meet the four-fifths rule and such a hypothetical change 
would bring the pass rates even closer together (81.25% 
Black pass rate divided by 89.57 % White pass rate equals 
90.71% ). Thus, under this hypothetical, while absolute 
equality would not be achieved, the point for practical 
significance is that a small shift in the pass rates would 
bring those rates much closer together. Therefore, under 
the present state of the law, and on these facts, it may not 
be maintained that the district court committed legal er­
ror in failing to find adverse impact from the differences 
in pass rates on the written component.12

2. The Oral Exam ination and the Perform ance  
Appraisal

Three of the seven named plaintiffs based their allega­
tions of discrimination on the alleged discriminatory im­
pact of the oral examinations (J.A. 95). However, by any 
measure and plaintiffs do not argue otherwise, the oral ex­
amination as well as the performance appraisal component 
did not have a disparate impact on Black police officers. 
In fact, on the oral, the pass rate for Blacks at 67.35 % was 
greater than the 57.58 % pass rate for Whites (J.A. 90). Fur­
ther, the average oral examination score of 73.46 for Blacks 
was higher than the mean score of 70.86 for Whites (J.A. 
90). Indeed, a test for statistical significance between White 
and Black mean scores on the oral examination yields a 
negative difference of 1.49 standard deviations.

12 Plaintiffs cite as error the district court’s analysis of practical 
significance of the pass rates arguing that statistical significance, if 
it exists, may not be nullified by a shift in pass rates (Br. at 34). While 
statistical significance may not be nullified in such a manner, in the 
wake of the Supreme Court’s determination that differences in selec­
tion rates must be “sufficiently substantial” in order to be legally signifi­
cant, it is entirely appropriate for a court to look at the “magnitude” 
of such differences in terms of actual numbers of individuals impacted 
by the process.



28

Also, on the performance appraisal, while there was no 
minimum score needed in order to be placed on the 
eligibility list, the mean score for Whites was 94.37 and 
for Blacks the mean score was 94.17 (J.A. 91). A difference 
of only . 16 standard deviations results from a comparison 
of mean Black and White scores on this component.

Therefore, since Blacks outperformed Whites both in 
pass rates and scores on the oral examination component 
and scored almost equally as well as Whites on the per­
formance appraisal component of the examination process, 
there was no adverse impact posed by those components 
of the process.

3. The Eligibility List

The multi-component examination process resulted in 
the ranking of the 310 successful candidates on the ver­
tical promotion list (J.A. 97). As of March 30, 1990, the 
date the eligibility list expired, 79 officers had been pro­
moted and the 85th name had been reached on the eligibili­
ty list (J.A. 91-92). As of that date, the promotion rate for 
Blacks was 7.9 % (5 divided by 63) and the promotion rate 
for Whites was 14.0% (70 divided by 499) (J.A. 124, 148).13 
The promotion rate for Blacks is thus 57 % of the promotion

13 In computing the promotion rates for the two groups, the plaintiffs 
would have included the six grandfathered individuals as well as the 
two individuals who retired before being offered promotion. The Port 
Authority defendants took the position that the grandfathered in­
dividuals should be excluded since they did not participate in the ex­
amination process, and the retirees should be excluded because they 
removed themselves from the selection process before being offered 
promotion (J.A. 124, 125, 147-148). However, in the court below, and 
in this Court as well (Br. at 4 ,9), plaintiffs agree with defendants’ 
assessment that, whatever methodology one selects for this computa­
tion, such selection will not materially affect the resolution of this issue 
(J.A. 94, 123).



29

rate for Whites. However, because of the small numbers 
(only 5 Blacks and 70 Whites selected) on which this 
analysis is based, the district court correctly determined 
that it would be appropriate to look at the statistical im­
pact of these differences.

Once more, the Uniform Guidelines have anticipated 
a result such as this and provide ample support for the 
court’s analysis. As the Port Authority’s expert, Dr. 
Abrams, noted in her affidavit, the Guidelines state that 
“ ‘[gjreater differences in selection rate may not constitute 
adverse impact where the differences are based on small 
numbers and are not statistically significant . . . .’ ” (J.A. 
41); 29 C .F .R . §1607.4D. Therefore, in analyzing the ac­
tual promotion rates, comparisons which involve small 
numbers, the district court correctly held that the proper 
approach is to perform a statistical analysis of the dispari­
ty in selection rates (J.A. 200). Performing standard devia­
tion analysis to measure the impact of the difference in 
promotion rates, that difference is 1.34 standard devia­
tions (J.A. 125, 148). As the court below held, “[tjhis is 
well below the level of 2 or 3 standard deviations” cited 
as significant by the courts (J.A. 200).

Indeed, a similar analysis of current promotions was 
sanctioned by this Court in K irkland  v. New York State 
Dept, o f  Correctional Services, 711 F.2d 1117, 1131 (2d 
Cir. 1983) (Correction Lieutenants). That case dealt with 
a claim that the rank ordering on a vertical list used for 
promotion to Correction Lieutenant had an adverse im­
pact on Blacks (id., at n.5). This court sustained the district 
court’s approval of a voluntary settlement of the case, af­
firming the district court’s finding that there existed a 
prim a fa c ie  case of discrimination on the basis of adverse 
impact. This Court summarized the district court’s analysis 
as follows:



30

“[The district court judge] determined that a 
prim a fa c ie  case of employment discrimination 
had been established after reviewing the statistics 
relevant to Exam 36-808 and its eligibility list.
552 F.Supp. at 670. Finding that the d ifferen ce  
betw een the percentage o f  minorities actually ap­
pointed  as o f  July 28, 1982 (9.0% ) and the  
percentage w hich w ould b e  expected to b e  ap ­
pointed from  a random selection am ounted to the 
level o f  5.86 standard deviations, [the judge] 
ruled that the statistics made out a prim a fa c ie  
case of Title VII discrimination under Castaneda 
v. Partida, 430 U.S. 482, 97 S.Ct. 1272, 51 
L.Ed.2d 498 (1977). Castaneda  stated that, in 
cases involving significant statistical samples, ‘if 
the difference between the expected value [from 
a random selection] and the observed number is 
greater than two or three standard deviations,’ 
a prim a fa c ie  case is established since the devia­
tion then could only be regarded as caused by 
some factor other than chance.” 711 F.2d at 1131 
(emphasis supplied, footnote omitted).

Therefore, the difference of 1.34 standard deviations for 
current promotions falls well below the 2 or 3 standard 
deviations cited by this Court in Kirkland. Moreover, it 
should be emphasized that, as we explained, this Court, 
following “recent Supreme Court pronouncements”, has 
noted that even a difference in selection rates of 2 or 3 stan­
dard deviations will not automatically be translated into 
a finding that plaintiffs have established a prim a fa c ie  case 
of discrimination. Ottaviani, 875 F.2d at 372-73.

Plaintiffs contend, however, that they are entitled, as 
a matter of law, to an inference of discrimination because 
the actual selection rates exceed the four-fifths rule (Br.



31

at 15). They fault the district court for looking further at 
the statistical significance of those rates. Once more, plain­
tiffs misinterpret the EEO C guidelines.

Those guidelines contemplate that differences in selection 
rates which do not satisfy the 80 % rule may, nonetheless, 
not amount to adverse impact where the differences are 
based on small numbers and are not statistically significant. 
29 C.F.R. §1607.4D. Here, as we have noted, faithful to these 
guidelines, the court below correctly determined that since 
the differences in question were based on small numbers, 
it would be appropriate to look at the statistical significance 
of the difference in the selection rates (J.A. 200).

Plaintiffs, however, argue that the numbers involved here 
are not as small as the numbers used in the example sup­
plied in the EEOC guidelines which are based on an ap­
plicant pool of 20 males and 10 females (Br. at 17). 
However, plaintiffs overlook the fact that the example given 
in the guidelines is just that, an “example”, and is not 
presented as determinative of the issue of what constitutes 
a small sample size. Plaintiffs would also seek to distinguish 
this case from the example in the guidelines (in the form 
of a question and answer), demonstrating that in a case 
of small numbers, a small shift in selection rates may result 
in a selection rate for minorities which is higher than that 
for majorities (Br. at 17). They note that, unlike that ex­
ample, “[i]n the present case, assuming that one or two of 
the promotees were black rather than white, the white pro­
motion rate would remain higher than the black promo­
tion rate” (Br. at 17-18). We would add that, if two of the 
promotees were Black instead of White, a comparison of 
promotion rates would bring the Black promotion rate 
within 80% of the White promotion rate.14 Thus, we

14 Thus, 7 Blacks divided by 64 equals a selection rate of 10.9 %. When 
that rate is compared to 68 Whites divided by 499 or a 13.6% promotion 
rate, the Black promotion rate is 80.14 % of the White promotion rate.



32

submit that plaintiffs’ hypothetical supports the decision 
of the court below to carefully scrutinize the magnitude 
of the differences at issue here and its determination that 
such differences are not sufficiently substantial.

Plaintiffs also argue that the district court committed 
error in failing to consider “additional” evidence of adverse 
impact other than the actual selection rates (Br. at 18). 
Plaintiffs would have a court look at the overall distribu­
tion of Black candidates on the eligibility list and assert 
that, in this case, that factor is determinative because 
Blacks were ranked in greater numbers at the bottom of 
the list. Once more, plaintiffs’ analysis misses the mark.

First, the overall ranks are obviously reflective of overall 
scores achieved on all components of the examination. 
Here, a test for statistical significance in the overall scores 
achieved on all components of the examination yields 2.43 
standard deviations. However, as the undisputed facts 
establish, an analysis of this difference in practical terms 
reveals that the results are in actuality much closer than 
this number might suggest, and amount to no more than 
2.43%  or what amounts to 2.2 written examination 
questions15 (J.A. 150). Since the Supreme Court has made 
clear that a disparity must be “sufficiently substantial”, 
the closeness of these scores in practical terms means that 
the difference in performance of Blacks and Whites on the 
overall examination is simply too close to warrant an in­
ference of discriminatory impact against Blacks.

Indeed, district courts in this circuit, in an effort to assess 
the magnitude of similar differences, have not ignored the

15 The overall mean score for Whites on the written and oral examina­
tion process and the performance evaluation process (mean final rank­
ing score) was 83.023%  and for Blacks that score was 80.597 % . The 
difference between those scores amounts to 2.43%  (J.A. 92, 156).



33

importance of practical significance. Thus, in Jackson  v. 
Nassau County Civil Serv. C om m ’n, supra , at 1162, the 
district court, although not using that terminology, analyzed 
the “practical significance” of a difference in Black and 
White scores which resulted in placement on an eligibili­
ty list for Community Service Assistant. In that case, the 
plaintiffs, as do the plaintiffs in this case, emphasized the 
fact that fewer Black candidates “achieved a score suffi­
cient to place him or her high enough on the eligibility 
list to be offered an appointment.” 424 F.Supp. at 1168. The 
Court rejected plaintiffs’ analysis since the claimed dispari­
ty in placement would have disappeared if two Black can­
didates had each answered one additional question cor­
rectly. See also, W ade v. New York Tel. Co., 500 F.Supp. 
1170, 1180 (S.D.N.Y. 1980) (failing to find an inference of 
racial discrimination due to the discharge of one or two 
extra minority employees when the total number of 
discharges is so small).

Second, plaintiffs’ reliance on comparative ranks is flawed 
because it does not fully take into account the fact that while 
310 individuals were placed on the eligibility list, in fact, 
only 79 positions became open for promotion. Indeed, an 
emphasis on average ranks or even average overall scores 
ignores the critical fact that, ultimately, Title VII neither 
requires that employers equalize scores or even equalize 
probabilities that all individuals from protected groups will 
ultimately be promoted, nor that employers keep promoting 
until selection rates for all racial groups are on complete 
parity. Bather, it requires that members of protected groups 
enjoy relatively equal opportunities for promotion to those 
positions that become available. Here, due to the Port 
Authority’s manpower needs, there were only 79 “Sergeant” 
positions which needed to be filled. Ultimately, what is most 
important is the fact that a significant number of Blacks 
performed well enough on the examination process to 
qualify themselves for promotion to these 79 spots.



34

Further, the cases cited by plaintiffs in support of their 
theory that comparative ranks present evidence of adverse 
impact are readily distinguishable. Plaintiffs claim that in 
Kirkland  v. New York State Dept, o f  Correctional Services, 
711 F.2d at 1122, this Court based its inquiry into disparate 
impact on the fact that “minority representation within 
the eligibility list’s rank-ordering system was dispropor­
tionately low at the list’s top and high at the list’s bottom” 
(Br. at 19). However, as we have noted, in ruling on the 
issue of adverse impact, Kirkland affirmed the lower court’s 
finding which was based on the facts showing that the dif­
ference between the number of minorities appointed and 
the expected number was 5.86 standard deviations. In 
sharp contrast, in the instant case, the difference is 1.34 
standard deviations (J.A. 200).

G ilbert v. City o f  L ittle R ock, A rk., 722 F.2d 1390 (8th 
Cir. 1983) (cited in Br. at 19-20, 23), is also an entirely dif­
ferent case on its facts from the one at bar. In Gilbert, the 
court of appeals remanded, for further proceedings, allega­
tions that a process for promotion of police officers to the 
ranks of Sergeant and Lieutenant had a discriminatory im­
pact on Black candidates. The plaintiffs argue that Gilbert, 
in which virtually all Black candidates were ranked too 
low on the eligibility list to attain the few promotions made 
from the lists certified each year, demonstrates the danger 
present in this case that “employers could easily use selec­
tion practices that cause discrimination and still evade en­
forcement of Title V II.” They assert that “[s]o long as an 
employer used a test to make only a modest number of pro­
motions, a paucity of minority promotions could not by 
itself be statistically significant” (Br. at 22).

First, plaintiffs’ argument in this regard (See also, Br. 
at 21) is confusing because it has never been contended 
that the numbers selected in this case, while small, are too 
small to be the subject of a reliable statistical analysis and



35

plaintiffs have stipulated to the accuracy of these underly­
ing statistics (J.A. 80-81). Second, in G ilbert, the court de­
termined that because of the serious allegations of pervasive 
racial harassment in the police force, on remand, “it is in­
cumbent upon the district court to consider the statistical 
evidence against the background of racial harassment at the 
Little Rock Police Department.” 722 F .2d at 1398. Signi­
ficantly, there are no similar allegations in this case, nor 
is there evidence that the Port Authority, as was the case 
in G ilbert, used the selection process at issue here, which 
has now expired, to avoid minority appointments. 722 F.2d 
at 1397. Indeed, in the court below, plaintiffs specifically 
withdrew their allegations of intentional discrimination. 
Also, in stark contrast to the situation in G ilbert (722 F.2d 
at 1397 n.8) in which the district court was faulted for not 
looking at the overall results including the ranking aspect 
of the promotion system, a significant number of Blacks 
were in fact ranked high enough to be promoted to 
Sergeant in the selection process under challenge.16

The plaintiffs noted in their Statement of Material Facts 
that “[i]f the Port Authority had actually made 145 pro­
motions, the calculated statistical significance of the dif­
ference in black and white promotion rates would have 
been greater than 2.0” (J.A. 94). This statement graphically 
demonstrates the fallacy in plaintiffs’ reliance on a com­
parison of average ranks. In fact, since 258 Whites and 
33 Blacks completed the entire examination process and 
were placed on the eligibility list, if the Port Authority had

16 Plaintiffs also attempt to glean support for their argument on the 
validity of a comparison of overall ranks from Underwood v . State of 
New York, 28 F .E .P . 922 (S.D.N.Y. 1980) (Br. at 19-20). In that case, 
however, in marked contrast to the facts presented here, a comparison 
of majority and minority pass rates on the ranking portion of the exam 
revealed a failure to meet the four-fifths rule and statistical disparities in 
pass rates between 20.476 and 22.312 standard deviations (n.10).



36

exhausted the entire list and promoted all those who were 
considered eligible for promotion, the promotion rate for 
Blacks (33/64 = 51.56% ) would have been virtually iden­
tical to the promotion rate for Whites (258/508 = 
50.79% ). When, as here, there are proportionally equal 
numbers of Blacks and Whites certified as eligible for pro­
motion and some Blacks performed better than others and, 
in addition, there are far more qualified individuals than 
there are positions to be filled, the focus should not be on 
rank but, rather, on whether there was a significant dif­
ference in the promotion rates for those positions that ac­
tually became available.

In sum, when pass rates are analyzed, neither the writ­
ten nor oral examination components had an adverse im­
pact on Black candidates. Further, when the same stan­
dard deviation analysis endorsed by the Courts to measure 
the impact of the written examination is applied to the ac­
tual promotions made from the Eligible List, no signifi­
cant adverse impact is demonstrated. The differences 
measured for actual selection rates fall well below the 2 
or 3 standard deviations which Courts have used as a bench­
mark for establishing a prim a fa c ie  case of discrimination. 
Furthermore, under most recent precedent, even dif­
ferences of 2 or 3 standard deviations are simply insuffi­
cient, standing alone, to automatically establish a prim a 
fa c ie  case. Most important, a significant number of Blacks 
performed well enough on the entire examination to be 
ranked high enough on the list to achieve promotion. 
Therefore, we submit that the district court was entirely 
correct in concluding that plaintiffs have utterly failed to 
show a “sufficiently substantial” difference in White and 
Black selection rates or performance rates which would 
raise an inference of discriminatory impact (J.A. 201).



37

CONCLUSION

For the foregoing reasons, this Court should affirm the 
judgment entered in favor of the defendants.

Dated: New York, New York 
May 17, 1991

On the Brief:

MILTON H. PACHTER 
Attorney fo r  Defendants-Appellees 

The Port Authority o f  New York 
and New Jersey

One World Trade Center - 66N 
New York, NY 10048 
(212) 435-6835

Arthur P. Berg 
Philip A. Maurer 
James Begley 
Carlene V. McIntyre



COUNSEL PRESS
11 EAST 36TH STREET, NEW YORK, NEW YORK 10016 

(212) 685-9800; (516) 222-1021; (914) 682-0992; (908) 494-3366

(119002)