Opposition Brief — San Francisco Police Officers Ass'n v. City & County of San Francisco

No. 92-1298

In The aga

Supreme Court of the United States

October Term, 1992

«

SAN FRANCISCO POLICE OFFICERS ASSOCIATION,

Petitioner,

CITY AND COUNTY OF SAN FRANCISCO, et al.,

OFFICERS FOR JUSTICE, et al.,

UNITED STATES OF AMERICA,

Respondents.

¢

On Petition For A Writ Of Certiorari

To The United States Court Of Appeals

For The Ninth Circuit

*

BRIEF IN OPPOSITION TO

PETITION FOR WRIT OF CERTIORARI

¢

Louise H. RENNE

City Attorney

GeorceE A. RILEY

Special Assistant

to the City Attorney

Counsel of Record

BARRON WEINSTEIN

MattTHew D. Davis

MARGARET W. BAUMGARTNER

Deputy City Attorneys

206 City Hall

San Francisco, CA 94102

(415) 554-6221

Attorneys for Respondents

COCKLE LAW BRIEF PRINTING CO., (800) 225-6964

OR CALL COLLECT (402) 342-283

>

dis

5

File

ves

TABLE OF CONTENTS

EAI S COE EPEC CAB inc ccc ccnansccneccsecan

REASONS FOR DENYING THE WRIT.............

I.

II.

ITI.

THE NINTH CIRCUIT’S APPROVAL OF THE

SLIDING BAND AS A FORM OF VOLUNTARY

AFFIRMATIVE ACTION WAS BASED ON

EXTENSIVE EVIDENCE AND EXPERT TESTI-

PUGNNES cRnkWnNeneevuns se ncnpunenacenueauaus so.

THE NINTH CIRCUIT’S DECISION IS CONSIS-

TENT WITH THE DECISIONS OF THIS AND

OTHER COURTS REGARDING THE USE OF

SOE GI ccc ivevacecasiaeeasinnscaseces

THE NINTH CIRCUIT’S INTERPRETATION OF

THE 1991 CIVIL RIGHTS ACT PROVIDES NO

BASIS FOR GRANTING THE WRIT...........

ee ws | Aree Ter eT Ie ETT TT ETT eee Tee

17

20

ii

TABLE OF AUTHORITIES

Cases

Afro-American Patrolmen’s League v. City of Atlanta,

re Ba. 2i)) Be ae -) eerrerererT TS 21, 22

Bridgeport Guardians v. City of Bridgeport, 933 F.2d

1140 (2d Cir.), cert. denied, Aah ip Ba

BOS HUE 60 vb Vivans sks deus sekas eee ah 8S 13, 20, 21

Castaneda v. Partida, 430 U.S. 482 (1977)............. 14

Donaghy v. City of Omaha, 933 F.2d 1448 (8th Cir.

1991), cert. denied, US. _,. Viz SAR. Fae

CUPID on hex i dnc cence sxedesawe ees cneneeeeeeee 22

Green v. Bock Laundry Machine Co., 490 U.S. 504

t) | Perr errr rere rrr Terr rr rr rrr err 24

Higgins v. City of Vallejo, 823 F.2d 351 (9th Cir.

1987), cert. denied, 489 U.S. 1051 (1989)......... 13, 22

Johnson v. Transportation Agency, 480 U.S. 616

Lo: ee eee errr peer e rrr re rrr eT: ere 22

Midlantic Nat'l Bank v. New Jersey Dept of Envtl.

Protection, G4 US. 406 CiGG6). «occ sos sevnunacsises 24

Officers for Justice v. Civil Service Comm’‘n, 371

Roepe. Tome OCD. Cal. BOGE. 605s cavixcisaneeeios: 5

Officers for Justice v. Civil Service Comm'n, 14 Empl.

Prac. Dec. (CCH) $7548 G00. Cal. 1977)... 5.60545 5

Officers for Justice v. Civil Service Comm’n, 14 Empl.

Prac. Dec. (CCH) 47549 (N.D. Cal. 1977)........... 6

Officers for Justice v. Civil Service Comm'n, 473

F.Supp. 801 (N.D. Cal. 1979), aff’d, 688 F.2d 615

(9th Cir. 1982), cert. denied, 459 U.S. 1217 (1983) ..... 2

Officers for Justice v. Civil Service Comm'n, 979 F.2d

ih, Se Bs errrrrerrrrr corr ee passim

iii

TABLE OF AUTHORITIES - Continued

Page

Price Waterhouse v. Hopkins, 490 U.S. 228 (1989)...... 24

San Francisco Police Officers’ Ass’n v. City and

County of San Francisco, 621 F.Supp. 1225 (N.D.

erie ound auc 6 bs escnesecadsccccucs 7

San Francisco Police Officers’ Ass'n v. City and

County of San Francisco, 812 F.2d 1125 (9th Cir.

eee wh 5 ss ova hsb es seccscescccscvccs 7

San Francisco Police Officers’ Ass’n v. City and

County of San Francisco, 842 F.2d 1126 (9th Cir.

EE ET ee 8

San Francisco Police Officers’ Ass'n v. City and

County of San Francisco, 869 F.2d 1182 (9th Cir.

1988), cert. denied, 493 U.S. 816 (1989).............. 8

Stuart v. Roache, 951 F.2d 446 (1st Cir. 1991), cert.

denied, __ U.S. __, 112 S.Ct. 1948 (1992)......... 22

United States v. Paradise, 480 U.S. 149 (1987)......... 15

University of California Regents v. Bakke, 438 U.S.

EE 13

Youngblood v. Dalzell, 804 F.2d 360 (6th Cir. 1986),

cert. denied sub nom., Cincinnati Firefighters Union

v. Youngblood, 480 U.S. 935 (1987)............00... 22

STATUTES

Civil Rights Act of 1991 (Pub. L. No. 102-166), 105

MEG ek sea c cn ecasccuccccsccsces passim

iv

TABLE OF AUTHORITIES - Continued

Page

OTHER

Uniform Guidelines On Employee Selection Pro-

cedures

Be ee oe ahaa svkea sas ae sted takanthinencnens 4

29 C.F.R. 1607.3(B)...... AWN SGRRESS SS Soe N ROH ERS: 17, 18

Congressional Reports

H.R. Rep. No. 102-40(II), 102nd Cong., 1st Sess.

65, reprinted in 1992 U.S. Code Cong. & Admin.

PGS a: CUE cree ccs oes oS RRS CR wee 25

STATEMENT OF THE CASE

The petition challenges a decision by the Ninth Cir-

cuit upholding the use of a “sliding band” by the City

and County of San Francisco (“City”). The City proposed

to use the sliding band to make certain promotions

within the San Francisco Police Department. The sliding

band permitted the Police Department to consider the

experience, training, disciplinary record and minority sta-

tus of each candidate whose promotional test score fell

within a precisely defined, narrow numerical range. This

process enabled the City to make some moderate pro-

gress toward satisfying the goals of a consent decree and

ending twenty years of federal litigation.

The sliding band affords greater precision for the use

of test scores than other methods previously approved by

courts. Rather than arbitrarily grouping numerical scores

into a level or grade from which candidates may be

chosen, the sliding band arranges scores based on the

standard concept of statistical significance. The sliding

band ensures that no candidate will be passed over in

favor of another whose test score is lower by a statis-

tically significant factor.

The sliding band permits measured and limited use

of affirmative action to overcome the discriminatory pat-

terns of past practices. The district court’s opinion, which

allows race or gender to be used as one of several factors

in choosing among candidates whose scores show no

reliable difference in their qualifications, rests on sound

scientific principles and extensive expert testimony. No

reported decision of this or any other court implies that

an employer must hire strictly in accordance with the

numerical order of test scores.

Contrary to petitioner’s assertion, the Ninth Circuit’s

opinion is limited to the unique circumstances of this

litigation. The petition, however, fails to present an accu-

rate account of the facts justifying the City’s proposal.

1. The Consent Decree.

In 1979, the district court approved a consent decree

(“Consent Decree” or “Decree”) ending years of bitter

litigation over the hiring and promotional procedures of

the San Francisco Police Department (“SFPD”). The trial

court’s decision was affirmed by the Ninth Circuit in

1982, and this Court denied a petition for writ of cer-

tiorari in 1983. Officers for Justice v. Civil Service Comm'n,

473 F.Supp. 801 (N.D. Cal. 1979), aff'd, 688 F.2d 615 (9th

Cir. 1982), cert. denied, 459 U.S. 1217 (1983).

The Decree settled two federal actions. In 1973, a

group of minority and female police officers called Offi-

cers for Justice (“OFJ”) joined several individuals and

organizations to file a suit against the City alleging racial

and sexual discrimination in the Police Department.

Shortly thereafter, the petitioner, the San Francisco Police

Officers Association (“POA”), intervened as a defendant.

In 1977, the United States filed a separate suit alleging

discrimination in the Department, and in early 1978 the

two actions were consolidated.

As noted by the Ninth Circuit in its opinion affirming

the Decree, the plaintiffs presented a “forceful statistical

demonstration” of prima facie discrimination:

[Minorities and women] appeared to be vastly

underrepresented in the upper ranks of the

department. Moreover, minorities and women

appeared to be concentrated in certain units,

bureaus, and subdivisions in the department,

while virtually absent in others. Plaintiffs

claimed these discrepancies resulted primarily

from the use of unvalidated written examina-

tions and various other class-biased screening

and selection devices.

Id. 688 F.2d at 618-19.

After the district court issued preliminary orders

enjoining certain employment practices, the consolidated

actions proceeded to trial in November, 1978. 473 F.Supp.

at 804. After two weeks of trial, the parties initiated

settlement discussions that resulted in the Consent

Decree. Id. Following a fairness hearing, the district court

approved the Decree in March 1979.

The Consent Decree contains a general prohibition

against “any practice which has the purpose or effect of

unlawfully discriminating against any employee [or

applicant for employment) because of such individual's

race, sex or national origin.” 473 F.Supp. at 811. By its

terms, however, the Decree “provides for specific, defin-

able and good faith efforts to be made [by the City] to

achieve certain goals for employment of women and

minorities . . . within specified periods of time.” Id. at

810. These goals apply to hiring and promotional prac-

tices:

In order to eradicate the present effects of past

hiring and promotion practices with respect to

minorities and women, the City shall adopt and

seek to achieve as a long-term goal the recruit-

ment, appointment and promotion of qualified

minorities and women in sufficient numbers so

as to increase substantially the minority and

female composition of the San Francisco Police

Department so that it more nearly reflects the

racial, ethnic and sexual composition of the rele-

vant labor force of the City and County of San

Francisco.

Id. at 813.

The Decree establishes a long-term goal of raising to

45% the minority representation in the Department. The

Decree also sets annual entry level goals of filling 50% of

all vacancies with minorities and 20% of all vacancies

with women. Id. For each of the promotional ranks of

assistant inspector, sergeant, lieutenant, and captain, the

Decree establishes a goal of appointing minorities and

women in proportion to their representation in the quali-

fied applicant pool. Id. at 814-15. The Decree provides

that the use of an examination will not justify the failure

to meet a goal unless the examination is shown to be

valid under the Uniform Guidelines On Employee Selection

Procedures, 29 C.F.R. 1607 (“Uniform Guidelines”). Id. at

813.

The Decree provides that it will continue in full force

and effect for a period of ten vears from its date of

approval. Id. at 820. However, due to the prolonged

delays caused by the litigation discussed below, the City

could not seek an order of dismissal when the Decree was

scheduled to expire in 1989. Thus, four years after its

scheduled termination, the Decree continues to govern

the employment practices of the SFPD.

2. Prior Litigation Over Promotions.

The instant petition concerns promotions to the ranks

of Q-35 Assistant Inspector (“Assistant Inspector”) and

Q-50 Sergeant (“Sergeant”), the first promotional ranks

above entry level police officer.1 Appointment to these

ranks have been the subject of federal litigation for two

decades.

In 1973, shortly after the OF] filed suit the district

court found that the 1971 Sergeant examination had a

disparate impact on minorities. The court ordered the

City to appoint one qualified minority from the examina-

tion list to the rank of Sergeant for every nonminority

selected. Officers For Justice v. Civil Service Comm'n, 371

F.Supp. 1328, 1342 (N.D. Cal. 1973).

In 1977, the district court extended its order requiring

“one for one” permanent promotions to the temporary

promotions of individuals to the rank of Sergeant. Officers

for Justice v. Civil Service Comm'n, 14 Empl. Prac. Dec.

(CCH) 47548 (N.D. Cal. 1977). The district court also

enjoined the City from making appointments from the

results of the 1976 Assistant Inspector examination

because the plaintiffs had raised serious questions

1 In the SFPD promotions to the ranks of Sergeant and

Assistant Inspector are made from the entry level rank of police

officer. Assistant Inspectors become Inspectors after completion

of two years in the rank. A person may hold dual rank. Thus, an

individual who is currently a Sergeant may compete for

appointment to the rank of Assistant Inspector, and an Inspector

may compete for the rank of Sergeant. Persons who achieve a

dual rank are referred to as Sergeant or Inspector depending on

their assignment to patrol or investigative duties.

regarding the adverse impact of that test against minor-

ities. Officers for Justice v. Civil Service Comm’‘n, 14 Empl.

Prac. Dec. (CCH) 97549 (N.D. Cal. 1977).

The early challenges to examination procedures were

resolved through the settlement that resulted in the Con-

sent Decree. The Decree sets forth the following specific

goals:

[T]he City shall, subject to the availability of

qualified minority and female applicants, adopt

and seek to achieve an annual goal of selecting

minorities and women for promotion to perma-

nent Sergeant and Assistant Inspector positions

in proportion to their representation in the qual-

ified applicant pool for each examination.

473 F.Supp. at 815. The Decree provides that the City may

not justify a failure to meet the above goal “unless the

[selection] procedure or examination is shown to be valid

under the Uniform Guidelines.” Id. at 813.

By an order dated October 1, 1979, the District Court

directed the City to establish a “Consent Decree Unit” to

administer and coordinate the City’s obligations under

the Decree. In consultation with the court-appointed

“ Auditor-Monitor,” the Consent Decree Unit was charged

with developing and administering examinations to meet

the Decree’s specific requirements.

Under the direction of the Consent Decree Unit, the

City administered Sergeant and Assistant Inspector

examinations in 1983. Each test comprised three parts:

multiple choice, written communication and oral exam-

ination. Each part was assigned a certain weight as a

percentage of a candidate’s final score. Both examina-

tions, however, produced “severe adverse impact for

minorities.” San Francisco Police Officers’ Ass'n v. City and

County of San Francisco, 621 F.Supp. 1225, 1227 (N.D. Cal.

1985).

A court-appointed expert reviewed the tests and

found that “insufficient attention was paid to the issue of

adverse impact.” Id. at 1228. The expert concluded that

the City did not explore alternatives, especially to the use

of the multiple choice test, that “may have minimized

such impact.” Id. Accordingly, the Civil Service Commis-

sion revised the system by which the examinations’ com-

ponents were weighed in order to reduce the adverse

impact. The multiple choice and written communication

components of the exam were scored on a pass-fail basis;

the oral component of the exam was made the sole rank-

ing device. Id.

The district court approved the reweighing system as

an appropriate means to remedy the adverse impact. Id.

at 1233. The Ninth Circuit reversed this holding, how-

ever, finding that the reweighing was not “a carefully

contoured affirmative action plan.” San Francisco Police

Officers’ Ass'n v. City and County of San Francisco, 812 F.2d

1125, 1132 (9th Cir. 1987). Changing the weights of the

exam components after the results were announced, the

appellate court found, dramatically altered a candidate’s

ability to prepare effectively for the examination. Id. at

1132. Once the results were reported the “City was

required either to validate its initial examination or, if it

could not, to devise and administer an alternative selec-

tion procedure that did not have an adverse impact.” Id.

at 1133. The Ninth Circuit directed the City to discard the

results of the reweighed examinations and administer a

new selection procedure. Id.

On rehearing, the Ninth Circuit refused to penalize

the beneficiaries of the reweighing plan by removing

them from their positions as Sergeants and Assistant

Inspectors. Because the City had abandoned the promo-

tional lists and the existing appointees would not be

replaced, the Ninth Circuit dismissed the appeal as moot

and withdrew the original opinion. San Francisco Police

Officers’ Ass'n v. City and County of San Francisco, 842 F.2d

1126 (9th Cir. 1988). On further rehearing, the Ninth

Circuit reinstated its original opinion but declined to

award any relief. San Francisco Police Officers’ Ass‘n v. City

and County of San Francisco, 869 F.2d 1182 (9th Cir. 1988).

This Court denied the POA’s petition for a writ of cer-

tiorari. San Francisco Police Officers’ Ass'n v. City and

County of San Francisco, 493 U.S. 816 (1989).

3. The Current Dispute Over Promotions.

During the litigation over the 1983 examinations, the

district court issued a “Supplemental Order” in October

1986 setting forth an accelerated schedule for examina-

tions. This order required the administration of two sets

of Sergeant and Assistant Inspector examinations: the

first set by December, 1986, and the second set by Decem-

ber, 1988.

The continuation of litigation in the Ninth Circuit

over the 1983 examinations, however, further postponed

the administration of the revised examinations. By an

order dated August 10, 1989, the district court directed

the City to administer Sergeant and Assistant Inspector

examinations by December, 1989. The order also required

the City to appoint 100 Sergeants and 60 Assistant Inspec-

tors from the results of these examinations.

The City engaged outside testing experts who took

extensive precautions to create valid examinations. The

experts performed a detailed and thorough job analysis

for each position. Based on these studies, the experts

specifically constructed the examinations to minimize the

possibility of adverse impact. Because the professional

literature shows that written and multiple choice exam-

inations tend to produce adverse impact, the test devel-

opers eliminated those components from the test designs.

Instead, the examinations consisted of elaborate “assess-

ment centers” in which candidates were required to per-

form the actual tasks assigned to Assistant Inspectors or

Sergeants. For example, each Assistant Inspector candi-

date interviewed actors posing as witnesses; each Ser-

geant candidate counseled an actor playing the role of a

subordinate officer.

Rather than require candidates to respond in writing,

candidates’ responses were tape recorded for scoring pur-

poses. The only writing required by the examinations was

the report writing actually required by each position. The

audiotapes were graded by racially mixed groups of

“raters.” A statistical analysis was performed to insure

that a rater’s scoring did not vary on the basis of the race

of the candidate. This analysis showed no such rater bias.

Despite the City’s sedulous efforts, the 1989 examina-

tions produced severe adverse impact with regard to

minorities if the candidates were selected in rank order

10

strictly according to test scores. The City, however, pro-

posed to use a “sliding band” as an alternative means to

select candidates from the examination results.

4. The Sliding Band.

Before the 1989 examinations, the applicants had

been informed that the City might use a “sliding band” to

meet the goals of the Decree. Each candidate was given a

brief explanation of the banding process that the City had

used for two years to satisfy the requirements of a similar

consent decree in the City’s fire department. The City

explained that “scores within a band are considered

equivalent” because “persons in the band are more or less

‘tied’ for rank order.” The candidates were also informed

that if banding were used, “some consideration [would]

be given to avoid adverse impact against minorities and

women.”

Based on fundamental statistical principles, the slid-

ing band quantifies the imprecision inherent in employ-

ment testing. The “standard error of measurement” for

any examination indicates the degree to which scores are

influenced by measurement error. In theory, each candi-

date has a “true score” that is not influenced by the errors

and chance factors present in examinations. A true score

is defined as the average score that would be achieved

over an infinite number of administrations of identical

examinations. The true score would emerge as extraneous

factors average out. The standard error of measurement

defines the limits of variance around an observed score

within which a candidate’s true score would fall with a

certain degree of probability.

11

The standard error of measurement is computed by

using a common formula that relates the standard devia-

tion of the scores actually produced by the examination

to the reliability coefficient for the examination.? The

result indicates a numerical range in which the differ-

ences between scores reflect, with a high degree of confi-

dence, measurement error rather than actual differences

in the knowledge, skills and ability of candidates.>

Scores that are separated by less than the width of

the band are considered substantially equal for purposes

of selection. Because the band is based on the relative

difference between scores, it has a function described as

“sliding.” Once the candidate with the highest score

within a band is selected, the band “slides” so that its

upper limit rests on the highest score remaining on the

list. The band then comprises all scores that are not

separated by a statistically significant factor.

2 The standard error of measurement is computed by multi-

plying the standard deviation of scores from the examination

times the square root of 1 minus the reliability coefficient (SEM

=SDx ~/T-RXX). The standard deviation measures the amount

of actual variance on either side of the mean score of a particular

examination. The greater the variance of score distribution

around the mean, the higher the standard deviation. On a nor-

mal, or standard, distribution, 95% of the scores fall within plus

or minus two standard deviations from the mean score. The

reliability coefficient measures the internal consistency among

various components of the examination as a measure of relevant

job qualifications.

3 The City multiplied the standard error of measurement by

a confidence factor of 1.96 to derive the final band width. Statis-

tical theory teaches that a factor of 1.96 times the SEM will

indicate the range in which a candidate’s true score will fall

with 95% probability.

12

For the tests in question, application of the formula

produced narrow bands. Each examination was graded

on a 100 point scale. The band for the Sergeant examina-

tion was 4.0 points; the band was 5.48 points for the

Assistant Inspector examination.

When the parties failed to reach a negotiated solu-

tion, the City filed a motion for a declaratory judgment

approving its proposal to use the sliding band. The City

requested permission to make 15 Sergeant promotions in

addition to the 100 promotions required by the Court’s

order. The first 100 promotions would be made in rank

order to meet any expectations created by the district

court’s earlier orders. The additional 15 Sergeant promo-

tions would be made using the band. The City also pro-

posed to use the sliding band to appoint no more than 7

Assistant Inspectors in addition to 60 appointed in rank

order.

The district court heard three days of expert testi-

mony on the City’s motion for declaratory relief. The City

called an expert on statistics, Dr. Bernard Sisken, to dem-

onstrate the validity of banding as a method of selecting

candidates. Dr. Sisken testified that the City’s proposal

was an appropriate use of statistics to define reliable

differences between the scores of candidates. The City

also called two experts in psychometrics, Dr. Sheldon

Zedeck and Dr. James Outtz, who agreed that the City’s

method of banding was substantially equally valid to

rank order selection for purposes of the Uniform Guide-

lines.

The POA called two experts, Dr. Ingram Olkin, an

expert on statistics, and Dr. Lawrence Meyers, an expert

in psychometrics. Although they disagreed with the

rear nes

13

City’s interpretation, both POA experts agreed on the

City’s computation of the band and the concept of signifi-

cant differences between scores as measured by the band.

As described above, the City’s initial banding pro-

posal relied on minority status as the sole criterion for

selection of candidates whose scores fell within the same

band. On June 6, 1991, the district court declared that the

City’s plan violated: the restrictions on race conscious

remedies set forth by this Court in University of California

Regents v. Bakke, 438 U.S. 265 (1978), and by the Ninth

Circuit in Higgins v. City of Vallejo, 823 F.2d 351 (9th Cir.

1987), cert. denied, 489 U.S. 1051 (1989). (Pet. App. at

48a-59a) The district court recognized, however, that

banding was an appropriate use of test scores and invited

the City to submit a modified proposal that included

other criteria. The Court urged the City to consider the

approach accepted by the Second Circuit in Bridgeport

Guardians v. City of Bridgeport, 933 F.2d 1140 (2nd Cir.),

cert. denied, ___ U.S. ___, 112 S.Ct. 337 (1991). (Id. at 58a)

On June 28, 1991, the City proposed certain addi-

tional factors to select among candidates whose scores

fell in the same band. On August 21, 1991, the district

court issued the order approving the City’s proposal. (Id.

at 15a — 47a) The court specifically approved the City’s

calculation of a band for each examination based on the

standard error of measure. (Id. at 31a) The court accepted

the City’s proposal to treat persons whose scores fall

below the lower limit' uf a band as significantly less

qualified than those persons whose scores are within the

band. (Id. at 32a) The court further found that “the City’s

proposed selection criteria are legally sound.” (Id. at 45a)

These criteria include professional conduct, education,

14

training and experience as well as the Department’s affir-

mative action goals. (Id. at 46a)

5. The Ninth Circuit Decision.

In an opinion by Chief Judge J. Clifford Wallace, a

panel of the Ninth Circuit unanimously affirmed the Dis-

trict Court’s order. Officers for Justice v. Civil Service

Comm’‘n, 979 F.2d 721 (9th Cir. 1992) (Pet. App. at la-14a).

The opinion noted that the district court had heard exten-

sive expert testimony in support of the plan, and that the

principle of statistical deviation underlying banding was

approved in Castaneda v. Partida, 430 U.S. 482 (1977). 979

F.2d at 727. The Ninth Circuit specifically rejected the

argument, advanced in the instant petition, that an

employer is required by federal law to hire in strict rank

order even when scores are not separated by a significant

difference:

The district court heard evidence that the test

scores themselves are imprecise measurements

of future job performance and that by banding

the scores and considering secondary criteria in

making promotion decisions, the City may

increase the probability that the most qualified

candidates will be promoted. The district court

did not clearly err in finding that banding, as

proposed by the City, is more valid, or at least

“substantially equally valid” to rank order pro-

motions.

Id. at 728.

The Court of Appeals also held that City did not

violate equal protection or Title VII through the limited

15

use of race or gender as factors together with other

relevant criteria to select candidates from within the

band. The Court held that in light of the “historical

context” of this litigation, the “City’s admission of past

discrimination” and the adverse impact of the examina-

tions, there was a strong basis in the evidence to satisfy

the standard for voluntary affirmative action under

United States v. Paradise, 480 U.S. 149 (1987). 979 F.2d at

726.

The Court of Appeals rejected the POA’s argument,

also repeated in the petition, that Congress outlawed all

forms of race conscious, affirmative action by enacting

the Civil Rights Act of 1991 (Pub. L. No. 102-166, 105 Stat.

1071 (1991) (the “1991 Act” or “Act”)). The POA argued

that such practices are prohibited by section 107 of the

Act, which makes unlawful employment practices in

which race or gender “is a motivating factor.” The Ninth

Circuit easily dismissed this argument, noting that sec-

tion 116 of the Act exempts lawful affirmative action

programs from the prohibitions in section 107.

¢

REASONS FOR DENYING THE WRIT

The Ninth Circuit’s decision reflects the unique fac-

tual history of this case. Rather than risk more intrusive

forms of court-ordered relief, the City entered into the

Consent Decree which mandates affirmative measures to

overcome the discriminatory results of past practices. The

appellate court approved the City’s proposal to meet the

terms of the Decree by using the sliding band as an

alternative to rank order selection. The decision does not,

16

as petitioner asserts, require employers who have no

similar history or legal obligations to use a sliding band

to overcome adverse impact.

Petitioner argues that the decision is an “unwar-

ranted federal intrusion” into the personnel practices of

state and local agencies. The City, however, was never

ordered to use the sliding band. The City sought court

approval for its proposal as a form of voluntary affirma-

tive action that was justified by the goals of the Consent

Decree and supported by the evidence of past discrimina-

tory practices which made the Decree necessary. The

sliding band enabled the Police Department to consider a

candidate’s experience and disciplinary record, among

other factors, in making promotions. What would be an

unwarranted federal intrusion is the result petitioner

seeks: a requirement that an employer hire in strict

numerical order from the results of a valid, but imperfect,

examination.

Petitioner’s attempt to gain review of the decision on

grounds relating to the 1991 Civil Rights Act is equally

unavailing. Whether the Act applies retroactively has no

bearing on the outcome; the Ninth Circuit held that the

City’s use of the sliding band under the unique circum-

stances of this case is lawful under the Act as well as

under the law in effect before the Act was passed. Peti-

tioner’s claim that the Act outlaws affirmative action flies

in the face of the statutory language and express Con-

gressional purposes.

17

I. THE NINTH CIRCUIT’S APPROVAL OF THE SLID-

ING BAND AS A FORM OF VOLUNTARY AFFIR-

MATIVE ACTION WAS BASED ON EXTENSIVE

EVIDENCE AND EXPERT TESTIMONY.

Petitioner’s argument that the Ninth Circuit’s deci-

sion is an improper intrusion into civil service procedures

conflates two separate objections. First, petitioner asserts

that the trial and appellate courts erred in the factual

determination that candidates selected through the slid-

ing band and candidates selected in strict rank order are

equally qualified. Second, petitioner claims that the deci-

sion will cause other employers to adopt “ad hoc” pro-

cedures that will undermine merit based selection. These

objections misrepresent the evidence before the trial court

and the holding of the Ninth Circuit.

The City did not propose the sliding band as a

method to select candidates whose qualifications would

“equal” those of candidates chosen in strict rank order

according to numerical scores. The City proposed to use

the sliding band as a method that was “substantially

equally valid” to strict rank order selection within the

meaning of section 1607.3(B) of the Uniform Guidelines.

The City presented the sliding band as a selection pro-

cedure that allowed the Police Department to consider

other qualifications if differences in scores did not indi-

cate reliable differences in ability.

As petitioner’s own expert testified, if the predictive

value of one selection method is not statistically signifi-

cantly lower than another method, the methods are “sub-

stantially equally valid” within the meaning of section

18

1607.3(B) of the Uniform Guidelines. The sliding band pro-

posal is based on this same concept of statistical signifi-

cance. Candidates are treated differently only if their

scores are separated by a statistically significant factor. If

there is such a significant difference, the employer can

assert, with a high degree of confidence, that the differ-

ence in scores will predict a difference in the actual

performance of the job in question.

Test scores that are not significantly different -- test

scores that fall within the band - cannot predict, with a

high degree of reliability, differences in actual job perfor-

mance. This view is commonly accepted in practice. As

the trial testimony showed, many aptitude tests report

- the range of possible variance in a candidate’s test score

due to measurement error. Such reports discourage

undue reliance on small numerical differences in scores.

The traditional system of assigning letter grades to test

results also reflects the view that small differences in

numerical scores are not reliable for purposes of distin-

guishing among students.

Letter grades, percentile ordering and other methods

of arranging scores into groups free employers from a

rigid system in which any difference in a numerical score,

no matter how small, would determine a decision about

hiring or promotion. Civil service systems, for example,

traditionally rely on the “rule of three” to give the

appointing authority the discretion to choose among the

top three scores on a list.

The sliding band affords a more precise use of test

scores than methods such as letter grades or the rule of

three that make arbitrary distinctions among test scores.

a

19

The rule of three, for example, permits an appointing

authority to select the second highest score even when

there is a statistically significant difference between the

first and second score. This result would not occur

through the use of the sliding band: scores are treated

differently only when there is a difference between scores

that indicates a reliable difference in the ability to per-

form the job. Thus, the sliding band preserves merit

based selection with appropriate recognition of the

imprecision and measurement error inherent in employ-

ment testing.

Under the sliding band procedure, if test scores do

not indicate a reliable difference in ability, an employer

may consider other factors that are related to reasonable

objectives. The City proposed to consider work history,

special experience and disciplinary records. Certainly

these are reasonable factors, especially in light of the

duties of a police officer. No court has ever suggested

that such criteria violate federal law.

The City’s procedure also permitted the Police

Department to consider the goals of the Consent Decree

in selecting candidates whose test scores indicate no reli-

able differences in ability. Such limited use of affirmative

action was Clearly justified, as the Ninth Circuit held, by

the evidence of discrimination resulting from earlier

reliance on unvalidated procedures and the City’s failure

to meet the goals of the Consent Decree. Officers for

Justice, 979 F.2d at 726-27.

The Ninth Circuit agreed with the district court,

which has overseen this litigation for twenty years, that

20

the Decree’s prohibition on discrimination did not pre-

clude the use of the sliding band. The Consent Decree’s

requirement that the City undertake good faith, affirma-

tive measures to reach the goals of the Decree clearly

contemplates forms of voluntary affirmative action such

as the sliding band. Id. at 727.

In short, the sliding band is predicated on the notion,

widely accepted in theory and practice, that insignificant

difference in numerical scores cannot reliably predict dif-

ferences in actual job performance. The City’s use of

banding to meet the goals of the Consent Decree was

supported by the particular history and evidence in this

case.

Il. THE NINTH CIRCUIT’S DECISION IS CONSIS-

TENT WITH THE DECISIONS OF THIS AND

OTHER COURTS REGARDING THE USE OF TEST

SCORES.

Petitioner incorrectly suggests that the Ninth Cir-

cuit’s decision contradicts the holdings of other courts

regarding the use of test scores, the interpretation of

consent decrees and the use of affirmative action to select

candidates whose examination scores indicate no reliable

difference in knowledge skills or ability. Petitioner, how-

ever, cites decisions that demonstrate no such conflict.

Petitioner relies on Bridgeport Guardians v. City of

Bridgeport, 933 F.2d 1140 (2d Cir.), cert. denied, __ U.S.

__, 112 S.Ct. 337 (1991), in support of the argument that

banding is appropriate only for poorly designed, unreli-

able examinations. Petitioner argues that the Second Cir-

cuit approved banding because the City of Bridgeport

21

“could not show that the results of the examination were

reliable enough to justify appointments in rank order.”

(Pet. at 13)

Petitioner misreads Bridgeport Guardians. In that case,

the City of Bridgeport had established that the challenged

exam was “content valid and justified by the City’s busi-

ness needs.” 933 F.2d at 1148. Nevertheless, the plaintiffs

“had sufficiently rebutted that evidence by showing that

the City could have elected to use banding in order to

alleviate the disparate racial effect of the examination

without disserving its legitimate needs.” Id. Thus, the

Second Circuit approved the use of the sliding band for a

reliable examination because the terms of a consent

decree and the evidence of disparate impact and past

discrimination justified that remedy.

Citing Afro-American Patrolmen’s League v. City of

Atlanta, 817 F.2d 719 (11th Cir. 1987), petitioner maintains

that the-Ninth Circuit’s decision deprives nonminorities

of enforceable rights under consent decrees. Petitioner

maintains that the decision will act as a disincentive to

public agencies to enter into settlements of Title VII

actions.

Petitioner’s reliance on Afro-American Patrolmen’s

League is also misplaced. In that case, the consent decree

specifically forbade the City of Atlanta from engaging in

affirmative, race conscious relief to meet the terms of a

consent decree. The explicit terms of the decree required

the City to validate examinations and hire from those

examinations without respect to the race or gender of

candidates. Id. at 724.

22

By contrast, the Consent Decree in the instant case

imposes no restriction on the use of test results. As the

Ninth Circuit noted, the Consent Decree contemplates the

type of affirmative relief prohibited by the decree in Afro-

American Patrolmen’s League. The Ninth Circuit concluded

that “the consent decree itself provides a proper rationale

for race conscious promotions and serves as a valid

defense against the [POA]’s reverse discrimination argu-

ments.” Officers for Justice, 979 F.2d at 727.

This Court has recognized that an employer is not

bound by federal law to select the highest scoring candi-

date; selecting a lower scoring but highly qualified candi-

date is proper to meet legitimate business goals including

the terms of a lawful affirmative action plan. Johnson v.

-——Transportation Agency, 480 U.S. 616 (1987). The sliding

band enabled the City to consider other lawful objectives

in a procedure that gives more weight to examination

scores than methods approved by other courts.

It is ironic that petitioner argues that the Ninth Cir-

cuit decision constitutes an unwarranted intrusion into

civil service systems and would discourage public enti-

ties from compromising Title VII actions by entering into

4 See Donaghy v. City of Omaha, 933 F.2d 1448 (8th Cir. 1991),

cert. denied, ___ U.S. ___, 112 S.Ct. 938 (1992) (approving promo-

tions of lower scoring police officers); Stuart v. Roache, 951 F.2d

446 (1st Cir. 1991), cert. denied, __ U.S. ___, 112 S.Ct. 1948 (1992)

(promoting officers to satisfy consent decree); Higgins v. City of

Vallejo, 823 F.2d 351 (9th Cir. 1987), cert. denied, 489 U.S. 1051

(1989) (city’s affirmative action plan is a proper basis to promote

lower scoring minority); Youngblood v. Dalzell, 804 F.2d 360 (6th

Cir. 1986), cert. denied sub nom., Cincinnati Firefighters v. Young-

blood, 480 U.S. 935 (1987) (consent decree based on statistical

disparities justifies promotions of lower scoring minority candi-

dates).

23

consent decrees. The City agreed to the Consent Decree to

avoid the possibility of more drastic forms of court-

imposed relief that might be justified after a finding of

intentional discrimination. Far from constituting interfer-

ence with a civil service system, the sliding band gives

the appointing authority some measure of discretion to

consider other qualifications of candidates while meeting

the goals of a lawful consent decree.

lil. THE NINTH CIRCUIT’S INTERPRETATION OF

THE 1991 CIVIL RIGHTS ACT PROVIDES NO

BASIS FOR GRANTING THE WRIT.

Petitioner argues that the Ninth Circuit’s decision

presents two important questions regarding the inter-

pretation of the 1991 Civil Rights Act. First, petitioner

maintains that the decision should be reviewed to con-

sider the question of the retroactivity of the 1991 Civil

Rights Act. Second, petitioner maintains that the Ninth

Circuit erred when it held that Congress did not outlaw

affirmative action when it passed the 1991 Civil Rights

Act. Neither argument presents appropriate grounds for

granting the writ.

Petitioner claims that this case presents a “unique”

opportunity to decide the issue of the retroactivity of the

1991 Civil Rights Act. This argument ignores the Ninth

Circuit’s actual holding. The appeals court held that it

was unnecessary to reach the issue of retroactivity

because it is clear that the provisions of the Act cited by

the petitioner did not change the law governing the use

of preferences in voluntary affirmative action. Officers for

Justice, 979 F.2d at 725. Thus, the Ninth Circuit’s decision

24

does not squarely present, for purposes of review by writ

of certiorari, the question of whether the Act is retroac-

tive: regardless of whether the law is or is not retroactive,

the decision remains the same.

In the appeal to the Ninth Circuit, petitioner argued

that the sliding band was prohibited by section 107 of the

Act. Section 107 codifies the result in Price Waterhouse v.

Hopkins, 490 U.S. 228 (1989). In that decision, a plurality

of this Court held that a plaintiff in a Title VII action will

prevail if she proves that an impermissible factor “played

a motivating part in an employment decision.” Id. at 250.

The decision expressly provided that it did not apply to

“the special context of affirmative action.” Id. at 239 n.3.

Nothing in section 107 indicates any congressional

purpose to apply the Price Waterhouse rule to the “special

context of affirmative action.” To the contrary, in section

116 of the Act, Congress stated that nothing in the Act

was intended to disturb affirmative action that is lawful

under existing case law. As the Ninth Circuit noted, peti-

tioner’s reading of the 1991 Act poses an internal incon-

sistency: Congress sought to protect affirmative action in

section 116 while outlawing it in section 107. Officers for

Justice, 979 F.2d at 725. Such a reading must be rejected.

“A party contending that legislative action changed

settled law has the burden of showing that the legislature

intended such a change.” Green v. Bock Laundry Machine

Co., 490 U.S. 504, 521 (1981). It is a fundamental rule of

statutory construction that “if Congress intends for legis-

lation to change the interpretation of a judicially created

concept, it makes that intent specific.” Midlantic Nat'l

25

Bank v. New Jersey Dept. of Envtl. Protection, 474 U.S. 494,

501 (1986). In the instant case, petitioner argues that the

1991 Civil Rights Act, while addressing seven specifically

identified Supreme Court decisions, also overturned two

decades of law on affirmative action by implication. Peti-

tioner simply cannot sustain its burden of showing the

section 107 prohibits the otherwise lawful use of prefer-

ences to meet the goals of a consent decree.> Therefore,

petitioner’s arguments regarding the 1991 Civil Rights

Act present no persuasive reason for review of the Ninth

Circuit’s decision.

> Petitioner also argues that the Ninth Circuit’s decision

violates the ban on “race-norming” in section 106 of the Act. In

“race-norming,” candidates were assigned a numerical score in_

relation to the exam performance of persons in the same racial

category. The final scores were used without regard to evidence

of prior discrimination; employers often had no knowledge that

the scores had been adjusted. As the legislative history demon-

strates, this arbitrary practice was not a form of affirmative

action designed to overcome past patterns of unlawful discrimi-

nation. See H.R. Rep. No. 102-40(II), 102nd Cong., 1st Sess. 65,

reprinted in 1992 U.S. Code Cong. & Admin. News 694, 750-51.

The sliding band is clearly not “race-norming” or any form

of score readjustment prohibited by section 106. Petitioner,

moreover, did not properly raise the issue on appeal, and the

Ninth Circuit declined to consider the question. Officers for Jus-

tice, 979 F.2d at 726. Accordingly, petitioner is precluded from

raising the issue in its petition.

26

CONCLUSION

For the reasons set forth above, the petition for writ

of certiorari should be denied.

DATED: March 4, 1993

Respectfully submitted,

Loutse H. RENNE

City Attorney

GeorceE A. RILEY

Special Assistant

to the City Attorney

Attorneys for Respondents