United States v. Georgia Power Company Comments on Brief Amicus Curiae

    IN THE
UNITED STATES COURT OF APPEALS 

FOR THE FIFTH CIRCUIT
No. 71 - 3447

UNITED STATES OF AMERICA,
Plaintiff-Appellant 
and Cross-Appellee,

- vs -
GEORGIA POWER COMPANY, et al..

Defendants-Appellees 
and Cross-Appellants

No. 71 - 3293
CHARLES KING, et al., and WILLIE MOREMAN,

Plaintiffs-Appellants,
- vs -

GEORGIA POWER COMPANY, et al.,
Defendants-Appellees.

COMMENTS ON THE BRIEF AMICUS CURIAE OF THE AMERICAN PSYCHOLOGICAL ASSO­
CIATION BY THE PRIVATE PLAINTIFFS- 

APPELLANTS

ALBERT J. ROSENTHAL 
GEORGE COOPER 
HARRIETT RABB

435 West 116th Street 
New York, N.Y. 10027

Of Counsel

HOWARD MOORE, JR.
ELIZABETH R. RINDSKOPF

75 Piedmont Avenue, N.E. Suite 1154 
Atlanta, Georgia 30303

ISABEL GATES WEBSTER75 Piedmont Avenue, N.E. Suite 1170 
Atlanta, Georgia 30303

JACK GREENBERG
WILLIAM L. ROBINSON
MORRIS J. BALLER

10 Colutiibus Circle - Suite 2030 
New York, N.Y. 10019



IN THE
UNITED STATES COURT OF APPEALS 

FOR THE FIFTH CIRCUIT
No. 71 - 3447

UNITED STATES OF AMERICA,
Plaintiff-Appellant 
and Cross-Appellee.

- vs -
GEORGIA POWER COMPANY, et al.,

Defendants-Appellees 
and Cross-Appellants,

No. 71 - 3293
CHARLES KING, et al., and WILLIE 
MOREMAN,

Plaintiffs-Appellants, 
- vs -

GEORGIA POWER COMPANY, et al.,
Defendants-Appellees.

COMMENTS ON THE BRIEF AMICUS CURIAE OF THE AMERICAN 
PSYCHOLOGICAL ASSOCIATION BY THE PRIVATE PLAINTIFFS-

APPELLANTS.

Pursuant to the court direction dated July 26, 1972, 
the Executive Committee of the Division of Industrial and Organ­
izational Psychology (Division 114), American Psychological 
Association (hereinafter referred to as the A.P.A.) filed a



brief in this case as amicus curiae. This brief discusses 
several of the salient points related to the testing issue in 
this case. Although the brief does not overtly support any 
party, the conclusions in it are consistent with and generally 
support the private plaintiff-appellants' position concerning 
the inadequacy of Georgia Power Company's test validation efforts. 
It should be abundantly clear to the court, especially with the 
added comments of this disinterested amicus curiae with particular 
expertise in the employment testing field, that the Company's 
efforts to validate its tests do not meet professional statisti­
cal standards and are thoroughly inadequate to justify the gross 
discriminatory impact of those tests. The following comments are 
s\ibmitted to direct the court's attention to those critical as- 
pacts of private plaintiffe-appellants' case confirmed by the
A.P.A. (hereinafter private plaintiffs-appellants will be re-

1/ferred to as plaintiffs).
1. Plaintiffs' major point is that the scoring and 

weighting of tests on a composite basis in Dr. Hite's validity 
study bears no resemblance to the actual multiple cutting score 
use of tests by Georgia Power in its employment procedures and

\/ Since A.P.A. does not purport to oppose or support the 
p>osition of any party to the litigation, we do not here either endorse or take issue with any of the A.P.A.'s 
statements. Rather, we restrict these comments to a few 
points showing the relationship of the A.P.A. statements to major aspects of the testing issue.

-  2 -



therefore provides no support for that actual use. (Plaintiffs 
Main Brief, pp. 35-39). The A.P.A. Brief solidly supports this 
point;

If a battery of tests is to be used, the 
method of data analysis should be consistent 
with the use to be made of the scores in 
actual employment practice.

The point is that it would be inappro­
priate to use the composite score technique 
for validation where the multiple cutting score requirement is used in practice, unless 
the purpose of the study were to consider 
adoption of the new technique if it proves sound. (A.P.A. Brief, pp. 17-18)

The Company, of course, did precisely what the A.P.A. says should
not be done.

2. Plaintiffs' second point is that even if Dr. Hite's 
study could be accepted as relevant to the Company's actual test 
use, the failure of the study to produce statistically signifi­
cant correlations at the traditional 5% level of confidence makes 
the study statistically meaningless. (Plaintiffs' Main Brief, 
pp. 39-41) On this point again the A.P.A. brief confirms 
appellants and rejects the company's claim that the 5%  level of 
confidence is improper.

"[I]t is not true that the 5% level of con­
fidence is too stringent." (A.P.A. Brief,P- 28)

It must be added in all candor that the A.P.A. brief does re­
cognize that lesser levels of confidence as low as 10% may some­
times be acceptable. (A.P.A. Brief, pp. 26-27) But the court's 
attention must be directed to p. 42 of plaintiffs' main brief 
where the actual levels of confidence attained by Dr. Hite

- 3 -



utilizing his preferred discriminant analysis technique are 
presented in the last column. Only for two out of the thirteen 
job classifications studied has even a 10% level of confidenceybeen attained.

3. Plaintiffs' final point is that Dr. Hite's study 
is in any event technically deficient and must be rejected be­
cause of these deficiencies. (Plaintiffs' Main Brief, pp. 43-44)
In particular, we noted Dr. Hite's "technique shopping" in a vain 
attempt to find a statistical procedure which would provide sig­
nificant correlations and his failure to cross-validate on another 
sample. (Plaintiffs' Main Brief, pp. 4 3-44) On this point the 
A.P.A. observes that if one validation technique is first attempted 
without success and then followed by a second technique for which 
significance is claimed, then before accepting the new results 
as truly significant they "should be verified by a repetition of 
the study on another sample" (A.P.A. Brief, p. 3?"). Dr. Hite 
concededly first attempted to validate Georgia Power Company's 
employment tests by using multiple regression analysis. Fail

One of the reasons, if not the principal reason, cited by Dr. Hite for his use of discriminant analysis was an alleged 
restriction of range in the samples he studied. However, the 
A.P.A. brief interestingly observes that the techniques of mul­tiple correlation and discriminant function analysis are 
"theoretically identical" (at pp. 34-35). For that reason, opting for discriminant analysis to escape the restriction of range 
problem is highly suspect. Furthermore, the A.P.A. brief notes 
not once but twice that recognized statistical corrections exist 
for situations vdiere the sample suffers from restricted range 
(at pp. 25 and 28), thus rejecting the Company's claim that re­
striction in range problems necessitate the acceptance of low levels of confidence.

- 4 -



ing in this, he switched to variants of discriminant function 
analysis on v^ich he placed primary reliance, except in the case of 
auxiliary equipment operators (where yet a fourth technique was 
relied upon). Putting aside arguendo the issue of whether Dr.
Hite's discriminant analysis results actually warrant their claimed 
validity, certainly these results have not been verified in the 
professional manner indicated by the A.P.A. As the A.P.A. brief 
indicates, without such verification, Georgia Power Company's 
claim of validation becomes no more than speculative.

CONCLUSION
While the A.P.A. brief does not overtly support any party 

in this case, it is not without its own bias. The bias is that of 
an organization whose members' livelihoods depend to a large ex­
tent on the use of tests and related procedures. While that bias 
must be considered in evaluating the A.P.A. brief, it seems clear 
that in the area of establishing professional standards for test 
validation, A.P.A. judgments must be given close attention. We 
commend to the court the views set out in the A.P.A. beief on the 
points summarized above —  going as they do to minimum technical 
and professional standards on which the A.P.A. has special expertise.

Respectfully siibmitted.

ALBERT J. ROSENTHAL 
GEORGE COOPER 
HARRIETT RABB435 West 116th Street 

New York, N.Y. 10027

HOWARD MOORE,/ JR.
ELIZABETH R. \aiNDSKOPF

75 Piedmont Avenue, N.E. Suite 1154 
Atlanta, Georgia 30303

ISABEL GATES WEBSTER75 Piedmont' Avenue, N.E. Suite 1170 
Atlanta, Georgia 30303

Of Counsel
JACK GREENBERG 
WILLIAM L. ROBINSON 
MORRIS J. BALLER10 Columbus Circle - Suite 2030 

New York, N.Y. 10019Attorneys for Private Plaintiffs-Appellants



CERTIFICATE OF SERVICE

The undersigned counsel for plaintiffs-appellants 
King, Moreman, et al. hereby certifies that on the day of 
September, 1972, he served copies of the foregoing Comments on 
the Brief Amicus Curiae of The American Psychological Association 
By The Private Plaintiffs-Appellants upon counsel of record 
for the other parties as listed below by placing said copies 
in the United States mail, airmail postage prepaid.

J. Lewis Sapp, Esq.,
1900 Peachtree Center Building 
230 Peachtree Street, N.W.
Atlanta, Georgia 30303
Robert L. Mitchell, Esq.
1841 First National Bank Building 
Atlanta, Georgia 30303
Stephen Glassman, Esq.
United States Department of Justice 
Civil Rights Division 
Washington, D.C. 20530
David Blasband, Esq.
Linden & Deutsch 
110 East 59th Street 
New York, New York
Lutz A . Prager
General Counsel's Office
Equal Employment Opportunity Commission
1800 G Street, N.W.
Washington, D.C. 20506
James D. Hutchinson, Esq.
1250 Connecticut Avenue, N.W. 
Washington, D.C. 20036

■ -----
Attotney for Pijivate Plaintiffs- 
Appellants, Kd̂ ng and Moreman, et al,