# Joint Appendix — Wards Cove Packing Co. v. Atonio

> Briefs, arguments, decisions, and more.

URL: https://www.frixlaw.com/law-library/documents/brief%3Amicro_IA40385012_0656%3A04

## Record

- **Collection:** Supreme Court brief
- **Document type:** Joint Appendix
- **Published:** January 1, 1989
- **Citation:** 490 U.S. 642

## Text

No. 87-1387

IN THE
Supreme Court of the United States

October Term, 1988

WARDS COVE PACKING COMPANY, INC.,
CASTLE & COOKE, INC.

Petitioners,

Vv.

FRANK ATONIO, et al.,
Respondents.

JOINT APPENDIX

Douglas M. Fryer*
Douglas M. Duncan
Richard L. Phillips
MIKKELBORG, BROZ,
WELLS & FRYER

Suite 3300

1001 Fourth Avenue Plaza
Seattle, Washington 98154
(206) 623-5890

Attorneys for Petitioners
* Counsel of Record

September 9, 1988

ehC 1) Legal Promters Seattle WA

Relevant Docket Entries.

Revised Pretrial Order ...

Deposition of Harold Brindle .

Deposition of Mervin W. Lessley

Deposition of Randall Lee
Mi lhol land . . * * . . . . . *

Deposition of Russell J. Rohrer.

Deposition of Marvin Clay Snyder

Testimony
Testimony
Testimony
Testimony
Testimony
Testimony
Testimony
Testimony
Testimony
Testimony
Testimony

Testimony

of
of
of
of
of
of
of
of
of
of
of

of

Clarke Kido .-...- =.
Frank Peters. ..- -
Stephen Noel Bird .

Richard Gurtiza .-

David Della . .

Orlando Bucsit.-

Andy Pascua .- .-
Michael Tardif. °
Gary Mamallo. ... .-
Mike Eddie Antonio. .
William T. Pascua .- .-

Shirley Jeanne Smith.

TABLE OF CONTENTS (Continued) TABLE OF CONTENTS (Continued)

Page Page
Testimony of H. J. Parrish. . .. . A-105 Testimony of Cheryl Tatom. . .. . A-407

Testimony of Jill Henderson . . . . A-112 Testimony of Charley Anderson. . . A-414

Testimony of Patrick Timothy Testimony of Robert Krapp. ... . A-422

ee se

.
>
!
~
~
>

Testimony of Richard Quirion ... A-432

Testimony of Joel Stromme . . . - - Ani23 Testimoriy of Eberle Mercer .... A-434

Testimony of Albert Belasario
Pastores, Jf. . «© « « © © «© « « « AeASS

A-127 Testimony of Frank Shucka. ... . A-447

Testimony of Earl Anderson .... A-438

Testimony of Eugene Baclig. ....

Testimony of Christopher Steele . . A-131 Testimony of James W. Yonker . . . A-453

Testimony of Phillip Fujii. . . . . A-137 Testimony of John R. Gilbert ... A-460

T -
Testimony of Cathlyn Hjorten. - . r A-141 Testimony of Winn F. Brindle . ° ° A 463

Testimony of Roderick E. Testimony of Larry L. DeFrance .. A-470
COMSEOR: CBee + 5 © Batt re Testimony of Jack Aiello ..... A-577
Testimony of Arsenio Eleccion - - - A-15! Testimony of Laurie Brindle

Testimony of Alec W. Brindle. A-156 Romine - +--+ + + + «+ + + + + + + A-586

Testimony of Warner Leonardo. - - - A-213 Testimony of Elroy Kowalski. . - - A-589

Testimony of John Lum. ...... A-614

.
.
.
>
'
N
w
~

Testimony of Donald E. Wise

Testimony of Dr. Albert Rees. - - - A=250

»
'
w
ra)
x

Testimony of Robert J. Flanagan
Testimony of Gary P. Latham .. - - A-391
Testimony of Ronald Baclig. . .- - A-402

Testimony of Alan Lew .... -+- + + A=403

-69> -iii-

RELEVANT DOCKET ENTRIES

3/20/74 Complaint filed.

5/5/82 Revised Pretrial Order
entered.

5/3 /82-

5/17/82 Trial conducted.

10/31/83 United States District

Court for the Western
District of Washington
opinion entered.

12/6/83 Order correcting opinion -
following nonjury' trial
and directing correction
of judgment entered.

8/16/85 Opinion in the Court of
Appeals affirming the
judgment entered.

11/18/85 Order withdrawing the
Court of Appeals opinion
an@ ordering rehearing en
banc entered.

9/2/87 Second opinion of the
Original panel of - the
Court of Appeals on remand
from the en banc court

entered.

11/12/87 Order Clarifying the
Opinion entered.

11/12/87 Petition for rehearing
denied.

12/23/87 Opinion of the en banc

Court of Appeals entered.

A-1

NO. 87-1387

IN THE SUPREME COURT
OF THE UNITED STATES

October Term 1988

WARDS COVE PACKING COMPANY,
CASTLE & COOKE, INC.,

Petitioners,
vs.
FRANK ATONIO, et al,

Respondents.

BRIEF OF PETITIONERS
APPENDIX

INC.,

UNITED STATES DISTRICT COURT
WESTERN DISTRICT OF WASHINGTON

FRANK ATONIO,
et al.,

Plaintiffs,
NO. C 74-145 S
vs.
REVISED
WARDS COVE PACKING PRETRIAL ORDER
COMPANY, INC.,

et al.,

Defendants.

me ee ee ee ee ee ee

[PTO, p. 2]

ADMITTED FACTS

(2) The following facts are

admitted by the parties. ...
x *&
[PTO, pp. 10, 14)

80. The majority of non-resident
cannery workers are lined up by the
cannery worker foremen after management
has estimated the number that will be

needed.

A-3

[PTO, p. 16]

93. At least during 1970 through
1975, Bumble Bee has not advertised in
media for applicants for any jbos, Nor
[sic] has it posted vacancies at the
cannery.

94. Non-resident cannery workers
are those whose off-season residence is
in the Lower 48. Some have been hired
directly by Bumble Bee in Astoria.

95. At least since 1970 Bumble Bee
has on occasion recruited resident
cannery workers from Alaska Native
villages through bush pilots. (Resident
Cannery workers are those whose off-
season residence is Alaska.)

96. At least since 1970 some non-
resident cannery workers have been hired
directly by Bumble Bee in Astoria. The

majority of non-resident cannery workers

A-4

dispatched to Bumble Bee through Local
37, ILWU have been Filipino.

x *
[PTO, p. 17]

102. At least since 1970 Red Salmon
has on occasion recruited some resident
cannery workers from Alaska Native
villages. The recruitment has been
through village leaders and bush pilots,

among others.

[PTO, p. 18]

104. The machinist foremen at Wards
Cove have on occasion recommended people
for machinist jobs, among other jobs.

105. Wards Cove has not advertised
in the media for jobs at least since 1970.
Nor has it posted vacancies at the cannery
at least from 1970 through 1977. Wards
Cove has called the Alaska State

Employment Service in recruiting resident

cannery workers, but hasn't done any
media advertising.

106. At least since 1970 resident
cannery workers at Wards Cove have
generally been hired for only portions of
the season, during intense periods of
canning.

107. At least since 1977 Wards Cove
has hired some non-resident cannery
workers directly at 88 East Hamlin,
Seattle, Washington.

108. The majority of non-resident
Cannery workers dispatched to Wards Cove
through Local 37, ILW = have been

Filipino.

[PTO, p. 19]

116. It has not generally been
defendants’ practice at any time since
1970 to post written notice at Bumble Bee,

Red Salmon, Wards Cove or Ekuk canneries

A-6

of job vacancies which develop during the

season. « «. «

[PTO, p. 20]

118. A company fisherman captain's
job can be learned through prior
experience as a company fisherman's

partner.

A-7

DISTRICT COURT

DEPOSITION OF HAROLD BRINDLE
[Dep., p. 2]

HAROLD BRINDLE, being first duly
sworn by the Notary Public to tell the
whole truth, deposed and said as follows:

* * ®
[Dep., p. 33]

Q. Did you ever discuss with him
the necessity of housing crews together?

A. No, but there are various
specific advantages for doing so.

Q. Is that for call out time?

A. Basically.

Q. That's the crux of the
advantage of housing crews together?

A. Yes.

Q. Sorters would ordinarily get
called out earlier than other cannery

workers; isn't that correct? The fish

come in and have to be sorted into the
bins before they are processed?

A. Basically.

x* kt
[Dep., p. 34]

Q. And they would get called out
earlier than the other cannery workers;
right?

A. On occasion.

* *«
[Dep., pp. 35-38]

Q. When you refer to call-out time
or call-out time as you understand the
usage of the term refers to that time of
the day when the employees are told to go
to work; is that correct?

A. Basically.

Q. If the tender comes in with a
lot of fish and it's early in the morning,
the sorters would get called out to start

the fish first; right?

A-9

A. Along with the beach gang to
unload the fish.

Q. And then later in the day,
perhaps, once it's in the bins, the
cannery workers would get called out; is
that correct?

A. Yes.

Q. And the machinists who are
assigned to particular departments would
get called out with them; is that correct?

A. Yes.

Q. At the same time, in other

A. Yes.

oO. Port engineers might get called
out much earlier in the day if a fishing
vessel had repairs that were needed; is
that right?

A. Right.

Q. On the whole the office staff
works regular hours, say, 9:00 to 5:00?

A. No.

A-10

Q. What are their hours?

A. Normally from 8:00 in the
morning through 9:00 in the evening.

Q. But on the whole they are not
subject to call out depending on when the
tenders come in; is that correct?

A. Most normally not.

Q. They have their regular hours
duties to do; correct?

A. Right.

Q. Quality control, when do they
do their work?

A. During the time of canning.

Q. They would get called out at
the same time as the machinists assigned
to departments and the cannery workers
and the sorters; is that correct?

A. That's right.

Q. Would the call-out time for all
the culinary crew members be the same?

A. No.

Q. Who would get called out first?

A-1l

A. The first cook.

Q. And then?

A. Probably the baker.

Q. And after that?

A. Everyone else.

Q. Would the waiters get called
out before the dishwashers?

A. No.

Q. Supervisors such as the beach
boss and the first machinist and the
cannery crew foreman, would they have
their own call-out times or would they be
called out at the same time as their
crews?

A. That's pretty hard to answer.

Q. Well, as a general rule?

A. Most normally if there's an

early call-out, they would be first.

* *£

A-12

~

DISTRICT COURT

DEPOSITION OF MERVIN W. LESSLEY
[Dep., Pp. 2]

MERVIN W. LESSLEY, being first duly
sworn by the Notary Public to tell the
whole truth, deposed and said as follows:

x *
[Dep., pp. 4-5]

Q. What did you start out as?

A. Machinist.

Q. What was the first year you
worked at Red Salmon?

A. 1960.

Q. Have you worked there
continuously since 1960?

A. For the canhing season, yes.

Q. That's what I meant. And
@uring the off season have you worked at
Take Union Terminal each season since
1960?

A. Yes.

A-13

[Dep., p. 6]
Q. When did you become a machinist
foreman at Red Salmon?

A. 1972.

[Dep., p. 6]

Q. From the 1972 season onward did
you have any responsibility in hiring
machinists for Red Salmon?

A. Well, somewhat.

Q. Who else had a role in hiring
machinists?

A. Mr. A. W. Brindle.

Q. Would you recommend machinists
to him?

A. Yes.

[Dep., p. 7]
Q. If there was an opening would
Brindle ask you if you knew somebody that

could fill the job?

A-14

A. Yes.

Q. And if you knew someone, would
you recommend the person?

A. Yes. If I thought they were
qualified I would, yes.

Q. If you didn't know someone,
would you ask around your friends?

A. Yes.

Q. Would you do anything else to
find prospective machinsts?

A. Well, there would be people
come in from off the street, you know.

Q. For the most part was it a
guestion of asking around among the
people you knew?

A. Well, no, I wouldn't say so.

Q. About half and half?

A. Maybe 60-40.

Q. Which is the 60?

A. Trying to find qualified people
on your own and taking people who came in

and applied. They would make out an

A-15

application. We would ivestigate their
background and their qualifications.

xk
[Dep., p. 12]

Q. You have heard the term:
"Filipino bunkhouse,” haven't you?

A. Yes.

Q. Have you heard it used at Red
Salmon?

A. I suppose I have.

Q. So far as you can tell, was it
commonly used at Red Salmon up until the
time of the fire?

A. Well, that was one way of
designating the different mess halls.

*x* &
[Dep., p. 14]

Q. Is there a building that is

still called the Filipino bunkhouse

there?

A-16

A. Well, it was called that
before. That's what it's called now, I
guess.

Q. Is there a building that is
called the Native bunkhouse there now?

A. It hasn't had a Native in it for

years.

[Dep., pp. 14-15]

Q. So, it's not called the Native
bunkhouse any more?

A. It's called Native bunkhouse
but there is nobody there.

Q. Nobody lives there at all?

A. No.

Q. Have you used the term
"Filipino bunkhouse"?

A. I have, yes.

Q. Have you used the term "Native
Bunkhouse"?

A. Yes. When I went to send

somebody over to do a job there. Like

A-17

with the plumbing and et cetera, et
cetera, et cetera. They have to know
where they are going and et cetera.

Q. Did A. W. Brindle use the term
"Native bunkhouse"?

A. I guess he has.

x *&
[Dep., p. 30]

Q. When you asked around for
machinists among your acquaintances,
those that you turned up were all white;
were they not?

A. Yes.

A-18

a

6 * The weighting used for each job
family is indicated by the last line in
the table, labeled WEIGHTING.

A-239

12. The columns in the table show
the results for each of the job families
[1]

at the canneries. The specific job
titles represented in each job family are
detailed in Dr. Price's affidavit. The
last three columns of the table are
aggregates of the other columns. For
example, the column labeled AT ISSUE
refers to the composite results for the
twelve at issue job families, i.e., all
except CANNERY and LABORER.

13. The first row in the table,
labeled HIRES shows the total number of
new seasonal hires for this cannery group
and job family for the years 1971 and
1980. The table indicates that there were
211 new seasonal hires in general skill
jobs during 1971 through 1980 at Wards

Cove and Red Salmon combined. There were

l. This table is organized by job
families. We have also prepared tables
Organized by departments. In general,
the comments I make here apply equally to
those tables organized by department.

A-240

709 new seasonal hires in the at issue
families, and a total of 2,047 new
seasonal hires for all jobs at these
canneries.

14. The second row of the table,
labeled ACT %, gives the percent of these
new seasonal hires who were white. For
the general skills jobs, this was
73.460%, i.e., 155 out of the 211 new
seasonal hires in the general skill jobs
were white.

15. The third row of the table,
labeled TARGET, indicates the percent of
the relevant labor force who are white.
This value is calculated using’ the
geographical weighting scheme mentioned
above and discussed by Dr. Rees’ and
Dr. Price's Affidavits.

16. The remainder of the table
deals with the issue of how close the
"actual" is to the “target". This issue

is central to the question of whether the

A-241

hiring ‘pattern at the canneries is
discriminatory. The Classical
statistical approach to this question is
to ask "Are the observed data so different
from what are expected from a random
selection (i.e., race blind) process,
that we must conclude that the process is
not random?" If so, we must look for
other explanations, one of which might be
race discrimination.

17. There are two primary reasons
that the “actual” and “target” will
seldom be the same. First, the canneries
hire only a small percentage of the
"relevant labor force" and, therefore,
there is sampling variation in the hiring
process. Second, detailed census data on
occupation and industry are only
available for a sample of the population
and, therefore, there is sampling
variation in any census-based estimate of

the “relevant labor force." Both of these

A-242

sources of sampling error need to be
considered in any comparison of hiring
patterns at the canneries to the
"relevant labor force.”

18. The next to last line in the
example table, labeled CMPS DEV gives the
results of our best comparison of the
actual to the target hiring percentages
for white workers at the Wards Cove
Packing Co. canneries. This row
indicates the number of "stancard
deviations" the actual hiring percentage
differs from the target hiring
percentage. These standard deviations
measure the variati.n in the difference
between the sample percentage and the
true population percentage which can be
attributed to chance variation. If che
hiring process is random with respect to
race, and if it were to be repeated a
large number of times, we would expect the

actual hiring percentage to be within

A-243

plus or minus 1.96 standard deviations of
the target 95 percent of the time. This
range of plus or minus 1.96 standard
deviations from the target percentage
contains what is known as a 95 percent
confidence interval. Values falling
within the range of the confidence
interval are consistent with the
hypothesis that hiring can be represented
by a random process, i.e., a race blind
process. Furthermore, values of 1.96 or
greater indicate that substantially more
white employees than expected were hired,
while values less than -1.96 indicate
that substantially fewer white employees
were hired than expected.

19. Inspection of the last row of
the table indicates only one column where
the value for CMPS DEV exceeds 1.96. The
value for TENDER is 2.230 standard
deviations, indicating that the

percentage of whites in the TENDER group

A-244

exceeds the expected value “by a
statistically significant amount.

20. What conclusions can be drawn
from these findings? There are two
possible explanations. The first is that
the canneries have not hired TENDER's in a
manner consistent with a random selection
process and, therefore, some other
explanation must be found for the hiring
pattern observed at Wards Cove Packing
Co.--for example, race discrimination or
perhaps skill requirements that we have
not adequately incorporated in our
estimates of the relevant labor force.
The second possible explanation is that
the hiring process is race blind, but that
we observed an unusual situation, @.9.,
that Wards Cove had an unusual and
unrepresentative set of applicants to
hire from. In fact, the consequence of
the statistical nature of the test

insures that approximately once in each

A-245

twenty times we apply this test we should
expect to find values of CMPS DEV that
exceed 1.96, even if the hiring process is
an absolutely random process. Since the
proper interpretation of the statistical
results is crucial to an understanding of

this case I will return to this subject

later.
* *
[pp. 1851-1852]
BY MR. ARDITI:
Q. I suppose what we have

established at this point is that when we
find one point nine six for larger

standard deviations, and the composite

deviation row, then we have a
statistically significant over-
representation for a statistically

unrepresentation of non-whites, is that
correct?

A. Yes, that would be correct.

A-246

Q. Now, when in the same row we
find minus one point nine six or more
standard deviation, it means that we have
a statistically significant over-

representation of non-whites, is this

correct?
A. That is correct.
Q. And, I realize that there is a

proper way to phrase this from a
statistician's point of view and an
improper way, but if it's one point nine
six standard deviations, what we are

really seeing there is a suspicion of

discrimination against non-whites,
correct?
A. I think that you are seeing

there is evidence that you don't have
random selection from a target of pool
with that particular target as to whether
or not that is discrimination or not is a

different issue.

A-247

Q. Okay, but when we find a plus
sign there we are talking of over-
representation of whites, correct?

A. That is correct.

Q. And, when we find a minus, we
are talking under-representation of
whites, correct?

A. That is correct.

Q. Now, one interpretation or one
possibility of finding a great deal of
minus signs with one point nine six
standard deviations or less, saying going
down to minus three or whatever, is that
the availability figure that you selected

is wrong, isn't that one interpretation

of it?
A. That would be one, yes.
> eS =
[p. 1853]

Q. Alright, let me give you an
example. Let's say that in hiring Alaska

residents, an employer hires primarily

A-248

from Alaska native villages, okay, rather
than in an even way from the state as a
whole, okay?

A. Alright.

Q. And Alaskan native villages are
more heavily non-white than as the state
as a whole, isn't that correct?

A. I assume.

Q. And the smallest breakdown that
you have on one in one hundred public use
sample for Alaska is statewide, correct?

A. That is correct.

Q. So, if an employer is actually
hiring and actually, or actually get
applicants from the native villages
rather throughout the state as a whole,
use op [sic] the one in one hundred public
use tape is going to under estimate the

percentage of non-whites, isn't it?

A-249

DISTRICT COURT

TESTIMONY OF DR. ALBERT REES
[pp. 1865-1884)

ALBERT REES, being first duly sworn,

on oath deposes and says:
* *

l. Let me begin by stating my
qualifications to be an expert witness in
this case. I received the Ph.D. in
Economics from the University of Chicago
in 1950. I have been Professor of
Economics at the University of Chicago
and at Princeton University and for many
years taught undergraduate and graduate
courses in labor economics at both these
institutions. I also supervised a large
number of doctoral dissertations in labor
economics at both institutions. I am the
author of three books in labor economics
and the coauthor of a fourth; two of these

books deal in part with racial

A-250

discrimination in labor markets. I have
also had many articles published in
leading journals, including the American

Economic Review, the Journal of Political

Economy, t.~ Industrial and Labor

Relations Review, the Journal of Human

Resources, and the Monthly Labor Review.

A full list of my publications is attached
to this affidavit.

2. The plaintiffs in this case
allege that the defendants engage in
racial discrimination against nonwhites
in hiring employees for certain seasonal
jobs in five of defendants' facilities in
Alaska (these will be referred to as
"class facilities"). The way in which
such allegations are ordinarily judged is
to compare the percentage of nonwhites
hired by defendants in the jobs at-issue
with the percentage of nonwhites in the
supply of labor qualified for and

potentially available to do the work in

A-251

question. For most manual jobs,
including skilled manual jobs, the
ordinary procedure for estimating the
percentage of nonwhites in the labor
force would be to examine the racial
composition of the qualified labor force
living within reasonable commuting
distance of the place of employment. For
some professional jobs, one might examine
the racial composition of the regional or
national pool of people qualified for the
job.

3. This case presents’ special
difficulties that render the usual
procedures inapplicable because most of
defendants' seasonal employees do not
live within the commuting distance of
defendants’ facilities. Many are hired
in the lower 48 states, largely from the
states of Washington and Oregon, and are
transported at the employer's expense to

Alaska (generally from Seattle) for the

A-252

summer packing season. Other seasonal
employees are hired from remote Alaskan
villages and transported within Alaska to
defendants’ facilities. While at the
defendants’ facilities, migrant workers
live in bunkhouses or apartments provided
by their employers.

4. To deal with this difficulty, I
have devised a special procedure to
estimate the racial composition of the
available labor force in the geographical
areas from which the defendants hire.
This procedure has been carried out under
my general direction by Dr. William G.
Price, Jr., and his associates.

5. The basic data source used in
this procedure is the one-in-one-hundred
Public Use Sample of the United States
Census of Population and Housing for
1970. This data source makes available on
computer tapes the detailed individual

records of age, sex, race, labor force

A-253

status, industry, and occupation of a 1%
sample of the population. The smallest
geographical areas by which these data
are classified is the county group; a
county group must have a population of
250,000 or more. In densely populated
areas, a county group may be a single
county; in sparsely populated areas, it
may be a whole state. Thus, King,
Snohomish, and Pierce Counties,
Washington, each constitute an entire
county group (Nos. 13501; 13502; and
13503). The whole state of Alaska is
county group 14801.

6. The first step in the
estimation procedure is to examine the
defendants’ employee records for those
who worked at any of defendants’ Alaska
facilities (except Icy Cape) in each of
the years 1971-1980. From these records
we determine each employee's race and

place of residence at the time of hire.

A-254

These places of residence for the whole
ten-year period are then grouped into the
county groups used by the Census. This
process permits us to derive weights to be
applied to the Census data in which the
weight given to each county group is
proportional to its importance in
providing the employees actually employed
by defendants in the years 1971-1980.
These weights can be derived for all hires
or for new hires, and can be derived
separately for particular facilities or
for particular occupations. In each
instance, the weights provide accurate
estimates of the geographical sources of
defendants’ labor supply.

7. In estimating labor supply for
particular departments or job families at
defendants’ facilities, we may use Census
data either for the labor force as a whole
or for parts of the labor force comprising

occupations which involve skills relevant

A-255

to particular job families. In my
opinion, the latter method, indicated by
the phrase "Skilled Workers," provides
the better basis of comparison for
defendants" hiring into the skilled job
families. However, for the “General
Skills" job family, and for comparisons
by department (without skill
differentiation), references to the
Census of the whole labor force ("All
Workers") is appropriate.

8. The more difficult part of the
procedure is toe estimate which
individuals in these geographical areas
are available for seasonal work. None of
the data in the Census tapes provide
Gefinitive information on this, nor would
any other statistical @ata source. The
Census data, however, can be used to
identify the classes of workers most
likely to be available for migrant

seasonal employment.

A-256

9. In some versions of the
procedure, I have assumed that workers
with full-time year-round employment
would not be willing to leave this
employment to take seasonal work even
where the seasonal work is very well paid.
This leaves several different categories
of workers available: the unemployed;
students and teachers who are free to work
during summer; and workers in other
seasonal industries whose seasons do not
necessarily coincide with the salmon
packing season. In addition, there are
workers in industries such as
construction who are not usually emloyed
steadily by any one employer. The sum of
these categories is the availability
class called “Relevant Seasonal.*

10. As an alternative assumption,
we have used the availability class of
everyone in the civilian labor force aged

18 and above. From either of these

A-257

availability classes the employer would
select those applicants best qualified by
skill and experience to fill the
vacancies not filled by returning workers
from previous seasons.

ll. In my judgment, the
availability class “Relevant Seasonal"
reflects the composition of defendants’
work force more accurately than the class
"Everyone in Work Force.” The two
assumptions, however, give roughly
similar results so that the choice
between them is usually not crucial.

12. we use these skill and
availability considerations to extract
from the Census data our best estimates of
the racial composition of the labor
supply available to the various jobs in
Alaska salmon canning industry. In doing
this, we do not claim to identify those
particular individuals in the Census

sample who are, in fact, willing and able

A-258

to work in the industry. It is certainly
true, for example, that not all college
students are willing to work in Alaska.
However, the best independent estimate we
have of the racial composition of those
college students who would be willing and
able to work in Alaska is the racial
composition of college students reported
in the Census.
“

13. The same argument applies to
our procedures for estimating the racial
composition of the labor force for any of
the at-issue jobs. For example, the
Census does not identify each of the
cannery machinist jobs as an occupation.
Nonetheless, the best available estimate
of the racial composition of persons
qualified to be cannery machinists is the
racial composition of those workers in
the Census sample reporting occupations
most similar to those in the machinist job

family. Based on Mr. DeFrance's

A-259

information, this would be those workers
who reported their occupations as those
in the machinist job family, Attachment 8
to Dr. Price's affidavit. We cannot
infer from this that all workers in these
occupations could be cannery machinists,
nor that all cannery machinists would
report one of these as their occupation
group. We simply use these
Classifications to construct our best
estimate of the racial composition of
people available to be cannery
machinists.

14. A final comment on the use of
the Census is that the month of the year
the Census was taken is not particularly
important to the validity of our study.
Our purpose is to estimate the racial
composition of the relevant labor force.
Since most workers do not change
occupations on a seasonal basis, the date

of the Census enumeration should have

A-260

little effect on the distribution of
sampled individuals across the
occupational categories.

15. In some versions of this
procedure the geographical patterns are
based on “all hires" or “all seasonal
hires;" in other versions the pattern is
based on “new hires” or “new seasonal
hires." The word “seasonal” indicates
that we have omitted employees who work in
defendants' facilities in the lower 48
states--the home offices of Wards Cove
Packing and of Columbia Wards Fisheries
and the Lake Union Terminals. These
workers are employed by defendants during
the season when the Alaska canneries do
not operate, and are therefore not
strictly migrant seasonal employees. The
aifference between “all hires" and “new
hires® is that in “new hires” each
individual is counted only if he aid not

work in the same facility on the same job

A-261

in the previous season, while in “all
hires" each employee is counted once for
every season in which he is employed, even
when he returns to the same job at the
same facility.

16. Defendants are generally
required by the terms of their union
contracts to rehire union employees from
the previous season who make a timely
written request to be rehired. In any
event, defendants would want to rehire a
satisfactory employee, particularly in a
skilled job, rather than take a chance on
an unknown applicant. For this reason,
versions of the analysis based on “new
hires" are preferable to those based on
"all hires." If there is any racial
discrimination, it would be reflected in
the initial hiring @ecisions, which are
the only decisions where the defendant
employers always exercise free choice.

The “all hires" versions of the procedure

A-262

also suffer because the multiple counting
of the same individual, which overstates
the number of hiring choices open to the
employer, can make random events appear
incorrectly to be statistically
significant.

17. The procedure we have devised
can be applied to all the workers in a
Gepartment or only to the skilled
workers. In the tables for “skilled
workers" (e.g., Exhibit A-278, Table 4
for each tab), the unskilled and hard-to-
Classify jobs in each department have
been removed and listed in a separate
column headed “general skilis.* The
remaining job families contain most of
the highly skilled, well paid jobs that
are the main subject of contention in this
liti vation.

18. Furthermore, the procedures can
use geographic weighting patterns (for

the Census data) based only on particular

A-263

jobs, based on the section of the job
(cannery workers and laborers are one
section, other employees are the other),
or based on the entire cannery workforce.
Of these three possibilities, the second
seems strongly preferable. Weighting by
individual jobs results in samples that
are too small to produce reliable
weights. A few individuals in the same
job hired from the same county group can
greatly inflate the relative importance
of that group. At the opposite extreme,
using the entire workforce ignores some
important differences between the
geographical origins of cannery workers
and laborers on the one hand and the
remaining workers on the other. In
particular, this version of the procedure
overweights Alaska, which provides a
larger portion of cannery workers and
laborers than of more skilled workers.

Since the cannery worker jobs are not at-

A-264

issue in this litigation, this seems
inappropriate. Weighting by section is
appropriate because almost all the
workers other than cannery workers and
laborers are hired at company
headquarters, whereas most cannery
workers and laborers are not.

19. The Census data gives the race
of each individual in the Census sample.
By following the procedures summarized
above, we have estimated the percentage
of nonwhites in the labor supply of
seasonal workers as defined by the
various geographical weights and
availability classes I have described.
We then compare the estimated percentage
of nonwhites in the labor supply with the
percentage of nonwhites actually hired by
defendants. Where these percentages
differ, as they ordinarily do, we use
Standard statistical tests to determine

whether or not the differences are small

A-265

enough to lead one to presume that they
could have occurred by chance. The
results of these comparisons and tests of
statistical significance are given in
Exhibits A-278 ("new seasonal hires") and
A-279 ("new hires").

20. The general conclusion I would
Graw from these exhibits is that there is
not a pattern or practice of racial
Giscrimination in hiring skilled workers
for defendants’ facilities. In many job
families the percentage of nonwhites
hired is larger than their percentage of
the relevant labor supply. In many more
job families, nonwhites are slightly
under-represented, but the differences
are not statistically significant and
could have occurred by chance with a high
probability. In a few job families of the
class facilities of two of the defendant
companies, nonwhites are under-

represented in comparison with the labor

A-266

force estimates, and the differences are
significant at the 5% level. Racial
discrimination is one possible
explanation of such differences.

21. The procedures used are more
reliable when applied to groups of
facilities rather than to _ individual
facilities. In the individual facilities
the sample size can be so small that even
relatively large disparities between the
percentage of nonwhites hired and the
percentage of nonwhites in the labor
force would not be statistically
significant.

22. For the defendant Wards Cove
Packing Company, the table that in my
judgment best enables one to test the
allegation of racial discrimination is
the one entitled "New Seasonal Hires,
Wards Cove Packing Company (Wards Cove
and red Salmon) for Skilled Workers

Weighted by Section of Job for

A-267

Availability Class Relevant Seasonal.”
(Exhibit A-278, Tab: WC&RS, Table 4).
This table shows 709 hires in the jobs at-
issue in this litigation. Of these 709
workers hired, 85.3% were white. The
percentage of whites in the labor supply
for these jobs is estimated as 89.9%,
which is listed on the line labeled
"target." Thus, in general, Wards Cove
Packing Company hired a larger percentage
of nonwhites in the jobs at-issue than the
1970 percentage of nonwhites in the
relevant seasonal labor force of the
geographical areas from which they draw
employees, when these geographical areas
are combined with weights representing
the area of origin of workers in all jobs
at-issue in all Alaska facilities (except

Icy Cape) of the three defendant

companies.

23. When we examine the columns of

this table, which represent job families,

A-268

separately, we find four in which the
percentage of whites hired was greater
than the percentage of whites in the labor
supply. These job families are
carpenter, fisherman, medical, anda
tender. We now examine the row labeled
"composite deviation” to determine
whether these differences are
statistically significant at the 5%
level, given the sampling errors inherent
in the process of hiring from a pool of
applicants and the 1% sampling of the
Census. This row tells us that in three
of the four cases the differences could
have occurred by chance (the composite
deviation is less than 1.96, the critical
level). In the remaining case, the tender
job family, the result would occur by
chance less than one time in twenty. I
conclude that nonwhites are significantly

under-represented in hiring in the tender

A-269

job family, which could be the result of
racial discrimination.

24. It might be argued that the two
Class facilities of Wards Cove Packing
Company should be examined separately, to
the extent different individuals do the
hiring at the two facilities. If we
examine the table for new seasonal hires
of skilled workers at the Wards Cove
cannery alone (Exhibit A-278, Tab: W,
Table 4), we find 318 hires in the jobs
at-issue, of which 89.623 percert were
white. The estimated percentage of
whites in the available labor supply is
(weighted by section) 89.949 percent, so
that in general, nonwhites are not under-
represented. Nonwhites are under-
represented in four job families:
fisherman, machinist, office, and tender,
but in none of these cases is the
Gifference statistically significauc at

the 5% level.

A-270

25. If we examine the corresponding
table for the facility at Red Salmon
(Exhibit A-278, Tab: RS, Table 4), we find
391 hires in the jobs at-issue. Of these
81.941% were white. The estimated
percentage of whites in the labor supply
is again 89.949%, so that in general,
nonwhites are not under-represented in
the jobs at-issue at Red Salmon. They are
under-represented in five job families:
carpenter, fisherman, medical, radio, and
tender, but only in the tender job family
is the difference statistically
significant at the 5% level. In the
tender job family the under-
representation of nonwhites could be the
result of racial discrimination.

26. If we examine the corresponding
tables for the ¢@efendant Columbia Wards
Fisheries (Alitak an@ Ekuk) (Exhibit
A-278, Tab: CWF: AK&EK), we find 862

hires in the jobs at-issue at the two

A-271

facilities taken together. Of these
71.694% were white. The percentage of
whites in the relevant seasonal labor
supply is again estimated at 89.945%, so
that for all the jobs at-issue taken
together the percentage of nonwhites
hired substantially exceeds that in the
labor supply. This is also true for each
job family separately, except for
"office" and "store and stockroom." In
neither of these is the under-utilization
of nonwhites statistically significant.
I conclude that there is no evidence in
this table to support the allegation that
nonwhites are under represented in at-
issue jobs at Columbia Wards Fisheries.
27. If we look at the corresponding
tables for the two facilities separately,
we obtain the same result. At Alitak
(Exhibit A-278, Tab: AK, Table 4), only
67.333% of the workers hired in the jobs

at-issue were white, substantially less

A-272

than the percentage in the available
labor supply. Nonwhites were under-
represented in two job families,
fisherman and office, but in neither case
is the difference statistically
significant. At Ekuk (Exhibit A-278,
Tab: EK, Table 4), 76.456% of the hires in
jobs at-issue were white, again well
below the percentage in the available
labor supply. Nonwhites were under-
represented in five job families:
carpenter, culinary, machinist, medical,
and storekeeper/stockroom, but none of
these differences is statistically
significant at the 5% level.

28. Next we examine the
corresponding table for South Naknek
(Exhibit A-278, Tab: SN, Table 4), the
only class facility operated by defendant
Castle & Cooke (Bumble Bee Seafoods).
Here we find 536 hires in the jobs at-

issue. Of these, 84.142% are white. The

A-273

percentage of whites in the relevant
seasonal labor supply is again 89.949%.
For at-issue jobs at South Naknek as a
group, nonwhites are not under-utilized.
Examining the separate job families,
however, we find seven where the
percentage of whites is greater than the
target percentage. These are beachgang,
carpenter, culinary, fisherman,
machinist, store and stockroom, and
tender. In two of these seven job
families (fisherman and machinist) the
differences are too large to have
occurred by chance more than one time in
twenty. I conclude that nonwhites were
significantly under-represented in hiring
in the fisherman and machinist job family
at South Naknek, which could be the result
of racial discrimination.

29. In considering whether the
defendants engage in ae pattern of

discrimination based on race, it may be

A-274

useful to look at the data for all their
facilities, and not merely at the "class
facilties" at which discrimination is
alleged by plaintiffs. This approach is
especially relevant to the extent that
defendants follow common hiring practices
and policies at all facilities. It also
results in larger sample sizes that
increase the power of the statistical
tests.

30. A broad view is obtained by
looking at all Alaska facilities in which
Wards Cove Packing Company has an
interest, which combines Wards Cove
Packing Company and Columbia Wards
Fisheries (except Icy Cape). Exhibit
A-278, Tab: WC: AL-IC, Table 4. For this
group of facilities, we examine the table
for new seasonal hires o” skilled workers
weighted by section of job. Id. There
were 2,789 hires in this group, of which

74.722% were white. The percentage of

A-275

whites in the relevant seasonal labor
supply was 89.949%, so that in this large
set of hires, nonwhites were
substantially over-represented. They
were under-represented in only’ three
small job families--administration,
medical, and radio--with a total of only
27 hires. In none of these was the
difference statistically significant at
the 5% level.

31. The same kind of comparison can
be made for another broad group of
facilities, Castle & Cooke, Inc.
interests, all Alaska facilities (except
Icy Cape). Exhibit A-278, Tab: C&C Al-IC,
Table 4. This overlaps the previous
grouping of facilities because it also
includes Columbia Wards Fisheries. For
the Castle & Cooke interests, we look at
skilled workers weighted by section of
job family for the availability class

"Relevant Seasonal.” Id. In this

A-276

grouping there were 2,616 hires in the
jobs at-issue, of which 73.777% were
whites. The percentage of whites in the
labor supply was again estimated at
89.949 percent, so that’ for these
facilities as a group, nonwhites are not
under-represented. They are under-
represented in only two job families,
administration and culinary, and_ in
neither of these cases is the difference
statistically significant at the 5%
level.

32. In Table 1, I have summarized
the results discussed so far for all at-
issue workers by facility. In Table 2, I
have summarized the results for the
facilities in Table 1 which had job

families where there are _ significant

positive differences between the actual

and target percentages of whites. (The

source for both tables is Exhibit A-278,

Table 4 for each Tab.) The remaining

A-277

facilities in Table 1 had no job families

with a significant positive difference.

A-278

Table 1
Summary of Statistical Results for New Seasonal Hires,
Skilled Workers Weighted by Section of Job Family
Availability Class Relevant Seasonal;
for All At-Issue Jobs

Facility No. of Actual Target
or Hires Percent Percent Composite

Group At-Issue White White Deviation
Wards Cove — 89.6 89.9 -0.17
Red Salmon j 391 81.8 89.9 -4.58
Alitak 450 67.3 89.9 -13.46
Ekuk 412 76.5 89.9 -7.78
South Naknek 536 84.1 89.9 -3.67

Wards Cove Class
Facilities 709 85.3 89.9 -3.20

Columbia Wards
Class Facilities 862 Tae? 89.9 -13.37

Wards Cove Interests
Alaska Operations 2,789 74.7 89.9 -14.27

Castle & Cooke, Inc.

Interests, Alaska
Operations 2,616 73.8 89.9 -15.02

A-279

Table 2
Summary of Statistical Results for New Seasonal Hires,
Skilled Workers Weighted by Section of Job Family,
Availability Class Relevant Seasonal; by Job Family*

Facility or No. of Actual Target
Group, and Hires in Percent Percent Composite
Job Family Job Family White White Deviation
Red Salmon, tender 108 96.3 87.4 2.04
South Naknek,

fisherman 70 100.0 83.1 2.69

machinist 76 100.0 91.1 2.49
Wards Cove Class

facilities, tender 296 95.3 87.4 2.23

* Job families shown are those with positive differences

that are significant at the 5% level.

A-280

33. The sample can be further
broadened by moving from “new seasonal
hires" to "new hires," which includes in
the cases considered workers who worked
in an Alaska facility during the packing
season and also in a base facility (Lake
Union Terminal, Wards Cove Home Office,
or Columbia Wards Home Office) at some
other time during the year. Most of the
workers added by this change are employed
by defendants on a year-round basis, but a
few may have been employed only briefly at
a base facility.

34. I have reexamined each of the
cases considered above, changing the
sample from "new seasonal hires" to "new
hires” (i.e., see Table 4 at the same Tab
in Exhibit A-279). In almost every case
the numbers change slightly. In only two
cases would the change affect my
conclusions, and these work in opposite

directions. At the Red Salmon facility of

A-281

Wards Cove Packing Company, I concluded
on the basis of new seasonal hires that
the under-representation of nonwhites in
the tender job family was statistically
significant at the 5% level. On the basis
of new hires, it would not be. Exhibit
A-27S9, Tab: RS, Table 4. This change in
the statistical significance of the
under-representation occurs despite the
fact that the sample size is only 108 new
seasonal hires, but 145 new hires.

35. At the Ekuk facility of
Columbia Wards Fisheries, I concluded on
the basis of new seasonal hires that the
under-representation of nonwhites in the
machinist job family was not
Statistically significant at the 5%
level. On the basis of new hires, it
would be statistically significant.
Exhibit A-279, Tab: EK, Table 4. The
change in statistical significance is

due, in part, to the increase in sample

A-282

size from 64 new seasonal hires to 79 new
hires.

36. The concept of the labor supply
for the skilled jobs advanced by the
expert witness for plaintiffs, Dr. Robert
J. Flanagan, is very different from what I
have used above. In general, he assumes
that the labor supply for the skilled jobs
in the salmon canning industry consists
of people who have previously worked in
the industry, including people who have
held unskilled jobs. His analysis
assumes, however, that some of the
skilled jobs require too much _ special
training or skill to be held by unskilled
workers.

37. Exhibits A-282 and A-283 are
called "“plaintiffs' version" of our
statistical analysis. In this version,
the jobs that Dr. Flanagan did not
include in his analysis (because the jobs

are too skilled to be performed by

A-283

unskilled workers or because he had no
opinion about them) have been moved out to
the various skilled job families. These
jobs as defined by Dr. Flanagan in his
deposition have been grouped together in
a separate column labeled “excluded.” We
can now reexamine the results for Wards
Cove Packing Company (Wards Cove and Red
Salmon) for New Seasonal Hires, Skilled
Workers Weighted by Section of Job,
Availability Class Relevant Seasonal,
using “plaintiffs' version" of the jobs
at-issue (Exhibit A-282, Tab: WC&RS,
Table 4). In this case, the comparisons
still show an under-utilization of
nonwhites in the tender job family, but it
is no longer statistically significant at
the 5% level. The same loss of
statistical significance occurs if we
examine the results for the tender job

family at the Red Salmon facility

A-284

separately, using "plaintiffs' version."
Exhibit A-282, Tab: RS, Table 4.

38. Using “plaintiffs' version" of
the comparison for Columbia Wards
Fisheries (Alitak and Ekuk) makes no
difference. (Exhibit A-282, Tab: CWF:
AK&EK, Table 4). In "“"plaintiffs'
version” as in "“defendants' version"
there is no job family where nonwhites are
significantly under-represented, and this
is also true when the two class
facilities, Alitak and Ekuk, are examined
separately. (Exhibit A-282, Tab: EK,
Table 4). When we use "“plaintiffs'
version" to make the corresponding
comparison for the Castle & Cooke (Bumble
Bee Seafoods) facility at South Naknek
(Exhibit A-282, Tab: SN, Table 4), the
results for the fisherman job family are
unchanged. Nonwhites remain
significantly under-represented. For the

machinist job family, the under-

A-285

representation of nonwhites is no longer

statistically significant.

39. Up to this point, I _ have

considered the areas from which
defendants recruit as a single
geographical labor market. In some

respects it may be desirable to divide
this market. The testimony of expert
witnesses for the plaintiffs, Dr. Robert
Flanagan and Dr. Shirley Smith, deals
largeley with migrant seasonal labor.
This does not correctly characterize some
of the Alaska residents hired by
defendants, namely those who live year-
round in the vicinity of defendants’
Alaska facilities, such as the cannery at
Wards Cove. Moreover, defendants’ union
contracts generally provide that
preference in hiring will be given to
Alaska residents. For this reason it is
useful to examine Alaska and the areas

from which defendants draw employees in

A-286

the lower 48 states as two separate
geographical labor markets.

40. Exhibit A-500 examines data for
those new hires who did not live in Alaska
when hired. It produced some conclusions
different from those obtained by
examining all new hires (i.e., from both
Alaska and the lower 48). In examining
all new hires, I concluded that Wards Cove
Packing Company, in its two class
facilities combined, under-represented
nonwhites among skilled workers in the
tender job family. If we look at the
table in Exhibit A-500 for Wards Cove
Packing Company (Wards Cove and Red
Salmon) for skilled workers hired outside
Alaska weighted by section of job for
availability class relevant seasonal, we
find a very slight under-representation
of nonwhites, which is not close to being
statistically Significant. Exhibit

A-500, Tab: WC&RS, Table 4. The reason

A-287

for this change is that the target
percentage of whites has risen--that is,
there is a higher percentage of whites in
the relevant seasonal labor force outside
Alaska than in the two geographical labor
markets combined.

4l. On the other hand, Exhibit
A-500 shows an under-representation of
nonwhites that did not appear in previous
exhibits. If we examine skilled workers
hired outside Alaska at the Ekuk facility
in all the jobs at-issue in this case, we
find that 97.361% are white, whereas the
estimated percentage of whites in the
available seasonal labor force outside
Alaska is 92.1408. Exhibic A-500, Tab:
Ek, Table 4. This difference is
statistically significant at the 58%
level.

42. In judging the meaning of this
under-representation of nonwhites among

skilled non-Alaskans at Ekuk, it should

A-288

be kept in mind that of the Alaskans hired
at Ekuk in the at-issue job families,
91.6% were nonwhite, and all of these were
Alaska Natives. Exhibit A-501, Table
2(A). This is why Ekuk does not show
under-representation of nonwhites when
both geographical markets are considered
together.

43. In examining new hires’ from
both labor markets at Ekuk, I concluded
that whites were under-represented in the
machinist job family to a statistically
significant degree. See, 435, supra.
When we examine the corresponding table
in Exhibit A-500, the under-
representation is no longer statistically
significant. Exhibit A-500, Tab: EK,
Table 4.

44. In examining new hires’ from
both labor markets at the South Naknek
facility of Castle & Cooke, I concluded

that nonwhites were under-represented in

A-289

the fisherman and machinist job families
to a statistically significant degree.
When we examine these job families for
non-Alaskan hires only, the under-
representation is still present but is no
longer statistically significant at the
5% level. Exhibit A-500, Tab: SN,
Table 4.

45. Exhibit A-501 examines’ the
hires of Alaska residents. Table 18 shows
the percentage of nonwhites in the
unweighted labor pool for Alaska. There
are 14.1% nonwhites in the whole labor
force in 1970 and 15.4% nonwhites in the
relevant seasonal labor force. Table 2A
shows the percentage of nonwhites among
new hires in the five class facilities.
In four of the five (Alitak, Ekuk, Red
Salmon, and South Naknek) the percentage
of nonwhites among Alaska residents hired
in the job families at-issue is far higher

than in the Alaska labor force as a whole,

A-290

ranging from 63.5% at Alitak to 91.6% at
Ekuk. All but one of the nonwhite Alaska
residents hired at these four facilities
were Alaska Natives. On the other hand,
the cannery at Wards Cove has only 7.4%
nonwhites among its new hires of Alaska
residents. Exhibit A-501, Table 2(A).
This cannery is located near the town of
Ketchikan, whose 1970 population was 12%
nonwhite. (See plaintiffs" Exhibit 480,
p. 29). It should also be noted (A-501,
Table 1A) that for Alaska residents the
percentage of nonwhites hired in the
defendants' Alaska operations (except Icy
Cape) is just as high for the at-issue
jobs (62.0 percent) as for cannery
workers and laborers (61.9 percent).

46. I turn now to the general
theory of the case presented by the expert
witnesses for the plaintiffs,
Dr. Robert J. Flanagan and Dr. Shirley

Jeanne Smith. As I have noted previously,

A-291

Dr. Flanagan considers the labor supply
to the skilled jobs to consist of all
workers who have previously worked in the
Alaska salmon industry. To measure the
racial composition of this labor supply,
he relies primarily on ae statistical
series complied by the old United States
Bureau of Fisheries for the years 1906-
1936 and by the United States Fish and
Wildlife Service for the years 1941-1955.
Using the last five years of this series,
he determines that the labor supply of
skilled workers to the Alaska salmon
canning industry is 47% nonwhite. This
percentage is far higher than my
estimates discussed above, which are in
the general neighborhood of 10% nonwhite.

47. The first objection to the data
used to form Dr. Flanagan's estimate is
that they are out of date. Even the last
year used, 1955, is fifteen years earlier

than the 1970 Census data on which I rely,

A-292

and long before the period covered by this
litigation.

48. Even if the data were up-to-
date, however, they would be
inappropriate. Dr. Flanagan assumes that
unskilled workers in the salmon industry
constitute the labor Supply to. the
skilled jobs because they can be
adequately trained for skilled work and
then promoted. There are some
industries, such as the basic steel
industry, in which this would be a correct
assumption. Such industries have three
Crucial characteristics: (1) that most
Skilled workers are in fact hired at the
unskilled level and are promoted from
below; (2) that the facilities operate
all year-round; and (3) that all
unionized employees belong the same
industrial union [sic]. The defendants’
canneries do not have any of these three

Crucial characteristics. Skilled workers

A-293

at defendants’ canneries are usually
hired directly into skilled jobs, so that
the labor supply to these jobs lies
largely outside the industry. The very
short packing season and the intense pace
of work while fish are running leaves
insufficient time for the present skilled
workers to train unskilled people to take
over their duties. Workers in different
departments of the canneries are
represented by different labor unions,
each with its own contract regulating
hiring and rehiring, so that we have a
craft union situation like that in the
construction industry, not an industrial
union situation like that in the basic
steel industry. In the circumstances of
the Alaska salmon canning industry, Dr.
Flanagan's theory is entirely
inappropriate.

49. The heavy representation of

nonwhites among migrant cannery workers

A-294

in defendants' facilities is in large
part a result of the influence of the
International Longshoremen and
Warehousemen's Union, Local 37, which is
predominately Filipino, and which
represents cannery workers at several of
the defendants' facilities.

50. The influence of Local 37 on
the racial composition of the labor force
at defendants’ facilities is shown in
Exhibit A-498, which shows the detailed
racial composition by union status of all
hires who were nonresidents of Alaska. Of
those nonresidents who were members of
Local 37, 68.9% were nonwhite and 58.4%
were Filipino (Exhibit A-498, Table 4).
For those who were members of other unions
or of no union, the percentage nonwhite
ranged from 3.1% to 11.1% and _ the
percentage of Filipino from zero to 1.6
percent. In my estimates of the labor

supply in the geographic areas from which

A-295

Gefendants draw employees, Filipinos
constitute only 0.477% of the relevant
seasonal labor force. Exhibit A-278,
Labor Pool Tables, Table 4.

Sl. Unweighted data on the
percentage of Filipinos in the population
and labor force of the broad areas from
which defendants draw employees are
available from the published volumes of
the Census of 1970. Some such data are
presented below in Table 3. Although the
percentages of Filipinos are generally
Slightly higher than in my weighted
estimates of labor supply, none of them

exceeeds one percent.

A-296

Table 3
Percentage of Filipinos in Selected
Population Groups and Areas, 1970

i

Area and Number (thousands) Percent
Group Filipinos Total Filipino

Total population

Alaska 1.3 300 0.43

California 135.2 19,953 0.68

Oregon 1.5 2,091 0.72

Washington 11.5 3,49 0.34
Labor force, 16 years and older

California 62.7 8,338 0.75

Washington 5.1 1,399 0.37

Seattle-Everett SMSA 3.5 611 0.57
Employed workers, Seattle-Everett SMSA

Craftsmen, foremen, and kindred 0.3 82 0.37

Operatives including transportation 0.5 66 0.76

Source: Census of Population, 1970.

A-297

52. For institutional reasons
connected with the history of the
industry and of Local 37, Filipinos are
greatly over-represented among cannery
workers. Dr. Flanagan's theory of labor
supply would require that an employer
whose work force in one department
greatly over-represents a particular
racial minority should over-represent
this minority correspondingly in all
other departments. I cannot see any basis
in labor economics for such a
requirement.

53. Dr. Flanagan has asserted that
defendants engage in occupational
segregation by confining nonwhite workers
to unskilled jobs. Exhibit A-499
indicates that to the extent this occurs,
it is the result of the influence of
Local 37. Table 1 of that exhibit
examines the racial composition of new

hires at the class facilities who were not

A-298

residents of Alaska, with members of
Local 37 excluded. Of 808 such hires in
the cannery worker and laborer
departments, 79, or 9.8% were nonwhite.
Of 2,001 such hires in the at-issue
departments, 168 or 8.4% were nonwhite.
These percentages are very similar.
Comparable results are obtained when all
Alaska operations of defendants are
considered. (Tables 4 and 9).

*'S4. Both Dr. Flanagan and Dr. Smith
state the view that nonwhites are heavily
over-represented in the labor supply of
seasonal migrant workers generally. This
view is based on their examination of
studies of migrant farm labor, of which
the most recent cited is Leslie Whitener

Smith and Gene Row, The Hired Farm Labor

Force of 1976, U.S. Department of

Agriculture Economics, Statistics and
Cooperatives Services, Agricultural

Economic Report No. 405, July 1978. This

A-299

report shows (Table A-16, p. 41) that a
preponderance of migrant farm workers who
traveled 1,000 or more miles to reach
their most distant work place in 1976 were
Hispanic. Minorities were not heavily
represented, however, among hired farm
workers who lived in Standard Federal
Region X (Alaska, Idaho, Oregon, and
Washington) in December 1976. Hispanics,
blacks, and other nonwhites were only 6%
of the 204,000 hired farm workers in the
region (Table 5, p. 14). (These figures
include nonmigrants).

55. I seriously question the
relevance of information on migrant farm
workers for seasonal migrant industries
in general and for the Alaska salmon
industry in particular. There are other
important seasonal migrant industries
besides agriculture, including heavy and
highway construction and the Eastern

seasonal resort hotel industry. Although

A-300

little data is available on the racial
composition of the workforce in seasonal
resort hotels, casual observation
indicates that it is predominately white.

56. D. Quinn Mills, in his book

Industrial Relations and Manpower _in

Construction (1972) states that

"construction crews on highways,
pipelines, transmission lines, and heavy
construction projects are mobile because
of the location of the projects and their
nature." (p. 138) The word mobile in this
quotation refers to geographical
mobility. He identifies five trades as
making up the heavy and highway branch of
the industry: carpenters, laborers,
operating engineers, cement masons, and
teamsters. (p. 14) The data he gives on
the racial composition of these five
trades in 1970 range from 5.0 percent
nonwhite for operating engineers to 30.3%

nonwhite for cement masons (p. 145).

A-301

Data for 1974 from the Federal Highway
Administration show 15% minorities among
equipment operators on Federal highway
projects. H. R. Northrup and 4H. G.

Foster, Open Shop Construction (1975),

p. 342.

57. The Alaska salmon’ industry
offers jobs to migrant workers that are
very different from migrant jobs in
agriculture. Workers are transported to
salmon canneries by plane or ship at the
employer's expense. Migrant agricultural
workers are usually transported to job
sites in dilapidated trucks or buses,
usually at their own expense. (See Lloyd

H. Fisher, The Harvest Labor Market in

California, 1953, pp. 57-58). They may

find no work available when they arrive,
and unlike cannery workers get no pay when
work is not available. |

58. Most important, salmon canning

is well paid, while migrant workers in

A-302

agriculture are paid very little. The
study by L. G. Smith and G. Rowe cited by
Dr. Shirley Smith tells us (p. 10) that
migrant farm workers in 1976 earned an
average of $1,807 for 82 days of farmwork.
This is $21.80 per day, or $153 for a
seven day week. In the same year, skilled
workers in the defendants' facilities in
Alaska were earning $400, $500, or even
$600 a week, much more than workers with
comparable skills in Seattle or Portland.
See, Table 4 attached. It is not at all
surprising that whites are not attracted
to low wage migrant farm work and,
therefore, that Hispanics with limited
job alternatives would constitute most of
the labor supply for this kind of work.
At the same time it is not surprising that
whites would be eager to work in the high-
wage salmon canning’ § industry. In
general, I find the testimony of Dr.

|
Flanagan and 0ODr. Smith of # limited

A-303

relevance to the circumstances of this
case.
x *
CROSS EXAMINATION
[pp. 1917-1929]
BY MR. ARDITI:

Q. Dr. Rees, Exhibits 278, and
A-278 thru A285 were prepared under your
supervision, were they not?

A. Yes, they were.

Q. And they reflect your
conclusions using different assumptions,
is that correct?

A. I draw my conclusions’ form
[sic] them, yes.

Q. Now, each of these tables is
arranged by facilities or a combination
of facilities, is that correct?

A. That is correct.

Q. Where hiring responsibilities,
is ultimately in the hands of each cannery

superintendent, you would look at the

A-304

percent non-white hired separately at
each cannery to determine whether there
was any discrimination, would you not?

A. Yes, under your assumption as
you have stated it, I would.

Q. And, in exhibits A-278 thru
A-285, that would be the first five tabs,
would it not?

A. Yes, it would.

Q. Now, within each tab of each of
each [sic] of those exhibits, you have
eight tables, correct?

A. That's correct.

Q. In tables five thru eight, you
take as the availability class everyone
in the work force or in the labor force,
correct?

A. Correct.

Q. Now, within any given
geographical area, most people in the

labor force would. prefer full-year

A-305

employment to seasonal employment, would
they not?

A. That's generally correct, yes.
It depends a little bit on the wage of the
seasonal employment. The higher the wage
in the seasonal employment, the more
people would apply.

Q. But generally, that is correct?

A. Generally, that is correct.

Q. Now, tables five thru _ [sic]
eight do not adjust for availability for
seasonal work, do they?

A. No, they assume that everyone
in the labor force is available.

Q. So is it fair to conclude that
to the degree that most people prefer
full-year employment to seasonal
employment, tables five thru [sic] eight
are not a good measure of availability?

A. I think tables one thru [sic]

four are a better measure, but I wouldn't

A-306

go so far as to say that five thru [sic]
eight are not a good measure.

Q. Now, within tables--or, in
tables on thru [sic] four, you make an
attempt to adjust for availability for
seasonal employment, correct?

A. Yes, we attempt to identify
those members of the work force who are
most likely to be interested in accepting
seasonal employment.

Q. Okay, but you make no
adjustment for those interested in
migrant work, do you, as opposed to
seasonal work.

A. I assume that those who are
interested in migrant seasonal work are a
subsection of those who are interested in
seasonal work generally.

Q. Okay, my questions is, you do
not, do you make a separate adjustment for

those interested in migrant seasonal work

A-307

as opposed to those interested in non-
migrant seasonal work, do you?

A. I know of no way of doing so
either through this data set or any other.

Q. Is your answer no?

A. Would you repeat the question,
counsel?

Q. In tables one thru [sic] four,
you do not make an effort to adjust for
people interested in migrant seasonal
work as opposed to non-migrant seasonal
work, do you?

A. No, I do not.

Q. You would agree, would you not
that in any given geographical area, most
people would prefer full-year fixed
location employment to migrant seasonal
employment, correct?

A. Well, that's true of most
people in the availability class. Every

one in the work force, I would not assume

A-308

that is true with most people in the other
availability class relevant to seasonal.

Q. I am asking you a general
question. Within any geographical area,
is it not true that most people prefer
migrant--I'm sorry--prefer full-year
fixed location employment to migrant
seasonal employment?

A. Yes, it is true. I thought that
had been asked and answered before,
counsel.

Q. In tables one thru [sic] four,
you focus on certain groups in looking at
your relevant seasonal availability
class, correct.

A. (No response).

Q. Certain occupational groups?

A. Yes, those tables are by job
family.

Q. Okay, what I am getting at is
how you developed your target figure for

those tables, okay?

A-309

A. Yes, well, that was described
yesterday in Dr. Price's testimony how we
did that, I will be happy to explain
further if that is helpful.

Q. You focused on certain
occupations, among other things, in
Geveloping your availability figures in
tables one thru [sic] four, correct?

A. Yes, in the availability
figures for the individual job families,
we focused on certain occupations.

Q. And certain industries as well?

A. The industries are taken into
account in defining the availability
class, not in defining the relevant
workers in a particular job family.

Q. Among the industries that you
focused on in your availability figures
in tables one thru [sic] four, was the
construction industry, correct?

A. That is correct.

A-310

Q. And another was those employed
as teachers?

A. Yes, in the education industry,
generally, not necessarily as teachers.

Q. And, another was those employed
in the canning of fruits and vegetables
and seafoods?

A. Correct.

Q. And, another was those engaged
in agriculture?

A. Correct.

Q. And then students were yet
another?

A. Yes.

Q. And the unemployed was yet
another, correct?

A. Correct.

Q. Okay, let's start with people
involved in construction. The Alaskan
salmon canning industry employment in
that industry peaks int he [sic]

summertime, correct?

A-311

A. That's correct.

Q. And, employment in the
construction industry also peaks in the
summertime, does it not?

A. Yes, in most parts of the
country. I don't know how big the season
is in the Pacific Northwest.

Q. But you'd expect it to be large
in the State of Alaska, wouldn't you?

A. If construction is started,
yes, very large.

Q. So far as the time of the year
goes, would it be fair to say that the
summer is that time of the year when
people in the construction industry would
be less available for work in the salmon
industry?

A. They may nevertheless be
between jobs during the summer because
employment in that industry is often for

short stretches.

A-312

Q. That wasn't my question. Let
me ask it again. Is it fair to say that
the summer would be that time of the year
when people in the construction industry
would be less available for work in the
salmon industry?

A. That would certainly be true in
Alaska. I would not know whether it is
true in Washington or Oregon.

Q. That is because you are not
familiar with the weather patterns here?

A. That's right. It was my
impression you have mild winters here and
they would not impede construction work.
You certainly know more about that than I
éo.

Q. Now, some construction trade
hire out of hiring halls, is that correct?

A. Yes, many of them do. The
unionized construction trade, not all

construction is union.

A-313

Q. Now, among the unionized
construction trades, hiring is often done
or dispatching is often done on the basis
of experienced weighted dispatched lists,
is it not?

A. I believe that it is. I'm not
very familiar with how hire halls work in
the construction industry.

Q. Well, is it your understanding
that people who have "X" hours of work in
the trade, in the locale, would be given
first preference, and people who have
smaller number of hours have a more
preference?

A. That is certainly one possible
arrangement, and I believe that it is
fairly common.

Q. So, if a person who had been
engaged in the construction industry left
the industry during the period of peak
development, namely the summer, he or she

might lose the opportunity to build up

A-314

preferences for dispatches in the latter
parts of the year, correct?

A. That's true. If the person is a
unionized construction worker.
Throughout this country, most
construction work in the residential
construction is non-union, and therefore,
it doesn't have a hiring hall and doesn't
have this kind of preference.

Q. But in your availability
figures you made no effort to weed out
non-union from union construction
workers, did you?

A. We can't because the census
doesn't identify union memberships.

Q. At least for unionized
construction workers, who work under this
particular type of dispatch system, the
loss of experience which could affect
employment opportunities in the latter

part of the year would be a reason not to

A-315

take employment in the salmon industry,
wouldn't it?

A. Yes, it might.

Q. Now, you have been a college
professor for quite some time, correct?

A. Well, I no longer am, but I was
for quite some time.

Q. And, let me ask you whether
you, yourself, say in the last twenty
years, were ever available, ever
available during the summer to go to work
in Alaska as a cannery worker?

A. No, sir, I was not.

Q. And one reason you found
employment in your profession is that was
during the summer?

A. Yes, I did most summers.

Q. And another is was that going
to Alaska would cause you a certain amount
of personal dislocation?

A. Yes, though I never lived in

that part of the country, counsel.

A-316

Q. Now, even putti.g aside the
question of wages, you could lose certain
benefits to your career if you took a
summer off to go to Alaska as opposed to
doing something like working on a
research paper, correct?

A. Early in my career that would
have been correct.

Q. And so there are a number of
reason [sic] why you, yourself, would not
have been available for work in Alaska,
correct?

A. Yes.

Q. And that is true of most other
college professors you know, isn't it?

A. Well, the ones I know tend to be
at the major universities and have the
best opportunities for summer work. I
would not want to genera.ize to all
college profec.ors and all little
colleges across the country. It is my

impression that professors in small

A-317

liberal colleges are quite frequently
available for other kinds of employment
during the summer.

Q. My question was that is true of
most of the other college professors you
know, isn't it?

A. Yes, it is.

Q. And in fact, to your knowledge,
that is true of most college professors on
the West Coast, isn't it?

A. I know very few college
professors on the West Coast.

Q. To your knowledge, that is true
of most college professors on the West
Coast, isn't it?

A. It is true of the ones I know.

Q. Okay, did you make any attempt
in devising your availability figure to
weed out college professors?

A. No, I don't think we should

have.

A-318

Q. Now, at least some high school
teachers are not available for migrant
seasonal work, is that correct?

A. In all parts of this
availability class some people are not
available for migrant seasonal work,
we're not attempting to identify a class
of people all of whom are available, we
are attempting to identify a class of
people from whom an employer might
reasonable [sic] be expected to draw some
employees.

Q. I gather that your answer is at
least some high school teachers are not
esebtebte for migrant seasonal
employment, correct?

A. That's correct.

Q. And you do not have a way of
determining exactly what percentage of
high school teachers are available and

what percentage are not available, do

you?

A-319

A. No, I do not.
Q. And in fact those available
could be comprised only of a small

percentage of high school’ teachers,
correct?

A. They could.

Q. Now, in the other industry that
you focused on in determining your
availability figures for table one
through four, there are at least some full
year employees, for example, a manager,
Or a mechanic in agriculture who might be
a full year employee, correct?

A. Yes, he might.

Q. And in the industries that you
looked at in determining your
availability figure for tables one
through four, you did not factor out those
who had full year employment as opposed to
those who had only seasonal employment,

correct?

A-320

A. That's correct. Well, there is
one exception to that, and that is the
unemployed who form a fairly substantial
part of that class, and my definition is
they were not employed at the time the
census was taken.

Q. I'm speaking of industry
Classifications.

A. Rather than the whole
availability class?

Q. Right.

A. With that qualification, it is
correct.

Q. Now, would it be fair to say
that whether a given student takes work in
the salmon industry during the summer
depends on whether or not better options
are available to him or her?

A. Yes, that's correct.

Q. And you did not adjust your
Gata to exclude students who had better

options, did you?

A-321

A. There is no way of identifying
in that data base or any data base that I
know of.

Q. Now, even among the unemployed
there would be people who are not
available for migrant seasonal work in
the this [sic] industry.

A. There might be some, sure.

Q. of your relevant seasonal
availability class, namely the ones in
table one through four, do you know what
percentage of those people were involved
in the construction industry?

A. I don't know off-hand. We
could probably provide that if it is
important.

Q. Do you know what percentage are
teachers?

A. No, I don't have a recollection
of how that available class devised a.nong

its various components.

A-322

Q. Now, in arriving at your
availability figures, you used the census
one and one hundred use tapes, did you
not?

A. That's correct.

Q. And the occupational detail in
those tapes appears only at the level of
county groups, is that right?

A. That's correct.

Q. The entire State of Alaska
forms one county group, doesn't it?

A. Yes, it does.

Q. Do you know what the percent
non-white is in the State of Alaska?

A. My recollection is that for the
total population of Alaska, it is about
fifteen percent.

Q. And that would be considerably
lower than the percent non-white in
Alaska native villages, wouldn't it?

A. Yes, very much so.

A-323

Q. Is it your understanding that
those Alaska residents employed at Ekuk,
Bumble Bee, Red Salmon, and Alitek [sic],
during the case period, came almost
entirely from Alaska native villages?

A. Yes, that is my understanding
with respect to the particular facilities
that you have named.

Q. You used various types of
weighting schemes in devising your
availability figures, correct?

A. Yes, we used two. We weight a
job and we weight by the section of the
job. My preference is for the second of
the two. The reason, would you like me to
explain the reason for that?

Q. Well, I would rather you just
answer the question, frankly. I
think--it is up to the court if the
court--

THE COURT: It could be covered on

re-direct.

A-324

BY MR. ARDITI:

Q. In weighting by job, then, you
used the residence of the people actually
hired in those jobs to determine the
‘weighting that should be given in the
availability figure, correct?

A. Yes, the people hired in those
jobs at all of the Alaskan facilities of
the three defendant companies, with the
exception of the facility at Icy Cape.

Q. Now, when you weighted by
section, you took as one section, cannery
worker and laborer, correct?

A. Correct.

Q. And you took as the other
section all jobs except cannery worker or
laborer, correct?

A. Correct.

Q. And when you devised
availability figures for each of the non-
cannery worker and non-laborer jobs, you

weighted by that section that was

A-325

comprised of jobs other than the cannery
worker and other than laborer, correct?

A. Yes, for each of the job
families.

Q. Now, where an employer recruits
can itself be discriminatory, can it not?

A. It can be.

Q. And where there is some reason
to believe that an employer's recruitment
practices are discriminatory, it would be
less preferable than in other cases to use
a weighting scheme based upon where
employees were hired, correct?

A. I have no reason to believe
that that's true in this case.

Q. My question was a general one,
and you can you answer it, or would you
like--

A. If I had a reason to believe
that an employer was deliberately
recruiting in some areas and not in others

in order to influence the racial

A-326

composition of his work force I would have
to take that under account.

Q. And you would also have to take
it into account if you had reason to
believe that the employer was
deliberately recruiting in heavily non-
white areas for low paying jobs or less
desirable jobs, and Geliberately
recruiting in predominately white areas
for high paying or more desirable jobs,
correct?

A. This is still a hypothetical
question?

Q. Yes, it is.

A. I would have to take those
facts into account.

Q. Now, in your table you attempt
to take a skills in the sense that on at
least some occasions you looked at census
occupational categories that you felt
were geared tc skill?

A. Yes, that is correct.

A-327

Q. Now, in selecting those
occupational categories, you relied
solely on Mr. DeFrance's analysis, is -
that correct?

A. That is correct.

* *«
[pp. 1929-1930]

Q. Is it fair to say then that your
statistics on the availability of non-
whites with the relevant skills rises and
falls with Mr. DeFrance's cross-mapping?

A. Yes, if Mr. DeFrance's cross-
mapping bore no resemblance to the skills
that were needed in these job families,
then we would be in trouble.

MR. ARDITI: Now, could the witness
be handed Exhibit A-278.

(The exhibit was handed to the
witness by the Bailiff).

BY MR. ARDITI:

Q. Do you have that in front of
you? |

A. Yes, I do.

A-328

[pp. 1931-1932]

Q. If you're more comfortable, we
can turn to Red Salmon, that is perfectly
fine, turn to Red Salmon and I'm going to
ask you to compare tables one and three.

A. Alright.

Q. Now, your availability figures
were given in the row labled _ [sic]
"target", right?

A. That is correct.

Q. And they are given as percent
white rather than percent non-white?

A. Correct.

Q. Is it fair to conclude by
comparing tables one and three that when
you adjust for skill, you find that there
are more non-whites, in other words the
availability figure for non-white
fishermen increases?

A. The target in table one is

eighty-nine point five zero six, in table

A-329

three it is eighty-two point six three

zero, yes, in that particular instance it
does increase, that is a possible result.

Q. Now, let's look at another
occupation or another catagory [sic].
Let's look at machinists.

A. The same two tables?

Q. The same two tables.

A. Yes, okay.

Q. Now, in those two tables, when
you adjust for skill, the availability of
non-whites goes down somewhat?

A. In table one, we have got the
same availability figure right cross the
tables and in table one the availability
is not being done by job family.

Q. I understand that, but in table
three it is, correct?

A. In table, it is [sic]. In the
case of machinists, the availability of

non-whites is lower, in table three, or it

A-330

Wy

is lower in table three than it is in
table one.

Q. Okay, so when you adjust for
skills with machinists, what you are
really doing is bring the availability
figure for non-whites down about three
percentage points, is that correct?

A. That's correct.

Q. Now, I would like to ask you
about instances in which you found
statistically significant under-
representation of non-whites, okay?

A. Yes.

[p. 1941)

Q. In fact, you find statistically
significant under-representation of non-
whites in machinist jobs and company
fishermen jobs at South Naknek in tables
one, two, three, four, five, six, seven

and eight of that exhibit?

A-331

A. Yes, I would say that
particular case, the evidence of under-
representation is fairly conclusive and
does not depend on which table you look
at.

Q. Now, in Exhibit A-280, you use
yet another method of counting hire,
correct?

A. Yes, 280 uses all hires that it
counts a person each time he appears on a
payroll, even if he has appeared in that
job, that facility in the previous
season.

Q. Let's see if we can't verify
that. It refers to all seasonal hires?

A. All seasonal hires, correct.

Q. So that would then indicate or
count as a hire each occasion an
individual appeared in a seasonal job,
correct?

A. Yes, correct.

A-332

Q. And as you understand it, the
deviation of hires for seasonal jobs here
is essentially a definition that
Dr. Flanagan was using?

A. That's correct.

Q. Let's look at the South Naknek
tab of Exhibit A-280, okay?

A. Yes.

{[pp. 1942-1946]

Q. Okay, let's go back to table
one of the South Naknek tab of Exhibit
280, A-280.

A. Correct.

Q. Now, in table--or in that
table, we find statistically significant
under-representation of non-whites, not
only in company fishermen and machinist
jobs, but also in carpenter and tender
jobs, correct?

A. Correct.

Q. And we find the same thing in

table two, correct?

A-333

A. That is correct.

Q. In table three, we find
statistically significant under-
representation of non-whites not only in
fishermen and machinist jobs, but also in
tender jobs, correct?

A. Correct.

Q. Now, in table four, table five,
and table six, and table eight, we also
find statistically significant under-
representation of non-whites not only in
fishermen and machinist jobs, but also
Carpenter and tender jobs, correct?

A. You skipped seven, but included
eight, is that correct?

Q. Yes.

A. Yes, that is correct.

Q. Now, let's turn to the Red
Salmon tab--strike that--let's go to
Wards Cove.

A. Wards Cove in A-280, I have it.

A-334

Q. Okay, on this table, table one,
we find statistically significant under-
representation of non-whites in
miscellaneous, machinists and tender
jobs, correct?

A. Correct.

Q. In table two, we find
statistically significant under-
representation of non-whites in
machinists and tender jobs, correct?

A. Correct.

Q. In table three, we find
statistically significant under-
representation of non-whites in
machinists jobs, correct?

A. Correct.

Q. And table four, we find it in
machinists and tender jobs, right?

A. Right.

Q. And in table five, we find
statistically significant under-

representation of non-whites in three

A-335

department, miscellaneous, machinist, and
tender, correct?

A. Correct.

Q. And in tabale [sic] six, we
fine it in machinist?

A. Yes.

Q. The same in table seven?

A. Yes.

Q. And in table eight, we find it
in both machinists and the tender jobs,
correct?

A. Table eight?

Q. Right.

A. I show for tender composite
deviation of one point nine zero five
which would not be significant at the five
percent level as to the significance.

Q. Okay, let's turn to Red Salmon.

A. Yes.
Q. In table one, we find
statistically--

A-336

THE COURT: Mr. Arditi, you don't
need to go through these, I can go in and
read these and see which is more than one
nine six if that is the point of
examination, is that the point of it?

MR. ARDITI: That is basically--

THE COURT: Well, I think I can
handle that.

BY MR. ARDITI:

Q. In Exhibit 281, you used a
Gefinition of “all hires", right?

A. Yes.

Q. And how does that differ from
the definition that you use for hires in
Exhibit 280?

A. Again, the difference is the
people who work during the winter at Lake
Union Terminals, they are included in
281, excluded in 280.

Q. Now, you mentioned at page five
of your affidavit that you prefer using

new hires to all hires, is that correct?

A-337

A. That's correct.

Q. And this is in essence saying
you have a different view of vacancies
than Dr. Flanagan?

A. That's right.

Q. Now, in any of your tables that
counts new hires or new seasonal hires,
you look only at hires which occured [sic]
from 1981 onward, is that correct?

A. 1971.

Q. I mean ‘71 onward, is that
right?

A. Yes, for the season 1971
through ‘80.

Q. And I think you say at page
five, lines twenty-eight through thirty
that you believe that this is acceptable
because if there is any racial
Giscrimination your method of counting
hires world show it in the initial hiring,

correct?

A-338

A. That's correct. If there is
Giscrimination in the period covered by
the litigation.

Q. Exactly. Now, if there were
Giscrimination in 1970, your figures
would not show it, correct?

A. If there had been
Giscrimination prior to 1970 that was
reflected in the composition of the 1970
labor force, yes.

Q. Now, it would be possible that
a policy of rehiring past employees in
their old jobs would--would or could
perpetuate discrimination that occurred
prior to 1971, isn't it?

A. Well, it is. May I state my
reasons for preferring new hires to all
hires?

MR. ARDITI: Well, that would be up
to the court again.

THE COURT: Alright, Mr. Duncan will

cover that.

A-339

THE WITNESS: Alright.
BY MR. ARDITI:

Q. Now, your definition of hires
Goes not take account of this possibility
that there may have been discrimination
in 1970 ich is perpetuated by a practice
of rehiring past incumbents in their old
job, does it?

A. That practice is required under
the defendants’ union contracts.

Q. Dr. Rees, that wasn't my
question. My question is--

THE COURT: Just answer the
question.

THE WITNES: Alright. Would you
restate the question, counsel?

BY MR. ARDITI:

Q. Your preferred method of
counting vacancies does not take into
account of possible discrimination in

1970 which is them [sic] perpetuated

A-340

through a practice of rehiring past
incumbents in their old jobs?

A. No, it does not.

**e
[pp. 1947-1955]

Q. Now, as TI understand these
tables, when you find something in the
composite deviation row, which is sinus
one point nine six or less, we have
statistically significant under-
representation of whites, is that
correct?

A. That‘*s correct.

Q. Let's start with--

THE COURT: Excuse me, just a minute.
Did you say minus one point six nine six
or less?

MR. ARDITI: Yes, it would go to--

THE WITNESS: You figure the bigger
minus number.

THE COURT: Alright.

A-341

MR. ARDITI: I wasn't sure how to
phrase that exactly.

THE COURT: Alright.

BY MR. ARDITI:

Q. Let's start with this
miscellaneous col. ~ there.

A. Okay.

Q. Well, first of all, to your
knowledge, miscellaneous jobs at Ekuk are
not covered by the Local 37 contract, are
they?

A. I don't have any knowledge of
that, I don't really know what has been
put in the miscellaneous job column at
Ekuk.

Q. Well, the minus six point six
nine six standard deviation would show a
highly significant under-representation
of non-whites--I'm sorry, a highly
significant under-representation of
whites, wouldn't it?

A. Yes, it would.

A-342

Q. Okay, I want you to assume
those are non-union jobs, okay?

A. Alright, I will assume that.

Q. That under-representation,
then, could not be attributable to any
union practices, could it, in hiring?

A. If those jobs are not covered
by any union, no, it would not.

Q. Let's look at the beachgang, it
is your understanding that that job is
covered by the Alaska Fishermen's Union
rather than any other?

A. I have looked at the
agreements, counsel, but I don't have
detailed recollection of just which jobs
at which facility are covered by the
facility. There are a lot of unions
involved here.

Q. Alright. We find highly
significant under-representation of
whites in the beachgang job, don't we?

A. Yes, we do.

A-343

Q. And if it is your understanding
that there are no Local 37 workers at
Ekuk?

A. Yes, that is my understanding.

Q. We find a _ statistical [sic]
significant under-representation of
whites in the culinary department,
correct?

A. Correct.

Q. And, we find _ =
storekeeper, stockroom department,
correct?

A. It's statistically significant.
It's not as pronounced as in the other
cases you have mentioned.

Q. And we find it in the tender
department, correct?

A. Seeetne.

Q. And we find it to an almost
alarming degree in the cannery worker
department, don't we?

A. Why is ic alarming?

A-344

Q. Well, it is minus thirty-five
point eight one zero standard deviation.

A. I find it large but not the
least alarming.

Q. What is the statistical
probability that you could come up with a
disparity that great between the actual
percent white hired and the target
percent white solely by chance?

A. If you were hiring these
workers solely by chance from the entire
area from which these defendants recruit,
that would be a highly improbable event.
In fact, we know, both you and I know that
is not the way cannery work is not
recruited at Ekuk. I can explain that if
you want me to, but it isn't alarming.

Q. Minus one point nine six
standard deviations would indicate that
something other than chance is happening
at the level of ninety-five percent

certainty, correct?

A-345

A. Yes.

Q. Okay--

A. That you're not hiring at
random from the whole geographical area
on which these tables are based.

Q. Okay, can you give me an idea of
what percent certainty you find at the
level of minus thirty-five point eight
one zero standard deviations?

A. Well, it's an event that would
happen by chance in a random draw from the
whole population, certainly, that's the
one time in ten thousand, and it is way
out of the tale of the probability
distribution. I'm sure Dr. Wise could
compute it for me, but I can't guess it.

Q. Now, is it fair to assume, just
in lay people's terms, what or when you
find these minus signs which are really
finding as a suspicion of discrimination

against whites?

A-346

A. You can call that if you like,
what you are finding is that that
particular work force is heavily non-
white.

Q. Okay, but your finding has such
a high degree that it raises the inference
of discrimination against whites,
correct?

A. Well, that's one way of
Gescribing it.

Q. Now, just generally speaking,
would it be fair to say that there is a
fairly high incidence of this
statistically significant under-
representation of whites in the job
tables that you prepared, the exhibits
you prepared?

A. Yes, it happens more often than
the reverse, than statistically
Significant discrimination agai: *t non-

whites.

A-347

Q. Ang it happens at Ekuk where
there is no Local 37, correct?

THE COURT: You are talking now about
cannery workers?

MR. ARDITI: I'm talking about all
Gepartments, and I will be more specific.
BY MR. ARDITI:

Q. It happens at a number of
Gepartments at Ekuk where there is no
Local 37 workers, correct?

A. Correct.

Q. It happens at Red Salmon, for
example, in labor departments were [sic]
workers are not represented by Local 37,
correct?

A. If you give me just a second to
get the--that is correct.

Q. It happens at Alite« [sic] in
miscellaneous, beachgang, carpenter,
tender and labor jobs, none of which are

covered by Local 37, correct?

A-348

MR. DUNCAN: Your Honor, I would
like to know which table we're looking at
when we're coming up with these
particular figures.

THE COURT: Airight, please refer to
the table, Mr. Arditi.

BY MR. ARDITI:

Q. Well, let’s start with takle

THE COURT: Of which cannery?
MR. ARDITI: Alitak, Exhibit A-278.
THE WITNESS: Okay.
BY MR. ARDITI:
Q. It happens at miscellaneous--
THE COURT: Wait a minute. Alright,
go ahead.
BY MR. ARDITI:
Q. It happens in miscellaneous,

beachgang, carpenter, culinary,
machinist, tender, laborer, correct?

A. Correct.

A-349

Q. And none of those would be
attributable to Local 37, would it?

A. No, some of them are
represented by other unions that have
clauses in their agreement giving
preferences to hiring of residents of
Alaska.

Q. Now, if your availability
figure for non-whites were hire, [sic] it
would reduce the incidence of this
statistically significant under-
representation of whites, wouldn't it?

A. Yes, it would.

Q. Ané, in fact one possible
interpretation of this pretcy substantial
incidence of statistically significant
under-representation of whites is that
your non-white availability figure is too
low, correct? Isn*’t that an

interpretation or possible interpretation

of it?

A-350

A. I think maybe too low for some
of the facilities on Bristol Bay which in
some unskilled job categories hire
heavily from native villages and not from
the region as a whole, but I have no way
of making separate availability figures
for those facilities from the census data
because the census date [sic] treats
Alaska as a unit.

Q. So, to that degree, your
reliance on the one and one hundred public
use sample tapes has produced for you
availability figures for non-whites in
the Bristol Bay canneries which are in
fact too low, correct?

A. They are too low for those
unskilled and semi-skilled jobs where
recruitment is in the native villages,
yes. They are not necessarily too low for
some of the skilled jobs where

recruitment is in the Lower Porty-eight.

A-351

Alright, but if going back to an
earlier section of our discussion today
you had reason to believe that there was
Geliberate recruitment of cannery workers
only in heavily non-white areas, and
Geliberate recruitment of other people
for other jobs in heavily white areas,
then that availability figure would be
too low, not only for what you lable [sic]
the unskilled jobs, but also for the other
jobs as well, correct?

A. I'm sorry, I lost that,
counsel, could you repeat it?

Q. If I understood your previous
response correctly, it was that one
possible interpretation of this under-
representation of whites was that your
availability figure for non-whites in
unskilled jobs in the Bristol Bay

canneries was too low, correct?

A-352

A. Yes, given the recruitment
patterns of those canneries for those
jobs.

Q. And you attributed that to the
fact that the census one and one hundred
public use sample tapes do not give fine
enough detail on a geographical level for
you to get an accurate assessment,
correct?

A. That is true within the State
of Alaska only. In the other states they
give a sufficient detail.

Q. And the reason, I believe, that
you limited that conclusion just to what
you call the unskilled jobs is that you
felt that that was the only place
where--I'm sorry--the reasons that you
limited that to what you called the
skilled jobs is that you believe that only
recruitment for the unskilled jobs was
done in Alaskan native villages, correct?

A. That's correct.

A-353

Q. Now, if you had reason to
believe that the recruitment practices at
least as far as selecting which location
to recruit for different jobs was
concerned, you had reason to believe that
the recruitment practices themselves were
Geliberately discriminatory, then you
would also have to suspect, wouldn't you,
that your use of the one and one hundred
public use sample tapes was also
inappropriate for what you call the
skilled jobs at the Bristol Bay cannery,
correct?

A. Tf I believed that, that would
be correct, yes.

Q. Thank you.

A. I have not--I happen not to
believe it, I think there are other ways
in which this can be explained, should be
explained.

Q. Let me ask you to draw on your

general expertise which I gather was

A-354

substantial as a labor economist.
Generally, in our particular society, we
find that more discrimination against
non-whites than we do against whites,
Gon*t we?

A. Yes, that's correct.

Q. And from your examination of
the wage information in this case, we
could conclude that the whites receive
the higher paying jobs as opposed to the
lower paying jobs, correct, on the whole?

A. On the whole, although the
cases you have just been citing are cases
of over-representation of non-whites in
skilled jobs so it is by no means a
uniform picture.

Q. That is not my question. If you
look at the higher paying departments at
the canneries, you will find they are more
heavily white than the lower paying
departments, correct?

A. That is in general correct.

A-355

Q. Okay, so when you were looking
at the distribution of non-white and non-
white workers at the canneries, you find
that the white workers generally get the
better paying jobs as opposed to their
non-white counterparts, correct?

A. There is a high proportion of
white workers in the better paying jobs,
yes.

Q. And based on the contact you
had in this case with management, you
would conclude that management is largely
white, is it not, at least at the highest
levels?

A. The representatives of
management I have met have been white,
yes.

Q. I gather from your tables that
you find there is a higher incidence of
statistically significant under-
representation of whites than of non-

whites, correct?

A-356

A. Taking all departments of the
cannery together, or the facilities into
account, that is correct, yes.

Q. Dr. Rees, does it jar your
common sense at all to conclude that in
light of the fact that society-wise,
there is a, there is greater
Giscrimintion [sic] against non-whites
than whites, in light of the fact that
whites generally get the better paying
jobs at the cannery than non-whites and in
light of the fact that management at the
upper levels of the canneries is white
rather than non-white, doesn't it jar you
[sic] common sense at least to some degree
to conclude that there is greater
discrimination against whites than non-
whites?

A. I haven't concluded there is
Giscrimination against whites, I have
concluded that whites are over-

represented in certain departments. We

A-357

haven't yet had an opportunity to discuss
why.

Q. Now, Dr. Rees, I gather from
your affidavit that it is your belief that
Local 37 in some way contributes to the
substantial number of non-whites in this
particular industry, is that correct?

A. Yes, in those facilities where
they represent the cnanery [sic] workers
and some of the culinary workers, they
clearly contribute to the higher
proportion of non-whites in those two

departments.

[pp. 1958-1959)

Q. Let me just ask you this,
Dr. Rees, in your view as a labor
economist, -have you ever encountered
situations where an employer tries to
blame his discrimination on a union?

A. To blame it on a union?

Q. Yes.

A-358

A. A few cases, yes.

Q. Now, is it not correct to say,
in your view, that an employer's
statements about the source of
discrimination are not always reliable
for just that reason, mainly that they
will try to blame it on another?

THE COURT: I think that question is
a question that I am going to have to
resolve, Mr. Arditi.

MR. ARDITI: He has written about
it.

THE COURT: Okay, well, if he has
then ask him about it.

BY MR. ARDITI:

Q. Have you written oon that
subject?

A. Yes, the statement you have
just made, I think is a quote or a
paraphrase from my textbook, "The

Beonemics of Work and Pay”.

A-359

[p. 1960)

Q. Now, in your view as a labor
economist, | one form of employment
Giscrimination occurs when an employer
succumbs to the prejudice of its
employees, correct?

A. Correct.

Q. Okay, and this can lead to
segregated work crews?

A. Can, yes.

Q. And an example of this is
hiring Black janitors to work at night by
no Blacks in production worker jobs,
correct?

A. Tha. would be an example, yes.

Q. And this minimizes contacts
between whites who may be prejudice and
non-whites, correct?

A. Correct.

Q. And another way of minimizing

this contact in a facility that provides

A-360

housing is to have separate housing for

whites and non-whites, correct?

A. Yes.
* + @#
{[p. 1969]
Q. Dr. Rees, what is the

probability in any given table, in
Exhibit A-278, that there would be one or
more out of ten job families with
more--with a positive one point nine six
standard deviation?

A. If you regard each job family
hiring into each job family as an
independent event, then Dr. Wise has
calculated that probability by twenty
percent. In one table out of five you
woulé expect to find such an event. That
is, simply by chance, simply as a random

of cards.

BY MR. ARDITI:

A-361

[pp. 1970-1973)

Q. Now, is it your understanding
that many of the native villages are
coastal villages?

A. I know a map has been
introduced and where the villages are, I
haven't examined it.

Q. Why don't you assume that many
of these villages are coastal villages,
and pueese assume further that many of the
residents of these villages grew up
around water and around boats. Under
those circumstances, when you weight by
areas from which the employer recruits
for that job, isn't it possible that you
are loosing [sic] a great many non-whites
in these areas who would have the skills
required for tender or company fisherman
jobs?

A. I think that is probably right,

that may very well be why in your earlier

cross-examination you discovered in the

-.- 362

tables some of those job families in which
non-whites were over-represented.

Q. Let me--you would use _ those
figures then for tender jobs, correct,
you would want to weight by cannery worker
hiring for such tenders?

A. No, no. If we had data, which
unfortunately we don't, which gave us
geographical detail within Alaska by
occupation, I would have wanted to use
those. They don't exist, and that would
show up these people that you are talking
about that have whose occupation is
fisherman and whose occupation is boat
caption [sic] or whatever.

° It would be fair to assume,
would it not, that if you weighted by the
entire cannery worker work force rather
than by specific jobs or job sections that
percent of non-white in your availability
figure would go up?

A. It goes up slightly, yes.

A-363

Q. And it makes at least some
sense to do that for jobs such as tender
and company fisherman which involve
marine work, correct?

A. Arguable.

Q. Now, is it your understanding
that the percent non-white and cannery
worker jobs at Ekuk is as high or higher
than at any of the other canneries at
issue?

A. I would have to check that. It
can be done very easily.

Q. Why don't you check it.

A. The cannery worker jobs at Ekuk
are approximately thirty-five percent
white. It is fairly low. I'm sorry, that
is Alitak. Ekuk, at Ekuk it's thirty-
three percent white or whites, it is a
couple of percentage points below Alitak.
South Naknek is’ forty-eight percent
white, cannery department. Red Salmon is

thirty-six percent white, Wards Cove,

A-364

sixty-nine percent white. So, yes, it is
one of the lowest.

Q. That is one of the lowest
percent white?

A. Percent white, yes.

Q. And correspondingly, it has the
highest percent non-white, correct?

A. That is correct.

Q. And that is the cannery that
does not use Local 37 labor, is that
correct?

A. That is correct.

Q. Now, if I understood your
testimony correctly, in your yiew, at
least if Local 37 were somehow cut out of
the picture, the percent Filipino and
other Asians would decline, correct?

A. It would decline gradually over
the years. That doesn't mean that
experienced workers wouldn't be rehired.

I assume that they would.

A-365

Q. But, you're not able to predict
the precise racial mix of those who would
come in and fill those jobs, are you?

A. No, I am not.

6. And at least some of those jobs
would be filled, in yoru [sic] view, by
non-whites who are not of Filipino or
Asian descent, correct?

A. Correct, that is especially

true in the Bristol Bay area.

x* ke

A-366

DISTRICT COURT

TESTIMONY OF ROBERT J. FLANAGAN
[p. 2064]

Q. Could you tell us what the
correction is.

A. The corrections are in the
second sentence of the second full
paragraph. Five lines down, cross out the
word "plaintiffs" and substitute the word
"defendants." Two words before that,
cross out the word "perform" and
substitute the following phrase: Based
on both the 47 percent figure, and the
high and low estimates of availability
developed. To be sure that's an insert,
let me read the corrected sentence
entirely just to be sure it's correct in
the record. The second sentence.
"Nevertheless, I note that statistical
tests, based on both the 47 percent figure

and the high and low estimates of

A-367

availability developed by defendant's
expert Dr. Albert Rees, indicate that
employment of non-white Carpenters
generally fell below the confidence
interval at the Bumblebee, Red Salmon,
Ekuk and Alitak canneries.
x* *&

[p. 2121 (a)-(b)]

ROBERT J. FLANAGAN, being first
Sworn, states:

Attached is a document entitled
Witness Statement of Robert J. Flanagan
which summarizes my testimony in this
case. I incorporate its contents into

this affidavit by this reference.

A-368

WITNESS STATEMENT OF
ROBERT J. FLANAGAN
Associate Professor of Labor
Economics and Director of the
Public Management Program
Graduate School of Business
Stanford University
Stanford, California 94305

I. Qualifications and Background

My general expertise is in the area
of labor economics. I have a B.A. degree
from Yale University and an M.A. and a
Ph.D. degree from the University of
California at Berkeley. All of these
degrees are in economics with a
specialization in labor economics and
include training in statistics and
econometrics. I am also a member of the
American Economics Association and the
Industrial Relations Research

Association.

A-369

[p. 2121(c)-(d)]

Labor supply: The number of individuals

who are available for work in a particular
Organization or industry. The most
common empirical analog to this concept

is the labor force which consists of the

employed plus those who are looking for
work in the industry but cannot find it
(the unemployed).

Work force: this is usually a synonym for

employment in an organization.

Seasonal work: work that occurs only part

of a year but usually in about the same
part (months or weeks) of each year. When
seasonal work occurs in areas that are
remote from major population centers, it
usually requires the services of migrant
workers.

Labor market discrimination occurs when

individuals who have equal ability to
perform a job but differ by race or sex

(or some other characteristic that is

A-370

unrelated to their productivity) are
treated differently in terms of pay or
employment opportunities.

Occupational segregation: eccurs when

individuals of a particular race (or sex,
etc.) are disproportionately employed in
lower-wage jobs and underemployed in
higher-wage jobs, given their
qualifications.

Analysis for this Litigation

I was asked by an attorney for the
plaintiffs in this case (1) to determine
the percent of the labor supply to the
defendant salmon canning companies that
was nonwhite for the period since 1970,
and (2) given the percent nonwhite in the
labor supply, to determine whether there
was discrimination in the allocation of

jobs to nonwhites at the defendant

canneries.

[p. 2121 (e)-(k)]

Nonwhite Availability

The choice of a source of data that
provides an accurate measure of the
availability of nonwhites for work in the
Alaska salmon canning industry is
dominated by two unusual characteristics
of employment in that industry:
(1) employment is highly seasonal, since
the canneries normally operate for only a
few months in the summer, and (2) most of
the canneries are in remote locations so
that migrant labor is required.
Seasonal, migratory work offers less job
security and greater disruption to family
life than do full-time, full-year jobs,
and for these reasons it is generally less
preferred. Judged by their job choices,
most workers prefer full-year employment.
Migrant, seasonal workers do not normally
have full-year employment opportunities.

Among the reasons for this is

A-372

i
}
4
;
*
;
4
>

discrimination in full-year employment
opportunities. As a result, the nonwhite
proportion of a migratory labor force is
usually considerably higher than the
nonwhite proportion of the full-year
labor force in the areas from which a
seasonal industry draws its workers, and
it is inappropriate to estimate nonwhite
availability to such an industry solely
from the racial mix of the full-year labor
force in those areas. Therefore, the
basic problem confronted in measuring the
availability of nonwhites for employment
in the Alaskan salmon canning industry
was to find a source of data that provided
a clear indication of availability for
seasonal migratory work.

With respect to statistics on labor
supply, there were three choices. One is
data from the 1970 Census of the
Population. This source provides

considerable data on the size and race

A-373

composition of the labor force
disaggregated by geographical area,
occupation, industry, and other
characteristics, and for this reason it
has been frequently used in determining
availability or labor supply in Title VII
litigation. However, this source has one
fatal defect as a measure of labor supply
for the Alaskan canning industry. The
Census data are dGominated by people in
full-year, full-time employment. For
most Title VII litigation, in which
discrimination is alleged in
organizations engaged in full-time, full-
year operations and employment, the use
of the Census of Population data is
appropriate. However, it is
inappropriate in the present case because
a seasonal, migrant labor force tends to
have a disproportionately large minority
component. The use of Census data will

therefore result in 3 substantial

A-374

understatement of the availability of
nonwhites for work in the industry.

The second alternative is data on
racial employment patterns developed and
published by the Equal Employment
Opportunity Commission from employment
data on the EEO1 forms submitted annually
by many organizations. This source
shares the flaw in the Census data, in
that it is dominated by full-time, full-
year employment, but has other debilities
as well. There is, for example, much less
occupational and geographic detail than
is available in the Census. In addition,
there are a number of shortcomings and
biases in the data that are tied to the
sample and data collection procedures and
that the EEOC itself acknowledges in the
volume in which the data are published.

The third and in my opinion the
preferable alternative is a long

statistical series on employment in the

A-375

entire Alaska salmon canning industry
prepared first (from 1906 to 1939) by the
U.S. Bureau of Fisheries and later (from
1941 to 1955) by the U.S. Fish and
Wildlife Service. (These data are in
Exhibits 61 and 62.) These data have the
advantage of showing the racial
composition of those individuals who
actually choose to work in an industry
characterized by seasonal, migrant work.
In examining these data, I have
considered three possible objections to
their use. First, they provide
information on industry employment, but
the standard measure of labor supply, the
labor force, includes the unemployed as
well as the employed. The number of
unemployed is usually a small fraction of
the number of employed in a labor market,
and, more importantly for the issues in
this litigation, the unemployment rate of

nonwhites generally exceeds the

A-376

unemployment rate of whites by a large
margin. Therefore, if information on the
unemployed were available, it is very
likely that it would raise the estimate of
the nonwhite percent of the labor force
available to the Alaska salmon canning
industry. The use of figures that are
limited to employment provide a
conservative estimate of nonwhite
availability from the standpoint of the
plaintiffs’ interests. The second
possible objection is that the data were
not collected after 1955. This would be a
particularly serious limitation if the
percent nonwhite in the industry was
highly variable over time or if it seemed
to exhibit a particular trend. But in
fact it was very stable. Over the last
seventeen years that the series was
published, the percent nonwhite stayed

around 50 £percent. The fact that

employment at the defendant canneries in

A-377

this case, in the Domingo v. Nefco case,

and in the Carpenter v. Nefco-Fidalgo

case was also around 47 percent nonwhite
during the 1970s indicates that’ the
nonwhite availability to the industry has
remained stable since 1955. The third
potential objection to using this data
source to estimate the nonwhite percent
in the labor supply to the Alaska salmon
canning industry is that the data do not
include information oon skill. In
recognition of this problem, I have made
adjustments for skill in my statistical
analysis. The analysis and the
adjustments will be discussed in the
following section.

In determining the percent nonwhite
available for work in the Alaskan salmon
canning industry I have used the data from
this third source for shoresman_ and
transporters. (They are summarized in

Exhibits 63 and 625.) In order to smooth

A-378

out minor annual fluctuations in the
data, I have taken the average percent
nonwhite for the last five years for which
the data were gathered, 1951-55. The
average percent nonwhite during those
years was 47 percent. Because of the
general stability in -? figure over a
much longer previous period, it appears
to be an appropriate (although somewhat
conservative, given the omission of
unemployment) estimate of the percent
nonwhite available for jobs in the
industry that do not require extensive
training.

Job Allocation Analysis

In order to reach an _ opinion
concerning whether there was evidence of
discrimination in the employment’. of
nonwhites in the defendant's’ salmon
canneries, I performed ae statistical
comparison of the percent nonwhite

available with the percent nonwhite

A-379

employed in various job categories at the
canneries. The standard of
nondiscrimination that is used in the
analysis is random selection, which means
that each individual in the labor pool has
an equal likelihood of being selected for
employment. The purpose of the
statistical analysis is to determine
whether any difference that may exist
between the percent nonwhite available to
the Alaska salmon canning industry (47
percent, as noted above) and the percent
nonwhite hired at any particular cannery
is consistent with random selection in
the allocation of jobs. In general, some
difference between the two figures would
be expected; that is, an individual
cannery would not be expected to match the
47 percent availability figure exactly.
This is because a single cannery hires
only a portion (or, in statistical terms,

a sample) of the industry's migrant,

A-380

seasonal labor force. Because a sample is
Only partially representative of the
population from which it is taken, the
characteristics of the sample (such as
the percent nonwhite) are unlikely to
match exactly the characteristics of the
underlying population. Consequently, in
the present case an employer's work force
will not precisely reflect the racial mix
of the industry-wide labor pool.
However, the difference between the
percent nonwhite employed by a cannery
and the percent nonwhite available (47
percent) must remain within certain
bounds to be consistent with random
selection (i.e., nondiscriminatory
hiring).

The nondiscriminatory range of
nonwhite employment is determined by
using standard statistical concepts of
the normal variation in a characteristic

such as the percent nonwhite associated

A-381

with random (nondiscriminatory)
selection. More specifically,
statisticians and economists frequently
use the standard deviation, a measure of
the normal variation of a characteristic
around the average value of the
characteristic, to construct a
"confidence interval” which is the range
around the availability figure (47
percent in the case of the salmon canning
industry) in which the percent nonwhite
hired is expected to fall most of the time
if an organization is practicing
nondiscriminatory hiring. The confidence
interval that is regularly used in the
social sciences is defined as plus or
minus two standard deviations around the
availability figure. (This is known as

the 95 percent confidence interval.)

* *

A-382

[p. 2121(1)-(p)]

The interpretation of the results of
the statistical analysis when hiring into
a simple job is at issue is as follows.
If the percent .nonwhite selected by an
Organization falls within the confidence
interval, the organization's selection
process is consistent with random or
nondiscriminatory selection. On the
other hand, it is unlikely that an
Organization would select a percent
nonwhite below the confidence interval if
it were selecting randomly. In this
Situation discrimination is likely. When
an organization practices occupational

segregation, i.e., the restriction of

nonwhite workers to low-wage jobs despite

the ability of some to hold high-wage
jobs, the employment of nonwhites will
tend to fall below the confidence

interval for high-wage jobs and may fall

A-3 83

above the confidence interval for low-
wage jobs.

Before examining the results of my
analysis of job allocation at the
defendant canneries, it is necessary to
address two further issues bearing on the
interpretation of the results:
(1) skill, and (2) the role of unions
representing workers at the canneries.
In analyzing the data for the defendant
salmon canneries, I have considered
adjustments that might be necessary for
skill and other qualifications for
employment in the indus

[Text truncated at 120,000 characters. The full text is on the page linked above.]

---

Source: Frix Law Library, https://www.frixlaw.com/law-library/documents/brief%3Amicro_IA40385012_0656%3A04. Public record. Not legal advice.
