Page 20 Portland Observer JUNE 22, 1989
RACIALLY BASED SCORE ADJUSTMENTS DEEMED JUSTIFIED FOR WIDELY USED
EMPLOYMENT TEST-------------------------------------------------------------------------------------------
There are statistical grounds to
continue to adjust the scores of mi
nority job seekers who lake the widely
used General Aptitude Test Battery
(GATB), an expert committee has
told die Deparunent of Labor. Such a
move is necessary to ensure that quali
fied blacks and Hispanics have the
same chances as white applicants of
being referred by the federal/state
Job Service program to employers,
the National Research Council com
mittee concluded.
While die GATB is as good a
predictor of job performance as are
other available employment tests, the
committee said, that ability is never
theless only “ modest.” Many appli
cants who could perform success
fully on the job score low on the test
and all low-scoring groups suffer from
the effects of this prediction error.
Blacks and Hispanics as groups
receive lower scores than do whites
and -w ithout score adjustm ents-
would be disproportionately excluded
from the Job Service job-referral pool,
the committee found.
The effect of score adjustm ents-
both as currently practiced by the
agency and as the committee would
like to see them m odificd-would be
to boost the scores of these minori
ties. The committee recommended
that the adjustment mechanism cho
sen by the Job Service be enough to
compensate for the impact of predic
tion errors made when referring
minority candidates to employers.
“ With no score adjustments, very
low fractions of minority-group
members will be referred for
employment,’’the committee said.
The result, it stressed, would be “ a
severe adverse impact on the em
ployment opportunities of those demo
graphic groups,” particularly if the
GATB is used as widely as proposed
by the Department of Labor (DOL).
Other referral criteria should be used
to supplement GATB test scores, it
added.
The general ability test is taken by
several hundred thousand of about
20 million Job Service applicants
annually.
MODEST VALIDITIES
Properly used, the GATB can help
provide a more able pool of appli
cants than employers would see in
the absence of testing, the committee
affirmed. Based on DOL research
involving more than 500 different
kinds of jobs, the committee con
cluded that the GATB has “ modest
but useful” validities for the types of
jobs typically handled through the
Job Service. However, it found that
the available evidence docs not sup
port the agency’s claim that the GATB
can be used to screen applicants for
all 12,000 job categories in the U.S.
economy.
Moreover, the committee labeled
as “ unfounded” the DOL claims that
broader use of the GATB would save
employers some S80 billion annu
ally through increased workforce ef
ficiency. Such claims were not based
on empirical evidence, it said, al
though it confirmed that individual
employers could realize economic
gains from GATB-based referral of
capable employees.
The study was requested by DOL,
which has been challenged by the
Department of Justice on the legality
of adjusting the scores of minority
test takers. The 14-member commit
tee was asked to help resolve techni
cal issues surrounding the validity of
the GATB as a test instrument. It was
charged especially to examine whether
the GATB could be used for a wider
range of job types and was asked to
study the impact of score adjustments;
it was not asked to resolve the legal
issues involved in the dispute.
The National Research Council is
the principal operating agency of the
National Academies of Sciences and
Engineering.
EXPANDED USE
At issue is how to refer the best
job candidates from the Job Service
(the state and local units of the U.S.
Employment Service, USES) to po
tential employers. Most of the job
orders handled by the Job Service are
for semi-skilled and skilled workers.
Some 19.2 million job applicants used
USES services in 1985, the most re
cent year for which data are avail
able. Only about 8 percent of those
applicants took the GATB, but DOL
officials would like to see the test
administered to virtually all candi
dates and used to help fill nearly all
request for applicants.
The GATB assesses verbal, nu
merical, perceptual, and psychomo
tor skills. However, because job
performance is a function o f com
plex human behavior, no test is likely
to predict it very well, the committee
said. GATB scores, for example, show
only modest correlations with super
visor ratings of job performance. Nev
ertheless, these correlations are simi
lar to those found for other general
ability tests used in employment set
tings, such as military entrance tests.(In
statistical terms, the correlation be
tween GATB scores and subsequent
supervisor evaluations of employees
ranges from 0.2 to 0.4; a perfect
correlation would be 1.0.)
MINORITY SCORES
The less predictive accuracy a test
has, the more likely it is that people
with low scores could nevertheless
do well on the job, the committee
explained. The GATB is no excep
tion, it found. The test produces both
“ false rejections’ ’ - low scorers who
could perform jobs well - and ‘ ‘ false
acceptances” - high scorers who
perform poorly on the job.
Minorities tend to receive lower
scores on the GATB than do major
ity-group job seekers.For example, a
black candidate taking the GATB in
the hopes of being referred to a job
requiring semi-skilled workers might
receive a raw score of 276. That
would pul him at the 50th percentile
(the “ median” score for all blacks
testing for that kind of job. The same
score of 276 is at the 32nd percentile
for Hispanic applicants (for whom
the median score si 295), but only at
the 22nd percentile for other, mostly
white, candidates (whose median score
is 308).
However, because the GATB is
such an imperfect predictor of job
performance, differences among
groups in actual job performance will
be much smaller than their differ
ences in test scores, the committee
said.
“ Our premise is that the inaccu
racy of the test should not unduly
affect the employment prospects of
able minority workers,” the com
mittee wrote. Therefore, it recom
mended that the Job Service con
tinue to construct its referral pools
using adjusted scores. The net effect
of these adjustments should be to
eliminate the disproportionate rates
of false rejection for those groups,
the committee said. At the test’s
current level of validity, referring
candidates of equivalent percentile
rankings within each of three racial
and ethnic groups - black, Hispanic,
and other - provides one way to
achieve the appropriate adjustments,
the committee suggested.
To give the employer the best
information available about prospec
tive employees, the committee rec
ommended that the Job Service in
clude with the referred client’s appli
cation both this within-group per
centile ranking as well as an ‘ ‘expec
tancy score.” An expectancy score
- based on the raw GATB score -
reflects the probability that a given
applicant will perform successfully
on the job as compared with all other
applicants who take the GATB.
IM PR O V IN G TH E TEST
The committee found other flaws
in the test that will have even more
impact if the testrcceivcs more wide
spread use. For example, the test is
highly “ speeded” - many of the
sections have such severe time limits
that the average person can finish
only about one-third of the question.
Test coaches can teach people to
randomly answer the remaining
questions in the last minute of the
test, thereby boosting their scores
but further reducing the test’s predic
tive abilities.
Refinement of the test to reduce
its speededness should begin imme
diately, the committee recommended.
Test security is a second problem
with the GATB because there arc
only two versions of the test. The
committee urged that additional forms
of the test be developed on a continu
ing basis so that no examinees bene
fit unfairly from prior knowledge of
the questions should copies of the
test leak into the public domain.
WHO SHOULD TAKE THE
TEST?
Many applicants will benefit from
taking the test,the committee said.
But it noted that the test is not a good
enough predictor of job performance
for test scores to be used in isolation.
“ Traditional and alternative referral
criteria” -su c h as experience and
education — should continue to be
used to supplement GATB scores.
Moreover, there are some client
groups - non-English speakers and
people with certain handicaps - for
whom standardized testing is inap
propriate, the committee noted. It
suggested that DOL take steps to
ensure that referrals are not made
“ automatically and soley” using
GATB scores.
“ We recommend that no job seeker
be obliged to take the GATB,” the
committee concluded; every office
that uses the GATB to refer candi
dates to jobs should maintain an ‘ ‘al
ternative referral path for those who
choose not to take the test.”
The report was edited by commit
tee chair John A. Hartigan.Eugene
Higgins professor of statistics apd di
rector, Statistical Computing Labo
ratory, Yale University, New Haven,
Conn.; and Research Council study
director Alexandra K. Wigdor.
SEARCH FOR
HOWARD
PRESIDENT
LAUNCHED
Howard University Board of Trus
tees chairman John E. Jacob recently
announced that a search committee
has begun work on the process of
seeking applications for the presi
dency of the university and making a
final recommendation for the appoint
ment of a new prcsidenL
Dr. James E. Cheek, president of
Howard since July 1,1969, announced
his retirement in April, effective June
30, 1989. Dr. Carlton P. Alexis,
executive vice president, will serve
as interim president, effective July 1.
“ The search committee will screen
all of the applications during August,
September, and October and will
interview the best candidates. Il is
anticipated that the board of trustees
will appoint the new president dur
ing the early part of 1990.”
TUNA
CARNATION
CHUNK LIGHT
IN
WATER
•
Yellowfin Tuna is
the lightest colored,
mildest flavored
species of
LIGHT MEAT TUNA
6V2 O Z .
TIN
EACH
THE FRIENDLIEST STORES IN TOWN
SINCE 1908
HOURS
WE RESERVE
THE RIGHT
TO LIMIT
QUAMTTT1ES
WEEKDAYS
8 am Io 9 pm
SUNDAYS
JO am to 7 pm
Specials Effective TUES. thru SUN.
JUNE 2Oth thru 25th, 1989
The Portland Bureau of
Environmental Services
provides sewage collection
and wastewater treatment
services to over 400,000
residents and businesses in
the Portland region.
O ur search for excellence
begins with you.
What are we looking for?
We are looking for career-oriented people
to staff our stores. People who want
to help us maintain, and build upon,
our outstanding reputation of quality,
value and customer service.
What else?
Potential. We are looking for people who
want to learn, grow and expand with us.
People to work in sales, operate our café
and receive stock. People with ideas and
initiative. People who believe that good
enough' never is.
People skills. We want people who
genuinely like others and enjoy helping
them. Those for whom friendliness
is second nature, and service
is a natural instinct.
What can we do for you?
earnings. A profit sharing program that’s
fully funded by our company. A 401 K plan.
A comprehensive health plan. And more.
Opportunity. We offer you the chance to
work with quality individuals in a quality
environment. And since it is our policy
to promote from within, we offer excellent
opportunities for career advancement.
It's our way of rewarding excellence.
We'd like to meet you. If you think you
have what we re looking for contact our
Personnel Office at Downtown Portland,
(503) 224-6666. ext. 205; Washington
Square, (503) 620-0555, ext. 205;
Clackamas Town Center (503) 652-1810,
ext. 205; Lloyd Center (503) 287-2444,
ext. 205; Vancouver Mall (206) 256-8666.
ext. 207; Salem Centre (503) 371-7710,
ext. 205; Downtown Rack (503) 243-1492,
ext. 211; or Clackamas Rack
(503) 654-5415, ext 490.
The rewards are many. Commission sales
that virtually let you determine your
W e are a group of profession
als proudly committed to
maintaining Portland's natural |
environment.
Choose a career
where you will find
your challenge in
helping us to meet
ours.
For information
regarding
em ploym ent
opportunities,
please contact
the City of Portland
Bureau of Personnel
248-4352
Bureau of
Environmental Services
The City of Portland
Bureau of Environmental!
Services
Proudly facing the environment's toughest
challenges, today.
Earl Blumenauer, Com m issioner
John Lang, Administrator
n o rd s tro m
• An Equal Opportunity Employer
PORTLAND OBSERVER
"The E yes and Ears of th e C om m unity"
288-0033
I