|
Treatment of Caucasian and African-American Applicants by
San Francisco Bay Area Employment Agencies;
Results of a Study Utilizing "Testers"
by
Ana P. Nunes and Brad Seligman
SUMMARY
The Testing
Project of The Impact Fund utilized "testers" to study
the impact of race on the prospect of securing an entry-level office
position through an employment agency in the San Francisco Bay Area.
Employment testing is a social science procedure that examines the
treatment of matched job applicants of different ethnicity who present
similar qualifications and interviewing styles to the same employer.
In this study, three pairs of research assistants applied for temporary
or permanent entry-level office work through employment agencies.
The three pairs of research assistants were comparable in all respects
except the variable being studied: race. Each team was composed
of one Caucasian and one African American research assistant.
Eight
of the seventeen agencies studied (47%) were observed to favor the
Caucasian applicant a majority of the time
(7 of these agencies favored the Caucasian applicant in every test).
Only four of the agencies (23.5%) consistently showed no preference
for the Caucasian applicant. When all tests are examined, the Caucasian
applicant was preferred 56% of the time, while the African American
applicant was preferred only 18% of the time. The balance of the
tests showed no clear preference. A "net rate" of adverse
treatment of African Americans (tests showing Caucasian preference
minus tests showing an African-American preference divided by total
tests) yields an adverse treatment rate of 38%.
TESTING
Matched pair
testing has long been used in both the social sciences and in the
struggle for civil rights. Sociologists and psychologists often
conduct empirical studies where they attempt to control all variables
except for the one they are studying. The use of testers is now
commonplace in the field of housing discrimination, and has been
endorsed by the United States Equal Employment Opportunity Commission,
and utilized by the Departments of Justice and Housing and Urban
Development as an appropriate research and civil rights investigative
tool. In order to provide reliable data, research assistants (the
"testers") must be carefully selected and matched with
similar research assistants (who possess similar interviewing styles
and demeanor), then trained to increase their already present similarities.
Paper credentials should be comparable in every respect. The goal
is to present two applicants who are similar in every respect except
for their race. Then, if the two research assistants are treated
differently, it is reasonable to attribute that difference in treatment
to the variable that is being tested: race.
THE TESTING
PROJECT
The Testing
Project was established by The Impact Fund, a nonprofit foundation,
in 1998 to conduct objective research concerning the prevalence
of discrimination. The Project is coordinated by an experienced
testing coordinator, who previously coordinated fair housing tests.
The Project has an Advisory Board that approves research projects
and general policies. Members of the Board include clinical law
professors, civil rights lawyers, and social scientists. The Board
determined that The Testing Project study the impact of race in
the employment agency industry.
INDUSTRY
TESTED
The employment
agency industry was chosen as the focus of this study for several
reasons. First, it is an important and rapidly growing segment of
the labor market. According to the Bureau of Labor Statistics, the
portion of the U.S. workforce employed by temporary agencies rose
from 0.5 percent in 1982 to 2.2 percent in 1997. Nationwide, staffing
companies employ 2.9 million people each day. Not only do employment
agencies offer gainful employment to many individuals, they are
also often the bridge to permanent employment. The ability of employers
to "test run" an employee is attractive in busy offices
where there is little time to screen applicants.
Earlier studies
in Washington D.C. and Chicago found a higher than average level
of disparate treatment of African-American applicants in such agencies
as compared to other sectors of the labor market. The Fair Employment
Council of Washington, D.C. found that its African-American research
assistants encountered less favorable treatment than their white
testing partners 66.7% of the time, while the average rate of disparate
treatment found in other industries was 24%. These earlier studies
were conducted during less than favorable economic conditions. The
average unemployment rate in San Francisco while this study was
taking place was 3.24%. The Testing Project was interested in seeing
what effect a stronger labor market might have on the treatment
of minority applicants.
The Testing
Project initially identified agencies in the San Francisco Bay Area
by using the National Association of Temporary and Staffing Services
lists, local newspapers, and the yellow pages. The project sought
to identify agencies which specialized in entry-level type office
work and that appeared to have a large employee base. Base line
research assistants — individuals that were not paired with other
testers — were then used to confirm that the agencies specialized
in entry-level work and to identify their procedures and requirements.
Ultimately, 17 agencies were selected from downtown San Francisco
and Oakland.
SELECTION
AND TRAINING OF RESEARCH ASSISTANTS
This study employed
three pairs of research assistants — two female pairs and one male
pair. Each pair consisted of one African-American individual and
one Caucasian individual. The selection of the pairs was laborious.
Ninety-three job applicants were screened to finally produce the
six research assistants used in this study. Each pair was matched
with regard to gender, age, level of articulateness, and general
manner. In addition, the pairs selected demonstrated a high level
of observational skills, general professionalism, and punctuality.
Once matched,
the research assistants participated in a three-day training process.
The training focused on four main topics:
1. Skills
necessary to be effective interviewees
The
research assistants were taught effective interview techniques
and then given the opportunity to practice those techniques
in many mock interviews. Test partners watched each other interview
and were instructed to take note of their partner’s demeanor
and style of interviewing. They were instructed to parallel
each other’s styles of interviewing in subsequent mock interviews.
The assistants were critiqued and coached after each interview
as to how to improve their interviewing skills as well as to
ensure that the testing pairs presented themselves in comparable
fashions. The assistants were also instructed to dress appropriately
for interviews, and comparably to each other.
2. Basic
information regarding the industry they would be testing
The research
assistants were told of the typical application process at the
employment agencies. This process usually included the completion
of a standard application form, a typing test, various software
application tests, and basic skills tests such as spelling and
grammar tests. Occasionally, a personality test would also be
administered. The Project Coordinator instructed the assistants
on how to perform on all these tests so as to keep their scores
comparable.
3. Development
of biographies and resumes
The
research assistants were aided by the Project Coordinator in
the development of biographies and resumes that presented comparable
qualifications and backgrounds. The research assistants’ resumes
described them as possessing approximately two years of college
education and comparable office work experience. Other necessary
office skills, such as typing speed and computer literacy, were
also emphasized. Each tester team’s credentials were nearly
identical with only slight variations so as not to arouse the
employers’ suspicions. The African-American assistants’ credentials
also reflected slightly more work experience than their Caucasian
test partners. The research assistants in reality were two recent
law school graduates, two college graduates, and two students
with two years of college experience.
4. Testing
procedures
The research
assistants were trained and instructed on the process of testing.
They were instructed to follow the Project Coordinator’s directions
and that all communication between the research assistant and
the employer was to be described to the Project Coordinator
as soon after the communication took place as possible. Detailed
written report forms were to be completed by the research assistants
as soon as possible after each activity took place. The assistants
were also instructed not to accept any job offer.
METHODOLOGY
The Project
Coordinator instructed the research assistants as to which employment
agencies they were to apply for work. The research assistants were
instructed to apply for entry-level office positions at each employment
agency.
The research
assistants’ initial contact with the employer was by phone, at which
time they attempted to schedule interviews. The African-American
research assistants scheduled their appointments prior to those
of the Caucasian research assistants. Each pair of research assistants
interviewed at each agency on the same day and within a few hours
of each other.
The research
assistants were trained to complete a detailed reporting form for
each interaction with the employer. In addition, the research assistants
were required to contact the Project Coordinator at each stage of
their interaction with the employer. All telephone calls from employers
to the research assistants were received by the phone bank maintained
at The Testing Project office. The research assistants were also
given references to furnish to the employment agencies which requested
references.
DOCUMENTATION
The research
assistants completed Test Report Forms after each test, documenting
their experiences from the time they arrived at the employment agency
until they left. Follow-up calls, correspondence, and all other
communication between the employer and research assistants were
documented by the research assistants or the Project Coordinator.
Voice mail messages left by employers (including reference checks)
were transcribed by the Project Coordinator; phone calls made by
the research assistants to the employers were documented by the
research assistants.
RESULTS
Seventeen employment
agencies were tested twice using one male pair of research assistants
and one pair of female research assistants. Those agencies where
two tests demonstrated a consistent preference for the Caucasian
applicant or no preference for the Caucasian applicant were not
tested again. The remaining eleven agencies were tested a third
time by a female pair of research assistants. In all, 45 tests were
conducted.
Each test result
was examined by the Testing Coordinator for substantive differences
in treatment and was categorized as either showing preference for
either the Caucasian or African American applicant, or no clear
preference. Generally, a preference was found where a research assistant
was offered a better paying, more responsible, or longer term position
than his/her test partner; was offered a position earlier than his/her
test partner; where an assistant was substantially more encouraged
in the hiring process than his/her test partner; or where lesser
qualification standards were applied to one assistant (i.e. less
tests or reference checks). No clear preference was found where
all favors and procedures were equally applied or where there were
conflicting differences in treatment that did not suggest consistent
treatment of one applicant over another. Summaries of each test,
and how they were characterized are attached as an appendix to this
report.
Examples of
test results showing a preference for the Caucasian applicant follow:
Employment
Agency #8 — Test 1
Caucasian
offered a position immediately; African-American never offered
a position.
The Caucasian
applicant was offered a position which started at $10 per
hour as a temporary position and would then transition into
a permanent position paying $26,000 per year plus benefits.
In contrast, the African-American applicant was told she would
need to polish her MS Excel skills to make her truly employable,
and that she should sign up with other employment agencies,
to ensure she would find work, and was never offered a job.
Employment
Agency #1 — Test 1
African-American
put through "more hoops"; offered lower paying job
than Caucasian test partner.
The Caucasian
applicant was not tested nor was a reference checked. She
was offered a temporary position as a receptionist at the
temporary agency for two to three months with a salary of
$10.50 per hour. The African-American applicant was tested
on various software applications and had one of her references
checked. She was offered a $10 per hour receptionist position.
Employment
Agency #1 — Test 2
African-American
told to "come back next week" — offered lower paying
job.
The Caucasian
applicant dropped by during walk-in hours and was tested and
offered a position paying $9.75 per hour. The African-American
applicant called Employment Agency #1 and was told that the
agency had walk-in hours. The applicant arrived at the agency
during walk-in hours and completed an application, but was
informed that he would need to make an appointment for the
following week to be interviewed and tested. The applicant
returned seven days later and was offered a position making
follow-up calls for an insurance company which paid $8 per
hour.
Overall, the
data may be summarized in the following statistics:
|
SUMMARY
OF RESULTS
(17
Agencies/45 Tests)
|
|
13
out of 17 agencies tested were observed preferring the Caucasian
applicant one or more times.
|
76.5%
of agencies tested
|
|
7 out
of 17 agencies preferred the Caucasian applicant every time.
|
41%
of agencies tested
|
|
8 out
of 17 agencies preferred the Caucasian the majority of the
time.
|
47%
of agencies tested
|
|
25
out of 45 tests showed preference for the Caucasian applicant.
|
56%
of all tests
|
|
8 out
of 45 tests showed a preference for the African- American
applicant.
|
18%
of all tests
|
|
"Net"
Rate of Differential Treatment of African American Applicants
(white preference tests minus black preference tests divided
by total tests).
|
38%
|
This study suggests,
that even in a strong labor market where one would expect to find
little
to no differential treatment, applicants are still treated differently
on the basis of their race. The nature of the differences in treatment,
however, may have been affected by the abundance of positions the
employment agencies were attempting to fill. With one exception,
all the testers were granted an interview and allowed to go through
the entire application process. The differences primarily appeared
in the quality and pay of actual job offers and the time which elapsed
before being offered a position.
CONCLUSION
The test results
show a clear pattern of preference in the employment agency industry
of the San Francisco Bay Area for the Caucasian job applicants,
despite the fact that the African-American applicants had similar
qualifications and interviewing skills. The differences in treatment
cannot be explained by any apparent objective factors other than
the race of the applicants. These results are particularly startling
given the virtual full employment economy of the area.
The rate of
differential treatment found in this study is somewhat lower than
the rate found in previous studies conducted in Chicago and Washington,
D.C. (38% versus 66.7%). This difference may be attributed in whole
or in part to the strong labor market and economy that the San Francisco
Bay Area, as well as the rest of the nation, has been experiencing
in the past year. The earlier studies were conducted in significantly
looser labor markets. Differences in the legal climate and employer
attitudes in the various geographic areas may also account for some
of the difference in results.
The results
lend support to the finding in the earlier studies that the employment
placement industry has a higher incidence of differential treatment
showing preference for Caucasian applicants than the average of
other various industries (24%). There are a few possible reasons
why this industry demonstrates a higher level of differential treatment.
It may be that one of the services offered by employment agencies
to their clients is to screen out "undesirables" as may
be defined by the client. Poor training or personal bias of employment
agency staff who conduct the interviews and have the power to assign
jobs to the applicants may also offer an explanation for the prevalence
of differential treatment in the hiring processes of employment
agencies. The potential role of unconscious bias, particularly when
screening a large number of applicants, cannot be discounted either.
A high level
of disparate treatment of African-American applicants at the entry
level of employment can have profound consequences over the course
of a career, because the initial job, and its salary level often
determine subsequent positions and levels of compensation. That
much of this treatment may be invisible to an African-American applicant
who has no knowledge of how Caucasian applicants are treated, does
not erase the economic consequences of discrimination, which are
magnified over time, as disparate career and salary paths are followed.
This report
is the first major study of the treatment of Caucasian and African
American applicants by employment agencies on the West Coast. It
should serve as a "wake up call" to all employment agencies
to examine their practices, employee training, and treatment of
applicants of all races and other protected characteristics. It
would be important to have follow up testing studies to examine
whether this wake up call has been heeded.
|
