Description
This paper reviews prior research on diversified mentoring relationships and formulates predictions regarding African
Americans’ access to public accounting mentors. Empirical analysis is based on responses from 116 African-
American public accounting employees, compared to 756 Caucasian responses from a prior study. Statistical results
indicate that African-American public accounting employees are less likely to obtain an informal mentor and perceive
greater barriers to obtaining mentors. For two mentoring functions (protection and assistance, and social support),
African-American respondents reported lower levels of support, compared to Caucasian respondents. Also, African-
American respondents indicated stronger intentions to leave public accounting
An examination of African Americans’ access to
public accounting mentors: perceived barriers and intentions
to leave
R.E. Viator *
Area of Accounting, College of Business Administration, Texas Tech University, Lubbock, TX 79409, USA
Abstract
This paper reviews prior research on diversi?ed mentoring relationships and formulates predictions regarding Afri-
can Americans’ access to public accounting mentors. Empirical analysis is based on responses from 116 African-
American public accounting employees, compared to 756 Caucasian responses from a prior study. Statistical results
indicate that African-American public accounting employees are less likely to obtain an informal mentor and perceive
greater barriers to obtaining mentors. For two mentoring functions (protection and assistance, and social support),
African-American respondents reported lower levels of support, compared to Caucasian respondents. Also, African-
American respondents indicated stronger intentions to leave public accounting. #2001 Elsevier Science Ltd. All rights
reserved.
Prior to passage of the Civil Rights Act of 1964,
African Americans faced substantial barriers to
entering the ?eld of public accounting in the Uni-
ted States. These barriers essentially resulted in the
exclusion of African Americans from the practice
of public accounting (Hammond & Streeter,
1994). The Civil Rights Act, which resulted from
the civil rights activism in the United States during
the 1950s and 1960s, created a legal environment
where organizations could be held accountable for
equal employment opportunity. However, racial
prejudice persisted and access to public accounting
employment continued to be denied to African
Americans with outstanding academic back-
grounds (Mitchell, 1969). Hammond (1997) docu-
ments that during the 1970s the public accounting
industry engaged in highly visible e?orts to lower
entry barriers to the profession by o?ering college
scholarships, internships, visiting professorships at
historically black colleges and universities, career
development seminars, and increased emphasis on
recruiting African Americans. Hammond argues
that these highly visible and public e?orts were
signals that the profession was in compliance with
the legal environment and public expectations, but
not re?ective of actual organizational attitudes
and behaviors.
1
0361-3682/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PI I : S0361- 3682( 01) 00005- 8
Accounting, Organizations and Society 26 (2001) 541–561
www.elsevier.com/locate/aos
1
Since the 1970s, public accounting ?rms have achieved
public recognition of organizational change regarding their
working environment for minorities. For example, Ernst &
Young was listed in The Best Companies for Minorities;
Employers Across America Who Recruit, Train, and Promote
Minorities (Graham, 1993).
* Tel.: +1-806-742-1444; fax: +1-806-742-3182.
E-mail address: [email protected]
Review of the electronic WEB sites for each of
the Big 5 public accounting ?rms indicates that
each acknowledges and supports diversity of
employment.
2
Also, most of the ?rms acknowl-
edge having mentoring programs or support an
environment that enables mentoring to take place.
Several of the ?rms have been cited in the business
press as outstanding places to work (Himelstein &
Forest, 1997; Hooks, 1996) and have formal men-
toring programs that target the retention and
promotion of women (Milano, 1999; Quintanilla,
1997).
In spite of these highly visible and well-adver-
tised programs for organizational change, the
question remains regarding whether African
Americans continue to face career development
barriers in the public accounting profession.
Longitudinal research over a 25-year period sug-
gests that while overt racist language has become
unacceptable in corporate America, ‘‘the belief in
stereotypes about people of color has not sig-
ni?cantly decreased’’ (Fernandez, 1999, p. 47).
Fernandez concludes that despite 30 years of
growing awareness about racism, rather than the
reality of racism changing, the language used to
express it has become more subtle: respondents in
the study would assert that people of color are
unquali?ed for a position rather than ‘‘they are
lazy.’’ Diversity awareness programs can o?er
training and education designed to enhance heal-
thy attitudes regarding diversity; however, organi-
zations can only invite individuals, rather than
require individuals, to work on their feelings con-
cerning people of other races (Hayles & Russell,
1997). In an extensive study of successful minority
executives, Thomas and Gabarro (1999, p. 27)
noted that majority members’ feelings of dis-
comfort when mingling with minorities, and their
concerns about personal risk, in?uence whether
majority members seek to include minority mem-
bers in social activities and informational net-
works, or support them when challenged by
others.
The speci?c issue examined in this paper is
whether African Americans are able to obtain
mentoring support in the public accounting
profession, given the social climate in corporate
America that continues to support prejudice-rela-
ted attitudes. Prior research has documented that
informal mentoring relationships are critical in
providing junior members with socialization sup-
port (Heimann & Pittenger, 1996; Koberg, Boss,
Chappell, & Ringer, 1994; Ostro? & Kozlowski,
1993) and career guidance (Dreher & Ash, 1990;
Fagenson, 1989; Kram, 1988; Whitely & Coetsier,
1993). Since development of informal mentoring
relationships relies on individual behavior (i.e.
potential mentors and prote´ ge´ s observe each other
over time as relationships form), African-Amer-
ican employees are likely to have limited access to
informal mentoring relationships if majority
members hold prejudice-related attitudes. In such
an environment, formal mentoring programs are
one alternative for providing African-American
employees with mentoring support. However, for-
mally assigned mentors tend to provide less sup-
port than informal mentors. For example, recent
studies have shown that mentoring relationships
generated by formal organizational programs
provide less career mentoring (Chao, Waltz, &
Gardner, 1992; Ragins & Cotton, 1999) and less
psychosocial support (Fagenson-Eland, Marks, &
Amendola, 1997; Ragins & Cotton, 1999). For
African Americans in public accounting, formal
mentoring programs may simply be another
super?cial o?ering: a highly visible signal that the
profession is in compliance with the legal envir-
onment and public expectations, but not re?ective
of actual attitudes and behaviors exhibited by
individual members.
Since the number of African Americans cur-
rently in public accounting is relatively small
(AICPA, 1997), African Americans must enter
into cross-racial mentoring relationships in order
to obtain mentoring support from in?uential ?rm
members (e.g. senior managers and partners who
exercise power in the ?rm). This paper examines a
theory of diversi?ed mentoring relationships, as
articulated by Ragins (1995, 1997a, b), to for-
mulate hypotheses regarding the expected experi-
ence of African Americans in obtaining career and
2
Electronic WEB sites for the Big 5 are: www.dttus.com,
www.pwcglobal.com, www.ey.com, www.arthurandersen.com,
and www.kpmg.com.
542 R.E. Viator / Accounting, Organizations and Society 26 (2001) 541–561
psychosocial support from potential cross-racial
mentors in public accounting. The paper considers
the mentorship experience of African Americans
in public accounting in regards to four issues: (1)
having an informal mentor; (2) perceived barriers
to obtaining a mentor; (3) the level of support
provided by mentors to African Americans; and 4)
intentions to stay or leave public accounting.
The remainder of the paper contains four sec-
tions. The ?rst section presents an overview of
diversi?ed mentoring relationships and identi?es
?ve testable hypotheses. The other sections discuss
research methods (instruments and participants),
statistical analyses, and ?ndings and implications
for future research.
2. Diversi?ed mentoring relationships
2.1. Barriers to diversi?ed mentoring relationships
Ragins (1997a) identi?es behavioral and per-
ceptual processes unique to diversi?ed mentoring
relationships, including cross-racial mentoring
relationships. These processes have a common
behavioral consequence: they can generate bar-
riers that restrict minority members in obtaining
mentoring support from majority members, espe-
cially potential mentors in positions of power and
in?uence. These processes include: (1) shared
identity and interpersonal comfort; (2) stereotyp-
ing and perceived competence;
3
and (3) visibility
and performance pressures.
4
The ?rst perceptual process is shared identity,
which is the cornerstone for building mentoring
relationships. Junior members tend to view poten-
tial mentors with admiration and respect concern-
ing their competency, power and in?uence. On the
other hand, mentors see potential prote´ ge´ s as per-
sons who would bene?t from advice and counsel,
and whose success would re?ect positively on the
mentor (Hunt & Michael, 1983; Kram, 1988).
Whereas mentors identify prote´ ge´ s as younger
versions of themselves, prote´ ge´ s see mentors as
representative of their future (Ragins, 1997b).
Accounting research has referenced the existence
of shared identity when describing how public
accounting prote´ ge´ s and mentors must ‘‘demon-
strate the right stu?’’ before mentoring begins
(Dirsmith & Covaleski, 1985).
A shared social identity is less likely to occur
when potential mentors and prote´ ge´ s are from
di?erent power groups (Deschamps, 1982), such
as those attributed to racial and ethnic back-
ground. For example, when Caucasian male man-
agers and African-American senior accountants in
the public accounting ?rms have di?erent social
and economic backgrounds they may simply not
perceive the other as an extension of himself/her-
self. Absent a shared identity, potential cross-
racial mentors and prote´ ge´ s have less of a basis for
developing mentoring relationships. Yet it is cri-
tical that minority members enter into diversi?ed
mentoring relationships, since majority members
are the ones who can exert power and in?uence in
behalf of minority members’ careers.
A second perceptual process that can generate
barriers to diversi?ed mentoring relationships is
stereotyping and perceived competence. As pre-
viously noted, potential prote´ ge´ s must possess a
level of competency that demonstrates they would
bene?t from potential mentors’ advice, counsel
and guidance. However, research has shown that
racial and gender-based stereotyping negatively
in?uences majority members’ perception of min-
ority members’ competence (Pettigrew & Martin,
1987). E?ective performance by minority members
is likely to be discounted and attributed to help
from others, rather than internal ability and e?ort
(Greenhaus & Parasuraman, 1993).
Experimental evidence of gender-based stereo-
typing has been found in a public accounting set-
ting where auditors, assessing the prospects for
career success of peers, assigned lower ratings to
?ctitious female cases (Anderson, Johnson, &
Reckers, 1994). However, a second experiment
reported in Johnson, Lowe, and Reckers (2000)
3
Ragins (1997a) discusses stereotyping as a separate per-
ceptual process that can a?ect development of diversi?ed men-
toring relationships. The construct is present in this paper in
regards to its e?ect on perceived competence.
4
Ragins (1997a) discusses workgroup support as a separate
behavioral process that can a?ect development of diversi?ed
mentoring relationships. The construct is presented in this
paper in regards to its e?ect on mentors’ concerns about the
visibility and performance of minority prote´ ge´ s.
R.E. Viator / Accounting, Organizations and Society 26 (2001) 541–561 543
found that o?ce size appears to moderate gender-
based stereotyping: audit seniors at large (small)
o?ces assigned higher (lower) performance ratings
to Caucasian females compared to Caucasian
males. Unexpectedly, African-American females
were consistently assigned higher performance
ratings than Caucasian males. The con?icting
results suggest that gender bias in public account-
ing performance evaluations may be moderated by
demographic variables, such as o?ce size and
race. However, when e?ective performance by
minority members is discounted by evaluators,
then there is a lower probability that competent
minority members will be recognized as potential
prote´ ge´ s, persons whose careers could bene?t from
advice and counsel.
Visibility and performance pressures are a third
perceptual and behavioral process that can gen-
erate barriers to obtaining diversi?ed mentoring
relationships. In providing career support, men-
tors tend to sponsor prote´ ge´ s in obtaining impor-
tant and highly visible assignments (Kram, 1988).
In public accounting, mentors support prote´ ge´ s’
careers by getting them assigned to important,
highly visible client engagements, assisting pro-
te´ ge´ s in managing the perceptions of other ?rm
members, and instructing prote´ ge´ s regarding busi-
ness strategies of the ?rm (Dirsmith & Covaleski,
1985; Dirsmith, Heian, & Covaleski, 1997).
Potential mentors incur additional risk in spon-
soring minority members for important and highly
visible assignments. Given that job outcomes
associated with minority members are likely to be
underestimated due to stereotyping and attribu-
tions (Ragins, 1997a), assigning minority members
to important client engagements simply increases
the likelihood that the engagement outcome will
be viewed negatively, thus attenuating the men-
tor’s status in the ?rm.
Although in?uential mentors have the ability to
assist and protect their prote´ ge´ s, such maneuver-
ing on behalf of minority member prote´ ge´ s risks
signaling that the mentor was a poor judge of
competency. Potential mentors may simply refrain
from supporting minority prote´ ge´ s in highly visi-
ble assignments, but this restraint is likely to be
recognized by minority prote´ ge´ s for what it is: an
absence of true mentoring. Other risks faced by
majority member mentors in selecting minority
prote´ ge´ s for highly visible and important assign-
ments include ostracism by peers (Jones, 1986)
and resentment from subordinates if the mentor is
perceived as helping minority members at the
expense of majority members (Ragins, 1997a).
In summary, the perceptual and behavioral
processes related to shared identity, perceived
competence, and performance pressure are expec-
ted to exert substantial barriers to diversi?ed
mentoring relationships. Both participants in
diversi?ed mentoring relationships face speci?c
obstacles and risks. Minority prote´ ge´ s struggle
with an absence of shared social identity (between
themselves and potential mentors) and must deal
with potential stereotyping of their competence;
majority mentors face threats to their prestige and
power if they place minority prote´ ge´ s in important
and highly visible assignments.
Barriers to diversi?ed mentoring relationships
are meaningful to the practice of public account-
ing. The empirical evidence suggests that public
accounting employees with mentors are more
likely to be attuned to ?rm politics (Dirsmith &
Covaleski, 1985), have lower intentions to leave
the ?rm (Viator & Scandura, 1991), and experi-
ence less role ambiguity in performing their jobs
(Viator, 2001). Thus, gaining access to mentors is
critical for employee advancement in the public
accounting profession. However, the relatively
small number of African Americans in the
accounting profession and the dominance of Cau-
casian males in upper management positions
(AICPA, 1997; Hammond, 1997; Mitchell & Flin-
tall, 1990) demands that African-American
employees enter diversi?ed relationships to
obtaining mentoring support. The perceptual and
behavioral processes described above suggest that
African-American public accounting employees
are likely to encounter barriers to diversi?ed men-
toring relationships. These barriers can be expec-
ted to result in speci?c behavioral outcomes, as
stated below.
H1 Compared to Caucasians, African-Amer-
ican employees in public accounting ?rms
are less likely to have informal mentors.
H2 Compared to Caucasians, African-Amer-
544 R.E. Viator / Accounting, Organizations and Society 26 (2001) 541–561
ican employees in public accounting ?rms
are more likely to perceive barriers to
obtaining a mentor.
The above hypotheses relate to the likelihood of
African-American employees obtaining public
accounting mentors. For those minority employ-
ees who do obtain a mentor, the perceptual and
behavioral processes outlined above are likely to
a?ect the level of mentoring support provided:
both career support and psychosocial support.
Career support provided by mentors includes
sponsoring prote´ ge´ s in obtaining challenging
assignments, promoting prote´ ge´ s’ visibility with
important clients and key managers and partners,
protecting prote´ ge´ s from potential con?icts with
managers and partners, and assisting prote´ ge´ s in
?nishing assignments and meeting deadlines
(Kram, 1988). Because of the potential risks
(described above) that Caucasian mentors assume
when sponsoring the assignment of minority pro-
te´ ge´ s to important clients and highly visible
engagements, Caucasian mentors are likely to
provide African-American prote´ ge´ s with less
career support than that provided to Caucasian
prote´ ge´ s. Unfortunately, the situation is not easily
remedied when African-American prote´ ge´ s seek
out other African Americans as mentors: these
mentors are likely to lack the organizational
power to assist and promote junior members’
careers (Ragins, 1997a). These two processes sug-
gest a third hypothesis.
H3 Compared to Caucasians, African-Amer-
ican prote´ ge´ s in public accounting receive
less career support from mentors.
Psychosocial support provided by mentors
includes providing acceptance and con?rmation of
prote´ ge´ s’ relationships with peers and supervisors,
counseling prote´ ge´ s concerning anxiety about
work and job performance, and providing perso-
nal friendship with informal exchanges, trust, and
con?dence. In addition to active social support,
mentors o?er role modeling, in which their atti-
tudes, values, and behavior provide a model for
prote´ ge´ s to emulate (Kram, 1988). A prime basis
for psychosocial support is the shared social iden-
tity and interpersonal comfort that Ragins (1997a)
identi?es as a potential barrier to obtaining men-
toring support in diversi?ed relationships. In other
words, the basic building block for developing
mentoring relationships is also a barrier for devel-
oping diversi?ed mentoring relationships. Minor-
ity members may respond to this barrier by
forming mentoring relationships with other min-
ority members, who are less likely than majority
members to be at relatively high organizational
levels. Empirical evidence suggests that same-race
mentoring relationships provide more psychoso-
cial support than cross-race relationships (Tho-
mas, 1990). Therefore, it is expected that the level
of psychosocial support reported by African-
American prote´ ge´ s will di?er from Caucasian
prote´ ge´ s, but will be moderated by the presence of
same-race mentors.
H4a African-American prote´ ge´ s with Caucasian
mentors receive less psychosocial support
compared to Caucasian prote´ ge´ s with Cau-
casian mentors.
H4b African-American prote´ ge´ s with African-
American mentors receive more psychosocial
support compared to African-American pro-
te´ ge´ s with Caucasian mentors.
5
2.2. Intentions to leave
Informal mentoring has been linked to organi-
zational socialization and related job outcomes.
Employees with mentors are likely to be more
satis?ed with their jobs and experience less work
alienation, such as social isolation (Koberg et al.,
1994). Organizational newcomers who develop
mentoring relationships tend to learn more
quickly about organizational issues, such as
proper channels of authority and organizational
policies and practices (Ostro? & Kozolowski,
1993). Also, mentoring has been associated with
higher rates of promotion (Dreher & Ash, 1990).
In public accounting settings, mentoring has
been identi?ed as a socialization process that
assists employees in understanding ?rm politics
5
Because of the rarity of Caucasian prote´ ge´ s with African-
American mentors, the complete interaction of prote´ ge´ race
with mentor race cannot be tested and, therefore, is not
addressed in this paper.
R.E. Viator / Accounting, Organizations and Society 26 (2001) 541–561 545
and preparing for promotion (Dirsmith & Cova-
leski, 1985). Public accounting employees without
a mentor have tended to experience higher turn-
over intentions (Viator & Scandura, 1991). If
African Americans face barriers to obtaining
informal public accounting mentors, then the
turnover intentions of African Americans are likely
to be higher than that experienced by Caucasians.
Without organizational processes to facilitate
the socialization of African Americans into public
accounting, it is unlikely that African Americans
will become attached to the organization. A rela-
tional demography perspective (Tsui, Egan, &
O’Reilly, 1992) suggests that the greater the dif-
ference between an individual and others in an
organization, the lower will be the individual’s
attachment to the organization, with lower psy-
chological commitment and higher intentions to
leave the organization. Given that African Amer-
icans make up a relatively small percentage of
public accounting employees (AICPA, 1997), they
are more likely to feel isolated and experience
‘‘being di?erent,’’ and thus, according to rela-
tional demography, less attached to the employing
?rm and more likely to leave.
Prior studies indicate that even high-achieving
African Americans are likely to experience career
roadblocks and become dissatis?ed with their
career opportunities (Jones, 1986). Backmon,
Clark, and Weisenfeld (1997) found that African
Americans, compared to Caucasians in a variety
of employment settings, indicated higher levels of
intentions to leave their current employer and
attributed their dissatisfaction to perceived lack of
advancement opportunities. Greenhaus, Para-
suraman, and Wormley (1990) reported that Afri-
can Americans holding managerial positions were
more likely to experience lower organizational
acceptance and less career satisfaction, compared
to Caucasians in similar positions.
Although African Americans may be assigned
mentors by formal organizational programs, it is
likely that such formally assigned mentors will not
provide as much career support and psychosocial
support compared to informal mentors (Ragins &
Cotton, 1999). Also, formally assigned mentoring
has not been directly linked to lower turnover
intentions in public accounting (Viator, 1999). The
preceding arguments suggest the following
hypothesis.
H5 Compared to Caucasians, African-Amer-
ican employees in public accounting experi-
ence higher turnover intentions.
3. Research methods
3.1. Participants
Potential participants received a cover letter
explaining the purpose of the study, a survey
questionnaire,
6
and a postage-free return envel-
ope. After three weeks, a follow-up reminder
postcard was sent to all potential participants.
In order to investigate the racial diversity issues
raised in this paper, a mailing list was obtained
from the National Association of Black Accoun-
tants (NABA) and consisted of those members
who were CPAs (n=927), or non-CPAs working
at large public accounting ?rms (n=381).
7
Returned surveys totaled 293 out of 1308 mailed,
representing a 22.4% response rate.
8
The relatively
low response rate from NABA members may be
6
Copies of the survey instrument are available from the
author.
7
Because of their relatively small number, all African
Americans employed by large public accounting ?rms were
included in this survey.
8
Following Oppenheim’s (1966) recommendation, several
statistical tests were performed to identify di?erences between
early and late respondents from the NABA mailing list. All
responses were coded by return-date. Since the total number of
respondents was relatively small, respondent groups were
formed by identifying a drop in the per-date response rate near
the median of the distribution, resulting in 59 early respondents
and 50 late respondents. Based on Chi-squared tests, no di?er-
ences were identi?ed regarding respondent organizational level,
gender, ?rm size, ‘‘having an informal mentor,’’ and ‘‘having a
formal mentor’’ (P>0.10). Based on regression analysis, no
di?erences were found for levels of perceived barriers (lack of
access, and lack of willingness), intentions to leave, and mea-
sures of psychosocial support (P>0.10). Only the beta coe?-
cient for career-related mentoring was marginally signi?cant
(=+1.23, P>0.07); however, it does not a?ect interpretation
of the results. See the non-signi?cant coe?cients for African-
American seniors and the di?erence between African-American
managers and Caucasian managers in the regression analysis of
career-related mentoring reported in Table 3.
546 R.E. Viator / Accounting, Organizations and Society 26 (2001) 541–561
attributed to their receiving other survey requests
in recent years (as indicated in participant feed-
back), resulting in limited participation in the
current study.
9
The response rate is similar to
Dreher and Cox (1996) whose survey also investi-
gated minority issues and mentoring relationships.
Some responses from NABA members are
omitted from the analyses reported in this paper.
First, a substantial number of responses were from
CPAs employed outside of public accounting
(n=156).
10
Second, other surveys were from
partners/directors at public accounting ?rms
(n=19) and support sta? employees (n=9). The
9
For the 493 NABA members working at ‘‘Big 5’’ public
accounting ?rms, 137 of the surveys were returned, represent-
ing a response rate of 27.7%, which was similar to the overall
NABA response rate.
10
Respondents working outside of public accounting are
not included in this study for two reasons. First, the compar-
ison data from the AICPA mailing list included only 25 Cau-
casian respondents that indicated current employment outside
of public accounting. Second, role pressure and turnover
intentions (one of the dependent variables examined in this
paper) are expected to di?er between public accounting and
non-public accounting practices.
Table 1
Participant demographics by organizational level
Participant organizational level Total participants
Sta? accountant Senior accountant Manager
NABA mailing list: African-Americans
Male 13 18 18 49
Female 24 28 8 60
AICPA mailing list: African-Americans
Male – 2 0 2
Female – 3 2 5
AIPCA mailing list: Caucasian
Male – 107 238 345
Female – 182 229 411
Total participants 37 340 495 872
African-American participants
Firm a?liation
‘‘Big 5’’ 100.0% 98.0% 75.0% 93.1%
Other National 0.0% 2.0% 3.6% 1.7%
Regional 0.0% 0.0% 21.4% 5.2%
% with informal mentor 72.9% 66.7% 60.7% 67.2%
% with formal mentor 83.8% 76.5% 39.3% 69.8%
Caucasian participants
Firm a?liation
‘‘Big 5’’ – 92.4% 89.7% 90.7%
Other National – 6.6% 6.0% 6.2%
Regional – 1.0% 4.3% 3.1%
% with informal mentor – 78.9% 83.9% 82.0%
% with formal mentor – 55.4% 35.1% 42.9%
R.E. Viator / Accounting, Organizations and Society 26 (2001) 541–561 547
demographic background for the remaining 109
participants employed in the public accounting is
presented in Table 1.
Responses from a previous study, in which the
participants were primarily Caucasian, served as
comparison data.
11
This mailing list was obtained
from the American Institute of CPAs (AICPA)
and consisted of members identi?ed initially as
employees of large public accounting ?rms. The
sample of 3000 members was strati?ed based on
gender, with an equal number of females and
males selected. Returned surveys included 902
participants, representing a 30% response rate.
12
Some responses were deleted for the following
reasons: incomplete surveys (n=13), not employed
in public accounting (n=25), partners/directors in
large ?rms (n=44), and members of racial/ethnic
groups other than Caucasian or African American
(n=54) or Caucasian prote´ ge´ s with cross-racial
mentors (n=3). Table 1 presents the demographic
background of the remaining 763 AICPA partici-
pants included in this study as comparison data.
Participants were required to identify: (1) whe-
ther they had current/recent mentors, and (2)
whether those mentors had been assigned to them
by a formal program at the ?rm, or were informal
mentoring relationships. As in other studies con-
sidering the e?ects of formal and informal men-
toring relationships, prote´ ge´ s who identi?ed
concurrent formal and informal mentoring rela-
tionships were instructed to consider their for-
mally assigned mentor when responding to
questionnaire items regarding support provided by
a mentor (Chao et al., 1992, Ragins & Cotton,
1999).
3.2. Measures
All scale items were scored on a ?ve-point Likert
scale from ‘‘strongly disagree’’ to ‘‘strongly
agree.’’ Pretests were conducted to ensure that
each scale item was applicable and re?ected
wording commonly used in the profession. In
some cases, minor wording changes were made
based on feedback received from 15 CPAs in four
o?ces of international public accounting ?rms.
Appendix A lists relevant questionnaire items,
along with their means, standard deviations, and
factor loadings.
3.2.1. Perceived barriers
The items assessing perceived barriers to men-
toring were adopted from Ragins and Cotton
(1991) and previously reported in Viator (1999).
Ragins and Cotton used principal component fac-
tor analysis to identify ?ve separate dimensions
for perceived barriers. Due to survey constraints,
13
this study adopted two of those factors: (1) lack of
access to potential mentors; and (2) lack of will-
ingness by potential mentors. In the Ragins and
Cotton study, these two factors demonstrated
strong construct validity in their predicted asso-
ciation with employee job rank, employee experi-
ence, and employee gender.
14
In the Viator study,
the factors had some association with mentor-type
and employee organizational level, but not
employee gender. Each factor was represented by
two questionnaire items, as presented in the
Appendix. The Cronbach Alpha for each factor
was greater than 0.80, suggesting relatively high
reliability, given 0.70 as a reasonable guideline
(Nunnally, 1978).
11
The two surveys were conducted at two di?erent points in
time, approximately one and half years apart. Detailed analyses
of data from the previous study are reported in Viator (1999)
and Viator (2001). Questionnaire items for the two surveys
were identical; however, survey instruments sent to NABA
members included the role reference ‘‘your current ?rm or
business’’ rather than simply ‘‘your current ?rm’’ in order to
accommodate responses from participants outside of public
accounting.
12
Kerlinger (1986) indicates that the expected response rate
from mail surveys is 20–40%.
13
As part of a broader study, the survey included some non-
mentoring questionnaire items. Response rate concerns limited
the length of the survey instrument and restricted the total
number of questions adopted.
14
One factor not selected lacked any association with pre-
dictor variables. Another factor, labeled ‘‘misinterpretation’’ by
Ragins and Cotton (1991), asked questions regarding potential
sexual innuendo. Concerns about sexual innuendo are likely to
generate some barriers to cross-gender mentoring relationships
(Ragins & McFarlin, 1990). These issues are not addressed in
the current study.
548 R.E. Viator / Accounting, Organizations and Society 26 (2001) 541–561
3.2.2. Mentoring functions
The mentoring function scale included 16 items
measuring two career support functions (career-
related mentoring, and protection and assistance)
and two psychosocial support functions (social
support and role modeling). These items were
adopted from scales used in previous mentoring
research (Chao et al., 1992; Dreher & Ash, 1990;
Noe, 1988; Scandura & Viator, 1994; Tepper,
1995; Turban & Dougherty, 1994; Viator, 2001).
Factor analysis, using an orthogonal rotation and
an eigenvalue-greater-than-1.0 criterion cuto?,
con?rmed the existence of four separate factors:
career-related, protection and assistance, social
support, and role modeling. As shown in the
Appendix, all items, except two social support
items, loaded on the expected factor, with factor
loadings greater than 0.50. After dropping the two
social support items, each factor had a Cronbach
alpha greater than 0.70, indicating acceptable
reliability.
3.2.3. Intentions to leave
Accounting studies have used a variety of mea-
sures of employee turnover intentions (Aranya &
Ferris, 1984; Aranya, Lachman, & Amernic, 1982;
Collins & Killough, 1992; Dillard & Ferris, 1979;
Harrell, Chewning, & Taylor, 1986; Harrell &
Stahl, 1984; Senatra, 1980). In the current study,
the turnover intentions scale was constructed from
two questionnaire items adopted from Collins and
Killough. The scale focused on ‘‘thinking about
leaving the ?rm’’ and ‘‘the probability of looking
for another job.’’ There is general support that
turnover intentions provide a viable indication of
subsequent actual turnovers (Steel & Ovalle,
1984). The Cronbach alpha was 0.85.
3.2.4. Demographic variables
A limitation of the current study is that partici-
pant race is not fully crossed with participant
organizational level: some NABA participants
were at the sta? level, but all AICPA participants
were at the senior accountant level or higher. In
order to facilitate data analysis presentation and
comparison of coe?cients, four dummy variables
are used to represent race and organizational level:
African-American Sta?, African-American Seniors,
African-American Managers, and Caucasian Man-
agers, (where Caucasian senior accountant is the
comparison category).
15
Other demographic variables likely to covary
with mentoring practices (Dreher & Ash, 1990;
Ragins & Cotton, 1991, 1999; Ragins & McFarlin,
1990) are included in the regression models.
Dummy variables identifying participant demo-
graphics include gender (0=male, and 1=female)
and national ?rm, regional ?rm (where ‘‘Big 5’’
international ?rm is the comparison category).
The regression analyses of barriers and intentions
to leave include dummy variables indicating the
presence or absence of informal mentors and formal
mentors.
Mentor demographic variables include dummy
variables identifying mentor race (0=Caucasian,
and 1=African American), mentor gender
(0=male, and 1=female), partner-mentor, senior
manager-mentor, other-mentor (where manager-
mentor is the comparison category), and the con-
tinuous variable length-of-mentoring-relationship.
The regression analyses of mentoring functions
include a dummy variable identifying formal men-
tor (0=informal, and 1=formal), as well as the
interaction of formal mentor with each racial
group. Only signi?cant interaction terms (e.g. the
interaction of formal mentor with African-Amer-
ican seniors) are reported in Section 3 below and
in Table 3. Prote´ ge´ s with concurrent formal and
informal mentors evaluated the level of mentoring
received from their formal mentor (as in Chao et
al., 1992; Ragins & Cotton, 1999).
4. Statistical analyses and results
Statistical test procedures included both chi-
square tests of association (for testing H1) and
linear regression analyses (for testing H2 through
H5). Speci?c tests of the directional hypotheses
are reported as one-tailed tests. Statistical results
15
Due to the relatively small number of African Americans
at the senior manager level (n=6), the variable manager com-
bines participants who identi?ed themselves as managers or
senior managers. Subsequent testing with an additional vari-
able for senior manager provided no substantive changes in the
results reported in this paper.
R.E. Viator / Accounting, Organizations and Society 26 (2001) 541–561 549
of the signi?cance of other demographic variables
are reported as non-directional two-tailed tests.
H1 stated that African Americans in public
accounting are less likely to have informal men-
tors. As previously noted, respondents indicated
whether they currently had an informal or formal
mentor, or both. A chi-square test performed on
the complete data set (n=872) was statistically
signi?cant (=13.73, P
This paper reviews prior research on diversified mentoring relationships and formulates predictions regarding African
Americans’ access to public accounting mentors. Empirical analysis is based on responses from 116 African-
American public accounting employees, compared to 756 Caucasian responses from a prior study. Statistical results
indicate that African-American public accounting employees are less likely to obtain an informal mentor and perceive
greater barriers to obtaining mentors. For two mentoring functions (protection and assistance, and social support),
African-American respondents reported lower levels of support, compared to Caucasian respondents. Also, African-
American respondents indicated stronger intentions to leave public accounting
An examination of African Americans’ access to
public accounting mentors: perceived barriers and intentions
to leave
R.E. Viator *
Area of Accounting, College of Business Administration, Texas Tech University, Lubbock, TX 79409, USA
Abstract
This paper reviews prior research on diversi?ed mentoring relationships and formulates predictions regarding Afri-
can Americans’ access to public accounting mentors. Empirical analysis is based on responses from 116 African-
American public accounting employees, compared to 756 Caucasian responses from a prior study. Statistical results
indicate that African-American public accounting employees are less likely to obtain an informal mentor and perceive
greater barriers to obtaining mentors. For two mentoring functions (protection and assistance, and social support),
African-American respondents reported lower levels of support, compared to Caucasian respondents. Also, African-
American respondents indicated stronger intentions to leave public accounting. #2001 Elsevier Science Ltd. All rights
reserved.
Prior to passage of the Civil Rights Act of 1964,
African Americans faced substantial barriers to
entering the ?eld of public accounting in the Uni-
ted States. These barriers essentially resulted in the
exclusion of African Americans from the practice
of public accounting (Hammond & Streeter,
1994). The Civil Rights Act, which resulted from
the civil rights activism in the United States during
the 1950s and 1960s, created a legal environment
where organizations could be held accountable for
equal employment opportunity. However, racial
prejudice persisted and access to public accounting
employment continued to be denied to African
Americans with outstanding academic back-
grounds (Mitchell, 1969). Hammond (1997) docu-
ments that during the 1970s the public accounting
industry engaged in highly visible e?orts to lower
entry barriers to the profession by o?ering college
scholarships, internships, visiting professorships at
historically black colleges and universities, career
development seminars, and increased emphasis on
recruiting African Americans. Hammond argues
that these highly visible and public e?orts were
signals that the profession was in compliance with
the legal environment and public expectations, but
not re?ective of actual organizational attitudes
and behaviors.
1
0361-3682/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PI I : S0361- 3682( 01) 00005- 8
Accounting, Organizations and Society 26 (2001) 541–561
www.elsevier.com/locate/aos
1
Since the 1970s, public accounting ?rms have achieved
public recognition of organizational change regarding their
working environment for minorities. For example, Ernst &
Young was listed in The Best Companies for Minorities;
Employers Across America Who Recruit, Train, and Promote
Minorities (Graham, 1993).
* Tel.: +1-806-742-1444; fax: +1-806-742-3182.
E-mail address: [email protected]
Review of the electronic WEB sites for each of
the Big 5 public accounting ?rms indicates that
each acknowledges and supports diversity of
employment.
2
Also, most of the ?rms acknowl-
edge having mentoring programs or support an
environment that enables mentoring to take place.
Several of the ?rms have been cited in the business
press as outstanding places to work (Himelstein &
Forest, 1997; Hooks, 1996) and have formal men-
toring programs that target the retention and
promotion of women (Milano, 1999; Quintanilla,
1997).
In spite of these highly visible and well-adver-
tised programs for organizational change, the
question remains regarding whether African
Americans continue to face career development
barriers in the public accounting profession.
Longitudinal research over a 25-year period sug-
gests that while overt racist language has become
unacceptable in corporate America, ‘‘the belief in
stereotypes about people of color has not sig-
ni?cantly decreased’’ (Fernandez, 1999, p. 47).
Fernandez concludes that despite 30 years of
growing awareness about racism, rather than the
reality of racism changing, the language used to
express it has become more subtle: respondents in
the study would assert that people of color are
unquali?ed for a position rather than ‘‘they are
lazy.’’ Diversity awareness programs can o?er
training and education designed to enhance heal-
thy attitudes regarding diversity; however, organi-
zations can only invite individuals, rather than
require individuals, to work on their feelings con-
cerning people of other races (Hayles & Russell,
1997). In an extensive study of successful minority
executives, Thomas and Gabarro (1999, p. 27)
noted that majority members’ feelings of dis-
comfort when mingling with minorities, and their
concerns about personal risk, in?uence whether
majority members seek to include minority mem-
bers in social activities and informational net-
works, or support them when challenged by
others.
The speci?c issue examined in this paper is
whether African Americans are able to obtain
mentoring support in the public accounting
profession, given the social climate in corporate
America that continues to support prejudice-rela-
ted attitudes. Prior research has documented that
informal mentoring relationships are critical in
providing junior members with socialization sup-
port (Heimann & Pittenger, 1996; Koberg, Boss,
Chappell, & Ringer, 1994; Ostro? & Kozlowski,
1993) and career guidance (Dreher & Ash, 1990;
Fagenson, 1989; Kram, 1988; Whitely & Coetsier,
1993). Since development of informal mentoring
relationships relies on individual behavior (i.e.
potential mentors and prote´ ge´ s observe each other
over time as relationships form), African-Amer-
ican employees are likely to have limited access to
informal mentoring relationships if majority
members hold prejudice-related attitudes. In such
an environment, formal mentoring programs are
one alternative for providing African-American
employees with mentoring support. However, for-
mally assigned mentors tend to provide less sup-
port than informal mentors. For example, recent
studies have shown that mentoring relationships
generated by formal organizational programs
provide less career mentoring (Chao, Waltz, &
Gardner, 1992; Ragins & Cotton, 1999) and less
psychosocial support (Fagenson-Eland, Marks, &
Amendola, 1997; Ragins & Cotton, 1999). For
African Americans in public accounting, formal
mentoring programs may simply be another
super?cial o?ering: a highly visible signal that the
profession is in compliance with the legal envir-
onment and public expectations, but not re?ective
of actual attitudes and behaviors exhibited by
individual members.
Since the number of African Americans cur-
rently in public accounting is relatively small
(AICPA, 1997), African Americans must enter
into cross-racial mentoring relationships in order
to obtain mentoring support from in?uential ?rm
members (e.g. senior managers and partners who
exercise power in the ?rm). This paper examines a
theory of diversi?ed mentoring relationships, as
articulated by Ragins (1995, 1997a, b), to for-
mulate hypotheses regarding the expected experi-
ence of African Americans in obtaining career and
2
Electronic WEB sites for the Big 5 are: www.dttus.com,
www.pwcglobal.com, www.ey.com, www.arthurandersen.com,
and www.kpmg.com.
542 R.E. Viator / Accounting, Organizations and Society 26 (2001) 541–561
psychosocial support from potential cross-racial
mentors in public accounting. The paper considers
the mentorship experience of African Americans
in public accounting in regards to four issues: (1)
having an informal mentor; (2) perceived barriers
to obtaining a mentor; (3) the level of support
provided by mentors to African Americans; and 4)
intentions to stay or leave public accounting.
The remainder of the paper contains four sec-
tions. The ?rst section presents an overview of
diversi?ed mentoring relationships and identi?es
?ve testable hypotheses. The other sections discuss
research methods (instruments and participants),
statistical analyses, and ?ndings and implications
for future research.
2. Diversi?ed mentoring relationships
2.1. Barriers to diversi?ed mentoring relationships
Ragins (1997a) identi?es behavioral and per-
ceptual processes unique to diversi?ed mentoring
relationships, including cross-racial mentoring
relationships. These processes have a common
behavioral consequence: they can generate bar-
riers that restrict minority members in obtaining
mentoring support from majority members, espe-
cially potential mentors in positions of power and
in?uence. These processes include: (1) shared
identity and interpersonal comfort; (2) stereotyp-
ing and perceived competence;
3
and (3) visibility
and performance pressures.
4
The ?rst perceptual process is shared identity,
which is the cornerstone for building mentoring
relationships. Junior members tend to view poten-
tial mentors with admiration and respect concern-
ing their competency, power and in?uence. On the
other hand, mentors see potential prote´ ge´ s as per-
sons who would bene?t from advice and counsel,
and whose success would re?ect positively on the
mentor (Hunt & Michael, 1983; Kram, 1988).
Whereas mentors identify prote´ ge´ s as younger
versions of themselves, prote´ ge´ s see mentors as
representative of their future (Ragins, 1997b).
Accounting research has referenced the existence
of shared identity when describing how public
accounting prote´ ge´ s and mentors must ‘‘demon-
strate the right stu?’’ before mentoring begins
(Dirsmith & Covaleski, 1985).
A shared social identity is less likely to occur
when potential mentors and prote´ ge´ s are from
di?erent power groups (Deschamps, 1982), such
as those attributed to racial and ethnic back-
ground. For example, when Caucasian male man-
agers and African-American senior accountants in
the public accounting ?rms have di?erent social
and economic backgrounds they may simply not
perceive the other as an extension of himself/her-
self. Absent a shared identity, potential cross-
racial mentors and prote´ ge´ s have less of a basis for
developing mentoring relationships. Yet it is cri-
tical that minority members enter into diversi?ed
mentoring relationships, since majority members
are the ones who can exert power and in?uence in
behalf of minority members’ careers.
A second perceptual process that can generate
barriers to diversi?ed mentoring relationships is
stereotyping and perceived competence. As pre-
viously noted, potential prote´ ge´ s must possess a
level of competency that demonstrates they would
bene?t from potential mentors’ advice, counsel
and guidance. However, research has shown that
racial and gender-based stereotyping negatively
in?uences majority members’ perception of min-
ority members’ competence (Pettigrew & Martin,
1987). E?ective performance by minority members
is likely to be discounted and attributed to help
from others, rather than internal ability and e?ort
(Greenhaus & Parasuraman, 1993).
Experimental evidence of gender-based stereo-
typing has been found in a public accounting set-
ting where auditors, assessing the prospects for
career success of peers, assigned lower ratings to
?ctitious female cases (Anderson, Johnson, &
Reckers, 1994). However, a second experiment
reported in Johnson, Lowe, and Reckers (2000)
3
Ragins (1997a) discusses stereotyping as a separate per-
ceptual process that can a?ect development of diversi?ed men-
toring relationships. The construct is present in this paper in
regards to its e?ect on perceived competence.
4
Ragins (1997a) discusses workgroup support as a separate
behavioral process that can a?ect development of diversi?ed
mentoring relationships. The construct is presented in this
paper in regards to its e?ect on mentors’ concerns about the
visibility and performance of minority prote´ ge´ s.
R.E. Viator / Accounting, Organizations and Society 26 (2001) 541–561 543
found that o?ce size appears to moderate gender-
based stereotyping: audit seniors at large (small)
o?ces assigned higher (lower) performance ratings
to Caucasian females compared to Caucasian
males. Unexpectedly, African-American females
were consistently assigned higher performance
ratings than Caucasian males. The con?icting
results suggest that gender bias in public account-
ing performance evaluations may be moderated by
demographic variables, such as o?ce size and
race. However, when e?ective performance by
minority members is discounted by evaluators,
then there is a lower probability that competent
minority members will be recognized as potential
prote´ ge´ s, persons whose careers could bene?t from
advice and counsel.
Visibility and performance pressures are a third
perceptual and behavioral process that can gen-
erate barriers to obtaining diversi?ed mentoring
relationships. In providing career support, men-
tors tend to sponsor prote´ ge´ s in obtaining impor-
tant and highly visible assignments (Kram, 1988).
In public accounting, mentors support prote´ ge´ s’
careers by getting them assigned to important,
highly visible client engagements, assisting pro-
te´ ge´ s in managing the perceptions of other ?rm
members, and instructing prote´ ge´ s regarding busi-
ness strategies of the ?rm (Dirsmith & Covaleski,
1985; Dirsmith, Heian, & Covaleski, 1997).
Potential mentors incur additional risk in spon-
soring minority members for important and highly
visible assignments. Given that job outcomes
associated with minority members are likely to be
underestimated due to stereotyping and attribu-
tions (Ragins, 1997a), assigning minority members
to important client engagements simply increases
the likelihood that the engagement outcome will
be viewed negatively, thus attenuating the men-
tor’s status in the ?rm.
Although in?uential mentors have the ability to
assist and protect their prote´ ge´ s, such maneuver-
ing on behalf of minority member prote´ ge´ s risks
signaling that the mentor was a poor judge of
competency. Potential mentors may simply refrain
from supporting minority prote´ ge´ s in highly visi-
ble assignments, but this restraint is likely to be
recognized by minority prote´ ge´ s for what it is: an
absence of true mentoring. Other risks faced by
majority member mentors in selecting minority
prote´ ge´ s for highly visible and important assign-
ments include ostracism by peers (Jones, 1986)
and resentment from subordinates if the mentor is
perceived as helping minority members at the
expense of majority members (Ragins, 1997a).
In summary, the perceptual and behavioral
processes related to shared identity, perceived
competence, and performance pressure are expec-
ted to exert substantial barriers to diversi?ed
mentoring relationships. Both participants in
diversi?ed mentoring relationships face speci?c
obstacles and risks. Minority prote´ ge´ s struggle
with an absence of shared social identity (between
themselves and potential mentors) and must deal
with potential stereotyping of their competence;
majority mentors face threats to their prestige and
power if they place minority prote´ ge´ s in important
and highly visible assignments.
Barriers to diversi?ed mentoring relationships
are meaningful to the practice of public account-
ing. The empirical evidence suggests that public
accounting employees with mentors are more
likely to be attuned to ?rm politics (Dirsmith &
Covaleski, 1985), have lower intentions to leave
the ?rm (Viator & Scandura, 1991), and experi-
ence less role ambiguity in performing their jobs
(Viator, 2001). Thus, gaining access to mentors is
critical for employee advancement in the public
accounting profession. However, the relatively
small number of African Americans in the
accounting profession and the dominance of Cau-
casian males in upper management positions
(AICPA, 1997; Hammond, 1997; Mitchell & Flin-
tall, 1990) demands that African-American
employees enter diversi?ed relationships to
obtaining mentoring support. The perceptual and
behavioral processes described above suggest that
African-American public accounting employees
are likely to encounter barriers to diversi?ed men-
toring relationships. These barriers can be expec-
ted to result in speci?c behavioral outcomes, as
stated below.
H1 Compared to Caucasians, African-Amer-
ican employees in public accounting ?rms
are less likely to have informal mentors.
H2 Compared to Caucasians, African-Amer-
544 R.E. Viator / Accounting, Organizations and Society 26 (2001) 541–561
ican employees in public accounting ?rms
are more likely to perceive barriers to
obtaining a mentor.
The above hypotheses relate to the likelihood of
African-American employees obtaining public
accounting mentors. For those minority employ-
ees who do obtain a mentor, the perceptual and
behavioral processes outlined above are likely to
a?ect the level of mentoring support provided:
both career support and psychosocial support.
Career support provided by mentors includes
sponsoring prote´ ge´ s in obtaining challenging
assignments, promoting prote´ ge´ s’ visibility with
important clients and key managers and partners,
protecting prote´ ge´ s from potential con?icts with
managers and partners, and assisting prote´ ge´ s in
?nishing assignments and meeting deadlines
(Kram, 1988). Because of the potential risks
(described above) that Caucasian mentors assume
when sponsoring the assignment of minority pro-
te´ ge´ s to important clients and highly visible
engagements, Caucasian mentors are likely to
provide African-American prote´ ge´ s with less
career support than that provided to Caucasian
prote´ ge´ s. Unfortunately, the situation is not easily
remedied when African-American prote´ ge´ s seek
out other African Americans as mentors: these
mentors are likely to lack the organizational
power to assist and promote junior members’
careers (Ragins, 1997a). These two processes sug-
gest a third hypothesis.
H3 Compared to Caucasians, African-Amer-
ican prote´ ge´ s in public accounting receive
less career support from mentors.
Psychosocial support provided by mentors
includes providing acceptance and con?rmation of
prote´ ge´ s’ relationships with peers and supervisors,
counseling prote´ ge´ s concerning anxiety about
work and job performance, and providing perso-
nal friendship with informal exchanges, trust, and
con?dence. In addition to active social support,
mentors o?er role modeling, in which their atti-
tudes, values, and behavior provide a model for
prote´ ge´ s to emulate (Kram, 1988). A prime basis
for psychosocial support is the shared social iden-
tity and interpersonal comfort that Ragins (1997a)
identi?es as a potential barrier to obtaining men-
toring support in diversi?ed relationships. In other
words, the basic building block for developing
mentoring relationships is also a barrier for devel-
oping diversi?ed mentoring relationships. Minor-
ity members may respond to this barrier by
forming mentoring relationships with other min-
ority members, who are less likely than majority
members to be at relatively high organizational
levels. Empirical evidence suggests that same-race
mentoring relationships provide more psychoso-
cial support than cross-race relationships (Tho-
mas, 1990). Therefore, it is expected that the level
of psychosocial support reported by African-
American prote´ ge´ s will di?er from Caucasian
prote´ ge´ s, but will be moderated by the presence of
same-race mentors.
H4a African-American prote´ ge´ s with Caucasian
mentors receive less psychosocial support
compared to Caucasian prote´ ge´ s with Cau-
casian mentors.
H4b African-American prote´ ge´ s with African-
American mentors receive more psychosocial
support compared to African-American pro-
te´ ge´ s with Caucasian mentors.
5
2.2. Intentions to leave
Informal mentoring has been linked to organi-
zational socialization and related job outcomes.
Employees with mentors are likely to be more
satis?ed with their jobs and experience less work
alienation, such as social isolation (Koberg et al.,
1994). Organizational newcomers who develop
mentoring relationships tend to learn more
quickly about organizational issues, such as
proper channels of authority and organizational
policies and practices (Ostro? & Kozolowski,
1993). Also, mentoring has been associated with
higher rates of promotion (Dreher & Ash, 1990).
In public accounting settings, mentoring has
been identi?ed as a socialization process that
assists employees in understanding ?rm politics
5
Because of the rarity of Caucasian prote´ ge´ s with African-
American mentors, the complete interaction of prote´ ge´ race
with mentor race cannot be tested and, therefore, is not
addressed in this paper.
R.E. Viator / Accounting, Organizations and Society 26 (2001) 541–561 545
and preparing for promotion (Dirsmith & Cova-
leski, 1985). Public accounting employees without
a mentor have tended to experience higher turn-
over intentions (Viator & Scandura, 1991). If
African Americans face barriers to obtaining
informal public accounting mentors, then the
turnover intentions of African Americans are likely
to be higher than that experienced by Caucasians.
Without organizational processes to facilitate
the socialization of African Americans into public
accounting, it is unlikely that African Americans
will become attached to the organization. A rela-
tional demography perspective (Tsui, Egan, &
O’Reilly, 1992) suggests that the greater the dif-
ference between an individual and others in an
organization, the lower will be the individual’s
attachment to the organization, with lower psy-
chological commitment and higher intentions to
leave the organization. Given that African Amer-
icans make up a relatively small percentage of
public accounting employees (AICPA, 1997), they
are more likely to feel isolated and experience
‘‘being di?erent,’’ and thus, according to rela-
tional demography, less attached to the employing
?rm and more likely to leave.
Prior studies indicate that even high-achieving
African Americans are likely to experience career
roadblocks and become dissatis?ed with their
career opportunities (Jones, 1986). Backmon,
Clark, and Weisenfeld (1997) found that African
Americans, compared to Caucasians in a variety
of employment settings, indicated higher levels of
intentions to leave their current employer and
attributed their dissatisfaction to perceived lack of
advancement opportunities. Greenhaus, Para-
suraman, and Wormley (1990) reported that Afri-
can Americans holding managerial positions were
more likely to experience lower organizational
acceptance and less career satisfaction, compared
to Caucasians in similar positions.
Although African Americans may be assigned
mentors by formal organizational programs, it is
likely that such formally assigned mentors will not
provide as much career support and psychosocial
support compared to informal mentors (Ragins &
Cotton, 1999). Also, formally assigned mentoring
has not been directly linked to lower turnover
intentions in public accounting (Viator, 1999). The
preceding arguments suggest the following
hypothesis.
H5 Compared to Caucasians, African-Amer-
ican employees in public accounting experi-
ence higher turnover intentions.
3. Research methods
3.1. Participants
Potential participants received a cover letter
explaining the purpose of the study, a survey
questionnaire,
6
and a postage-free return envel-
ope. After three weeks, a follow-up reminder
postcard was sent to all potential participants.
In order to investigate the racial diversity issues
raised in this paper, a mailing list was obtained
from the National Association of Black Accoun-
tants (NABA) and consisted of those members
who were CPAs (n=927), or non-CPAs working
at large public accounting ?rms (n=381).
7
Returned surveys totaled 293 out of 1308 mailed,
representing a 22.4% response rate.
8
The relatively
low response rate from NABA members may be
6
Copies of the survey instrument are available from the
author.
7
Because of their relatively small number, all African
Americans employed by large public accounting ?rms were
included in this survey.
8
Following Oppenheim’s (1966) recommendation, several
statistical tests were performed to identify di?erences between
early and late respondents from the NABA mailing list. All
responses were coded by return-date. Since the total number of
respondents was relatively small, respondent groups were
formed by identifying a drop in the per-date response rate near
the median of the distribution, resulting in 59 early respondents
and 50 late respondents. Based on Chi-squared tests, no di?er-
ences were identi?ed regarding respondent organizational level,
gender, ?rm size, ‘‘having an informal mentor,’’ and ‘‘having a
formal mentor’’ (P>0.10). Based on regression analysis, no
di?erences were found for levels of perceived barriers (lack of
access, and lack of willingness), intentions to leave, and mea-
sures of psychosocial support (P>0.10). Only the beta coe?-
cient for career-related mentoring was marginally signi?cant
(=+1.23, P>0.07); however, it does not a?ect interpretation
of the results. See the non-signi?cant coe?cients for African-
American seniors and the di?erence between African-American
managers and Caucasian managers in the regression analysis of
career-related mentoring reported in Table 3.
546 R.E. Viator / Accounting, Organizations and Society 26 (2001) 541–561
attributed to their receiving other survey requests
in recent years (as indicated in participant feed-
back), resulting in limited participation in the
current study.
9
The response rate is similar to
Dreher and Cox (1996) whose survey also investi-
gated minority issues and mentoring relationships.
Some responses from NABA members are
omitted from the analyses reported in this paper.
First, a substantial number of responses were from
CPAs employed outside of public accounting
(n=156).
10
Second, other surveys were from
partners/directors at public accounting ?rms
(n=19) and support sta? employees (n=9). The
9
For the 493 NABA members working at ‘‘Big 5’’ public
accounting ?rms, 137 of the surveys were returned, represent-
ing a response rate of 27.7%, which was similar to the overall
NABA response rate.
10
Respondents working outside of public accounting are
not included in this study for two reasons. First, the compar-
ison data from the AICPA mailing list included only 25 Cau-
casian respondents that indicated current employment outside
of public accounting. Second, role pressure and turnover
intentions (one of the dependent variables examined in this
paper) are expected to di?er between public accounting and
non-public accounting practices.
Table 1
Participant demographics by organizational level
Participant organizational level Total participants
Sta? accountant Senior accountant Manager
NABA mailing list: African-Americans
Male 13 18 18 49
Female 24 28 8 60
AICPA mailing list: African-Americans
Male – 2 0 2
Female – 3 2 5
AIPCA mailing list: Caucasian
Male – 107 238 345
Female – 182 229 411
Total participants 37 340 495 872
African-American participants
Firm a?liation
‘‘Big 5’’ 100.0% 98.0% 75.0% 93.1%
Other National 0.0% 2.0% 3.6% 1.7%
Regional 0.0% 0.0% 21.4% 5.2%
% with informal mentor 72.9% 66.7% 60.7% 67.2%
% with formal mentor 83.8% 76.5% 39.3% 69.8%
Caucasian participants
Firm a?liation
‘‘Big 5’’ – 92.4% 89.7% 90.7%
Other National – 6.6% 6.0% 6.2%
Regional – 1.0% 4.3% 3.1%
% with informal mentor – 78.9% 83.9% 82.0%
% with formal mentor – 55.4% 35.1% 42.9%
R.E. Viator / Accounting, Organizations and Society 26 (2001) 541–561 547
demographic background for the remaining 109
participants employed in the public accounting is
presented in Table 1.
Responses from a previous study, in which the
participants were primarily Caucasian, served as
comparison data.
11
This mailing list was obtained
from the American Institute of CPAs (AICPA)
and consisted of members identi?ed initially as
employees of large public accounting ?rms. The
sample of 3000 members was strati?ed based on
gender, with an equal number of females and
males selected. Returned surveys included 902
participants, representing a 30% response rate.
12
Some responses were deleted for the following
reasons: incomplete surveys (n=13), not employed
in public accounting (n=25), partners/directors in
large ?rms (n=44), and members of racial/ethnic
groups other than Caucasian or African American
(n=54) or Caucasian prote´ ge´ s with cross-racial
mentors (n=3). Table 1 presents the demographic
background of the remaining 763 AICPA partici-
pants included in this study as comparison data.
Participants were required to identify: (1) whe-
ther they had current/recent mentors, and (2)
whether those mentors had been assigned to them
by a formal program at the ?rm, or were informal
mentoring relationships. As in other studies con-
sidering the e?ects of formal and informal men-
toring relationships, prote´ ge´ s who identi?ed
concurrent formal and informal mentoring rela-
tionships were instructed to consider their for-
mally assigned mentor when responding to
questionnaire items regarding support provided by
a mentor (Chao et al., 1992, Ragins & Cotton,
1999).
3.2. Measures
All scale items were scored on a ?ve-point Likert
scale from ‘‘strongly disagree’’ to ‘‘strongly
agree.’’ Pretests were conducted to ensure that
each scale item was applicable and re?ected
wording commonly used in the profession. In
some cases, minor wording changes were made
based on feedback received from 15 CPAs in four
o?ces of international public accounting ?rms.
Appendix A lists relevant questionnaire items,
along with their means, standard deviations, and
factor loadings.
3.2.1. Perceived barriers
The items assessing perceived barriers to men-
toring were adopted from Ragins and Cotton
(1991) and previously reported in Viator (1999).
Ragins and Cotton used principal component fac-
tor analysis to identify ?ve separate dimensions
for perceived barriers. Due to survey constraints,
13
this study adopted two of those factors: (1) lack of
access to potential mentors; and (2) lack of will-
ingness by potential mentors. In the Ragins and
Cotton study, these two factors demonstrated
strong construct validity in their predicted asso-
ciation with employee job rank, employee experi-
ence, and employee gender.
14
In the Viator study,
the factors had some association with mentor-type
and employee organizational level, but not
employee gender. Each factor was represented by
two questionnaire items, as presented in the
Appendix. The Cronbach Alpha for each factor
was greater than 0.80, suggesting relatively high
reliability, given 0.70 as a reasonable guideline
(Nunnally, 1978).
11
The two surveys were conducted at two di?erent points in
time, approximately one and half years apart. Detailed analyses
of data from the previous study are reported in Viator (1999)
and Viator (2001). Questionnaire items for the two surveys
were identical; however, survey instruments sent to NABA
members included the role reference ‘‘your current ?rm or
business’’ rather than simply ‘‘your current ?rm’’ in order to
accommodate responses from participants outside of public
accounting.
12
Kerlinger (1986) indicates that the expected response rate
from mail surveys is 20–40%.
13
As part of a broader study, the survey included some non-
mentoring questionnaire items. Response rate concerns limited
the length of the survey instrument and restricted the total
number of questions adopted.
14
One factor not selected lacked any association with pre-
dictor variables. Another factor, labeled ‘‘misinterpretation’’ by
Ragins and Cotton (1991), asked questions regarding potential
sexual innuendo. Concerns about sexual innuendo are likely to
generate some barriers to cross-gender mentoring relationships
(Ragins & McFarlin, 1990). These issues are not addressed in
the current study.
548 R.E. Viator / Accounting, Organizations and Society 26 (2001) 541–561
3.2.2. Mentoring functions
The mentoring function scale included 16 items
measuring two career support functions (career-
related mentoring, and protection and assistance)
and two psychosocial support functions (social
support and role modeling). These items were
adopted from scales used in previous mentoring
research (Chao et al., 1992; Dreher & Ash, 1990;
Noe, 1988; Scandura & Viator, 1994; Tepper,
1995; Turban & Dougherty, 1994; Viator, 2001).
Factor analysis, using an orthogonal rotation and
an eigenvalue-greater-than-1.0 criterion cuto?,
con?rmed the existence of four separate factors:
career-related, protection and assistance, social
support, and role modeling. As shown in the
Appendix, all items, except two social support
items, loaded on the expected factor, with factor
loadings greater than 0.50. After dropping the two
social support items, each factor had a Cronbach
alpha greater than 0.70, indicating acceptable
reliability.
3.2.3. Intentions to leave
Accounting studies have used a variety of mea-
sures of employee turnover intentions (Aranya &
Ferris, 1984; Aranya, Lachman, & Amernic, 1982;
Collins & Killough, 1992; Dillard & Ferris, 1979;
Harrell, Chewning, & Taylor, 1986; Harrell &
Stahl, 1984; Senatra, 1980). In the current study,
the turnover intentions scale was constructed from
two questionnaire items adopted from Collins and
Killough. The scale focused on ‘‘thinking about
leaving the ?rm’’ and ‘‘the probability of looking
for another job.’’ There is general support that
turnover intentions provide a viable indication of
subsequent actual turnovers (Steel & Ovalle,
1984). The Cronbach alpha was 0.85.
3.2.4. Demographic variables
A limitation of the current study is that partici-
pant race is not fully crossed with participant
organizational level: some NABA participants
were at the sta? level, but all AICPA participants
were at the senior accountant level or higher. In
order to facilitate data analysis presentation and
comparison of coe?cients, four dummy variables
are used to represent race and organizational level:
African-American Sta?, African-American Seniors,
African-American Managers, and Caucasian Man-
agers, (where Caucasian senior accountant is the
comparison category).
15
Other demographic variables likely to covary
with mentoring practices (Dreher & Ash, 1990;
Ragins & Cotton, 1991, 1999; Ragins & McFarlin,
1990) are included in the regression models.
Dummy variables identifying participant demo-
graphics include gender (0=male, and 1=female)
and national ?rm, regional ?rm (where ‘‘Big 5’’
international ?rm is the comparison category).
The regression analyses of barriers and intentions
to leave include dummy variables indicating the
presence or absence of informal mentors and formal
mentors.
Mentor demographic variables include dummy
variables identifying mentor race (0=Caucasian,
and 1=African American), mentor gender
(0=male, and 1=female), partner-mentor, senior
manager-mentor, other-mentor (where manager-
mentor is the comparison category), and the con-
tinuous variable length-of-mentoring-relationship.
The regression analyses of mentoring functions
include a dummy variable identifying formal men-
tor (0=informal, and 1=formal), as well as the
interaction of formal mentor with each racial
group. Only signi?cant interaction terms (e.g. the
interaction of formal mentor with African-Amer-
ican seniors) are reported in Section 3 below and
in Table 3. Prote´ ge´ s with concurrent formal and
informal mentors evaluated the level of mentoring
received from their formal mentor (as in Chao et
al., 1992; Ragins & Cotton, 1999).
4. Statistical analyses and results
Statistical test procedures included both chi-
square tests of association (for testing H1) and
linear regression analyses (for testing H2 through
H5). Speci?c tests of the directional hypotheses
are reported as one-tailed tests. Statistical results
15
Due to the relatively small number of African Americans
at the senior manager level (n=6), the variable manager com-
bines participants who identi?ed themselves as managers or
senior managers. Subsequent testing with an additional vari-
able for senior manager provided no substantive changes in the
results reported in this paper.
R.E. Viator / Accounting, Organizations and Society 26 (2001) 541–561 549
of the signi?cance of other demographic variables
are reported as non-directional two-tailed tests.
H1 stated that African Americans in public
accounting are less likely to have informal men-
tors. As previously noted, respondents indicated
whether they currently had an informal or formal
mentor, or both. A chi-square test performed on
the complete data set (n=872) was statistically
signi?cant (=13.73, P