Mentoring in public accounting firms: An analysis of mentor-protégé relationships, mentorship functions, and protégé turnover intentions

Mentoring in public accounting firms: An analysis of mentor-protégé relationships, mentorship functions, and protégé turnover intentions

~ ) Pergamon Accountfn~ Organtratfomand Soctety,VoL 19, No. 8, lap. 717-734, 1994 Copyright© 1994 ElsevierScienceLtd Printed in GreatBritain.All rig...

2MB Sizes 0 Downloads 10 Views

~ ) Pergamon

Accountfn~ Organtratfomand Soctety,VoL 19, No. 8, lap. 717-734, 1994

Copyright© 1994 ElsevierScienceLtd Printed in GreatBritain.All rights reserved O361-3682/94 $7.00+O.O0

MENTORING IN PUBLIC ACCOUNTING FIRMS: AN ANALYSIS OF MENTOR-PROTI~GE RELATIONSHIPS, MENTORSHIP FUNCTIONS, AND PROTEG]~ TURNOVER INTENTIONS*

T. & SCANDURA University o f M i a m i and 1~ E. VIATOR

University o f Kentucky

From in-depth interviews with public accounting per~otmel, Dtrsmith & Covaleski [Account/n~ Organ/zatfon$ and Soc/e0~(1985) pp. 149-169] concluded that mentorin4gexists in large public accounting firmsand benefitsthe mentor, ~ and the firm.Basedon a nationalsurveyofpublic accountingemployees, the current study uses quantitative data maly~ to identify public accounting mentoring fimction~ their effect on employee m o v e r intentions, and their associationwith specific- - o n a l variables(prot~-~ otgauimtionsl level,p r o t ~ gender, mentor's position, and audit firmstructure). The study found that public accounting mentoring consists of three separate functions: social suplx~ career development, and role modeling Path analysis indicated that employees with lower remover intentions received more career development support from their mentor and had a partner as a mentor. The level of social support provided by partner mentors tended to be less than that provided by manager mentors. Femaleprot~'g~s,who tended to have managers as mentors, received more social suppo~ when their mentor was also female. The role modeling function did not differ across ~ o n a l variables.The results of the study indicate that while social support is a key factor in definingthe m e n ~ process for public accounting employees, the career development function is associated with lower prot6"g~turnover intentions.

A c c o u n t i n g b e h a v i o r a l r e s e a r c h has d e m o n s t r a t e d that e m p l o y e e s w h o i n t e n d t o leave p u b l i c a c c o u n t i n g e x p e r i e n c e g r e a t e r role ambiguity a n d role conflict ( S o r e n s e n & Sorensen, 1974; Senatra, 1980), are m o r e likely to have u n m e t e x p e c t a t i o n s r e g a r d i n g organizational s e t t i n g a n d specific j o b c o n t e n t ( D e a n et aL, 1988), e x h i b i t l o w e r o r g a n i z a t i o n a l c o m m i t m e n t ( A r a n y a e t aL, 1 9 8 2 ) a n d e n c o u n t e r i n c r e a s e d o r g a n i z a t i o n a l - p r o f e s s i o n a l conflict ( A r a n y a & Ferris, 1984). Prior r e s e a r c h has s u g g e s t e d that e m p l o y e e t u r n o v e r i n t e n t i o n s c a n b e r e d u c e d t h r o u g h i m p r o v i n g c e r t a i n lines

of formal c o m m u n i c a t i o n s : the advising o f u n d e r graduate accounting students, the counseling of public accounting employees, and the comm u n i c a t i o n s b e t w e e n staff a c c o u n t a n t s a n d j o b s u p e r v i s o r s ( R h o d e et a t , 1977; Senatra, 1980; D e a n et aL, 1988; G r e g s o n , 1990). T h e s e r e c o m m e n d a t i o n s focus o n m o d i f y i n g formal c o m m u n i c a t i o n networks a n d overlook potential benefits w h i c h m a y b e o b t a i n e d from utilizing n o n f o r m a l c o m m u n i c a t i o n n e t w o r k s , s u c h as m e n t o r i n g o f lower-level e m p l o y e e s . K r a m ( 1 9 8 5 ) defines a m e n t o r as a n o l d e r a n d m o r e e x p e r i e n c e d e m p l o y e e w h o advises,

* The authors acknowledge the helpful comments of Adrian Harrell, Frank Collins,Jean Cooper, and Doug Poe on earlier versions of this paper. 717

718

T.A. SCANDURAand R. E. VIATOR

counsels, and otherwise enhances the career development of younger and relatively inexperienced employees ("prot6g6s"). Hunt & Michael ( 1 9 8 3 ) suggest that employees w h o have a m e n t o r experience greater job success and have a better chance of being promoted. In interviews with over 100 public accountants, Dirsmith & Covaleski (1985) found evidence that mentoring exists in public accounting. The authors concluded that, although public accounting mentoring relates to the performance of audit tasks, it serves to instruct the individual on the politics and p o w e r within the organization and provides a vehicle for socializing the individual into the profession. Organizational studies have identified three functions performed by mentors: coaching prot~g6s on career development, providing social support, and serving as a role model. The relative importance of these functions in public accounting mentoring has not been previously examined. While coaching prot6g6s on career development may lead to highly visible outcomes, such as prot6g6 retention and promotion, other mentoring functions, which relate to the socialization of the individual into the firm and the profession, may be necessary for the career development function to occur. This study uses quantitative data collection methods to identify the functions performed by public accounting mentors, assess the relative importance of each mentoring function, and examine their association with employee turnover intentions. The paper first reviews the type of mentoring functions identified in organizational studies using open-ended interviews and quantitative data collection techniques. Next, the paper discusses organizational variables which are likely to impact the level of mentoring received by public accounting prot6g6~s. The paper then presents results from a mentorship survey of 841 public accounting managers and staff employees in large national firms. The Results section of the paper discusses the mentoring functions identified, their relative importance in the mentoring process, the manner in which mentoring functions differ across specific organizational variables, and the association of mentoring functions with employee turnover intentions.

THEORETICAL ISSUES

Mentorship functions Kram (1983, 1985) conducted an extensive and detailed study of mentoring by interviewing mentor-prot~g~ pairs in a public-sector organization. From content analysis of these interviews, Kram ( 1 9 8 3 ) concluded that mentors provide two separate and distinct functions: a career development function and a psychosocial function. The career development function included such activities as the mentor sponsoring the prot~g~ in work assignments and promotional opportunities, enhancing the prot6g6's exposure and visibility, and providing the prot6g6 with political support and protection. The psychosocial function included an active role, w h e r e the mentor provided the prot~g6 with social acceptance and personal friendship, and a passive role, where the mentor served as a role model for the prot~g~ to learn appropriate organizational behavior. A subsequent study by Kram & Isabella ( 1 9 8 5 ) provided additional evidence from open-ended interviews that mentoring involves both a career development and psychosocial function. While prot6g6s are expected to benefit from both the career development function and the psychosocial function, the two mentoring functions serve different purposes. The career development function focuses on preparing the prot6g~ for career advancement, while the psychosocial function serves to clarify the prot6g6's sense o f identity and develop a greater sense of c o m p e t e n c e and serf-worth (Schockett & Haring-Hidore, 1985; Noe, 1988; Olian et aL, 1988). Several organizational studies have used quantitative data collection techniques and factor analysis to identify the functions performed by mentors and to determine the relative importance of each function (Burke, 1984; Schockett & Haring-Hidore, 1985; Douglas & Schoorman, 1988; Noe, 1988; Olian etaL, 1988; Dreher & Ash, 1990; Burke et aL, 1991; Scandura, 1992). Noe ( 1 9 8 8 ) and Schockett & Harin-Hidore ( 1 9 8 5 ) reported that the psychosocial function they identified explained more

MENTORING IN PUBLIC ACCOUNTING than four times as m u c h o f the mentoring variance as did the career development function. In an experimental study, Olian et aL ( 1 9 8 8 ) found that potential prot~g6s placed the mentor's perceived ability to address various psychological and social needs above career enhancing functions. Burke et aL ( 1 9 9 1 ) found that managers provided significantly more psychosocial functions, but not career development functions, to their prot6g6s as compared to typical subordinates. These findings suggest that the psychosocial function o f mentoring is very important in defining a mentoring relationship° While the mentoring process may support and enhance the prot6g6s career development, the psychosocial function provided b y mentors serves as the foundation for the mentoring process~ These findings from organizational studies seem to contradict some of the conclusions by Dirsmith & Covaleski ( 1 9 8 5 ) regarding the relative importance o f various mentoring functions in public accounting. From open-ended interviews with public accounting partners and employees, the authors found that "the issue of socialization was not interpreted to be important at the lower ranks as it relates to mentoring" (Dirsmith & Covaleski, 1985, p. 160). They attributed this to two unique characteristics o f public accounting: ( 1 ) the technical orientation of public accounting practice, and ( 2 ) the extensive formal socialization process associated with the large public accounting firms (training programs, office socials, etc.). However, the social support provided by a mentor, in the form of serving as a "role model", was considered important in the development and p r o m o t i o n of manager prot6g~s, w h e r e "at the manager level, there was a trend away from counselling o n doing toward behaving and being partnerlike" (Dirsmith & Covaleski, 1985, p. 160). Dirsmith and Covaleski reported that "the consensus of partners and managers was that the raison d'6tre for mentoring was for the

719

career management of the prot6g~", w h e r e the m e n t o r instructed the prot6g6 o n office politics, assisted in managing the prot6g~s visibility, and handled the prot6g6's promotional opportunities (Dirsmith & Covaleski, 1985, p. 161). All o f these activities relate to the career development function identified in other studies o f m e n t o r i n g (Roche, 1979; Burke, 1984; Kram, 1985; Kram & Isabella, 1985; Noe, 1988). Undoubtedly, these career development activities are the focal point for public accounting mentoring; however, the participants in the Dirsmith and Covaleski study may have discounted the importance o f the psychosocial function in establishing a mentoring relationship. The activities associated with the psychosocial function provide a way for the prot6g~ to get to know the mentor and serve to enhance the prot6g6's serf-esteem and sense of identity. Given the results from prior organizational studies of mentoring, the relative importance ofmentoring functions can be stated as a directional hypothesis. HI. In public accounting, the psychosocialfunction, as compared to the career development function, explains more variance in the underlying factor structure of mentoring. Dirsmith & Covaleski ( 1 9 8 5 ) identified role modeling as a separate mentoring function and noted its importance in the development o f manager prot6g~s aspiring to be partners. Using quantitative data analysis techniques, Burke ( 1 9 8 4 ) and Scandura ( 1 9 9 2 ) found evidence that role modeling is a separate and unique mentoring function, passive in nature and different from other psychosocial processes w h e r e the m e n t o r offers "active" social support (i.e. going to lunch and exchanging confidences ). If role modeling in public accounting is a separate mentoring function, its association with prot6g~ outcomes and organizational variables may differ w h e n compared to an "active" social support function) Based on the results from prior research, it is e x p e c t e d that role modeling

t The psychosocial function identified by Kram (1983) included both active and passive processes. We use the terms "social support" to identify active psychosucialprocesses and "role modeling" to identifypassivepsychosocialprocesses.

720

T. A. SCANDURAand 1L E. VIATOR

that social support may be more important to female prot6g6s compared to male prot6g6s. Although Dreher & Ash ( 1 9 9 0 ) and Ragins & H2. In public accounting, role modeling is a separate McFarlin ( 1 9 9 0 ) found essentially no prot6g6 mentoring function, distinct from an "active" social gender differences regarding mentoring funcsupport function. tions or mentoring roles, both Burke ( 1 9 8 4 ) and Noe ( 1988 ) found that female prot6g6s reported Mentoring functions are expected to have a receiving more social support from their mentor differential impact on employee career outcomes. compared to male prot6g6s. Hunt & Michael ( 1 9 8 3 ) suggest that mentoring None of the prior studies examined mentoring "aids in the development of managerial talent" in a male-dominated profession, such as public and that former prot6g6s tend to be "less accounting. In such professions, female prot6g6s mobile" (implying that former prot6g6s are less may not receive the necessary mentoring for likely to leave the organization). This study two reasons. First, female prot6g6s in a maleexplores the association between mentoring dominated profession may simply receive less functions and employee turnover intentions, mentoring than male prot6g6s. Second, as and examines whether the strength of mentoring indicated by Burke ( 1 9 8 4 ) and Noe (1988), functions differs based on p r o t 6 ~ organizational female prot6g6s may actually prefer more social level, prot6g6 gender, mentor's position, and support than male prot6g6s, yet they may not receive such social support from male mentors audit firm structure. in a male-dominated profession. Given that Burke also found that female mentors provided Protdg~ organizational level Prior studies, such as Rasch & Harrell (1990), more social support than male mentors, it is have found a negative relationship between reasonable that female prot6g6s in public employee organizational level (measures of accounting may form mentoring relationships "tenure") and employee turnover intentions. In with female mentors in order to receive the addition to having a direct negative affect preferred higher level of social support. Given o n employee turnover intentions, employee a shortage of potential female mentors in public organizational level may indirectly affect turn- accounting, the likelihood of female prot6g6s over intentions through mentoring functions. having cross-gender mentoring relationships That is, the level of mentoring functions may increases and the interaction of this effect with differ based on prot6g6 organizational level. For mentoring functions is uncertain. example, Dirsmith & Covaleski ( 1 9 8 5 ) found that role modeling was more important in the Mentor's position Hunt & Michael (1983) reported that mentors mentoring of manager prot6g6s, w h e r e mentors would demonstrate "how to behave" rather than tend to be at high organizational levels, possess articulate "what to do". Thus, the strength of power, and are willing to support and protect any detected mentoring functions may vary by their prot~g6s. Burke ( 1 9 8 4 ) suggested that w h e n the mentor is at a high level in the prot~g~ organizational level. organization, the prot~g6 gains more from the mentoring relationship. Given the benefit of Prot#gd gender It is possible that public accounting female having a mentor relatively high in the organizaprot6g6s may differ from male prot6g6s in the tion, it is expected that mentor's position will type of mentoring they receive, and this directly affect the strength of mentoring functions difference may be related to the higher turnover and prot6g6 turnover intentions. The mentor's position is likely to be a function rate experienced by females in public accounting (Pillsbury et aL, 1989; Rasch & HarreU, 1990). of prot6g6 organizational level and prot6g6 Some prior organizational research indicates gender. Dirsmith & Covalesld ( 1 9 8 5 ) reported

in public accounting is a unique mentoring function.

MI~NTOR1NGIN PUBLICACCOUNTING

721

m(-) vrot~a Organiz&tional Level

H4(+) "" .........

Mentor '• Pos~tLon

\7

'

.....

i

H.%)

I

~Turnover

In~ent £ons [ /

Ig¢(+)

, J~(+) Isoclal support [

Audit Firm Structure

[career ]neveZo~ment

H.~-) t

! ! $ ( + ~ [R°La R°deling [

H~-)

Prot6g~ Gender

Fig 1, Proposed model of memor-protC~ rehtlonships and mentortng functions. that partners tend to mentor managers, and managers tend to mentor seniors; thus, public accounting protCLg~ at higher organizational levels are more likely to have a partner as a mentor. In regards to prot~g6 gender, Pillsbury et aL (1989) suggested that w o m e n in public accounting are less likely than men to form a relationship with a "successful" mentor. Ragins (1989) identified several reasons w h y male mentors may be reluctant to sponsor female p r o t ~ s , including perceiving female prot6g6s as a greater risk than their male counterparts, being more comfortable with developing a professional and personal relationship with another male, and perceiving females as mothers and spouses rather than executive peers. Given that public accounting partners are predominantly male, these barriers to male-mentor/ female-prot6g6 mentoring relationships suggest that female proteges are less likely than male protc~g6s to have a public accounting partner as a mentor.

Audit structure Bamber etaL (1989) found that auditors from firms using structured audit methodologies experienced less role conflict and less role ambiguity. The systematic difference of role conflict and role ambiguity across audit firms may affect the degree of mentoring needed and provided in structured firms versus unstructured firms. That is, audit firms using unstructured methodologies are expected to provide more mentoring to compensate for increased role conflict and role ambiguity. In addition to the direct effect of audit firm structure on mentoring functions, audit firm structures may act as a moderating variable, influencing the strength of the assoc~tion between organizational variables and mentortng functions. For example, staff-level prot~g6s at unstructured firms may need and receive more mentoring than staff-level prot6g6s at structured firms, whereas manager-level prot~a~s at unstructured firms may receive the same degree of

722

T.A. SCANDURAand R. E. VIATOR

mentoring as managers at structured firms. In a similar manner, audit firm structure may moderate the strength of the association between mentor's position and mentoring functions. This study explores the effect of audit firm structure on mentoring functions by using Kinney's ( 1 9 8 6 ) classification of audit firm structure and identifying whether respondents worked at unstructured firms, intermediate structured firms, or structured firms.

Proposed relationships concerning prot#gd turnover intentions Based on the previous discussion, Fig. 1 presents the proposed relationships between the major variables. Most variables (prot6g6 organizational level, prot6g6 gender, mentor's position, and mentoring functions) are hypothesized to directly affect individual turnover intentions. Prot6g6 organizational level and prot~g6 gender are expected to indirectly affect turnover intentions through the mentor's position and mentoring functions (career development, social support, and role modeling), all intervening variables. Also, the mentor's position and audit firm structure are expected to indirectly affect turnover intentions through the three mentoring functions. The proposed relationships suggest several directional hypotheses concerning the relationship between organizational variables, mentoring functions, and prot6g6 turnover intentions. These general hypotheses are presented below and referenced in Fig. 1. H3. Prot6g6turnover intentions are directly and negatively affected by prot6g6 organizational level, prot6g6 gender, the mentor's position, and mentoring functions (career development, social support, and role modeLing~ H4. Mentor'sposition is directly and positivelyaffected by prot~g6 organizational level and prot6g6 gender. H5. Mentoring functions (career development, social support, and role modeling) are directly and positively

affected by prot6g6 organizationallevel, prot6g6 gender, mentor's position, but negatively affected by audit firm structure. Prior discussion suggested that audit firm structure may be a moderating variable which influences the strength of the association between organizational variables and mentoring functions. This relationship is not graphically presented in Fig. 1, but can be stated in a specific directional hypothesis. H6. As audit firm structure increases, the strength of the relationship between organizational variables and mentoring functions decreases.

RESEARCH METHOD

Subjects The study obtained a mailing list of 3000 public accounting employees firom the American Institute of Certified Public Accountants. The mailing list represented a random sample of managers and staff accountants from large public accounting firms, including all "Big 8" public accounting firms 2 and other large national firms, which is the population examined in prior studies of public accounting employee turnover intentions. Each potential participant was sent a survey packet which included a cover letter, survey questionnaire, and a postage paid return envelope. A total of 1024 questionnaires were returned, representing a 34% response rate. Of the 1024 completed survey instruments, 28 (or 2.7% ) were from respondents whose promotion to partner had not been reflected in the AICPA mail/tag list. These 28 ]responses were omitted from the data analysis performed by this study. Also omitted from the analysis were 106 respondents who had already left public accounting3 and

2The study was conducted prior to any mergers by the "Big8" public accounting firms. 3 Prior research of public accounting turnover has tended to focus on employee turnover intentions rather than on differences between employees who have left the firm and those who remain. This approach is followed in our study since only 10% of the respondents (106 out of 1024) had actually left public accounting.

MENTORING IN PUBLIC ACCOUNTING 49 u n u s a b l e responses. 4 T h e r e m a i n i n g 841 r e s p o n d e n t s w e r e f r o m e m p l o y e e s in large n a t i o n a l p u b l i c a c c o u n t i n g firms, i n c l u d i n g all o f t h e "Big 8" firms, w i t h t h e n u m b e r of p a r t i c i p a n t s r a n g i n g f r o m 6 4 to 170 p e r p u b l i c a c c o u n t i n g firm. T a b l e 1 p r e s e n t s d e s c r i p t i v e statistics for the 841 u s a b l e responses. 5

Survey questionnaire T h e s u r v e y q u e s t i o n n a i r e first asked t h e p a r t i c i p a n t s to identify w h e t h e r t h e y h a d experie n c e d working relationships with co-workers w h o had helped their career and had had a significant a n d positive effect o n their c a r e e r m o b i l i t y in t h e f i r . 6 P a r t i c i p a n t s w h o a n s w e r e d "yes" w e r e c o n s i d e r e d to have e x p e r i e n c e d a m e n t o r i n g relationship.7 These participants w e r e asked to r e p o r t h o w m a n y of t h e s e developmental relationships they had experienced and to d e s c r i b e t h e i r m e n t o r ( s ) in t e r m s o f t h e i r o r g a n i z a t i o n a l position, g e n d e r , age, a n d durat i o n o f t h e m e n t o r i n g relationship. T h e s e p a r t i c i p a n t s t h e n c o m p l e t e d the q u e s t i o n n a i r e s e c t i o n s c o n t a i n i n g the m e n t o r s h i p i n s t r u m e n t and questions relating to personal characteristics a n d c a r e e r mobility. Participants w h o a n s w e r e d

723

" n o " t o the first q u e s t i o n w e r e classified as n o t h a v i n g e x p e r i e n c e d a m e n t o r i n g relationship. These participants answered only those sections of t h e q u e s t i o n n a i r e w h i c h r e l a t e d t o t h e i r p e r s o n a l characteristics a n d c a r e e r mobility. T h e m e n t o r s h i p s u r v e y i n s t r u m e n t was o n e s e c t i o n of the m u l t i - p a r t q u e s t i o n n a i r e a n d c o n t a i n e d 20 Likert-type q u e s t i o n s b a s e d u p o n activities p r e v i o u s l y r e p o r t e d i n t h e literature. 8 These questions related to the three mentorship f u n c t i o n s ( c a r e e r d e v e l o p m e n t , social support, a n d role m o d e l i n g ) . T h e q u e s t i o n s c o v e r e d b o t h w o r k a n d n o n - w o r k activities a n d a d d r e s s e d m e n t o r s h i p activities r a n g i n g from the m e n t o r taking a p e r s o n a l i n t e r e s t i n t h e prot~g~'s c a r e e r to t h e m e n t o r socializing w i t h the prot~g~ after work. T h e w o r k - r e l a t e d b e h a v i o r s i n c l u d e d m e n t o r s giving advice o n t h e prot~g~'s p r e s e n t job, as w e l l as advice o n c a r e e r p r o g r e s s a n d promotional opportunities. Non work-related behaviors i n c l u d e d sharing of personal p r o b l e m s a n d confidences. T h e q u e s t i o n s w e r e d e s i g n e d to reflect the n a t u r e o f p u b l i c a c c o u n t i n g a n d p r e t e s t e d u s i n g thirty CPAs from the local offices o f "Big 8" p u b l i c a c c o u n t i n g firms. T h e s u r v e y q u e s t i o n n a i r e c o n t a i n e d items

4 The unusable responses included 33 participants who identified a "peer" as their mentor. Because this low rate results in very few observations in some subsamples, these 33 responses were omitted from the following statistical analyses. The conclusions of this paper are not affected by omitting these 33 participants from the data set. s Subject to limitations imposed by promised confidentiality, the authors agree to provide additional information from the survey data. Copies of the survey instrument itself may be obtained by writing directly to the authors. 6 The survey question identifying the respondent as having experienced a mentoring relationship is taken from Kram's (1983) study. Although Kram used open-ended interviews to identify mentoring relationships and explore mentoring functions, the "pivotal question" which identified the relationship(s) to be examined further was "Is there anyone among those that you have mentioned today that you feel has taken a personal interest in you and your development?" (Kram, 1983, p. 612). 7 It is important to emphasize that this study used an indirect question to identify whether the participants had experienced a mentoring relationship. This insured that the participants were describing a "nonformal" relationship rather than one established within the formal organizational structure, such as the relationship with an "advisor" or "counselor". Also, the relationships identified in this study are based on employee perceptions and are considered to be "perceived" mentoring relationships. s The survey questionnaire was limited to four pages in order to obtain a relatively high number of responses (n -- 1024). This feature of the survey design precluded obtaining additional information concerning other employee behaviors and attitudes. Also, the mentorship instrument was designed to detect the existence of "general" mentorship functions, and did not include items which may have further distinguished general and specific functions, such as items relating to audit practice career development versus audit business career development.

T. A. SCANDURAand R. E. VIATOR

724

TABLE 1. Descriptive statistics for respondents Category Respondent background

Reported having a mentor Yes No ( n = 651 ) ( n - 190)

Organizational level Staff accountant Male

All respondents ( n -- 841)

237

87

Female

12___99

53

324 182

Total

366

140

5O6

Male Female

198 8"7

36

Total

285

50

234 101 335

Education level Bachelor's Master's

478 173

187 3

Manager

1.._44

Age Mean Minimum Maximum S.D. Years experience Mean Minimum Maximum S.D.

relating to career aspirations and turnover intentions. For the analysis reported in this paper, participants w e r e classified as having a "staying" intention if they reported that the position they would like to hold in five years g a s either their current position or a higher position in the firm. They w e r e classified as "leaving" if they reported other intentions such as expecting to hold a position in a different firm, a position in industry, or b e c o m i n g a sole practitioner. Accounting studies have used s o m e w h a t different measures o f employee turnover intentions (Dillard & Ferris, 1979; Senatra, 1980; Aranya et aL, 1982; Harrell & Stab_l, 1984; Aranya & Ferris, 1984; Harrell et aL, 1986). The current study was interested in capturing the respondent's career aspiration and therefore asked the respondent to identify the position he/she would like to hold in five years, a question similar to Dillard & Ferris (1979). However, to test for construct validity, a second turnover

29.5 22 45 4.07

29.3 22 45 4.48

5.16 1 13 2.38

3.26 1 11 1.94

intentions scale was constructed from three questionnaire items which asked the participant to rate on a five-point scale his/her expectations about being p r o m o t e d at the firm, his/her expectations o n leaving the firm to get ahead, and ~ e r expectations on b e c o m i n g a partner at the firm. This turnover intentions scale was also used to statistically test the relationships p r o p o s e d in Fig. 1. The results of this testing w e r e similar to those reported in Fig. 2, providing evidence of construct validity.

RESULTS

Mentorship funaiom The 20-item mentorship instrument was factor analyzed using varimax orthogoml rotation procedures and resulted in the final three-factor solution presented in Table 2. Five items w e r e d r o p p e d owing to low or ambiguous factor loadings. A criterion cutoff point o f 0.50 for

725

MENTORING IN PUBLIC ACCOUNTING TABLE 2. Varimax factor Ioadings Mentorship scale

Mentorship item 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

Mentor takes a personal interest in m y career. Mentor has placed m e in important assignments. Mentor gives m e special coaching o n t h e job. Mentor advises m e about promotional opportunities. I share personal p r o b l e m s with mentor. Mentor helps m e coordinate professional goals. I socialize with m e n t o r after work. I try to m o d e l m y behavior after mentor. I admire m e n t o r ' s ability to motivate others. I e x c h a n g e confidences with mentor. I r e s p e c t m e n t o r ' s knowledge o f t h e accounting profession. 12. I consider m e n t o r to b e a friend. 13. I respect m e n t o r ' s ability to teach others. 14. Mentor has d e v o t e d special time a n d consideration to m y career. 15. I often go to l u n c h w i t h mentor.

Factor 1 Social support

Factor 2 Career development

Factor 3 Role modeling

0.20

0.71

o.05

-0.13

0.64

0.03

0.21

0.62

0.22

0.25 0.73 0.07 0.78 0.17 0.10 0,:7"7 0.00

0.57 0.14 0.54 0.03 0.03 0.O6 0.10 O.14

0.02 0.14 0.23 0.OO 0.69 O.80 0.23 0.55

0.70 0.18 0.25

0.24 0.25 0.73

0.24 0.69 0.13

0.65

0.22

0.00

accepting items was used, as suggested by goals, and devoting special time and consideration Nunnally (1978). The identified factors were to the prot6g6's career. This six-item scale had a readily interpretable as the three previously Cronbach alpha of 0.75 and was associated with reported mentoring functions: career develop- 11% of the total mentorship variance. The third factor related to role modeling ment, social support, and role modeling. The items that loaded strongly on the first activities and included the prot6g6 modeling factor were activities where prot6g6s share his/her behavior after the mentor, admiring the more personal aspects of their lives with their mentor's ability to motivate others, respecting mentor ("social support" activities). These the mentor's knowledge of the accounting items included sharing personal problems with profession, and respecting the mentor's ability the mentor, socializing with the mentor after to teach others. These items represent prot~g6's work, exchanging confidences with the mentor, perceptions of their mentor's style (a passive considering the mentor to be a friend, and going form of mentoring) rather than personal interto lunch with the mentor. This five-item scale actions between prot6g6.s and mentors (active had a Cronbach alpha of 0.81 and was associated forms of mentoring). This four-item scale had a Cronbach alpha of 0.70 and was associated with with 30% of the total mentorship v a r i a n c e . 9 The items that loaded strongly on the second 9% of the mentorship variance. The social support factor is associated with factor related to career development activities and included the mentor taking a personal the largest explained variance, indicating the interest in the prot6g6's career, placing the relative importance of social support in defining prot~g6 in important assignments, giving special the mentoring process. Certain socialization coaching on the job, advising about promotional activities (such as the prot6g6 sharing personal opportunities, helping to coordinate professional problems with the mentor and exchanging

9 The principal c o m p o n e n t analysis p r o d u c e d 16 factors. Only t h e t h r e e factors described in t h e text had eigenvatues greater than one. These three factors a c c o u n t e d for 5096 o f t h e m e n t o r s h i p variance; the remaining factors, associated with t h e u n e x p l a i n e d variance, w e r e individually insignificant.

T. A. SCANDURAand R. E. VIATOR

726

TABLE3. Zero-order correlation coefficientsfor model variables 1 1.

2. 3. 4. 5.

6. 7. 8.

Turnover intention ( 1 = Leave) (O = Stay) Mentor's position (1 -- Partner) (0 = Manager) Prot~g6organ, level (1 -- Manager) (o = Staff) Prot~g~gender (1 -- Male) (O = Female) Audit firm structure (3 = Structured) (2 = Intermediate) (1 = Unstructured) Socialsupport Career development Role modeling

2

3

-0.09**

0.14"**

--

-0.09"*

0.10"*

0.07*

0.Ol

4

5

6

7

--0.10"**

0.04 -0.05

-0.04 -0.27*** -0.05 -o.1 l*** -0.05 -0.07* -0.06 -0.03 -0.06

-0.06

0.05 -0.05 -0.02

0.06 0.05 0.05

-0.43*** 0.36*** 0.41"**

* p < o.10; ** p < 0.05; *** p < 0.01. confidences) are apparently critical in defining the mentoring relationship. This finding agrees with prior organizational studies using quantitative methods to explore the underlying factor structure of mentoring (Schockett & HaringHidore, 1985; Noe, 1988; Olian et aL, 1988; Burke et aL, 1991). Although no statistical tests are possible, a three-to-one ratio in the percentage of explained variance suggests that the social support function in public accounting is relatively important in explaining mentoring: public accounting mentoring implies engaging in social support activities. Thus, w e conclude that there is support for Hypothesis 1. The results of factor analysis indicate that, in public accounting, the role modeling function is a separate mentoring function, distinct from an active social support function. The finding of three separate mentoring functions is not trivial. The conclusions reached by Dirsmith & C o ~ e s k i ( 1 9 8 5 ) had suggested that role modeling is a separate and distinct mentoring function in public accounting; however, a separate orthogonal factor for the role modeling function was not found by either Noe (1988), Shockett & Haring-Hidore (1985), or Dreher & Ash ( 1990 ), even though these studies used a scale which

included role modeling items. The results o f this study suggest that, as Dirsmith and Covaleski described in their study, part of the process in public accounting mentoring involves the prot~g~ observing h o w the mentor behaves in certain situations. We conclude there is support for Hypothesis 2.

Correlations Table 3 presents Pearson Product Moment correlations for the variables studied. As shown in this table, turnover intentions were significantly and inversely related to mentor's position, prot~g~ organizational level, and prot~g~ gender (with females more likely to have intentions to leave the firm). With respect to mentoring functions, only the career development function was significantly and negatively related to turnover intentions. Audit firm structure was marginally ( p < 0.10) and positively related to turnover intentions.

Path analysis Path analysis, or structural equation modeling (Pedhazur, 1982)was used to further investigate the direct and indirect relationships proposed in Fig. 1. For those equations which had a

MENTORING IN PUBLIC ACCOUNTING

727

Protege

Organizational Level

E

~

Mentor's

[

/

~

Posi.t:~on

I

,

@.06

~ ~ntentions Turnover

Z

/

I

[,oo ,l so, o= I

I Audit Firm structure

Prot~g~ Gender

Fig. 2. Empirical model of mentor-prot6g~ relationships and mentoring functions.

dichotomous dependent variable, logistic regression (legit) was used to calculate the partial regression coefficients (path coefficients). 1° Ordinary least squares regression was used for those equations with continuous dependent variables (i.e. the three mentoring functions). Coefficients not statistically significant (alpha = 0.05) were removed, resulting in the revised model presented in Fig. 2. The Fig. 2 revised model indicates that most of the proposed relationships from Fig. 1 were supported. The first hypothesis tested in path analysis is Hypothesis 3, which implies that prot6g6 organizational level, prot6g6 gender, mentor's position, and the three mentoring functions directly affect prot6g6 turnover intentions. Except for the social support and role

modeling functions, path analysis indicated that the specified variables directly affected prot6g~ turnover intentions, supporting Hypothesis 3. Consistent with prior research on public accounting employee turnover intentions (Pillsbury et aL, 1989; Rasch & HarreU, 1990), the results indicate that prot6g6s at lower organizational levels and female prot6g6s are more likely to leave the firm. Prot6g6s whose mentor is a manager in the firm, rather than a partner, are also more likely to leave the firm. Dirsmith & Covaleski (1985) reported that certain career development activities (such as assisting the prot6g6 in being assigned to the "right" jobs )were considered by the interviewees to be most important in the mentoring process (p. 161). The path analysis presented in this

~oOrdinary least squares regression requires that the error terms be normally distributed with constant variance. Such assumptions are unlikely to be true when the dependent variable is dichotomous. The sensitivity of ordinary least squares parameter estimates has led to recommendations for using logistic regression, which has fewer underlying assumptions (Aldrich & Nelson, 1984).

728

T. A. SCANDURAand IL E. VIATOR

study indicates that higher levels of c a r e e r d e v e l o p m e n t support w e r e associated with prot6g6 intentions to stay with the firm. It is interesting to note that the career d e v e l o p m e n t function directly affected prot~g~ turnover intentions, whereas the social support function, which explained a m u c h larger percentage o f the mentorship variance, did not. This poses an interesting question as to the relative importance o f each function. The results do not imply that the social support function is unimportant. In fact, the results of factor analysis indicate that the social support function is a critical factor in defining the mentoring process. In public accounting, the mentoring of less experienced employees includes activities where the prot~g6 shares personal p r o b l e m s with the mentor, socializes with the m e n t o r after work, exchanges confidences with the mentor, and considers the m e n t o r to be a friend. These social support activities may need to exist in order that career d e v e l o p m e n t activities, which are linked to prot6g6 career outcomes, can occur. Hypothesis 4 p r o p o s e d that prot6g6 organiza. tional level and prot6g6 gender indirectly affect turnover intentions through the m e n t o r ' s position. The results of path analysis provide support for this hypothesis. Manager prot6g~s, versus staffprot6g6s, w e r e m o r e likely to have a partner as a mentor. This finding agrees with that r e p o r t e d by Dirsmith & Covaleski (1985). Also, male prot6g6s w e r e m o r e likely to have a partner as a mentor, indicating that female prot6g~s may face barriers in obtaining mentorship from partners in public accounting organizations. Hypothesis 5 p r o p o s e d that mentoring functions are directly and positively affected b y prot6g6 organizational level, prot6g~ gender, m e n t o r ' s position, yet negatively affected b y audit firm structure. For the career d e v e l o p m e n t

function, the initial regression model indicated no significant effects. However, audit firm structure was hypothesized to b e a potential moderating variable, w h e r e an interaction b e t w e e n audit firm structure and the other variables could affect the level of mentoring received. The linear regression m o d e l was rerun and included an interaction b e t w e e n audit firm structure and each of the o t h e r variables. A significant effect was found for the interaction of audit firm structure with prot~g~ organizational level. To interpret the meaning of this interaction effect, separate regression models w e r e run for each prot~g6 organization level. The results indicate that as the level o f audit firm structure increased ( f r o m unstructured to intermediate, then structured) career developm e n t support increased in the staff prot~g~ group ( b e t a = 0.149, p < 0.005), but not the manager prot~g~ group ( b e t a = - 0 . 0 9 7 , p >

o.lo). The significant association in the staffprot~g~ group is different than hypothesized. Staff prot6g~s at firms with m o r e audit structure received m o r e ( n o t less) career d e v e l o p m e n t support, c o m p a r e d to staffprot~g6s at firms with less audit structure. Staff prot~g6s at firms with m o r e audit structure rated their m e n t o r higher on such activities as taking a personal interest in their career, placing t h e m in important assignments, giving t h e m special coaching on the job, advising t h e m about promotional opportunities, helping t h e m to coordinate their professional goals, and devoting consideration to their career, tl Hypothesis 5 also p r o p o s e d that the social support function is directly and positively affected b y prot6g6 organizational level, prot~g~ gender, m e n t o r ' s position, and audit firm structure. Path analysis indicates tl~at only m e n t o r ' s

t t Another way of viewing the moderating effect of audit firm structure is to examine the association between prot~'g6 organizational level and the career development function, for each level of audit firm structure. The association between prot~g~ organizationallevel and the career development function is not significantfor prot6"g~sat unstructured audit firms 07 <0.05), but does exhibit a significant negative correlation for prot6-g~ at intermediate audit firms (r -- -0.17, p < 0.05) and structured audit firms (r = -21, p < 0.01). These results indicate that for intermediate and structured audit firms, prot~g~ at lower organizational levels (staff/seniors) receive more career development mentoring than prot~g6"s at higher organizational levels (managers).

MENTORING IN PUBLIC ACCOUNTING TABLE 4. Analysis of female prot~g~ social support by mentor gender

Mentor gender Female mentor Male mentor

Number of female

Mean level of social support

prot~g6"s

(scale ----25)

33 183

18.55 16.56 F = 5.52 p < 0.02

position had a direct effect on the social support function; however, this effect was inverse from the hypothesized effect. Prot6~g~s with a partner m e n t o r tended to receive less social support than prot6g6s with a manager as a mentor. Although they did not receive as much social support, prot6g~s with a partner m e n t o r did have lower intentions to leave the firm. Female prot6ges c o m p a r e d to male prot6g6s gave their mentors similar ratings for all three mentoring functions, including the social support function. This is especially interesting since prior research (Burke, 1984; Noe, 1988) had found that female protc~g¢~s in various organizations received more social support than male prot~g6s. The level of mentoring received may be a function of w h e t h e r the mentoring relationship is a cross-gender relationship. Therefore, additional analysis o f variance (ANOVA) was performed to determine if the level of mentoring differed based o n the interaction of m e n t o r gender with prot6"g~ gender. The interaction term was not sigmfi"cant for the social support function ( o r any o t h e r mentoring function). However, this lack of statistical significance may b e a function of the relatively small sample size for some cross-gender mentoring groups. Since the key issue is w h e t h e r the level o f social support ( o r any other mentoring function) differs for female prot,~g6s with male mentors versus female prot6g6s with female mentors, additional tests (t-tests) were performed o n the reduced data set (female prot6g6s only). As presented in Table 4, female prot6g6s w h o identified a female m e n t o r reported receiving significantly higher levels of social support, compared to female prot6g~s w h o had identified a male mentor.

729

Given the results of Burke ( 1 9 8 4 ) and Noe (1988), it appears that public accounting female prote'g6s with female mentors receive the e x p e c t e d higher level o f social support, whereas female proteges with male mentors did not. This seems to suggest that male mentors in public accounting may need to be m o r e attentive to the social support they provide their female prot6g~s. However, increased social involvement in cross-gender mentoring relationships poses problems for the potential male mentor and female prot~g~ because o f potential sexual innuendos and office gossip (Fitt & Newton, 1981; Collins, 1983; Clawson & Kram, 1984; Kram, 1985; Ragins, 1989). Finally, Hypothesis 5 proposed that the role modeling function is directly and positively affected by prot~g~ organizational level, prot~g¢~ gender, mentor's position, and audit firm structure. Path analysis indicates that none o f these variables had a direct affect o n the role modelling function. The regression models w e r e re-run to include audit firm structure as a moderating variable and no significant effects w e r e found. Although Dirsmith & Covaleski ( 1 9 8 5 ) found that role modeling was most important to manager prot6g~s, this study ( using quantitative data collection methods ) found that manager prot~g6s compared to staff prot6g6s rated their mentors similarly regarding role modeling activities. The results from this study suggest that a mentor's role modeling is as important to lower level prot~g6s as it is to manager level prot~g6s.

Additional analysis o f turnover intentions The completed survey instruments included 106 respondents w h o w e r e not currently employed in public accounting These offered the study an opportunity to examine the construct validity associated with the turnover intentions variable. Construct validity focuses o n the extent to which a measurement performs in accordance with logical expectations. It was e x p e c t e d that responses from individuals w h o had left public accounting would be similar to those currently in public accounting w h o indicated an intention to leave, yet different

730

T.A. SCANDURAand R. E. VIATOR

TABLE5. Descriptive analysis of the construct validity for prot6g6 turnover intentions

Descriptor variable Prot6g6 organizational level Staff Manager Prot~g~ gender Female Male Mentor's position Manager Partner Career development function (high score -- 30)

Prot6g6s Prot6g6s Prot~g6s who intend who intend who IeR to stay to leave the firm (n = 425) (n = 226) (n=32)

52.7% 47.3%

62.0% 38.0%

65.6% 34.4%

30.1% 69.9%

38.9% 61.1%

40.6% 59.4%

35.5% 64.5% 25.1

46.9% 53.1% 24.4

46,9% 53.1% 24.1

from those w h o indicated an intention to stay. Additional tests were performed on the association between turnover status (intend to stay, intend to leave, and left the firm) and explanatory variables (prot6g~ organizational level, prot6g6 gender, mentor's position, and mentoring

functions). Prior to obtaining their current position, 54 of the 106 respondents had worked for another employer after leaving public accounting. 12 These 54 respondents were omitted from the additional analysis. Of the remaining 52 respondents, 20 indicated that they had not experienced a mentoring relationship and thus did not complete the survey sections concerning mentoring functions and mentor's position. The remaining 32 respondents indicated that they had experienced a mentoring relationship while in public accounting (they identified either a partner or manager w h o had helped their career). The responses of these 32 individuals w e r e used to examine the construct validity of the turnover intentions variable. Because of the low n u m b e r of usable responses in this category, the additional analysis is descriptive in nature

and does not include any additional statistical tests. Table 5 presents these descriptive data. The profile of prot~g6s w h o had left their public accounting firm was very similar to those w h o indicated an intention to leave public accounting, yet clearly different from those prot~g6s w h o indicated an intention to stay in public accounting. With regard to prot~g6 organizational level, the percentage of prot6g6s with a staff level position was 65.6% for those w h o had left their firm and 62.0% for those w h o intended to leave their firm, yet only 52.7% for those w h o intended to stay with their firm. With regard to prot~g~ gender, the percentage of prot6g6s w h o were female was 40.6% for those w h o had left their firm and 38.9% for those w h o intended to leave their firm, yet only 30.1% for those w h o intended to stay with their firm. In a similar manner, the percentage of prot6g6s w h o indicated a manager as their mentor was 46.9% for those w h o had left their firm and 46.9% for those w h o intended to leave their firm, yet only 35.5% for those w h o intended to stay with their firm. Initial tests had indicated that prot6g6s w h o intended to leave their firm received lower levels of career development support, as compared to prot6g~s w h o intended to stay with their firm. As reported in Table 5, prot6g~s w h o had left their firm reported a level of career development support (24.1) which was similar to those w h o intended to leave their firm (24.4), yet different from those w h o intended to stay with their firm (25.1). In fact, prot6g6s w h o had left their firm reported receiving the lowest level of career development support, while those w h o indicated an intention to stay with their firm reported the highest level of career development support. In summary, the analysis presented in Table 5 supports the construct validity of the prot6g~ turnover intentions variable. The survey design raises questions concerning the construct validity of the identified "perceived"

~2This findingsuggests that the MCPAmailinglist database contains outdated employmentstatus on some of its members. Of the 1024 questionnaires returned in this study, 54 (or 5.3%) were from members who had changed jobs a second time since their employmentin public accounting

MENTORING IN PUBLIC ACCOUNTING

mentoring relationships. In this study, 651 of 841 respondents w e r e classified as having identified a mentoring relationship (see Table 1). It was expected that having a mentor in public accounting would be associated with intentions to stay with the firm. Sixty-five per cent of the 651 respondents classified as having experienced a mentoring relationship indicated an intention to stay with their firm, whereas only 29% of the 190 employees reporting no mentoring relationship indicated an intention to stay. t3 These frequencies were statistically different (Chi-square = 17.26, p <0.0001). Thus, as expected, the identified "perceived" mentoring relationships were associated with lower employee turnover intentions. This analysis provides some support for the construct validity of the mentoring relationships discussed in this study.

DISCUSSION Some limitations of this study should be mentioned. First, while the study obtained a relatively large number of usable responses ( n = 841), the responses were obtained from a mail survey sent to 3000 public accounting employees. It is possible that the results presented in this paper are affected by a response bias; however, the demographic information presented in the paper suggests that the background of the participants in this study is representative of the more general population of public accounting employees. A second limitation is that the study used a single, indirect measure to determine whether participants did or did not have a mentor. While this measure is based on Kram's (1983) pivotal question for identifying mentoring relationships, it is possible that this question did not properly include and exclude participants based on whether they had

731

experienced a nonformal mentoring relationship. The data analysis presented in this paper are limited by the use of this indirect question to identify mentoring relationships. Based on the analysis presented in this paper, public accounting mentoring appears to consist of three separate functions: social support, career development, and role modeling. The social support function explained the largest percentage of the mentorship variance (more than that explained by the career development function), indicating that part of being a mentor involves engaging in social support activities such as sharing personal problems and exchanging confidences. The finding of a separate role modeling function suggests that the psychosocial function of public accounting mentoring also includes a passive socialization process, where prot~g6s learn by observing the mentor, as opposed to actively socializing with the mentor. Of the three mentoring functions, only the career development function was associated with lower prot6g~ turnover intentions. Prote'g6s who reported receiving higher levels of career development support also indicated lower turnover intentions. However, the relative importance of the social support function should not be overlooked. Although the social support function is not directly linked to prot6g6 career intentions, it seems to be essential in defining the mentoring process and providing a basis for the career development function to occur. The level of career development support did not differ across organizational variables, except that staff proteges received more career development support as the level of audit firm structure increased. Although this effect was in the opposite direction than expected, the result suggests that higher audit firm structure is not only associated with lower role conflict and role

13 It is possible that e m p l o y e e s w h o had n o t e x p e r i e n c e d a m e n t o r i n g relationship ( w h o w e r e unable to identify a coworker w h o h a d helped their career and s u p p o r t e d their career mobility) m i g h t have had a greater tendency to decline participating in this study. This potential response bias could have affected t h e "no m e n t o r i n g relationship" frequencies reported in this study. Potential response bias and other study limitations are further e x a m i n e d in t h e Discussion section of this paper.

732

T . A . SCANDURA and R. E. VlATOR

ambiguity (as reported by Bamber et aL, 1989), but it is also associated with additional mentoring provided to lower-level employees. Based on the results from prior studies (Burke, 1984; Noe, 1988), it was initially expected that female prott~g6s would receive higher levels of social support from their mentors. In this study, the level of social support did not differ across prot6qg~ gender (female prot6g6s reported receiving as much social support as male protc~g6s); however, female prot£~g6s with a female mentor did report a higher level of social support compared to female protc~g6s with a male mentor. Given the results o f prior studies, it seems reasonable that female protc~gc~smay actually prefer more social support than do male prot6g6s, yet may have difficulty obtaining this higher level of social support in a male-dominated profession. The level of role modeling did not differ by either prot6g6 gender or prot6g6 organizational level. This suggests that role modeling, as measured in this study, is perceived to be equally important to public accounting proteges, regardless o f organizational level or gender. As expected, having a partner as a mentor was significantly associated with lower intentions to leave the firm. Manager prot6g~s (rather than staff proteges) and male proteges (rather then female prot6g~s) w e r e more likely to have a parmer mentor. The disparity across prot~g~ gender may be difficult to reduce since female proteges seem to prefer relatively higher levels of social support compared to male proteges, yet partner mentors tend to provide less social support to their proteges. This study raises additional questions concerning the dynamics of the mentoring process. Active social support from a mentor appears to be relatively important in defining a mentoring

relationship, yet the social support function was not directly associated with lower prot~g6 turnover intentions. It may b e that a mentor's social support impacts other variables, such as the prot6g~s self-esteem, self-efficacy, influence orientation, or achievement motivation, which in turn influences prot6g6 career outcomes. Given the relative importance o f social support to female prot6g~s, additional research may focus on the manner in which female mentors versus male mentors provide the additional social support preferred by female prot6gc~s. Accounting research should also take a closer look at barriers to cross-gender mentoring relationships, including concerns about sexual innuendos and sexual harassment. The analysis of the effect of mentoring functions across audit firm structure indicated that higher audit firm structure was associated with higher levels of monitoring. Given the results of this study and Bamber et aL (1989), future research should examine w h e t h e r higher levels of mentoring are directly associated with lower role conflict and role ambiguity. It may be that social support indirectly affects prot6g~ career outcomes through either lower role conflict and role ambiguity or lower workrelated stress. Additional work is needed to understand which organizational characteristics facilitate or hinder the initiation and cultivation of mentoring relationships. Several national public accounting firms are attempting to either formalize mentoring programs or provide rewards to those individuals w h o devote time and effort to nonformal mentoring relationships. Future research may examine w h e t h e r formal mentoring programs provide benefits similar to nonformal mentoring and what extrinsic rewards support and encourage nonformal mentoring.

BIBLIOGRAPHY Aldrich, J. H. & Nelson, F. C., Linear Probability, Logit and Probit Models ( Beverly Hills, CA: Sage, 1984 ). Aranya, N. & Ferris, K. R., A Reexamination of Accountants Organizational-Professional Conflict, Accounting Rev/ew (January 1984) pp. 1-15.

MENTORING IN PUBLIC ACCOUNTING Arany~ N., Lachman, IL & Amernic, J., Accountants' Job Satisfaction: A Path Analysis, Accounting Organizations and Society (1982) pp. 201-215. Bamber, E. M., Snowtmll, D. & Tubbs, IL M., Audit Structure and Its Relation to Role Conflict and Role Ambiguity: An Empirical Investigation, Accounting Review (April 1989) pp. 285-299. Burke, R. J., Mentors in Organizations, Group and Organization Studies (1984) pp. 353-372. Burke, R. J., McKenna, C. S. & McKeen, C. A., How do Mentorships Differ from Typical Supervisory Relationships, Psychological Reports (1991) pp. 459-466. Clawson, J. G. & Kram, ICE., Managing Cross-gender Mentoring, Business Horizons (May/June 1984) pp. 22-32. Collins, N. W., Professional Women and Their Mentors (Englewood Cliffs, NJ: Prentice-Hall, 1983). Dean, R. A., Ferris, K. IL & Konstans, C., Occupational Reality Shock and Organizational Commitment: Evidence from the Accounting Profession,Accounting Organizations and Society ( 1988 ) pp. 235-250. Dillard, J. F. & Ferris, IC 1L, Sources of Professional Staff Turnover in Public Accounting Firms: Some Further Evidence, Accounting Organizations a n d Society ( 1979 ) pp. 179-186. Dirsmith, M. W. & Covaleski, M. A., Informal Communications, Nonformal Communications and Mentoring in Public Accounting Firms, Accountin~ Organizations a n d Society ( 1985 ) pp. 149--169. Douglas, C. A. & Schoorman, F. D., The Impact of Career and Psychosocial Mentoring by Supervisors and Peers, Paper presented at the 1988 National Academy of Management meetings in Anaheim, CA

(1988). Drcher, G. F. & Ash, IL A., A Comparative Study of Mentoring among Men and Women in Managerial, Professional and Technical Positions,Journal of Applied Psychology (1990) pp. 539-546. Fitt, L W. & Newton, D. A., When the Mentor is a Man and the Protege is a Woman, Harvard Business Review(March/April 1981)pp. 55-60. Gregson, T., Communication Satisfaction: A Path Analytic Study of Accountants Afftliated with CPA Firms, Behavioral Research in Accounting (1990) pp. 32-49. Harrell, A. M. & Stahl, M. J., McClelland's Trichotomy of Needs Theory the Job Satisfaction and Work Performance of CPA Firm professionals, Accounting Organizations and Society (1984) pp. 241-252. Harrell, A., Chewning, E. & Taylor, M., Organizational-Professional Conflict and the Job Satisfaction and Turnover Intentions of Internal Auditors, Auditing. A Journal of Practice and Theory (Spring 1986) pp. 109-121. Hunt, D. M. & Michael, C., Mentorship: A Career Training and Development Tool,Academy ofManagement Review (1983) pp. 475--485. Kinney, W. IL, Jr, Audit Technology and preferences for Auditing Standards,Journal of Accounting and Economics (March 1986) pp. 73--89. Kram, ICE., Phases of the Mentor Relationship, Academy of ManagementJournal (1983) pp. 608-625. Kram, IC E., Mentoring at Work: Developmental Relationships in Organizational Life (Glenview, IL: Scott, Foresman, & Co., 1985). Kram, ICE. & Isabella, L. A., Mentoring Alternatives: The Role of Peer Relationships in Career Development, Academy of ManagementJournal (1985) pp. 110-132. Noe, R. A., An Investigation of the Determinants of Successful Assigned Mentoring Relationships, Personne/ Psychology (1988) pp. 457-479. Nunnally, J. C., Psychometric Theory (New York: McGraw-Hill, 1978). Olian, J. D., Carroll, S. J., Giannantonio, C. M. & Feren, D. B., What do Proteges Look for in a Mentor? Results of Three Experimental Studies,Journal of Vocational Behavior (1988) pp. 15-37. Pedhazur, E., Multiple Regression in Behavioral Research: Explanation and Prediction ( New York: Holt, Rinehart and Winston, 1982). Pillsbury, C., Capozzoli, L & Ciampa, A., A Synthesis of Research Studies Regarding the Upward Mobility of Women in Public Accounting, Accounting Horizons (March 1989) pp. 63-70. Ragins, B. R., Barriers to Mentoring: The Female Manager's Dilemma, Human Relations (1989) pp. 1-22. Ragins, B. 1L & McFarlin, D., Perceptions of Mentor Roles in Cross-gender Mentoring Relationships, Journal of Vocational Behavior (1990) pp. 321-339. Rasch, R. H. & Harrell, A, The Impact of Personal Characteristics on the Turnover Behavior of Accounting professionals, Auditing: A Journal of Practice and Theory (Spring 1990) pp. 90-102. Rhode, J. G., Sorensen, J. E. & Lawler, E. E., III, Sources of Professional Staff Turnover in Public Accounting Firms Revealed by the Exit Interview, Accounting Organizations and Society (1977) pp. 165-175. Roche, G. IL, Much Ado about Mentors, Harvard Business Review (Jan/Feb 1979) pp. 14--24. Scandura, T. A., Mentorship and Career Mobility: An Empirical Investigation,Journal of Organizational Behavior (1992) pp. 169-174.

733

734

T.A. SCANDURA and 1~ E. VIATOR Schockett, M. & Harding-Hidore, M., Factor Analytic Support for Psychosocial and Vocational Mentoring Functions, Psychological Reports (1985 ) pp. 627--630. Senatra, P. T., Role Conflict, Role Ambiguity, and Organizational Climate in a Public Accounting Firm, Accounting Review (October 1980) pp. 594-.-603. Sorensen, J. E. & Sorenson, T. L, The Conflict of Professionals in Bureaucratic Organizations, Administrative Science Quarterly (March 1974) pp. 98-106.