Does mentor support increase women's career advancement more than men's? The differential effects of career and psychosocial support.

Abstract:

Based on past research on mentoring, this study examined whether mentor career support helped women's career advancement more than it did men's, and more than psychosocial (emotional) support did. The sample of 3220 Australians, chiefly from the public sector and finance and business service industry, was surveyed twice, a year apart. Mentor career support increased women proteges' advancement more than it did men's, whereas psychosocial support generally reduced women's advancement more than it did men's. Unexpectedly, female mentors, not males, had the strongest effects, both helping and hindering their protege's advancement. Mentor support was not related to men's advancement. Why mentoring was differently related to men's and women's advancement was discussed.

Keywords:

MENTORING; WOMEN IN MANAGEMENT; GENDER DIFFERENCES; ADVANCEMENT.

1. Introduction

Mentoring is thought to assist both men's and women's career advancement, but to be especially helpful for women because they may encounter more barriers to their advancement than men do (Burke & McKeen 1990; Noe 1988; Ragins 1989, 1999; Wallace 2001; Wanberg, Welsh & Hezlett 2003). Indeed, reviews show that organisations have established some mentor programs specifically to help women and others underrepresented in management to advance, acknowledging that women may have more family, organisational, and interpersonal barriers to their hierarchical advancement than men do (Clutterbuck & Ragins 2002; Blake-Beard 2001). This study assesses whether mentoring helps women to advance in their careers more than it does men. If it does, then the emphasis on mentoring lbr women to reduce their under-representation in management would appear worthwhile.

It is not known from reviews of the results of prior studies whether mentor support is related to gender differences in advancement (Noe, Greenberger & Wang 2002; O'Neill 2002). This may be because mentor support comprises two separate functions (career and emotional/ psychosocial) which scholars propose have different links to advancement (Kram 1985; Ragins 1997). The two functions appear not yet to have been tested separately for their links to gender differences in advancement. If career support from mentors is most beneficial to women's advancement, rather than psychosocial support, then more emphasis can be placed on career support to help women advance. The aim of this study is to assess whether mentor career support assists women's career advancement more than it does men's and whether it has more positive effects than psychosocial support does. Reviewers of the literature strongly recommend that research examine the link of mentoring (mentor support, dyad gender) to gender differences in career advancement because of the dearth of studies in the area (Allen, Eby, Poteet, Lentz & Lima 2004; Wanberg, Welsh & Hezlett 2003).

2. Mentor Support

2.1 Mentor Support and Career Advancement

Mentor support has been conceptualised as comprising career and psychosocial support (Kram 1983; Noe 1988; Ragins 1999). Role modelling is not included in the present study as a distinct function of mentoring separate from its inclusion in psychosocial support. Allen et al's (2004, p. 128) recent review concluded that 'the extant theoretical and empirical research is clear that career and psychosocial support serve as the primary distinct and reliable overarching operationalizations of mentoring'. Career support entails mentors sponsoring their proteges for advancement, coaching them, providing challenging assignments, and protecting them and making them visible in organisations, and is therefore argued to primarily help proteges' hierarchical advancement (Allen et al. 2004; Kram 1985; Ragins 1997; Ragins & Cotton 1999; Wanberg, Welsh & Hezlett 2003). Psychosocial support entails mentors providing their proteges with friendship and acceptance, counselling them, and acting as role models, and therefore is argued to primarily help proteges' emotional well-being and personal growth (e.g. self-worth) and, secondarily, their career advancement (Allen et al. 2004; Kram 1985; Ragins 1997; Ragins & Cotton 1999; Wanberg, Welsh & Hezlett 2003).

Career support would therefore be more related than psychosocial support to career advancement, as Allen et al's (2004) meta-analysis cautiously inferred from the few relevant studies. In cross-sectional studies, career support was related to career advancement or success more than psychosocial support was, which usually was not related (Ensher, Thomas & Murphy 2001; Orpen 1995; Scandura 1992; Waters, McCabe, Kiellerup & Kiellerup 2002). However, Day and Allen (2004) found neither career nor psychosocial support was related to the salary and promotions of public servants of a single municipality, though having a mentor was. Overall, mentor career support appears related to career advancement more than psychosocial support is (Allen et al. 2004), supporting the propositions made in relation to mentoring (Kram 1985; Ragins 1997; Ragins & Cotton 1999).

2.2 Gender Differences

Mentors are thought to be essential to women's career advancement (Burke & McKeen 1990; Ragins 1989, 1999; Ragins & Sundstrom 1989), and may enhance women proteges' career outcomes more than men's (Noe, Greenberger & Wang 2002). Why might such gender differences occur? Scholars suggest that women need more interventions than men, including mentoring, to help them advance in the hierarchy in their organisations because they have more obstacles to their advancement than men do (Ragins 1989, 1999; Ragins & Sundstrom 1989; Tharenou 1997; Wallace 2001). Tharenou's (1997, 1999) reviews provided empirical support for women incurring more barriers to their managerial career advancement than men do. The major barriers women incurred were gender discrimination, male hierarchies, and lack of informal networks that assist advancement, which helped men advance in management more than women. Moreover, studies have supported the view that women needed a mentor more than men did to advance and that men advanced without mentors due to not needing them as much (Bahniuk, Dobos & Kogler Hill 1990; Burt 1998; Schor 1997). Hence, more barriers and more need may explain why mentoring helps women to advance more than men. The notion that mentor support may assist women's career advancement more than it does men's does not mean, however, that men do not need mentors to advance.

Why should mentor career support and psychosocial support affect women's and men's career advancement differently? The two mentoring functions differ, and in a way that suggests they may have different results for men and women. Because, compared to men, women incur more gender discrimination, more dissimilarity from male managerial hierarchies, and lack informal networks for advancement (Tharenou 1997, 1999), they may need to be sponsored, provided with challenging assignments, and coached more than men to advance. Sponsoring a female protege gives her visibility, credibility, legitimacy, and access to her sponsor's networks (Burt 1998). Burt (1998) showed that mentor sponsoring was not needed by men to advance but it was by women. Coaching a woman on how to advance and to deal with the organisation helps her negotiate the minefield for promotion. Providing her with challenging assignments gives her the preparation, visibility, and credibility for advancement (Mainiero 1994; Tharenou 1997, 1999). Although a mentor's provision of psychosocial support will help a woman's emotional well-being, it will not directly help her advance and may be an inappropriate focus. Indeed, focusing on a woman's emotions and on internal psychological factors (e.g. acceptance) may disadvantage her advancement compared to focusing on actions that can directly assist her career and that focus on external factors (e.g. through sponsoring her).

2.3 Evidence for Different Links" of Mentor Support to Career Advancement by Gender

Indeed, when mentor career and psychosocial support are not differentiated, tests for gender differences in advancement give mixed results. The rare studies are cross-sectional in design. The results of one group of studies measuring general mentor support combining career and psychosocial support suggest mentor support may not be especially helpful to women's advancement or may even have negative effects. Dreher and Ash (1990) found that mentor support combining career and psychosocial support was not differently related to men's and women's salary, promotions, and chance of promotion. Lyness and Thompson (2000) found significant gender differences for US bank executives. Mentor support had negative links with executive women's career success but not executive men's. Lyness and Thompson (2000) did not explain the negative link, but their mentor support scale measured psychosocial support more than career support. In a study of women's advancement in Australian banks, Metz and Tharenou (2001) found that general mentor support combining career and psychosocial support was mostly not, and for one sample was negatively, related to managerial advancement, a finding for which no explanation was given. It is not known if the results were affected by combining possible positive links for career support with advancement for women with no or negative links for psychosocial support. Ragins (1999) recommended that studies differentiate the two types of mentoring to test gender differences in outcomes.

By contrast, a second group of studies suggests that mentor career support helps women advance more than men, and is the critical feature of mentoring. Burt (1998) found in a high technology firm that women advanced more to executive levels and earlier when they had strong ties to a strategic sponsor, unlike men whose advancement was predicted by large, informal, weakly connected networks. Burt (1998) concluded that women needed legitimacy and credibility to advance in management, which strong ties to a strategic sponsor gave them, whereas men already had that legitimacy and credibility because they were men. Therefore, to advance, women needed mentor sponsorship more than men. Johnson and Scandura (1994) found that mentor coaching was positively related to women accountants' salaries but not men's, whereas friendship and role modelling were not related to either's salary. Johnson and Scandura (1994) proposed that the positive result for coaching and nonsignificant result for the psychosocial functions arose because coaching provides on-the-job training, but did not explain why coaching did not affect men. Coaching may have enabled women to advance through learning the ropes and dealing with the environment for advancement that they face, which men needed to do less than women to advance. Schor (1997) found that women executives, unlike men counterparts, reported that mentors were important to their advancement throughout their careers. The functions that emerged inductively from the interviews were all career functions directly enhancing advancement. The women described their mentors as advisors, sponsors, teachers, path pavers, and career guides, but did not mention psychosocial functions. The men rarely mentioned mentors in relation to their advancement, describing them only as role models. The present study extends prior studies by testing if career and psychosocial support will have different links to men's and women's advancement (salary, promotion, managerial level), with mentor career support rather than psychosocial support having positive links.

In sum, mentor career support should increase women's career advancement more than it does men's, and more than psychosocial support does. Because it does not directly assist advancement, psychosocial support is unlikely to explain gender differences in advancement or would do so less than career support. Psychosocial support is likely to be unrelated to women's advancement and may be negatively related (Lyness & Thompson 2000) due to its focus. It focuses on emotions and internally, unlike career support, which focuses directly on career actions and externally. This study extends prior studies by using a longitudinal design to assess these links, because mentoring is likely to have lagged rather than simultaneous links with advancement. Wanberg, Welsh and Hezlett's (2003) review and a search of more recent publications since then shows that most studies of the impact of mentoring on career outcomes are cross-sectional and that no longitudinal studies have considered the interaction between gender and mentoring on career outcomes. Wanberg, Welsh and Hezlett (2003) and Allen et al. (2004) pointed out that it is important to show that mentoring precedes career success outcomes in time and that very few studies have used such designs. The present study appears to be the first to provide a longitudinal test of the interactive effect between gender and mentoring on career outcomes. Based on the arguments given, it is posed that:

H1: Mentor career support will increase women's career advancement (H1a), and more than it does men's (H1b).

H2: Mentor psychosocial support will be less related to women's career advancement than career support (H2a), and may reduce or not be related to it (H2b).

H3: Mentor psychosocial support will be less related to gender differences in career advancement than will mentor career support.

3. Interaction with the Gender of the Mentor

The links proposed so far have ignored the gender of the mentor, which may affect the impact of a mentor's career support on a protege's advancement. Proteges of male mentors may gain greater career outcomes than proteges of female mentors. Using cross-sectional designs, studies show that both male and female proteges of male mentors have higher salaries than those of female mentors (Dreher & Chargois 1998; Dreher & Cox 1996; Wallace 2001; Weill & Kimball 1996). Other studies have given mixed results dependent on the type of measure of a male mentor (Ragins & Cotton 1999) or of career advancement (Ensher, Thomas & Murphy 2001; Day & Allen 2004). Why might the proteges of male mentors have higher salaries than the proteges of female mentors? Ragins (1989, 1997, 1999) argued that male mentors provide more career support than female mentors, including sponsoring their proteges to high-ranking positions, because men have more power in organisations than women do. Indeed, although Ragins and Cotton (1999) found nonsignificant results, Sosik and Godshalk (2000) found that male mentors provided more career support than female mentors, who provided more psychosocial support (role modelling) than male mentors. Hence, the proteges of male mentors may advance more than those of female mentors, because male mentors provide more career support and have more power to leverage it.

Examining the links of mentor support to gender differences in career advancement requires consideration of the interaction of the mentor's and protege's gender with the type of mentor support (Ragins 1999). Reviews show that studies have not examined such links (Allen et al. 2004; Wanberg, Welsh & Hezlett 2003). For women proteges, career support from a male mentor may assist their advancement more than from a female mentor. This is because male mentors have more power than female mentors to affect their protege's advancement through the exercise of their support (Ragins & Sundstrom 1989). There is the greater possibility of differential effects by the gender of the mentor for women than men because women have male and female mentors whereas most men have male mentors (O'Neill 2002; Ragins 1989, 1999). Men have been found to advance more than women whether they had a mentor or not, but women without mentors to advance the least of all groups (Bahniuk, Dobos & Kogler Hill 1990). It is not that mentoring does not have an impact on men's careers, but that it may have more of an impact on women's. Hence, women with male mentors may gain greater advancement from their mentor's career support than women with female mentors (Ragins & Cotton 1999). The effect will not be as strong for men as they do not vary in mentor gender as much and do not need mentor career support for advancement as much as women do.

However, some evidence suggests female mentors may be best for women's career development. Reviews of mixed evidence indicate that proteges in same-gender mentoring relationships may gain advantages over those in cross-gender relationships (Noe, Greenberger & Wang 2002; O'Neill 2002; Ragins 1999). The reviews concluded that women with female as opposed to male mentors can learn more strategies for dealing with barriers to their advancement (e.g. work-life balance), gain more social support including role modelling, and gain more career support especially with challenging assignments. Hence, there may also be a countering force against the advantages of a male mentor's career support, with positive effects for women having female mentors. Despite the conflicting possibilities, overall, it is proposed that:

H4: Mentor career support will increase career advancement more for women proteges when the mentor is a male rather than a female.

4. Method

4.1 Respondents and Data Collection

The sample for this study was selected to comprise employees at lower and middle organisational levels to obtain those who could still advance in their careers and thus for whom mentoring may be most relevant. Both public sector (Australian Public Service, APS) and private sector employees were sought. The private sector employees were chiefly from the finance, property and business services industry, which was thought to be most comparable to the APS. Stratified sampling was used to select men and a comparable number of women by level within organisations. The study was part of a larger study explaining career advancement. At Time 1 (T1), surveys were mailed to 10,820 employees for return in prepaid envelopes. The return rate was 52% (n = 5627; 2614 women, 3013 men) of whom 83% (4670) volunteered for follow up. At Time 2 (T2), which was a year later, 323 were lost to mailing (e.g. incorrect addresses, left employment, maternity leave), leaving 4347 of whom 79% responded (3434: 1593 women, 1841 men).

Table 1 describes the Time 1 sample. They were mostly from 20 to 49 years old with an average age of 30 to 34 years, and were chiefly subordinates up to middle managers.

Employees were asked if they had a mentor, defined as a higher-ranking person with advanced experience and knowledge who has played a committed role in the development of their careers (cf. Kram 1985; Noe, Greenberger & Wang 2002; Ragins 1999; Scandura & Williams 2001) and, if so, to select that person. Employees' advancement over their careers was of interest in this study. Hence, no time restriction was placed on when the relationship with the mentor had occurred, similarly to prior studies that thus included both current and past mentors (Bahniuk, Dobos & Kogler Hill 1990; Corzine, Buntzman, & Busch 1994; Day & Allen 2004; Dreher & Ash 1990; Dreher & Chargois 1998; Dreher & Cox 1996; Johnson & Scandura 1994; Kirchmeyer 1998; Lyness & Thompson 2000; Wallace 2001; Weill & Kimball 1996). Chao (1997; Chao, Walz, & Gardner 1992) found little difference in results, including in career outcomes, between a sample of current and past proteges and the sub sample of current proteges. Chao (1997) concluded that the advantage of mentored groups on income endures over the long term, as did Rosenbaum (1984). He found the positive effect of sponsors on the promotions of early career employees flowed on to their later advancement. Advancement benefits arise from mentoring beyond the duration of the relationship, partly through the skills proteges learn and the signals given of their ability and potential (Chao 1997; Kram 1985; Rosenbaum 1984). Therefore, it was felt appropriate to use both past and current mentors to predict increases in advancement over the next year.

At Time 1, 89.2% of the respondents reported a mentor. Of those 5019, 4166 volunteered to be followed up. The respondents reported at Time 1 that mentors were their immediate supervisor (37%), another senior person in the organisation (30%) or outside of it (3%), a past boss (25%), or another category of person (4%). Recent studies have shown that direct supervisors of subordinates may be their mentors. For example, 50% of Ragins, Cotton and Miller's (2000) proteges reported their mentors were their immediate supervisors, 42% of Scandura and Williams (2004), and 85% of Sosik and Godshalk's (2000) and Burke and McKeen's (1997) samples.

Chi-square tests of the Time 1 data showed that the Time 2 nonrespondents differed from respondents by being younger [[chi square] (9, 5001) = 184.92, p = 0.00], more likely to be single [[chi square] (1, 5010) = 12.67.17, p = 0.00], and less educated [[chi square] (9, 5001) = 186.58, p = 0.00]. They had worked fewer years [[chi square] (6, 5004)= 91.20, p = 0.00], held lower level occupations [[chi square] (7, 5003)= 370.46, p = 0.00], and worked in the private sector more than the public sector [[chi square] (1, 5010)= 475.64, p = 0.00] and in larger rather than smaller organisations ([chi square] (4, 5006) 136.05, p = 0.00].

However, they did not differ on gender [[chi square] (1, 5010) = 6.42, ns], having a mentor [[chi square] (1, 5010) = 0.01, ns], or mentor gender [[chi square] (1, 5010) = 3.62, ns]. Nonrespondents had a lower level of advancement over their careers than respondents. They had lower salaries, [chi square] (9, 5001) = 243.67, p = 0.00; lower managerial levels, [chi square] (7, 5003) = 76.93, p = 0.00; and fewer managerial promotions, [chi square] (5, 5005) = 99.14, p = 0.00. Nonrespondents had a higher level of current and prospective advancement than respondents. They were promoted more last year, [chi square] (1, 5010) = 7.27, p = 0.01, had less time since promotion, [chi square] (4, 5006) = 21.80, p = 0.00, and a greater chance of promotion, [chi square] (4, 5006)= 11.07, p = 0.03), probably because they were younger than respondents. To help cater for the differences between nonrespondents and respondents, the analyses controlled for earlier advancement and demographic variables.

4.2 Measures

4.2.1 Career Advancement At Times 1 and 2, six outcomes were used in order to capture career advancement as fully as possible. They were predicted separately. They are different aspects of advancement, as their intercorrelations show (Table 2), and have been shown not to be related in the same way (e.g. promotions with salary and level) for women as for men (e.g. Cox & Harquail 1991; Ragins & Cotton 1999; Stroh, Brett & Reilly 1992).

Three outcomes measured the level of advancement an employee had over their career. Employees were asked their salary range using a 10-point item scored from 1, under $15,000 to 10, over $95,000. They described their managerial level using Tharenou's (2001) 8-point item (1, nonsupervisor/nonmanager; 2, first-level supervisor; 3, lower manager; 4, middle manager; 5, senior manager; 6, executive; 7, senior executive/divisional head; 8, CEO). They gave the number of managerial promotions they had in their careers on a 6-point item from 1, none to 6, nine or more. Based on Cox and Harquail (1991), managerial promotions were defined as promotion from one level in the managerial hierarchy to a higher one (e.g. from subordinate to supervisor, middle to senior manager).

It was not possible to verify the self-reports of salary, level, and promotions but they have been found highly related to company reports (Turban & Dougherty 1994). There was support for the convergent and discriminant validity of the three measures in the present study. Salary, managerial level, and managerial promotions were more related to other measures of advancement indicating responsibility (number of people supervised, 0.43, 0.58, 0.43; number of years supervising others, 0.50, 0.63, 0.72; level of occupation type, from managerial and administrative to unskilled, -0.54, -0.51, -0.36) than to measures less, or spuriously, related to advancement (number of organisational levels, -0.13, -0.11, 0.05; organisation size, -0.14, -0.13, 0.04; respondent age, 0.36, 0.37, 0.46; company tenure, 0.24, 0.22, 0.34).

The other three outcomes measured current and prospective advancement. Employees reported if they had been promoted in the last year (coded 1, no; 2, yes). They described how long they had been without promotion from 1, less than 1 year to 5, 10 or more years, using Gattiker and Larwood's (1989) item. They rated their chance of promotion over the next 5 years from 1, no chance to 5, very good chance, using Stout, Slocum and Cron's (1988) item. As expected, time since promotion, chance of promotion, and promoted last year were more related to years in the position (0.46, -0.28, 0.26) than to time to the position (0.14, -0.06, -0.09).

4.2.2 Mentor Support At Time 1, employees who reported a mentor rated his/her support from 1, strongly disagree to 7, strongly agree on the 21, 7-point item mentor support scale of Ragins and McFarlin (1990, pp. 328-9; Ragins & Cotton 1999, p. 550). The scale measured the career functions of sponsor, challenging assignments, and coach, and the psychosocial functions of role model, counsel, acceptor, and friend. Confirmatory factor analysis using LISREL 8.30 (Joreskog & Sorbom 2000) showed that 1-, 2- (career, psychosocial support) and 7 factor models exceeded .90 on absolute fit indices. The fit indices for the 1-, 2-, and 7 factor models were: GFI = 0.970, 0.986, 0.972; AGFI = 0.964, 0.980, 0.965; RMSEA = 0.0694, 0.0505, 0.0677. The 2-factor model had the highest absolute fit. Improvements in fit were assessed between the 2 and 7-factor models by differences in chi-squares, as recommended (Anderson & Williams 1992). The 2-factor model improved the fit over the 7-factor model, [DELTA][chi square] ([DELTA]df = 20) - 2346.90, p = 0.00, and was used. The 2-factor solution had the highest fit, corresponded to the major conceptualisations from propositions made about mentoring (Allen et al. 2004; Kram 1985; Noe 1988; Noe, Greenberger & Wang 2002; Ragins 1999), and arose in the other confirmatory factor analysis conducted of mentor support (Tepper, Shaffer, & Tepper 1996). Hence, it was felt justified to lose the distinctions among the seven mentor dimensions making up the original scale. Career support averaged the nine items for sponsor, challenging assignments, and coach ([alpha] = 0.85). Psychosocial support averaged the 12 items for friend, acceptance, role model, and counsel ([alpha] = 0.91).

4.2.3 Mentor Gender Gender of the mentor was coded as men (1) and women (2).

4.2.4 Protege Gender Protege gender was coded as women (1) and men (2).

4.2.5 Control Variables Ragins (1999) has recommended stringent controls for examining the links of mentoring to gender differences in outcomes, which Noe, Greenberger and Wang (2002) and Wanberg, Welsh and Hezlett (2003) supported in their reviews. The present study followed Ragins' guidelines, and also allowed for the control of individual characteristics (e.g. ability, personality) that Wanberg, Welsh and Hezlett (2003) say provide for alternative or spurious explanations of the link of mentoring to career success. Ragins (1999) recommended the control of particular variables that affect mentoring functions and outcomes and vary by gender. They were the employee's organisational rank (i.e. managerial level), organisational tenure, education, age, and position, and organisation size and type. As Ragins (1999) showed, this is because, if not controlled, significant relationships are found for mentor variables that are due to other influences. Ragins (1999) also recommended, when assessing the relationship between gender and mentoring functions and outcomes, the control of specific mentoring-related variables that may differ according to gender. They included the length of the mentoring relationship, whether the mentor is the supervisor or not, and whether the organisation has a mentoring program or not.

The control variables also allowed for the control of gender differences in characteristics of the sample. The men and women differed (see table 1), as shown by t-tests. A negative t-test for the control variables measured at T1 that follow indicates that women have a lower score on that attribute than men (women were coded, 1; men, 2), which they consistently had.

All the t-tests in the brackets that follow showed significant gender differences at p = 0.00. The controls took into account the type of sector (coded 1, public; 2, private; t = -4.33) and the organisation's size (measured from 1, fewer than 1000 employees to 5, 8000 employees or greater; t = -4.65). Also controlled were the demographic variables of protege age (measured from 1, 20-24 to 9, 60 to 64 years; t = -5.36) and organisational tenure (measured from 1, fewer than 5 years to 8, greater than 35 years; t = -12.89). The type of position the employee held was controlled. The level of the respondent's occupation type was the average of the scores given to both their position and occupation using the single-digit codes of the Australian Standard Classification of Occupations. The scores are interval data (Department of Employment and Industrial Relations 1987) assessing occupational skill level, from 1, managers and administrators to 8, labourers (t = 3.97, meaning women were at lower occupational levels than men). Managerial level (t = -4.33) was used to control the respondent's rank (Ragins 1999) when it was not the dependent variable.

Human capital was controlled by education (measured from 1, some secondary school to 9, masters/PhD; t = 0.67, ns) and training and development (t = -10.40) which averaged Tharenou's (2001) 6, 7-point items, which both predict advancement. The trait of masculinity (t = -3.05, p = 0.002), an instrumental orientation focused on getting the job done or problem solved, was controlled because it predicts managerial advancement (Tharenou 1997) and mentor support (Scandura & Ragins 1994) and differs by gender (Bern 1981). It averaged the 10, 7-point items of the short form of the Bern Sex Role Inventory (Bern 1981).

The study controlled the three mentoring-related variables Ragins (1999) suggested. Formal mentor program was the extent respondents' organisations provided access to formal mentoring programs (measured from 1, not at all to 7, always; t = -2.26, p < 0.05). Years relationship was how long the relationship was with the selected mentor (measured from 1, less than 1 year to 6, 5 or more years; t = -5.98, p = 0.00). Mentor level was whether the mentor was the respondent's supervisor (coded 1) or a more senior person (coded 2) (t = 0.65, ns).

4.3 Methods of Analysis

Moderated hierarchical regression analysis was used to test the hypotheses. The predictor variables were all measured at Time 1 and the dependent variables all measured at Time 2. Logistic regression initially was used to predict promoted last year, because it was a binary variable. Because the results were very similar to those using linear regression, the latter are reported for consistency with the results for the other five career outcomes. The independent variables (mentor support, mentor gender, and protege gender) were each entered (after the controls), followed by the 2- and 3-way interactions to test the interactive effects proposed. All 2- and 3-way interactions were entered to take into account their effects.

The order of entry of the predictors was: (a) the controls; (b) mentor career support, psychosocial support, and mentor gender; (c) protege gender; (d) the 2-way interactions between each mentor variable and protege gender (testing H1b, H3); and (e) the 3-way interactions between the type of mentor support and protege and mentor gender (testing H4). In addition, to predict the increase in career advancement, the regressions were rerun predicting the T2 career advancement measures controlling their T1 counterparts in the first step. If the interaction between gender and a particular variable is significant, the beta weights for that variable are significantly different for men and women. Separate regressions were run for men and women to examine the support for H1 and H2 for the propositions made for women alone (H1a, H2) and to interpret the interactions (H1b, H3, H4).

5. Results

Descriptive statistics showed some differences in regard to the mentoring of women and men. At Time 1, women reported having a mentor more than men but only by 2%, [chi square] (1, 5627) = 4.92, p = 0.03 (90%, 88%). Women reported, as much as men, [chi square] (1, 5010) = 0.42, ns, that their mentor was their boss (38% women, 37% men) or a more senior person (30%, 32%). Women had female mentors more than men did, [chi square] (1, 5010) = 300.16, p = 0.00. The percentage of men with male mentors was 43%, of men with female mentors was 7%, of women with male mentors was 32%, and of women with female mentors was 18%. Women reported receiving more mentor career support (t = 4.09, p = 0.00; X = 4.87 women, X = 4.72 men) and psychosocial support (t = 5.12, p = 0.00; X = 5.34 women, X = 5.14 men) than men.

Male mentors did not provide more career support than female mentors and female mentors provided more psychosocial support especially to female proteges. Respondents reported the same amount of career support from male as female mentors respectively (t = -0.45, ns; X = 4.80; X = 4.79) irrespective of protege gender (F = 0.00, ns). Proteges reported more psychosocial support from female than male mentors respectively (t = -5.67, p = 0.00; X = 5.37, X = 5.20) in interaction with protege gender (F = 11.91, p < .001). Female proteges of female mentors (X = 5.45) had more psychosocial support than male proteges of female mentors (X = 5.29) and female (X = 5.18) and male (X= 5.13) proteges of male mentors.

Table 2 gives the means, standard deviations, alpha coefficients, and correlations for the total sample. Table 3 gives the results for the regression analysis predicting the T2 career advancement measures for the total sample without Time 1 advancement controlled. Table 4 predicts the increase in advancement by controlling T1 advancement, which appeared to capture most of the effect of the control variables (cf. table 3). Tables 5 (not controlling the T1 advancement measure) and 6 (controlling the T1 advancement measure) provide the results of the regressions run separately for men and women.

Formal tests were made of multivariate multicollinearity, resulting in high tolerances (generally > 0.70), except for training and managerial level, and low variance inflation factors (all [less than or equal to] 2). Some pairs of variables were correlated above 0.50: training and occupational level; managerial level with occupational level or training; and career and psychosocial support. Hence, each variable in a pair was omitted systematically in reruns of the regression analyses. The same results were obtained for the hypotheses with the variables omitted or included. Hence, multicollinearity was not a problem and the results were robust. Although there were multiple significance tests made with a large sample, the confidence level of p < 0.05 was kept because of the stringent controls and the time lag for prediction of a year.

5.1 Hypothesis 1: Mentor Career Support

H1 posed that mentor career support would increase women's career advancement (H1a), and more than men's (H1b). For H1a to be supported, career support needed to have significant beta weights in the results of the regression equations run for women (tables 5, 6). For H lb to be supported, the 2-way interactions between career support and protege gender needed to add significant variance to the main effects and have significant beta weights (tables 3, 4).

Overall, H1 was supported. As shown in tables 5 and 6, supporting H1a, career support positively predicted women's career advancement a year later for 5 of the 6 career advancement measures (not managerial level). As shown by the significant interaction effects and increments in variance (tables 3, 4), supporting H1b, career support and protege gender significantly interacted to predict salary (table 4 only), managerial promotions, time since promotion, chance of promotion, and promoted last year. A significant interaction for a particular variable shows the beta weights for men and women for that variable significantly differ from each other. The separate regressions run for men and women (tables 5, 6) showed significant beta weights for the prediction by career support for women but not men (with one exception). The two beta weights to be given in each bracket stand for the prediction by career support of, first, Time 2 advancement without Time 1 advancement controlled (table 5) and, second, Time 2 advancement with Time 1 advancement controlled (table 6). Beyond organisational, job, and individual factors, career support positively predicted women proteges' salaries (B = 0.05, p < 0.05; B = 0.03, ns), increased their managerial promotions (B = 0.06, p < 0.05; B = 0.05, p < 0.05), reduced their time since promotion (B = -0.15, p = 0.00; B = -0.12; p = 0.00), and increased their chance of promotion (B = 0.13, p = 0.00; B = 10; p = 0.00) and promotion last year (B = 0.11, p = 0.00; B = 10; p < 0.01), more than for men.

5.2 Hypotheses 2 and 3: Mentor Psychosocial Support

H2 proposed that mentor psychosocial support would be less likely than mentor career support to increase women's career advancement (H2a), and may not be related to it or may reduce it (H2b). For H2 to gain support, there would need to be nonsignificant or negative beta weights for psychosocial support (H2b), thus not explaining career advancement positively as did career support (H2a). H2 overall was supported. As tables 5 and 6 show for the regression equations run for women, supporting H2b, mentor psychosocial support either had nonsignificant links to career advancement (chance of promotion, promoted last year) or reduced career advancement (salary, managerial promotions, managerial level, time since promotion). Hence, supporting H2a, women's career advancement was not positively explained by psychosocial support, as it was by career support (tables 5, 6).

H3 proposed that mentor psychosocial support would be less related to gender differences in career advancement than would mentor career support. For H3 to be supported, the 2-way interactions between psychosocial support and protege gender needed to be nonsignificant or less frequently significant than for career support.

H3 was not supported. Protege gender interacted with psychosocial support to predict salary (table 4 only), managerial promotions, managerial level, time since promotion, chance of promotion (table 4 only), and promoted last year (table 3 only). The interactions explained significant increments in variance except for salary and, in table 4, managerial level. There were as many significant interactions for protege gender with psychosocial support as career support. The separate regressions run for men and women gave significant beta weights for psychosocial support for women and not men (with one exception), but not when predicting chance of promotion and promoted last year. A significant interaction effect for a variable shows the beta weights for men and women significantly differ from each other. The pair of beta weights that follow in the brackets give, first, the beta weight when Time 1 career advancement was not controlled (i.e. table 5) and second when it was (table 6). Beyond organisational, job, and individual factors, mentor psychosocial support reduced women proteges' salaries (B=-0.09, p = 0.00; B = -0.06, p = 0.001), managerial promotions (B = -0.12, p = 0.00; B 0.10, p = 0.00) and managerial level (B = -0.03, ns; B = -0.04, p< 0.05), and increased their time since promotion (B = 0.06, p < 0.05; B = 0.07, p < 0.05), more than for men.

In sum, mentor career support increased women's career advancement, especially their current and prospective promotion, more than men's. Career support did not increase men's advancement. Psychosocial support reduced women's advancement, and more than for men.

5.3 Hypothesis 4: Mentor Gender and Career Support

H4 proposed that career support would increase career advancement more for women proteges when the mentor is a male rather than a female. If H4 was supported, 3-way interactions for career support would be significant. To interpret a significant 3-way interaction, the regressions need to be rerun for the four groups: women with male or female mentors and men with male or female mentors. If H4 were supported, career support would predict career advancement more for women with male mentors than female mentors.

Not supporting H4, five of the six career advancement measures did not have significant 3-way interactions between protege gender, mentor gender, and career support (tables 3, 4). A significant 3-way interaction was found for promoted last year. The direction of the effect did not support H4 but favoured female mentors. The regressions were rerun separately for the four groups split by the gender of the mentor and protege. The first beta weight given in the bracket does not control for Time 1 promoted last year, the second does. Career support for women proteges from female mentors increased their promotion last year (B = 0.20, p = 0.00; B = 0.19, p = 0.001). This was more than career support for women proteges from male mentors, which either had no effect (B = 0.06, ns) or decreased their promotion last year (B = -0.08, p =0.02), and more than career support for male proteges from male mentors (B = 0.02, ns; B = 0.02, ns) or female mentors (B0.05, ns; B = -0.06, ns). An unhypothesized finding was that psychosocial support for women proteges from female mentors decreased their promotion last year (B = -0.13, p = 0.02; B = -0.13, p = 0.02) more than for the three other groups (women with male mentors [B = 0.02, ns; B = 0.03, ns], or men with male [B = 0.02, ns; B = 0.02, ns] or female mentors [B = 0.02, ns; B = 0.02, ns]).

The lack of 3-way interactions with mentor gender may be due to the unbalanced distributions in the mentor/protege gender groups and thus decreased statistical power (Ragins 1999). Although the 3-way interactions were not significant for the other outcomes, the results for the four groups indicated similar results for women with female mentors as for promoted last year, though not significantly more than for the other groups. Table 7 provides a summary of the results for the links of career or psychosocial support with career advancement for the four gender dyads. The positive links of career support and the negative links of psychosocial support consistently arose for women with female mentors, inconsistently for women with male mentors, and overall not for men with male or female mentors. For women proteges of female mentors, career support from their mentor significantly increased, whereas psychosocial support significantly decreased, their career advancement (e.g. managerial promotions, promotion last year, time since promotion). Similarly, for women proteges of male mentors, career support had positive links by increasing their salaries and reducing their time since promotion but, dissimilarly, had negative links by decreasing their managerial levels and chance of promotion. Similar to their counterparts with female mentors, psychosocial support from a male mentor significantly decreased women's salaries and managerial promotions. The gender of the mentor generally was not related to the link of career or psychological support to the career advancement for men.

6. Discussion

This study is the first to provide a comprehensive, systematic test assessing whether mentor career support, unlike psychosocial support, helps women advance in their careers more than men. The study is the first to examine the links between gender and mentoring and career advancement in which mentoring precedes career advancement and where comprehensive alternative explanations of career advancement are controlled in order to show whether mentoring is a unique driver of career outcomes, as Wanberg, Welsh and Hezlett (2003) advocate to be done.

The study makes two key contributions not found in previous studies. First, for this sample, supporting arguments based on propositions made about mentoring (Kram 1983) and difficulties women face for managerial advancement (Tharenou 1997), the results demonstrate how mentoring can especially help women advance compared to men--through their mentors' career support. The results support the oft-repeated views of scholars that mentors can be especially helpful for women to assist with their lack of hierarchical advancement compared to men's (Burke & McKeen 1990; Noe 1988; Ragins 1989, 1999; Wallace 2001). But only if the mentor support is career support. For this sample, mentor psychosocial support does not help, and may even reduce, women's advancement and more than men's. Second, the gender of the mentor helps explain how career and psychosocial support relate to women's career advancement. But it is female mentors, not male, that are most relevant. In this sample, career support from a female mentor most helps women to advance, but this needs to be balanced against the at-times negative influence of a female mentor's psychosocial support.

6.1 Explanations of the Findings

6.1.1 Links of Mentor Support to Women's Career Advancement For this sample, beyond relevant individual, job, and organisational control variables, their mentors' career support increases women's promotion more than men's, unlike the results of some past studies. This may be because past studies that have shown no gender differences in the effects of mentoring on career advancement have measured only the presence of a mentor, not the type of support provided (Corzine, Buntsman & Busch 1994; Fagenson 1989), or have combined career and psychosocial support (Dreher & Ash 1990). Scholars propose that mentor career support, not just the presence of a mentor, comprises actions that facilitate the protege's hierarchical advancement. Hence, the studies could not provide a theoretically-based test of the effect of mentor functions on career advancement (Kram 1985; Ragins 1989, 1997). The results of the present study support those of other studies that indicated that women may benefit more from mentor career support than men do (Bahniuk, Dobos & Kogler Hill 1990; Burt 1998; Johnson & Scandura 1994; Schor 1997).

Why does mentor career support help women advance more than men? Substantial evidence consistently shows that women have more obstacles to their hierarchical advancement than men do (Tharenou 1997, 1999). Hence, women may need sponsorship, challenging work, and coaching to gain promotion more than men do (Burt 1998; Ragins 1999). Career support for women from female mentors translates most into advancement, perhaps because women proteges gain from being sponsored, challenged and coached by someone like themselves who has incurred the particular difficulties women can face. For men, mentor support does not seem to matter as much to their advancement as for women (Bahniuk, Dobos & Kogler Hill 1990; Burt 1998; Schor 1997), and so the type of mentor support and the gender of the mentor are less critical than for women.

Psychosocial support does not help women's advancement in this sample and may reduce it, possibly explaining past negative links for women (Lyness & Thompson 2000; Metz & Tharenou 2001). It is reasonable to suggest, as shown by the present results, that psychosocial support may not help women climb the managerial hierarchy. Psychosocial support focuses on the woman's emotions and well-being, that is, inwardly, rather than outwardly to help her directly deal with and overcome obstacles in the external environment to her advancement, as sponsorship, coaching, and challenging assignments do. The links for psychosocial support appear to be particularly marked for women proteges, not men proteges. Men may simply not receive that level of psychosocial support because their obstacles do not prompt it.

The effects are heightened when both the protege and the mentor are women. The results indicate that the women of this sample with female mentors have both an advantage and a disadvantage with respect to career advancement. When they have a female mentor, the career support their mentor provides translates into promotion. This may be because, as some evidence suggests (Noe, Greenberger & Wang 2002; O'Neill 2002; Ragins 1999), women proteges learn more career strategies on how to advance from same gender mentors, which women need more than men need. However, it is not clear in this study why women do not gain more advancement from the career support of their male mentors. Ragins' (1999) argument was that male mentors provide more career support than do female mentors, and thus could advance their proteges more. In the present study, as in Ragins and Cotton's (1999), male mentors do not provide more career support than female mentors. Hence, the women of this sample may not have gained more from a male mentor, perhaps explaining the lack of superiority of having a male mentor in this study. Future research needs to examine why a female mentor's career support, rather than a male mentor's, enhances a female protege's career advancement. The results do not suggest that male mentors are more powerful advancers of others than female mentors, as has been suggested (Ragins 1999).

In contrast to career support, the psychosocial support women mentors provide to their female proteges reduces their promotion. This may be because the pairing provides the most psychosocial support of all pairs and the focus of psychosocial support does not help a protege's hierarchical advancement (Kram 1985; Ragins 1999). Women mentors who provide substantial psychosocial support could also be providing inappropriate role modelling to their female proteges. Future research, however, needs to examine why mentor psychosocial support, especially from a female mentor, may reduce women's career advancement. Women who receive more psychological support from their mentors may be different from other women (e.g. less independent) and elicit more psychosocial support from their female mentors or be attracted to female mentors who offer more psychosocial support than career support.

6.1.2 Links of Mentor Support to Men's Career Advancement By contrast to the women's advancement, mentor career and psychosocial support overall are not related to men's advancement for this sample. Some other studies have also found that mentor support was not related to men's managerial level, income, and promotion (Burt 1998; Johnson & Scandura 1994; Lyness & Thompson 2000). Prior studies with positive results for the presence of a mentor on men's levels have been cross-sectional (Kirchmeyer 1998) making comparison with the present longitudinal results difficult. By contrast, it should also be noted that, although mentor support is not related to men's career advancement for this sample, having a male mentor does increase men's managerial levels (as it does women's, tables 5, 6). Because most male proteges have a male mentor, mentoring does thus help men advance in their careers compared to having a female mentor.

Why is mentor support not related to the advancement of the men in the present study? First, the results for the prediction of men's advancement, as shown in tables 5 and 6, indicate that men are likely to advance irrespective of support from a mentor, as found (Bahniuk, Dobos & Kogler Hill 1990; Burt 1998; Schor 1997). The results for men show that the type of organisation (the private sector, larger organisations) and men's human capital (organisational tenure, training and development, masculine gender role traits) most explain men's advancement. Second, the results for the total sample (tables 3, 4) show that, even after rendering the samples as comparable as possible, men still advance more than women. The results are consistent with the strong evidence for a male advantage for hierarchical advancement (Tharenou 1997), though it is possible the incremental explanation by gender is due to some other, uncontrolled factor. However, overall, the men of this sample appear not to advance from mentor career support as do the women. The reasons appear to be that the men advance from organisational and human capital factors and perhaps from being men, and need mentor career support less than women do to advance.

For the total sample, mentoring does enhance advancement when the mentor is male rather than female, which is the common situation for men. Mentor and protege gender significantly interact to predict some of the advancement measures (less time since promotion, promotion more last year) (tables 3, 4). In diversified mentoring theory, Ragins (1999) suggested that the most advantageous pairing for protege outcomes is a male mentor with a male protege, because of the power of men mentors and the advantages of gender similarity. With respect to the significant interactions, the results (tables 5, 6) indicate that men who have male mentors advance more in terms of having less time since promotion and being promoted more last year than women do. In addition, as shown in tables 5 and 6, male proteges (as well as female) of male mentors have higher managerial levels a year later than proteges of female mentors. Hence, mentoring can help the men of this sample advance. Unlike women, it is not through specific career support actions, which men may not need, but through an alliance with a male senior person.

6.5 Limitations, Future Research, and Practice

This study has several limitations, including how generalizable the results are. Respondents were chiefly from the public sector and finance and business service industries. It is possible that the positive links of career support for women more than men arose because of the industries surveyed. The public sector and finance industries have been found worldwide to promote interventions designed to assist women in their advancement into management more than other industries do (International Labour Organization 1997; Wirth 2001). Mentoring may therefore increase the advancement of women in the public sector and finance industries, and female mentors may have more power than in other sectors. Such schemes may not be especially designed to help men advance in those industries, hence not affecting their advancement. The positive links of career support for women more than men proteges may not occur in other industries that do not facilitate women's advancement as much. The study's results thus require replication in other industries.

There may be alternative explanations for the results. The greater link of mentors' career support to women's than men's advancement may be because women report more support and give more credit to career help than men do (Ragins 1999). The career support men gain from mentors will thus not be as related to their advancement as to women's.

The nonsignificant and negative links for mentor psychosocial support may occur because the psychosocial support received was of low quality. The emphasis placed on defining a mentor for respondents as one who plays a committed role in their career may have caused the psychosocial aspects to be downplayed and contributed to the findings about psychosocial support. There was no measure of protege satisfaction with the psychosocial support to assess the quality of the psychosocial support received. It is possible that women who received psychosocial support were less likely to advance because they had poor quality psychosocial support. If the protege was part of a formal mentoring program, the psychosocial support may have been of low quality because the nature of the program was about career advancement and not a personal relationship. However, whether the organisation had a formal mentor program was controlled in the analyses. Future research needs to measure the protege's satisfaction with their mentor psychosocial support, the quality of the support received, and whether the mentoring they gained was part of a formal program to explain the link between psychosocial support and women's career advancement.

Moreover, the negative links for psychosocial support and positive links for career support may reflect women's advancement affecting the mentor support they receive as well as support affecting their advancement. This study did not examine reverse effects. Mentors may respond with more psychosocial support to women who are not advancing, resulting in negative links. Studies have shown that proteges with higher compared to lower ability and potential are selected more by mentors and gain more career support (Allen, Poteet & Russell 2000; Green & Bauer 1995; Olian, Carroll & Giannantonio 1993; Ragins, Townsend & Mattis 1998). Ability and potential may be reflected in current and recent promotion and so affect who gains mentor support and of what kind, thus partly explaining the positive links of mentor career support with women's advancement. Moreover, women who need career development and seek that from their mentors may be different from the women who need psychosocial support. They may be more likely to be independent and not need their mentors as much as the women who report their mentors provided friendship and counselling. Bidirectional tests are needed to examine if women's advancement or other characteristics affect the type of mentor support they receive and partly explain the positive links for career support and negative links for psychosocial support.

Methodological issues also limit the findings. The study is self-report, and thus reliant on perceptions. Even though there was substantial evidence for their validity, the measures of career advancement were not objective.

Moreover, some of the mentors were past and others current. Hence, the time frame of the study is disrupted because Time 1 responses may have referred to the time the data were collected or some unknown time in the past. Posthoc analyses were conducted, repeating the regressions for the proteges whose mentors were their current boss versus a past boss. The results showed that mentor career support from immediate bosses was more related to women's advancement than from past bosses. Psychosocial support was negatively related to advancement for women mentors whether bosses were current or past, not suggesting any effect. Hence, for the total sample, there may have been lower magnitudes for the results for mentor career support than if the mentors were all current mentors.

Conceptually, future research needs to examine the reasons career support helps, and psychosocial support may hinder, women's advancement, especially when their mentors are female. Although this study followed propositions that argued for the influence of the two types of support according to the specific functions they provide, examination is needed of the exact causes of the effects found. The role that psychosocial support plays for women's advancement especially needs examination as it is emphasized in some studies of senior women as important to their self-confidence and persistence to advance (Ragins, Townsend & Mattis 1998).

This study is the first longitudinal study to examine whether there are gender differences in the prediction by mentor support and gender of proteges' career advancement. It takes into account possible alternative or spurious explanations of gender differences in career advancement by controlling relevant organisational, job, and individual characteristics to show the unique contribution of mentoring and its relative importance. For this sample, mentor career support explains women's career advancement more than it does men's and mentor psychosocial support either does not assist women's advancement, as for men's, or has a negative effect. The effects are strongest for women with female mentors. Male mentors may also help their proteges (male or female) advance more than female mentors do.

What recommendations can be made for proteges, mentors, and organisations from these results? Australian women still form only a low 25% of managers and administrators, much the same as 30 years ago (Australian Bureau of Statistics, 2003). Results from a census of Australian women in leadership revealed women executive managers have reached 8.8% (up 0.4% from 2002) and the number of women board directors has increased from 8.2% in 2002 to 8.4% in 2003 (EOWA, 2003). Thus there is still much work to be done in promoting the careers of women in leadership.

With respect to women proteges, women need to seek career support from their mentors in order to advance. When women who seek advancement choose female mentors, they need to capitalize on the career support they can provide, but perhaps be wary of the psychosocial support, which may be less helpful to their advancement. Proteges may be best to have a network of diverse mentors as suggested by scholars (e.g. Higgins & Kram 2001; Janasz, Sullivan & Whiting 2003), because both female mentors who provide career support and male mentors in general can help women advance. With respect to recommendations for mentors, to assist women to advance, mentors need to focus on career support for their proteges. If the protege is seeking to advance and the mentoring relationship is focused on helping the protege advance, women mentors need to not focus on psychosocial support, even though it may have positive effects on their protege's well-being. With respect to recommendations for organisations, providing mentor career support will have small positive effects on women's advancement, and, as shown (table 6), perhaps comparably to women increasing their human capital (training and development) and working in high skill level occupations and in the public sector. Hence, mentor career support should be one of a number of developmental initiatives encouraged to assist women to advance. Overall, the results suggest that mentor career support can help women's advancement more than men's and that mentor career support is a useful strategy to help women's hierarchical advancement, though its effects will be small.

(Date of receipt of final transcript: October 7, 2004. Accepted by Sharon Parker, Area Editor.)

Table 1

Description of Sample at Time 1 by Percentage Frequencies and
Means on Demographic Items

Item                               Total    Women    Men

Age [chi square] (7) = 133.94
  20-24                            14         8       6
  25-29                            21        12       9
  30-34                            18         9       9
  35-39                            17         8       9
  40-44                            14         6       8
  45-49                             9         4       5
  50-54                             5         2       3
  55-59 (none 60 or over)           2         1       1
  X                                 3.52      3.27    3.77
Fulltime work tenure [chi square] (7) = 504.21
  < 5 years                        16         9       7
  5 up to 10 years                 24        15       9
  10 up to 15 years                21        12       9
  15 up to 20 years                14         6       8
  20 up to 25 years                11         4       7
  25 up to 30 years                 7         2       5
  30 or more years                  7         2       5
  X                                 3.28      2.8     3.77
Organisation tenure [chi square] (6) = 430.31
  < 5 years                        43        24      19
  5 up to 10 years                 27        15      12
  10 up to 15 years                13         6       7
  15 up to 20 years                 7         2       5
  20 up to 25 years                 5         1       4
  25 up to 30 years                 3         1       2
  30 or more                        2         0       2
  X                                 2.21      1.82    2.58
Spouse [chi square] (1) = 33.72
  Spouse                           65        30      35
  No spouse                        35        19      16
  X                                 1.35      1.38    1.31
Dependent children [chi square] (4) = 236.65
  None                             59        33      26
  One or more                      14         7       7
  Two                              18         7      11
  Three                             7         2       5
  Four or more                      2         0       2
  X                                 2.05      1.58    2.02
Salary [chi square] (4) = 449.18
  $15,001-25,000                   18        13       5
  $25,001-35,000                   40        20      20
  $35,001-45,000                   22        10      13
  $45,001-55,000                   12         5       7
  $55,001 and over                  8         2       6
  X                                 3.57      3.23    3.91
Number subordinates [chi square] (1) = 125.27
  None                             52        29      23
  1 to 5                           43        19      24
  More than 6                       5         2       3
  X                                 1.66      1.5     1.71
Education level [chi square] (6) = 61.25
  Some secondary school             2         1       1
  10 years completed               16         9       7
  12 years completed               32        14      18
  Technical college course/
  Or Associate diploma/ diploma    16         8       8
  Undergraduate degree             22        11      11
  Honours/graduate diploma          8         5       3
  Masters degree, PhD               4         2       2
  X                                 4.58      4.6     4.56
Organisation size [chi square] (4) = 273.47
  Up to 1000                       32        17      15
  1001-2000                         6         3       3
  2001-4000                        11         6       5
  4001-8000                         8         4       4
  >8000                            42        19      23
  X                                 3.2       3.1     3.33
Industry [chi square] (1) = 28.49
  Finance, property & business     38        18      21
  Public administration            49        25      24
  Community services                8         5       3
  Other                             5         2       3
Employer sector [chi square] (1) = 28.49
  Public sector                    59        30      29
  Private sector                   41        18      22
  X                                 1.41      1.37    1.44
Occupation type [chi square] (4) = 168.39
  Managers                         19         7      12
  Professionals                    30        15      15
  Paraprofessionals                13         6       7
  Clerks                           35        19      16
  Other                             3         2       1
  X                                 3.06      3.32    2.96
Organisation provides formal mentor program
  [chi square] (6) = 32.84
  Not at all                       17         9       8
  Almost never                     17         8       9
  Seldom                           21        11      10
  Sometimes                        27        13      14
  Frequently                       12         5       7
  Almost always                     4         2       2
  Always                            2         1       1
  X                                 3.21      3.17    3.20
Managerial level [chi square] (4) = 198.53
  Subordinates                     41        24      17
  Supervisors                      21        10      11
  Lower managers                   14         5       9
  Middle managers                  16         7       9
  Senior managers, executives       8         3       5
  X                                 2.33      2.11    2.54
Length of mentor relationship [chi square] (5) = 46.89
  Up to a year                     21        12       9
  1 up to 2 years                  27        13      14
  2 up to 3 years                  20        10      10
  3 up to 4 years                   9         5       5
  4 up to 5 years                   6         3       3
  5 or more years                  17         7      10
  X                                 4.00      3.85    4.14

Note: Chi-square tests indicated there were significant differences
between men and women on all variables at p = 0.00. Some categories
in items have been combined.

Table 2

Means, Standard Deviations, Alpha Coefficients, and
Correlations Between the Variables

                                         Correlations

    Variable                M       SD      1       2       3

Time 1

 1. Sector                   1.41    0.49
 2. Organisation size        3.21    1.75    0.21
 3. Age                      4.52    1.86   -0.16   -0.13
 4. Tenure                   2.21    1.51    0.10    0.20    0.45
 5. Occupation level         3.06    1.61    0.27    0.21   -0.28
 6. Education                4.58    2.35   -0.23   -0.20    0.03
 7. Training                 3.83    1.62   -0.32   -0.17    0.42
 8. Masculinity              4.91    0.92    0.10   -0.01    0.04
 9. Formal mentor program    3.21    1.52    0.08    0.13    0.01
10. Years relationship       3.06    1.73    0.00   -0.10    0.30
11. Mentor level             1.37    0.48    0.03    0.03    0.00
12. Career support           4.80    1.07    0.06   -0.05   -0.09
13. Psychosocial support     5.25    0.94    0.08   -0.01   -0.07
14. Mentor gender            1.25    0.43   -0.18   -0.05   -0.10
15. Protege gender           1.51    0.50    0.07    0.07    0.14
16. Salary                   3.58    1.41    0.00   -0.14    0.36
17. Promotions               2.21    1.14    0.04    0.03    0.46
18. Managerial level         2.33    1.52    0.02   -0.14    0.37
19. Time since promotion     2.39    1.07   -0.06   -0.02    0.33
20. Chance promotion         3.08    1.31    0.04    0.11   -0.31
21. Promotion last year      1.24    0.43    0.02    0.06   -0.17

Panel B: Time 2

22. Salary                   3.96    1.42    0.11   -0.11    0.25
23. Promotions               2.41    1.13    0.10    0.04    0.41
24. Managerial level         2.64    1.61    0.11   -0.12    0.28
25. Time since promotion     2.51    1.10   -0.11   -0.06    0.33
26. Chance promotion         2.95    1.30    0.07    0.10   -0.35
27. Promotion last year      1.24    1.43    0.06    0.10   -0.23

                                         Correlations

    Variable                4       5       6       7       8

Time 1

1.  Sector
2.  Organisation size
3.  Age
4.  Tenure
5.  Occupation level        -0.11    0.93
6.  Education               -0.25   -0.35
7.  Training                 0.17   -0.51    0.20    0.79
8.  Masculinity             -0.01   -0.12    0.02    0.22    0.87
9.  Formal mentor program    0.07    0.00   -0.02    0.06    0.04
10. Years relationship       0.24   -0.17   -0.03    0.28    0.11
11. Mentor level            -0.02    0.04   -0.02   -0.02    0.06
12. Career support          -0.09   -0.04    0.00    0.10    0.15
13. Psychosocial support    -0.05   -0.04   -0.01    0.09    0.18
14. Mentor gender           -0.18    0.08    0.01   -0.06   -0.08
15. Protege gender           0.25   -0.14   -0.01    0.13    0.04
16. Salary                   0.24   -0.54    0.31    0.52    0.18
17. Promotions               0.37   -0.36   -0.02    0.52    0.21
18. Managerial level         0.23   -0.51    0.18    0.57    0.26
19. Time since promotion     0.37   -0.02   -0.05    0.09   -0.09
20. Chance promotion        -0.21   -0.03    0.10    0.03    0.18
21. Promotion last year     -0.12    0.05   -0.02   -0.05    0.07

Panel B: Time 2

22. Salary                   0.15   -0.49    0.31    0.45    0.20
23. Promotions               0.28    0.30   -0.05    0.52    0.25
24. Managerial level         0.15   -0.46    0.16    0.53    0.29
25. Time since promotion     0.36   -0.08   -0.04    0.12   -0.10
26. Chance promotion        -0.24    0.01    0.05   -0.03    0.16
27. Promotion last year     -0.17    0.16   -0.01   -0.15    0.05

                                         Correlations

    Variable                9       10      11      12      13

Time 1

1.  Sector
2.  Organisation size
3.  Age
4.  Tenure
5.  Occupation level
6.  Education
7.  Training
8.  Masculinity
9.  Formal mentor program
10. Years relationship      -0.01
11. Mentor level            -0.01    0.09
12. Career support           0.16    0.05   -0.19    0.85
13. Psychosocial support     0.14    0.18   -0.04    0.54    0.91
14. Mentor gender           -0.05   -0.12    0.01    0.00    0.08
15. Protege gender           0.03    0.08    0.01   -0.08   -0.11
16. Salary                   0.05    0.26   -0.01    0.05    0.04
17. Promotions               0.08    0.24    0.00    0.04    0.05
18. Managerial level         0.06    0.28   -0.01    0.07    0.11
19. Time since promotion    -0.07    0.21   -0.02   -0.17   -0.12
20. Chance promotion         0.14   -0.13    0.02    0.17    0.16
21. Promotion last year      0.06   -0.09    0.02    0.10    0.08

Panel B: Time 2

22. Salary                   0.05    0.20   -0.04    0.06    0.04
23. Promotions               0.06    0.22    0.01    0.04    0.05
24. Managerial level         0.03    0.25    0.01    0.05    0.09
25. Time since promotion    -0.05    0.19   -0.02   -0.13   -0.09
26. Chance promotion         0.11   -0.16    0.02    0.15    0.11
27. Promotion last year      0.03   -0.13    0.04    0.06    0.04

                                         Correlations

    Variable                14      15      16      17      18

Time 1

1.  Sector
2.  Organisation size
3.  Age
4.  Tenure
5.  Occupation level
6.  Education
7.  Training
8.  Masculinity
9.  Formal mentor program
10. Years relationship
11. Mentor level
12. Career support
13. Psychosocial support
14. Mentor gender
15. Protege gender          -0.24
16. Salary                  -0.19    0.24
17. Promotions              -0.17    0.25    0.48
18. Managerial level        -0.15    0.14    0.63    0.56
19. Time since promotion    -0.06    0.07    0.09    0.05    0.05
20. Chance promotion        -0.02    0.10    0.04    0.00    0.02
21. Promotion last year      0.04   -0.02   -0.07   -0.04   -0.03

Panel B: Time 2

22. Salary                  -0.19    0.19    0.86    0.41    0.58
23. Promotions              -0.17    0.20    0.45    0.76    0.54
24. Managerial level        -0.15    0.08    0.58    0.50    0.83
25. Time since promotion    -0.05    0.08    0.12    0.07    0.09
26. Chance promotion        -0.01    0.07   -0.02   -0.03   -0.05
27. Promotion last year      0.03   -0.03   -0.15   -0.10   -0.15

                                         Correlations

    Variable                19      20      21      22      23

Time 1

1.  Sector
2.  Organisation size
3.  Age
4.  Tenure
5.  Occupation level
6.  Education
7.  Training
8.  Masculinity
9.  Formal mentor program
10. Years relationship
11. Mentor level
12. Career support
13. Psychosocial support
14. Mentor gender
15. Protege gender
16. Salary
17. Promotions
18. Managerial level
19. Time since promotion
20. Chance promotion        -0.36
21. Promotion last year     -0.53    0.22

Panel B: Time 2

22. Salary                   0.06    0.08   -0.06
23. Promotions               0.02    0.02   -0.02    0.44
24. Managerial level         0.03    0.03   -0.02    0.60    0.53
25. Time since promotion     0.54   -0.40   -0.28    0.05    0.02
26. Chance promotion        -0.33    0.61    0.18    0.04    0.01
27. Promotion last year     -0.19    0.28    0.16   -0.08   -0.06

                                    Correlations

    Variable                24      25      26      27

Time 1

1.  Sector
2.  Organisation size
3.  Age
4.  Tenure
5.  Occupation level
6.  Education
7.  Training
8.  Masculinity
9.  Formal mentor program
10. Years relationship
11. Mentor level
12. Career support
13. Psychosocial support
14. Mentor gender
15. Protege gender
16. Salary
17. Promotions
18. Managerial level
19. Time since promotion
20. Chance promotion
21. Promotion last year

Panel B: Time 2

22. Salary
23. Promotions
24. Managerial level
25. Time since promotion     0.01
26. Chance promotion        -0.01   -0.38
27. Promotion last year     -0.05   -0.57    0.27     --

Note: Correlations of 0.04 are significant at p < 0.01 and of 0.05 are
significant at p < 0.001. Alpha coefficients are in the diagonal.
Dashes indicated where alpha coefficients could not be calculated.

Table 3

Prediction by Time 1 Mentoring and Gender of Time 2 Career Advancement
a Year Later

                              Salary                   Promotions

                                   [DELTA]                  [DELTA]
Variable                 B        [R.sup.2]       B        [R.sup.2]

Sector                 0.25 ***                 0.19 ***
Organisation size     -0.02                     0.10 ***
Age                    0.00                     0.20 ***
Tenure                 0.03                    -0.02
Managerial level       0.28 ***                 0.25 ***
Occupation level      -0.22 ***                -0.03
Education              0.21 ***                -0.12 ***
Training               0.19 ***                 0.36 ***
Masculinity            0.03                     0.08 ***
Formal mentor
  program              0.01                    -0.01
Mentor level          -0.03 *                   0.00
Years relationship     0.02                    -0.02
[DELTA] [R.sup.2]                 0.474 ***                0.461 ***
Career support
  (CS)                 0.03                     0.00
Psychosocial
  support (PS)        -0.06 ***                -0.02
Mentor gender (MG)    -0.05 ***                -0.04 *
[DELTA] [R.sup.2]                 0.006 ***                0.002 **
Protege gender
   (PG)                0.06 ***                 0.06 ***
[DELTA] [R.sup.2]                 0.003 ***                0.003 ***
CS x protege
  gender              -0.01                    -0.05 **
PS x protege
  gender               0.03                     0.08 ***
MG x protege          -0.02                    -0.02
  gender
[DELTA] [R.sup.2]                 0.001                    0.004 ***
CS x PG x MG           0.02                     0.00
[DELTA] [R.sup.2]                 0.000                    0.000
PS x PG x MG          -0.03                    -0.01
[DELTA] [R.sup.2]                 0.000                    0.000
[R.sup.2]                         0.484 ***                0.470 ***
Df                                213,125                  213,141

                              Managerial Level            Time since
                                                          Promotion

                                   [DELTA]                  [DELTA]
Variable                 B        [R.sup.2]       B        [R.sup.2]

Sector                 0.33 ***                -0.08 ***
Organisation size     -0.04 *                  -0.08 ***
Age                    0.06 ***                 0.17 ***
Tenure                -0.00                     0.30 ***
Managerial level       --                      -0.02
Occupation level      -0.28 ***                 0.03
Education              0.04                     0.00
Training               0.42 ***                -0.02
Masculinity            0.14 ***                -0.08 ***
Formal mentor
  program             -0.01                    -0.03
Mentor level          -0.01                    -0.02
Years relationship     0.05 ***                 0.09 ***
[DELTA] [R.sup.2]                 0.457 ***                0.198 ***
Career support
  (CS)                -0.06 ***                -0.06 **
Psychosocial
  support (PS)         0.02                    -0.01
Mentor gender (MG)    -0.05 **                  0.01
[DELTA] [R.sup.2]                 0.003 ***                0.005 ***
Protege gender
   (PG)               -0.06 ***                 0.06 *
[DELTA] [R.sup.2]                 0.003 ***                0.000
CS x protege
  gender              -0.01                     0.06 **
PS x protege
  gender               0.05 **                 -0.07 ***
MG x protege          -0.01                     0.04 *
  gender
[DELTA] [R.sup.2]                 0.002 *                  0.004 ***
CS x PG x MG          -0.03                     0.03
[DELTA] [R.sup.2]                 0.000                    0.000
PS x PG x MG           0.01                    -0.03
[DELTA] [R.sup.2]                 0.000                    0.000
[R.sup.2]                         0.465 ***                0.208 ***
Df                                203,133                  213,115

                                                        Promotion Last
                              Chance Promotion               Year

                                   [DELTA]                  [DELTA]
Variable                 B        [R.sup.2]       B        [R.sup.2]

Sector                 0.02                    -0.00
Organisation size      0.09 ***                 0.08 ***
Age                   -0.30 ***                -0.11 ***
Tenure                -0.15 ***                -0.10 ***
Managerial level      -0.01                    -0.03
Occupation level      -0.05 *                   0.07 ***
Education              0.01                     0.02
Training               0.08 ***                -0.04
Masculinity            0.13 ***                 0.07 ***
Formal mentor
  program              0.08 ***                 0.02
Mentor level           0.03 *                   0.04
Years relationship    -0.09 ***                -0.04
[DELTA] [R.sup.2]                 0.192 ***                0.083 ***
Career support
  (CS)                 0.07 **                  0.04
Psychosocial
  support (PS)         0.04                     0.02
Mentor gender (MG)    -0.02                     0.01
[DELTA] [R.sup.2]                 0.007 ***                0.003 *
Protege gender
   (PG)                0.13 ***                 0.03
[DELTA] [R.sup.2]                 0.015 ***                0.001
CS x protege
  gender              -0.05 *                  -0.06 **
PS x protege
  gender               0.04                     0.05 *
MG x protege          -0.04 *                   0.00
  gender
[DELTA] [R.sup.2]                 0.002 *                  0.002 *
CS x PG x MG           0.03                    -0.05 *
[DELTA] [R.sup.2]                 0.000                    0.001
PS x PG x MG           0.03                     0.06 *
[DELTA] [R.sup.2]                 0.000                    0.002 *
[R.sup.2]                         0.217 ***                0.091 ***
Df                                213,122                  213,155

Note: * p < 0.05; ** p < 0.01; and *** p < 0.001.

Table 4

Prediction by Time 1 Mentoring and Gender of Change in Career
Advancement from Time 1 to Time 2

                               Salary                  Promotions

                                     [DELTA]                   [DELTA]
     Variable            B          [R.sup.2]      B          [R.sup.2]

Time 1 dependent
  variable              0.79 ***                  0.58 ***
[DELTA][R.sup.2]                    0.736 ***                 0.584 ***
Scctor                  0.15 ***                  0.12 ***
Organisation size       0.01                      0.04 ***
Age                    -0.05 ***                  0.08 ***
Tenure                 -0.04 ***                 -0.07 ***
Managerial level        0.03                      0.10 ***
Occupation level       -0.06 ***                  0.01
Education               0.05 ***                 -0.07 ***
Training                0.07 ***                  0.20 ***
Masculinity             0.02 **                   0.05 ***
Formal mentor
  program               0.00                     -0.02
Mentor level           -0.03 ***                  0.00
Years relationship     -0.01                     -0.00
[DELTA][R.sup.2]                    0.027 ***                 0.053 ***
Career support (CS)    -0.00                     -0.00
Psychosocial
  support (PS)         -0.03 *                   -0.02
Mentor gender (MG)     -0.01                     -0.02
[DELTA][R.sup.2]                    0.001 *                   0.001 *
Protege gender         -0.01                      0.01
[DELTA][R.sup.2]                    0.000                     0.000
CS x protege gender    -0.02 *                   -0.04 **
PS x protege gender     0.03 *                    0.06 ***
MG x protege gender    -0.00                      0.01
[DELTA][R.sup.2]                    0.000                     0.002 ***
CS x PG x MG            0.01                      0.01
[DELTA][R.sup.2]                    0.000                     0.000
PS x PG x MG           -0.02                     -0.01
[DELTA][R.sup.2]                    0.000                     0.000
[R.sup.2]                           0.764 ***                 0.640 ***
df                                  22,3124                   22,3140

                          Managerial Level        Time since Promotion

                                     [DELTA]                   [DELTA]
     Variable            B          [R.sup.2]      B          [R.sup.2]

Time 1 dependent
  variable              0.72 ***                  0.42 ***
[DELTA][R.sup.2]                    0.695 ***                 0.286 ***
Scctor                  0.16 ***                 -0.07 ***
Organisation size      -0.01                     -0.06 ***
Age                    -0.02                      0.10 ***
Tenure                 -0.05 ***                  0.17 ***
Managerial level        --                        0.01
Occupation level       -0.07 ***                 -0.00
Education               0.00                      0.00
Training                0.15 ***                 -0.02
Masculinity             0.06 ***                 -0.05 ***
Formal mentor
  program              -0.02 *                   -0.01
Mentor level            0.00                     -0.01
Years relationship      0.01                      0.03
[DELTA][R.sup.2]                    0.031 ***                 0.054 ***
Career support (CS)    -0.04 **                  -0.03
Psychosocial
  support (PS)         -0.01                     -0.00
Mentor gender (MG)     -0.02 *                    0.01
[DELTA][R.sup.2]                    0.001 ***                 0.001
Protege gender         -0.06 ***                  0.01
[DELTA][R.sup.2]                    0.003 ***                 0.000
CS x protege gender    -0.01                      0.06 ***
PS x protege gender     0.03 *                   -0.06 ***
MG x protege gender     0.00                      0.03 *
[DELTA][R.sup.2]                    0.000                     0.003 ***
CS x PG x MG           -0.01                      0.02
[DELTA][R.sup.2]                    0.000                     0.000
PS x PG x MG            0.01                     -0.03
[DELTA][R.sup.2]                    0.000                     0.000
[R.sup.2]                           0.731 ***                 0.345 ***
df                                  21,3132                   22,3114

                          Chance Promotion        Promotion Last Year

                                     [DELTA]                   [DELTA]
     Variable            B          [R.sup.2]      B          [R.sup.2]

Time 1 dependent
  variable             -0.52 ***                  0.11 ***
[DELTA][R.sup.2]                    0.376 ***                 0.024 ***
Scctor                  0.03                     -0.00
Organisation size       0.03 *                   -0.08 ***
Age                    -0.15 ***                 -0.10 ***
Tenure                 -0.09 ***                 -0.10 ***
Managerial level       -0.03                     -0.04
Occupation level        0.04                      0.07 **
Education               0.02 *                    0.02
Training                0.04 *                   -0.03
Masculinity             0.06 ***                  0.06 ***
Formal mentor
  program               0.03 *                    0.02
Mentor level            0.02                      0.04 *
Years relationship     -0.04 **                  -0.04 *
[DELTA][R.sup.2]                    0.040 ***                 0.070 ***
Career support (CS)     0.04                      0.01
Psychosocial
  support (PS)          0.01                      0.00
Mentor gender (MG)     -0.01                      0.03
[DELTA][R.sup.2]                    0.001                     0.001
Protege gender          0.05 ***                 -0.06 **
[DELTA][R.sup.2]                    0.002 ***                 0.001
CS x protege gender    -0.05 **                   0.05 *
PS x protege gender     0.04 *                   -0.00
MG x protege gender    -0.03                     -0.05 *
[DELTA][R.sup.2]                    0.002 **                  0.002 *
CS x PG x MG           -0.02                     -0.05 *
[DELTA][R.sup.2]                    0.000                     0.001
PS x PG x MG            0.03                     -0.06 *
[DELTA][R.sup.2]                    0.000                     0.002 *
[R.sup.2]                           0.422 ***                 0.101 ***
df                                  22,3121                   22,3154

Note: * p < 0.05; ** p < 0.01; and *** p < 0.001.

Table 5

Prediction by Time 1 Mentoring and Gender of Time 2 Career Advancement
a Year Later for Women and Men Separately

                                 B                         B
                               Salary                  Promotions

Variable               W            M            W            M
Sector                  0.11 ***     0.33 ***     0.13 ***     0.21 ***
Organisation size       0.00        -0.02         0.07 ***     0.13 ***
Age                    -0.01         0.01         0.11 ***     0.30 ***
Tenure                  0.04         0.03        -0.01        -0.07 **
Managerial level        0.28 ***     0.28 ***     0.29 ***     0.23 ***
Occupation level       -0.27 ***    -0.17 ***    -0.07 *       0.00
Education               0.19 ***     0.23 ***    -0.13 ***    -0.12 ***
Training                0.16 ***     0.20 ***     0.40 ***     0.32 ***
Masculinity             0.03         0.04 *       0.09 ***     0.08 ***
Formal mentor
  program               0.02        -0.01        -0.02        -0.01
Mentor level           -0.01        -0.04         0.01        -0.04 *
Years relationship      0.02         0.02         0.02         0.00
[DELTA][R.sup.2]        0.466 ***    0.486 ***    0.423 ***    0.483 **
Career support (CS)     0.05 *       0.02         0.06 *      -0.03
Psychosocial
  support (PS)         -0.09 ***    -0.05        -0.12 ***     0.05
Mentor gender (MG)     -0.05 **     -0.04        -0.04        -0.03
[DELTA][R.sup.2]        0.008 ***    0.003 *      0.009 ***    0.002
[R.sup.2]               0.474 ***    0.489 ***    0.433 ***    0.483 **
Df                     15,1557      15,1563      15,1562      15,1574

                                 B                         B
                          Managerial level        Time since promotion

Variable               W            M            W            M
Sector                  0.31 ***     0.33 ***    -0.02         0.14 ***
Organisation size      -0.00        -0.07 ***    -0.07 **     -0.08 ***
Age                     0.06 **      0.06 *      -0.15 ***     0.19 ***
Tenure                 -0.02         0.02         0.20 ***     0.35 ***
Managerial level        --           --          -0.01        -0.01
Occupation level       -0.36 ***    -0.20 ***     0.01         0.03
Education               0.03         0.05 *       0.02         0.02
Training                0.42 ***     0.39 ***    -0.05        -0.00
Masculinity             0.12 ***     0.16 ***    -0.08 ***     0.07 **
Formal mentor
  program              -0.03         0.01         0.00        -0.06 *
Mentor level            0.01        -0.02        -0.04        -0.01
Years relationship      0.04 *       0.06 **     -0.11 ***     0.05 *
[DELTA][R.sup.2]        0.452 ***    0.464 **     0.128 ***   0.258 ***
Career support (CS)    -0.05        -0.06 *       0.15 ***    -0.01
Psychosocial
  support (PS)         -0.03         0.06 *       0.06 *      -0.06
Mentor gender (MG)     -0.04 *      -0.04 *      -0.02         0.02
[DELTA][R.sup.2]        0.006***     0.004 **     0.014***     0.003
[R.sup.2]               0.458***     0.468 **     0.142***     0.262 ***
Df                     14,1560      14,1569      15,1546      15,1564

                                                           B
                                 B                   Promotion last
                          Chance promotion              year (a)

Variable               W            M            W            M
Sector                 -0.06 *       0.08 **     -0.07 *       0.05
Organisation size       0.09 ***     0.09 ***     0.06 *       0.10 ***
Age                     0.26 ***    -0.35 ***    -0.11 ***    -0.12 ***
Tenure                  0.10 ***    -0.16 ***    -0.06 *      -0.15 ***
Managerial level       -0.01        -0.01        -0.05        -0.03
Occupation level        0.04        -0.04         0.11 ***     0.05
Education               0.00         0.01         0.04         0.00
Training               -0.11 **      0.06 *      -0.03        -0.03
Masculinity            -0.13 ***     0.13 ***     0.06 *       0.08 **
Formal mentor
  program              -0.05 *       0.10 ***     0.01         0.03
Mentor level           -0.02         0.05 *       0.02         0.06 *
Years relationship      0.11 ***    -0.06 *      -0.08 **     -0.00
[DELTA][R.sup.2]        0.159 ***    0.260 ***    0.077 ***    0.102 ***
Career support (CS)    -0.13 ***     0.03         0.11 ***    -0.00
Psychosocial
  support (PS)         -0.01         0.04        -0.02         0.02
Mentor gender (MG)     -0.00        -0.04        -0.02         0.03
[DELTA][R.sup.2]        0.015 ***    0.004 *      0.009 ***    0.001
[R.sup.2]               0.175 ***    0.265 ***    0.086 ***    0.103 ***
Df                     15,1555      15,1562      15,1572      15,1578

Note: W = women, M = men; and
* p < 0.05; ** p < 0.01; and *** p < 0.001.

Table 6

Prediction by Time1 Mentoring and Gender of Change in Career
Advancement from Time 1 to Time 2 for Women and Men Separately

                                 B                         B

                               Salary                  Promotions

     Variable              W            M            W            M

Time 1 dependent
  variable              0.79 ***     0.76 ***     0.57 ***     0.56 ***
[DELTA][R.sup.2]        0.729 ***    0.726 ***    0.553 ***    0.579 ***
Sector                  0.12 ***     0.15 ***     0.10 ***     0.12 ***
Organisation size       0.04 *      -0.01         0.04 *       0.06 ***
Age                    -0.02        -0.07 ***     0.04         0.13 ***
Tenure                 -0.02        -0.02        -0.04 *      -0.09 ***
Managerial level        0.03         0.04 *      -0.12 ***     0.10 ***
Occupation level       -0.11 ***    -0.03 *      -0.05 *       0.04 *
Education               0.02         0.08 ***    -0.10 ***    -0.06 ***
Training                0.04 *       0.07 ***     0.18 ***     0.19 ***
Masculinity             0.02         0.04 **      0.06 ***     0.05 **
Formal mentor
  program               0.01        -0.01        -0.01        -0.02
Mentor level           -0.02        -0.04 **      0.02         0.01
Years relationship     -0.02        -0.01         0.01        -0.03
[DELTA][R.sup.2]        0.020 ***    0.035 ***    0.052 ***    0.059 ***
Career support (CS)     0.03        -0.02         0.05 *      -0.03
Psychosocial
  support (PS)         -0.06 ***    -0.01        -0.10 ***     0.04
Mentor gender (MG)     -0.00        -0.01        -0.03        -0.01
[DELTA][R.sup.2]        0.002 *      0.001        0.007 ***    0.001
[R.sup.2]               0.750 ***    0.761 ***    0.612 ***    0.640 ***
Df                     16,1556      16,1562      16,1561      16,1573

                                 B                         B

                          Managerial level        Time since promotion

     Variable              W            M            W            M

Time 1 dependent
  variable              0.70 ***     0.73 ***     0.36 ***     0.47 ***
[DELTA][R.sup.2]        0.671 ***    0.713 ***    0.202 ***    0.356 ***
Sector                  0.18 ***     0.14 ***    -0.02        -0.10 ***
Organisation size       0.02        -0.03 *      -0.05 *      -0.06 **
Age                    -0.01        -0.03        -0.10 ***     0.09 ***
Tenure                 -0.05 ***    -0.03         0.09 ***     0.22 ***
Managerial level        --           --           0.01         0.01
Occupation level       -0.15 ***    -0.02        -0.03        -0.00
Education              -0.03         0.02         0.00         0.00
Training                0.13 ***     0.14 ***    -0.04        -0.01
Masculinity             0.06 ***     0.06 ***    -0.06 *      -0.05 *
Formal mentor
  program              -0.02        -0.02         0.02        -0.03
Mentor level            0.01         0.01        -0.02         0.01
Years relationship      0.01        -0.01         0.07 **     -0.00
[DELTA][R.sup.2]        0.038 ***    0.029 ***    0.041 ***    0.064 ***
Career support (CS)    -0.02        -0.04 *      -0.12 ***     0.04
Psychosocial
  support (PS)          0.04 *       0.02         0.07 *      -0.04
Mentor gender (MG)     -0.02        -0.02        -0.02         0.03
[DELTA][R.sup.2]        0.004 ***    0.001        0.007 **     0.002
[R.sup.2]               0.713 ***    0.744 ***    0.250 ***    0.421 ***
Df                     15,1559      15,1568      16,1545      16,1563

                                 B                         B

                                                     Promotion last
                          Chance promotion              year (a)

     Variable              W            M            W            M

Time 1 dependent
  variable             -0.49 ***    -0.53 ***     0.10 ***     0.10 ***
[DELTA][R.sup.2]        0.327 ***    0.419 ***    0.020 ***    0.029 ***
Sector                 -0.02         0.05 *      -0.06 *       0.04
Organisation size       0.04         0.03 *       0.06 *       0.09 ***
Age                    -0.13 **     -0.18 ***    -0.10 ***    -0.10 **
Tenure                 -0.06 *      -0.09 ***    -0.06 *      -0.14 ***
Managerial level       -0.03        -0.03        -0.05        -0.03
Occupation level       -0.02        -0.04         0.11 ***     0.05
Education              -0.04        -0.01         0.04         0.00
Training                0.05         0.03        -0.03        -0.03
Masculinity             0.08 ***     0.05 *       0.05 *       0.07 **
Formal mentor
  program               0.00         0.06 ***     0.01         0.02
Mentor level            0.01         0.03         0.02         0.05 *
Years relationship     -0.06 **     -0.02        -0.07 **      0.00
[DELTA][R.sup.2]        0.037 ***    0.050 ***    0.068 ***    0.083 ***
Career support (CS)     0.10 ***    -0.01         0.10 **     -0.01
Psychosocial
  support (PS)         -0.04         0.02        -0.03         0.02
Mentor gender (MG)      0.01        -0.02        -0.02         0.02
[DELTA][R.sup.2]        0.006 ***    0.001        0.007 ***    0.001
[R.sup.2]               0.370 ***    0.464 ***    0.096 ***    0.112 ***
Df                     16,1554      16,1561      16,1571      16,1577

Note: W = women, M = men; and
* p < 0.05. ** p < 0.01. *** p < 0.001.

Table 7

Summary of the Prediction of Career Advancement by Dyads
Varying by Protege and Mentor Gender

     Dyads          Salary        Managerial        Managerial
                                   Promotion           Level

                                Career Support

Male protege/     Not           Not significant   Decreased
male mentor       significant   or decreased
Male protege/     Not           Not significant   Not significant
female mentor     significant
Female protege/   Increased     Not significant   Decreased
male mentor
Female protege/   Not           Increased         Not significant
female mentor     significant

                                Psychosocial Support

Male protege/     Decreased     Not significant   Increased or
male mentor       or not                          not significant
                  significant
Male protege/     Not           Not significant   Not significant
female mentor     significant
Female protege/   Decreased     Decreased         Not significant
male mentor
Female protege/   Not           Decreased         Not significant
female mentor     significant

     Dyads          Time Since         Chance of       Promoted Last
                     Promotion         Promotion           Year

                                    Career Support

Male protege/     Not significant   Not significant   Not significant
male mentor
Male protege/     Not significant   Not significant   Not significant
female mentor
Female protege/   Decreased         Decreased         Not significant
male mentor                                           or decreased
Female protege/   Decreased         Increased         Increased
female mentor

                                    Psychosocial Support

Male protege/     Not significant   Not significant   Not significant
male mentor

Male protege/     Not significant   Not significant   Not significant
female mentor
Female protege/   Not significant   Not significant   Not significant
male mentor
Female protege/   Increased         Not significant   Decreased
female mentor

Note: Where there are two results in a column, the second is for the
analyses where the Time 1 measure of the dependent variable was
controlled.

References

Allen, T.D., Eby, L.T., Poteet, M.L., Lentz, E. & Lima, L. 2004, 'Career benefits associated with mentoring for protdges: A meta-analysis', Journal of Applied Psychology, vol. 89, pp. 127-36.

Allen, T.D., Poteet, M.L. & Russell, J.E.A. 2000, 'Protege selection by mentors: What makes the difference?', Journal of Organizational Behavior, vol. 21, pp. 271-82.

Anderson, S.E. & Williams, L.J. 1992, 'Assumptions about unmeasured variables with studies of reciprocal relationships', Journal of Applied Psychology, vol. 77, pp. 638-50.

Australian Bureau of Statistics, 2003, 'Labour force', Australia: Commonwealth Government Printer, catalogue no. 6203.0, Canberra.

Bahniuk, M.H., Dobos, J. & Kogler Hill, S.E. 1990, 'The impact of mentoring, collegial support, and information adequacy on career success', Journal of Social Behavior and Personality, vol. 5, pp. 431-51.

Bern, S. 1981, Bem Sex-Role Inventory, Consulting Psychologists Press, NY.

Blake-Beard, S. 2001, 'Taking a hard look at formal mentoring programs', Journal of Management Development, vol. 20, pp. 331-45.

Burke, R.J. & McKeen, C.A. 1990, 'Mentoring in organizations: Implications for women', Journal of Business Ethics, vol. 9, pp. 317-32.

Burke, R.J. & McKeen, C.A. 1997, 'Benefits of mentoring relationships among managerial and professional women', Journal of Vocational Behavior, vol. 51, pp. 43-57.

Butt, R.S. 1998, 'The gender of social capital', Rationality and Society, vol. 10, pp. 5-46.

Chao, G.T. 1997, 'Mentoring phases and outcomes', Journal of Vocational Behavior, vol. 51, pp. 15-28.

Chao, G.T., Walz, P.M. & Gardner, P.D. 1992, 'Formal and informal mentorships', Personnel Psychology, vol. 45, pp. 619-36.

Clutterbuck, D. & Ragins, B.R. 2002, 'Mentoring and Diversity', Butterworth Heinemann, Oxford.

Corzine, J.B., Buntzman, G.F. & Busch, E.T. 1994, 'Mentoring, downsizing, gender and career outcomes', Journal of Social Behavior and Personality, vol. 9, pp. 517-28.

Cox, T.H. & Harquail, C.V. 1991, 'Career paths and career success in the early career stages of male and female MBAs', Journal of Vocational Behavior, vol. 39, pp. 54-75.

Day, R. & Allen, T.D. 2004, 'The relationship between career motivation and self-efficacy with protege career success', Journal of Vocational Behavior, vol. 64, pp. 72-91.

Department of Employment and Industrial Relations, 1987, Australian Standard Classification of Occupations, Australian Government Publishing Service. Canberra.

Dreher, G.F. & Ash, R.A. 1990, 'A comparative study of mentoring among men and women in managerial positions', Journal of Applied Psychology, vol. 75, pp. 539-46.

Dreher, G.F. & Chargois, J.A. 1998, 'Gender, mentoring experiences, and salary attainment among graduates of an historically black university', Journal of Vocational Behavior, vol. 53, pp. 401-16.

Dreher, G.F. & Cox, T.H. 1996, 'Race, gender and opportunity', Journal of Applied Psychology, vol. 81, pp. 297-308.

Ensher, E.A., Thomas, C. & Murphy, S.E. 2001, 'Comparison of traditional, step-ahead, and peer mentoring on proteges' support, satisfaction, and perceptions of career success', Journal of Business and Psychology, vol. 15, pp. 419-38.

EOWA 2003, Census of Women Board Directors and Women Executive Managers, Australian Bureau of Statistics 6150.0--July 2003. Retrieved July 12, 2004, from http://www.eeo.gov.au/Australian_Women_In_Leadership_Census/ 2003_Australian_Women_In_Leadership_Census.asp.

Fagenson, E.S. 1989, 'The mentor advantage: Perceived career/job experiences of proteges versus nonproteges', Journal of Organizational Behavior, vol. 10, pp. 309-20.

Gattiker, U.E. & Larwood, L. 1989, 'Career success, mobility and extrinsic satisfaction of managers', The Social Science Journal, vol. 26, pp. 75-92.

Green, S.G. & Bauer, T.N. 1995, 'Supervisory mentoring by advisors', Personnel Psychology, vol. 48, pp. 537-61.

Higgins, M.C. & Kram, K.E. 2001, 'Reconceptualising mentoring at work: A developmental network perspective', Academy of Management Review, vol. 26, pp. 264-88.

International Labour Organization, 1997, Breaking the Glass Ceiling." Women in Management, International Labour Office, Geneva.

Janasz, S.C., Sullivan, S.S. & Whiting, W. 2003, 'Mentor networks and career success: Lessons from turbulent times', Academy of Management Executive, vol. 17, pp. 78-91.

Johnson, N.B. & Scandura, T.A. 1994, 'The effects of mentoring and sex-role style on male-female earnings', Industrial Relations, vol. 33, pp. 263-74.

Joreskog, K.G. & Sorbom, D. 2000, LISREL 8, SPSS, Chicago.

Kidd, M.P. & Shannon, M. 1996, 'The Gender Wage Gap: A Comparison of Australia and Canada', Industrial and Labor Relations, vol. 49, pp. 729-46.

Kirchmeyer, C. 1998, 'Determinants of managerial career success', Journal of Management, vol. 24, pp. 673-92.

Kram, K.E. 1983, 'Phases of the mentor relationship', Academy of Management Journal, vol. 26, pp. 608-25.

Kram, K.E. 1985, Mentoring at Work, Scott, Foresman, Glenview, IL.

Lyness, K.S. & Thompson, D.E. 2000, 'Climbing the corporate ladder: Do female and male managers follow the same route?', Journal of Applied Psychology, vol. 85, pp. 86-101.

Metz, I. & Tharenou, P. 2001, 'Women's career advancement', Group and Organization Management, vol. 26, pp. 312-42.

Mainiero, L. 1994, 'Getting anointed for advancement: The case of executive women', Academy of Management Executive, vol. 8, pp. 53-67.

Noe, R. 1988, 'Women and mentoring', Academy of Management Review, vol. 13, pp. 65-78.

Noe, R., Greenberger, D.B. & Wang, S. 2002, 'Mentoring: What we know and where we go', Research in Personnel and Human Resources Management, vol. 21, pp. 129-73.

Olian, J.D., Carroll, S.J. & Giannantonio, C.M. 1993, 'Mentor reactions to proteges: An experiment with managers', Journal of Vocational Behavior, vol. 43, pp. 266-78.

O'Neill, R.M. 2002, 'Gender and race in mentoring relationships', in Mentoring and Diversity, eds. D. Clutterbuck & B.R. Ragins, Butterworth Heinemann. Oxford, pp. 1-22.

Orpen, C. 1995, 'The effects of mentoring on employees' career success', The Journal of Social Psychology, vol. 13, pp. 667-68.

Ragins, B.R. 1989, 'Barriers to mentoring', Human Relations, vol. 42, pp. 1-22.

Ragins, B.R. 1997, 'Diversified mentoring relationships in organizations: A power perspective', Academy of Management Review, vol. 22, pp. 482-521.

Ragins, B.R. 1999, 'Gender and mentoring relationships', in Handbook of Gender in Organizations, ed. G.N. Powell, Oaks, Sage, CA, pp. 347-70.

Ragins, B.R. & Cotton, J. 1999, 'Mentor functions and outcomes', Journal of Applied Psychology, vol. 84, pp. 529-50.

Ragins, B.R., Cotton, J. & Miller, J.S. 2000, 'Marginal mentoring', Academy of Management Journal, vol. 43, pp. 1177-94.

Ragins, B.R. & McFarlin, D.B. 1990, 'Perceptions of gender roles in cross-gender mentoring relationships', Journal of Vocational Behavior, vol. 37, pp. 321-39.

Ragins, B.R. & Sundstrom, E. 1989, 'Gender and power in organizations', Psychological Bulletin, vol. 105, pp. 51-88.

Ragins, B.R., Townsend, B. & Mattis, M. 1998, 'Gender gap in the executive suite', Academy of Management Executive, vol. 12, no. 1, pp. 28-42.

Rosenbaum, J. 1984, Career Mobility in a Corporate Hierarchy, Academic Press, NY.

Scandura, T.A. 1992, 'Mentorship and career mobility: An empirical investigation', Journal of Organizational Behavior, vol. 13, pp. 169-74.

Scandura, T.A. & Ragins, B.R. 1994, 'The effects of sex and gender-role orientation on mentorship in male-dominated occupations', Journal of Vocational Behavior, vol. 43, pp. 251-65.

Scandura, T.A. & Williams, E.A. 2001, 'An investigation of moderating effects of gender on the relationships between mentorship initiation and protege perceptions of mentoring functions', Journal of Vocational Behavior, vol. 59, pp. 342-63.

Scandura, T. A. & Williams, E. A.2004, 'Mentoring and transformational leadership: The role of supervisory career mentoring', Journal of Vocational Behavior, vol. 65, pp. 448-68.

Schor, S.M. 1997, 'Separate and unequal', Business Horizons, vol. 40, no. 5, pp. 51-8.

Sosik J.J. & Godshalk, V.M. 2000, 'The role of gender in mentoring', Journal of Vocational Behavior, vol. 57, pp. 102-22.

Stout, S.K., Slocum, J.W. & Cron, W.L. 1988, 'Dynamics of the career plateauing process', Journal of Vocational Behavior, vol. 32, pp. 74-91.

Stroh, L.K., Brett, J.M. & Reilly, A.H. 1992, 'All the right stuff", Journal of Applied Psychology, vol. 77, pp. 251-60.

Tepper, K., Shaffer, B. & Tepper, B. 1996, 'Latent structure of mentoring function scales', Educational and Psychological Measurement, vol. 56, pp. 848-57.

Tharenou, P. 1997, 'Managerial career advancement', in International Review of Industrial and Organizational Psychology, eds. C. Cooper & I. Robertson, Wiley, NY, pp. 39-93.

Tharenou, P. 1999, 'Gender differences in advancing to the top', International Journal of Management Reviews, vol. 2, no. 1, pp. 1-22.

Tharenou, P. 2001, 'Going up? Do traits and informal social processes predict advancing in management?', Academy of Management Journal, vol. 44, pp. 1005-17.

Turban, D.B. & Dougherty, T.W. 1994, 'Role of protege personality in receipt of mentoring and career success', Academy of Management Journal, vol. 37, pp. 688-702.

Wallace, J.E. 2001, 'The benefits of mentoring for female lawyers', Journal of Vocational Behavior, vol. 58, pp. 366-91.

Wanberg, C.R., Welsh, E.T. & Hezlett, S.A. 2003, 'Mentoring research: A review and dynamic process model', Research in Personnel and Human Resources Management, vol. 22, pp. 39-124.

Waters, L., McCabe, M., Kiellerup, D. & Kiellerup, S. 2002, 'The role of formal mentoring on business success and self-esteem in participants of a new business start-up program', Journal of Business and Psychology, vol. 17, pp. 107-21.

Weill, P.A. & Kimball, P.A. 1996, 'Determinants of salaries earned', Journal of Organizational Behavior, vol. 17, pp. 657-78.

Wirth, L. 2001, Breaking Through the Glass' Ceiling, International Labour Office, Geneva.

Phyllis Tharenou, Division of Business and Enterprise, University of South Australia, City West Campus, North Terrace SA 5000. Email: [e-mail]