Toward Improving the Effectiveness of Formal Mentoring Programs: Matching by Personality Matters

Size: px
Start display at page:

Download "Toward Improving the Effectiveness of Formal Mentoring Programs: Matching by Personality Matters"

Transcription

1 579567GOMXXX / Group & Organization ManagementMenges research-article2015 Article Toward Improving the Effectiveness of Formal Mentoring Programs: Matching by Personality Matters Group & Organization Management 2016, Vol. 41(1) The Author(s) 2015 Reprints and permissions: sagepub.com/journalspermissions.nav DOI: / gom.sagepub.com Christine Menges 1 Abstract Organizations establish formal mentoring programs to advance personal and professional development, but not all relationships between mentors and protégés deliver these results. Based on the similarity-attraction paradigm, it is proposed that protégés receive more career and psychosocial support if mentors and protégés have similar personalities. A test of the model with data from a sample of 68 mentor protégé dyads of a formal mentoring program showed, first, that career support is linked to mentor and protégé similarity in the personality trait openness to experience and, second, that psychosocial support for protégés is linked to mentor and protégé similarity in openness to experience and conscientiousness. It is concluded that matching mentors with protégés on two specific personality traits, openness to experience and conscientiousness, enhances the outcomes of mentoring relationships for protégés. Keywords mentoring, personality, human resource management 1 Anglia Ruskin University, Cambridge, UK Corresponding Author: Christine Menges, Anglia Ruskin University, East Road, Cambridge CB1 1PT, UK. christine.menges@anglia.ac.uk

2 Menges 99 Traditional mentoring refers to the provision of career and psychosocial support by an experienced mentor to a relatively less experienced protégé. Research has consistently demonstrated that mentoring provides many benefits to protégés. In organizational settings, several reviews and meta-analyses have shown the positive effect of mentoring on both objective career outcomes (e.g., promotions and salary) and subjective career outcomes (e.g., career and job satisfaction; Allen, Eby, O Brien, & Lentz, 2008; Allen, Eby, Poteet, Lentz, & Lima, 2004; Kammeyer-Mueller & Judge, 2008; Ng, Eby, Sorensen, & Feldman, 2005; Underhill, 2006; Wanberg, Welsh, & Hezlett, 2003). To reap the benefits of mentoring, formal mentoring programs have been initiated across a wide range of contexts, including corporate and educational settings (Allen, Eby, & Lentz, 2006). In formal mentoring relationships, organizations match less experienced protégés with more experienced and knowledgeable mentors who are expected to provide career support (e.g., sponsorship and challenging assignments) and psychosocial support (e.g., counseling and encouragement) to enhance the protégés career and personal development (Baugh & Fagenson-Eland, 2007; Finkelstein, Allen, Ritchie, Lynch, & Montei, 2012; Kram, 1985; Noe, Greenberger, & Wang, 2002; Wang, Tomlinson, & Noe, 2010; Weinberg & Lankau, 2011). Despite the prevalence and potential importance of formal mentoring programs, there is, however, relatively little empirical research on how those programs can effectively achieve the benefits inherent in mentoring. Although such research has been growing in recent years, our knowledge of informal mentoring still outweighs what we know about formal mentoring. Formal mentoring relationships differ from informal mentoring in several ways: Formal mentoring relationships are initiated through an organizational program that assigns mentors to protégés and then facilitates and supports the relationship within the assigned dyad (Wanberg et al., 2003). Informal mentoring relationships, in contrast, develop spontaneously and without involvement from an organization (Chao, Walz, & Gardner, 1992), usually on the basis of mutual identification, interpersonal comfort, and sympathy (Kram, 1985; Ragins, 2002). Formal mentoring relationships, therefore, are initiated and controlled by a third party (i.e., the organization), whereas informal mentoring relationships are self-organized. Formal and informal mentoring also differ in duration and structure (Ragins & Cotton, 1999). Formal relationships are developed for a specific period of time and with prescribed goals for the protégés. In informal mentoring relationships, protégés and mentors decide themselves which goals they want to pursue and in which time frame (Allen & Eby, 2003). The research that exists on formal mentoring programs suggests that such programs are not as effective as informal mentoring, because the naturally

3 100 Group & Organization Management 41(1) occurring attraction that is typical of informal relationships is often lacking in formal relationships (Blake-Beard, O Neill, & McGowan, 2007; Turban & Lee, 2007). One cannot force the natural attraction and desire to work together that are essential for successful mentoring relationships (Chao et al., 1992; Fagenson-Eland, Marks, & Amendola, 1997; Ragins & Cotton, 1999). For example, a qualitative study examining relational problems in formal organizational mentoring programs found that mismatches regarding personality, values, and working style were the most common mentoring problems for both protégés and mentors (Eby, McManus, Simon, & Russell, 2000). Notably, such problems lessen the benefits of mentoring; protégés experiencing mentoring problems report more stress, stronger intentions to leave the organization and the mentoring relationship, less relational learning and job satisfaction, more psychological job withdrawal, and more depressed mood than those who do not experience problems in mentoring (Eby & Allen, 2002; Eby & McManus, 2004). To avoid these problems in formal mentoring relationships and to reap the benefits of informal mentoring relationship, it is of critical importance to match mentors and protégés in ways that resemble or at least facilitate some of the natural attraction that drives informal mentoring relationships. A key question for formal mentoring program managers is, therefore, what matching criteria they could effectively use. Extant research provides minimal guidance. Calls for research on the matching process in formal mentoring relationships have sounded, particularly in respect to matching criteria (Blake-Beard et al., 2007) but the answers thus far have left many questions open. What adds to the challenge is that the information regarding the criteria that are utilized in the matching process needs to be available before the mentoring relationship can begin. In an endeavor to identify useful criteria for the matching process in formal mentoring programs, this study examines whether mentor and protégé personality traits matter. The guiding hypothesis, based on the similarityattraction paradigm (Byrne, 1971), is that protégés will receive better mentoring support (both in terms of career and psychosocial support) in formal mentoring programs if their personality is similar, rather than dissimilar, to the personality of their mentor. The research thus examines how protégé and mentor personality characteristics interact with one another (Chandler, Kram, & Yip, 2011; Eby & Allen, 2002; Turban & Lee, 2007) and how dyadic characteristics (i.e., mentor protégé personality similarity) affect the extent to which protégés receive career and psychosocial support from mentors. The findings obtained in a field study with 68 mentor protégé matched pairs who participated in a formal mentoring program suggest a nuanced pattern of effects according to which, intriguingly, only a few distinct personality trait

4 Menges 101 similarities and not personality similarity in general matter for effective matching between protégés and mentors. Matching in Formal Mentoring Programs Common matching processes observed in organizations by which mentoring partners are put together to establish formal mentoring relationships are administrator-assigned matching (i.e., program coordinators match mentors and protégés based on their personal assessments of the likelihood that an interpersonal connection between the mentoring partners will form), choicebased matching (i.e., the mentoring partners choose each other during the matching process), and assessment-based matching (i.e., mentoring partners are assessed along specific dimensions and then matched based on the results of the assessments; Blake-Beard et al., 2007). Research shows that mentors and protégés input into the matching process in formal mentoring relationship relates to perceived program effectiveness, mentor commitment, and program understanding (Allen et al., 2006). However, it has remained an open question which particular criteria should be assessed and used to match formal mentoring pairs in ways that facilitate beneficial mentoring outcomes (Blake-Beard et al., 2007). Matching criteria that have been considered include demographic characteristics, particularly gender and race. Pertinent research found that racial and gender similarities result in more fruitful mentoring relationships (Allen & Eby, 2004; Feldman, Folks, & Turnley, 1999; Lankau, Riordan, & Thomas, 2005; McKeen & Bujaki, 2007; Ragins, 1997a, 1997b; Scandura & Williams, 2001; Thomas, 1993). What has rarely been considered, however, is that there is a deeper psychological layer, beneath the surface characteristics of race and gender, that may, too, be of considerable importance for the matching process. Specifically, the idea that similarities in mentors and protégés personalities may matter for the mentoring support that the protégé receives has been neglected despite pertinent calls from mentoring scholars to examine the personality characteristics not just of each mentoring partner but also of the mentoring dyad (Chandler et al., 2011; Eby & Allen, 2002; Turban & Lee, 2007). Personality and Mentoring Researchers have lamented that relatively little is known about the importance of personality characteristics in formal mentoring relationships both in general, as well as in respect specifically to the dyadic matching process that precedes the formation of such relationships (Turban & Lee, 2007). Personality is defined as the relatively stable dispositions of individuals that contribute to

5 102 Group & Organization Management 41(1) consistency in their thoughts, behaviors, and emotions (Funder, 2001; Leary, 1999). Decades of research extracted five dimensions of personality (i.e., traits) along which individuals tend to systematically differ or converge. These Big Five dimensions are openness to experience, conscientiousness, extraversion, agreeableness, and neuroticism. These traits form the five-factor model (FFM) of personality that has been accepted by many scholars as a comprehensive framework for organizing a wide range of personality dispositions (Digman, 1996; Goldberg, 1993; McCrae & Costa, 1996). Personality traits that fall outside of this framework include, for example, self-monitoring, locus of control, and Type A or Type B personality. Mentoring researchers interested in personality have mainly focused on the effects of an individual s (i.e., a protégé s or mentor s) personality characteristics on the involvement in and the quality of mentoring relationships. Relevant findings in informal mentoring settings include that individuals who engaged in mentoring relationships tended to be open to experience, conscientious, and extraverted (Niehoff, 2006) and that individuals with internal loci of control, high self-monitoring, and low neuroticism were more likely to initiate and receive mentoring (Turban & Dougherty, 1994). Further work concerning informal mentoring relationships identified that extraversion and Type A personality relate positively to protégé-initiated informal mentoring relationships (Aryee, Lo, & Kang, 1999). The same study also found that protégés individual extraversion was positively related to mentoring received, suggesting that extraverted protégés with their gregarious and sociable disposition tended to engage more frequently in interactions with their mentors (Aryee et al., 1999). Waters (2004) found that informally established mentoring pairs in which mentors and protégés both had high levels of agreeableness, openness, and extraversion (and those in which protégés had high conscientiousness) hold a shared view of the relationship regarding, particularly, the provision of psychosocial support. Similarly, Bozionelos (2004) found that, in informal mentoring relationships, mentors individual openness to experience relates positively to the mentoring provided, showing that individuals with broad interests and receptivity to new experiences and ideas are more likely to provide mentoring support. In a formal mentoring setting, however, Wanberg, Kammeyer-Mueller, and Marchese (2006) found no significant relationship of individual s openness to experience and mentoring support received. Similarity in Mentoring Relationships The research on the individual personality characteristics of each mentoring partner in mentoring relationships notwithstanding, very little is known about the mentoring pair as an interactive dyad of personalities (Turban & Lee,

6 Menges ). Because each mentoring partner s personality contributes in a synergistic way to the relationship, what results from the relationship is likely to constitute more than the sum of its parts. In a contribution to extant mentoring research on personality, this study advances the core idea that the similarity between the mentors and the protégés personalities matters for the mentoring support protégés receive. This idea stems from the similarity-attraction paradigm (Byrne, 1971; Newcomb, 1956), which suggests that people have more positive interactions with those who are similar to them than with those who are dissimilar. The rationale for this argument is that similarities invoke an attraction dynamic whereby individuals with similar attitudes and personalities accentuate each other s positive attributes and derive positive social identity from viewing and treating each other favorably, whereas individuals with dissimilar attitudes and personalities tend to view and treat each other less favorably. A common counter-argument is that differences in attitudes and personalities create benefits through dissimilarity that could facilitate attraction, too (Dryer & Horowitz, 1997; Schimel, Pyszczinski, Greenberg, O Mahen, & Arndt, 2000); yet research shows largely that similarity is more desired by the relationship partners and more facilitative to beneficial outcomes than dissimilarity (Caspi & Herbener, 1990; Gattis, Berns, Simpson, & Christensen, 2004; Luo & Klohnen, 2005; Robins, Caspi, & Moffitt, 2000). For example, Dijkstra and Barelds (2008) showed that individuals desire romantic partners who resemble themselves in terms of personality. In support of the similarityattraction hypothesis, correlational analyses revealed that individuals own personality scores related to those of their ideal romantic partner. Researchers interested in similarities between (any) relationship partners have made a distinction between actual similarity (i.e., the degree to which one is actually similar to another individual) and perceived similarity (i.e., the degree to which one believes oneself similar to another). Although Byrne (1971) originally proposed that actual similarity produces attraction, others have more recently argued that it is perceived similarity that produces attraction (e.g., Condon & Crano, 1988; Hoyle, 1993; Ptacek & Dodge, 1995). And yet, numerous studies suggest that actual similarity in personality affects not only how relationships develop but also how effective and fulfilling they are. For example, married and dating couples with actual similar personalities are more satisfied with their relationships than those with dissimilar personalities (Gattis et al., 2004; Luo & Klohnen, 2005; Robins et al., 2000). Also, actual personality similarities between supervisors and subordinates have been shown to advance the career progression of subordinates (Schaubroeck & Lam, 2002). Deluga (1998) found that subordinate supervisor actual similarity in conscientiousness was positively associated with subordinate in-role behavior. The

7 104 Group & Organization Management 41(1) underlying idea is that actual similarity in conscientiousness forms a mutually beneficial relationship, in which both understand each other s responsibilities and can mutually anticipate the behavior of the other (Meglino, Ravlin, & Adkins, 1991). Similarly, Ashkanasy and O Connor (1997) found that actual similarity in achievement values (a construct close to conscientiousness) was significantly related to higher leader member exchange. For mentoring relationships, both perceived similarity (i.e., one mentoring partner s perception of the other mentoring partner s personality) and actual personality similarity (i.e., both mentoring partners independently report similar personality profiles) matter, but the majority of research concerns perceived similarity (Burke, McKeen, & McKenna, 1993; Madia & Lutz, 2004; Owen & Solomon, 2006). For example, Madia and Lutz (2004) found that perceived similarity in extraversion and conscientiousness had a positive impact on mentors expressed intention to remain in a mentoring relationship. There are some exceptions that considered actual personality similarity. Godshalk and Sosik (2003) found that mentor s and protégé s actual similarity in learning-goal orientation (a trait that is related to, but narrower than openness to experience) resulted in more career support received in informal mentoring relationships a finding that was replicated for formal relationships (Egan, 2005). Although perceived similarity is reportedly more important than actual similarity in determining mentoring relationship outcomes (Montoya, Horton, & Kirchner, 2008; Turban & Jones, 1988), perceived similarity cannot be used for the matching process in formal mentoring programs, because matching has to occur prior to the participants actually meeting. What could be used, however, is actual similarity between mentoring partners, as actual personality can be assessed with standard personality measures prior to the commencement of a formal mentoring relationship. A considerable limitation in our knowledge about formal mentoring relationships is that we know very little about whether the match in actual (rather than perceived) personality between protégés and mentors matters (beyond learning-goal orientation) for the mentoring support that protégés receive from mentors. Here it is suggested that similarity in personality along all five major personality dimensions openness to experience, conscientiousness, extra-version, agreeableness, and neuroticism will make it easier for mentors and protégés to establish relationships in a formal mentoring program and to yield the benefits associated with mentoring. Because both individuals mentors and protégés enter the relationship created by a formal program as strangers, interacting with someone with similar personality traits will make it easier to establish rapport, will enhance a sense of trust and confidence, and will reinforce the favorable self-image of both mentors and protégés. According to the similarity-attraction paradigm (Byrne, 1971; Newcomb, 1956), similarity

8 Menges 105 facilitates these benefits regardless of the personality trait in question. The underlying rationale, according to reinforcement-affect accounts (Clore & Byrne, 1974), is that personality similarity bolsters the interaction partners opinions, beliefs, and feelings and thus evokes a subtle positive affective response that triggers attraction and facilitates interaction. According to uncertainty reduction accounts, personality similarity also increases predictability, thus allowing interaction partners to gain trust in one another and to communicate more smoothly and with less effort (Berger & Calabrese, 1975). Empirical research showed that even similarity on neuroticism a personality trait that could be seen as not particularly socially desirable can carry relational benefits: Employees received higher peer-assessed performance ratings to the extent that employees level of neuroticism matched their supervisor s level of neuroticism (Strauss, Barrick, & Connerley, 2001). Based on these theoretical and empirical considerations, the expectation in this research is that actual personality similarity in all five major personality traits will benefit the mentoring relationship. Once a high-quality mentoring relationship is established, mentors are likely to provide valuable career and psychosocial support to protégés (Kram, 1985). Career support concerns the help mentors provide to facilitate their protégés career ambitions. Mentors can help protégés, for example, to find the right job, advance in the hierarchy of an organization, or prepare for a leadership role. Psychosocial support is provided when a relationship exists in which mentor and protégé experience acceptance and confirmation through interaction. Mentors can support protégés, for example, by listening empathically, by helping them through tough times, and by giving them self-confidence. In summary, a similarity-attraction perspective is brought to bear upon the question as to what matching criteria could be effectively used in formal mentoring programs to enhance the career support and the psychosocial support that protégés receive as part of the mentoring relationship. Taking the considerations above into account, the hypotheses are as follows: Hypothesis 1: Protégés and mentors similarity in personality relates positively to the extent to which protégés receive career support. Hypothesis 2: Protégés and mentors similarity in personality relates positively to the extent to which protégés receive psychosocial support. Method Sample and Data Description The present study was conducted in a formal mentoring program of a medium-sized Swiss business school. This program starts every fall semester,

9 106 Group & Organization Management 41(1) lasts for 2 years, is voluntary, and is available for both undergraduate (i.e., bachelor) and graduate (i.e., master and PhD) students. The mentors are alumni of the business school who volunteer to take on students as protégés. The goal of the program is to enhance students personal and career development. A mentoring program manager administers the program, which involves coupling mentors and protégés at the beginning of the program, ensuring mentors and protégés meet once per semester or more, and resolving any concerns raised by mentors or protégés. For the year in which this study was conducted, 241 mentoring pairs were formed according to students preferences. From a list of available mentors, the students ranked their preferences based on each mentor s job title and company, and most students were paired with one of their top three choices. The study consisted of two measurement points (Time 1 and Time 2). First, 6 months after the mentoring relationships were initiated, all 241 mentor protégé pairs were invited via to complete a web-based questionnaire that captured the independent variable (i.e., personality). Second, 6 months later (i.e., 12 months after the start of the mentoring program), all mentors and protégés were invited via to complete a second web-based questionnaire that captured the dependent variables (i.e., psychosocial support and career support). Only those mentoring tandems were retained in the sample in which both partners (i.e., mentor and corresponding protégé) completed the survey at both time points. The participation was voluntary and for research purposes only; participants received no compensation for completing the surveys. The final sample consisted of 68 mentoring pairs (response rate: 28%). The sample of the mentors included 44 (65%) men and 24 (35%) women averaging 43 years of age (SD = 10.13). Most mentors were Swiss (78%). The sample of the protégés included an equal number of men (34) and women (34) and averaged 24 years of age (SD = 3.77). Most protégés were Swiss (74%). Of the 68 protégés, 51 (75%) were undergraduates, 17 (25%) were graduates, and most (71%) had received one of their three mentors of choice. Race/ethnicity was not assessed. Measures Personality. The Big Five personality traits (openness to experience, conscientiousness, extraversion, agreeableness, and neuroticism) were assessed with Saucier s (1994) Big Five measure. Each personality trait was assessed with eight adjectives; for example, openness for experience was assessed with adjectives including creative, intellectual, and imaginative. Both mentors and protégés rated the extent to which each adjective described themselves on a 5-point scale ranging from 1 (strongly disagree) to 5 (strongly

10 Menges 107 agree). The internal consistencies for protégés ratings were α =.77 (openness), α =.67 (conscientiousness), α =.64 (extraversion), α =.71 (agreeableness), and α =.84 (neuroticism). The internal consistencies for mentors ratings were α =.63 (openness), α =.75 (conscientiousness), α =.68 (extraversion), α =.69 (agreeableness), α =.76 (neuroticism). The alpha values were within the common range reported in recent work (e.g., Solomon & Jackson, 2014) utilizing Saucier s (1994) Big Five measure. Mentoring support. Mentoring support was assessed by the protégés at measurement point T2. Protégés provided responses to a four-items measure of career support (Cronbach s α =.79) and an eight-items measure of psychosocial support (Cronbach s α =.89) on a 5-point scale ranging from 1 (strongly disagree) to 5 (strongly agree). All items are listed in the appendix. The measure used to assess career support and psychosocial support that protégés received was developed specifically for the student sample of this study. Established mentoring support measures mainly focus on mentoring in corporate settings. Such measures assess career support as the extent to which, for example, sponsorship, exposure-and-visibility, coaching, protection, and challenging work assignments are provided. These aspects of career support do not apply to a student-alumni mentoring program. Rather, career support in this context is more likely to include assistance in career planning, job search, and application processes. Therefore, the measure to assess career support in this study was adopted to capture these particular ways of support. Psychosocial support is assessed in established measures as role modeling, acceptance and confirmation, counseling, and friendship (Kram, 1985). In the context of an alumni-student mentoring program, psychosocial support also concerns the emotional help students at their young age need to deal with the demands of their studies. The measure was established following pertinent recommendations (Hinkin, 1998). First, 23 items were generated based on informal interviews with mentors and protégés who had completed the program in the past, on literature about student mentoring (D Abate, 2010), and on established mentoring function scales such as the Mentoring Function Items by Noe (1988) and the Career Support Scale by Riley and Wrench (1985). Next, a pre-test study was run with an independent sample of 212 protégés in the mentoring program cohort of the year prior to the one that was used for the main data collection. Participants were asked to rate the 23 items on a 5-point scale ranging from 1 (strongly disagree) to 5 (strongly agree). The response rate was 47.6%. Then, an exploratory factor analysis (EFA) was conducted to extract the relevant factors. A principal components analysis with oblique rotation identified a two-factor solution with items loading on psychosocial

11 108 Group & Organization Management 41(1) support and career support. Only those items that clearly loaded on a single factor and had factor loadings of.60 and above were retained (Nunnally, 1976), resulting in 6 items for career support and 12 items for psychosocial support. In the next step, these 18 items were administered to the 241 protégés invited to participate in the main study at measurement point T1. Of those invited, 102 protégés (42%) provided complete responses at measurement point T1, and the data of all those protégés were utilized (independent of whether these protégés also responded at measurement point T2) for the purpose of a confirmatory factor analysis (CFA) to quantify the goodness of fit of the factor structure (Long, 1983) and to determine which items to retain for the final measurement of mentoring support. Upon inspection of factor loadings, only items with factor loadings above.60 were kept (Nunnally, 1976) and so 4 items were retained to measure career support and 8 items to assess psychosocial support. The results of the final CFA for the proposed two-factor structure indicated a reasonable fit (χ 2 = 87.06, df = 51, p <.001, comparative fit index [CFI] =.94, root mean square error of approximation [RMSEA] =.08, Tucker Lewis index [TLI] =.91) and compared favorably with a onefactor structure (χ 2 = , df = 52, p <.001, CFI =.82, RMSEA =.14, TLI =.73, Δχ 2 = 74.56, p <.001). Thus, the remaining 4 items for career support and 8 items for psychosocial support listed in the appendix were used to assess the mentoring functions at measurement point T2. Control variables. Previous studies on mentoring found demographic variables to be significantly related to mentoring support (Fagenson-Eland, Marks, & Amendola, 1997); therefore, the genders, ages, and nationalities of mentors and protégés were included as control variables. To account for similarity effects on demographic variables, the analyses also controlled for whether the mentoring relationship consisted of a same gender dyad (i.e., man man, woman woman) or a cross-gender dyad (i.e., woman man) and for whether mentor and protégé were either both Swiss or foreign, or had different nationalities. In addition, it was controlled for whether protégés were at the undergraduate or graduate level and for whether protégés had or had not been assigned one of their three preferred mentors. It was also controlled for how frequently the mentoring dyad had met in person, by asking the protégés at measurement point T2 for how often they had met their mentor in person. The answer options, on a 5-point scale, were 0, 1, 2, 3, and more than 3 times. Because the quality of the psychosocial and career support that protégés received could depend on the mentor s satisfaction with the mentoring relationship, this confounding variable was also included as a control variable. The mentor s satisfaction with the protégé was measured using Ragins

12 Menges 109 and Cotton s (1999) four-item scale (Cronbach s α =.82). Sample items are My protégé is someone I am satisfied with and My protégé has been effective in his/her role. Items were rated on a 5-point scale ranging from 1 (strongly disagree) to 5 (strongly agree). Data Analysis The hypotheses were tested by running a hierarchical regression analysis per personality trait (Aiken & West, 1991; Cohen & Cohen, 1983). In Step 1, the control variables were entered; in Step 2, mentors and protégés individual personality traits were entered; and in Step 3, the interaction term between mentors and protégés personality traits was entered. All variables were z-standardized prior to the analyses, as recommended by Dawson and Richter (2006). Results Descriptive Statistics Table 1 presents means, standard deviations, and bivariate correlations for mentors and protégés for all variables. The protégé sample consisted of undergraduate and graduate students. A check for differences between these students on the variables assessed showed no differences except for age, with the undergraduates (M = 23 years, SD = 3.46) being younger than the graduates (M = 27 years, SD = 2.35), t(66) = 5.05, p <.001, and extraversion, with the undergraduates (M = 3.68, SD =.66) being more extraverted than the graduates (M = 3.32, SD =.56), t(66) = 2.01, p <.05. Hypothesis Tests The hypotheses suggested that similarities in the personality traits between mentor and protégé benefit two mentoring functions, career support and psychosocial support, as assessed by the protégé. There was partial support for the hypotheses, revealing that distinct similarities in personality traits matter. Specifically, similarities in openness to experience and conscientiousness significantly affected the mentoring support received by the protégés, whereas agreeableness, neuroticism, and extraversion had no effects. Openness to experience and mentoring support. In support of Hypothesis 1, the results presented in Table 2 show that the dyadic interaction between mentor openness and protégé openness related significantly to the extent to which

13 Table 1. Descriptive Statistics for Mentors and Protégés. Variables M SD Mentors 1. Gender Age Nationality Satisfaction with relationship Openness Conscientiousness ** 7. Extraversion Agreeableness **.11.29* 9. Neuroticism **.30*.25* Protégés 1. Gender Age Nationality ** 4. Educational level **.27* 5. Preferred choice Meeting frequency Openness Conscientiousness Extraversion *.24* * Agreeableness * Neuroticism **.38** Career support *.34**.17.24* Psychosocial support ** ** Note. n = 68 mentors; n = 68 protégés. Dummy-coded variables: gender (0 = female, 1 = male); nationality (0 = foreign, 1 = Swiss); education level (0 = undergraduate, 1 = graduate); preferred choice (0 = no, 1 = yes). *p <.05. **p <.01. ***p <

14 Menges 111 Table 2. Career and Psychosocial Support Regressed on Similarity in Openness. Career support Psychosocial support Variables entered Step 1 Step 2 Step 3 Step 1 Step 2 Step 3 Mentor age Protégé age Mentor gender Protégé gender.28.31*.30* Paired gender Mentor nationality Protégé nationality Paired nationality Mentor s satisfaction * Protégé educational level * Protégé preferred choice * Meeting frequency.34*.30*.34*.35*.39**.42** Mentor openness.30*.25* Protégé openness *.30* Mentor openness Protégé.30*.27* openness R * ΔR 2.11*.07*.06.06* Note. n = 68 mentoring pairs. Standardized regression weights are shown. p <.10. *p <.05. **p <.01. ***p <.001. career support was provided (β =.30, p <.05), even when the individual effects of mentor and protégé personality, as well as age, gender, gendermatch, nationality, nationality-match, meeting frequency, mentor s satisfaction with the protégé, and protégé s educational level and mentor allocation were taken into account. To interpret the interaction effect of mentor openness and protégé openness on career support, the interaction graph was plotted in Figure 1. This graph shows that protégés low in openness received more career support from mentors who also scored low in openness than from mentors who scored high in openness. Similarly, protégés high in openness received more career support from mentors who also scored high in openness than from mentors who scored low in openness. Thus, Hypothesis 1 was supported in respect to openness to experience. The similarity between mentor and protégé explained an additional 7% of the variance in career support as evidenced by a significant change in R 2 (ΔR 2 =.07, p <.05).

15 112 Group & Organization Management 41(1) Figure 1. Career support as a function of mentor openness and protégé openness. Hypothesis 2 predicted that similarity in personality would be positively related to psychosocial support. Table 2 shows, in support of Hypothesis 2, that the dyadic interaction between mentor openness and protégé openness explained the extent of psychosocial support protégés received (β =.27, p <.05) beyond the effects of control variables. Upon inspection of the interaction graph, however, the similarity in openness appears to matter only when the mentor is low in openness. Specifically, as Figure 2 shows, mentors who scored low in openness provided protégés scoring also low in openness with more psychosocial support than protégés scoring high in openness. When mentors were high in openness, however, protégés openness scores did not matter for psychosocial support. That is, protégés received the same amount of psychosocial support from mentors who scored high in openness regardless of the protégé s openness score. A significant change in R 2 suggests that the similarity between mentor and protégé on openness adds 6% explained variance in psychosocial support (ΔR 2 =.06, p <.05). Conscientiousness and mentoring support. Hypothesis 1 suggested that similarity in conscientiousness would be positively related to career support. However, analyses showed that the interaction between mentor conscientiousness and protégé conscientiousness did not significantly predict the extent of career support, and the R 2 change at Step 3 was not significant. Thus, in respect to conscientiousness, Hypothesis 1 was not supported.

16 Menges 113 Figure 2. Psychosocial support as a function of mentor openness and protégé openness. Hypothesis 2 suggested that similarity in conscientiousness would be positively related to psychosocial support. The results presented in Table 3 support Hypothesis 2 by showing that the interaction between mentor conscientiousness and protégé conscientiousness significantly predicted psychosocial support (β =.27, p <.05), even when control variables were accounted for. Figure 3 presents a graphical plot of the interaction between similarity in conscientiousness and psychosocial support. Protégés low in conscientiousness received more psychosocial support from mentors who also scored low in conscientiousness than from mentors who scored high in conscientiousness. Similarly, protégés high in conscientiousness received more psychosocial support from mentors who also scored high in conscientiousness than from mentors who scored low on conscientiousness. A significant R 2 change indicated that similarity in conscientiousness contributed 6% to the variance explained in psychosocial support (ΔR 2 =.06, p <.05). Apart from these findings concerning the hypotheses, it is worth noting that two other variables had important effects on the mentoring relationship. First, meeting frequency affected both career support and psychosocial support: The more often the mentoring pair met, the more protégés benefitted from the relationship, as evidenced in significant bivariate correlations (career support: r =.34, p <.01; psychosocial support: r =.37, p <.01), these relationships remained significant in the regression models. Second, preferred choice of mentor related positively to career support (r =.27, p <.05);

17 114 Group & Organization Management 41(1) Table 3. Career and Psychosocial Support Regressed on Similarity in Conscientiousness. Career support Psychosocial support Variables entered Step 1 Step 2 Step 3 Step 1 Step 2 Step 3 Mentor age Protégé age Mentor gender Protégé gender.28.29*.29* Paired gender Mentor nationality Protégé nationality Paired nationality Mentor s satisfaction * Protégé educational level Protégé preferred choice Meeting frequency.34*.31*.30*.35*.35*.31* Mentor conscientiousness Protégé conscientiousness Mentor conscientiousness Protégé.06.27* conscientiousness R * ΔR * Note. n = 68 mentoring pairs. Standardized regression weights are shown. p <.10. *p <.05. **p <.01. ***p <.001. this relationship too remained significant or marginally significant in the regression models. Discussion This study showed that similarities in specific personality characteristics between mentors and protégés have an effect on the support that protégés receive from mentors in a formal mentoring program. Specifically, this study found that similarity in openness to experience improved the career support that the protégés received from their mentors, and similarities in openness to experience and conscientiousness enhanced the psychosocial support that protégés received. But just as important as those significant findings are, is the insight that similarity in extraversion, agreeableness, and neuroticism had no effect. It seems, therefore, that the general effect for the benefits of

18 Menges 115 Figure 3. Psychosocial support as a function of mentor conscientiousness and protégé conscientiousness. personality similarity that one would expect based on the attraction-similarity paradigm (Byrne, 1971) does not hold for formal mentoring relationships. Instead, in these specific relationships, distinctly, similarities in openness to experience and conscientiousness matter. Theoretical and Practical Contributions This research holds theoretical and practical contributions both to relational theories concerning similarity attraction and, more specifically, to mentoring theory and practice. First, in a challenge to relational theories advancing a generally beneficial effect of personality similarity on relationship outcomes (Byrne, 1971; Newcomb, 1956), the findings here suggest specific effects: Similarities on two personality traits openness to experience and conscientiousness matter for mentoring relationships, but similarities on other personality traits extraversion, agreeableness, and neuroticism do not matter. These differentiated results corroborate other recent research that challenges the alleged general positive effect of personality similarity. For example, a study on initial interactions between strangers found that these interactions benefitted from similarity in levels of extraversion, but suffered if both interaction partners were similarly low on agreeableness. Similarities concerning other personality traits had little effects on the interactions (Cuperman &

19 116 Group & Organization Management 41(1) Ickes, 2009). In a study about friendship networks, it was discovered that people selected friends with similar levels of agreeableness, extraversion, and openness, but not necessarily with similar levels of conscientiousness and neuroticism (Selfhout et al., 2010). Apparently, the benefits that flow from similarity depend on whether similarity matters in a given context. Thus, this research, together with other recent findings, contributes to advancing a more nuanced understanding of the attraction-similarity paradigm. This research contributes to a better understanding of formal mentoring relationships and provides formal mentoring managers with guidance for the matching process by which protégés are assigned to mentors. It responds to calls for research to examine how protégé and mentor personality characteristics interact with one another (Chandler et al., 2011; Eby & Allen, 2002; Turban & Lee, 2007) and to calls for research on the matching criteria that could effectively be used in the matching process of formal mentoring programs to facilitate beneficial mentoring outcomes (Blake-Beard et al., 2007). Of particular importance is the focus of this study on actual personality in the context of formal mentoring relationships. Whereas prior research established the importance of perceived personality similarity (Turban & Lee, 2007), this research demonstrated the importance of actual personality similarity. The distinction between actual and perceived similarity is essential in the context of formal mentoring programs as matching has to occur prior to the first meeting of the mentoring partners. Actual similarity can be assessed beforehand, so it can be used for the matching process, whereas perceived similarity cannot. Furthermore, the findings corroborate existing studies on perceived similarity in personality that found perceived similarity to have a positive effect on several mentoring outcome variables, such as time spent with protégés (Burke et al., 1993) and intention to remain in the mentoring relationship (Madia & Lutz, 2004). Thus, the current study both adds to and extends existing research highlighting the importance of similarity in personality. This study also raises vital questions for mentoring research. Why is it that protégés receive more mentoring support in formal mentoring relationship if their mentor shares similar, rather than dissimilar, levels of openness to experience and conscientiousness with them? And why is it that extraversion, agreeableness, and neuroticism had no effect? Although this research did not empirically examine the mechanisms that bring about the beneficial effects of similarity, some thoughts can be advanced to tentatively explain these intriguing findings. Why does similarity in openness to experience matter for career and psychosocial support? Openness to experience involves intellectual curiosity, creativity,

20 Menges 117 imagination, open-mindedness, and attentiveness to emotions (McCrae & Costa, 1996). If there are discrepancies at these levels, the mentoring dyad partner with greater openness to experience might be more curious, aim for more creativity, or push the other to convey more, particularly when it comes to emotions, whereas the less open partner may feel overwhelmed by the other partner s intellect and creativity, or be less daring and more selective about shared information. These differences, in turn, could complicate the mentoring relationship and breed conflict, leading the mentor to withdraw career support from the protégé. In contrast, if both mentoring dyad partners share the same levels of curiosity and creativity and are comfortable about the amount of emotions conveyed, then the mentoring partners are likely to establish a mutual understanding and feel at ease in each other s presence. Indeed, a recent study of dyadic interactions between strangers showed that when both strangers scored similarly on openness (i.e., high/high or low/low), the strangers were better able to recognize that their counterpart was attempting to accommodate their own behavior to fit in and facilitate the relationship. In contrast, strangers who scored differently on openness (i.e., high/low) attributed a lower amount of behavioral accommodation to partners (Cuperman & Ickes, 2009). Therefore, it seems that mentors and protégés who share similar levels of openness to experience are likely to better understand each other s efforts to make the mentoring relationship work, particularly given that both mentoring partners usually enter the relationship within a formal mentoring program as strangers. As a result, mentors are likely to provide career support to the protégé. Psychosocial support, too, benefits to the extent that mentors and protégés share similar levels of low openness to experience, but is unaffected in mentoring relationships in which the mentor has high openness to experience (as seen in Figure 2). Such mentors, it seems, can accommodate both protégés low in openness to experience and those high in openness. This finding is consistent with Wanberg et al. s (2006) suggestion that high openness to experience is expected to similarly predispose mentors to be more willing to mentor an individual that is not a mirror reflection of themselves (pp ). However, Wanberg et al. did not find empirical support in their study of formal mentoring relationships even though Bozionelos (2004) had already reported a generally positive effect of mentors openness to experience on a number of mentoring outcomes in informal mentoring relationships. These inconsistencies may be due to the interaction effect revealed here. As the findings of the current study show, psychosocial support is not only linked to the mentor s level of openness to experience, but it also depends on the protégés level of openness. A low level of openness on the side of the mentor is

21 118 Group & Organization Management 41(1) associated with decrements in psychosocial support anticipated by scholars in earlier work only when the protégé is high in openness, but not when the protégé is low in openness. Why does similarity in conscientiousness matter for psychosocial support, but not career support? Conscientiousness refers to a tendency to show self-discipline, act dutifully, and be organized, task-focused, and persistent (McCrae & Costa, 1996). Mentors and protégés with a similar natural understanding of duties and disciplines are likely to get along better than those with dissimilar dispositions; thus, protégés benefit in terms of psychosocial support if they match their mentors level of conscientiousness. This finding is concordant with an earlier study showing that perceived similarity in conscientiousness had a positive impact on mentors expressed intention to remain in a mentoring relationship (Madia & Lutz, 2004). If both parties invest similar amounts of effort and persistence, the relationship is likely to be one of mutual understanding. Mentors and protégés with similar levels of conscientiousness might thus reinforce and encourage each other. In such relationships, mentors are likely to offer psychological nurturance and social support in ways that the protégé appreciates. Indeed, Waters (2004) found mentoring pairs in which both had high levels of actual conscientiousness were likely to hold a shared view of the mentorship. In contrast, if one of the mentoring partners has a stronger task focus and invests more effort than the other, the relationship will likely become burdened by feelings of inequity, thus deteriorating the felt relationship quality for the protégé. For example, if one partner is task-focused but the other is not, then this difference dilutes the meaning of the relationship, causes controversy over the direction and purpose of mentoring meetings, and thus breeds psychosocial conflict. The psychosocial benefits of similarity in conscientiousness do not extend to the domain of career support that is likely to benefit little if both parties are slacking. Better mentoring outcomes are likely to be achieved if at least one of the partners in the mentoring relationship is organized, persistent, and focused on the task. Research showed, for example, that protégés with an internal locus of control (a trait indicative of conscientiousness) were more likely to receive career support than others (Turban & Dougherty, 1994), apparently, independent of mentors level of conscientiousness. Perhaps, then the relational benefits of being similar that facilitate psychosocial support may be compromised by the instrumental benefits for career support that could be derived from the conscientiousness of even just one mentoring partner, leading overall to a null effect of conscientiousness similarity on career support.