Journal of Business Research 64 (2011) 363–370
Contents lists available at ScienceDirect
Journal of Business Research
Effects of mentoring functions on receivers' organizational citizenship behavior in a Chinese context: A two-study investigation Ho Kwong Kwan a,1, Jun Liu b,⁎, Frederick Hong-kit Yim c,2 a b c
Department of Management, Drexel University, Philadelphia, PA 19104, United States School of Business, Renmin University of China, PR China Department of Marketing, Hong Kong Baptist University, Kowloon Tong, Hong Kong
a r t i c l e
i n f o
Article history: Received 1 June 2008 Received in revised form 1 October 2009 Accepted 1 March 2010 Available online 7 May 2010 Keywords: Mentoring functions Organizational citizenship behavior Relationship quality China
a b s t r a c t This research examines the relationship between mentoring functions received by employees and their organizational citizenship behavior (OCB), and the moderating effects of perceived quality relationships with mentoring function providers in China. Results of Study 1 from a sample of 385 supervisor–subordinate dyads reveal that role modeling received by protégés positively relates to their OCB, and perceived mentoring relationship quality moderates the association between career support received and protégés' OCB. Moreover, nonprotégés and protégés who receive high levels of overall mentoring functions perform more OCB than protégés who receive low levels of mentoring functions. Results of Study 2 from 258 supervisor–subordinate dyads show that role modeling received by subordinates positively relates to their OCB directed at individuals (OCBI) and at the organization (OCBO) while career support received positively relates to OCBO. Additionally, leader–member exchange moderates the linkages of role modeling with OCBI and OCBO. © 2010 Elsevier Inc. All rights reserved.
1. Introduction Over the past 20 years, there has been a steady growth of research focusing on mentoring and organizational citizenship behavior (OCB), respectively (Farh et al., 2004; Fletcher and Ragins, 2007). Mentoring functions including career support, psychosocial support, and role modeling strongly provide positive effects on receivers' career success (Scandura and Ragins, 1993). Additionally, OCB is an extra-role and discretionary behavior not directly and explicitly recognized by the formal reward system in the organization including such actions as helping other employees actively and working conscientiously (Farh et al., 2004). Interestingly, an intersection between these two important research streams has been identiﬁed. For example, Tepper and Taylor (2003) show that supervisors' mentoring behavior positively relates to subordinates' OCB. Moreover, Donaldson et al. (2000) demonstrate that the high quality of mentoring relationships perceived by employees facilitates their OCB. In spite of the above encouraging ﬁndings, ﬁve research gaps about mentoring functions and OCB can be identiﬁed. First, there is a paucity of empirical research exploring the impacts of all mentoring functions received by employees on their OCB despite the multidimensional nature of mentoring functions. For instance, the study of Tepper and ⁎ Corresponding author. Tel.: + 86 10 82500496. E-mail address: firstname.lastname@example.org (J. Liu). 1 Tel.: + 1 267 322 0017. 2 Tel.: + 852 3411 7526. 0148-2963/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.jbusres.2010.04.003
Taylor (2003) is limited to a composite measure of mentoring functions. Second, moderating effects have not been examined within this research domain. Third, the association between mentoring functions and OCB has not been examined in the context of mentoring relationships. Fourth, much of the research on mentoring functions and OCB has been based on U.S. samples (Bozionelos and Wang, 2006; Farh et al., 2004). Finally, mentoring research in China has focused on white-collar respondents and it is interesting to conduct studies employing other occupational groups such as blue-collar workers (Bozionelos and Wang, 2006). To address these gaps, this research explores the main effects of different dimensions of mentoring functions on employees' OCB, and the moderating effects of relationship quality in China across two studies. To compare and generalize our ﬁndings in different contexts, we conducted Study 1 employing blue-collar workers in a mentorprotégé context and Study 2 using white-collar employees in a supervisor–subordinate context. The current investigation contributes to the literature on mentoring functions and OCB in several ways. First, it takes a ﬁrst step toward understanding the role of different dimensions of mentoring functions in enhancing employees' OCB. Second, moderating effects of relationship quality are examined according to relational cultural theory. The kernel of relational cultural theory is that human growth takes place mainly in a context of relational connection with others, in contrast to independence and autonomy (Fletcher and Ragins, 2007), underlining the role of relational interactions in the human development process. Both givers and receivers are essential components of mentoring
H.K. Kwan et al. / Journal of Business Research 64 (2011) 363–370
functions, and each party feels responsible to contribute to the growth and development of the other. The efﬁcacy of mentoring functions on outcomes is thus mostly based on the quality of interactions between providers and receivers. Finally, our study employing both blue- and white-collar respondents in China is critical. Scholars have raised the concern that the mentoring and OCB ﬁndings from white-collar respondents in the West may not be generalized to Chinese societies because of the cultural and job role issues (Bozionelos and Wang, 2006; Farh et al., 2004). It is thus timely to assess whether effects of mentoring functions on OCB established in the West hold in China. 2. Conceptual background and hypotheses Generally, three mentoring functions are identiﬁed, namely career support, psychosocial support, and role modeling (Scandura and Ragins, 1993). First, career support functions aim to advance receivers' career involving competence demonstration, understanding enhancement, and promotion nomination. Second, psychosocial support functions target at receivers' subjective needs and identity sense by providing a platform for receivers to share their feelings and personal issues. Third, role modeling functions allow experienced employees to be the ones who are exemplary for their less experienced employees to observe and imitate. These three mentoring functions can be provided by mentors and supervisors. To understand the effect of mentoring on employees' OCB, we develop hypotheses below. Hypothesis 1 focuses on mentoring functions provided by mentors, whereas Hypotheses 2–5 emphasize those performed by both mentors and supervisors. 2.1. Mentoring functions received and OCB Mentoring relationship is an interpersonal relationship between an older and more experienced individual (mentor) and a younger and less experienced individual (protégé) (Kram, 1985). Based on social exchange theory (Blau, 1964), mentoring relationships may act as a vehicle to facilitate protégés' OCB. Organ (1977) contends that employees who are inequitably over-rewarded tend to demonstrate OCB. In fact, mentoring functions received can be viewed as a privilege by protégés because a mentor helps protégés handle their work and enhance their personal learning (Kram, 1985). Moreover, a mentor provides career support, psychosocial support, and role modeling functions which promote protégés' career success (Richard et al., 2002). It is likely that protégés who beneﬁt from mentoring relationships may reciprocate through OCB. However, not all protégés receive the same levels of mentoring functions. A negative experience may appear when protégés receive low levels of mentoring functions and they may believe that they are losers but not beneﬁciaries. To restore equity, protégés who receive low levels of mentoring functions tend to decrease their OCB demonstration.
successful relationships could increase their satisfaction levels (Allen et al., 1999; Brown et al., 2008). This socialization perspective is consistent with the ﬁndings that the relationship between career support functions received by employees and their job satisfaction is signiﬁcantly positive (Lankau and Scandura, 2002). As satisfaction has long been an important antecedent of OCB (Organ, 1977), we propose: Hypothesis 2. Career support received by employees is positively related to their OCB. Psychosocial support received may have positive impact on employees' OCB by enhancing their organizational commitment and perceived procedural justice. Psychosocial support can increase receivers' organizational commitment when givers and receivers share their feelings (Richard et al., 2009). These personal networks tend to attach receivers more closely to the organization (Baugh and Scandura, 1999). Moreover, psychosocial support received can enhance perceived procedural justice because psychosocial support relates to the politics aspect of socialization that provides information concerning power structures and work relationships in the organization (Allen et al., 1999). Junior employees who lack work experience are less likely to understand the functioning of an organization. Experienced employees can explain the rationale behind organizational procedures and their explanation mitigates junior employees' unrealistic expectation of what the organization should give, thereby promoting procedural justice (McManus and Russell, 1997). In fact, the relationship between procedural justice perceived by employees and their OCB is signiﬁcantly positive (Tepper and Taylor, 2003). Hypothesis 3. Psychosocial support received by employees is positively related to their OCB. Mentors and supervisors can foster employees' OCB by serving as role models. This process can be explained by social cognitive theory asserting that people learn various behaviors through observation and imitation of role models (Bandura, 1977). By observing behaviors from mentors or supervisors, employees establish their schemas that direct their behaviors when interacting with people (Ragins and Verbos, 2007). Indeed, informal mentoring provision has been conceptually framed as a speciﬁc form of OCB because the degree of mentoring provision is discretionary and unrelated to formal reward systems (McManus and Russell, 1997). This argument is consistent with ﬁndings that levels of mentoring functions provided by mentors positively relate to their OCB levels (Jandeska and Kraimer, 2005), and mentors tend to be higher in helpfulness than nonmentors (Allen, 2003). Hypothesis 4. Role modeling received by employees is positively related to their OCB.
2.2. Moderating effects of relationship quality Hypothesis 1. Protégés who receive high levels of mentoring functions perform higher levels of OCB than (a) nonprotégés and (b) protégés who receive low levels of mentoring functions. (c) Nonprotégés perform higher levels of OCB than protégés who receive low levels of mentoring functions. The current investigation proposes that the three mentoring functions received by employees relate to their OCB performance individually. Career support received may inﬂuence employees' OCB by enhancing their skills and satisfaction. By acquiring instructions from experienced people, employees can develop their managerial and technical skills (Burke, 1984). In fact, OCB is positively associated with skill levels (George and Jones, 1997) and perceived ability to help (Spector and Fox, 2002). Moreover, career support function provides career direction and socialization which enhance receivers' job satisfaction: for example, socialization that helps employees establish
Relationship quality refers to a global assessment of mutual beneﬁts, effectiveness, and satisfaction with the relationship (Allen and Eby, 2003). Because relational mentoring facilitates trust, respect, empowerment, and effective communications, high-quality relationships are more likely to beneﬁt receivers than low-quality relationships (Fletcher and Ragins, 2007). This contention is consistent with a ﬁnding showing how much protégés can beneﬁt from mentoring is inﬂuenced by the development of mentoring relationship based on trust and openness (Dymock, 1999). As hypothesized above, employees' OCB positively relates to the level of beneﬁts they gain from mentoring functions: Employees will perform more OCB when they beneﬁt more in a high-quality mentoring relationship. Moreover, trust moderates social exchange relationships between OCB and its antecedents in the Chinese context (Wong et al., 2005).
H.K. Kwan et al. / Journal of Business Research 64 (2011) 363–370
Research suggests that people tend to observe and imitate others particularly when the role model is trusted (Swap et al., 2001). When employees believe that their role models are trustworthy, they imitate their respected ones more and display more OCB. Hypothesis 5. The relationships linking employees' (a) career support, (b) psychosocial support, and (c) role modeling received with their OCB are stronger when relationship quality is higher than when relationship quality is lower. 3. Methods (Study 1)
3.2.1. Mentoring functions Mentoring functions were gauged with 9 items developed by Scandura and Ragins (1993). Each mentoring function was assessed by 3 items. The original measurement scale had 15 items. We removed 6 items (3 from career support, 2 from psychosocial support, and 1 from role modeling) because they did not ﬁt with the company condition assessed in the pilot study. For instance, given that the company scheduled the lunch time and location for its employees, we dropped the item “I often go to lunch with my mentor.” A sample item was “My mentor has devoted special time and consideration to my career.” Cronbach's alphas for career support, psychosocial support, and role modeling were 0.86, 0.75, and 0.79, respectively.
3.1. Sample and procedure To test the above hypotheses in a mentor–protégé context, we administered surveys to employees working in a private manufacturer located in China. Two sets of questionnaires were designed for subordinates and their supervisors. Supervisors assessed the OCB levels of one to three subordinates, while subordinates reported whether they had mentors, the mentoring functions received, and the relationship quality, as well as responded to the control variables. This study targeted at all technical workers, totally 1128 people. The job responsibilities of the technical workers were to reﬁne existing models and production process, and to design and make new models and products. All technical workers and their supervisors acquired the survey packets during work hours from data collectors who were trained and led by one of the authors. Survey questionnaires were coded before distribution and an ofﬁcer of human resources department assisted to record the identity numbers and the respondents' names to match supervisor–subordinate dyads. In total, 474 and 451 complete and usable surveys were sent back by mail from subordinates and supervisors resulting in response rates of 42.0% and 40.0%, respectively. After matching pairs of supervisor– subordinate, we obtained 385 usable dyads that comprised the sample for the study. Approximately 62.6% subordinate respondents were male. On average, subordinate respondents were of age 24.78 years (SD = 5.66 years) and tenure 3.92 years (SD = 3.35 years). About 9.4% had a bachelor or postgraduate degree, while 43.4% graduated from technical schools and 34.3% graduated from high schools. Also, 13.0% only received primary school or had no formal education. For supervisor respondents, about 66.1% were male. On average, they were of age 34.53 years (SD = 5.41 years). To evaluate whether respondents had a mentoring relationship, we utilized a well-developed deﬁnition of mentor (Ragins et al., 2000). Mentors may include past senior colleagues and peers in different organizations. Because the company did not implement any formal mentoring programs, we regarded all mentoring relationships as informal. To address the issue of multiple mentors, respondents were told to identify their most inﬂuential mentor. Finally, 240 (62.3%) subordinate respondents were identiﬁed as protégés. Of those protégés, 155 (64.6%) were male. The average age was 22.0 years (SD = 2.9 years) and tenure was 2.6 years (SD = 1.7 years). Protégés reported that 186 (76.3%) mentors were male, and we found that 182 (75.8%) mentoring relationships were same-gender and 58 (24.2%) were cross-gender. 3.2. Measures All scales employed in this study were originally developed in the West. Except the OCB measure that has a Chinese version (Lam et al., 1999), we translated all scales into Chinese. Two Chinese doctoral students in management conducted back-translation to ensure the translation quality (Brislin, 1980). All constructs described below with the exception of control variables were measured on a ﬁve-point Likert-type scale ranging from (1) “strongly disagree” to (5) “strongly agree.”
3.2.2. OCB OCB was assessed with 15 items representing 3 dimensions originally developed by Podsakoff et al. (1990). This study applied altruism, conscientiousness, and civic virtue to represent OCB because these 3 dimensions are common in the West and in China (Farh et al., 2004). A representative item was “S/he is always ready to help or to lend a helping hand to those around him/her.” Cronbach's alphas for altruism, conscientiousness, and civic virtue were 0.87, 0.89, and 0.93, respectively. Because our main focus is the construct rather than the dimensions and studies suggest that OCB is best regarded as a single composite factor (Wong et al., 2008), we aggregated the scores for the three dimensions to form an overall composite measure of OCB. Cronbach's alpha was 0.81. 3.2.3. Relationship quality The quality of mentoring relationship perceived by protégés was measured with 5 items developed by Allen and Eby (2003). The original scale was responded by mentors. This study modiﬁed the scale to ﬁt with protégés' reporting. A sample item was “The mentoring quality between my mentor and I was very effective.” Cronbach's alpha was 0.88. 3.2.4. Control variables We controlled for protégés' demographic variables including gender (coded 0 = male and 1 = female), age, education levels, and organizational tenure (years) because those variables may impact OCB (e.g., Gregersen, 1993). Education levels were hierarchically coded as (1) no formal education, (2) primary school, (3) high school, (4) technical school, (5) bachelor degree, and (6) postgraduate degree. Based on Ragins et al.'s study (2000), we measured a series of mentoring variables and considered same-gender and cross-gender mentoring relationships. However, we found that all the control variables about mentoring were not signiﬁcantly related to protégés' OCB. To preserve power, we only controlled for the demographic variables of protégés in our further analysis. 4. Results (Study 1) We conducted a preliminary analysis of the OCB data considering the potential for data non-independence because in our study, 161 supervisors rated 385 subordinates. We calculated the design effect and obtained 1.54, which is below the conventional cutoff of 2 (http//: www.statmodel.com). Hence, we believe that supervisor-ratings were relatively independent and did not affect the ﬁndings signiﬁcantly. We applied structural equation modeling with LISREL 8.54 to test our measurement model in which the three mentoring functions, relationship quality, and OCB were included. The model yielded an acceptable ﬁt to the data: χ2 (109) = 221.53, p b 0.001; RMSEA = 0.061; SRMR = 0.052; IFI = 0.97; CFI = 0.97; and NNFI = 0.96. The average variance extracted (AVE) ranged from 0.527 to 0.708, and all these estimates exceeded the square of the correlation between the factors making up each pair ranging from 0.001 to 0.423, providing evidence of convergent and discriminant validity. Further, we ran a one-factor
H.K. Kwan et al. / Journal of Business Research 64 (2011) 363–370
Table 1 Means, standard deviations, scale reliabilities, and correlations. Variable
1. 2. 3. 4. 5. 6. 7. 8.
4.24 3.67 3.73 3.50 3.89 3.62 3.63 3.23
0.85 0.75 0.88 0.74 0.74 0.77 0.90 0.93
0.71 0.53 0.57 0.59 0.60
0.40 0.42 0.42 0.10 0.13
Career support Psychosocial support Role modeling OCB Relationship quality Altruism Conscientiousness Civic virtue
(0.86) 0.36⁎⁎ 0.48⁎⁎ 0.21⁎⁎ 0.37⁎⁎ 0.18⁎⁎ 0.21⁎⁎ 0.15⁎
(0.75) 0.52⁎⁎ 0.25⁎⁎ 0.23⁎⁎ 0.19⁎⁎ 0.19⁎⁎ 0.24⁎⁎
(0.79) 0.25⁎⁎ 0.35⁎⁎ 0.18⁎⁎ 0.22⁎⁎ 0.23⁎⁎
(0.81) 0.05 0.83⁎⁎ 0.84⁎⁎ 0.87⁎⁎
(0.88) -0.02 0.07 0.06
(0.87) 0.54⁎⁎ 0.61⁎⁎
Note: N = 240. M = mean; SD = standard deviation; AVE = average variance extracted; LSV = largest shared variance. Cronbach's alpha appears in the brackets along the diagonal. ⁎ p b 0.05 (two-tailed). ⁎⁎ p b 0.01 (two-tailed).
model, yielding an unacceptable ﬁt. Table 1 shows the correlation matrix, means, standard deviations, AVE, and the largest shared variance of all the variables. A univariate analysis of variance (ANOVA) was applied to test Hypothesis 1a, b, c to evaluate the difference of the OCB demonstration among nonprotégés and protégés who receive high versus low levels of overall mentoring functions. We used a median split to divide the protégés into two groups (high vs. low). Consistent with our expectations, the difference among nonprotégés and protégés receiving high versus low levels of mentoring functions regarding OCB was signiﬁcant (F(6, 385) = 4.11, p b 0.01). The levels of OCB performed by nonprotégés (M = 3.60) and by protégés who received high levels of mentoring functions (M = 3.65) were signiﬁcantly higher than the level of OCB performed by protégés who received low levels of mentoring functions (M = 3.28), supporting Hypotheses 1b, c. However, the difference of the OCB performance between nonprotégés and protégés receiving high levels of mentoring functions was insigniﬁcant, failing to support Hypothesis 1a. Hierarchical regression was conducted to test Hypotheses 2–4 (shown in Table 2). Hypotheses 2–4 state that mentoring functions received are associated with protégés' OCB. Model 2 of Table 2 indicates that as predicted, role modeling was positively related to OCB (β = 0.16, p b 0.05), supporting Hypothesis 4. However, career support (β = 0.10, n.s.) and psychosocial support (β = 0.12, n.s.) were not signiﬁcantly related to OCB, failing to support Hypotheses 2 and 3.
To test Hypothesis 5, we introduced the interaction terms into our regression model. Model 4 in Table 2 presents the results. Relationship quality had signiﬁcant moderating effects on the associations of OCB with career support (β = 0.24, p b 0.05) but not with psychosocial support (β = 0.02, n.s.) and role modeling (β = −0.13, n.s.), providing support for Hypothesis 5a but not for b and c. The nature of the signiﬁcant interaction was examined by plotting values plus and minus one standard deviation from the means of career support and relationship quality. Fig. 1 demonstrates that as expected, when levels of relationship quality were high, career support was positively related to OCB. In contrast, when levels of relationship quality were low, career support was not signiﬁcantly related to OCB. The results of Study 1 provide support to several hypotheses by demonstrating that role modeling received is positively associated with the protégés' OCB and that the relationship between career support received and OCB is moderated by relationship quality. Yet, there are several limitations in Study 1. First, all respondents were from unskilled workers in a private company in a mentor–protégé context. Samples from white-collar employees in state-owned enterprises in the supervisor–subordinate context are needed. Second, we collected data at the same time. Third, research shows that OCB can be divided into OCB directed at individuals (OCBI) and at the organization (OCBO) whose antecedents are not the same (Williams and Anderson, 1991). Finally, we did not control for any cultural variables that may impact our results. To address the above-mentioned limitations, we conducted Study 2. 5. Methods (Study 2)
Table 2 Results of regression analyses.
5.1. Sample and procedure Dependent variable: OCB Model 1 Model 2
Control variables Protégé's gender − 0.11 Protégé's age − 0.04 Protégé's education levels 0.12 Protégé's organizational tenure − 0.04 Independent variables Career support Psychosocial support Role modeling Moderating variable Relationship quality Interactions Career support × relationship quality Psychosocial support × relationship quality Role modeling × relationship quality F-value 1.62 R2 0.03 2 ΔR N = 240. ⁎ p b 0.05 (two-tailed). ⁎⁎ p b 0.01 (two-tailed).
− 0.14⁎ − 0.07 0.13 − 0.02
− 0.13⁎ − 0.07 0.14 − 0.03
− 0.13⁎ − 0.06 0.13 − 0.05
0.10 0.12 0.16⁎
0.12 0.13 0.18⁎
0.12 0.14 0.20⁎
Respondents were white-collar employees and their direct supervisors in two state-owned gas and petroleum enterprises in China. Three waves of data collection were performed to acquire the predictors, moderator, and consequences at different time points. In the ﬁrst-wave survey (T1), the subordinates provided information on
− 0.06 0.24⁎ 0.02
4.58⁎⁎ 0.12 0.09⁎⁎
4.27⁎⁎ 0.13 0.01
− 0.13 3.87⁎⁎ 0.16 0.03⁎ Fig. 1. The moderation effect of relationship quality between career support received and OCB.
H.K. Kwan et al. / Journal of Business Research 64 (2011) 363–370
their own demographics, mentoring functions received, and control variable (i.e., power distance). In the second-wave survey (T2, 4 months after T1), the subordinates were asked to report their relationship quality with supervisors (i.e., leader–member exchange). In the third-wave survey (T3, 4 months after T2), their supervisors reported subordinate OCBI and OCBO. We ﬁrst acquired a list of 443 subordinates and 120 corresponding supervisors. In wave one, 332 subordinate questionnaires were returned. After deleting incomplete cases, a total of 321 usable questionnaires were obtained. Four months later, questionnaires were distributed to the 321 subordinates (as one subordinate had left, 320 subordinates responded to the questionnaires). In total, 273 questionnaires were returned among which 268 were usable. In wave three, questionnaires were distributed to their 105 supervisors (we conﬁrmed from the human resources department that all subordinates did not change their supervisors across Times 1, 2, and 3). In total, 230 subordinates and 102 supervisors returned the questionnaires. After deleting incomplete cases and matching subordinates with their supervisors, the ﬁnal sample consisted of 258 matched supervisor– subordinate dyads, including 258 subordinates and 102 supervisors. Each supervisor rated one to four subordinates. The valid response rates were 58.2% for subordinates and 97.1% for supervisors. Of the 258 subordinates, 50.0% were male. The average age was 36.57 years (SD = 6.43) and the average organizational tenure was 9.97 years (SD = 5.10). About 36.4% had a bachelor degree and 10.5% had a postgraduate degree, while 53.1% graduated from technical schools or received lower education. Of the 102 supervisors, 59.8% were male. Supervisor average age was 38.52 years (SD = 6.28) and the average organizational tenure was 10.95 years (SD = 4.93). 5.2. Measures 5.2.1. Mentoring functions Mentoring functions were gauged with 15 items developed by Scandura and Ragins (1993). We replaced ‘mentor’ with ‘direct supervisor’ to ﬁt with the supervisor–subordinate context. A sample item was “My direct supervisor gives me special coaching on the job.” Cronbach's alphas for career support, psychosocial support, and role modeling were 0.87, 0.83, and 0.86, respectively. 5.2.2. OCBI and OCBO OCBI and OCBO were assessed with 7 items, respectively, that were developed by Williams and Anderson (1991). A sample item was “S/he helps others who have been absent.” Cronbach's alphas for OCBI and OCBO were 0.88 and 0.90, respectively. 5.2.3. Leader–member exchange (LMX) LMX was measured with a 6-item scale that was developed by Scandura and Graen (1984). A sample item was “I would characterize my working relationship with my direct supervisor as highly effective.” Cronbach's alpha was 0.90.
5.2.4. Control variables We added a six-item measure of power distance as a control variable that was originally developed by Dorfman and Howell (1988). A sample item was “Managers should make most decisions without consulting subordinates.” Cronbach's alpha was 0.90. 6. Results (Study 2) We conducted an analysis of OCBI and OCBO for data nonindependence because 102 supervisors rated 258 subordinates. We calculated the design effect and obtained 1.16 for OCBI and 1.21 for OCBO, showing that supervisor-ratings were relatively independent. We tested a proposed seven-factor CFA model, in which the three mentoring functions, LMX, OCBI, OCBO, and power distance were included. The model yielded a good ﬁt to the data: χ2 (758) = 997.92, p b 0.001; RMSEA = 0.034; SRMR = 0.036; IFI = 0.98; CFI = 0.98; NNFI = 0.98. The AVE ranged from 0.515 to 0.607, and all these estimates exceeded the square of the correlation between the factors making up each pair ranging from 0.084 to 0.476. Moreover, we ran a ﬁve-factor model combining all three mentoring functions into a composite measure, yielding an acceptable ﬁt to the data: χ2 (769)= 1,437.59, p b 0.001; RMSEA = 0.068; SRMR = 0.045; IFI = 0.95; CFI = 0.95; NNFI = 0.94. However, the AVE of the overall mentoring function scale was only 0.40, which was below the basic requirement of 0.50 (Fornell and Larcker, 1981). Hence, the convergent validity of the overall mentoring function measurement may be problematic which provides further support for the multidimensional nature of the mentoring functions. Furthermore, we ran a one-factor modeling combining all constructs and a six-factor model combining OCBI and OCBO. The ﬁt indices were not as good as those of the proposed model. Table 3 displays the correlation matrix, means, standard deviations, AVE, and the largest shared variance of all the variables. Hierarchical regression was conducted to test Hypotheses 2–4 (shown in Tables 4, 5). The second models of Tables 4 and 5 indicate that as predicted, role modeling was positively associated with OCBI (β = 0.16, p b 0.05) and OCBO (β = 0.15, p b 0.05), supporting Hypothesis 4. Moreover, career support was positively related to OCBO (β = 0.16, p b 0.05) but not to OCBI (β = 0.03, n.s.), partially supporting Hypothesis 2. Moreover, psychosocial support was not signiﬁcantly related to OCBI (β = 0.08, n.s.) and OCBO (β = −0.01, n.s.), failing to support Hypothesis 3. To test Hypothesis 5, we introduced interaction terms into our regression model. The fourth models in Tables 4, 5 present the results. LMX had signiﬁcant moderating effects on the associations of role modeling with OCBI (β = 0.23, p b 0.01) and OCBO (β = 0.30, p b 0.001), supporting Hypothesis 5c. However, LMX did not provide any moderating effects on the associations of career support with OCBI (β = 0.00, n.s.) and OCBO (β = −0.07, n.s.), as well as on those of psychosocial support with OCBI (β = −0.06, n.s.) and OCBO (β = 0.02, n.s.), failing to support Hypothesis 5a, b. Figs. 2 and 3 demonstrate that as expected, when levels of LMX were high, role modeling was
Table 3 Means, standard deviations, scale reliabilities, and correlations. Variable
1. 2. 3. 4. 5. 6. 7.
3.29 3.31 3.42 3.17 3.20 3.69 3.68
0.58 0.67 0.73 0.72 0.70 0.64 0.71
0.40 0.54 0.56 0.56 0.61 0.52 0.57
0.06 0.45 0.45 0.30 0.08 0.48 0.48
Overall mentoring functions Career support Psychosocial support Role modeling LMX OCBI OCBO
1 (0.91) 0.85⁎⁎ 0.82⁎⁎ 0.82⁎⁎ 0.39⁎⁎ 0.23⁎⁎ 0.25⁎⁎
2 (0.87) 0.57⁎⁎ 0.49⁎⁎ 0.32⁎⁎ 0.16⁎⁎ 0.23⁎⁎
(0.83) 0.54⁎⁎ 0.34⁎⁎ 0.19⁎⁎ 0.16⁎⁎
(0.86) 0.33⁎⁎ 0.22⁎⁎ 0.22⁎⁎
(0.90) 0.26⁎⁎ 0.23⁎⁎
Note: N = 258. Cronbach's alpha appears in the brackets along the diagonal. M = mean; SD = standard deviation; AVE = average variance extracted; and LSV = largest shared variance. LMX = leader–member exchange; OCBI = organizational citizenship behavior directed at individuals; and OCBO = organizational citizenship behavior directed at the organization. ⁎⁎ p b 0.01 (two-tailed).
H.K. Kwan et al. / Journal of Business Research 64 (2011) 363–370
Table 4 Results of regression analyses. Dependent variable: OCBI
Control variables Subordinate gender Subordinate age Subordinate education levels Subordinate organizational tenure Power distance Independent variables Career support Psychosocial support Role modeling Overall mentoring functions Moderating variable LMX Interactions Career support × LMX Psychosocial support × LMX Role modeling × LMX Overall mentoring functions × LMX F-value R2 ΔR2
− 0.06 − 0.05 0.07 0.09 0.02
− 0.05 − 0.04 0.04 0.09 0.01
− 0.03 − 0.06 0.04 0.09 − 0.00
− 0.02 − 0.08 0.03 0.09 0.04
− 0.05 − 0.04 0.04 0.09 0.00
− 0.03 − 0.06 0.03 0.09 − 0.00
− 0.03 − 0.07 0.03 0.10 0.02
− 0.00 0.05 0.13
0.01 0.07 0.12
0.03 0.08 0.16⁎
3.69⁎⁎ 0.09 0.03⁎⁎
0.14⁎ 3.90⁎⁎⁎ 0.11 0.02⁎
0.00 − 0.06 0.23⁎⁎ 0.77 0.02
2.22⁎ 0.07 0.05⁎⁎
3.01⁎⁎ 0.10 0.03⁎⁎
3.32⁎⁎⁎ 0.14 0.04⁎⁎
2.74⁎ 0.06 0.04⁎⁎
N = 258. LMX = leader-member exchange and OCBI = organizational citizenship behavior directed at individuals. ⁎ p b 0.05 (two-tailed). ⁎⁎ p b 0.01 (two-tailed). ⁎⁎⁎ p b 0.001 (two-tailed).
positively related to OCBI and OCBO. In contrast, when levels of LMX were low, role modeling was not signiﬁcantly related to OCBI and OCBO. 7. Discussion This study examines whether mentoring functions received by Chinese employees inﬂuence their OCB, speciﬁcally in a mentor– protégé (Study 1) and a supervisor–subordinate (Study 2) context.
Our results are consistent with social cognitive theory that role modeling received by employees positively relates to their supervisor-rated OCB in Study 1 and to their OCBI and OCBO in Study 2 and that career support positively relates to OCBO in Study 2. Moreover, with respect to relational cultural theory, we found that mentoring relationship quality moderates the relationship between career support received and protégés' OCB in Study 1, and that LMX moderates the relationships between role modeling received and subordinates' OCBI and OCBO in Study 2. Finally, consistent with social
Table 5 Results of regression analyses. Dependent variable: OCBO
Control variables Subordinate gender Subordinate age Subordinate education levels Subordinate organizational tenure Power distance Independent variables Career support Psychosocial support Role modeling Overall mentoring functions Moderating variable LMX Interactions Overall mentoring functions × LMX Career support × LMX Psychosocial support × LMX Role modeling × LMX Overall mentoring functions × LMX F-value R2 ΔR2
− 0.01 − 0.04 0.08 0.09 0.05
0.01 − 0.03 0.03 0.08 0.03
0.02 − 0.04 0.03 0.08 0.02
0.02 − 0.06 0.02 0.08 0.07
0.00 − 0.02 0.04 0.08 0.03
0.02 − 0.04 0.04 0.08 0.02
0.02 − 0.06 0.03 0.10 0.05
0.16⁎ − 0.01 0.15⁎
0.13 − 0.04 0.13
0.15⁎ − 0.01 0.11
− 0.07 0.02 0.30⁎⁎⁎ 0.64 0.01
2.56⁎ 0.08 0.07⁎⁎
2.91⁎⁎ 0.10 0.02⁎
4.21⁎⁎⁎ 0.17 0.07⁎⁎⁎
N = 258. LMX = leader-member exchange and OCBO = organizational citizenship behavior directed at the organization. ⁎ p b 0.05 (two-tailed). ⁎⁎ p b 0.01 (two-tailed). ⁎⁎⁎ p b 0.001 (two-tailed).
3.17⁎⁎ 0.07 0.06⁎⁎⁎
3.50⁎⁎ 0.09 0.02⁎
0.20⁎⁎ 4.47⁎⁎⁎ 0.13 0.04⁎⁎
H.K. Kwan et al. / Journal of Business Research 64 (2011) 363–370
Fig. 2. The moderation effect of LMX between role modeling received and OCBI.
Fig. 3. The moderation effect of LMX between role modeling received and OCBO.
exchange theory, nonprotégés and protégés who receive high levels of overall mentoring functions perform more OCB than protégés who receive low levels of overall mentoring functions in Study 1. Inconsistent with our predictions, we could not ﬁnd direct linkages of career and psychosocial support with protégés' OCB in Study 1 and associations of career support with subordinates' OCBI as well as those of psychosocial support with OCBI and OCBO in Study 2. These results may be explained in several ways. First, a meta-analysis shows that psychosocial support relating to psychological states mainly inﬂuences employees' affect rather than behaviors (Eby et al., 2008). Hence, psychosocial support tends to inﬂuence employees' job satisfaction rather than OCB. Second, career support relates to OCBO but not to OCBI because from the view of employees, mentors and supervisors may represent the organization providing career functions. Receivers may feel more of a sense of obligation to give back to the organization than to individuals. Another unexpected result is that nonprotégés and protégés who receive higher levels of overall mentoring functions perform similar levels of OCB. It may be because nonprotégés can obtain mentoring functions from other social networks in Chinese societies (Bozionelos and Wang, 2006). We could not ﬁnd any moderating effects of relationship quality on the association between psychosocial support and OCB. It is possible that career and role modeling functions are less emotionally involved than psychosocial functions, and thus relationship quality plays a more important role in the impact of career and role modeling functions received. Moreover, relationship quality moderates the linkage of role modeling with OCB in Study 2 but not in Study 1. The reason may be that relationship quality is more important for role modeling functions in the supervisor–subordinate condition than in the mentor-protégé condition. Employees actively establish informal mentoring relationships which may imply that mentors and protégés like each other or have common attributes. However, the relationship between supervisors and subordinates is formally assigned by the organization without any emotional foundation. Hence, relationship quality plays a relatively more critical role in supervisor–subordinate relationships. Finally, career support does not have main effects on OCB in Study 1 but have effects on OCBI in Study 2. Our ﬁrst explanation is that we did
not divide OCB into OCBI and OCBO in Study 1 which masked the effect of career support on OCBI. Another explanation is that Study 2 applied an 8-month longitudinal design that provided enough time for employees to digest what they had received and obtained obvious beneﬁts from career functions. The results of this study make important contributions to the literature on mentoring and OCB in four ways. First, past research on mentoring and OCB ignored the multidimensionality of mentoring functions. This study found that role modeling received signiﬁcantly relates to employees' OCB and that career support relates to subordinates' OCBI. Second, we reveal the signiﬁcant moderating effects of relationship quality on the relationships of career support and role modeling received with OCB. As noted above, relationship quality is critical to variables with low levels of emotions involved. This ﬁnding may be extended to other work relationships such that relationship quality can positively moderate the relationship between emotion-absent variables and behavioral outcomes. Third, we uncover the negative results of receiving low levels of mentoring functions. Our study shows that protégés low in mentoring functions received perform less OCB than nonprotégés, drawing attention to the cost of mentoring. Finally, this study extends the scope of mentoring investigation in China by including unskilled workers to demonstrate that career support and role modeling received impact employees' OCB. Our ﬁndings can extend existing theories and empirical ﬁndings from a developed Western nation to a transition economy. We are also cognizant of the limitations associated with this research. First, the mentoring system in this study is informal. Second, our respondents were from manufacturing ﬁrms compared to the Western counterparts. It is necessary to test our hypotheses in a formal mentoring setting within various industries. Despite these limitations, the present study has demonstrated that career support and role modeling received play a key role in employees' OCB. Many studies show that mentoring functions facilitate employees' career success (Kram, 1985). The reciprocation from employees, however, has not been adequately addressed. If receiving mentoring functions has positive implications for employees' OCB, organizations will be more likely to promote mentoring. Acknowledgements The work by Jun Liu was supported by the Program for New Century Excellent Talents (NCET) in Chinese Universities. The authors would like to thank the guidance from Jeffrey Greenhaus, Chun Hui, Frank Linnehan, and Laci Rogers on early drafts of this paper. References Allen TD. Mentoring others: a dispositional and motivational approach. J Vocat Behav 2003;62:134–54. Allen TD, Eby LT. Relationship effectiveness for mentors: factors associated with learning and quality. J Manage 2003;29:469–86. Allen TD, McManus SE, Russel JEA. Newcomer socialization and stress: formal peer relationships as a source of support. J Vocat Behav 1999;54:453–70. Bandura A. Self-efﬁcacy: toward a unifying theory of behavioral change. Psychol Rev 1977;84:191–215. Baugh SG, Scandura TA. The effects of multiple mentors on protégé attitudes toward the work setting. J Soc Behav Pers 1999;14:503–21. Blau P. Exchanges and power in social life. New York: Wiley; 1964. Bozionelos N, Wang L. The relationship of mentoring and network resources with career success in the Chinese organizational environment. Int J Hum Resour Manag 2006;17:1531–46. Brislin RW. Translation and content analysis of oral and written materials. In: Triandis HC, Berry JW, editors. Handbook of cross-cultural psychology, vol. 2: methodology. Boston: Allyn & Bacon; 1980. p. 389–444. Brown BP, Zablah AR, Bellenger DN. The role of mentoring in promoting organizational commitment among black managers: an evaluation of the indirect effects of racial similarity and shared racial perspectives. J Bus Res 2008;61:732–8. Burke RJ. Mentors in organizations. Group Organ Stud 1984;9:353–72. Donaldson SI, Ensher EA, Grant-Vallone EJ. Longitudinal examination of mentoring relationships on organizational commitment and citizenship behavior. J Career Dev 2000;26:233–49.
H.K. Kwan et al. / Journal of Business Research 64 (2011) 363–370
Dorfman PW, Howell JP. Dimensions of national culture and effective leadership in patterns. Adv Int Comp Manage 1988;3:127–50. Dymock D. Blind date: a case study of mentoring as workplace learning. J Workplace Learn Employ Couns Today 1999;11:312–7. Eby LT, Allen TD, Evans SC, Ng T, DuBois DL. Does mentoring matter? A multidisciplinary meta-analysis comparing mentored and non-mentored individuals. J Vocat Behav 2008;72:254–67. Farh J-L, Zhong C-B, Organ DW. Organizational citizenship behavior in the People's Republic of China. Organ Sci 2004;15:241–53. Fletcher JK, Ragins BR. Stone center relational cultural theory: a window on relational mentoring. In: Ragins BR, Kram KE, editors. The handbook of mentoring at work: research, theory and practice. Thousand Oaks, CA: Sage; 2007. p. 373–99. Fornell C, Larcker DF. Evaluating structural equation models with unobservable variables and measurement error. J Mark Res 1981;18:39–50. George JM, Jones GR. Organizational spontaneity in context. Hum Perform 1997;10:153–70. Gregersen HB. Multiple commitments at work and extrarole behavior during three stages of organizational tenure. J Bus Res 1993;26:31–47. Jandeska KE, Kraimer ML. Women's perceptions of organizational culture, work attitudes, and role-modeling behaviors. J Manag Issues 2005;17:461–78. Kram KE. Mentoring at work: developmental relationships in organizational life. Glenview, IL: Scott, Foresman; 1985. Lam SSK, Hui C, Law KS. Organizational citizenship behavior: comparing perspectives of supervisors and subordinates across four international samples. J Appl Psychol 1999;84:594–601. Lankau MJ, Scandura TA. An investigation of personal learning in mentoring relationships: content, antecedents, and consequences. Acad Manage J 2002;45:779–90. McManus SE, Russell JEA. New directions for mentoring research: an examination of related constructs. J Vocat Behav 1997;51:145–61. Organ DW. A reappraisal and reinterpretation of the satisfaction–causes–performance hypothesis. Acad Manage Rev 1977;2:46–53. Podsakoff PM, MacKenzie SB, Moorman RH, Fetter R. Transformational leader behaviors and their effects on followers' trust in leader, satisfaction, and organizational citizenship behaviors. Lead Quart 1990;1:107–42.
Ragins BR, Verbos AK. Positive relationships in action: relational mentoring and mentoring schemas in the workplace. In: Dutton J, Ragins BR, editors. Exploring positive relationships at work: building a theoretical and research foundation. Mahwah, NJ: Lawrence Erlbaum; 2007. p. 91-116. Ragins BR, Cotton JL, Miller JS. Marginal mentoring: the effects of type of mentor, quality of relationship, and program design on work and career attitudes. Acad Manage J 2000;43:1177–94. Richard OC, Taylor EC, Barnett T, Nesbit MFA. Procedural voice and distributive justice: their inﬂuence on mentoring career help and other outcomes. J Bus Res 2002;55: 725–35. Richard OC, Ismail KM, Bhuian SN, Taylor EC. Mentoring in supervisor-subordinate dyads: antecedents, consequences, and test of a mediation model of mentorship. J Bus Res 2009;62:1110–8. Scandura TA, Graen GB. Moderating effects of initial leader–member exchange status on the effects of a leadership intervention. J Appl Psychol 1984;69:428–36. Scandura TA, Ragins BR. The effects of sex and gender role orientation on mentorship in male-dominated occupations. J Vocat Behav 1993;43:251–65. Spector PE, Fox S. An emotion-centered model of voluntary work behavior: some parallels between counterproductive work behavior and organizational citizenship behavior. Hum Resour Manage Rev 2002;12:269–92. Swap W, Leonard D, Shields M, Abrams L. Using mentoring and storytelling to transfer knowledge in the workplace. J Manage Inf Syst 2001;18:95-114. Tepper BJ, Taylor EC. Relationships among supervisors' and subordinates' procedural justice perceptions and organizational citizenship behaviors. Acad Manage J 2003;46:97-105. Williams LJ, Anderson SE. Job satisfaction and organizational commitment as predictors of organizational citizenship and in-role behaviors. J Manage 1991;17:601–17. Wong Y-T, Wong C-S, Ngo H-Y, Lui H-K. Different responses to job insecurity of Chinese workers in joint ventures and state-owned enterprises. Hum Relat 2005;58: 1391–418. Wong C-S, Law KS, Huang G-H. On the importance of conducting construct-level analysis for multidimensional constructs in theory development and testing. J Manage 2008;34:744–64.