Convergence of In-Group Leader-Member Exchange Dealings and Mentoring - Essay Example

Running Head: DISSERTATION Dissertation of the of the Executive Summary This study explored the theoretical convergence of in-group Leader-Member Exchange dealings and mentoring by investigating how participants described the two types of relationships with supervisors. In-and out-group leader-member relationships were compared on measures designed to identify mentoring relationships. Reported patterns of social power used with in-group members and with protégés were also compared. Seventy-two employed people completed the Leader-Member Exchange scale. Career Support Scale, and power scales. Results indicated that in-group members described their relationships with their supervisors as a mentoring one whereas out-group members did not. A similar pattern of social influence was also reported by those who were identified as in-group members and as protégés. In-group members and mentored individuals reported greater use of expert and referent power than out-group members and non-mentored people. Further, out-group members and non-mentored individuals both reported a greater use of coercive power than did in-group members and protégés. Table of Contents Leader- Members Relationship 1. Introduction Research interest in mentor/protégé relationships developed during the 1980s. It arose largely without articulated conceptual links to other topics in social/organizational psychology such as leadership, social influence, pro-social behavior, or social exchange. The first studies of this developmental relationship were descriptive and exploratory, such as Krams early research (2003) which identified two functions that mentors provide their protégés: career and psychosocial. More recent research (Noe, 1998a, 458) has examined the process and outcome of formal assigned mentor-protégé pairings and, in particular, the difference in outcomes between formal and naturally developing informal mentoring relationships (Chao, Walz, & Gardner, 2002, 619). Noe (1998b, 68) reviewed work on mentoring as it pertains to womens career development. He recommended clarifying the nature of the mentoring construct, suggesting several research questions about the development and effectiveness of these relationships. Noe also stated that "the use of leader-member exchange theory may further understanding of the dynamics involved in the formation of mentorships" (p. 73). Consistent with Noes recommendations, the present study was designed to investigate the convergence of mentoring relationships with another construct in social/organizational psychology: high-quality leader-member exchange (LMX). To explore the convergence of LMX and mentoring, we propose that in-group members (those with high-quality exchange relationships) possess protégé-like relationships with their supervisors, whereas out-group members relationships with their supervisors do not have the qualities of mentoring relationships. 1.1 Leader-Member Exchange Leader-member exchange (LMX) focuses on the quality of the relationship between a leader and individual members of the group. According to Graens model (1996, 1201), a leader behaves differently with each individual group member. This contrasts with traditional theories which imply that leaders behave the same towards all members of the group. The LMX model proposes that two sub-groups develop, each characterized by a different type of exchange with the leader. The ingroup consists of the "informal assistants" and the out-group consists of the remainder of the work group. Liden and Graen (2000, 451) assert that in-group members "make contributions that go beyond their formal job duties and take on responsibility for the completion of tasks that are most critical to the success of the unit…[and out-group members]… perform the more routine, mundane tasks of the unit and experience a more formal exchange with the supervisor" (p. 452). Furthermore, as a reward ingroup members "receive more attention, support, and sensitivity from their supervisors" (p. 452). 1.2 Mentoring The mentor-protégé relationship appears to be very similar in quality to in-group exchange. Based on her study of mentor-protégé pairs, Kram (2005) proposed that mentors engage in two types of roles with respect to their protégés: career-enhancing and psychosocial. Career-enhancing functions include sponsorship, exposure/visibility, coaching, protection, and providing challenging assignments. Psychosocial functions consist of role modeling, acceptance/confirmation, counseling, and friendship. Empirical support for these dimensions has been summarized by Chao et al. (2002, 621). Kram (2000), Phillips-Jones (2002, 22), and others have described types of help or benefits mentors provide to protégés. These include teaching the protégé technical aspects of the organization and profession, which may include its political workings and facilitating competence by exhibiting trust and confidence in the protégé. Other types of help are sharing information that is confidential and sensitive, encouraging and praising the protégé, and highlighting the protégés accomplishments. 1.3 Power Mentors relationships with protégés can be understood by viewing these types of help as sources of social power as conceptualized by French and Raven (1999, 150) and Raven (2002, 217). For example, mentors exercise expert power over protégés when teaching them technical aspects of the organization and when sharing confidential information (McGeorge, 23). Mentors who exhibit trust and confidence in protégés may develop referent power, and those who encourage and praise protégés make use of reward power. French and Ravens (1999, 155) treatment of the bases of social power has been the dominant framework of power used in organizational behavior research. Five bases of power were originally identified: reward, coercive, legitimate, expert, and referent. Since Ravens 1995 study, informational power has also been described. 1.4 Leader-Member Exchange (LMX), Mentoring, and Power Evidence for the similarity of in-group and mentoring relationships can be seen in two different aspects of the relationships: the benefits received by in-group members/ protégés and the pattern of social influence used with in-group members/ protégés. Benefits. The benefits received by in-group members and mentored individuals are very similar. Graen and Cashman (1995, Press) stated that "in-group members [high LMX relationships] received greater latitude in developing their roles, more inside information, greater influence in decision making, stronger support for their actions, and more consideration for their feelings than did out group members" (p. 146). Additional benefits are increased attention, support and sensitivity received from supervisors by in-group members (Liden & Graen, 2000, 457). A second similarity between in-group and mentoring relationships is the pattern of social influence used by supervisors with in-group members and by mentors with protégés. Graen and Cashman (1995, Press) found that members and leaders reported greater use of referent and expert power, and less use of coercive power, with in-group members than with out-group members. The nature of the mentoring relationship suggests that referent and expert power may be prominent sources of influence with protégés, whereas coercive power may more frequently be used with non- protégés. 1.5 Hypotheses Four hypotheses will be tested in the study. First, individuals who describe themselves as being in high LMX relationships (the in-group) will also describe themselves as being mentored. Thus, individuals LMX scores will be positively correlated with scores on a measure of mentoring (the Career Support Scale, CSS; Riley & Wrench, 1995, 375). Second, people categorized as possessing high LMX relationships will report more characteristics of a mentoring relationship than individuals categorized as having low LMX relationships. Thus, individuals categorized as in-group members will have higher scores on the Provisions, Emotion, and Self-Concept subscales of the CSS than individuals categorized as being out-group members. Because both in-and out-group members would be expected to perceive their supervisors as having access to more organizational resources than they, scores on the Resources subscale of the CSS are not predicted to differ between in-and out-groups. Third, following Graen and Cashmans (1995, Press) results, individuals categorized as in-group members will report a greater use of referent and expert power by their supervisors than individuals categorized as out-group members. Further, in-group members will report less use of coercive power by their supervisors than out-group members. Fourth, those categorized as being in a mentoring relationship will report a pattern of supervisory influence similar to that reported by individuals categorized as possessing high LMX relationships: greater use of referent and expert power and less use of coercive power than reported by those in a non-mentoring relationship. 2. Methodology 2.1 Participants The participants were mid-level human resource managers recruited through a local professional organization and through contacts at two public-sector and two private-sector organizations located in the Southeast. One hundred twenty-two packets were distributed and 72 were returned for a response rate of 59%. Of the 70 people who indicated their sex, 41 % (n = 29) were women and 59% (n = 41) were men. The mean age of the sample was 34 years (SD = 9.6). The mean length of tenure with the current organization was six years (SD - 7.3), with half the sample reporting tenure of three years or less. Participation was both voluntary and anonymous. No participant indicated that he or she had taken part in a formal mentoring program. 2.2 Instruments Leader-member exchange (LMX) was measured with the Scandura and Graen (1994, 1588) scale consisting of seven 4-point Likert-type items which address aspects of the leader-member relationship such as mutual communication and support. For example, items ask about how well the respondents supervisor "recognizes your potential" or "understands your problems and needs." Following the recommended methodology (G.B. Graen, personal communication, 1991), individuals item scores were summed and the totals divided at the median of 22.5, with high scores indicating in-group membership and low scores indicating out-group membership. The LMX scale showed high internal consistency (Cronbachs α = .87). Similar results were found by Scandura and Graen, who reported Cronbachs α = .86 at time 1 and .84 at time 2. Mentoring was assessed using the Career Support Scale (CSS) developed by Riley and Wrench (1995, 377). The CSS uses a 5-point Likert-type format with 29 items describing the respondents role partner in terms of personal qualities and behaviors. Sample items ask about whether the person "sets challenging performance standards…to follow" and "assists…in learning the technical aspects of (the) job." Five response alternatives range from "strongly disagree" to "strongly agree" with high scores indicating agreement with the item. Scoring of the CSS as done by Riley and Wrench provides two groups of individuals: those truly mentored and those who are not truly mentored. Someone who is truly mentored has an average item score totaling 3.5 or higher (on a 5-point scale) on each of the four subscales. This method of scoring appeared inappropriate for this study because the Resources subscale was not included in our hypotheses and because of the conventional methodology for testing relationships with LMX. Thus the scoring of the CSS was modified. For correlational analyses in this study, the CSS measure was treated as a continuum by summing the items to compute the subscale scores and the total CSS score. For other analyses, individuals whose CSS score was above the median of 100.5 were categorized as mentored while individuals with scores below the median were categorized as not mentored. The CSS showed high internal consistency (Cronbachs a = .96). Power was assessed by the scale developed by Hinkin and Sehriesheim (1999, 565) which consists of twenty 5-point Likert-type items, with four items for each base of power. For example, respondents indicate whether the role partner can "increase my pay level" or "give me good technical suggestions." Response alternatives range from strongly agree to strongly disagree, with high scores indicating agreement. Hinkin and Sehriesheim state, "the new scales have built-in content validity (cf. Nunnally, 1998, 58) and that they have demonstrated reasonable internal consistency reliability, factor structure, discriminant validity, and criterion-related validity" (p. 566). Scoring of the scales consisted of summing individual items for each base of power. In the current study, results indicated that the power scales maintained high internal consistency (Cronbachs αs for reward = .81, coercive = .87, legitimate = .89, expert = .84, and referent = .92). These internal consistencies are similar to the results obtained by Hinkin and Schriesheim (1999, 567). 2.3 Procedure Participants were given a packet containing information and materials needed to participate in the study. The packets were distributed to the members of the professional organization at their monthly meeting and were mailed to the employees who were recruited at their work sites. Each packet contained a cover letter; a form for requesting results; a demographic sheet; the LMX, CSS, and power scales; and an addressed, stamped envelope to return the materials to the researcher. 3. Results There were no sex effects on any of the measures except the reward power scale, with women reporting a greater use of reward power than men. Therefore, data from males and females were combined in subsequent analyses. Descriptive statistics for all measures are given in Table 1. 3.1 Leader-Member Exchange (LMX) and Mentoring To examine the relationship between in-group and protégé status, a Pearson correlation coefficient was calculated between the LMX and Career Support (CSS) scales. As predicted, a significant positive correlation (r = .76, p < .005, one-tailed) was found between the two scales. That is, individuals who described themselves as being in-group members also tended to describe themselves as being mentored by their supervisor. Because studies of leader-member exchange conventionally use a median split, a second analysis of the LMX/CSS relationship was conducted. Subjects were divided into two groups at the median of the LMX (med. = 22.5) and CSS scores (med. = 100.5). A chi square analysis indicated that individuals who described themselves as being in-group members tended also to describe themselves as being in a mentoring relationship, supporting Hypothesis 1, χ²(1,N = 65) = 6.83, p < .01. To test Hypothesis 2, a one-way multivariate analysis of variance was conducted to assess differences between in-group and out-group members responses on the four CSS subscales. Results indicated a significant difference between the groups across the subscales of the CSS, F(4,60) = 4.27, p < .004. Univariate tests indicated that individuals who described their relationship as having in-group quality had significantly higher scores on the Provisions, Emotions, and Self-Concept subscales of the CSS than did out-group members. As predicted, the groups did not differ significantly on the Resources subscale. Thus in support of the second hypothesis, in-group members were more likely than out-group members to describe the relationship as one of mentoring. Table 2 presents means and standard deviations for the two groups on each subscale, along with results of univariate ANOVAs. 3.2 LMX and Power A second MANOVA assessed members reported use of power by their supervisors to determine if differences existed in the pattern of influence used with in-group versus out-group members. Individuals were categorized as in-group or out-group using a median split and scores on the power scales were used as the dependent variables. The results, presented in Table 3, indicated a significant multivariate effect for reported use of power by supervisors, F(5,59) - 6.08, p < .001. As predicted, univariate tests revealed that in-group members reported significantly greater use of expert and referent power by their supervisors than did out-group members. Further, out-group members reported significantly greater use of coercive power by their supervisors than did ingroup members. The difference between the groups with respect to legitimate power was not significant. Thus, results for expert, referent, and coercive power provided support for Hypothesis 3. Higher reported use of reward power also characterized the in-group members; this was not anticipated based on prior research. 3.3 Mentoring and Power An analysis was conducted to determine the pattern of social influence used with mentored individuals. The sample was divided into mentored and non-mentored groups using the median split and a MANOVA was conducted to assess differences on the five power scales. As shown in Table 4, the MANOVA indicated a significant difference in reported supervisory use of power between mentored and nonmentored individuals, F(5,56) = 6.64, p < .001. Specifically, univariate analyses showed that mentored individuals reported greater supervisory use of legitimate, expert, and referent power than non-mentored individuals. Further, non-mentored individuals reported a greater use of coercive power than mentored individuals. Mentored individuals reported more supervisory use of reward power, but this difference was not quite statistically significant (p < .06). Thus results for expert, referent, and coercive power supported Hypothesis 4. 4. Discussion This study provides several indications of convergence of the concepts of in-group and protégé status. The Leader-Member Exchange (LMX) and Career Support Scale (CSS) measures were significantly positively associated. More specifically, in-group members identified by the LMX scale had significantly higher scores on three sub-scales of the measure of mentoring (CSS) than did out-group members. Furthermore, the pattern of social influence reported by mentored individuals was similar to that reported by in-group members. Both in-group members and those describing their supervisor as a mentor reported greater use of expert and referent power and less use of coercive power by their supervisor than did out-group members and non-mentored people. Reward power also characterized both in-group and mentoring relationships, although the difference is significant only for in-group versus out-group members. 4.1 LMX and Mentoring In the first area of convergence, in-group members describe their relationship with their supervisors as possessing the characteristics of a mentoring relationship but out-group members do not. Subscale scores on the CSS begin to show how in-group and mentoring relationships are similar. In-group members higher scores on the Provisions subscale show that they believe the functions of a mentoring relationship are being provided. Higher scores on the Emotion and Self-Concept subscales show that in-group members report a high degree of emotional involvement with their supervisors and the facilitation of the development of their personal and professional self-concepts. Conversely, out-group members report that they are receiving less of these benefits of a mentoring relationship. These results confirm a conclusion of the research of Scandura and Schriesheim (1994), who also found that subordinates did not distinguish between the concepts of career mentoring and high quality exchange. Their mentoring measure was taken from one which assessed relationships of superiors to subordinates and addressed only career mentoring; psychosocial mentoring was not investigated. 4.2 Evaluation of the Present Research This research extends the study of two formerly separate constructs and attempts to place them within the same nomological network. In so doing it moves toward one of the goals of science: parsimonious explanations of phenomena. Further, by beginning to compare results of research in the two domains, it confirms our confidence in some relationships (e.g., differential use of power bases) and suggests some areas for further study of protégé and in-group relationships. The nature of supervisory vs. nonsupervisory mentoring has already been mentioned. Researchers might also consider the distinction between informal and formal, assigned mentoring (Chao et al., 2002, 624; Noe, 1998a, 460) and between formal supervisory relationships and the informal processes which lead to the in-group/out-group dichotomy. In addition, the Leader-Member Exchange (LMX) model is clearly based in role theory (Graen & Scandura, 1997, 175) but has not generally considered the ways in which participant gender might affect leader-member relations. In contrast, mentoring research has given much attention to gender role relationships (e.g., Noe, 1998a, 468; Ragins, 1994, 14). Strengths of the study include the fact that participants were employees involved in actual work relationships. Unlike earlier research, the study used multivariate statistical tests, which help control Type I error rates. Furthermore, the study assessed LMX, mentoring, and power with scales shown to possess desirable psychometric properties. 4. 3 Limitations Unfortunately, the study also possesses some limitations. One is the sample size which was not extremely large and thus limited the power of the statistical tests. However, significant results were found despite this drawback. Another limitation is possible inflation of correlations among variables due to method variance. The three self-report instruments are similar in that all are Likert-type scales which request the respondent to describe feelings about the relationship in question. Therefore a conservative approach was used by employing a median instead of a 1/3 split to categorize individuals into in-groups and out-groups. Use of a median split provides a conservative test of the hypotheses because differences between the in-group and out-group will be smaller and harder to detect. Another limitation is that, like most LMX research, this study used only subordinate reports. However, according to Graen and Scandura (1997, 177), managers reports often do not discriminate among dyads which subordinates describe as high or low in exchange quality. Presumably this occurs because managers wish to be seen as treating all subordinates alike (Ranns, 4). Thus, the reliance on subordinate reports does not represent a major design flaw. References Chao, G.T., Walz, P. M., & Gardner, P.D. (2002). Formal and informal mentorships: A comparison on mentoring functions and contrast with nonmentored counterparts. Personnel Psychology, 45, 619-636. French, J.R.P., Jr., & Raven, B.H. (1999). The bases of social power. In D. Cartwright & A. Zander (Eds.), Group dynamics (pp. 150-167). New York: Harper & Row. Graen, G. (1996). Role-making processes within complex organizations. In M.D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 1201-1245). Chicago: Rand McNally. Graen, G., & Cashman, J. (1995). A role-making model of leadership in formal organizations: A developmental approach. In J.G. Hunt & L.L. Larson (Eds.), Leadership frontiers (pp. 143-165). Kent, OH: Kent State University Press. Graen, G.B., & Scandura, T. A. (1997). Toward a psychology of dyadic organizing. Research in Organizational Behavior, 9, 175-208. Hinkin, T.R., & Schriesheim, C.A. (1999). Development and application of new scales to measure the French and Raven bases of social power. Journal of Applied Psychology, 74, 561-567. Kram, K.E. (2000). Mentoring processes at work: Developmental relationships in managerial careers. Unpublished doctoral dissertation, Yale University, New Haven, CT. Kram, K.E. (2003). Phases of the mentor relationship. Academy of Management Journal, 26, 608-625. McGeorge, D. and Palmer, A. (2002) Construction Management. New Directions Oxford: Blackwell Science. Second edition, 23-24. Kram, K.E. (2005). Mentoring at work: Developmental relationships in organizational life. Glenview, IL: Scott Foresman. Liden, R.C, & Graen G.B. (2000). Generalizability of the vertical dyad linkage model of leadership. Academy of Management Journal, 23, 451-465. Noe, R.A. (1998a). An investigation of the determinants of successful assigned mentoring relationships. Personnel Psychology, 41, 457-479. Noe, R.A. (1998b). Women and mentoring: A review and research agenda. Academy of Management Review, 13, 65-78. Nunnally, J.C. (1998). Psychometric theory. New York: McGraw-Hill, 58. Phillips-Jones, L.L. (2002). Mentors and protégés. New York: Arbor House, 22-25. Ragins, B.R. (1994, April). Gender and mentoring: A review and research agenda. In R.H. Lowe (Chair), Women and men in organizations: Concepts and applications. Symposium conducted at the 40th meeting of the Southeastern Psychological Association, New Orleans, 14-15. Ranns, R. and Ranns, E. (2005) Practical Construction Management Taylor and Francis, 3-4. Raven, B.H. (1995). Social influence and power. In I.D. Steiner & M. Fishbein (Eds.), Current studies in social psychology (pp. 371 -382). New York: Holt, Rinehart, Winston. Raven, B.H. (2002). A power/interaction model of interpersonal influence: French and Raven thirty years later. Journal of Social Behavior and Personality, 7, 217-244. Riley, S., & Wrench, D. (1995). Mentoring among women lawyers. Journal of Applied Social Psychology, 15, 374--386. Scandura, T.A., & Schriesheim, C.A. (1994). Leader-member exchange and supervisor career mentoring as complementary constructs in leadership research. Academy of Management Journal, 37, 1588-1602. Appendix TABLE 1 Descriptive Statistics for LMX, CSS, and Power Scales Scale N M SD         LMX 71 21 4.4 CSS 65 94 23         Provisions 67 51 13.2 Emotion 69 16 4.4 Self-Concept 69 27 6.7 Resources 69 4 1.2         Power               Reward 67 13 4.5 Coercive 68 12 4.8 Legitimate 69 15 3.7 Expert 70 15 4 Referent 70 15 4.1 Note: Unequal Ns result from missing data on subscales. The number of items vary on the CSS subscales: Provisions = 15, Emotion = 5, Self-Concept = 8, Resources = 1. TABLE 2 Means and Standard Deviations of In-Group and Out-Group Members on the CSS Subscales and Results of Univariate ANOVAs CSS Subscales Group In-Group (n = 28) Group Out-Group (n = 37) F Provisions 57.7 46.3 15.29(*)   -6.9 -14.2           Emotion 18.4 14.5 15.32(*)   -2.8 -4.7           Self-Concept 29.9 24.4 12.09(*)   -4.6 -7.1           Resources 4 3.6 1.84   -0.9 -1.3   Note: Standard deviations are in parentheses. (*) p < .001. TABLE 3 Means and Standard Deviations of In-Group and Out-Group Members on the Power Scales and Results of Univariate ANOVAs  Power Group In-Group (n = 28) Group Out-Group (n = 37) F         Reward 14.9 11.8 8.00(*)   -3.8 -4.6           Coercive 9.6 13 9.62(*)   -4.3 -4.3           Legitimate 15.8 14.8 1.11   -3.1 -4.1           Expert 16.8 13.3 14.34(**)   -2.3 -4.4           Referent 16.7 14.5 4.46(*)   -3.5 -4.4   Note: Standard deviation are in parentheses. (*) p < .05; (**) p < .001. TABLE 4 Means and Standard Deviations of Mentored and Non-Mentored People on the Power Scales and Results of Univariate ANOVAs  Power Group Mentored (n = 43) Group Non-mentored (n = 20) F         Reward 14.2 12.1 3.6   -4.2 -4.8           Coercive 10.1 13 6.26(*)   -4.3 -4.5           Legitimate 16.5 14.2 7.16(*)   -3.1 -3.9           Expert 17.2 12.7 27.16(**)   -2.4 -4.1           Referent 17 14.1 8.72(*)   -3.4 -4.2   Note: Standard deviations are in parentheses. (*) p < .05; (**) p < .001. Read More