$6.81:- 0 11415%%U IGAN STATE UNIVERSITY LIBRARIE WWWHWWW 3 _1293 00571 3262 LIBRARY Michigan State University This is to certify that the thesis entitled DIFFERENTIAL LEADER-MEMBER EXCHANGE INFLUENCES ON COMMUNICATION AND ROLE STRESSORS presented by Eric George Zook has been accepted towards fulfillment of the requirements for Wegree in OMM NI ATION I \W mm Major professor Date __MAY_8_._1L9_B.9__ 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution ‘¢.¥t"‘ IV1ESI_J RETURNING MATERIALS: Place in book drop to LIBRARIES remove this checkout from ”- your record. FINES will be,charged.if book is returned after the date stamped below. AI“ '3‘?“ Q 3 ‘12-'53- DIFFERENTIAL LEADER-MEMBER EXCHANGE: INFLUENCES ON COMMUNICATION AND ROLE STRESSORS BY Eric George Zook A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Communication 1989 >< F!) 5679 ABSTRACT DIFFERENTIAL LEADER-MEMBER EXCHANGE: INFLUENCES ON COMMUNICATION AND ROLE STRESSORS BY Eric George Zook This paper examines the role of Leader-Member Exchange (Graen, 1976) on individuals' perceptions of role stress in the workplace. It is argued that differences in leader-member relationships influence the communication exchange of leaders and members (i.e., supervisors and subordinates) in terms of social support and participation in decision making. In turn, these are linked to perceptions of role ambiguity and role conflict on the part of the member. A general model testing these relationships provided a fair fit to the data, but failed to predict either role conflict or ambiguity. Subgroup analyses did reveal differences however. Specifically, employees with low job tenure perceived less role ambiguity in the face of informational support, while the role ambiguity of employees with supervisory responsibilities was reduced primarily by their ability to participate in decision making. In general, it appears that the impact of LMX on role stress is mediated by communication variables. To Sabrina, whose existence makes mine brighter iii ACKNOWLEDGMENTS Several people helped greatly in the production of this work. My greatest thanks and appreciation go to Kathy Miller, my chair. Her endless patience, good humor and dogged questioning helped to make this a much stronger work, as well as a less arduous task. Appreciation also goes to the other members of my committee: Jim Stiff and Frank Boster. Their ‘questioning provided strong insights for the extension of this work and helped reinforce in me that improvement in thought is always possible. Judy Lyles made the data collection possible, despite the burden on her own research. For that, and her comraderie which helped maintain sanity in the process, I thank her. My parents also deserve thanks for their unfailing support despite the different road I travel. Their love and encouragement underscore my accomplishments. And finally, I am grateful to my new wife Sabrina, whose dogged determination to move to another state and explore life provided further motivation to complete this work. Furthermore, her intellectual curiosity and outlook provide constant enjoyment and challenge. iv TABLE OF CONTENTS Chapter 1. THEORY AND LITERATURE REVIEW . . . . . . . Role Theory . . . . . . . . . . . . . . . Leader-Member Exchange . . . . . . . . . . LMX Effects on Dyadic Communication . . . Communication Effects on Role Stress . . . Social Support . . . . . . . . . . . Sources of Support . . . . . . . Types of Support . . . . . . . . Participation in Decision Making . . A General Model of LMX, Communication and stress 0 O O O O O O O O O O O O O O O O 0 METHODS . . . . . . . . . . . . . . . . . sample C O O O O O O O O O O O O O O O O 0 Research Procedures . . . . . . . . . . . Operationalization . . . . . . . . . . . . Leader-Member Exchange . . . . . . . Social Support . . . . . . . . . . . Participation in Decision Making . . Role Stress . . . . . . . . . . . . . Analysis . . . . . . . . . . . . . . . . . RESULTS 0 O O O O O O O O O O O O O O O O Confirmatory Factor Analysis . . . . . . . Original LMX Scale . . . . . . . . . Standard Scales . . . . . . . . . . . Page 12 16 17 18 19 23 24 28 28 28 30 30 31 31 31 36 38 38 39 4O Chapter Path Analysis . . . . . . . . . . . General Model . . . . . . . . . Models for High- and Low-Tenure Employees . . . . . . . . . . . Models for Supervisory and Non- Supervisory Employees . . . . . 4 O DISCIJSSION O O O O O O O O O O O O 0 Measurement Models . . . . . . . . . Path Models . . . . . . . . . . . . Limitations and Directions for Future ResearCh O O O C O O O O O O O O O 0 Summary . . . . . . . . . . . . . . List of References . . . . . . . . . vi Page 42 47 49 57 65 65 66 69 71 76 Table 1. LIST OF TABLES Leader-Member Exchange Scale . . . . Supervisor Social Support Scale . . Participation in Decision Making Scale Role Ambiguity and Role Conflict Scale Scale Items and Factor Loadings . . Correlations Correlations for Entire Sample . . . for Low-Tenure Subgroup . Correlations for High-Tenure Subgroup Correlations Subgroup . . Correlations for for Non-Supervisory Supervisory Subgroup Vii Page 32 33 34 35 43 46 51 52 59 60 LIST OF FIGURES Figure Page 1. General Model . . . . . . . . . . . . . . . 26 2. General Model for Entire Sample . . . . . . 48 3. General Model for Low-Tenure Subgroup . . . 53 4. Revised Model for Low-Tenure Subgroup . . . 54 5. General Model for High-Tenure Subgroup . . . 55 6. Revised Model for High-Tenure Subgroup . . . 56 7. General Model for Non-Supervisory Subgroup . . . . . . . . . . . . . . . . . . 61 8. Revised Model for Non-Supervisory Subgroup . . . . . . . . . . . . . . . . . . 62 9. General Model for Supervisory Subgroup . . . 63 10. Revised Model for Supervisory Subgroup . . . 64 viii CHAPTER ONE: THEORY AND LITERATURE REVIEW "The particular social self of a man is his image in the eyes of his own ‘set' which exalts or condemns him as he conforms or not to certain requirements." (James, 1892; sexist language in original) In setting forth his ideas concerning the "social self," William James marked the beginning of a long line of thought and research on the relationship between individuals and societies (Neiman & Hughes, 1951). In modern day parlance, however, James' social self is referred to as role. The concept of role has since become a major component of study in a variety of academic fields, most notably anthropology, sociology, psychology, management and communication. This paper provides a brief summary of the basic tenets of role theory, culminating in the more recent emphasis on dyadic linkage as espoused in Graen's (1976) theory of Vertical Dyadic Linkage (VDL). Role theory typically focuses on role stresses which arise due to the role forces produced by a person's role set. However, the position argued in this paper is that role stress is mediated by the differing amounts and quality of communication patterns associated with supervisor-subordinate dyads of differential quality. A general model is developed which delineates the impact of differential supervisor-subordinate quality on communication aspects of the relationship, and the influence of the latter on perceived role stress. 01 eo As noted above, the concept of role has been espoused as the fundamental building block of society. That is, role represents the location where the expectations of other social actors combine with the expectations and actions of individual persons and result in human behaviors (Parsons, 1965, 1951; Goode, 1959; Merton, 1957). Linton (1936) defines role in this fashion as the behavior oriented to fulfilling the patterned expectations of others. It is this conception more than any which has spawned the vast amount of research on examining the strains associated with the reckoning of all these expectations. Goode (1959) summarizes many of the early ideas concerning role strain. He examines a variety of role strains which individuals might experience as well as methods available for reducing these strains. In addition, he discusses the transactional nature of role relationships, applying an economic bargaining model as a way of understanding the role allocations of individuals. Though on a more general level, this discussion raises ideas which Graen (1976) later emphasized in developing his ideas on VDL. In the 1960s, emphasis shifted from the macro level of society to formal organizations. Due largely to the influential works of Kahn, Wolfe, Quinn, Snoek and Rosenthal (1964) and Katz and Kahn (1966), role research within organizations centered on the role strains of role conflict and role ambiguity. Role conflict is defined by Katz and Kahn (1966) as "the simultaneous occurrence of two (or more) role sendings such that compliance with one would make more difficult compliance with the other," (p. 184). Neither or these works, however, presents so clear a definition of role ambiguity. Kahn, et al., (1964) describe role ambiguity as uncertainty about what should be done to adequately fulfill one's role; specifically, they focus on uncertainty regarding how one's supervisor evaluates performance, opportunities for advancement, scope of responsibility and expectations of others regarding one's performance. The most recent review of the research on role conflict and ambiguity was conducted by Van Sell, Brief and Schuler (1981). Using the role episode model developed by Kahn, et a1. (1964), these authors reported numerous inconsistencies across studies and a great deal of variance in different employee responses to these stressors. Particularly, in reference to the relationship between role sender-focal person relationships and interpersonal factors, Van Sell, et a1. (1981) note a limited amount of research. However, studies in this area do suggest "that the structuring and supportive behavior of role senders (e.g., supervisors and co-workers), power of role senders, their functional importance to the focal person and the communication frequency between the role sender and focal person influence the focal person's perceptions of role conflict and ambiguit," (Van Sell, et al., 1981, p.55). It is suggested that greater understanding of how leadership style affects role stressors is but one area in need of further study. Several researchers (Baird, 1969; Schriesheim & Murphy, 1976) have investigated this area, but have measured "general" leadership style such as structuring and consideration behavior (Blake & Mouton, 1964). Graen and his colleagues (Graen, 1976; Dansereau, et al., 1975) refer to this as the Average Leadership Style (ALS) approach and question two of its fundamental assumptions: (1) that all subordinates under a given manager are sufficiently homogenous to be considered as a single entity, and (2) that a manager interacts in essentially the same manner with each of his or her subordinates. To counteract these flaws, Dansereau, Graen, and Haga (1975) developed their VDL approach, more recently labeled Leader-Member Exchange (LMX). This approach focuses on how role negotiation yields relationships of differential quality among supervisor-subordinate dyads. This distinction seems particularly germane to a consideration of role stressors, since role ambiguity and role conflict can be linked back to this negotiation process and may be influenced by the quality of the dyadic relationship (Graen & Johnson, 1973). The present study seeks to apply Graen's (1976) work on LMX to role ambiguity and role conflict in the workplace. The theory and research behind LMX are developed more fully below. Leader—Member Exchange The Leader-Member Exchange model grew out of initial research by Graen, Dansereau, and Minami (1972) and has been presented as an alternative to the traditional view of leadership as a single-style approach (see Campbell, Dunnette, Lawler & Weick, 1971). Graen and his colleagues (Dansereau, et al., 1975; Graen, et al., 1972) chose to apply the ideas of role negotiation as discussed by Kahn, et al. (1964) and Katz and Kahn (1966), to the leadership process. The construct of role negotiation developed by Kahn, et al. (1964) correlates with Goode's (1959) ideas of economic pricing. At a minimum, Kahn, et al., (1964) argue that role negotiation will occur due to the problematic nature of communication; sent role expectations never equal received expectations due to encoding and decoding errors. Furthermore, the focal person has his or her own ideas of appropriate role behavior and will seek to define the role to allow maximal pursuit of personal goals. Graen (1976) expands these ideas to argue that as a result of role negotiation, managers will develop relationships of differential quality with the various subordinates with which he or she works. Therefore, managers will have different styles of leadership with different subordinates rather than a general blanket approach to leading. Graen (1976) further argues that two major forces appear to cause this phenomenon: (1) the desire of subordinates to personalize their jobs, and (2) the desire of managers to engender extra assistance they cannot formally require under an employee contract. While no empirical support for these causal forces has been presented, it is apparent from research that differentiations in leader-member relationships do occur. According to a LMX perspective, all subordinates will seek to negotiate changes in their role as defined in the job description, if only by virtue of their attempt to translate the description into reality. However, managers will refrain from allowing too many subordinates to depart from the formal contract for a number of reasons (lack of personal resources to develop and maintain closer relations, inability or unwillingness to trust all subordinates to the same extent, etc.). These desires and concerns interact in the first four to six weeks of supervisor-subordinate interaction to stamp the relationship with a particular character (Graen, 1976). Longitudinal studies have been consistent in demonstrating the stability of these relationships (Graen, 1976; Dansereau, et al., 1975). Thus, Graen (1976) identifies two subordinate groups with distinct types of leader-member exhange relationships. He labels these the "in-group" and the "out-group." Graen (1976) reserves the concept of leadership to characterize in-group relationships. The relationship between a manager and an out-group member is labeled as su erviso , and is built primarily on the formal employment contract. Thus, out-group subordinates are expected to do little more than fulfill basic job requirements. A leadership relation, on the other hand, moves the interaction between a manager and subordinate to a more informal level where subordinates are allowed greater latitude in defining both the scope and activities of their job. Research on the LMX distinction has yielded a number of intriguing findings (Dienesch & Liden, 1986: Graen, 1976; Dansereau, et al., 1975). In-group members do indeed report receiving greater freedom in defining their job responsibilities, more information, more managerial support and consideration, and greater influence in the decision-making process. In-group members also express more positive attitudes toward the intrinsic outcomes of their work, their interpersonal relationships with the manager, the technical competency of the manager, and the value of job performance rewards. Finally, managers were more likely to evaluate in-group members' role behavior as corresponding to expectations of appropriate role behavior and also viewed these members as more dependable. Despite these findings, however, Dienesch and Liden (1986) identify a number of weaknesses in their 1975; Graen, et al., 1972) chose to apply the ideas of role negotiation as discussed by Kahn, et al. (1964) and Katz and Kahn (1966), to the leadership process. The construct of role negotiation developed by Kahn, et al. (1964) correlates with Goode's (1959) ideas of economic pricing. At a minimum, Kahn, et al., (1964) argue that role negotiation will occur due to the problematic nature of communication; sent role expectations never equal received expectations due to encoding and decoding errors. Furthermore, the focal person has his or her own ideas of appropriate role behavior and will seek to define the role to allow maximal pursuit of personal goals. Graen (1976) expands these ideas to argue that as a result of role negotiation, managers will develop relationships of differential quality with the various subordinates with which he or she works. Therefore, managers will have different styles of leadership with different subordinates rather than a general blanket approach to leading. Graen (1976) further argues that two major forces appear to cause this phenomenon: (1) the desire of subordinates to personalize their jobs, and (2) the desire of managers to engender extra assistance they cannot formally require under an employee contract. While no empirical support for these causal forces has been presented, it is apparent from research that differentiations in leader-member relationships do occur. According to a LMX perspective, all subordinates will seek to negotiate changes in their role as defined in the job description, if only by virtue of their attempt to translate the description into reality. However, managers will refrain from allowing too many subordinates to depart from the formal contract for a number of reasons (lack of personal resources to develop and maintain closer relations, inability or unwillingness to trust all subordinates to the same extent, etc.). These desires and concerns interact in the first four to six weeks of supervisor-subordinate interaction to stamp the relationship with a particular character (Graen, 1976). Longitudinal studies have been consistent in demonstrating the stability of these relationships (Graen, 1976; Dansereau, et al., 1975). Thus, Graen (1976) identifies two subordinate groups with distinct types of leader-member exhange relationships. He labels these the "in-group" and the "out-group." Graen (1976) reserves the concept of leadership to characterize in-group relationships. The relationship between a manager and an out-group member is labeled as su erviso , and is built primarily on the formal employment contract. Thus, out-group subordinates are expected to do little more than fulfill basic job requirements. A leadership relation, on the other hand, moves the interaction between a manager and subordinate to a more informal level where subordinates are allowed greater latitude in defining both the scope and activities of their job. Research on the LMX distinction has yielded a number of intriguing findings (Dienesch & Liden, 1986: Graen, 1976; Dansereau, et al., 1975). In-group members do indeed report receiving greater freedom in defining their job responsibilities, more information, more managerial support and consideration, and greater influence in the decision-making process. In-group members also express more positive attitudes toward the intrinsic outcomes of their work, their interpersonal relationships with the manager, the technical competency of the manager, and the value of job performance rewards. Finally, managers were more likely to evaluate in-group members' role behavior as corresponding to expectations of appropriate role behavior and also viewed these members as more dependable. Despite these findings, however, Dienesch and Liden (1986) identify a number of weaknesses in their review of the LMX literature. Four of these are relevant to the present study and are discussed below. The first problem plaguing the LMX literature noted by Dienisch and Liden (1986) is the conceptual confusion surrounding the leader-member exchange construct. Leader-member exchange has been conceptualized by at least six distinct definitions. Originally, LMX was defined in terms of the negotiation latitude granted to subordinates. However, over the years the definition has taken a variety of forms: degree of trust between leader and member, degree of perceived equity of exchange in the relationship by both leader and member, subordinate competence, degree of loyalty between leader and member, degree of mutual influence, and the amount of interpersonal affect between leader and member (Dienisch & Liden, 1986). Given this variety, the latter authors have suggested that LMX is a multidimensional construct made up of: (1) perceived contribution to the relationship, (2) loyalty (public support for goals and personal character of other), and (3) affect (mutual personal attraction among the dyadic members). A fourth dimension which appears to be tapped by typical LMX scales is recognition of personal accomplishments. An 10 original scale was developed and used in this study to assess possible multidimensionality. Conceptual confusion has contributed to a second important weakness of the extant LMX literature--the lack of validation for scales purporting to measure leader-member exchange relationships. Over 18 years of research, LMX has been measured via 2-item, 4-item, 5-item, 7-item, 10-item, and 12-item scales, none of which has been psychometrically assessed. Consideration of the face validity of these measures suggests the measurement of dimensions of leader-member relationships beyond the negotiating latitude that a member possesses in defining his or her role. A recent study by Duchon, Green, and Taber (1986) correlated sociometric assessments of LMX with one of the traditional self-report measures noted above and found general agreement between the results of both methods. However, this provides only indirect support for the validity of these scales. Psychometric evaluation of existing scales measuring the LMX construct is necessary to move LMX research beyond the current measurement obstacle. Third, few conceptual or empirical arguments exist to support a dichotomous interpretation of the leader-member exhange construct, though almost all the 11 research to date has treated it as such, or followed an "in-middle-out" trichotomy (Dienisch & Liden, 1986). This seems part and parcel of the conceptual muddle discussed above. Given the wide range of negotiating outcomes possible in role-making episodes it seems more reasonable to consider the nature of the exchange relationship as a continuous variable. Dienisch and Liden's (1986) final criticism of the LMX literature is the need for more extensive study of organizational outcome variables as they relate to leader-member exchange relationships, particularly performance. The current study attempts to expand our knowledge of these relationships through the examination of the effect of LMX distinctions on communicative behavior and role stress in the workplace. Though related work considering role rejectors and role acceptors has been done in the past (Johnson & Graen, 1973), no research has explicitly considered the LMX distinctions which arise via the negotiation phase and influence perceptions of role conflict and ambiguity. The most likely avenue by which LMX impacts these perceptions is through differences in the communication patterns across variations in dyadic quality. The next section delineates how LMX relates to several communication 12 variables, which are central to the role-making process. LMX Effects on Dyadic Comauaication Much of Graen's (1976) work on role negotiation and behavior expands on the research and theory of Kahn, et al., (1964) and Katz and Kahn (1966). It is no different when exploring the impact of LMX quality on dyadic communication. Katz and Kahn (1966) note for instance, the near certainty that more communication will occur among some dyads than others. Among the reasons posited for this occurrence are personality characteristics of individuals, role prescriptions, and proximity and other constraints attached to positional placement in the group's communication structure. It requires but a small shift in focus to argue that as a member of a work group (albeit the formal leader), a manager will develop different patterns of communication with various subordinates. The concept of LMX therefore, provides us with a way of understanding how communication might be structured and differences arise across dyads. One of the most consistently reported relationships between LMX quality and communciation is 13 in the area of supervisory social support. Graen (1976) presents two studies in which subordinates in higher quality relations reported receiving higher job latitude, more information, more support of their actions, and greater consideration of feelings and needs than did out-group members. Liden and Graen's (1980) test for LMX generalizability with first-line supervisors and subordinates reveal consistent findings; high LMX subordinates reported receiving greater attention, support, and interpersonal sensitivity from their supervisors, in comparison with those in low qualty LMX dyads. Similar findings also appear in Graen, Liden, and Hoel (1982): "Those members reporting higher quality exchanges described leaders who frequently talked to them about the details of their job performance, about their personal and work problems, and about ways to improve their effectiveness," (p.871). This contrasts with low quality exchange members who reported that leaders seldom or never talked to them about their effectiveness. A second communication variable of interest as a mediating variable between LMX and role stress is participation in decision making. Fewer studies of LMX have considered this variable, though Graen (1976) did 14 find that high LMX subordinates reported greater influence in the decision process. In the only study to specifically examine the decision making process, Scandura, Graen, and Novak (1986) found an interaction effect between subordinate LMX and productivity: the greatest involvement was reported among subordinates with high quality LMX/high productivity and the lowest among those with low quality LMX/low productivity. The results discussed in both sections above are not suprising given the close relationship between supervisor and subordinate in high quality LMX dyads. The subordinate's input into decision making would be highly valued by supervisors who see the employee's behavior as more "role appropriate" (Graen, 1976). Similarly, supervisors can be expected to provide disproportionate levels of social support to high quality relationships in the work unit, since this would likely yield a greater return on investment. It is important to distinguish here between the LMX relationship itself and the communication variables associated with that relationship. As conceptualized in this thesis, leader-member exchange is a relational quality assessment which serves to arrange relationships in a hierarchy of importance or value. Such distinctions are easily made between strangers, 15 acquaintances, friends, family and lovers. Though more subtle in organizational settings, LMX research has consistently revealed the existence of such a relational hierarchy (Dansereau, et al., 1975; Graen, 1976). Social support and participation in decision making, on the other hand, are both communicative behaviors. As such, they can be applied across relationships in differing quantities and qualities. As a relationship grows more important, we are likely to engage more frequently in communicative behaviors of social support, and in the business setting, greater amounts of participation in decision making functions. This change in communicative patterns can most likely be attributed to the increasing trust associated with members of valued relationships. This distinction between relational quality and its outcomes is important. As the quality difference reverberates through dyadic interaction, it will produce specific changes in the communication patterns which characterize the relationship. These different patterns in turn, affect both the amount of perceived role stress as well as the way individuals cope with it. 16 Communigatloa Effects on Role Stress The influence of communication on perceptions of role stress can be attributed to the uncertainty which underlies both role ambiguity and role conflict. The former stressor is primarily uncertainty about various aspects of role behavior. The latter, however, also contains a minimum implication of uncertainty as to which role expectations are most important and vital as well as whether or not one will be able to fulfill them. In this context, PDM and social support can be seen as mechanisms which aid individuals in reducing, or at least managing, their uncertainty. Sutton and Kahn (1987) discuss the need for employees to predict, understand, and control their work environment as a manner of avoiding various strains. Anything which can help organizational members develop appropriate causal relationships to guide their actions will be useful in this endeavor. Clearly, information about existing or potential stressors in the work place can increase one's perception of personal control and and may be provided through either PDM or social support. Social support also has emotional and instrumental dimensions which can address a breadth of strain situations. The 17 process whereby this is accomplished is examined separately for each variable in the following discussion. Social Supporp The term social support constitutes a broad range of meaning and has become the focus of a large body of research in its own right. The majority of this literature has focused on how social support functions to help individuals in situations of stress: parents of children with cancer (Chesler & Barbarin, 1984), patients with cancer (Wortman, 1984; Wortman & Dunkel-Shetter, 1979), persons who have lost a spouse, child or other relationally close individual (see review in Albrecht & Adelman, 1987) and persons suffering physical or mental ailments (DiMatteo & Hays, 1981). Our understanding of social support has also been applied to a variety of organizational stressors; Ray's (1987) review of the literature in this area reveals positive relationships with successful socialization, better and more useful performance appraisals, and appropriate adaptation to organizational changes. While a variety of social support definitions appear in the literature, Albrecht and Adelman (1987) build a definition focusing on its role in reducing 18 uncertainty; social support is "the verbal and nonverbal communication between recipients and providers that reduces uncertainty about the situation, the self, the other, or the relationship, and functions to enhance a perception of personal control in one's life," (p.19). Sources of Support. Within organizations, the main sources of support are direct supervisors and co-workers (Ray, 1987). Jayaratne and Chess (1984), hOwever, have identified the supervisor as the Bay support source. Furthermore, Miller, Ellis and Zook (1988) report a strong relationship between supervisor support and role stress (using a combined measure of role conflict and ambiguity) for both support staff and caregivers at a large psychiatric hospital. In general, the relationship between supervisor support and role ambiguity appears to be most consistent (Seers, Mcgee, Serey & Graen, 1983; Jayaratne & Chess, 1984) although LaRocco, House and French (1980) report a negative correlation with role conflict as well. These findings are consistent with the view that supervisors have greater access to organizational resources and control over altering conditions of role ambiguity and role conflict (Graen, 1976). If the supervisor is the source of the problem and can be made 19 sufficiently aware of this, he or she is most capable of resolving the issue. If the source comes from other organizational members with whom the subordinate interacts, the supervisor is organizationally situated with the power and authority to seek resolution of the matter. Given this, and the focus of the current study on LMX, only supervisor support is considered. Types of Support. The research on social support has identified three basic types of support: cognitive, emotional and instrumental (Jacobson, 1986). ' ive apppprt is information, knowledge and/or advice that helps an individual understand and deal effectively with his or her world. This may range from helping individuals engage in reframing situations to teaching people new skills with which they may better control their environment (Albrecht & Adelman, 1987). In keeping with the general usage in much of the support literature, this form will be labeled as informational support through the remainder of this thesis. Emotiopal support reinforces an individual's sense of self-worth through messages of admiration, respect and love. The acceptance provided by such support allows individuals to maintain some sense of personal respect and control despite the stressful situation in which he or she is currently enmeshed. 20 lastrpmaptal apppprt refers to goods and services that help an individual solve specific problems. These activities include such things as shopping for a sick friend, taking care of a neighbor's pets while they're on vacation and bringing food to families who have recently suffered the loss of a loved one. Jacobson (1886) also discusses relationships among support types and posits a linear model ranging from emotional through informational to instrumental. Little research has been done in this area however. Still, it seems likely that a distinction can be made between emotional and informational support on the one hand, and instrumental support on the other. This distinction can be understood as a degree of ease in providing the support. That is, emotional and information support are constantly and instantly at one's disposal, since they are based primarily on relational and situational knowledge of the other by a support giver; this knowledge allows appropriate response to persons in stress situations. Instrumental support, on the other hand, may likely require the martialing of resources not readily available, therefore requiring greater effort of coordination. It also is typically more demanding on one's time and energy and therefore most likely in 21 closer personal relationships. Thus, the extent to which emotional and informational support are forthcoming increases the trust and positive affect in the relationship such that one is more likely to give instrumental support which typically requires more of oneself. It is posited here then that emotional and informational support are precursors to instrumental support. The next step is to link the impact of support type with role ambiguity and role conflict. In spite of consistent findings of negative relationships between social support and stress in a variety of areas, no research to date has investigated their varying impact on role conflict and role ambiguity. Therefore, the extent to which formal hypotheses may be stated is limited, though some preliminary statements appear warranted. Given the nature of the role stressors under consideration, informational support should have a strong impact on role ambiguity. Since the problem is an unclear definition of one's role, additional information can clarify the role, thus decreasing one's level of stress. Informational support may also help individuals in role conflict by establishing priorities among conflicting roles and requirements, but this possibility is does not warrant a full hypothesis. 22 Emotional support probably performs a buffering role for these stressors. That is, it can provide employees with a sense of acceptance and self-worth while they "learn the ropes" and gain greater control over the organizational environment. This is no doubt valuable, but without appropriate informational support, nothing is done to eradicate the source of the problem; the stressors still lurk in the environment. Thus, while emotional support should increase with LMX quality, it is unlikely that it will directly impact either role conflict or role ambiguity, other than through an indirect path through instrumental support. The role of instrumental support will function primarily in situations of role conflict. Such situations can be dealt with by temporarily assigning another coworker to help an individual with conflicting roles, or creating a new position to unburden an employee with too many role demands. It seems unlikely that role ambiguity is either directly or indirectly affected by instrumental support. Indeed such "support" might be perceived as threatening if it implied that the individual had no idea of the job and was incapable of performing in a satisfactory manner. 23 Thus, social support from one's supervisor should prove an invaluable form of communication for influencing employee perceptions of role stress. This underscores the importance then of supervisory support being a function of LMX differences across dyads in a work group. Participation in Qeciaiop Makipg The second communicative variable which is likely to affect role stress is participation in decision making. A meta-analysis by Miller and Monge (1986) provides evidence that both worker satisfaction and productivity can be positively influenced by employee involvement in the decision-making process. Their results suggest that participation works to some extent through a cognitive process in which workers become more productive and satisfied due to increased information and personal control (i.e., less uncertainty). Jackson (1983) provides further support for a link between PDM and the role stressors incorporated in this study. She manipulated participation at a hospital outpatient facility by increasing the number of meetings held with work groups; unit supervisors received training on how to run effective meetings and given a list of important potential topics to address 24 in these meetings. The results showed strong negative relationships between PDM and both role conflict and ambiguity after six months. These stressors in turn strongly influenced employees' emotional stress and via that, overall job satisfaction, absenteeism and turnover intention. The importance of limiting role conflict and ambiguity is thus emphasized, as is PDM as a method for accomplishing this. Jackson's study (1983) is of further note in that its findings are consistent with the distinction made in the present study between participation in decision making and social support. Though there is a potential overlap between informational support and participation in decision making, these constructs can be distinguished in that participation is focused on the ability of subordinates to wield influence in the decision making process while informational support involves the acquisition of sufficient information to perform one's task. Thus, it is the difference between use of information and the acgpisition of information. A General Model of LMXl Communication and Stress Given the research and theory reviewed above, a general model can be developed which incorporates the variables of interest. Specifically, the leader-member 25 exchange relationship serves as the exogenous variable. This is conceptualized as the quality of relationship between each leader-member dyad. LMX quality then affects the extent to which subordinates receive both social support from their supervisor and opportunities to participate in the decision making process. Preliminary support for these linkages is found in Graen (1976). As described above, participation in decision making is hypothesized to have a negative relation with perceptions of role ambiguity and role conflict. Schuler and Jackson (1986) point out that "the more participation allowed employees, the more likely they will be able to get a clearer understanding of what is expected and what is rewarded (Schuler, 1980). Consequently, the less uncertainty and stress there is for the individual" (p. 215). This is in keeping with Miller and Monge's (1986) cognitive model of participation. Given larger amounts of each social support type, it is expected that members of high LMX dyads will report lower amounts of both role stressors as well. Specifically, it is hypothesized that informational support will be negatively related to role ambiguity and that instrumental support will be negatively related to role conflict. Emotional support serves 26 HUHAhZOU NAOM UZHMdZ ‘ ZOHmHUNQ 2H MHHDUHmZ< MAOM Hmommbm . \\ ZOHH§KOMZH 1 lil NUSUNM mmmzmzlmmndma 4 Ill Hmommbm Aom .o ouowfim 57 Mode 3 for Su e iso and Nonsu rviso Em lo e s In testing the general model at the supervisory and nonsupervisory levels, correlations among the variables for each group were once again computed. The corrected and uncorrected correlations, along with means and standard deviations for each group are presented in Tables 9 and 10. The hypothesized general model tested with the nonsupervisory group is presented in Figure 7. As with both cases above, the model provided a good fit to the data (average squared error = .03) but contained nonsignificant path coefficients. The final model for non-supervisory employees is presented in Figure 8. Again, both forms of role stress fall out of this model, leaving only the relations between LMX and the communication variables. As with the general model tested on full data, and the long-term employee model, the path coefficients between these elements are highly significant. Overall, the model fits well as all reproduced correlations are within sampling error of the actual correlations and average squared error is .01. The hypothesized general model tested with supervisory employees is presented in Figure 9, with path coefficients and standard errors. The model 58 presents a good fit to the data (average squared error = .02) but contains a number of nonsignificant paths. These were dropped from the model which was reanalyzed. The final path model for the supervisory group is presented in Figure 10. Role ambiguity is predicted by participation in decision making in this model. The model fits the data well. None of the reproduced correlations deviate from the actual correlations by more than sampling error, and the average squared error is .008. 59 Table 9: Correlations for Non-Supervisory Subgroupé ITEMS 1 2 3 4 5 6 7 I """ 33'"'II3""III""II2'"'III""IBI"'IIII 2 .81** -- .87 .81 .57 -.17 -.31 3 .71** .78** -- .91 .71 -.20 -.23 4 .70** .74** .84** -- .70 -.13 -.20 5 .63** .46** .58** .58** -- -.21 -.14 6 -.00 -.14 -.17 -.11 -.16 -- .78 7 -.11 -.27 -.20 -.18 -.14 .62** -- Means 17.87 24.69 25.98 21.75 10.21 13.73 21.18 S.D 4.76 4.46 8.14 9.10 2.31 3.56 5.98 Leader-Member Exchange Supervisor Emotional Support Supervisor Informational Support Supervisor Instrumental Support Participation in Decision Making Role Ambiguity Role Conflict \lO‘U‘IbUNH II II II II II II II 6 Correlations in the lower half of the matrix are uncorrected for attenuation; corrected correlations appear in the upper half. 60 Table 10: Correlations for Supervisory Subgroupé ITEMS 1 2 3 4 5 6 7 I """ I:'""III""III""TI2""TEI'"ITSI"TIE" 2 .7l** -- .63 .44 .29 -.12 -.02 3 .71** .57** -- .83 .34 -.20 .22 4 .52** .40* .77** -- .45 -.19 .15 5 .49** .23 .28 .37* -- -.46 -.04 6 -.07 -.10 -.17 -.16 -.34 -- .48 7 .14 -.02 .19 .14 -.03 .38 ~- Means--I7:75--24:58-_25:36--25:06—-II:64---I5:50--24:22 S.D 3.43 3.46 6.69 6.14 1.62 3.40 4.85 * p < .01 Leader-Member Exchange Supervisor Emotional Support Supervisor Informational Support Supervisor Instrumental Support Participation in Decision Making Role Ambiguity Role Conflict \imUlwal-I' II II II II II II II 9 Correlations in the lower half of the matrix are uncorrected for attenuation; corrected correlations appear in the upper half. 61 no. e um unmofiwwcmfim « oszdz ZOHmHUma zH - onsuumnmiaoz Mom Hmpoz Hmumsou .n ouswflm 62 mo. e um uamUAMHawflm « UZHMuondmlcoz mom Hmpoz pomfi>om .m ounwfim 63 UZHMdz NAOM mH I no. e um unmowwflcwwn « ‘0 ZOHmHumn 2H . . onHuoanm How Hero: Hmuosmo .m munMMm 64 EHBHafil whom; no. e um mandamanwfim « ll. oszhomnm now Home: vomfi>om .oH mummfim Chapter IV: DISCUSSION This chapter discusses the implications of the analyses presented in Chapter III. First, the results of confirmatory factor analyses on the original LMX scale and the standard scales are discussed. Second, the implications of the final path models for each subgroup are presented. Finally, limitations of this study and directions for future research are discussed. Measurement Models Several interesting results emerge from the confirmatory factor analyses of the scales used in this research. First, Graen, Novak and Sommerkamp's (1982) measure of the leader-member exchange concept was confirmed as a unidimensional factor rather than as a multidimensional construct. This finding has several implications. First, the confirmation of this model gives us increased confidence about the validity of past research on leader-member exchange and provides us with the possibility of reasonably comparing the current research to these previous efforts. Second, these data tend to refute conceptual arguments for a multidimensional LMX construct. Our analyses suggest that LMX should simply be viewed as the overall quality 65 66 of relationship in supervisor-subordinate dyads. Perhaps it is too difficult for most people to distinguish specific dimensions of their relationship with another. Coupled with halo effects, it may not be too surprising that multiple dimensions are not found. However, given the heuristic value of a multidimensional approach to LMX as well as the possibility of greater precision, further research along these lines is encouraged, with particular focus on scale construction. Another finding of particular value concerning the measurement model is the empirical distinction between LMX and the various forms of social support. Thus, conceptualizing LMX as a general level of relational quality does not make it so broad as to incorporate communication elements. This distinction speaks to the potential value of a heavier communication focus for examining LMX influences in the work setting. Path Models The primary theoretical finding of this research is that leader-member exchange does have very strong effects on the communication variables of social support and participation in decision making. This was true for the sample as a whole as well as across all the subgroups. Thus, subordinates who have high 67 quality LMX report receiving greater amounts of social support and PDM. These findings are consistent with research results presented in Chapter I (Scandura, Graen, & Novak, 1986; Graen, Liden & Hoel, 1982: Graen, 1976) Examination of the impact of these communication variables on role stress revealed an impact only on role ambiguity for the supervisory and low tenure subgroups. Consistent with the uncertainty reduction framework within which this research was situated, informational communication had the greatest impact on reducing role ambiguity. However, this information was acquired in different ways depending on the subgroup under investigation; low-tenure employees relied on supervisory informational support while supervirsors relied on participation in decision making. The former finding corresponds somewhat with Fisher's (1985) findings that emotional and informational support from supervisors helped facilitate positive adjustment outcomes (e.g., satisfaction, performance, commitment) for newcomers. For the supervisory group, the finding that PDM impacted on role ambiguity is consistent with past research which emphasizes the importance of structure controls for the reduction of job-related strain (Jackson, 1983; Miller, et al., 1988). 68 The use of different information sources by these two groups may be understood in terms of the level of activity with which the focal person acquires information. As discussed in Chapter I, PDM can be construed as a process whereby employees may legitimately seek information necessary for role clarification, and potentially, be aggressive in this pursuit. In contrast to this active model, information support is typically passive. In the case of new employees, an inability to define appropriate information needs is part of the general ambiguity problem they face. Additionally, while supervisors may encourage new employees to ask plenty of questions, concerns for impression management and an unwillingness to approach the supervisor may limit the employees' active search for information. The fact that no significant relationships were found between the communication variables and role conflict suggests that reduction of this stressor may rely more on the individual's adaptation capacity. As used in this research, both instrumental support and PDM were argued to be important as a means of prioritizing conflicting roles. Given unsupportive results, two explanations are possible. First, this hypothesized prioritizing may not actually take place. 69 Research focusing more directly on this is required for any definitive response. Secondly, even if such prioritizing takes place, it may not reduce the felt role conflict of the focal person. Further, prioritizing as a method of stress reduction assumes that all role senders could agree on a single hierarchical scheme, a highly unlikely assumption. In general, it is not surprising to find that the greatest impact on role ambiguity appears to be tenure on the job. Role ambiguity differences between the tenure groups, while not fully significant, did result in differences in the final models. As already noted, no significant relations between any of the communication variables and role ambiguity are found for high-tenure employees. However, the relatively long job stay (two years) of the "low" tenure group in this study reveals that role ambiguity is important long into one's transition to a new job. Thus, the important relationships between LMX and the communication variables which limit role ambiguity could help to alleviate negative outcomes such as absenteeism and turnover. Limitations and Directions for Suture Research This study had several limitations which provide suggestions for improving or extending this research. 70 This section will discuss these limitations and propose useful avenues for further exploration of the relationship between supervisor-subordinate relationships and role stress. First, the self-report nature of the data, though important for considering perceptions of relationships, roles, and communication in the workplace, sets some limitations on the assessment of communicative behavior. Future research could usefully consider the mechanisms through which actual interaction or participation in communication networks influences the role definition process in organizations. A second drawback with the use of self-report data concerns the reliance on only one portion of the supervisor-subordinate dyad for purposes of defining the relational quality. This presents an obvious limitation in the form of missing data from each employee's supervisor. Without descriptions of the relationship by these supervisors, the possibility of bias on the part of subordinates is more likely. A third limitation is the relatively low levels of role ambiguity and, especially, role conflict reported by respondents in this study. It is possible that occupations characterized by higher levels of situational and interpersonal stress would reveal 71 different relationships among these constructs. Future research should replicate these findings in a wide range of organizational and occupational settings. A final suggestion for future study concerns the findings of previous LMX research. These past studies have revealed a number of organizational outcome variables associated with LMX quality: turnover (Graen, Orris, and Johnson, 1973), job satisfaction (Graen, 1976; Dansereau, et al., 1975), and potentially performance (see review in Dienesch & Liden, 1986). It may very well be that these findings are mediated by the communication differences recorded in the present study. For instance, differences in information about job requirements and methods might result in differences in productivity. Further, communication can be seen as an important variable in influencing a person's perceptions about his or her job, thereby influencing job satisfaction. Reexamination of these findings from a communication perspective could do much to improve our understanding of the manner in which LMX affects such organizational outcomes. Summary This research investigated the mediating nature of communication as a way of understanding the impact of LMX on role conflict and ambiguity. The results of 72 this study present a number of both theoretical and practical applications. Theoretically, this study greatly extends the current state of knowledge concerning the conceptualization and operationalization of the LMX construct. The suggestion that LMX might be multidimensional was expanded and explored here without positive results. Thus, a unidimensional approach to LMX receives support. This is given further support by the validation of an original LMX scale. The scale assessed was prototypical of those used in past LMX research and its validation allows greater faith in these research findings. This research also succeeds to some extent in showing LMX effects as being mediated by communication variables. The connection between LMX and communication is strong and unequivocal. While this study was able to show further effects only for role ambiguity, these relations were quite strong. This provides support for a communication explanation of LMX results in the workplace. For the practitioner, these results suggest the importance of managerial awareness of the relational quality associated with their various subordinates. While it is unlikely that a manager can produce high 73 quality relationships with all subordinates, it is important to avoid blatant favoritism that might stir up resentment in the work unit. Though past research reveals that low LMX members can identify coworkers with high LMX relationships with the supervisor, (Graen, Liden & Hoel, 1982) it has not been shown whether this causes great resentment. However, this may be due to the fact that such low LMX members leave the organization (Graen, Orris, & Johnson, 1973). Further research on the extent to which variability in LMX relationships affects the stability of overall work group relations would be useful in this regard. In relation to role ambiguity, managers should be particularly attuned to the information needs of low-tenure employees. Managers may need to take a much more active role in feeding information to new employees for purposes of orienting them to both their new role and the organizational environment in which they will operate. This means doing more than sending the employee to a general orientation meeting in which formal organizational policies are presented. The employee also needs information about the informal organizational rules, specifically within his or her workgroup. As time goes on, employees may gain enough understanding to know what information is needed and 74 where to get it; until then, information needs to be routed their way in a concious manner. The finding that employees in supervisory positions reported greater levels of both role conflict and role ambiguity is also important information for managers. This occurrence may be linked to the greater responsibilities for others built into supervisory positions. Whatever the cause, the main lesson here is the need for such employees to have an active role in defining role expectations. This is afforded through participation in the decision making process which allows them seek new and updated information about their responsibilities as the work environment changes, thereby alleviating role ambiguity. For example, if a new efficiency study is being considered, its exact requirements in terms of resources will typically be made explicit. If supervisors are involved in the planning process, they are afforded the opportunity to gain information about how this study will affect their role expectations. The problem here is why role conflict is not also aided by such participation. Presumably, if supervisors are able to seek information about role requirements and changes, they should be able to perceive conflicting demands and raise concerns for 75 discussion. As noted above, role conflict may not be amenable to reduction by communication. Katz and Kahn (1966) have noted that role conflict is an inherent quality of life: the goal is not so much to alleviate role conflict as to keep it down to manageable levels. In short, the fact that supervisors must coordinate a variety of human actors, each having their personal role expectations, places them in a position where conflicts will accrue, whether between two subordinates, or between upper management and his or her work group. Therefore, managers should be attuned to the level of role conflict being experienced by supervisors, and intervene when role conflict begins to overwhelm. Greater investigation needs to focus on how much role conflict is too much, however, for this advice to be useful. LIST OF REFERENCES LIST OF REFERENCES Albrecht, T.L. & Adelman, M.B. (1987). Qommgpicatiag apgial_§pppgrr. Beverly Hills: Sage. Adelman, M.B., Parks, M.R. & Albrecht, T.L. (1987). Supporting friends in need. In T.L. Albrecht and M.B. Adelman (Eds.), Qommpnigaring social support (pp. 105-125). Beverly Hills: Sage. Baird, L.L. (1969). A study of the role relations of graduate students. Spgrpal of Edugarional Pershglggr. £2: 15-21- Blake, R.R. & Mouton, J.S. (1964). The managerial grig. Houston: Gulf Publishing Co. Campbell, J.P., Dunnette, M.B., Lawler, E.E., & Weick, K.E. (1970). Managerial behavior, performance, and erfecrivaness. New York: McGraw Hill. Chesler, M. A. & Barbarin, O. A. (1984). Difficulties of providing help in a crisis: Relationships between parents of children with cancer and their friends. gourpal or Sggial Isapaa, AS (4), 113-134. Dansereau, F., Jr., Graen, G. & Haga, W. J. (1975). A vertical dyad linkage approach to leadership within formal organizations--a longitudinal investigation of the role making process. Qrganizational Behavior and fluman Performance, 1;. 46-78. Dienesch, R.M. & Liden R.C. (1986). Leader-member exchange model of leadership: A critique and further development. Agademy of Management R v'ew, ll, 618-634. Duchon, Green & Taber (1986). Vertical dyad linkage: A longitudinalasessment of antecedents, measures, and consequences. Jo nal of A lied Ps cholo , 11 (1) , 56-60. Fiske, S.T. & Taylor, S.E. (1984). Social cognition. Reading, MA: Addison-Wesley. Fisher, C. D. (1985). Social support and adjustment to work: A longitudinal study. Journal of Mana me t, ll(3), 39-53. 76 77 Ganster, D.C., Fusiler, M.R., & Mayes, B.T. (1986). Role of social support in the experience of stress at work. ourna of A 1 ed Ps cholo , ll, 102-110. Goode (1959). A theory of role strain. American Sociological Reviey, SS, 483-496. Graen, G. (1976). Role-making processes within complex organizations. In M.D. Dunnette (Ed.), Handbogk of industrial ang organizational psychology (pp. 1201-1245). Chicago: Rand McNally. Graen, G., Dansereau, F., & Minami, T. (1972). Dysfunctional leadership styles. Qrganizational Sahavior and Human Perfornance, 1, 216-236. Graen, G., Liden, R., & Hoel, W. (1982). Role of leadership in the employee withdrawal process. Journal of Applied Psychology, a1, 868-872. Graen, G., Novak, M. & Sommerkamp, P. (1982). The effects of leader-member exchange and job design on productivity and satisfaction: Testing a dual attachment model. Organizational Behavior and numan Performance, SQ, 109-131. House, J.S. & Cottington, E.M. (1986). Health and the workplace. In L.H. Aiken & D. Mechanic (Eds.), Applicatipn pf spcial scienca to clinical medicine ang nealtn pglicy (pp.392-415). New Brunswick: Rutgers University Press. Hunter, J.E. (1980). Factor analysis. In P.R. Monge and J.N. Cappella (Eds.), Multivariate technignes in nunan gpnnunigarion resaarch (pp. 229-258). New York: Academic Press. Hunter, J.E. & Gerbing, D.W. (1982). Unidimensional measurement, second order factor analysis, and causal models. Researcn in Qrganizational Senavior, A, 267-320. Hunter, J.E. & Lim, T.S. (1987). LIMSTAT. Unpublished manuscript, Department of Communication, Michigan State University, E. Lansing, MI. 78 Jackson, S.E. (1983). Participation in decision making as a strategy for reducing job-related strain. Journal of Appliad Psychology, SS (1), 3-19. Jacobson, D.E. (1986). Types and timing of support. gournal pf Healrn ana Sgcial Sehavior, 21, 250-264. James, W. (1892). angnplggy. New York: Henry Holt. Jayaratne, S. & Chess, W.A. (1984). The effects of emotional support on perceived job stress and strain. The Journal of Applied Behavioral Science, an (2), 141-153. Johnson, T. W. & Graen, G. (1973) Organizational assimilation and role rejection. Organizational Senavipr and nnman Perfornanga, lg, 72-87. Kahn, R.L., Wolfe, D.M., Quinn, R.P., Snoek, J.D. & Rosenthal, R.A. (1964). Qrganizational stress. New York: Wiley. Katz, D. & Kahn, R.L. (1966). The social psychology of organizations. New York: Wiley. LaRocco, J.M., House, J.S., & French, J.R.P., Jr., (1980). Social support, occupational stress and health. Journal of Health ang Social Behavior, 21, 202-218. Liden, R. C. & Graen, G. (1980). Generalizability of the vertical dyadic linkage model of leadership. Academy of Managemenr Journal, 3. 451-465. Linton, R. (1936). The study of man. New York: Appleton-Century-Crofts, Inc. Merton, R. K. (1957). The role-set: Problems in sociological theory. Ina British Journal or Sociology, a, 106-120. Miller, K.I. & Monge, P.R. (1985). Social information and employee anxiety about organizational change. Human Communication Rasaarcn, ll, 365-386. 79 Miller, K.I., Ellis, B.H., Zook, E.G. & Lyles, J. (1988). An integrated model of stress, burnout and communication in the workplace. Michigan State University, East Lansing, MI. (In review). Neiman, L. J., & Hughes, J. W. (1951). The problems of the concept of role--a re-survey of the literature. Spcial Eorces, SQ, 141-149. Norusis, M. J. (1986). SPSS-PC+. Chicago: SPSS, Inc. Parsons, T. (1965). Structure and process in modern 1 societies. New York: The Free Press. Parsons, T. (1951). The social system. Glencoe, Ill: The Free Press. Ray, E.B. (1987). Supportive relationships and occupational stress in the workplace. In T.L. Albrecht and M.B. Adelman (Eds.), Communicating apgial_§nppgrr (pp. 172-191). Beverly Hills: Sage. Rizzo, J.R., House, R.J., & Lirtzman, S.T. (1970). Role conflict and ambiguity in complex organizations. Administrative Science Quarterly, lS, 150-163. Scandura, T. A., Graen, G. B., & Novak, M. A. (1986). When managers decide not to decide autocratically: An investigation of Leader-Member Exchange and decision influence. Journal of Applied Psychology, 1l(4), 579-584. Schriesheim, C.A. & Murphy, C.J. (1976). Relationships between leader behavior and subordinate satisfaction and performance: A test of some situational moderators. Journal of Applied Esyghglogy, fil, 634-641. Schuler, R.S. (1980). Definition and conceptualization of stress in organizations. Organizational Behavior and Human Eerformance, 2S, 184-215. Schuler, R.S., Aldag, R.J., & Brief, A.P. (1977). Role conflict and ambiguity: A scale analysis. Organizational Benavior and Human Performance, an, 119-128. 80 Schuler, R.S. & Jackson, S.E. (1986). Managing stress An uncertainty through PHRM practices: interpretation. Acadeny pf Management Journal, 22, 331-340. Seers, A., McGee, G.W., Serey, T.T., & Graen, G. The interaction of job stress and social (1983). support: A strong inference investigation. 273-284. Acagemy of Management gournal, 2Q (2), R.I. & Kahn, R.L. (1987). Prediction, Sutton, understanding, and control as antidotes to In J.W. Lorsch (Ed.), organizational stress. Sandbook of organizational pehavior (pp.272-285). Englewood Cliffs, NJ: Prentice Hall. Van Sell, M., Brief, A.P., & Schuler, R.S. (1981). Integration of Role conflict and role ambiguity: the literature and directions for future research. Human Relations, SA (1), 43-71. Vroom, V. (1960). Some personality determinants of the affects of participation. Englewood Cliffs, NJ: Prentice Hall. Wortman, C. (1984). Social support and the cancer patient: Conceptual and methodological issues. Cancer, SS, 2339-2360. Wortman, C. & Dunkel-Schetter, C. (1979). Interpersonal relationships and cancer: A theoretical analysis. Journal of Social Issues, aa(1), 120-125.