.. . , . : 2..., . . . ;.In... ‘IIfl'g| I _ m u ;. :r.....f....,. E z: 3.: '5... 11:. . .1: . ‘.... .....V. «Luv ,nui,.\.u..?.....u..m.n .. A .:, ., ... _: :3 2: .. THESES 2_ 7’6 CC " y This is to certify that the dissertation entitled Why Do Subordinates Lie To Their Superiors? A Model Of Organizational Deception presented by Anne Patricia Hubbell has been accepted towards fulfillment of the requirements for PH.D . Communication degree in (/W Walk Major professor Date WI/XZMDD MS U i: an Affirmative Action/Equal Opportunity Institution 0-12771 LIBRARY Michigan State University PLACE IN RETURN BOX to remove this checkout from your record. To AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE mm 11 203202 Now 5; 2003 2 11m mus-p.14 WHY DO SUBORDINATES LIE TO THEIR SUPERIORS? A MODEL OF ORGANIZATIONAL DECEPTION By Anne Patricia Hubbell A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Communication 2000 ABSTRACT WHY DO SUBORDINATES LIE TO THEIR SUPERIORS? A MODEL OF ORGANIZATIONAL DECEPTION By Anne Patricia Hubbell Literature in organizational deception has long focused on why individuals in organizations distort upward communication. Yet, few have tried to model antecedents of upward distortion. The purpose of this research is to create and test such a model. The model will test organizational, dyadic, and individual-level variables and their influence on a superior’s trustworthy behaviors and the quality of the relationship between a subordinate and a superior. It is proposed that subordinates who believe their superior is trustworthy and report having a high quality relationship with the superior, will be more likely to communicate truthful information to the superior. If, however, the superior has not demonstrated trustworthiness and the relationship is of low quality, various other messages will be communicated, none completely honest. This reciprocity is evident in exchanges between superiors and subordinates (F ulk & Mani, 1986) and is well explained within the framework of Social Exchange Theory (Blau, 1964) and Agency Theory (see Eisenhardt, 1989). Thus, Fulk and Mani’s (1986) model of upward information distortion, Social Exchange Theory, and Agency Theory will be used to explain the links between the antecedent variables and both trustworthy behaviors of a superior and the quality of the relationship with a superior and how these two variables influence the veracity or lack of truthfulness in a message. A model will be developed from this discussion. Path analysis and multiple regression will be used to test the model. C0pyright by Anne Patricia Hubbell 2000 DEDICATION I dedicate this work to my husband, Max Leuenberger and to Sorbach, my beloved tabby. They have been my greatest admirers and I love them dearly. iv ACKNOWLEDGMENTS My deepest thanks goes to the individuals who helped me gain access to organizations. Particular thanks goes to Shaila Gaffney, Cindy Kamerad, Jenn Heisler, and Bill Donohue, Ph.D. Without their help I would not have had such excellent sources of data. Also, sincere appreciation goes to the individuals within the companies who set up collection times and ensured that participants were there. Without a knowledgeable and supportive committee and dissertation director, this work would not have possible. My director, Caryn Medved, Ph.D., always found time to counsel me in writing, data collection, and data analysis. She untiringly went through revision after revision and strove to teach me how to be a better writer — or at least a much more consise one! Without knowing it she has taught me much more. She has taught me how to value the process, and to learn from every setback. She has also demonstrated the values and actions of an excellent mentor. When I work with students I will think of her as a model for my behaviors. My excellent committee also proved to be invaluable to the completion of this dissertation. John Wagner, Ph.D. was a constant source of information for the literature review and data analysis. Actually, the project began in a class I took from him so special appreciation for his involvement in the process is needed. Bill Donohue, Ph.D. is always there to help me find information I need and to make me laugh. Vernon Miller, Ph.D. has also been there from the beginning and special thanks go to him for helping me think through the typology of messages this is based upon. Thank you for your time, support, and knowledge. Also, my friends who I have neglected but who are always there for a laugh and sometimes to dry tears: Betty Lafrance, Liz Jarrard, Marge Barkman, Rebecca Chory, and Dana Mastro. And my family who taught me to value education and to work hard at anything I put my mind to. Most especially my father who always believed in me, even when I did not do well in school because I talked too much. He has been the one person who has loved and believed in me my whole life and has made me the person I am today. Finally, and yet extremely important, I would like to thank my husband, Max Leuenberger, who was an editor, supporter, and chocolate covered raisin provider. Never have I met a more patient and loving individual. He has been my greatest gift in this world. A couple days ago I was re-reading a letter that my Babusia (great-grandmother) wrote to my mother. It talked about living life with integrity and honesty. One of the greatest lessons that I have learned in graduate school is the importance of these virtues. I study deception because of the lack of integrity and honesty I have experienced in the world. In the process of trying to understand these transgressions, I hope to have enlightened others and to continue to do so. vi TABLE OF CONTENTS LIST OF TABLES ..................................................................................................... ix LIST OF FIGURES ................................................................................................... x CHAPTER 1 INTRODUCTION ..................................................................................................... 1 CHAPTER 2 LITERATURE REVIEW, MODEL CREATION, AND HYPOTHESES ................ 5 Overview ........................................................................................................ 5 Fulk and Mani’s Model of Information Distortion ........................................ 5 Agency Theory and Social Exchange Theory ............................................... 6 Behaviors of Superiors leading to Increased Trust by Subordinates ............. 8 Quality of the Relationship with a Superior - LMX ...................................... 12 Organizational Structure ................................................................................ 14 Organizational Trust ...................................................................................... 16 Impression Management Behaviors ............................................................... 18 Ascendancy Desires ....................................................................................... 20 Autonomy ...................................................................................................... 21 Interdependence ............................................................................................. 22 Creating a Typology of Deceptive Messages ................................................ 23 Types of Deceptive Messages ........................................................................ 24 Model of Organizational Deception and Truth .............................................. 29 Hypotheses Begin .......................................................................................... 30 CHAPTER 3 METHODS ................................................................................................................ 35 Pretest ............................................................................................................. 35 Experiment ..................................................................................................... 42 CHAPTER 4 RESULTS .................................................................................................................. 47 Overview ........................................................................................................ 47 Experiment ..................................................................................................... 47 Organizations One and Two - Credit Unions ................................................ 47 Organization Three — Automobile Supplier ................................................... 48 Organization Four — Public Relations Company ........................................... 48 Analysis of Scales .......................................................................................... 49 Analysis of the Likelihood of Use Instrument ............................................... 54 Comparison of Major Variables among the Four Companies ....................... 54 Testing the Hypotheses .................................................................................. 55 Post Hoc Analyses ......................................................................................... 67 Analysis of Potential Confounds ................................................................... 68 vii Analysis of the Manipulation Check .......................................................................... 70 CHAPTER 5 DISCUSSION AND IMPLICATION OF FINDINGS .............................................. 72 Contributions of the Current Research .......................................................... 72 Moderator Variables ...................................................................................... 76 Post Hoc Analysis of Acendancy Desire, or Ambition .................................. 77 Theoretical Explanations ............................................................................... 78 Limitations of the Current Research .............................................................. 80 Future Directions ........................................................................................... 81 APPENDIX A INSTRUMENTS ........................................................................................................ 85 Scales for Trust in a Superior ......................................................................... 85 Quality of the Relationship with 3 Superior Scale ......................................... 86 Antecedents to Trust in a Superior and Relational Quality Scales ................ 87 Moderator Variable Scales ............................................................................. 88 APPENDIX B LIKELIHOOD OF USE SCALE ............................................................................... 89 APPENDIX C CORRELAT IONS OF ALL RELEVANT VARIABLES ......................................... 91 APPENDIX D HYPOTHETICAL SITUATION ............................................................................... 92 APPENDIX E HIGH AND LOW LEVELS OF AUTONOMY AND INT ERDEPENDENCE AND THEIR INFLUENCE ON TRUST AND MESSAGE SELECTION .............. 93 APPENDIX F HIGH AND LOW LEVELS OF AUTONOMY AND INTERDEPENDENCE AND THEIR INFLUENCE ON RELATIONAL QUALITY AND MESSAGE SELECTION .............................................................................................................. 94 REFERENCES .......................................................................................................... 96 viii LIST OF TABLES Table 1: Reliability Estimates and Numbers of Items .............................................. 42 Table 2: Scale Reliabilities and Number of Items .................................................... 50 Table 3: Six Dimensions ofTrust in a Superior ................... 51 Table 4: Factor Loadings for Trust Scale ................................................................. 52 Table 5: Means and Standard Deviations for All Variables ..................................... 56 Table 6: Tests of the Trust Model in Path ................................................................. 58 Table 7: Total Variance by Models When Looking at Moderator Variables ........... 58 Table 8: Regression Analysis of Trust on Message Selection .................................. 61 Table 9: Tests of the Quality of the Relationship Model in Path .............................. 64 Table 10: Total Variance Explained by Models When Looking at Moderator Variables ................................................................................................... 64 Table 11: Regression Analysis of the Quality of the Relationship on Message Selection .................................................................................................... 66 Table 12: Regression Analysis of Ambition’s Relationship with Message Selection .................................................................................................... 68 ix Figure 1: Figure 2: Figure 3: Figure 4: Figure 5: Figure 6: LIST OF FIGURES Model of Organizational Deception and Truth ......................................... 30 Expected Correlations ............................................................................... 53 Multi-Trait Matrix ..................................................................................... 53 Path Coefficients for Model with Trust as a Central Variable .................. 57 Path Coefficients for Model with Relational Quality as Central Variable ...................................................................................................... 63 Ambition as Antecedent to Message Selection ......................................... 67 Chapter 1 INTRODUCTION “Strict truth telling and truth seeking is at odds with a life lived among other humans” (Kingwell, 1993, p. 389) All languages have some similarities. Argyle (1969), asserts that all languages share a “common structure” (p. 65). In every language, words become sentences and grammatical rules are followed (Argyle, 1969). Also in every language, there is an opportunity to change the words within the sentences to misrepresent the information. Thus, there is always a chance to lie. In fact, research has shown that we lie a large percentage of the time within normal conversations (F iedler and Walka, 1993; Taylor et al., 1994; Turner et al., 1975). In our organizational lives, opportunities to deceive also exist and deceptive message behavior in the workplace is the focus of the following study. Deceptive message behavior (e. g., Buller & Burgoon, 1996; Buller et al., 1996; McComack, 1992; McComack et al., 1992) and deception in the workplace (e. g., Athanassiades, 1973, 1974; Gaines, 1980; Fulk & Mani, 1986; Mellinger, 1956; Muchinsky, 1977; O’Reilly & Roberts, 1974; Roberts & O’Reilly, 1974a; Zand, 1972) have previously been studied from a communication perspective while related work in the areas of organizational behavior (e.g., Bhattacharya et al., 1998; Lewicki et al., 1998; Rousseau et al., 1998) and organizational psychology (e. g., Sheppard & Sherman, 1998) have focused more broadly on issues of trust in the organization, not specifically deceptive message behavior. An important goal of this study is to bring together research that has broadly focused on trustworthy superior-subordinate relationships and research on the influence of relational trust to determine what effect these two variables have the deceptiveness of messages. Thus, a model is developed and tested that explores such relationships. Agency Theory (see Eisenhardt, 1989) and Social Exchange Theory (Blau, 1964) provide the theoretical framework for model development. Whitener et a1. (1998) assert that both theories explain the relational nature of trustworthy behavior of a superior on the trustworthy behavior of a subordinate. Agency Theory is based on economic self- interests. Both superiors and subordinates wish to maximize their rewards and minimize their costs within their relationship (Whitener et al., 1998). Social Exchange Theory (Blau, 1964; Whitener et al., 1998) claims a reciprocity exists between individuals. So, if a superior acts in a trustworthy manner, the subordinate will reciprocate by also enacting trustworthy behaviors (Whitener et al., 1998). Together, the two variables will be central to the development of the model that will be tested here. The current study is important to research in the area of organizational deception because it attempts to better understand and model previously studied variables along with introducing new variables. It includes variables from prior research on organizational distortion, such as ascendancy desires (e.g., Athanassiades, 1973, 1974; Gaines, 1980; O’Reilly, 1978; Roberts & O’Reilly, 1974) and trust in a superior (e.g., A Athanassiades, 1973, 1974; Fulk & Mani, 1986; Gaines, 1980; Mellinger, 1956; Muchinsky, 1977; O’Reilly & Roberts, 1974; Roberts & O’Reilly, 1974a; Zand, 1972). It also includes new variables called for by F ulk and Mani (1986) (e.g., the quality of the relationship, organizational trust, and impression management behaviors). The current research also contributes to our understanding of organizational deception through its additional focus on relational trust and through the provision of theoretical explanations for this effect that are desperately needed in the literature (see critique of interpersonal deception literature by Stiff, 1996). Trust is a variable found to have strong effects on the truthfulness of messages (Hubbell, 1999b) and the loss of trust has been shown to have a positive effect on the propensity to distort information (Gaines, 1980; Roberts & O’Reilly) and on deceptive behavior (Mellenger, 1956; Zand, 1972). Yet, an in-depth consideration of the influence of trust on the creation of truthful or deceptive messages is lacking. The current model theoretically explains the effects of trust and also seeks to understand the antecedents of trust. It will also demonstrate the influence of trust and distrust on truthful and deceptive messages. Understanding why subordinates trust or distrust is useful information to organizations because they are motivated to decrease the amount of distorted information communicated upward (Hildebrand, 1998). If an organization can understand why subordinates distort information given to their superiors, they can teach superiors behaviors that will increase subordinates’ desires to communicate truthfully. Some behaviors make a situation more conducive to truth-telling, even when telling the truth is not entirely comfortable (Hubbell, 1999). If organizations can better understand the factors influential to truth-telling, they can work to increase these factors. For example, if high trust in a superior is found to be a strong determinant to truthful communication, organizations can strive to help superiors foster trust from subordinates. . Finally, trust could be one element of a larger issue. Many interpersonal communication scales incorporate trust as an element that helps to describe the intimacy level of a relationship (Burgoon, 1984). The majority of past research in organizational deception has examined trust but not any other relational element that may influence information distortion. Thus, the first purpose of the current study is to determine if the quality of the relationship has a greater influence than trust alone on deceptive or honest messages. The second purpose is to create a better measure of trust in a superior and then to test the influence of trust on the truthfirlness or deceptiveness of information communicated to the superior. A third purpose is to determine how specific antecedent variables influence the variables of trust and relational quality. The following study consists of five sections. First, a review of the literature will be presented, including a previously tested model of organizational distortion, theoretical underpinnings, and elements of the proposed model. Second, the model will be developed and hypotheses stated. Thirdly, the methods used in examining this model are outlined. Fourth, the analysis of results are presented. Fifth, a discussion of potential implications of this study and future directions for research are provided. Chapter 2 LITERATURE REVIEW, MODEL CREATION, AND HYPOTHESES verview This chapter reviews the literature on (1) F ulk and Mani’s ( 1986) model of organizational distortion, (2) social exchange and agency theory contributions to studying trust and deceptive message behaviors, (3) the construct of superior-subordinate trust, (4) antecedents to trust, and (5) deceptive message behaviors. It also proposes the a model of organizational deception and the hypotheses and research questions focused on testing the model. Fulk and Mani’s Model of Information Distortion Fulk and Mani (1986) proposed and tested a model that portrayed three major influences on upward distortion: downward communication from the superior, role stress, and situational or contextual variables. The strongest link was the frequency of communication between superiors and subordinates. When superiors communicated frequently with subordinates, subordinates responded with increased communication. Also when superiors withheld information, subordinates responded with increased distortion and withholding of information. Thus, F ulk and Mani’s (1986) research demonstrated that the reciprocal nature of the frequency and quality of communication. Although, F ulk and Mani’s work provides a starting point for this study's exploration of deceptive communication, it lacks an overarching theoretical basis to explain the important and reciprocal nature of trust. Therefore, two theories are proposed to explain such effects: Agency Theory and Social Exchange Theory. Agmcy Theory and Social Exchange Theory Whitener et a1. (1998) propose Agency Theory and Social Exchange Theory as two useful theoretical explanations for the development of trust between a superior and a subordinate. These theories provide (1) essential information for the conceptualization and operationalization of relational trust and relational quality and (2) theoretical justifications for the links among trust and relational quality and various antecedent and outcome variables of interest to organizational deception scholars. Agency Theog. Although often thought of as a macro-level perspective, Eisenhardt (1989) claims Agency Theory is also applicable to a dyadic relationship. According to Whitener et a1. (1998), superiors and subordinates try to “structure the relationship to protect their own interests” (p. 514). In essence, individuals attempt to “maximize individual utility and ...both parties seek to minimize risks associated with the relationship” (p. 515). The benefits and risks in the relationship between a subordinate and superior, however, are more economically based than they are in Social Exchange Theory. The superior is interested in maximizing performance and the subordinate is interested in maximizing personal benefits (e. g., a raise). Further, a superior’s ability to accurately monitor or control a subordinate’s performance can be predicated on gaining truthful information from employees. This situation demonstrates the primary problem identified by Agency Theory. The problem is the difference in goals and attitudes between two individuals - a principle (superior) and an agent (subordinate) (Eisenhardt, 1989; Whitener et al.,1998). One way to reduce the disparity is to develop the level of trust between management and employees. As trust grows, perceived risk decreases, and, the superior may give the subordinate more responsibility. The subordinate, trusting that he/she will be rewarded, works harder. The quality of the relationship could also influence the perceived disparity in desires. According to Duck (1977) and Neirneyer et al. (1996), individuals are more attracted to individuals who are similar to them cognitively (Duck, 1977) or in their perceptions of others (N eimeyer et al., 1996). The individuals who we like are those individuals who are more like us, or at least we perceive them to be. Therefore, the quality of the relationship with another person could be the result of the perceived similarity. Those superiors who seem similar to the subordinates could be perceived as having similar desires. In terms of deceptive communication, if the subordinate does not trust or have a good relationship with the superior, he/she may be less likely to communicate problem information. The relationship, or belief in the manager because of the relationship, makes it “safe” for the subordinate to be truthful. Such reciprocity of behaviors is also a part of Social Exchange Theory (Blau, 1964; Homans, 1961, 1974). Social Exchange Theom. The basis of the Social Exchange Theory (Blau, 1964; Homans, 1961, 1974) is reciprocity and the mental assessment of costs and rewards. When superiors and subordinates exchange rewards, individuals may be prompted to positively evaluate the trust level in and the quality of the relationship. Whitener et al. (1998) describe this as follows: First, social exchanges may involve extrinsic benefits with economic value (e.g., information and advice) or intrinsic benefits without any direct objective economic utility (e. g., social support)... Second, whereas benefits in economic exchanges are formal and often contracted explicitly, such benefits are rarely specified in a priori or explicitly negotiated in social exchanges. Thus providing benefits is a voluntary action. Finally, ...the exchange of benefits involves uncertainty... Consequently, relationships evolve slowly, starting with the exchange of relatively low-value benefits and escalating to hi gher-value benefits as the parties demonstrate their trustworthiness (p. 515). As trust develops, individuals become more likely to take risks (Whitener et al., 1998). They become more likely to communicate information that could make them vulnerable. Also, Hubbell (1999b) found that subordinates are more likely to tell the truth to superiors they trust. Perhaps, then, subordinates who learn over time to trust their superiors will be more likely to communicate sensitive information in a truthful, rather than deceptive fashion. Further, if trust is one element of the quality of the relationship, the relationship should also drive the propensity of individuals to be honest or distort information. As the quality of the relationship develops, individuals should be less likely to distort information. Thus, Social Exchange Theory and Agency Theory will both provide the current study with a strong theoretical basis for measuring not only trust, but also linkages to antecedent and outcome variables. Our attention now turns to the concept of trust and the particular behaviors which engender trust from subordinates. Behaviors of Superiors Leading to Increased Tmst by Subordinates In the current research, the level of trust between a subordinate and a superior is an individual-level variable measured as a participant’s perception of trust. It is an important variable to consider when looking at deceptive message creation since trust has been found to be a strong predictor of truthful and deceptive messages (Hubbell, 1999b). The construct of trust will be described in terms of (1) the types of trust, (2) how trust will be defined in the current study, and (3) supervisors’ behaviors that lead to increased trust. Types of Trust. Rousseau et al. (1998) assert that there are three different types of trust. First, “Deterrence-based trust” (Rousseau et al., 1998, p. 398), is trust as a result of sanctions against untrustworthy behavior. Deterrence-based trust, however, is not included in the current model. The current model is based on Whitener et al.’s (1998) description of the elements that activate trust in a superior, and deterrence is not one of the factors listed by Whitener et al. (1998). Thus, it will not be considered in the current model. The second type of trust, “institutional-based trust” (p. 400) occurs when organizations “act as broad supports for the critical mass of trust that sustains further risk taking and trust behavior” (p. 400). Institutional trust is one form of trust represented in the current model as an antecedent to trust between a superior and a subordinate. It is believed that organizations that act in ways that elicit trust from employees will also motivate managers to enact trustworthy behaviors toward subordinates. Trust related to superior and subordinate relationships is the third type of trust discussed by Rousseau et al. (1998). “Relational trust” (p. 399) is a type of trust that develops over time between individuals. It is typified by “beliefs in the positive intentions of the trustee but also... the absence of negative intentions” (p. 400). The relational aspect of trust will be the primary focus of the current model. The following section elaborates further on the definition of trust to be used in the present study. Definition of Trust. A fundamental element of trust is risk. land (1972) claims that trust is a risk, it is making oneself vulnerable to an individual who cannot be controlled (Butler, 1999). Rousseau et al. (1998) also emphasize the importance of risk. It is the process of “exchange, risk taking, and successful fulfillment of expectations” (p. 399) that results in increased risk and increased resource exchange. Along with the perception of risk, there is an expectancy that the individual who is trustworthy will not take advantage of a vulnerable or risky situation. Rotter (1967) states that trust is “an expectancy held by an individual or group that the word, promise, verbal or written statement of another individual or group can be relied upon” (p. 651). Butler and Cantrell (1984) take this expectation a step further by claiming that individuals become trustworthy through their consistent use of honesty and integrity. Rousseau et al. (1998) incorporate risk, vulnerability, and expectations of future behavior in their defnrition of deception. This definition will guide the current research: Trust is a psychological state comprising the intention to accept vulnerability based upon positive expectations of the intentions or behavior of another (p. 395). Trust, therefore, is predicated by prior relational experience. The superior must enact behaviors which assure the subordinate he/she is acting in a positive manner and will continue to do so. Behaviors that elicit trust are discussed next. Tmstwoghy Behaviors. Whitener et al. (1998) includes integrity, honesty, and consistency in their description of a superior’s trustworthy behaviors. They claim five factors (or dimensions) effect a subordinate’s perception of a superior’s trustworthiness. First, “Behavioral Consistency” (p. 516) emphasizes the importance of positive actions over time. As subordinates begin to learn through repeated exposure that a superior will act in a predictable and positive manner, the subordinates will feel they are better able to predict behaviors and trust will increase. The second dimension of trustworthiness, according to Whitener et al. (1998), is “Behavioral Integrity” (p. 516). Whitener et al. (1998) emphasize Dasgupta’s (1988) two actions that lead to the perception of integrity. 10 These behaviors are: “(1) telling the truth and (2) keeping promises” (Whitener et al., 1998,p.516) The third dimension is “Sharing and Delegation of Control” (Whitener et al., 1998, p. 517). Such a dimension is very similar to the autonomy variable that is considered in the current research to be a moderating variable between the antecedent variables and trust in a superior and the quality of the relationship with the superior. Thus, the belief that a subordinate has an increased or decreased ability to control his/her situation is included as an influence on the veracity of a message, however, it is not considered as an element of trust. “Communication” (p. 517), the fourth dimension, and “Demonstration of Concern” (p. 517), the fifth dimension, may demonstrate a strong influence on the distortion of upward communication. According to Whitener et al. (1998), the three elements of commrurication that influence trustworthiness are: “(1) accurate information, (2) explanations for decisions, and (3) openness” (p. 517). In terms of demonstrating concern, trustworthiness is influenced by: “(1) showing consideration and sensitivity for employees’ needs and interests, (2) acting in a way that protects employees’ interests, and (3) refraining from exploiting others for the benefit of one’s own interests” (p. 517). Thus, four dimensions — behavioral consistency, behavioral integrity, communication, and demonstration of concern - should demonstrate an effect on the honesty or dishonesty of the messages communicated upward to superiors. Because of the reciprocal nature of commrmication between subordinates and superiors (F ulk & Mani, 1986), if superiors communicate honestly or in a way that reflects trust, subordinates will reciprocate by speaking honestly, or in a trustworthy manner. Honesty is an aspect of behavioral integrity that Whitener et al. (1998) claim to represent 11 trustworthy behavior. Using Social Exchange Theory, those superiors who enact behaviors that increase trust by a subordinate will receive more honest messages. The robust effect of trust in a superior on the ways information is communicated upward has been demonstrated by a moderate amount of research (e. g., Gaines, 1980; Hubbell, 1999b; Roberts & O’Reilly, 1974a). Individuals in high-trust conditions create more truthful messages (Hubbell, 1999b). Yet, when trust is low, one type of deception (withholdings, or white lies) has been found to increase (Gaines, 1980; Roberts & O’Reilly, 1974a). Therefore, individuals in high trust relationships are less likely to use withholdings and are more likely to tell the truth. Other researchers have not linked trust to specific message types, however, they have demonstrated the strong influence of trust on deceptive behavior and upward distortion. land (1972) claims that low trust will lead a subordinate to “conceal or distort relevant information” (p. 230). This is supported by Mellinger (1956) who asserts that distrusting subordinates will act evasively, aggressively, or compliantly, so as to deceive their superiors. Yet, trust may only be one part of a variable that could exert greater influence on the veracity of communicated information — the quality of the relationship. Qpalig pf the Relationship with a Superior - LMX The relationship between a supervisor and subordinate is strongly influenced by the emotional bond between them (Costigan et al., 1998). One measurement tool that has been used extensively to assess this bond is the Leader Member Exchange scale (LMX). LMX “evolved from the 2-item Negotiating Latitude measure” (Graen & Scandura, 1987, p. 190) which Dansereau et al. (1972) created to the multiple-item measures now used to assess LMX. The LMX scale has been used to evaluate the way that leaders and members communicate in order to sustain their roles (Dienesch & Liden, 1986) and their 12 relationship. Dienesch & Liden (1986) describe this as follows: It (LMX) is based on the concept that role-development will inherently result in differentiated role definitions and, therefore, in varied leader- member exchanges (p. 621). This type of role-development creates an in-group and out-group between leaders and their followers. In-group members have been found to have higher quality relationships with their superiors. Such relationships have been described as having “high trust, interaction, support, and formal/informal rewards” (Dienesch & Liden, 1986, p. 621). What is important to note is that the quality of the relationship with a superior has been found to influence communication patterns (Fairhurst, 1993). For example, subordinates with high quality relationships with superiors are more likely to use language that establishes a common ground with the superior (Fairhurst, 1993). Out-group members have a different type of relationship with their supervisors. Such relationships are described as being “characterized by low trust, interaction, support, and rewards” (Dienesch & Liden, 1986, p. 621). Fairhurst and Chandler (1989) describe the out-group members’ relationships to superiors as being “characterized by the use of formal authority, low trust, support and rewards” (p. 215). The subordinate’s perception of whether he/she is in the out-group or in-group can have a significant effect on job performance. Those individuals who have a higher quality relationship with their supervisors: report having more job responsibilities (Liden & Graen, 1980), contribute more significantly to their work groups (Liden & Graen, 1980), are more satisfied with their job (Schriesheim et al., 1992; Vecchio et al., 1986), are satisfied with supervisors (Gerstner & Day, 1997), and report more perceived l3 similarity with their managers (Kozlowski & Doherty, 1989). Also, those subordinates who have a higher quality relationship with their supervisors tend to have better performance appraisals (Deluga & Perry, 1994) and are more likely to be promoted (Wakabayashi et al., 1988). The reciprocity described in Social Exchange Theory could easily describe these effects. It could be postulated from the above findings that when there is a high quality relationship, both sides enact behaviors to maintain the relationship. Also, Agency Theory would argue that the subordinate maximizes his/her benefits in a high quality relationship. Thus, individuals who report having a high-quality relationship with a supervisor may be more inclined to be truthful with potentially harmful information than will those individuals who believe that they have a low quality relationship with their supervisors. Those with a low quality relationship perceive the relationship as not being beneficial to their organizational survival or success and may work to thwart it by being purposefully evasive. Therefore, the importance of trust in a superior and the relationship with the superior are key to the quality of communication between a subordinate and superior. Next, this review moves to an exploration of potential antecedents to the trustworthy behaviors and a high quality relationship between a subordinate and superior. WM Organizational structure has demonstrated mixed effects on the quality of the communication between subordinates and superiors (J ablin, 1987). According to J ablin (1987), there are four primary dimensions that dominate the research on organizational structure. These are as follows: (1) configuration (e. g., span of control, organizational size), (2) complexity l4 (vertical and horizontal), (3) formalization, and (4) centralization (p. 391). Of these four, two elements (configuration and complexity) are of primary interest in the current research. They both demonstrate some influence on the quality of the communication between a subordinate and a superior and, they are unbiased measures. mggp'zational Size. The size of an organization is generally accepted as the total number of employees (full and part-time) within the organization. According to Jablin (1987), research results on the relationship between the size of an organization and communication quality have been mixed (see Karsarda, 1974; Porter & Roberts, 1976). However, numerous studies have demonstrated trends, without significance, that when size increases, “the quality of communication (e.g., exchange pattern, openness) between superiors and subordinates decreases” (Green et al., 1983; Jablin, 1982; Klauss & Bass, 1982; Snyder & Morris, 1984; cited in Jablin, 1987, p. 397). In terms of how messages are influenced, Erez and Rim (1982) found that “middle managers in very large organizations (6,000 or more employees) tend to use ‘rational’ tactics more with their bosses than managers in small organizations (less than 599) employees)” (J ablin, 1987, p. 397). The superiors from the larger organization tended to use less personal, more out- group, ways of communicating. Individuals within a large organization may feel depersonalized and distrustful of the organization. They are “just a number” and are unable to influence in their work environment or their day-to-day operations. They may begin to believe that their voices are unheard and thus, trust and the quality of the relationship may decrease. Leading to a decrease in the quality of the upward communication between a subordinate and superior. A second structural element of an organization that could influence the quality of the communication within an organization is the horizontal complexity of the 15 organization. Horizontal Complexig. According to Jablin (1987), there have been several operationalizations of horizontal complexity: “number of occupational specialties, degree of extra-organizational professional activity, and number of organizational departments” (p. 402; also see Hage, 1974; Hage et al., 1971). Each operationalization has yielded different types of effects (J ablin, 1987). A different operationalization, by Bacharach and Aiken (1977) depicted horizontal complexity as the number of departments. As horizontal complexity increased, the “frequency of subordinate’s upward, lateral, and total commmrication” (Jablin, 1987, p. 402) decreased. Just as the increase in size increases depersonalization, increasing the number of departments may have the same effect. A subordinate may no longer identify with the organization that has a large number of departments. If they see their manager as an organizational representative, their trust in and quality of the relationship with the manager could decrease. Along with the structural assessments of an organization, there are perceptual elements of organizational climate that may influence trust in a superior and the quality of the relationship with the superior. The individual-level, or perceptual construct of organizational trust is one such element. annizational Trust Organizational trust from a subordinate’s perspective is defined by Gilbert and Tang (1998) as follows: Organizational trust is a feeling of confidence and support in an employer; it is the belief that an employer will be straightforward and will follow through on commitments (p. 322). Organizational trust has been found to be an important element in work 16 relationships (Chan, 1997; Gilbert & Tang, 1998; Lewis & Weigert, 1985). From a climate perspective, trust has been examined as low level employees’ perceptions of their CEO and top management (Fox, 1974). It also has been found to influence a subordinates’ perceptions of the organization (Muchinsky, 1977), their faith in the possible goal attainment of the corporation (Gilbert & Tang, 1998), job satisfaction (Driscoll, 1978), and their belief that the organization will act in their best interest (Gilbert and Tang, 1998). When organizational trust is lacking, subordinates begin enacting an “us against them” attitude that ultimately can affect employee performance (Gilbert & Tang, 1998). It also may increase a subordinate’s propensity to distort information. If individuals do not feel they will be treated fairly and cannot rely on their CEO and upper management to make decisions which will be mutually beneficial, they may be less likely to communicate sensitive information upward to their superiors. Subordinates may perceive the consequences of truthful communication as greater than telling a lie. According to Agency Theory, by telling the truth, rewards could be withheld or sanctions could be given. Organizational trust will be used in this research to assess its direct effect on relational trust (e. g., trustworthy behaviors of a superior) and the quality of the relationship between the subordinate and superior, as well as its indirect influence on truthful or deceptive message selection. Along with the perceptual and structural assessments of organizational climate, the individual-level factors of impression management and ascendancy desires also effect the distortion or truthfirlness of information communicated to a superior. l7 Imprgssjpn Mgagemgt Behaviors Impression management is an attempt to create positive impressions in other people (Rosenfeld, Giacalone, & Riordan, 1994, see Giacalone & Rosenfeld, 1989, 1991). Impression management has been defined from an organizational perspective as follows: Impression management, as defined by Giacalone (1989), refers simply to the tendency to present oneself in a socially desirable way to others - especially others in authority. Impression management has also been characterized as the defining and playing of roles in organizational life, and individuals ‘who fail to recognize this aspect of organizational life run the danger of performing poorly, or unwittingly being relegated to lesser roles as extras or understudies’ (Gardner, 1992, p. 17) (taken from Mendenhall & Wiley, 1994, p. 605). Thus, individuals with effective impression management behaviors can influence the perceptions of others, particularly in relation to trustworthy behaviors. A positive impression is particularly relevant to the quality of the relationship between a subordinate and superior. Support for this link is demonstrated in research showing impression management as positively related to LMX (Dockery & Steiner, 1990; Wayne & Green, 1993), a measure of the quality of the exchange between a subordinate and a superior. For example, Wayne and Green (1993) demonstrated a positive relationship between LMX and other-focused impression management behaviors, such as sincere flattery and favor rendering (Wayne & Liden, 1995). Similarly, Wayne and Liden (1995) found that supervisor-focused impression management behaviors increased the superior’s belief in his/her similarity with the 18 subordinate and liking of the subordinate. Ultimately, this leads to more positive performance appraisals (Wayne & Liden, 1995). Such findings can be explained through the lens of Social Exchange Theory. Those individuals (superiors or subordinates) who are able to effectively use other- focused impression management behaviors should be able to engender a positive feeling from the other. Also, since other-focused impression management behaviors emphasize trying to help the other person (e. g., favors) and trying to make them feel better about themselves and their performance at work (e.g., flattery), such behaviors may develop higher levels of trust between individuals. For example, imagine a subordinate who often does favors for a superior. The superior learns from this behavior that the subordinate can be “counted on” to help out. Thus, the superior’s trust in the subordinate increases. As the superior’s trust in the subordinate increases, the subordinate reciprocates by increasing trust in the superior. The superior may also use other-focused impression management behaviors to increase trust. For example, the superior writes a very flattering remark about the subordinate on a performance appraisal form. According to Agency Theory, the subordinate will believe that the praise is a reward for hard work and trust in the superior increases. However, another variable that may influence these variables is ascendancy. Ascendancy desires may motivate a subordinate to enact other-focused impression management behaviors so they can increase the level of trust a superior has of him/her and increase the quality of the relationship with the superior. Also, ascendancy desires have been shown to have some influence on upward distortion. 19 As anc Desires Ascendancy or achievement needs represent a subordinate’s desire for upward mobility. According to Athanassiades (1974), it is “a drive or striving to ascend, to rise above others in a hierarchy” (p. 199). Turner (1967) argues that the need to achieve comes from the “desire to excel in accordance with standards of excellence” (p. 147). So, to excel is not enough, one must also. achieve excellence. Gaines (1980) combined the two ideas into the belief that “aspiration is instead a twofold phenomena involving desire for upward mobility and achievement” (p. 931). Athanassiades (1973, 1974) and Read (1962) found a positive relationship between ascendancy desires and distortions. Yet, Gaines (1980), demonstrated the opposite to be true. Hubbell (1999b) attempted to explain the difference between the two findings. She claims that the difference is simply one of level of specification. Gaines (1980) includes differing types of messages with varying amounts of distortion and Athanassiades (1973, 1974) and Read (1962) simply asks if subordinates if they distort information. In a separate test, Hubbell (1999b) found partial support for Gaines (1980) in that individuals in a high ascendancy condition tend to use more truthful messages than individuals in a low ascendancy condition. Ascendancy desires may indirectly influence trustworthy behaviors of a superior through their influence on behaviors that lead to increased trust (i.e., other-focused impression management behaviors). According to Agency Theory, a subordinate seeks to “minimize risks” (Whitener et al., 1998, p. 515) to the relationship and to also increase benefits. In order to accomplish this, the subordinate tries to demonstrate his/her utility to the superior and to benefit from this demonstration (Whitener et al., 1998). The same may be true for the quality of the relationship. The subordinate may want to increase the 20 quality of the relationship to get promoted and enacts behaviors (i.e., other-focused impression management behaviors) to increase his/her “likability.” All of the above antecedents to trust (i.e., organizational size, trust, and complexity, as well as impression management behaviors) in a superior and the quality of the relationship with the superior may be moderated by another variable, a subordinate’s perception of job autonomy. Autopgmy Autonomy is the amount of independence individuals have in their work i that enables them to have control over and make decisions about how they will accomplish their goals (Hackman and Oldham, 1975). According to Breaugh (1989), it is distinct from the amount of interdependence individuals have with superiors and Breaugh argued that: Turner and Lawrence (1965) viewed autonomy (‘the amount of discretion the worker is expected to exercise in carrying out assigned work activities’) solely in terms of worker latitude or control. They posited another task attribute, ‘required interaction,’ to encompass worker interdependence/independence (p. 1035). In a 40 study meta-analysis of the influence of autonomy, Spector (1986) found that “autonomy is strongly related to employee turnover, absenteeism, performance, motivation, physical ailments, emotional distress, and job satisfaction” (Breaugh, 1989, p. 1033). Past organizational deception research (Athannassiades, 1973, 1974) has demonstrated that level of autonomy negatively influences the propensity to modify information communicated to superiors. However, the influence of autonomy on the quality of upward communication 21 may be better explained as indirect, rather than direct. It could moderate the relationship between the antecedent variables and trust in a superior or relational quality. Individuals with higher levels of autonomy may also have greater motivation and ability to perform impression management behaviors well. The result would be to increase the effect of impression management behaviors on trust and the quality of the relationship. The relationship between a superior’s other-focused impression management behaviors and trust in a superior and the quality of the relationship with the superior could be influenced in a similar fashion. Higher levels of autonomy, as perceived by a subordinate, might lead to an increase in the effect of impression management behaviors on trust and the quality of the relationship. A second moderator variable occurs between both the trust in a superior and the relationship between a superior and subordinate and their influence on message selection, interdependence. lpterdgpendepce Interdependency in the current research is the extent to which individuals depend on each other for task and social resources in order to fulfill their job requirements. According to Whitener et al. (1998), the interdependence between a superior and a subordinate may change the relationship between a superior’s trustworthy behavior and the subordinate’s resulting trustworthy behavior. For example, a superior may be very honest with a subordinate. If the relationship is also highly interdependent, this trustworthy behavior should demonstrate a stronger effect than it would in a less interdependent relationship. So, in return, the subordinate will be more likely to be 22 honest with the superior that he/she has a more interdependent relationship with. The link between the quality of the relationship with a superior and message selection should also be similarly influenced. Highly interdependent superior- subordinate dyads must communicate more to coordinate the sharing of resources. According to Fiske and Taylor (1991 ), mere proximity to individuals (being around them a lot) is likely to increase our attraction to them. The highly interdependent dyad, therefore, could also be described as one in which there is a higher quality relationship than in a less interdependent dyad. The greater the relationship, the less likely the subordinate would be to distort information. In summary, past research leads us believe that organizational trust, impression management behaviors, organizational structure, and ascendancy desires are all antecedents to relational trust and the quality of the relationship with a superior. They are all moderated by the level of autonomy experience by a subordinate. The outcomes of trust and the quality of the relationship are specific subordinate's message choice behaviors. Message selection is moderated by the level of interdependence in an individual’s work. The following section will outline the five message choice categories to be predicted from relational trust and quality as well as how Social Exchange Theory and Agency Theory provide a rationale for predicting message choice. Creating a Typology of Deceptive Messages The deceptive message literature in the areas of organizational deception (e. g., Fulk & Mani, 1986) and interpersonal deception (e. g., McComack, 1992; McComack et al., 1992) provide the background for the model's specific predictions of message choice. There are three schools of thought in the organizational deception literature - strategic ambiguity (e.g., Eisenberg, 1984), distortion of information (e. g., Athanassiades, 1973, 23 1974; Gaines, 1980; Roberts & O’Reilly, 1974a, 1974b), and total deception or falsification of relevant information (e.g., Grover, 1993; Taylor, 1911). The primary difference among the three perspectives is in how each believes information is changed within a message. For example, strategic ambiguity focuses on a message being vague or communicating multiple meanings (Eisenberg, 1984). The information distortion perspective, however, has introduced different message types, none of which are ambiguous (see Athanassiades, 1973, 1974; Gaines, 1980; McEvoy, 1990; Mellinger, 1956; Roberts & O’Reilly, 1974a, 1974b; Swanson, 1996; Zand, 1972). The third perspective - the total deception perspective - primarily examines messages that are complete distortions of the truth (e. g., Grover, 1993, Taylor, 1911). In a comparison among these three approaches, Hubbell (1999b) asserts that they all define deception in a similar manner. All three emphasize the intentional manipulation of relevant information and this is what makes a message distorted, or deceptive. Thus, only that information that is important or relevant information, when purposefully altered, is deceitful. Based on the fact that each perspective shares a common definition of deception, Hubbell (1999b) created a typology of deceptive messages that combined the three perspectives and tested them for their appropriateness and honesty. The following section will describe the different types of deceptive messages to be explored in this research. Types pf Dgggptivg Messages The types of messages included in the typology have been taken from Grice’s (1989) maxims of conversations (Hubbell, 1999b). Using McComack’s Information Manipulation Theory (IMT, 1992) as a model, each of the types of messages within the three perspectives are compared to potential violations of Grice’s (1989) maxims. 24 Violations of the maxims would mean violating conversational expectations regarding the Quantity, Quality, Relevance, and Manner of information communicated in a message (Grice, 1989). The first of the message types described by Hubbell (1999b) represent a violation of Quantity. It also represents the first of three message types taken from the information distortion perspective (see Hubbell, 1999b) and it is called “withholdings.” Withholdings — Violations of Quantig. Withholding messages contain manipulations of the total amount of information that a subordinate communicates to a superior (Roberts & O’Reilly, 1974a). Withholdings represent a violation of Grice’s (1989) maxim of Quantity in that some, but not all, of the relevant information is communicated. Withholdings have also been called gatekeeping messages (Roberts & O’Reilly, 1974a). Researchers have tried to create distinct operationalizations for withholding and gatekeeping messages, yet these operationalizations tend not to describe mutually exclusive categories. For example, Fulk and Mani (1986) wanted withholdings and gatekeepings to represent different message types, yet, they used similar operationalizations for each. In their research, assessments of the two messages emphasize the amount of distortion. The only difference between the two is that gatekeeping refers to the to_tal M of information not communicated to superiors and withholding refers to the holding back of useful information. Since in the definition of deceptive messages given here, only the distortion of relevant information constitutes deception and what information is withheld in gatekeeping messages is unknown, gatekeepings cannot be considered deceptive messages. Only withholdings remain as representative of deception. A third message considered to be similar to withholdings are called “sieving” 25 messages (Gaines, 1980, p. 936). Such messages demonstrate a bias toward being positive or negative (Gaines, 1980). What is unknown is whether the preponderance of messages are either positively or negatively valenced. Gaines (1980) also does not indicate if these biases are related to any potential effects. However, other research has demonstrated a bias to communicate favorable information upward (O’Reilly & Roberts, 1974). Thus, sieving messages may be similar to withholding messages in that some information (the favorable kind) will be communicated upward and the unfavorable information will be filtered out of the message. Despite the seemingly different operations of withholding distortions, it does seem that some commonalties exist. First, sievings and withholdings seem to represent the same type of deceptive message. Second, favorableness and relevance of information within a message can be distorted. This is unlike distortions where all of the relevant information is changed. Distogigns — Violations of Qpalig. Distorted messages are completely false and involve “actively changing the nature of the information transmitted” (F ulk & Mani, 1986, p. 484). Distortions as a message type represent a complete falsification of information. They are violations of Grice’s (1989) maxim of Quality. The maxim of Quality dictates that information be completely truthful and that relevant information should not be modified (Grice, 1989). Because distortions completely modify the relevant information, they are considered to be Quality violations. Although the term “distortion” comes from the information distortion perspective, this type of message is also contained within the third school of thought in the organizational deception literature -— the total deception perspective. Research in this area has not been as descriptive of specific message types, however, Grover (1993), believes 26 that information can be manipulated in two ways. It can be completely distorted or it can be a simple twist of the truthful information. The complete distortion is similar to the “distortion” message just discussed and represents a violation of Grice’s (1989) Quality maxim. The slight twist of the truth is similar to the withholding message described above and is an example of a violation of Grice’s (1989) Quantity maxim. Up to this point, only violations of Quality and Quantity have been described. There are two other message types that represent violations of Manner and Relevance. Ambigpgus Messages - Violations of Manner. Eisenberg’s (1984) description of ambiguous messages is very similar to what have been referred to as equivocations in the interpersonal communication literature (Bavelas, 1983; Bavelas, 1985; Bavelas et al., 1988). Such messages are vague and can also communicate multiple meanings. Similarly, Eisenberg (1984) sees strategically ambiguous communication as including vague or equivocal language. This type of message is a violation of Grice’s (1989) maxim of Manner. It is the manner in which the information is communicated that makes it unclear because individuals either have to infer a meaning or choose from multiple potential meanings. One particularly interesting example of an equivocally ambiguous message is the use of metaphors within organizations. A commonly used organizational metaphor is that of a corporation being a “family” (e. g., Grant & Oswick, 1995; Smith & Eisenberg, 1987). The family metaphor can represent many different meanings. For example, a family may be close and have traditional values but, in times of crisis, should be accepting of change. The organization can thus exploit any of these particular meanings at any time - if so desired. If organizational members buy into the family metaphor, they may also unknowingly buy into all of its potential meanings. 27 So, up to this point there are three types of messages that can be considered deceptive - withholdings, distortions, and ambiguous messages. There is one remaining that represents a violation of Grice’s (1989) maxim of Relevance. Avoiding or Chapging the Subject Messages — Viplatiops of Relevgge. When individuals avoid a sensitive topic by changing the subject, they are violating Grice’s (1989) maxim of Relevance. The maxim of Relevance dictates that only information relevant to the situation should be communicated. Although messages that avoid a subject are not included in any of the three organizational deception perspectives, Hubbell (1999b) included them based on their violation of Relevance (Grice, 1989) and because they have been found to be an important message type in other deception research (e.g., McComack et al., 1992, Hubbell, 1999a). Hubbell (1999b) found them to be an alternative message type, yet, they were not often used. Although all the deceptive message types have been found to be used to communicate difficult information, unexpectedly the truth also has been shown to be used quite often (Hubbell, 1999b). Hubbell (1999b) found individuals often create truthful messages - particularly in both high trust and high ascendancy, or ambition, conditions. Such information may be useful for organizations. Organizations may feel that it is important to understand the factors that lead to the creation of truthful messages in addition to deceptive messages. By understanding such factors, organizations can place more emphasis on cultivating them so that subordinates are unafraid of communicating relevant but potentially harmful information. All of the variables and messages were combined into a model whose purpose was to predict when individuals would choose certain types of messages. 28 M of anizational Dec tion and Truth The purpose of this section is to place the variables previously discussed into a model that predicts when individuals will choose to distort information in particular ways. Two parallel models are constructed with different central variables. In one version of the model, trust in a superior is tested as the primary antecedent to message selection. In the second version, the quality of the relationship is tested as antecedent to message selection. The first set of hypotheses will be centered on trust and the second set on the quality of the relationship. First, the general model will be introduced and illustrated. Following a description of the model, each of the specific links and the related hypotheses will be described. Based on the prior discussion regarding antecedent variables, types of deceptive messages, and deceptive tactics, the following model is proposed (see Figure 1). 29 Autonomy Interdependence v « I o ' Organizational ++ El Trust + . /:: Ambiguous I Ascendancy Trust in 3 Superior Desrres and the Quality of — Withholding + the Relationship ' + with the Superior Subordinate Other- ’ lDistortion Focused Impression Superior’s Other- Focused Impression + Organizational Structure F igm 1. Model of Organizational Deception and Truth anm'zational Tgst and last in a Superipr. Muchinsky (1977) found that trust in a superior is related to trust in the organization. Also, when organizational trust is violated, subordinates begin to feel antagonism toward the corporation. The superior, often viewed as an agent of an untrustworthy organization, would be considered untrustworthy as well. Thus, if individuals believe they can trust their organization, this should lead to positive perceptions of trust for their supervisors. Therefore, the following hypothesis is postulated: H1: Organizational trust will be positively related to a subordinate’s trust in their superior. Sirrrilarly, the organizational structure should also influence the trust in the superior. As 30 discussed above, size and organizational complexity may have a negative influence on the trust in a superior. Thus the following relationships are hypothesized: H2: The size of the organization will be negatively related to the level of trust in a superior. H3: The organizational complexity of an organization will be negatively related to the level of trust in a superior. Other links described by the model also influence trust in a superior. They include impression management behaviors of the superiors and subordinates. Impressipp Mgragement. As previously discussed, impression management behaviors that are other-focused (i.e., benefit a superior) are believed to be positively related to eliciting trustworthy behaviors on the part of the superior. One particular type of other-focused impression management should have the strongest positive effects - ingratiations. Ingratiations are behaviors focused on increasing another person’s liking of the ingratiator (Dockery & Steiner, 1990; Wayne et al., 1994). They include actions such as favor-rendering and flattery (Wayne & Liden, 1995). Behaviors such as these should have a positive effect on trustworthy behaviors of a superior, as just discussed. Based on this, the following hypothesis is proposed: H4: Subordinates’ other-focused impression management behaviors will be positively related to the level of trust between subordinates and superiors. Due to the fact that the relationship between the subordinate and superior is dyadic and the subordinate’s impression management behaviors are believed to influence perceptions of a superior, the reverse should be true as well. Thus, the following hypothesis is proposed: 31 H5: Superiors’ other-focused impression management behaviors will be positively related to the level of trust between subordinates and superiors. One additional individual-level variable, ascendancy desires, has a direct effect on other- focused impression management behaviors, and thus, will demonstrate an indirect effect on the trustworthy behaviors of a superior. Ascendancy Desires. An individual with high ascendancy desires may make a decision to improve the relationship with a supervisor. To accomplish this, the subordinate may enact other-focused impression management behaviors. Therefore: H6: With trust as the central variable in the model, ascendancy desires will demonstrate a positive influence on the other-focused impression management behaviors of subordinates. All of the variables up to this point have been discussed in relation to their direct or indirect influence on a subordinate’s trust in a superior. Yet, the influence trust on message selection has not been described. Thus, the following section will elaborate on these effects. last in a Superior — Trustworthy Behaviors of a Superigr. Trust has been found to have a strong positive relationship to the use of truthful messages (Hubbell, 1999b). Thus, high levels of trust should lead to more truthful messages and low levels of trust should be lead to deceptive message choices. Therefore, the following hypothesis is proposed: H7: High trust in a superior will be positively related to truth-telling and negatively related to deceptive message selection. Although it is believed that trust in a superior is positively related to truthful message creation, it will have different effects on the creation of other messages included 32 in the model. Hubbell (1999b) found a strong negative relationship between trust in a superior and the use of distortion messages. Further, the other message types demonstrated weak negative relationships with trust in prior research (Hubbell, 1999b). Therefore, the following hypothesis is proposed: H8: Distortions will be chosen more often than the other deceptive messages in situations of low trust in a superior. The Quality of the Relationship with a Superior. The quality of the relationship with a superior is also believed to be influenced by the same antecedent variables as trust. Therefore, the following hypotheses are posed: H9: Organizational trust will be positively related to the quality of the relationship between a subordinate and superior. H10: The size of the organization will be negatively related to the quality of the relationship between a subordinate and superior. H11: The organizational complexity of an organization will be negatively related to the quality of the relationship between a subordinate and superior. H12: Subordinates’ other-focused impression management behaviors will be positively related to the quality of the relationship between a subordinate and superior. H13: Superiors’ other-focused impression management behaviors will be positively related to the quality of the relationship between a subordinate and superior. Ascendancy desires will again be tested and thus, 33 H14: With the quality of the relationship as the central variable, ascendancy desires will demonstrate a positive influence on the other-focused impression management behaviors of subordinates. The quality of the relationship with a superior should also demonstrate the same type of influence on message selection, as does trust in a superior. The following hypotheses depict such relationships: H15: A high quality relationship will be positively related to truth-telling and negatively related to deceptive message selection. H16: Distortions will be chosen more often than the other deceptive messages in situations of low quality relationships with superiors. The Role of the Moderator Variables. Although exploratory, the potential effects of the two moderators will also be investigated in the model. The following research questions will be answered: R1: Does autonomy moderate the relationship between the antecedent variables and the level of trust in a superior and does it moderate the relationship between the antecedent variables and the quality of the relationship between a subordinate and superior? R2: Does interdependence moderate the relationship between the level of trust and message selection and does it moderate the relationship between the quality of the relationship and message selection? The tests of the hypotheses and methods for answering the research questions will be the focus of the following chapter. 34 Chapter 3 METHODS Pretest m Twenty nine students from a large Midwestern university were recruited in their communication classes. They were offered extra credit for their participation. Emcedure. Before any questionnaires were distributed, participants were assured of the complete confidentiality and anonymity of responses. The volunteer nature of their participation and freedom to withdraw at any point was also communicated. After such verbal assurances, participants received written verification of the confidential and voluntary nature of the study along with a brief description of the research. Once they decided to participate, they were asked to sign a consent form and return it to the researcher before receiving any materials. After participants completed the consent forms, the consent forms were placed in a sealed envelope which only the primary researcher could open. The research participants were then given a questionnaire. They were told the purpose of the survey was to better understand organizational relationships. After completion of the survey, respondents were asked to place it in a second sealed envelope. They were then told the purpose of the research, any questions were answered, and they were thanked for their time. Inst;ru_nent. Items in the pretest measured a superior’s trustworthy behaviors, a subordinate’s perceived autonomy, superior and subordinate other-focused impression 35 management behaviors, a subordinate’s ascendancy desires, trust in the organization, the lie acceptability scale, and demographic information. Scales to Measure Variables. The variables were measured in the pretest to test the reliability of the scales. Trust in the Organization. Organizational trust was measured using ten items. One reason for this was that there were several scales that could be used to assess organizational trust. However, they often have few items and some of the items differ substantially among the scales. This may be due to the fact that there are different dimensions of trust in an organization. For example, Lumata’s (1988) measurement of trust looked at the belief in the believability of upper management as well as whether individuals in general within the organization will “do things behind your back.” These two questions appeared to measure different aspects of trust in an organization - one looked at behaviors of upper management and the other at behaviors of all individuals within the corporation. Therefore, three items assessing organizational trust were taken from (Lumata, 1988). The three items were (Lumata, 1988, p. 57):: Item 1: Employees here feel you can trust top management Item 2: When top management here says something you can really believe it is true Item 3: People in this organization will do things behind your back Lurnata (1988) found a reliability of .77 for the three items. To better understand these dimensions of organizational trust as well as determine if other dimensions exist, seven items were added to Lumata’s (1988) questions. Two of them were based on the description of important elements of organizational trust, as discussed earlier. Two more came from Hubbell (1999b) and adapted from items used to 36 assess trust in superiors. The last three were fiom Cook and Wall (1980). All seven of items were as follows: From description of organizational trustzz Item 1: In my organization, upper management follows through on commitments Item 2: Actions taken by upper management are in my best interest From Hubbell (1999b): Item 3: Upper management makes good decisions Item 4: Upper management is fair with employees From Cook and Wall (1980): Item 5: Upper management at my firm is sincere in its attempts to meet the workers’ point of view Item 6: Upper management can be trusted to make sensible decisions for the finn’s future Item 7: I feel quite confident that the firm will always try to treat me fairly (the word “Upper” was added to items 3 and 4) (p. 50). Subordinate ’s Other-Focused Impression Management. Wayne and Liden (1995) found that some other-focused impression management strategies had a more positive influence on LMX than other impression management behaviors. Other-focused impression management behaviors are called ingratiations and include favor-rendering and flattery (Dillard et al., 1989; Wayne & Liden, 1995). An example of the questions that will be asked (see Appendix A for the other items) is, “How often do you compliment your supervisor on the quality of his/her work?” Superior ’3 Other-Focused Impression Management. To assess the perceived amount of ingratiatory behaviors a superior makes toward a subordinate, the questions 37 used for subordinate other-focused impression management behaviors were altered to reflect a superior’s actions. For example, one item was, “How often does your manager seem to go out of his/her way to say something nice to you? ” Autonomy as a Part of a Job Design. Hackman and Oldham (1975a, 1976b) measured autonomy as a factor within their Job Diagnostic Survey. Reliability estimates of .66 and .64 have been found for the 1975(b) version. The three items from this version to be used here were: Item 1: How much autonomy is there in your job? That is, to what extent does your job permit you to decide on how to go about the work? Item 2: The job denies me any chance to use my personal initiative or judgement in carrying out the work. Item 3: The job gives me considerable opportunity for independence and freedom in how I do the work. Ascendancy Desires. Hubbell (1999b) created three items based on earlier operationalizations of the desire for upward mobility within the organizational deception literature. One item was deleted from Hubbell’s (1999b) analyses because it did not ' correlate highly with the others. The two remaining items were correlated (r = .59) and were maintained for analysis. The items were as follows: Item 1: I would like to be promoted Item 2: I consider myself to be ambitious The two items appeared to measure the ascendancy desire of an individual. Yet, there was a second element of ascendancy desires - the desire for achievement or to excel (Turner, 1967). Therefore, four items from a Need for Achievement Scale used by Miller et al. (1999) were added to the two items. The Need for Achievement Scale has been 38 adapted from a five-item scale created by Steers and Braunstein (1976). Four items were retained in Miller et al.’s (1999) analysis, with a reliability of .58. An example item was: “I do my best work when my job assignments are fairly diflicult. " Ascendancy, along with the other individual-level and organizational-level variables, was hypothesized in the current research to influence the trustworthy behaviors exhibited by a superior, increasing the level of trust in that superior by a subordinate. Trust in 0 Superior. Four dimensions of trust in a superior were previously discussed. The first dimension was Behavioral Consistency. The predictability of behavior over time would be indicative of how trustworthy a manager was believed to be. Five questions similar to the following were used to assess the dimension: Item 1: My manager’s behavior is very predictable Item 2: My manager is very consistent in his/her behavior The second dimension was Behavioral Integrity. According to Whitener et al. (1998), two behaviors tell a subordinate a superior has integrity. These were “telling the truth” and “keeping promises” (Whitener et al., 1998, p. 516). An example of the five questions used to assess truth-telling was, “I would expect my supervisor to tell me the truth” (adapted from the Specific Interpersonal Trust Scale, J ohnson-George & Swap, 1982). An example of one of the five questions that were used to evaluate keeping promises was, “I can completely rely on my supervisor to keep promises made to me ” (adapted from the Trust Scale, Rempel & Holmes, 1986). Behavioral consistency and integrity were the first two dimensions of trust, communication was the third. The focus of the third dimension was the manner and quality of information communicated to subordinates and the subordinates’ perceptions of their ability to communicate openly with superiors. There were three elements which 39 represent the communication dimension of trustworthy behavior. The first element of the communication was “accurate information” (Whitener et al., 1998, p. 517). If a superior communicates accurate information to a subordinate, he/she would be perceived as more trustworthy. An example of the five items used to rate this element was as follows: “I believe my superior is very precise in his/her communication with me. ” The second element of the communication dimension was the superior’s propensity to explain decisions. The example question for this element was as follows: “My superior always explains his/her reasons making decisions. ” The third element of communication was openness. An example item for this element was: "Does your manager make you feel free to talk to him/her? ” (taken from Miller et al., 1999). The three elements that represent the third dimension, were similar to the elements used to assess the fourth dimension — demonstration of concern. Again, three elements comprise the demonstration of concern dimension. The first was a superior’s demonstration of concern and sensitivity to their subordinates. An example item for the element was as follows: ”Is your manager genuinely interested in your welfare? " (adapted from Miller et al., 1999). The second element was a superior’s protection of a subordinate’s interests. The example item for this dimension was, “My superior will always make efl'orts to protect my interests. " The third element for demonstration of concern was a superior’s ability to avoid hurting the subordinate in order to benefit the superior. For example, the superior could get an idea from the subordinate and claim that idea as his/her own. The superior would benefit and the subordinate would not. The following was an example question for this element: "I feel that I can trust my manager with information which could be held against me ” (adapted from Miller et al., 1999). All of the above dimensions of trust were to be evaluated in the experimental 40 phase to determine the influence they exhibited on either the manipulation or lack of manipulation of information communicated to a superior. However, one antecedent was believed to moderate the relationship between trust and the choice of a deceptive or honest message - interdependence. Level of Interdependence. Interdependency in the current research was the extent to which individuals depend on each other for task and social resources in order to fulfill their job requirements. An example question that was used to assess this variable was, "My manager and I often share resources in order to accomplish tasks. " Table 1 illustrates the resulting number of items for each scale and the reliability estimates found in the pretest (N=29). 41 Table 1. Reliabilig Estimates and Numbers of Items * Scales Reliability Number of Items Assessments of Trust in a Superior: Consistency .75 4 Keeping Promises .94 5 Telling the Truth .89 4 Interest in Subordinate’s Welfare .95 6 Protection of Subordinate’s Interests .93 5 Would Not Hurt Subordinate .88 5 Communicate Accurately .91 5 Explains Decisions .70 3 Openness to Communication .92 4 Impression Management Behaviors .91 5 Of Superior Assessment of Relationship: Leader-Member Exchange Scale .93 7 Assessments of Organizational Trust: Organizational Trust Scale .95 9 Subordinate Assessments of Him/herself: Ambition .64 5 Autonomy .81 5 Impression-Management Behaviors .87 5 Interdependence .90 6 *All of the scales can be found in Appendix A W Recmitment. A cross-sectional design was used. Participants from four different organizations responded to the surveys. In order to motivate participants to complete the surveys, they were informed of a prize drawing. Each individual who completed the survey had a chance to win a gift certificate to a large grocery chain. 42 The number of participants needed to support the refutation of the null hypothesis was estimated with a power analysis. A moderate correlation of .30 was selected for the analysis. Based on this, for a power estimate of .85 (r = .30, 2-tailed significance of .05), 97 individuals would be needed. Therefore, the target number of individuals for this research was 100. The resulting number of participants was 99. Prpcedure -— gm Site in Four Organizations. Participants from four different organizations participated in this phase. In three of the four organizations, individuals were given the consent form and survey by the researcher. The experiment took place either during a staff meeting or during assigned times. For example, in the largest organization (Auto Supply Company), individuals were asked to come into a training room. They were assigned a time to attend based on their work group. In the two smallest companies (Credit Unions), survey collection took place during morning staff meetings. In the fourth organization (Public Relations Company), surveys were placed in mailboxes and participants were asked to return the surveys to their Human Resources Department and to place the consent forms and questionnaires in separate envelopes. Participants first received the consent form and were informed of the confidentiality of their responses. After completion of the consent form, respondents received the questionnaire. It took approximately 20 minutes for each survey to be completed. When surveys were turned in to the researcher, the respondents were thanked for their time. After all surveys were collected, a name was drawn from a hat for the winner of the gift certificate and the winner was notified. Instrument. The pretested survey items were contained in the first section of the questionnaire. It was joined with an additional section where more information on message creation, message selection, and additional information was gathered. In the 43 additional section, respondents first read a hypothetical situation. The situation was taken from Hubbell (1999b) who found a strong effect for trust and a moderate effect for ascendancy desires on message generation (see Appendix D). The situation was slightly changed in this research. Participants were told to try to consider themselves and their managers in the situation. They also received the hypothetical situation after being asked to respond to all of the scales measuring the antecedent variables. This made the issues raised by the questions more present in the respondents’ minds as they thought about the hypothetical situation. After reading the hypothetical situation, participants were asked to read several alternative messages and rate how likely it would be for them to have used those messages as well. There was an important purpose for this measure. It could create a continuous measurement of message types. Past research has considered only categorical data - only certain types of messages have been considered (see Hubbell, 1999b). Categorical data is useful but makes it difficult to test the entire model. Thus, analyses of continuous data (likelihood of use of messages) was the focus of the current research. The messages that were assessed for likeliness of use came from Hubbell (1999b). Two coders evaluated the messages and placed them into the message type categories (Hubbell, 1999b). The messages chosen to be used here received identical ratings fiom both coders, therefore, they were demonstrated to consistently be seen as representing a particular message type. The following includes examples of each of the messages: Complete Truth: “1 saw Jesse take the office supplies.” Withholding: “You are right, I have noticed some things are missing.” Ambiguous: “Things always seem to come up missing around here. They tend to turn up eventually.” Changing the subject: “That reminds me, I was just reading an article about how to keep an office supply inventory. Do you want to see it?” Distortion: “I am sorry, I haven’t noticed anything missing at all. I don’t know what could be happening.” Respondents were given 5 examples of each of the message types and asked how likely it might be that they would use each of the messages. They also were told that it would be entirely acceptable for them to perceive several of the messages as viable and that they may see a message similar to the one they have already produced. However, they were instructed to not turn back and change the message they created in response to the situation. Along with the scales assessing the independent variables and the hypothetical situation, participants were asked to answer additional questions for manipulation checks, potential confounds, and demographic items. Manipulation Check. Participants were asked two questions pertaining to whether they were able to place themselves in the given hypothetical situation. Demographic Information. Participants were asked several questions regarding their amount of hours worked a week, age, ethnicity, sex, and current occupation. Potential Confounds. One potential confound was an individual’s level of lie acceptability. Some individuals could be more sensitive to deceptive behavior than others. To rule out the bias caused by extreme levels of lie acceptability (e.g., it is not at all acceptable or it is acceptable to lie), the Lie Acceptability Scale (Levine et al., 1992) was used to assess whether individuals’ messages were influenced by the potential bias. 45 Another potential confound was a recent major organizational event, such as restructuring. Restructuring is an event that could alter most of the antecedent variables in the model. Organizational trust may be changed due to subordinates seeing colleagues being let go. Ascendancy desires could be thwarted due to the fact that subordinates have to releam what opportunities are available, and how to gain access to them. Relational quality would also be effected because a new superior may be put in place and/or the subordinates’ decreased organizational trust and ascendancy desires would lead to decreased relational quality. Therefore, to rule this confound out of the current research, respondents were asked if their organization has been restructured in the last year. A third potential confound was the amount of time a subordinate had worked with his/her current superior. Trust is a relational component that develops over time. Thus, subordinates may have a low level of trust due to not having long enough time to evaluate the superior’s consistent trustworthy behaviors. The fourth potential confound was an individual’s level in the organization. Because individuals who are managers may be biased in perceptions of other managers, their responses may have to be controlled for. Thus, all individuals were asked if they were cm'rently in a managerial position. 46 Chapter 4 RESULTS Ma The following chapter will describe the specific analyses which were performed on these data. The description will begin with the analysis of the scales. This will be followed by an exploration of potential differences among the four organizations. Next, the two versions of the model will be tested. Experiment Ninety nine individuals from four different organizations participated in the current research. The mean age was 35.71 (Median = 32, Mode = 29) and the majority worked at least 40 hours a week (Mean = 41.83). The majority of respondents were caucasian (N=76), with a small amount of individuals in the following groups: hispanic (N=3), native arnerican (N=2), other (N=4), african arnerican (N=8), and asian arnerican (N=2). The average tenure individuals had in the organization was 3.40 years and the amount of time with their current manager was less (Mean = 1.96 years). The following section will describe each of the organizations. annigations me and Two — Credit Unions Two of the organizations were credit unions, focused on service to families. The credit unions were approximately 50 miles apart and had different ownership. The professional emphasis of both credit unions was on the customer and on how they could provide the right services that would attract individuals. One of the organizations had 13 employees and the other had 18, and both had two levels in their hierarchy -— the CEO 47 (owner) was one level and the employees were the second. annigatipn Three - Automotive Supplier This company was part of a larger company that orginated in Germany. They work autonomously from their German counterpart, however, they were ultimately accountable to the German branch. In the Metropolitan Detroit complex there were two buildings and 115 employees, some of which (8) were employees of another organization and under contract with the auto supplier. There were 5 layers in the hierarchy of the company. Within the two buildings, parts were created and tested. Mass production of the parts took place in Ohio and Mexico. Therefore, most of the participants (N=52) in the survey were engineers and technicians, with some administrative staff and upper management. The supplier took great pride in all of the awards it had won from various US and Japanese (Toyota) auto manufacturers. They were also proud of the parts they have created for race cars. annization Four — Public Relations Qompany The individuals within this corporation (participants=16, total N of company=53) were also part of a larger conglomerate. The organization that owns them, however, was more diversified than the German company just discussed. They had ownership in advertising, PR, and many other types of organizations. The public relations branch was located in the Detroit Area and was connected physically to one of the advertising divisions. They considered themselves as having 7 layers because of all of the differences between positions, yet, many of them considered the manager of the PR firm to be their manager. Therefore, they did not have an official “chain of command.” Also, many of them had offices outside of their facility. They worked for a large automobile company and had an office in one of the public relations departments (organized by car 48 group) there. Now, having discussed each of the organizations, the following section will begin the analysis of the scales, followed by analysis of the messages and the likelihood of their use instrument. Then, the hypotheses and research questions will be addressed. Analysis of Scales Content Validity. Content validity is concerned with how well items represent the construct they are expected to represent. A way to test for content validity is to assess the unidimensionality of a scale. Confirrnatory Factor Analysis (CFA) was used to test the internal consistency and parallelism for each scale. Some items were deleted from some of the scales to create unidirnensional solutions. All of the scales, except for the one measuring ambition, demonstrated unidimensionality and parrellism. Table 2 illustrates the reliability estimates number of items for each scale: 49 Table 2. Scale Reliabilities and Number of Items Scales Reliability Items Assessments of Trust in a Superior: Consistency .80 3 Keeping Promises .88 5 Telling the Truth .90 4 Interest in Subordinate’s Welfare .91 6 Protection of Subordinate’s Interests .90 5 Would Not Hurt Subordinate .85 4 Communicate Accurately .93 5 Explains Decisions .85 4 Openness to Communication .86 4 All Trust Measures Combined .98 40 Impression Management Behaviors .90 5 Of Superior Assessment of Relationship: Leader-Member Exchange Scale .90 7 Assessments of Organizational Trust: Organizational Trust Scale .90 8 Subordinate Assessments of Him/herself: Ambition .67 4 Autonomy .78 5 Impression-Management Behaviors .78 3 Interdependence .89 6 A second order dimensionality was expected to occur for the items assessing a superior’s trustworthy behaviors. Four dimensions of the scale were expected, one from each of the dimensions of trust - behavioral consistency, behavioral integrity, communication, and demonstration of concern. Confirrnatory factor analysis demonstrated a second order unidimensional solution. The correlations within the 50 dimensions were higher than they were among the dimensions and all of the inter-item correlations were approximately .50 or higher. Yet, the predicted factors were not completely supported. Six dimensions, instead of four, were demonstrated. The dimensions are depicted in Table 3 and Table 4 contains the factor loadings. Table 3. Six Dimensions of Trust in a Superior Scale Reliability Communication: .93 Communicate Accurately Behavioral Integrity: .94 Consistent Conununication Honesty Keeps Promises Explains Decisions .85 Openness .86 Would not Hurt Subordinate .85 Demonstration of Concern: .95 Protection of Interests Interest in Welfare 51 Table 4. Factor madmgs for :1 rust Scale Be- Not Hurt Demon. Com- havioral Explains Sub- of Factors munication Integrity Decisions Openness ordinate Concern Communication 1 .00 Behavioral Integrity .83 1.00 Explains Decisions .82 .74 1.00 Openness .82 .88 .77 1.00 Not Hurt Subordinate .73 .84 .62 .84 1.00 . Demon. of Concern .80 .87 .79 .92 .81 1.00 Three of the original dimensions remained — communication, behavioral integrity, and demonstration of concem — but three new dimensions appeared to emerge — explanation of decisions, openness of communication, and not hurting the subordinate. These results will be further discussed in the final chapter. Construct Validig. Construct validity is concerned with the relationships between variables. It is demonstrated by the convergent and divergent validity of the construct with similar and dissimilar constructs. The purpose is to determine if constructs correlate with other constructs they are similar to and not with dissimilar constructs. In this research, the trust scale was tested for convergent and discriminate validity. Such a test was necessary because the scale had been created for the current research, it had no prior tests of reliability or validity. According to the multi-trait, multi- method matrix (Campbell & Fiske, 1959), there was an expected pattern of correlations among scales. Because the data was primarily gathered using the same method, the multi-method portion of the matrix cannot be used here to validate the Trust Scale. However, the expected pattern of correlations for the multi-trait matrix are in Figure 2. 52 (1) the reliability diagonal Highes Correlations (2) the same trait — same method (3) the convergent, but different trait - same method (4) the divergent different trait - same method Lowest Correlations F igpre 2. Expected correlations Figure 3 illustrates the multi-trait matrix for the current set of data. Trust in General LMX Ambition Superior Trust Trust in Superior .98 General Trust" .88 .92 LMX ._8_Q .7_l .90 Ambition .1 8 .19 .20 .67 General Trust is a scale created to test the Trust in Superior Scale and includes items like, “I trust my manager.” See Appendix A Reliability Diagonal = In Bold Same Trait/same method = In Italics Convergent — Different, but Similar, Trait/same method = Upderlipeg Divergent — Different, not Similar, Trait/same method = Normal Figege 3. Multi-Trait Matrix All of the correlations within the multi-trait matrix are as predicted, except for the reliability of the Ambition Scale. However, the correlations of the Ambition Scale with the other scales demonstrates that it acts as the divergent trait, its correlations with them are small and nearly identical. Thus, the Ambition Scale lends partial support for the construct validity of the Trust Scale. The Trust Scale seems to behave as predicted relative to measurements of other divergent and convergent traits. The construct validity, then, appears to be supported. To see all of the correlations between all of the variables, see Appendix C. 53 Analysis of the Likelihood of Use Instrumgit There were five likelihood of use or message selection scales. They were all tested with confirmatory factor analysis. The scale that assessed the likelihood of individuals selecting ambiguous messages had a reliability estimate of .83 and contained four items. The scale representing changing the subject messages had a reliability of .79 and again, four items. The complete lie, or distortion, scale contained four items and had a reliability of .87. For the truth selection scale, the reliability was .92 for five items and the withholding scale had a reliability of .76 and three items. Originally all of the scales contained five items but items detrimentally influencing internal consistency were deleted. Therefore, the internal consistency and reliability of the five scales was considered acceptable for further analysis. However, to ensure that message selection was not influenced by company membership, comparisons among the companies had to be assessed. Comparison of Major Variables among the Four Companies. AN OVA analyses were used to assess whether the variables that are antecedent to trust and relational quality demonstrated different effects among the four corporations. Unfortunately, because of the low number of individuals within three of the organizations (less than 20), there was only a slight chance that any of the relationships would be significant with these organizations. As would be expected, none of the relationships for these three organizations were found to be significant. However, company three (the automotive supplier company) had 53 participants and thus, the chance for significance among the variables was more likely. The variables that were found to significantly influence relational quality were: superior’s other focused impression management behaviors (SS = 1.53, df = 1, MS = 1.53, F = 11.90, p = 54 .001), interdependence (SS = 3.51, df= 1, MS = 1.53, F = 27.20, p = .0001), and organizational trust (SS = 1.25, df= 1, MS = 1.25, F = 9.72, p = .003). Organizational size and levels were constant and could not demonstrate an influence. The antecedents to trust also were found to demonstrate main effects on trust. The two which were significant were: superior’s other focused impression management behaviors (SS = .65, df = 1, MS = .65, F = 4.49, p = .04), and interdependence (SS = 1.27, df = 1, MS = 1.27, F = 8.78, p = .005). Again, organizational size and levels were constant and could not demonstrate an influence. The comparison among the corporations was flawed due to the low number of individuals within the analyses. To ensure that effects were not due to differences, the potential confound, company membership, was analyzed and will be discussed within the confound analysis section. Assuming that the organizations demonstrated similar effects among the major variables, the hypotheses were tested and answers for the research questions were sought. Teeiing I119 Hypotheses With the inclusion of a continuous measurement of message choice (i.e., likelihood of message use), it becomes possible to test the fit of the entire model. Thus, path and regression analyses were conducted on these data. Path analyses were used to test the entire model and regression was used to better understand some of the relationships of the antecedents on trust and the quality of the relationship. The following Table (Table 5) illustrates the distribution of the variables in terms of their means and standard deviations. 55 Table 5. Means and Standard Deviations for All Variables Std. Variables Num.Valid Missing Mean Deviation Age 92 7 35.71 10.90 Ambiguous Message 95 4 2.04 .75 Ambition 96 3 4.02 .54 Autonomy 97 2 4.06 .60 Avoidance Message 99 0 2.64 .76 Distortion Message 98 1 1.83 .80 General Trust 97 2 3.96 .69 Interdependence 99 0 3.70 .66 Lie Acceptability 96 3 4.02 .68 Organizational Trust 94 5 3.54 .60 Relational Quality 97 2 3.84 .66 Subordinate IM 95 4 2.26 .49 Superior IM 94 5 2.35 .54 Trust 87 12 3.81 .58 Truthful Messages 97 2 2.48 1 .12 Withholding Messages 98 1 2.67 .85 The correlations among the variables were entered into the path program created by Hunter and Hamilton (1995) (see Appendix C for all correlations). The program tests for a goodness of fit index for the model and calculates the significance of the goodness of fit. To test for the moderators, the sample needed be separated into two levels for each moderator (high and low) and then tested at each level. Supplemental regression analyses of the effect of the antecedent variables on trust and the quality of the relationship were also completed. Simultaneous regression of all antecedents was used due to the lack of theoretical justification for ordering the variables and using hierarchical regression. To determine the effects of trust and the quality of the relationship, each of the likely messages was regressed onto the two variables. Discussion of the findings will begin with trust as the center of the model and then will i talk about the quality of the relationship as the center of the model. 56 Trust. The proposed model and path coefficients were as follows (see Figure 4, to see the model at high and low levels of autonomy and interdependence see Appendix E): Autonomy“ Low High V Organizational Trust '25 .47 Organizational Size'23\.31> -.ll -.30 Trust Organizational Complexity a Interdependence“ Low High V Ambiguous Superior Focused Impression Management .44 . 12 Ambition Subordinate Focused Impression Management % Withholding “24 Distortion Change Subject *Where there is only one path coefficient, there is no moderating effect. Em Path Coefficients for Model with Trust as Central Variable The model without the moderator variables met the goodness of fit requirements and was significant (xz=114.55, df=66, p=.001). However, when the model was tested at high and low levels of autonomy and interdependence, it did not meet the requirements and was not significant. Such findings can be explained in relation to the issue of power. Although the entire model fits, when the sample is divided to look at the moderators, power is lost and finding significance becomes less likely. Table 6 includes the chi 57 square and significance levels for each moderator test of the model. Table 6. Teste of the fifrust Model in Path Model x2 p Low Autonomy 65.41 .50 High Autonomy 71.03 .31 Low Interdependence 69.66 .36 High Interdependence 64.43 .53 For the regression analysis, the antecedent variables regressed onto trust produced an R2 of .63 and some changes occurred when the moderator variables were assessed (see Table 7). Table 7. Total ngance Explained by Models When Looking at Moderator Variables Moderator R2 Low Autonomy .55 High Autonomy .63 Low Interdependence .57 High Interdependence .68 More of the variance is explained by the higher levels of the moderators. This may indicate that the higher levels of the moderators have a stronger influence on the 58 relationships they effect than the lower levels of the moderators. The following discussion of hypotheses and research questions are based on the above models. & The hypothesis that organizational trust was positively related to a subordinate’s trust in a superior was supported. The above model demonstrates that organizational trust and a subordinate’s trust are positively related but that the relationship is moderated by the amount of autonomy experienced by the subordinate. At higher levels of autonomy, organizational trust had a stronger positive influence on the subordinate’s trust in his/her organization. Regression analysis further demonstrated that organizational trust demonstrates a significant relationship ([3=.25, p=.007) with trust in a superior. Therefore, the hypothesis was supported. Q The hypothesis that the size of an organization would be negatively related to the trust in a superior was not supported. Path analysis demonstrated the relationship to be in the opposite direction — it was positive. Further, regression analysis demonstrated that the relationship was not significant (B=.O9, p=.3 8). & The complexity of the organization was also believed to be negatively related to the trust in a superior. Regression analysis demonstrated support for the effect of complexity on trust (B=-.l9 p=.04). Path analysis further demonstrated that the relationship between organizational complexity and trust is positively moderated by autonomy such that at high levels of autonomy, more complexity has a more detrimental effect on trust in a superior. H_4_: A subordinate’s other-focused impression management behaviors toward a superior were believed to demonstrate a positive influence on trust in a superior. Again, regression analysis demonstrated that the relationship was insignificant (B=.06, p=.5 3). 59 The path coefficients also illustrated a weak relationship. Thus, the hypothesis was not supported with these data. H_5:_ A superior’s other-focused impression management behaviors was postulated to have a positive effect on trust in that superior. Such an effect was supported by the regression analysis (13:34, p=.001). Thus, if subordinates perceived their superior as using impression management behaviors, the superior would be trusted more by subordinates. A moderator effect of autonomy was not found in either path or regression. & Ascendancy desires, or ambition, was believed to positively influence impression management behaviors of a subordinate. The low path coefficient representing the relationship (.12) refuted the hypothesis. Regression analysis of ascendancy desires on subordinate impression management behaviors also demonstrated no significant influence (B=.09, p=.42). However, the exact role of ambition receives firrther exploration in the post hoc analysis section. Lil This hypothesis claims that high levels of trust would be associated with increases in truth-telling and decreases in trust would result in more deceptive message usage. The following table (Table 8) illustrates the standardized beta weights for each of the likelihood of use messages on trust. 60 Table 8. Regression Analysis of Trust on Message Selection Message B p R2 Ambiguous -.32 .003 .10 Change Subject -.21 .05 .04 Distortion -.23 .03 .05 Truth . 1 7 . 1 1 .02 Withholding -.05 .67 .002 The influence of trust on the likeliness of use of messages was only significant on the ambiguous, changing the subject, and distortion messages, not on the withholding and truthfirl messages. Because truthful messages were unrelated, the hypothesis failed. _H_8; This was the hypothesis that predicted that low levels of trust would result in a greater use of distortions than any other message. Due to the finding that ambiguous messages were more negatively related to trust than distortions, the hypothesis was not completely supported. However, partial support for the hypothesis could be demonstrated. At low levels of interdependence, the influence of trust on distortion was increased to slightly above the influence of trust on ambiguous message selection. This indicated that distortion messages would be one of the most likely messages used when trust and interdependence were at low levels. ii The question of whether autonomy moderated the relationship between the antecedent variables and the level of trust in a superior was demonstrated on two of the antecedent variables that demonstrated a significant effect on trust — organizational trust 61 and organizational complexity. Its moderating effects were positive: so that at higher levels of autonomy, the relationship between the two variables and trust in a superior was increased. Q The question of whether interdependence moderated the relationship between the level of trust and message selection did gain some support. Of the three messages significantly influenced by trust, only one’s relationship with trust was moderated by interdependence. The original relationship between trust and distortion messages was negative. When interdependence was added to the relationship, the negative relationship between trust and distortion messages was increased. Thus, interdependence exacerbated the relationship between trust and distortion messages. The following section will discuss the model with the quality of the relationship as the central variable. Quality of the Relationship. The proposed model and path coefficients are as follows (Figure 5, see Appendix F for high and low levels of autonomy and interdependence): 62 Autonomy“ Interdependence“ Low High Low High Organizational Trust\,1{A V . 3 6 Truth .42 _ -_ Amb' Organizational Siz '1‘ .27 , * 15 a lguous 6\ Quality of 15 the ' a . . Relationship 1 l WltthIdlng -.O6 - .20 '- Organizational -.28 Complexity Distortion . 1 5 Superior Focused Impression -.35 Management Change Tn Subject Ambition Subordinate Focused Impression Management *Where there is only one path coefficient, there is no moderating effect. F igge 5. Path Coefficients for Model with Relational Quality as Central Variable The model without the moderator variables met the goodness of fit requirements and was significant (x2=127.98, df=66, p=.0001). However, when the model was tested at high and low levels of autonomy and interdependence, it did not meet the requirements for goodness of fit. Table 9 includes the chi square and significance levels for each moderator test of the model. 63 Table 9. Test th uali of the Relationshi Modeli Path Model x p Low Autonomy 62.38 .60 High Autonomy 85.23 .06 Low Interdependence 65.95 .48 High Interdependence 71.39 .30 For the regression analysis, the R2 for all of the antecedent variables on trust was .68. Table 10 illustrates the variance explained for each of the moderator analyses. Table 10. Tegl Variance Explained by Models When Looking at Moderaior Variablee Moderator R2 Low Autonomy .77 High Autonomy .56 Low Interdependence .76 High Interdependence .80 All but one (high autonomy) of the regressions with the split moderators demonstrated the ability to explain more of the variance than the original model. The following responses to the hypotheses and research questions were based on the coefficients in the model and the regression analysis of all of the antecedents. 64 11% The hypothesis that organizational trust would be positively related to the relationship with a superior was supported by regression analysis (B=.24, p=.005). Further, path analysis demonstrated that the relationship was moderated by the level of autonomy. At higher levels of autonomy, organizational trust had a more positive influence on the quality of the relationship. ifl; The size of the organization was believed to be negatively related to the quality of the relationship. The relationship was insignificant (B=.11, p=.25) and therefore, the hypothesis was not supported. H_1_1; The complexity of the organization was also believed to be negatively related to the quality of the relationship. Although a negative relationship was found, it was not significant (B=-. 13, p=.13), demonstrating the hypothesis to be false. fl1_2_:_ A subordinate’s other-focused impression management behaviors were believed to be positively related to the quality of the relationship. The relationship was not significant (B=.06, p=.47) and thus, the hypothesis was not supported. fill; A superior’s other-focused impression management behaviors was believed to influence the quality of the relationship. The hypothesis was supported in that the relationship between the two variables was significant (B=.33, p=.0001). Also, the path analysis demonstrated no moderating effect from high or low levels of autonomy. _l$ Ascendancy desires, or ambition, were believed to demonstrate a positive effect on a subordinate’s other-focused impression management behaviors. The path correlation again was low between the two variables (.12), thus again refuting the relationship between the two variables. m A high quality relationship was believed to positively influence truthful message selection and decrease deceptive message selection. The following table (Table 65 11) illustrates the standardized beta weights for each of the likelihood of use messages on the quality of the relationship. Table 11. Regieeeiop Apalysie of the Qualiiy of the Relationship on Message Selection Message I3 P R2 Ambiguous -.13 .20 .02 Change Subject -.06 .59 -.01 Distortion -. 17 .1 1 .02 Truth . 19 .06 .04 Withholding .12 .25 .01 None of the relationships between the quality of the relationship and the selection of messages were significant. Therefore the hypothesis was not supported. HE The hypothesis that distortions would be used more than other deceptive messages in low quality relationships was not supported due to lack of significance. fl; Research question one explored the role of autonomy in moderating the relationship between the antecedent variables and the quality of the relationship with a superior. Of the two variables found to have a significant influence on the relationship, organizational trust was moderated by the level of autonomy. At high levels of autonomy, organizational trust demonstrated a stronger positive relationship with relational quality. RZ_:_ The question of the role of interdependence on the relationship between the quality of the relationship and message selection was found to be irrelevant due to the 66 fact that the quality of the relationship showed no significant effect on message selection. Although all of the results indicated no influence from the quality of the relationship, yet, there were still questions regarding the one variable that should have some influence, ambition. To better understand the effects of ambition, post hoc analyses were conducted. Post Hec Anaylses Several analyses were conducted on ambition to determine its influence. First, regression and path analysis were employed to determine the influence of ambition on the selection of messages. Figures 6 illustrates the path coefficients that were found. -.14 Truth _'01 Ambiguous Ambition _ 08 Withholding 1 Distortion Change Subject /2 V Figme 6. Ambition as Antecedent to Message Selection The coefficients were not very high for any of the relationships between ambition and message selection. The model, however, did meet the goodness of fit requirements 08:51.09, df=10, p=.0001). Regression analysis for each of the messages on ambition, or ascendancy desires, was also completed. Table 12 contains the results of the analysis. Table 12. Reggession Analysis of Ambition’s Relationehip with Message Selection Message I3 P Ambiguous -.01 .91 Change Subject .10 .35 Distortion -.001 1 .00 Truth -. 14 . l 8 Withholding -.08 .45 Ambition demonstrated very little effect on message selection. Yet, there may be confounds that could influence message selection. &elysis 9f. Eotentiel Confounds Four potential confounds were analyzed in the current research — lie acceptability, restructuring of the organization, years with a superior, and managerial status. One additional potential confound —the company an individual works in - emerged during the research and was examined as well. The analysis of the potential confounds began by looking at their correlations with all of the relevant variables. If uncorrelated with any of the variables, the confound was considered unimportant. If correlated with any of the variables, the confound was firrther tested. Three organizational variables and lie acceptability were examples of variables 68 that were correlated with significant variables and thus, received “further testing. Years with a superior, management position, the company variable, and lie acceptability were all added into a regression equation to test for their potential effects. To test these three potential confounds, first, the company variable was added to the original set of antecedent variables, then years with a superior was added, followed by whether an individual was a manager in the company. Trust was the dependent variable in the regression equations used for the first set of analyses. When the company variable was added to the regression on trust, the R2 increased from .63 to .64. Overall, the company variable did not demonstrate a significant effect (B=-.47, p=.10). Then, when the years with a superior was added to the new regression equation, the R2 was unchanged (.64) and years with a superior did not demonstrate much effect on trust (B=-.02, p=.80). An individual’s status with the company (whether .he/she was a manager or not) also did not effect the R2 (.64) and did not have a significant influence on trust (B=.006, p=.94). Finally, when lie acceptability was added to the regression equation, the R2 was again virtually unchanged (.66) and the effect on trust was insignificant (B=.12, p=.14). The second set of regression analyses were utilizing the quality of the relationship. When the company variable was added to the regression on the relationship, the R2 went from .68 to .69 and yet again, the company variable did not demonstrate significance (B=-.l9, p=.48). Then, years with a superior was added to the equation, increasing the R2 to .72. Also, the years with a superior demonstrated a significant effect on the quality of the relationship with a superior (B=.20, p=.004). Based on these analyses, there appears to be a relationship between the years an individual has been with a manager and the quality of the relationship with that manager. Whether an individual was a manager or not, however, did not seem to influence the quality of the relationship 69 with a direct superior. The R2 when an individual’s status was added to the equation was virtually unchanged (.73) and the relationship proved to be insignificant (B=.06, p=.35). The effect when lie acceptability was added to the equation also demonstrated no effect on the R2 (.73) and the effect of lie acceptability was insignificant (B=.04, p=.61). The effect of lie acceptability, however, was still unclear. Therefore, correlations between lie acceptability and the selected messages were calculated. Two of the messages were related with lie acceptability — ambiguous messages (r=-.27, p=.05) and distortion messages (r=-.20, p=.05). Although significant, only two of the selected messages demonstrated any important relationship with lie acceptability. The potential confounding effect of lie acceptability was thus believed to not be systematic. The final potential confound was company restructuring. The variable was uncorrelated with all of the relevant variables. It was thus dismissed due to having no potential influence on the results. Analysis of the Manipulation Check Two questions asked respondents if they had a difficult time placing themselves in the situation. The mean for both questions was slightly above the mid-point of three (3.13 and 3.27). The respondents, then had a slightly difficult time placing themselves in the situation. The primary researcher talked with several of the respondents when they were having trouble with the hypothetical situation. Often it was the generality of the situation that posed a problem. Individuals wanted more information about the co- worker they could potentially tell on. They also wanted more information about the relationship with the superior and the importance of the supplies to the company. Such suggestions may indicate that individuals may require more information regarding a situation. Also, many individuals had a difficult time creating a message, yet, could 70 easily pick messages out when presented with them. So, when they responded to the plausibility of the situation, they may have been responding with their frustration over the difficulty of creating a message. The implications of these results will be the subject of the following chapter. 71 Chapter 5 DISCUSSION AND‘IMPLICATIONS OF FINDINGS There are four main contributions that this study makes to the literature on organizational deception. First, it demonstrates that the quality of the relationship does not influence message selection as well as does trust in a superior. Second, the significant effects of trust in a superior support its importance in the study of organizational relations. The third contribution is the initial development of a valid measure of trust. The fourth contribution is the discovery of significant antecedent variables for both relational quality and trust in a superior. Each of the contributions will be discussed within this section. This will be followed by a discussion of the moderator variables, the emergent role of ambition, and the role of theory. Finally, the major limitations of the research will be explored and implications for future research examined. Contribpiiene ef the Current Research First, it had been initially proposed that trust could be one element of the quality of the relationship and that relational quality may have a more pronounced effect on message selection than would trust alone. Such is not the case. Trust appears to have some influence on message selection and the quality of the relationship demonstrates no significant effect. This is an important contribution because organizations often focus on the quality of the relationship and not on the specific underlying dimensions. If the quality of the information communicated is crucial for an organization, the current research indicates that it needs to address a specific element of a relationship and not try 72 to influence all of the dimensions. The specific element, and second main contribution, indicated in the current research is trust in a superior. The potential influence of trust in a superior makes it important to the study of organizations (see Jones and George, 1998). Whitener et al. (1998) argue that there is a resurgence of interest in trust within organizations — primarily from an interpersonal level. Trust can influence the “quality of communication, performance, citizenship behavior, problem solving, and cooperation” (Whittier et al., 1998, p. 513). It could be argued that the current findings demonstrate support for the effect of trust on the quality of communication, citizenship, and cooperation. Deceptive behavior is not usually considered to be cooperative or to demonstrate good citizenship within a corporation. It also represents low quality communication. If trust interacts with the level of interdependence between individuals to influence increases in their deceptive behavior, it is demonstrating a significant effect on these important organizational variables. Thus, showing support for the potentially significant role of trust in organizational research. The measurement of trust is the third contribution of the current research. Although an important construct, it has not been well or consistently measured. The current measurement contains three of the dimensions (communication, behavioral integrity, and demonstration of concern) proposed by Whitener et a1. (1998) and three additional dimensions (explains decisions, openness, and not hurting the subordinate). Tests of internal consistency and validity demonstrate that the measurement too] well represents the construct of trust in a superior. They also support the predicted second order unidimensional solution for the measurement tool. Therefore, although the dimensions require further exploration with multiple methods, this initial test 73 demonstrates that the measure is a solid and much needed tool for assessing trust. The fourth contribution of the current research is the discovery of significant antecedents to trust and relational quality. Although the quality of the relationship and trust demonstrated different effects on the messages, they are both influenced by two of the same antecedent variables — organizational trust and a superior’s use of impression management behaviors. Organizational trust is positively related to both the quality of the relationship and the level of trust in a superior. Further, it is moderated by the level of autonomy a subordinate experiences. At high levels of autonomy, organizational trust becomes more positively related to trust in a superior and the relationship with the superior. At low levels of autonomy, the positive relationship will decrease. It appears that when subordinates have a lot of autonomy, their feelings of trust in their superior and the relationship with the superior are strongly effected by the trust they have in their organization. Thus, if organizations are moving to more autonomous work environments they may be wish to consider ways to increase employee trust from an organizational level. This could mean an increase in the development of policies that demonstrate concern for the employee and increased communication. Such changes may prompt individuals to be more honest and straightforward in their communication with immediate supervisors. In terms of the influence of a superior’s other—focused impression management behaviors, they are highly correlated with trust and the quality of the relationship and are not moderated by autonomy. Such a finding makes sense because other-focused behaviors have been found to be associated with increases in liking. It seems likely that when impression management behaviors are used, the quality of the relationship should increase. For the link with trust, trust is measured in part by behaviors that are beneficial 74 to the subordinate’s welfare. The impression management behaviors are ones that could make the subordinate feel that the superior is concerned for his/her welfare, leading to an increase in trust. A third antecedent had been found to influence trust but not relational quality - organizational complexity. As organizations increase in size, subordinates demonstrate less trust in immediate superiors. Also, the relationship is moderated by autonomy such that more autonomy leads to a stronger negative relationship between organizational complexity and trust in a superior. This may be because with increasing autonomy and complexity, individuals have less interaction with their superiors and as discussed in Chapter 2, individuals tend to gravitate toward individuals who they interact with more often. Also, if there is more interaction, there may be more opportunity for superiors to enact behaviors that will elicit trust from subordinates. Other antecedent variables did not demonstrate the effects hypothesized, yet, the failure of the other antecedents provides useful information as well. First, organizational size failed to demonstrate a significant influence on trust and relational quality. Yet, there is organizational size only reached 115 people in the largest company. As discussed previously, this may be considered to be a small company, not a large company, therefore there was not a large enough company for comparison among the organizations used here. The impression management behaviors of a subordinate also did not demonstrate any effect on the two major variables — trust and quality of the relationship. The path coefficients between subordinate impression management behaviors and trust and the quality of the relationship are among the lowest. It is then hard to believe that the variable could be improved so that it would make a significant contribution. Simply put, 75 the behaviors of the subordinate does not increase trust or relational quality — in the mind of the subordinate. Because of the reciprocal nature of the trust and the quality of the relationship, future research could include these behaviors, but measure them from the perspective of the superior. It may be that they influence the superior’s behavior which, in turn, influences the level of trust in that superior and the quality of the relationship with him/her. This leads to the next topic for discussion, the moderators of the relationships among variables. Moderator Variables Autonomy was believed to influence the relationships between the antecedents and organizational trust and the quality of the relationship. A significant moderating effect was found only for the relationships between organizational trust and organizational complexity and trust and relational quality. But, such findings are important because they demonstrate that giving individuals more freedom to interpret their work makes their trust in the organization more important. Also, with more autonomy, the negative influence of organizational complexity on trust in a superior becomes more pronounced. Such variables deserve future research because they may have important implications to organizations. With the increase of individuals working in their homes or outside of the traditional office environment, such an issue may become more relevant to organizations. If individuals do not have strong trust in their organization, their trust in their superior may decrease. If they exist in a complex organization and have high autonomy, their trust in their superior will further decrease. Thus, future research should continue to explore the effect of autonomy on organizational trust and organizational complexity’s relationship with trust in a superior. 76 Interdependence also was found to be an important moderator variable. It demonstrates a negative effect on the selection of the most deceptive message (distortions). At low levels of interdependence, trust is more negatively associated with the selection of distortion messages. Thus, their use increases when trust and interdependence are low. When trust and interdependence are at high levels, it becomes less likely that a distortion will be selected. Such findings indicate that if organizations work to increase the amount of trust and interdependence between superiors and subordinates, complete distortions will be chosen less by subordinates talking to their superiors. Future research can seek to better explain the relationships because by understanding how interdependence influences the relationship between trust and deception, ways to increase interdependency’s influence can be found. Post Hoc Analysis of Ascendancy Desire, er Ambition Further exploration of one of the other variables also was performed in the current research. Ascendancy desire, or ambition, had been believed to be antecedent to a subordinate’s other-focused impression management behaviors. Such is not the case in the current research. So, ambition’s influence on message selection received post hoc analyses. The findings demonstrate that ambition has little influence on the selection of messages. Although it is still unclear why such effects were found, the findings actually may help clear up some of the debate on the influence of ambition. Athanassiades (1973, 1974) and Read (1962) argue that ascendency is a strong predictor of the distortion of information, while Gaines (1980), Roberts and O’Reilly (1974a), and O’Reilly (1978) believe it to have little effect. The more current research coupled with findings here and by Hubbell — ambition not having a strong influence on message creation (1999b) - illustrate that perhaps ambition does not influence the distortion of messages to superior 77 and firrther exploration of its effects may not be necessary. Theoreiical Explanations Although several contributions of future research have been explored, there is one large implication that still needs to be addressed —- the link to theory. A very important element of the current research is the application of theory. Theory is noticeably missing in deception research in both the interpersonal and organizational deception literature. The application of two theories — Social Exchange Theory and Agency Theory - helps to explain the links in the current model as well as to provide empirical support for further exploration of these theoretical frameworks in organizational deception research. According to Whitener et al. (1998), Social Exchange Theory explains the reciprocal nature of trust in that if a superior enacts trustworthy behaviors, a subordinate will respond with trustworthy behaviors. Although truthful messages are not significantly linked to trust, when trust increased, ambiguous and avoidant messages decreased. Also, when interdependence is low and trust in a superior is high, individuals are less likely to choose complete lies, or distortions. Such findings appear to support the use of Social Exchange Theory for explaining the effect of trust on message selection. When individuals trust their managers they tend to decrease their preference for deceptive messages, thereby reciprocating with trustworthy behaviors. Social Exchange Theory seems to well explain the influence of trust on message selection, Agency Theory helps explain the antecedents to trust. Agency Theory is based on the premise that individuals work to serve their own purposes — to maximize their gain and minimize their loss — in every interaction (Whitener et al., 1998). The strong relationship between a the impression management behaviors of a superior and trust in the superior demonstrate that subordinates like receiving flattery and their manager doing 78 favors for them. Such behaviors benefit the subordinate and maximize their gain in the relationship. They respond by increasing their trust in the giver of the flattery and favors — the superior. The strong role of autonomy in the influence of organizational trust and organizational complexity demonstrate that subordinates are strongly influenced by their ability to make choices about their work. Those with high autonomy, or those with a choice in how they do their work, may be considered to have already received a great benefit. They have their independence. As discussed earlier, individuals who are more autonomous may be so because others have trusted them to be successful on their own. They reciprocate by trusting in the organization and their superiors more. In terms of complexity, autonomy also influences its effect on trust in a superior. When individuals are more autonomous, organizational complexity demonstrates an increasingly negative relationship with trust in a superior. This may be because individuals who are highly autonomous benefit from membership in smaller organizations. They may have more freedom and less supervision when there are less levels and managers. For example, individuals within the credit unions (the companies with two levels), individuals considered themselves to be highly interdependent and they had high levels of trust in their superiors. Yet, the companies were very small (less than 20 people). Therefore, there may be more opportunity for independence (a perceived benefit) in an organization with fewer levels. With increased independence and decreased size, trust in a superior will still be negatively related to complexity, but much less so. The importance of theory cannot be overlooked, however, there are some limitations that cause us to interpret the findings with caution. Limitations of the research will be the subject of the following section. 79 Limiietions ef the Current Research The use of survey data in organizational research is controversial. Self-report measures like this have often “been used too frequently to address research questions that they are unable to answer” (Spector, 1994, p. 385). However, there has been support for ' the use of self-reports when explaining perceptions of individuals (Schmitt, 1994; Spector, 1994), creating other hypotheses that can be further tested in the field (Spector, 1994), and delving into new areas of research (Schmitt, 1994). Elaborating on the last point, according to Schmitt (1994), it is appropriate to use self-report data in new area of research simply to determine “what is worthwhile bothering with” (p. 395). Therefore, it is reasonable to use surveys as the primary source of data in these particular situations. However, the results must be interpreted with care. Support for the hypotheses indicate areas of future study, they do not represent complete acceptance of the relationships. To demonstrate if the relationships are valid, they must be researched further within the field with other data sources. A second limitation to this study concerns the specific hypothetical situations used in the survey. Individuals knew little about the co-worker they were to talk to the manager about. Several of the individuals asked for more information about the co- worker. They wanted to know about the relationship with that person and if the person had stolen from the company before. Without knoWing more about the person, it appears that some individuals find it difficult to assess whether they would turn the co-worker in or not. To address this in future research, individuals could be given more details regarding the situation and the co-worker. This leads to a third limitation of the hypothetical situation — context. It is very likely that context can alter how we chose to manipulate information. Tests of similar 80 messages in interpersonal research have demonstrated that what is more appropriate in one situation, is considered inapprOpriate in another (see Hubbell, 1999a; McComack et al., 1992). Thus, more testing using different hypothetical situations may alter the use of certain message types. Such testing is an example of an implication for future reseach. One final limitation that can be addressed by further research is the sample size. Although 99 individuals participated, when testing moderators in path analysis, the sample had to be split. This resulted in difficulty in finding significant effects. Future research using significantly more participants (200 or more) could better test the moderators and demonstrate significance. F ridge Direetions The current research demonstrates the importance of theory in organizational deception research in terms of explaining effects. It also illustrates the variables that could and should be tested using different situations and different methods. The following section will describe how future research can build on and address these issues. The first suggestion for future research is to expand the theoretical underpinnings of the current research. Future research in organizational deception could attempt to link other theories to the current model to determine if they better explain the effects of trust. For example, Negotiated Order Theory (Donohue & Roberto, 1993) posits that the type of a relationship which exists between individuals will influence whether they are able to resolve difficult situations (such as hostage negotiation). Perhaps if subordinates and superiors have certain types of relationships they may be more or less likely resolve a difficult situation through open and honest communication. They may not feel as compelled to lie. Such issues could be answered with future research. A second direction for future research is to more closely explore the link between 81 trust and the veracity of communication. Because trust is believed to be constructed over time, longitudinal analyses would also be appropriate. Such research could examine how trust changes over time and if such changes alter the veracity of communication from subordinates to superiors. Other future directions with testing of trust and the honesty, or dishonesty, of messages, could explore the role of trust in a subordinate and its influence on how honest a superior is to a subordinate. Superior and co-worker honesty and its relationship to trust could also be examined because if communication is reciprocal, those who communicate more truthful information should receive more truthful information in return. Third and finally, another direction for future research resulting from this work is to continue to validate and refine the measure of trust used in the study. The most logical next step for testing the validity of the measurement tool is to test it using multiple methods. Focus group research and structured interviews could be conducted. Participants could be asked about the dimensions found here to determine if they represent the construct of trust in their minds. Longitudinal analysis also could help test the trust measurement. If it alters over time and does so in relation to similar variables, it would be acting as predicted. Further, common method variance could be avoided by testing different parts of the scale in separate surveys. If the surveys are given to an individual within a couple days of each other and the items on the two surveys do not correlate, they may not be representing the same variable. Instead, the high correlations found here could be attributed to common method variance instead of to a highly reliable second order unidimensional scale. By better understanding the conditions leading to truthful and deceitful communication, organizations will be better able to understand how to confront and 82 reduce situations that result in the communication of distorted information. Also, if organizations are shown the importance of variables such as trust, perhaps the way they relate to their associates will change. With a change focused on increasing trust, perhaps other beneficial effects will occur, other than decreasing deceptive communication or behavior. So, understanding why subordinates tell the truth or lie is a relevant and important task for research in organizational deception. 83 APPENDICES 84 APPENDIX A INSTRUMENTS Scales for Trust in a Superior: Behavioral Consietency: My manager’s behavior is very predictable My manager’s behavior is consistent I usually know what my manager will do next My manager behaves reliably Behavieral Integp'g (keeping pgomises and telling ibe gm): Promises: I think that my manager does what he/she says they will do I think my manager is fair My superior keeps his/her word My manager keeps his/her promises I can rely on my supervisor to keep his/her promises Honesty: I trust my manager to be honest with me My manager is sincere in his/her dealings with me I believe my superior is straight-forward I believe my manager tells the truth Dmenetration ef Concegp: Genuine interest in the welfare of the subordinate: My manager is genuinely interested in my welfare I know my manager cares about me My manager makes me feel that things I tell him/her are really important My manager is interested in helping me in any way possible I think my manager tries to help me when he/she can My manager listens to me Protection of the subordinate’s interests: My superior will always make efforts to protect my interests My manager works to firrther my interests I believe my manager is interested in me doing well I believe my manager does things that will further my interests I know that my manager will work to protect me Not Hurt the Subordinate: I feel free to discuss with my manager any difficulties I may have in my job, without it 85 being held against me later I know that my manager will not tell others information that could be harmful to me My manager does not claim my ideas as his/her own My manager would not sacrifice me to benefit his/herself My manager would “stab me in the back” I think my manager would communicate harmful information about me if he/she would benefit from it ommunication: Accuracy of information: I believe my superior is very precise in his/her communication with me I believe my manager communicates accurately with me Usually, my superior clearly communicates what he/she wants I usually understand what my manager wants My superior’s communication is usually accurate Explanation of decisions: My superior always explains his/her reasons for making decisions I am usually told why things are done the way they are in my department I usually understand why things are done the way they are in my department My superior talks to me about why some decisions are made Openness: I can tell my manager anything I can freely discuss job-related ideas and feelings with my manager My manager listens to me when I tell him/her about things that are bothering me It is “safe” to say what I am really thinking to my manager Additional Ceneral Trust Measure: I trust my manager I have confidence in my manager I believe in my manager I believe my manager has integrity Quality of the Relationship with a Superior Scale: Leader-Member Exchange Scale: Do you usually feel that you know where you stand, that is, do you usually know how satisfied your immediate supervisor is with what you do? How well do you feel that your immediate supervisor understands your problems and needs? How well do you feel that your immediate supervisor recognizes your potential? Regardless of how much formal authority your immediate supervisor has built into his or her position, what are the chances that he or she would be personally inclined to use power to help you solve problems in your work? 86 Again, regardless of the amount of formal authority your immediate supervisor has, to what extent can you count on him or her to “bail you out” at his or her expense when you really need it? I have enough confidence in my immediate supervisor that I would defend and justify his or her decision if he or she were not present to do so. How would you characterize your working relationship with your supervisor? Antecedents to Trust in a Superior and Relational Quality Scales: annizational Trust: People in this organization will do things behind your back When top management here says something you can really believe it is true Actions taken by upper management are in my best interest In my organization, upper management follows through on commitments Upper management is fair with employees Upper management makes good decisions I feel quite confident that the firm will always try to treat me fairly Upper management at my firm is sincere in its attempts to meet the workers’ point of v1ew Upper management can be trusted to make sensible decisions for the firrn’s future _A_ipbition: I would like to be promoted I consider myself to be ambitious I try to perform better than my co-workers I try very hard to improve on my past performance at work I try to avoid any added responsibilities on my work Impression-Management Behaviors of Subordinate Toward Superior: How often do you compliment your supervisor on the quality of his/her work? How often do you do favors for your manager? How often do you compliment your supervisor? How often do you do something nice for your supervisor without being asked to? How often do you go out of your way to say something nice to your manager? Impression Management Behaviors ef Superior Toward Subordinate: How often does your manager compliment you? How often does your supervisor compliment you on the quality of your work? How often does your manager do favors for you? How often does your supervisor do something nice for you without being asked to? How often does your manager go out of his/her way to say something nice to you? 87 Moderator Variable Scales: Interdependence: My manager and I often share resources in order to accomplish tasks My manager and I depend on each other I believe that my manager and I influence each other when making decisions I think that my manager and I cooperate with each other My superior and I often pool resources on projects When accomplishing a task, I believe my superior’s opinions influence me and my opinions influence him/her Autonomy: I have a lot of autonomy in my job, that is, my job permits me to decide on my own how to go about the work The job denies me any chance to use my personal judgement in carrying out the work The job gives me considerable opportunity for independence in how I do the work I have independence from other people’s control I am able to do my job without constant supervision 88 APPENDIX B LH~ = l- 3 9 ° 3 2 ~2- 2 a5 a .‘3 .0 .o C m I: .0 _ u f8 .. 5 .5 ._ on .99 .2 8 3 E t! a 0 'a .g - - . eiea'ei,aa ::~2§S§§-—i Variables < z < < < 2 u 5 o .5 .3 :3 g i :2 (I) t- 1- 3 2 Ambiguous Messages 83 Ambition -.01 .67 Autonomy -.11 .16 .78 Avoidance Messages .47 .10 -.10 .79 Company .18 .29 -.06 .00 1.00 Distortion Messages .52 .00 - 13 .14 .15 .87 General Trust -.26 .19 .46 -.23 -.24 -. 16 .98 Interdependence -.24 .46 .46 -.20 -.14 -. I4 .54 .89 Levels .14 2| -.01 -.02 .96 .12 -.l7 -.14 1.00 Lie Acceptability -.22 .10 .21 .01 -.25 -.20 .35 .28 -.23 .79 Organizational anst -.13 . 18 .35 .05 -.18 -. 18 .47 .25 -.11 .29 .90 Relational Quality -. l 3 .20 .54 ~06 -.23 -. l 7 .71 .66 -. l 7 .34 .48 .90 Subordinate 1m -.13 .09 .26 -.07 -.08 -. l 7 .37 .40 .00 .20 .25 .45 .78 SuperiorIM -.14 .11 .19 -.02 -.18 -.01 .52 .42 -.12 .24 .29 .61 .53 .90 Tmst -.32 .18 .47 -.21 ~32 -.23 .89 .61 -.25 .44 .50 .80 .43 .59 .98 Truthful Messages -.27 - 14 .14 -.17 -.28 -.33 .20 .20 - 19 .07 .28 .19 .23 .09 .17 .92 Withholding Messages .19 -.08 -.05 .38 -.15 -.10 -.05 -.04 -.10 -.06 .10 .20 .18 .10 -.05 .37 .76 91 APPENDIX D HYPOTHETICAL SITUATION One aftemoon when coming into the office after lunch, you ran into a co-worker putting some office supplies into his/her car. You asked the co-worker what he/she was doing and the co-worker said, “I really need this stuff and I plan on replacing it after payday.” You are upset by what the co-worker had done and doubt that he/she will buy new materials. ' Later that day, your manager comes up to you and says: “I’ve noticed that office supplies seem to be disappearing at an alarming rate. I am pretty concerned about this. Do you know what could be happening to them?” 92 APPENDIX E HIGH AND LOW LEVELS OF AUTONOMY AND INTERDEPENDENCE AND THEIR INFLUENCE ON TRUST AND MESSAGE SELECTION High Autonomy and Interdependence High Autonomy High Interdependence Organizational Trust .47 Truth .40 .31 Organizational Size \ Trust -. l 7 -.30 . . Organizational , —> Distortion Complexity Low levels of Autonomy and Interdependence Low Autonomy Low Interdependence O ' t' IT t rgamza lona rus Truth .25 .01 Trust Organizational Size \23, -.37 . . -.11 -\p . . Organrzatronal . Distortion Complexity 93 APPENDIX F HIGH AND LOW LEVELS OF AUTONOMY AND INTERDEPENDENCE AND THEIR INFLUENCE ON RELATIONAL QUALITY AND MESSAGE SELECTION High Autonomy and High Interdependence High Autonomy High Interdependence Organizational Trust .42 .36 Organizational Siz ~27 the ~28 Truth “—2,, Distortion Relationship "35 Organizational "20 a . ghgpg: Complexity u 16° Low Autonomy and Low Interdependence Low AItonomy Low Interdependence Orgarnzatlonal Trust '17 Truth .07 01' . . a1 S. 11 -.11 D anrzatlon z -- ' ' g l \ Quality Of f a lstortlon the Relationship 22 Ch . . -.06 fl) ange Orgamzatlonalfi -—> Subject Complexrty 94 REFERENCES 95 References Argyle, M. (1969). Social interaction. Chicago: Aldine Publishing Company. Athanassiades, J. C. (1973). The distortion of upward communication in hierarchical organizations. Academy of Management Journal, 16 (2), 207-226. Athanassiades, J. C. (1974). An investigation of some communication patterns of female subordinates in hierarchical organizations. Humg Relatiens, 27 (3 ), 195-209. Bacharach, S. B., & Aiken, M. (1977). Communication in administrative bureaucracies. Academy of Management ioumal, 20, 356-377. Bavelas, J. B. (1983). Situations that lead to disqualification. flipped Cemmmicatiep Researeb, 2, 130-145. Bavelas, J. B. (1985). A situational theory of disqualification: Using language to ‘leave the field.’ in J. Forgas (Ed.). Lepgpage and sociel situations. New York: Springer. Bavelas, J. B., Black, A., Chovil, N., & Mullet, J. (1988). Equivocation as a solution to conflicting interpersonal goals. Paper presented at the 9th Annual Conference on Discourse Analysis. Temple University, Philadelphia. Bhattacharya, R., Devinney, T. M., & Pillutla, M. M. (1998). A formal model of trust based on outcomes. Academy ef Management ,lo_u_rpal, 23, 459-472 Blau, P. M. (1964). Exchange in power in social life. New York: Wiley. Bond, C. F., Kahler, K. N., & Paolicelli, L. M. (1985). The miscommunication of deception: An adaptive perspective. ioumal of Experimental Psychology, 21, 331- 345. ‘ Breaugh, J. A. (1989). The work autonomy scales: Additional validity evidence. Humep Relations, 11, 1033-1056. Buller, D. B., & Burgoon, J. K. (1996). Interpersonal deception theory. Communicatien Theog, 3, 203-242. Buller, D. B., Burgoon, J. K., Buslig, A., Roiger, J. (1996). Testing interpersonal deception theory: The language of interpersonal deception. W 268-289. Burgoon, J. K., Hale, J. L. (1984). The fundamental topoi of relational communication. Communication Monogpephe, 51, 193-214. 96 Butler, J. K. (1999). Trust expectations, information sharing, climate of trust, and negotiation effectiveness and efficiency. Croup and mganizational Mmegemem, 2_4, 217-238. Butler, J. K., & Cantrell, R. S. (1984). A behavioral decision theory approach to modeling dyadic trust in superiors and subordinates. Psychological Reports, 55, 19-28. Campbell, D. T., & F iske, D. W. (1959). Convergent and discriminate validation by the multitrait-multimethod matrix. Psychological Bulletin, 56, 81-108. Chan, M. (1997). Some theoretical propositions pertaining to the context of trust. International ,leumal of annizaiional Analysis, 5. 227-248. Cohen, J., & Cohen, P. (1983). Applied multiple reggession/Correlatiep malyeis fer the behavioral sciences (2nd Edition). London: Lawrence Erlbaum Associates, Publishers. Cook, J ., & Wall, T. (1980). New work attitude measures of trust, organizational commitment and personal need non-fulfillment. Journal ef Cecupational Psychology, 53, 39-52. Creed, W. E. D., & Miles, R. E. (1996). Trust in organizations: A conceptual framework linking organizational forms, managerial philosophies, and the opportunity costs of controls. In R. M. Kramer & T. R. Tyler (Eds.), Trust in organizaiiepe', Froniiere ef theeiy and research. Thousand Oaks, CA: Sage, 16-38 Dasgupta, P. (1988). Trust as a commodity. In D. Gambetta (Ed.), Trust: Making md breaking eooperative relations. Cambridge, MA: Basil Blackwell, 49-72. Deluga, R. J ., & Perry, J. T. (1994). The role of subordinate performance and ingratiation in leader-member exchanges. Group and annizational Managemeni, i2, 67-86. Dillard, J. P., Segrin, C., & Harden, J. M. (1989). Primary and secondary goals in the production of interpersonal influence messages. Communicatiop Monogr_aphs, 56, 19-37. Dienesch, R. M., & Liden, R. C. (1986). Leader-member exchange model of leadership: A critique and further development. Academy of Manegement Joymei, 12, 61 8-634. Dockery, T. M., & Steiner, D. D. (1990). The role of the initial interaction in leader-member exchange. rou and anizational dies 1 395-413. Donohue, W. A., & Roberto, A. J. (1993). Relational development as Negotiated Order in hostage negotiation. Hman Commimication Research, 20, 175-198. 97 Driscoll, J. W. (1978). Trust and participation in organizational decision making as a predictor of satisfaction. Academy of Management Jeumal, 21, 44-56. Duck, S. W. (1977). Inquiry, hypothesis, and the quest for validation. In S. Duck (Ed), Theory and practice in intemersonal ettraction. New York: Academic Press. Edwards, A. L., & Abbott, R. D. (1973). Measurement of personality traits: Theory and technique. In P. H. Mussen & M. R. Rozenzweig (Eds.). Annual Review ef Psychology, 24, 241-278. Eisenberg, E. M. (1984). Ambiguity as strategy in organizational communication. Commpnicatien Monomhe, S1, 227-242. Eisenhardt, K. M. (1989). Agency theory: An assessment and review. Aeedemy ef Mepagemem Review, 14, 57-74. Erez, M., & Rim, Y. (1982). The relationship between goals, influence tactics, and personal and organizational variables. Hman Reletiope, 35, 871-878. F airhurst, G. T. (1993). The leader-member exchange patterns of women leaders in industry: a discourse analysis. Communicatien Monogmphs, bQ, 321-351. Fairhurst, G. T., & Chandler, T. A. (1989). Social structure in leader-member interaction. Cemimicatien Monogpaphs, 5b, 215-236. Fiedler, K., & Walka, I. (1993). Training lie detectors to use nonverbal cues instead of global heuristics. Hpman Cemmunication Researcb, 2, 199-223. Fiske, S. T., & Taylor, S. E. (1991). Seejdlpemmmb, New York: McGraw- Hill, Inc. Fox, A. (1974). BMW London: Faber and Faber. Fulk, J ., & Mani, S. (1986). Distortion of communication in hierarchical relationships. Communicatiop Yearbook, 9, 483-510. Gaines, J. H. (1980). Upward communication in industry: An experiment. Human Reletions, 12, 929-942. Gardner, W. L. 111. (1992). Lessons in organizational dramaturgy: The art of impression management. nganizational D3mamice, 21, 33-46. Gerstner, C. R., & Day, D. V. (1997). Meta-analytic review of leader-member exchange theory: Correlates and construct issues. umal f A 827-844. 98 Giacalone, R. A. (1989). The presentation of self in evegyday iife. Garden City, NY: Doubleday/Anchor. Giacalone, R. A., & Rosenfeld, P. (Eds.). (1989). impression managemmi in the organization Hillsdale, NJ: Lawrence Erlbaum. Gilbert, J. A., & Tang, T. L. (1998). An examination of organizational trust antecedents. delic Personnel Managemeni, 27, 321-338. Grant, D. & Oswick, C. (1995). An Afterward: Is there anything more to be said about metaphor? In D. Grant & C. Oswick (Eds.). Metaphor mid ergmizetione. London: Sage publications. Green, S. G., Blank, W., & Liden, R. C. ( 1983). Market and organizational influences on bank employees’ work attitudes and behaviors. al f A l' Psychelogy, 68, 298-306. Grice, P. (1989). Studies in me way ef werde. Cambridge, MA: Harvard University Press. Grover, S. L. (1993). Lying, deceit, and subterfuge: A model of dishonesty in the workplace. mganizational Scienee, 4, 478-495. Hackman, J. R., & Oldham, G. R. (1975a). The development of the job diagnostic survey. ioumal of Applied Psyehology, eo, 159-170. Hackman, J. R., & Oldham, G. R. (1975b). Work redesign. Reading, MA: Addison-Wesley Publishing Company. Hage, J. (1974). Communicetion and erganizational centrol: Cybemeiiee in health and welfare settings. New York: John Wiley. Hage, J ., Aiken, M., & Marrett, C. B. (1971). Organization structure and communications. American Socielegical Review, 26, 860-871. Hildebrand, C. (1998). Beware the weak links. CIQ,11, 58-66. Homans, G. C. (1961). Social behavior: Its elementmy forme. New York: Harcourt, Brace and World. Homans, G. C. (1974). Social behavior: Its elementeg femie, (Rev. ed,), New York: Harcourt, Brace, Jovanovich. Hubbell, A. P. (1999a). “I love your family - they are just like you”: Lies we tell to lovers and perceptions of their honesty and appropriateness. Paper presented at the International Communication Association Annual Convention, San Francisco, June, 1 999. 99 Hubbell, A. P. (1999b). The integration of three perspectives of organizational deception: A typology of deceptive messages. Paper presented at the National Communication Association Annual Convention, Chicago, November, 1999. Hunter, J. E., & Hamilton, M. A. (1995). Path: A progmrp in basiee. J ablin, F. M. (1982). Formal structural characteristics of organizations and superior-subordinate communication. Human Communicaiion Resemeb, 8, 338-347. Jablin, F. M. (1987). Formal organizational structure. In F. M. Jablin, L. L. Putnam, K. H. Roberts, & L.W. Porter (Eds). Handbook ef organizationel communication/An interdisciplinm perspective. London: Sage Publications. Johnson-George, C., & Swap, W. C. (1982). Measurement of specific interpersonal trust: Construction and validation of a scale to assess trust in a specific other. J0me] of Remonalim and Sociel P§yeholegy, 4;, 1306-1317. Jones, G. R., & George, J. M. (1998). The experience and evolution of trust: Implications for cooperation and teamwork. Academy of Management Review, 2;, 53 1- 546. Karsarda, J. D. (1974). The structural implications of social system size: A three-level analysis. mericmi Seeieiegjeel Review, 39, 19-28. Kimberly, J. R. (1976). Organizational size and the structuraliest perspective: A review, critique, and proposal. Administrative Seience Quarterly, 21, 571-597. Kingwell, M. (1993).. Is it rational to be polite? The Journal ef Rm'iosephy, S, 387-404. Klauss, R., & Bass, B. M. (1982). Intemersonal cemunicatiep ip ergem'eetiepe. New York: Academic Press. Kozlowski, S. W., & Doherty, M. L. (1989). Integration of climate and leadership: Examination of a neglected issue. Jemal bi Applied Psyebelegy, :24, 546- 553. Levine, T.R., McComack, SA, and Baldwin Avery, P. (1992). Sex differences in emotional reactions to discovered deception. Commrication Quarterly, 40, 289-296. Lewis, J. D., & Weigert, A. (1985). Trust as a social reality. Seeial Fomee, b3, 967-985. Lewicki, R. J ., McAllister, D. J ., & Bies, R. J. (1998). Trust and distrust: New relationships and realities. Academy ef Management Review, 23, 438-458. Liden, R. C., & Graen, G. (1980). Generalizability of the vertical dyad linkage model of leadership. Academy of Management Jedmal, 23, 451-465. 100 Lurnata, M. F. (1988). Commdm'cation g'goup membership: Determinete bf elimate perceptions and werk satisfactien in organizations. Unpublished dissertation at Michigan State University, Department of Communication. Marrow, A. J ., Bowers, D. G., & Seashore, S. R. ( 1967). Management by participation: Creating a climate for personal and organizatienal developmeni. New York: Harper & Row Publishers. McCauley, D. P., & Kuhnert, K. W. (1992). The professional manager. New York: McGraw-Hill. McComack, S. A. (1992). Information Manipulation Theory. We Monogmphs, 59, 1-16. McComack, S. A., Levine, T. R., Solowczuk, K. A., Torres, H. I., and Campbell, D. M. (1992). When the alteration of information is viewed as deception: An empirical test of information manipulation theory. WM, 59, 17-29. McEvoy, G. M. (1990). Public sector managers’ reactions to appraisals by subordinates. Public Personnel Management, 1212), 201-212. Mellenger, G. D. (1956). Marx, Durkheim, mid Weber: Fermetiene efrnodern social thougbt. London: Sage Publications. Mendenhall, M. E., & Wiley, C. (1994). Strangers in a strange land: The relationship between expatriate adjustment and impression management. Americm Behavioral Scientist S], 605-620. Miller, V. D., Johnson, J. R., Hart, 2., & Peterson, D. L. (1999). A test of antecedents and outcomes of employee role negotiation ability. Jepmal of Applied Communication Researeb, 27, 24-48. Muchinsky, P. M. ( 1977). Organizational communication: Relationships to organizational climate and job satisfaction. Academy of Managemmii Jeurnm, 20, 592- 607. Neimeyer, R. A., Brooks, D. L., & Baker, K. D. (1996). Personal epistemologies and personal relationships. In D. Kalekin Fishman & M. Walker (Eds.), lbe construetion of gioup realiiiee. Malabar, FL: Krieger. O’Reilly, C. A. III. (1978). The intentional distortion of information in organizational communication: A laboratory and field investigation. Humep Relab'ene, 21, 173-193. Porter, L.W., & Roberts, K. H. ( 1976). Communication in organizations. In M. Dunnette (Ed.), Handboek of induetrial and organizatiopal psvchology. Chicago: Rand McNally. 101 Read, W. H. (1962). Upward communication in industrial hierarchies. H_umep Relations 15 3-15. Roberts, K. H. & O’Reilly, C. A., III. (1974a). Failures in upward communication in organizations: Three possible culprits. Academy ef Management Jemal, 17, 205-215. Roberts, K. H. & O’Reilly, C. A, HI. (1974b). Measuring organizational communication. Journal of Applied Psychology, 59, 321-326. Rosenfeld, P., Giacalone, R. A., & Riordan, C. A. (1994). Impression management theory and diversity: Lessons for organizational behavior. ;A_m_ri_mi Behavioral Scientiei, 31, 601-604. Rosseau, D. M., Sitkin, S. B., Burt, R. S., & Camerer, C. (1998). Not so different after all: A cross-discipline view of trust. Academy ef Mmegmmt Review, 23, 393- 404. Rotter, J. B. (1967). A new scale for the measurement of interpersonal trust. Jemal ef Pemonalim, 25, 651-665. Schmitt, N. (1994). Method bias: The importance of theory and measurement. Journal of annizatienel Behavior, 15, 393-398. Schriesheim, C. A., Neider, L. L., Scandura, T. A., & Tepper, B. J. (1992). Development and preliminary validation of a new scale (LMX-6) to measure leader- member exchange in organizations. E uca iona d P chol 'ca Measur 135-147. Sheppard, B. H., & Sherman, D. M. (1998). The grammars of trust: A model and general implications. Academy ef Management of Review, 22, 422-437. Smith, R. C. & Eisenberg, E. M. (1987). Conflict at Disneyland: A root- metaphor analysis. Communicetion Monogmphs, 54, 367-3 80. Snyder, R. A., & Morris, J. H. (1984). Organizational communication and performance. Jedmal ef Applied Psychology, b2, 461-465. Spector, P. E. (1994). Using self-report questionnaires in OB research: A comment on theuse of a controversial method. Jegmai of Qrgmiizmional Behevieg, 15, 385-392. Stiff, J. B. (1996). Theoretical approaches to the study of deception: Comments on the interpersonal deception theory. Communication Theom, 2, 289-296. 102 Steers, R. M., & Braunstein, D. N. (1976). A behaviorally-based measure of manifest needs in work settings. Journal of Vocational Behavior, 9, 251-266. Taylor, F. W. (1911). The prineiples of ecientiiic management New York: W. W. Norton & Company. Taylor, L., Gittes, M., O’Neal, E. C., & Brown, S. (1994). The reluctance to expose dangerous lies. Journal of Applied Social Psychology, 24, 301-315. Turner, J. H. (1967). Experimental environments and the emergence of achievement motivation in adolescent males. Sociomet_ry, 33, 147-165. Turner, R. E., Edgley, C., & Ohnstead, G. (1975). Information control in conversations: Honesty is not always the best policy. The Rmeas Jeumei ef Seciology, 69-89. Ullrnan, J. B. (1996). In B. G. Tabachnick & L. S. Fidell (Eds). LR'mg multivariate etaiisties (Third edition). New York: HarperCollins College Publishers. Vecchio, R. P., Griffeth, R. W., & Hom, P. W. (1986). The predictative utility of the vertical dyad linkage approach. The Journal of Seeial Psychology, 12b, 615-617. Wakabayashi, M., Graen, G., Graen, Michael, & Graen, Martin. (1988). Japanese management progress: Mobility into middle management. Journal of Applied Psyeholegy, 72, 217-227. Wayne, S. J., & Green, S. A. (1993). The effects of leader-member exchange on employee citizenship and impression management behavior. Humgp Relations, 4d, 1431- 1440. Wayne, S. J ., & Liden, R. C. (1995). Effects of impression management on performance ratings: A1. Academy ef Management Jeumel, 28, 232-260. Whitener, E. M., Brodt, S. E., Korsgaard, M. A., & Werner, J. M. (1998). Managers as initiators of trust: An exchange relationship framework for understanding managerial trustworthy behavior. Academy of Management Review, 22, 513-530. Zand, D. E. (1972). Trust and managerial problem solving. Administrative Science Qdarterly, 17, 229-240. 103