Citation
Linkages among Relationship Maintenance Strategies, Relationship Quality Outcomes, Attitude, and Behavioral Intentions

Material Information

Title:
Linkages among Relationship Maintenance Strategies, Relationship Quality Outcomes, Attitude, and Behavioral Intentions
Creator:
KI, EYUN-JUNG
Copyright Date:
2008

Subjects

Subjects / Keywords:
Canaries ( jstor )
Factor analysis ( jstor )
Mathematical variables ( jstor )
Modeling ( jstor )
Multilevel models ( jstor )
Perceptual organization ( jstor )
Psychological attitudes ( jstor )
Public relations ( jstor )
Statistical models ( jstor )
Trust ( jstor )
City of Gainesville ( local )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Eyun-Jung Ki. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
2/28/2007
Resource Identifier:
649814497 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

LINKAGES AMONG RELATIONSHIP MAINTENANCE STRATEGIES, RELATIONSHIP QUALITY OUTCOMES, ATTITUDE, AND BEHAVIORAL INTENTIONS By EYUN-JUNG KI A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2006

PAGE 2

Copyright 2006 by Eyun-Jung Ki

PAGE 3

To my beloved father, No-Sam Ki, who passed away in the summer of 2004; my dearest mother, Kyoung-Soon Lee; my love, Hyoungkoo Kh ang—their endless love and support made it possible to be where I am now.

PAGE 4

iv ACKNOWLEDGMENTS I have been extremely fortunate by ha ving numerous wonderful people who have supported and loved me endlessly during this long journey. Thanks to them, my journey was much happier and more satisfying than I had expected or planned. First, I am sincerely grateful to my mentor and dissertation chair, Dr. Linda Childers Hon, the Al and Effie Flanagan Prof essor and Senior Associate Dean for the College of Journalism and Comm unications at the University of Florida. She has been with me every step of the way to support, en courage, and guide me. She inspired me to be the kind of teacher and human being I wa nt to be. Whenever I encountered an obstacle or lost momentum, she was the one who cured my broken heart and made me regain focus and return to my work. Her warm heart, enthusiasm , and dedication have been truly inspirational and will never be forgotten. She has also provided me with several golden opportunities, including teach ing and research experiences, funding opportunities, and chances to earn several pres tigious awards. It was also a great honor for me to coordinate with her to write and publish several articles and to complete my dissertation research under her gu idance. I deeply appreciate a ll that she has done for me. I would like to express my gratitude to all my dissertation committee members. I am thankful to Dr. Spiro Kiousis, whose knowle dge and insight were particularly helpful in building and improving my theoretical framework. He also encouraged me emotionally and greatly assisted me in the j ob hunting process. Dr. Juan-Carlos Molleda read my dissertation thoroughly, providing de tailed comments that improved the quality

PAGE 5

v of my dissertation; he was always more than willing to offer help and provided academic and emotional support. Dr. Debbie Treise was always willing with a warm heart to be my help and supporter. Dr. Steven S hugan, who was an outside member on my dissertation committee, offered me great guidan ce, especially with my data analysis. Additionally, I gratef ully acknowledge a re search sponsorship from the Public Relations Department of Florida Farm Burea u. I would like to especially thank Edward Albanesi and Rod Hemphill, who have always been more than willing to offer me generous support and immediate responses, de spite their busy schedules. Without their support and generosity, I would not have been able to conduct the mail survey for my dissertation. Several other faculty members deserve recognition: Dr. Bill Chamberlin, my initial advisor in the doctoral program, has continuously offered me support and encouragement. He motivated me to emulate his nurturing behavior as a teacher, both in and out of the classroom. I also appreciate that he made it very easy to open my heart to him. Dr. Mary Ann Ferguson always attempte d to find the best opportunities for me and allowed me the privilege of working with her on a number of projects. Dr. Kathleen Kelly, our former department chair, provide d me with great teaching opportunities and offered her warm heart. I greatly appreciate all that these faculty members have done for me. The Korean Gators from the College of Journalism and Communications have encouraged me emotionally to stay focuse d. Without their support, my long-lasting journey could not have been completed. Al so, three Korean faculty members in the college; Drs. Chang-Hoan Cho, Youjin Choi, and Hyojin Kim have greatly supported me.

PAGE 6

vi The Korean Gator Tennis Team members also deserve recognition. My favorite days of the week have been either Thursday or Friday, when we get together to play tennis, because this time is always a joy for me. I appreciate the other team members’ patience with my poor performance as well as their personal care fo r me. I will always remember the pleasant times I shared with them. Needless to say, I am grateful to all of my colleagues, professors, and the staff members at the College of J ournalism and Communications. Their warm hearts and kind words made my journey much easier and more rewarding. Some people outside the University of Fl orida also deserve to be mentioned. I must express my gratitude to Professor John Ledingham at Capital University. He has provided a great deal of assi stance during my dissertation wr iting process and responded immediately whenever I asked questions. I also appreciate Dr. Sung-Un Yang who answered my frequent questions about statis tical analysis and continuously encouraged me. I feel a deep sense of gratitude for my dear friend Roxanne Watson, who has been my dissertation buddy. We have shared both the agony and joys of the research and writing process, and her encouragement and concern for my success has meant a great deal to me. I will always fondly remember the movie nights and workout sessions we shared. I also appreciate Ana-Klara Hering who provided ma rine style workout training as well as her encouragement duri ng my dissertation process. All of my friends all over the world have helped me not to get lost during my painful and lonely dissertation journey, especially Hyun Jung Yun, Youngshin Hong,

PAGE 7

vii Jung Yun Chang, YoungKyung Min, Yungjun Sung, and Eun Hee Moon. They always made my days much brighter and easier in so many ways. Unfortunately, I lost my beloved father, No-Sam Ki, in the summer of 2004 during my first year as a doctoral student. Although he is no longer with me on earth, I have felt that he has taken care of me, and I know he w ill continue to be with me in both happiness and sorrow. I am sure that he will be at my graduation ceremony in spirit and will be proud of what I have achieved thus far. My special thanks go to my mother, Kyoungsoon Lee, who has been my best consultant, friend, and pillar of support throughou t my life. I have been so lucky and happy to be her daughter. My achievements ar e the culmination of her life-long belief in me and my abilities and her c onfidence that I would attain success in my life. I also would like to thank my younger brother, Ta e-Yeon Ki, who has shown his love and concern during my dark hours. Last but not least, I woul d like to express my special thanks to Hyoungkoo Khang, my beloved fianc and soul mate, for his advi ce, encouragement, belief in me, and his everlasting love, especially during the dark hour s when I faltered. He constantly kept me focused on the light at the end of the tunnel. He is the most wonderful thing that has ever happened to me. The best experience we have shared has been trav eling together on our Ph.D. journeys and both successf ully receiving Ph.D.s from the College of Journalism and Communications at the University of Florida. GO GATORS!

PAGE 8

viii TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................ivLIST OF TABLES............................................................................................................xiiLIST OF FIGURES...........................................................................................................xvABSTRACT....................................................................................................................xvi i CHAPTER 1 INTRODUCTION........................................................................................................1Purpose of the Study.....................................................................................................4Significance of the Study..............................................................................................52 LITERATURE REVIEW...........................................................................................12Public Relations as Relationship Management...........................................................12Definitions of Organizati on-Public Relationships...............................................13Stages of Developing Relationships....................................................................15Antecedent...........................................................................................................17Part I: Relationship Maintenance Strategies...............................................................18Access...........................................................................................................23Positivity.......................................................................................................24Openness/Disclosure....................................................................................25Sharing of Tasks...........................................................................................26Networking...................................................................................................27Assurances....................................................................................................28Part II: Relationship Quality Outcomes......................................................................29Relationship Quality Outcome Dimensions for This Study................................34Control Mutuality................................................................................................35Satisfaction..........................................................................................................36Trust.....................................................................................................................37Commitment........................................................................................................38Linkages among the Four Relational Dimensions...............................................40Satisfaction and trust....................................................................................41Trust and commitment.................................................................................41Part III: A Comprehensive Model Linki ng Relationship Quality, Attitude, and Behaviors...............................................................................................................42

PAGE 9

ix Organization-Public Relati onships as Perception...............................................42Attitude................................................................................................................44Behavior Intentions.............................................................................................45Hierarchy of Effects............................................................................................46The Low-Involvement Hierarchy........................................................................50Comprehensive Model................................................................................................52Summary of Research Qu estions and Hypotheses.....................................................523 METHODOLOGY.....................................................................................................54The Process of Developing Measurement Scales.......................................................55Population and Samples......................................................................................58Quantitative Research Approach.........................................................................60Pretest...........................................................................................................62Mail survey...................................................................................................63Measures..............................................................................................................64Relationship maintenance strategies............................................................64Relationship quality outcomes.....................................................................66Attitude.........................................................................................................68Behavioral intentions....................................................................................68Involvement..................................................................................................69Demographic information............................................................................70Reliability and validity test..........................................................................70Data Reduction and Data Analysis.............................................................................72Statistical Procedures for Data Analysis.............................................................72Multiple Regression Analysis..............................................................................74Structural Equation Modeling.............................................................................75Model Fit.............................................................................................................764 RESULTS...................................................................................................................78Description of Samples...............................................................................................78Response Rate.....................................................................................................78Demographics......................................................................................................79Descriptive Statistics..................................................................................................82Relationship Maintenance Strategies..................................................................83Relationship Quality Outcomes...........................................................................87Attitude and Behavioral Intentions......................................................................92Reliability of Initial Measurement Items.............................................................93Measurement Models..................................................................................................94Factor Analysis....................................................................................................94Relationship maintenance strategies............................................................96Relationship quality outcomes...................................................................100Confirmatory Factor Analysis...........................................................................103Reliability and Validity Test.............................................................................107Correlation Analysis.................................................................................................109Multiple Regression Analysis...................................................................................111

PAGE 10

x Testing the Model of Relationship between Relationship Maintenance Strategies and Relations hip Quality Outcome...............................................118Best Indicator of Overall Relationship Quality.................................................122Structural Equation Modeling...................................................................................124Relationship among Relationship Quality Outcome Indicators........................125Linkages among Relationship Quality Outcome, Attitude, and Behavioral Intentions........................................................................................................127Linkages among Relationship Quality Pe rception, Attitude, and Behavioral Intentions of Low-Involved Group................................................................129Comprehensive Model Linking Maintenance Strategies, Relationship Quality Outcome, Attitude and Behavioral Intentions.....................................................131Post Hoc Analysis.....................................................................................................135Revised Model Linking Relationship Quality Outcome, Attitude, and Behavioral Intentions.....................................................................................135The Comprehensive Model Linking Rela tionship Maintenance Strategies, Relationship Quality Outcome, Attitude, and Behavioral Intentions............1375 DISCUSSION AND CONCLUSION......................................................................139Summary of Results..................................................................................................139Descriptive Statistics.........................................................................................140Relationship maintenance strategies..........................................................140Relationship quality outcomes...................................................................140Correlation Analysis..........................................................................................141Measures of Relationship Maintenance Strategies and Relationship Quality Outcomes.......................................................................................................142Causal Links between Relationship Main tenance Strategies and Relationship Quality Outcome............................................................................................144Best Indicator of Relationship Qual ity Outcome of Overall Relationship Quality............................................................................................................146Influential Order among Relationship Indicators..............................................147Linkages among Relationship Perception, At titude, and Behavioral Intentions147The Comprehensive Model...............................................................................148Post Hoc Analysis..............................................................................................150Linkage among relationship quality per ceptions, attitude, and behavioral intentions on relationship length...........................................................150The comprehensive model on length of relationship.................................150Theoretical and Manage rial Implications.................................................................151New Measure of Relationship Maintenance Strategies.....................................151More Refined Measure of Relationship Quality Outcomes..............................153Sequential Order of the Relationship Indicators...............................................153Best Indicator of Overall Relationship Quality.................................................155The Ways in Which Communication Activ ities as Relationship Maintenance Strategies Contribute to Relationship Quality Outcomes..............................156The Ways in Which Relationship Qualit y Outcomes Contribute to Attitude and Behavior..................................................................................................161

PAGE 11

xi A Comprehensive Model Linking Rela tionship Maintenance Strategies, Relationship Quality Outcomes, Attitude, and Behavior...............................164Limitations and Future Research.......................................................................167Conclusion.........................................................................................................171APPENDIX A SURVEY INVITATION..........................................................................................172B COVER LETTER.....................................................................................................173C SURVEY QUESTIONNAIRE.................................................................................174D INFORMED CONSENT FORM..............................................................................178E FOLLOW-UP POSTCARD CONTENT..................................................................179F COVER LETTER SAMPLE OF REPLACEMENT MAIL.....................................180LIST OF REFERENCES.................................................................................................181BIOGRAPHICAL SKETCH...........................................................................................200

PAGE 12

xii LIST OF TABLES Table page 3-1 Relationship maintenance st rategies measurement items........................................653-2 Relationship quality out come measur ement items...................................................673-3 Model fit criteria.......................................................................................................774-1 Response rate calculation.........................................................................................794-2 Sample demographic description.............................................................................814-3 Descriptive statistics for re lationship maintenance strategies..................................864-4 Descriptive statistics of relationship quality outcomes............................................904-5 Descriptive statistics of att itude and behavioral intentions......................................934-6 Results of reliability tests of initial measurement items..........................................944-7 Factor loadings for access........................................................................................964-8 Factor loadings for positivity...................................................................................974-9 Factor loadings for openness....................................................................................974-10 Factor loadings for sharing of tasks.........................................................................984-11 Factor loadings for networking................................................................................984-12 Factor loadings for assurances.................................................................................994-13 Factor loadings for control mutuality.....................................................................1004-14 Factor loadings for satisfaction..............................................................................1014-15 Factor loadings for trust.........................................................................................1014-16 Factor loadings for commitment............................................................................1024-17 Confirmatory factor analysis of relationship maintenance strategies....................104

PAGE 13

xiii 4-18 Fit measures for relationship mainte nance strategies measurement model...........1054-19 Confirmatory factor analysis of relationship quality outcomes.............................1064-20 Fit measures for relationship qua lity outcome measurement model......................1074-21 Correlation matrix between relationship maintenance st rategies and relationship quality outcomes....................................................................................................1124-22 Stepwise regression analysis of relationship maintenance strategies on relationship quality outcomes.................................................................................1164-23 Stepwise regression analysis of relationship quality outcomes on overall relationship.............................................................................................................1174-24 Path model of relationship maintenan ce strategies and relationship quality outcomes.................................................................................................................1204-25 Path model of relationship quality outcome on overall re lationship quality..........1234-26 Path model of the relationship among relationship quality outcome indicators....1264-27 Fit measures of the relationship amo ng relationship quality outcome indicators..1264-28 Tested model linkages among relationsh ip quality perception, attitude, and behavioral intentions..............................................................................................1284-29 Fit measures of linkages among rela tionship quality perception, attitude, and behavioral intentions..............................................................................................1284-30 Tested model linkages among percepti ons of relationship quality perception, attitude, and behavioral inte ntions on low-involvement........................................1304-31 Fit measures of linkages among rela tionship quality outcome, attitude, and behavioral intentions on low-involved public........................................................1314-32 Testing comprehensive model................................................................................1334-33 Fit measures of comprehensive model...................................................................1334-34 Tested model linkages among relationsh ip quality outcome, attitude, and behavioral intentions on the length of the relationship..........................................1364-35 Fit measures of linkages among rela tionship quality outcome, attitude, and behavioral intentions on the length of the relationship..........................................1364-36 Testing the comprehensive model on the length of the relationship......................1374-37 Fit measures of the comprehensive model on the length of the relationship.........138

PAGE 14

xiv 5-1 Final measurement items for relationship maintenance strategies.........................1435-2 Final measurement items for relationship quality outcomes..................................143

PAGE 15

xv LIST OF FIGURES Figure page 1-1 Measurement model of relati onship maintenance strategies......................................61-2 Measurement model of re lationship quality outcomes..............................................71-3 Linkages among relationship maintenan ce strategies and re lationship quality outcomes.....................................................................................................................71-4 Linkages among relationship quality outcomes.........................................................81-5 Sequential model 1: Relationship qual ity perception, attitude, and behavioral intentions....................................................................................................................81-6 Sequential model 2: Relationship qual ity perception, behavi or, and attitude............81-7 Linkages among relationship maintena nce strategies, relationship quality outcomes, attitude, and behaviors..............................................................................82-1 Stages and forms of relationships.............................................................................162-2 Simplified version of stages and forms of relationships..........................................172-3 Measurement model of relati onship maintenance strategies....................................302-4 An initial conceptual model linking relationship maintenance strategies and relationship quality outcomes...................................................................................312-5 Proposed model for linkages am ong relationship quality outcome.........................422-6 Proposed model linking relationship quali ty perception, attitude, and behavioral intentions..................................................................................................................492-7 Proposed model linking relationship quali ty perception, behaviors, and attitude....523-1 Major steps for devel oping a measurement scale ....................................................554-1 Measurement model of relati onship maintenance strategies..................................1054-2 Measurement model of re lationship quality outcomes..........................................107

PAGE 16

xvi 4-3 Initial model of the relationship between relationship maintenance strategies and relationship quality outcomes.................................................................................1184-4 Final model linking relationship maintena nce strategies and relationship quality outcomes.................................................................................................................1224-5 Path model of relationship quality ou tcomes on overall relationship quality........1244-6 Path model of the relationship among relationship quality outcome indicators....1264-7 Initial model linking relationship quality outcomes, attitude and behavioral intentions................................................................................................................1294-8 Revised model linking relationship quali ty outcomes, attitude and behavioral intention..................................................................................................................1294-9 The model tested model linkages among relationship quality outcomes, attitude, and behavioral intentions on low-involvement......................................................1314-10 Comprehensive model linking maintena nce strategies, relationship quality outcomes, attitude, and behavioral intentions........................................................1345-1 Relationships among relationship main tenance strategies and relationship quality outcomes....................................................................................................1455-2 The influential sequence among rela tionship quality outcome indicators.............1475-3 Final model linking relationship quality perception, attitude, and behavioral intentions................................................................................................................1485-4 Comprehensive model linking relationsh ip maintenance strategies, relationship quality outcomes, attitude, and behavioral intentions............................................149

PAGE 17

xvii Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy LINKAGES AMONG RELATIONSHIP MAINTENANCE STRATEGIES, RELATIONSHIP QUALITY OUTCOMES, ATTITUDE, AND BEHAVIORAL INTENTIONS By Eyun-Jung Ki August 2006 Chair: Linda Childers Hon Major Department: Journalism and Communications The purpose of this study was to expl ore two central con cepts—relationship maintenance strategies and re lationship quality outcomes—f ocusing especially on their measures and causal relationships. In additi on, this study investigat es the relationships between or among: the influence of order in relationship percepti ons, attitudes, and behavioral intentions; and relationship ma intenance strategies, relationship quality outcomes, attitudes, and behavioral intentions. To accomplish the research purposes, this study examined the relationship between Florida Farm Bureau, as an organization, and its members, as the organization’s strategic public. The two proposed measures, relationshi p maintenance strategi es, and relationship quality outcomes, were found to be reliable and valid. These developed measures can help practitioners better understand how to main tain or cultivate rela tionships with their target publics.

PAGE 18

The causal relationships between the ma intenance strategies and relationship quality outcomes indicate that the four strate gies of access, positivity, sharing of tasks, and assurances represent effective, pro active approaches that organizations can implement to maintain or cultivate relationships with their strategic publics. However, an organization’s use of openness and networki ng was found not to improve the public’s perception of any relationship quality outcome. For the four relationship quality dimensi ons used in this study—control mutuality, satisfaction, trust, and commitment—current me mbers’ perceptions of relationship trust and commitment were influential to their posi tive attitude toward th e organization. More importantly, the present study shows that the public’s perception of commitment can also directly engender supportive behavior towa rd the organization am ong members of a key public. The research was also designed to empirica lly test a model that posited the linkages among relationship maintenance strategies, relationship quality outcomes, supportive attitude, and behaviors toward an orga nization among members of a key public. The current study proves that relationship mainte nance strategies can effectively generate quality relationships between an organizati on and its publics. Furthermore, a public’s supportive attitude and behaviors toward the organization can arise from positive perceptions of relationship quality. xviii

PAGE 19

1 CHAPTER 1 INTRODUCTION Understanding how to effectively manage relationships between organizations and their target publics has influenced public relations scholars and practitioners for more than a decade. Quality relationships with publics result in several positive public relations effects, such as im proving organizational effectiven ess (Dozier, L. A. Grunig, & J. E. Grunig, 1995; L. A. Grunig, J. E. Grunig, & Dozier, 2002; Hon, 1997; Huang, 1999); resolving conflicts between the or ganization and its publics (Huang, 1997); affecting publics’ attitudes, evaluations, and behaviors (Bruning, 2002; Ki & Hon, 2007); developing positive public relations strategi es (Huang, 2004); and enhancing corporate reputation (Bridges & Nelson, 2000; Hutton, Goodman, Alexander, & Genest, 2001). This relationship-centered perspective shifts the traditional function of public relations (e.g., disseminating information and attaining publicity) to th e more meaningful purpose of cultivating stable, long-term, and quality relationships between an organization and its publics. J. E. Grunig and Hunt’s (1984) defi nition of public relations as the management of communication between an organization and its publics fu rther focuses on relationship management, rendering public relations a more valuable practice. Numerous scholars devote their studies to the field of organization-public relationships. Sallot, Lyon, Acosta-Alzuru, and Jones (2003) found th at during the last decade, relationship theory has been the second most frequently used perspective in public relations scholarship. Ledingham ( 2003) proposed relationship management as a general theory for public rela tions based on Littlejohn’s cr iteria—usefulness, parsimony,

PAGE 20

2 elegance, and empirical support—and offered a useful framework for the study, teaching, and practice of public relations. Studies of organization-public relationships include three stages: 1) antecedents of relationships, 2) relationship maintenance strate gies, and 3) relationship quality outcomes. The first category, antecedents of relations hips, describes the underlying reasons why organizations establish relati onships with specific public s (Broom, Casey, & Ritchey, 1997). Antecedents of relationships are define d as “social and cultu ral norms, collective perceptions and expectations, needs for resour ces, perceptions of unc ertain environment, and legal/voluntary necessity” (Broom et al., 1997, p. 94). The second category, relationship maintenance strategies, covers the strategies utilized to maintain and cultivate those relationships. Finally, re lationship outcomes are the consequences, or measures of relationship quali ty, that are produced by effec tive relationship maintenance (J. E. Grunig & Huang, 2000). In effect, relationship maintenance strategies lead to quality relationship quality outcomes. Scholars posit that several relationship maintenance strategies (e.g., access, positivity, openness, sharing of tasks, networking, and assurances) can produce better relationship quality outcomes (e.g., contro l mutuality, satisfaction, trust, and commitment) (J. E. Grunig & Huang, 2000; Hon & J. E. Grunig, 1999). Some organization-public relationshi p studies have attempted to measure relationships and refine existing measurement scales (e.g., Br uning & Ledingham, 1999; Ferguson, 1984; L. A. Grunig, J. E. Grunig & Ehling, 1992; J. E. Grunig & Huang, 2000; Hon & J. E. Grunig, 1999; Huang, 1997, 2001a; Jo, 2003, 2006; Kim, 2001) . Because these studies focused on relationship quality outcome scale development, they fail to emphasize the

PAGE 21

3 establishment of measurements for relations hip maintenance strategies or how proposed relationship maintenance strategies infl uence relationship quality outcomes. Relationship maintenance strategies are thought to impact positive relationship quality outcomes. Interpersona l relationship literature presen ts consistent findings on the association between relationship maintena nce strategies and relationship quality outcomes (Canary & Stafford, 1992, 1993; Stafford & Canary, 1991). For example, positivity has been identified as the most e ffective relationship maintenance strategy in sustaining control mutuality. Likewise, assuranc es are consistently noted as an important predictor of commitment (Canary & Sta fford, 1992, 1993; Stafford & Canary, 1991). This research examines the extent to which maintenance strategies predict relational outcomes by adopting this interper sonal relationship pers pective as well as suggestions noted by public relations schol ars (e.g., J. E. Grunig & Huang, 2000; Hon & J. E. Grunig, 1999). Knowing which maintenan ce strategies affect relationship quality outcomes and how they do so is useful to publ ic relations in both th eory and practice. However, prior to exploring the influence of maintenance strategies on relationship quality outcomes, a relationship maintenance strategies scale should be developed based on Spector’s framework (1992), which provi des a model for creating multiple-item measurement scales. The existing relationshi p quality outcomes also must be refined based on the steps suggested by Spector (1992). Although most scholars and practitioners agree on the importance of key relationship quality outcomes (e.g., contro l mutuality, satisfaction, trust, and commitment), important questions about the li nks between these indica tors have not been addressed properly. After examining relations hip theory in other disciplines, scholars

PAGE 22

4 suggested that some of the relationship i ndicators actually preced e others (Ki & Hon, 2005, 2007; Jo, 2003, 2006). For instance, they pr oposed that satisfaction might be an antecedent of trust and trust a predictor of commitment. Thus, this study explores the links between different rela tionship quality outcomes. A public relations program evaluati on measures how an organization’s communication impacts a specific or target pub lic’s perceptions of its relationship with the organization (Bruning & Ledingham, 2000). Scholars have investigated the association between relationship quality outcom e, attitude, and behaviors or behavioral intentions. Specifically, Ki and Hon (2007) explored a causal m odel linking a public’s relationship perceptions, attitude toward the organization, and behavior al intentions using the standard hierarchy of effects model (perception attitude behavior). They found that among the six indices proposed by Hon and J. E. Grunig (1999), publics’ relational perceptions of satisfaction and control mutua lity were significant predictors of attitude, which influences behavioral inte ntions. The extant hierarchy of effects literature explains that different influential se quences among perception, attitude, and behavior occur in low-involvement situations. In other word s, a public with low-involvement in an organization-public relationship might go throug h different sequences. Therefore, this study attempts to find a more applicable model to measure relationships with lowinvolvement publics. Purpose of the Study The purpose of this study is to explore two concepts: relational maintenance strategies and relati onal outcomes. In addition, this study will investigate the causal relationships between or among relationship maintenance strategies, the outcomes of quality organization-public rela tionships, as well as public a ttitude and behaviors.

PAGE 23

5 Therefore, seven models—two measuremen t models and five causal relationship models—are proposed. Two measurement models: A measurement model of relationship maintenance strategies—access, positivity, openness, sharing of tasks, networking, and assurances A measurement model of relationship quality outcomes—control mutuality, satisfaction, trust, and commitment Five causal models (struc tural equation models): A model to test linkages among rela tionship quality outcome indicators. A model to test which and to what extent relationship maintenance strategies affect relationship quality outcomes. Two models to test the linkages between relationship quality outcome perceptions, attitude, and behaviors. Sequential order 1: relationship perception attitude behavior Sequential order 2: relationship perception behavior attitude A comprehensive model that links relations hip maintenance strategies, relationship quality outcomes, attitude, and be haviors toward an organization. Significance of the Study This research is original and significant in severa l ways. First, this study contributes to theoretical developments explicating the conceptual definition and operational measures of relationship maintena nce strategies. These findings will benefit scholars and practitioners, not only in the fiel d of public relations, but also those studying interpersonal relations as well as interorganizationa l relations. This is the first study to empirically measure relationship maintenance strategies—which are essentially the dayto-day activities of practitione rs at the program level—and li nk these strategies to their impacts on organizations. Thus, this re search will provide insights into how

PAGE 24

6 organizations should direct their public relatio ns efforts to maintain and cultivate positive relationships with their key publics. Figure 1-1. Measurement model of re lationship maintenance strategies

PAGE 25

7 Control Mutuality Trust Satisfaction Commitment Control Mutuality 1 Control Mutuality 2 Control Mutuality 3 Control Mutuality 4 Turst 1 Trust 2 Trust 3 Trust 4 Satisfaction 1 Satisfaction 2 Satisfaction 3 Satisfaction 4 Commitment 1 Commitment 2 Commitment 3 Commitment 4 Figure 1-2. Measurement model of relationship quality outcomes Relationship Maintenance Strategies Relationship Quality Outcome Note : Arrow indicates the direction of impact. Figure 1-3. Linkages among relationship mainte nance strategies and relationship quality outcomes

PAGE 26

8 Satisfaction Trust Commitment Note : Arrows indicate the direction of influence. Figure 1-4. Linkages among rela tionship quality outcomes Relationship Quality Perception Attitude Behavior Figure 1-5. Sequential model 1: Relationship quality percepti on, attitude, and behavioral intentions Relationship Quality Perception Behavior Attitude Figure 1-6. Sequential model 2: Relationship quality percep tion, behavior, and attitude Relationship Maintenance Strategies Positive Attitude Supportive Behaviors Relationship Quality Outcome Figure 1-7. Linkages among relationship main tenance strategies, relationship quality outcomes, attitude, and behaviors Second, finding the causal links between re lationship maintenance strategies and relationship quality outcomes will provide gui delines for how an organization should use each strategy to impact specific relationship quality outcomes. On a pragmatic level, the results of this study will al so benefit organizations inte rested in employing effective

PAGE 27

9 relationship maintenance strategies and improving target publics’ perceptions of organization-public relationships. Third, the existing relationship quality outco me measurements are refined in this study to ensure the robustness of scale th rough reliability and validity using multiple scale development procedures. Relationship quality outcome scales differ from study to study and only a handful of studies have atte mpted to measure the reliability and validity of the relationship quality outcome scal es (e.g., Huang, 2001b; Jo, 2006; Kim, 2001). The few studies on organization-public rela tionship measurement are insufficient to standardize scales that will prove useful to sc holars and practitioners. To develop reliable and valid measurement scales, several separate studies that test and apply the measures for standardization are necessa ry. Furthermore, applying re lationship measures to other types of organization-public relationships through multiple scale development procedures may help achieve more accurate relationship measures. Likewise, proposed relationship quality outcome measures should be statisti cally tested through exploratory factor analysis and confirmatory factor analysis. Fourth, as previously suggested by scholar s and several relational marketing studies (Anderson & Narus, 1990; Ganesan, 1994; Garbarino & Johnson, 1999), this study empirically determines sequential order among the relationship quality outcome dimensions—satisfaction, trust, and commitment (Jo, 2003, 2006; Ki & Hon, 2005, 2007). Identifying links betw een relationship indicators can provide a more solid framework for measuring organization-public relationships. More importantly, a solid framework may provide public relations profe ssionals with more practical information.

PAGE 28

10 Knowing the relationship quality outcome deve loped within a specific public would help professionals establish more stra tegic public relations programs. Fifth, this study empirically tests a mode l for linking relationship perceptions, attitude, and behavioral intentions, or what Ki and Hon (2007) tested using structural equation modeling (SEM). It has been suggest ed that positive, long-term relationships are valuable to organizations because these relationships are more likely to encourage supportive behaviors (e.g., sales, donations, favorable legislation, and high performance among employees), while preventing unsupportiv e behaviors (e.g., boycotts, picketing, litigation, and government re gulation) among publics (L. A. Grunig et al., 2002). Although scholars and practitioners have argue d that effective relationship management is the litmus test for successful public relati ons, few studies have empirically determined how positive relationships drive attitude and behavior. The models proposed can potentially provide compelling evidence of th ese linkages and move theory and practice forward in terms of evaluating the effec tiveness of public relations programs. Additionally, this study tests an alternative model suggested by the existing hierarchy of effects literature: that rela tionship perception directly a ffects behaviors which impacts attitude (relationship perception behavior attitude) for low-involvement publics. Lastly, the results from the empirical da ta determining the causal sequences among relationship maintenance strategies, relationshi p quality outcomes, a ttitude, and behavior can provide a more comprehens ive model of public relations effectiveness measurement. Criticized for their inapplicabil ity to practice, public relatio ns scholarship theories have overemphasized image and perception. However, the proposed comprehensive model can be used by public relations professionals for evaluation.

PAGE 29

11 The remainder of this dissertation is divide d into four chapters. In Chapter 2, the literature review, consis ts of three parts: relationship maintenance strate gies, relationship quality outcomes, and relevant research regarding the proposed models. Chapter 3, methodology, covers the data collection proce ss including the pilot test and mail survey, while Chapter 4 presents the results of the co llected data analysis. Chapter 5 discusses the theoretical and practical implications of the scales and proposed models, limitations of the scales and models, and suggestions for future research.

PAGE 30

12 CHAPTER 2 LITERATURE REVIEW In this chapter, the aforementioned centr al concepts (e.g., relationship maintenance strategies, relationship quality outcomes, attitude, and behavior) are clarified. A general overview of relationship management will be provided including diverse definitions of organization-public relationships before explai ning each concept. In addition, indicators of these constructs will be discussed. For example, the six indicators—access, positivity, openness, sharing of tasks, ne tworking, and assurances—will be explained in the section covering relationship maintenan ce strategies. The section an alyzing relationship quality outcomes explain the four indicators—cont rol mutuality, satisfaction, trust, and commitment—followed by a review of attitude and behavior. Finally, previous research detailing the different sequential orders among these variables is explained at the end of this chapter. Public Relations as Re lationship Management The focus on relationship management in publ ic relations research dates back to Ferguson’s (1984) assertion that the relationship between an organization and its strategic publics should be the central unit of analysis for public relations scholarship and its practice. Many public relations scholars have since adopted this perspective. Ehling (1992) contended that this relationship pers pective is a marked departure from the original theory and practice of public relations, which wa s the manipulation of public opinion, to the cultivation of relationships. Practitioners often use the language of relationship building in their definitions of public relations effectiveness (Hon, 1997),

PAGE 31

13 and many scholars in public relations have en thusiastically embraced this shift (Broom, Casey, & Ritchey, 2000; Bruning & Ledi ngham, 1999, 2000; Coombs, 2000; Cutlip, Center, & Broom, 2000; J. E. Grunig, 1993; J. E. Grunig & Huang, 2000; Heath, 2001; Ledingham & Bruning, 1998; Leichty & Springston, 1993; Wilson, 2001). Likewise, Dozier, L. A. Gr unig, and J. E. Grunig (1995) emphasized that “the purpose and direction of an organizati on (its mission) is affected by [its] relationship with key constituents (publics) in the orga nization’s environment” (p. 85). Broom and Dozier (1990) proposed that this relational pers pective changes th e value of public relations initiatives from measures of comm unication output to measures of behavioral outcomes, therefore providing a basis for ev aluating public relations effectiveness. Definitions of Organizatio n-Public Relationships One objective of the current study is to m easure relationship maintenance strategies and refine existing measures of relationship quality outcomes. To develop reliable and valid measures, this study adopted the multiple-item measurement suggested by Spector (1992).1 Specifying the domain of the construct is the first step in the suggested procedure for developing solid measures. In this stage, the researcher must specify the definition in terms of what is included and what is excluded. Therefore, the first act in developing multiple-item measures of orga nization-public relati onships is defining conceptual specifications for organization-pub lic relationships. A consistent and clear definition of the relationsh ip helps to develop a valid operational measure of organization-public relationships. 1 Each procedure is explained in next chapter in detail.

PAGE 32

14 Analyzing academic research articles about organization-pub lic relationships published between 1984 and 2004, Ki and Shin (2006) discovered that only one in four articles contained a definition of organizati on-public relationships and that definitions varied from scholar to scholar. Review ing interpersonal, interorganizational relationships, psychotherapy, and systems th eory, Broom, Casey, and Ritchey (1997) concluded: The absence of a useful definition [of relationship ] precludes measurement of organization-public relationships and forces both scholars and practitioners alike to measure one part or another and make pot entially invalid inferences about the relationship. The absence of a fully explicated conceptual definition of organization-public relationships limits th eory building in pub lic relations. (p. 96) After Broom, Casey, and Ritchey’s (1997) as sertion, public relati ons scholars have developed a variety of defin itions of organization-public re lationships. Berko, Rosenfeld, and Samovar (1997) defined a relationship as: The connection that exists when (1) the interactants are awar e of each other and take each other into account, (2) there is some exchange of influence, and (3) there is some agreement about what the nature of the relationship is and what the appropriate behaviors are given the na ture of the relationship. (p. 448) Hung (2005) noted that “organization-public relationships arise when organizations and their strategic publics ar e interdependent and this interdependence results in consequences to each other that organizations need to change” (p. 396). This definition focuses on interaction, interdependence, and in fluences both sides of the relationship. Ledingham and Bruning (1998) de fined organization-public rela tionships as “the state which exists between an organization and its key publics, in which th e actions of either can impact the economic, social, cultural or poli tical well being of the other” (p.62). This definition highlights the diverse impact that both parties can exert on a relationship. Thomlison (2000) suggested another definition, stating that a relationship is “a set of

PAGE 33

15 expectations two parties ha ve for each other’s behavior based on their interaction patterns” (p. 178). Furthermore, Broo m, Casey, and Ritchey (2000) defined organization-public relationships as follows: Organization-public relationships are repr esented by the patterns of interaction, transaction, exchange, and linkage between an organization and its publics. These relationships have properties that are distinct from the identities, attributes, and perceptions of the individuals and social collectivities in the relationships. Though dynamic in nature, organization-public rela tionships can be de scribed at a single point in time and tracked over time. (Broom et al., 2000, p. 18) As the different definitions of organization-public relationships illustrate, scholars emphasized interdependence, interaction, a nd impact between two parties or more. Looking at the existing definitions, this study defines the organization-public relationship as “the state in which each party relies on th e other party’s resources and each party is affected by the other.” The focus of this st udy, however, is relationship quality outcomes rather than relationship itself . Relationship quality outcomes ar e defined as “factors that determine or characterize successful rela tionships between an organization and its strategic publics.” Stages of Developing Relationships Although definitions of organization-public relationships differ slightly, scholars seem to agree that such relationships cons ists of three stages—antecedents, relationship maintenance strategies, and re lationship quality outcomes. Applying Broom, Casey, and Ritchey (1997) model, J. E. Grunig a nd Huang (2000) extended and developed a conceptual framework describing these stag es of developing relationships—antecedents, maintenance strategies, and relationship quality outcomes. Figure 2-1 shows the components of the three stage model. Figure 2-2 represents a simplified version of the model of public relationships.

PAGE 34

16 Situational Antecedents Organization affects public (O1 P1) Public affects organization (P1 O1) Organization-public coalition affects another organization (O1 P1 O2) Organization-public coalition affects another public (O1 P1 P2) Organization affects an organizationpublic coalition (O1 O2P2) Multiple organizations affect multiple publics (Oi Pi) Maintenance Strategies -Symmetrical Access Positivity Openness (disclosure) Sharing of tasks Networking Assurances Relationship Outcomes Control Mutuality (Joint acceptance of degree of symmetry) Commitment (Interdependence, loss of some autonomy) Satisfaction Trust Figure 2-1. Stages and forms of relationships (Adapted from J. E. Grunig & Huang, 2000, p. 34).

PAGE 35

17 Antecedents of Relationships Relationship Maintenance Strategies Relationship Quality Outcome Figure 2-2. Simplified version of st ages and forms of relationships Antecedent The antecedents to relationships are gene rally defined as “sources of change, pressure, or tension on the sy stem derived from the environment” (Broom et al., 1997, p. 94). At the antecedent level, an organi zation should identify with which publics the organization needs to develop a relationship. To identify an organization’s strategic publics, public relations practit ioners should examine their si tuations and develop public relations program objectives to communicate with the identified publics. Publics are groups of individuals who are formally or in formally impacted by an organization or vice versa. Applying situational th eory, J. E. Grunig and Hunt (1984) categorized publics into three groups: 1) latent public (a group that does not recognize a situ ation as problematic), 2) aware public (a group that has moved from a latent stag e and recognizes the problem), and 3) active public (a group th at is active and organizes to discuss and do something about the situation). The other two stages—relationship maintena nce strategies and relationship quality outcomes—are considered process and out come objectives respectively. Process objectives are strategies used to cultivate and develop relati onships and they result in outcome objectives, such as relationship quality.

PAGE 36

18 Part I: Relationship Maintenance Strategies Scholars measure the relationships betw een an organization and its strategic constituencies since relationship quality outcomes are the most meaningful means of evaluating public relations effectiveness. Se veral studies show that organizations that develop positive, long-term relationships with their publics are more effective and achieve their organizational goa ls (Dozier et al., 1995; L. A. Grunig et al., 2002; Hon, 1997; Huang, 1999). Furthermore, a positive relational outcome is dependent on an organization’s effort to cultivate and maintain positive relationships. However, how and to what extent communication activi ties produce positive organization-public relationships has not been investigated. Public relations scholars have identif ied various relationship maintenance strategies that are more likel y to produce positive relations hip quality outcomes. Some studies focus on relationship maintenance stra tegies in online environments. The first such study by Ki and Hon (2006) explored how corporations utilized relationship maintenance strategies—access, positivity, open ness, sharing of tasks, and networking— through their Web sites. They found that openness and access ar e the most commonly used relationship maintenance strategies among organizational Web sites. In another study that applied relational maintenance strategies to new technology, Kelleher and Miller (2006) concentrated on organizational blogs as a communication channel between an organization and its strategic publics. They discovered that perceived relational maintenance strategies, especially co nversational human voice and communicated relational commitment, were significantly associ ated with relationship quality outcomes. However, these studies are limited to online communication. So far, no study has

PAGE 37

19 empirically examined how relationship mainte nance strategies can be measured and how they affect relational quality outcomes. Public relations scholars adopted several relationship maintenance strategies from research on interpersonal relationships, such as those between husbands and wives or friendships between two indivi duals, and applied them to public relations studies. The main categories of relationship maintenan ce strategies derived from interpersonal communication were identified as (a) positivity (interacting with partners in a cheerful, uncritical manner); (b) openness (directly discussing the nature of the relationship and disclosing one’s desires for the relationship); (c) sharing of tasks (performing one’s responsibilities such as household chores); (d) social networks (relying on common affiliations and relations); and (e) assurances (communicating one’s desire to continue the relationship) (Canary & Stafford, 1992). Various scholars proposed that conflict re solution theories would be useful to developing other relationship maintenance strategies (Grunig, J. E. & Huang, 2000; Huang, 1997). Huang’s disserta tion (1997) examined conflict management strategies (i.e., integrative,2 distributive,3 and avoidance/non-confrontational strategy4) and their 2 Putman (1990) explained that integrative strategy “aims to reconcile the interests of both parties, reach joint benefits, or attain ‘win-win’ goal s through open information exchange and joint decision making” (p. 3). 3 Contrary to integrative strategy, distributive stra tegy aims to solve pure conflicts of interest. It is defined as “a fixed-sum feature by efforts to maximize gains and minimize losses within a ‘win-lose’ or self-gain orientation” (Putman, p. 3). 4 Avoidance or non-confrontational strategy discour ages the expression of conflict. This strategy includes the following behaviors such as refusing existence of conflict, avoiding disagreement (Putman & Wilson, 1982), and restricting discu ssion of conflict with an opponent (Morrill & Thomas, 1992).

PAGE 38

20 association with relati onship quality outcomes.5 Her study assumed that positive, stable organization-public relationshi ps decreased conflicts betw een an organization and its strategic publics. Huang discovered that control mutuality and trust reduced the likelihood of conflicts. However, conflict mana gement strategies and sub-strategies are not the focus of this study. Hon and J. E. Grunig (1999) suggested that the relationship maintenance strategies used in interper sonal relationships can be a pplied to organization-public relationships by changing the focus of comm unication strategies from individuals to publics. These maintenance strategies are more likely to affect relationship quality outcomes, such as control mutuality, satisfac tion, trust, and commitment (J. E. Grunig & Huang, 2000; Hon & J. E. Grunig, 1999). A lthough assumptions about the linkages between relationship maintenance strategies and relationship quality outcomes have been posited, no empirical research tests how and which identified relationship maintenance strategies affect relationship quality outco mes. Furthermore, no research has been conducted to determine whether these suggest ed strategies are e qually effective at producing positive relational outcomes. Before exploring the links between re lationship maintenance strategies and relationship quality outcomes, it is necessary to develop measurement scales for the maintenance strategies. As Spector (1992) proposed, the first step in developing multiple-item measurement scales is to defi ne the construct. Also, definitions are fundamental since they delineate the na ture of, and guide boundaries for, the 5 Conflict management strategies and their associa tion with relational outcomes are explained in detail in the following reference. Huang, Y.-H. (1997). Public relations strategies, relational outcomes, and conflict management strategies. Unpublished doctoral dissertation, University of Maryland, College Park, MD.

PAGE 39

21 phenomenon under investigation (Canary & Stafford, 1994). This study defines relationship maintenance strategi es as “any organizational beha vioral efforts that attempt to sustain or cultivate relati onships with strategic publics.” This study also posits that positive evaluations of these strategies among strategic publics lead to better relational outcomes. In interpersonal relationships, relation ship maintenance efforts between two individuals (either romantic c ouples or friendships) were found to be a reliable predictor of some relationship characteris tics. First, relational maintenance efforts were primarily related to relational stabi lity, satisfaction, and commitmen t (Dindia & Canary, 1993). Second, relationship maintenance involves not on ly a stage of relati onal development, but the dynamic processes involved in re lating (Canary & Stafford, 1994). Maintaining and cultivating organizati on-public relationships is a goal for organizations that desire l ong-term, stable, and satisfying relationships with their key publics. Without relationship maintena nce, the desired re lationship between organizations and their key publics could dissolve. Hon and J. E. Grunig (1999) highlighted the necessity of relationship ma intenance in public relations conveying that most public relations professi onals retain knowledge that ha s “something to do with how to communicate with publics, in order to ma intain a relationship with those publics” (p. 13). Relationship maintenance strategies in or ganization-public relationships have been derived from interpersonal communications since public relations studies suggest that interpersonal relationship pers pectives can be applied to public relations (Thomlison, 2000; Toth, 2000; Wood, 1995). Both in terpersonal and or ganization-public

PAGE 40

22 relationships share several key relational featur es such as trust (L. A. Grunig et al., 1992), control mutuality (Burgoon & Hale, 1984, 1987; Canary & Spitznerg, 1989; Canary & Stafford, 1992; Ferguson, 1984; Stafford & Canary, 1991), and commitment (Aldrich, 1975, 1979; Burgoon & Hale, 1984, 1987; Canary & Spitzberg, 1989; Canary & Stafford, 1992). Interpersonal communication scholars res earched how different communication strategies influence the devel opment, maintenance, and dissolution of relationships, or the effects of communication strategies on re lational outcomes. For example, Stafford and Canary (1991) investigated what kinds of communication strategies romantic couples use to maintain their relationships. Th ey discovered the following five relational maintenance strategies: (a) positivity (such as attempts to make the relationship enjoyable for both), (b) openness (such as disclosure of thoughts and feelings), (c) assurances (of love and commitment), (d) networking (havi ng common friends), and (e) shared tasks (taking joint responsibility for household task s). They found that these five relational maintenance strategies helped establish a pos itive relationship and fo ster key relationship outcomes (Stafford & Canary, 1991). Several studies in the field of interpersona l relationship studies prove that relational strategies are associat ed with perceptions of equitable relationships (Canary & Stafford, 1992). In public relations, th e ideal relationship between an organization and its publics is perceived as fair and equ itable by both parties. Likewise , scholars proposed that the five maintenance strategies used in interp ersonal relationships ar e analogous to public relations strategies in the two-way symmetri cal model of public rela tions (J. E. Grunig, 1989; J. E. Grunig & L. A. Grunig, 1992; J. E. Grunig & White, 1992; J. E. Grunig &

PAGE 41

23 Huang, 2000; Ki, 2003; Ki & Hon, 2006). Scholars have regarded these strategies as the most effective for maintaining and fosteri ng positive, stable relationships in public relations (J. E. Grunig & Huang, 2000; Hon & J. E. Grunig, 1999). These relationship maintenance strategies are considered process indicators that are useful as communication strategies for producing desi rable relationship quality outcomes. In this study, relationship ma intenance strategies are considered precursors used to maintain or cultivate relationships between an organization and its publics. These are suggested as the most effectiv e strategies for developing qua lity relationships. The six strategies adopted from Hon and J. E. Grunig (1999) are discussed in detail. Access Access is not included in interpersonal re lationships literature. Hon and J. E. Grunig (1999) suggested access as one relations hip maintenance strategy and identified access as follows: Members of publics or opinion leaders provi de access to public relations people. Public relations representa tives or senior managers provide representatives of publics similar access to organizational deci sion-making processes. Either party will answer telephone calls or read letters or e-mail me ssages from the other. Either party is willing to go to the othe r when they have complaints or queries, rather than taking negative reactions to third parties. (Hon & J. E. Grunig, 1999, p. 14) Access is a strategy that a party (either a public or an organization) uses to reach the other party and express or share their opin ions and thoughts. Since the focus of this study is what an organization does to maintain and sustain a relationshi p with its strategic public, this study only deals with maintenance strategies of the organization. Based on Hon and J. E. Grunig’s guidelines, this study de fines access as “the degree of effort that an organization puts into provi ding communication channels or media outlets that assist its strategic publics in reaching it.”

PAGE 42

24 Positivity Interpersonal communication studies defi ne positivity as “attempts to make interactions pleasant” (Canary & Stafford, 1994, p. 15). In Canary and Stafford’s study, positivity consisted of cheerful and nice be havior, courteous and polite communication, and uncritical behaviors toward partners (Canary & Stafford, 1994). Positivity has consistently been revealed as an essentia l predictor of control mutuality (Canary & Stafford, 1992, 1993; Dainton, 1991; Stafford & Canary, 1991). In addition, positivity was found to be a proactive strategy in c onstructive maintenance action (Guerrero, Eloy, & Wabnik, 1993). Hon and J. E. Grunig (1999) adopted positivity from interpersonal communications and applied it to public relations. They conceptualized positivity as “anything the organization or publics do to make the re lationship more enjoyable for the parties involved” (p. 14). Positivity is equitable to the “Be Unconditionally Constructive” principle for developing re lationships espoused by Fisher and Brown (1988). Research on interpersonal relations consistently finds that maintenance strategies are correlated with relati onal outcomes (Canary & Stafford, 1992, 1993; Stafford & Canary, 1991). First, positivity is identified as an essential predictor of control mutuality (Canary & Stafford, 1992, 1993; Stafford & Canary, 1991) since communicating in a cheerful, courteous, and polite manner is more likely to encourage cooperation from the other party and help preser ve interdependence in a re lationship. In addition, an interpersonal relationship study revealed that perceptions of one’s positivity influence relational satisfacti on (Dindia, 1989). Since this study regards re lationship maintenance stra tegies as organizational behavioral efforts designed to result in positive relationship quality outcomes, positivity

PAGE 43

25 is defined as “the degree to which member s of publics benefit from the organization’s efforts to make the relationship more enjoyable for key publics.” Openness/Disclosure Openness, which is also called disclosu re, is a concept studied actively by researchers of interpersonal communications. Openness wa s operationalized as “direct discussion about the nature of the relationshi p and setting aside times for talks about the relationship” (Canary & Sta fford, 1994, p. 12). Examples of openness in interpersonal communication are: “We share things with each other that no one else knows,” “We discuss our problems in the rela tionship,” “We often talk about how things used to be,” “I try to provide advice through past experience,” and “I rely on her for advice” (p. 15). Guerrero, Eloy, and Wabnik (1993) found that openness represents proactive and constructive maintenance act ions like positivity. Openness in public relations involves disclosing “thoughts and feelings among parties” in a relationship (Hon & J. E. Gr unig, 1999, p. 14). More specifically, openness is the condition in which both organizations and publics are open and honest with each other and more than willing to share their opi nions about how they think, what concerns or problems they have, and how satisfied or dissatisfied they are w ith each other (L. A. Grunig et al., 2002). Furthermore, openness is a vital component of trust (Dimmick, Bell, Burgiss, & Ragsdale, 2000). J. E. Grunig a nd Huang (2000) further explained that this strategy corresponds with the symmetrical mode l and that it leads to positive relational outcomes. Openness can be helpful to equall y distributing power in a relationship (Bok, 1989). J. E. Grunig and Huang (2000) suggested that monitoring openness can be an effective gauge of relationship quality. In a detailed example based on J. E. Grunig and

PAGE 44

26 Hunt’s (1984) study, AT&T community l eader teams under evaluation in the 1970s, tended to interact more openly with effective teams than with less effective teams. In addition, J. E. Grunig and Hua ng (2000) suggested that public relations managers could quantify openness through evaluating sugges tions, complaints, inquiries, and other contacts from members of publics, the media, government, or leaders of activist groups registered with the organization. In an organization-public relationship, schol ars have continuously revealed that openness is an important indicator of rela tionship quality outcomes. Ledingham and Bruning (1998), for instance, asserted that ope nness leads to satisfyi ng relationships. L. A. Grunig, J. E. Grunig, and Ehling (1992) argued that openness is an essential dimension for evaluating relationship quali ty with an organization’s strategic constituencies. This study defines openness as “an organization’s efforts to provide information about the nature of th e organization and what it is doing.” Sharing of Tasks Studies on interpersonal rela tionships have examined the sharing of tasks between couples who share household duties and perf orm shared responsibilities (Canary & Stafford, 1994; Stafford & Canary, 1991). Fu rthermore, studies have shown that an example of this strategy is the mutual perfor mance of routine tasks and chores (Canary & Stafford, 1994). Sharing of tasks is a consis tent and important pr edictor of relational characteristics such as control mutuality, commitment, liking, and satisfaction. It significantly explains control mutuality and liking (Canary & Stafford, 1994) as well as commitment and satisfaction (Stafford & Ca nary, 1991) in relationships between two individuals. Other studies also have emphasized the impor tance of sharing tasks for relational satisfaction (Huston, McHale, & Cr outer, 1986; Wilmot & Sillars, 1989). As

PAGE 45

27 Guerrero, Eloy, and Wabnik (1993) has suggest ed, the sharing of tasks represents a proactive and constructive maintenance strategy. This strategy can be applied to the pub lic relations environment. For example, publics and organizations could share ta sks such as reducing pollution, providing employment opportunities, making a profit, and staying in business, in the interests of either the organization, publics, or both (J . E. Grunig & Huang, 2000). This relationship maintenance strategy is conceptualized as “o rganizations’ and publics’ sharing in solving joint or separate problems” (Hon & J. E. Grunig, 1999, p. 15). Sharing of tasks could be assessed through social responsibility reports which explain the degree to which an organization has made an effort to share proble ms or issues of interest with publics. By adopting Hon and J. E. Grunig’s conceptualiz ation, this current st udy defines sharing of tasks as an “organization’s efforts to share in working on projects or solving problems of mutual interest between an organization and it s publics.” Thus, shar ing of tasks entails the organization’s willingness to carry on its and its publics’ responsibilities so that both may achieve their interdependent goals. Networking The term network or networking is commonly used for the structure of ties between actors in a social system. The actors can be individuals, organizations , industries, or even nation states (Nohria & Eccl es, 1992). Networking can be formed through conversation, friendship, kinship, authority, economic excha nge, information exchange, or anything that builds the basis of a relationship (N ohria & Eccles, 1992). In interpersonal relationships, networking means the practice of spending time w ith mutual friends to gain support and make the relationship enjoyable (Canary & Stafford, 1994). Networking is performed through shared explanation, contro l mutuality, and liking (Canary & Stafford,

PAGE 46

28 1994). Moreover, Guerrero, Eloy, and Wabni k (1993) has suggested that networking should be proactive and nurture cons tructive maintenance behaviors. Expanding this idea to organization-public relationships, organizations can build networks with the same groups that netw ork with their publics such as unions, community groups, activist groups, and envi ronmentalists (Hon & J. E. Grunig, 1999). And as J. E. Grunig and Huang (2000) ha ve suggested, one way networking can be evaluated is through contacts with networks of activist groups. Adopting Hon and J. E. Grunig’s (1999) definition, this study defi nes networking as “the degree of an organization’s effort to build networks or coalitions with the sa me groups that their publics do, such as environmentalis ts, unions, or community groups.” Assurances In the interpersonal relationship literature , assurances are behaviors that imply an expression of love between tw o individuals (Canary & Staffo rd, 1994). In the literature on organization-public relationships, assu rances occur when “each party in the relationship attempts to assure the other that it and its concerns are legitimate and to demonstrate that it is committed to maintaini ng the relationship” (L. A. Grunig et al., 2002). Using assurances as a relationship maintenance strategy, an organization can reinforce how it values its strategi c publics (Hon & J. E. Grunig, 1999). Assurances as a maintenance strategy ha ve been found to be most effective in nurturing commitment between two individuals (Canary & Stafford, 1992, 1993; Stafford & Canary, 1991). For example, Canary a nd Stafford (1992) studied the association between relationship maintenance strategies a nd relational outcomes in married couples. They found that the perception of assurances most affected the commitment of both the husbands and wives. In thei r studies, assurances conti nuously appeared as a strong

PAGE 47

29 predictor of trust in interp ersonal relationships (Canary & Stafford, 1993; Stafford & Canary, 1991). This study defines assurances as “any efforts by an organization to assure its strategic publics that they and their concerns are attended to.” The measurement of relationship maintenance strategies and their impact on relationship quality outcomes in the field of public relations has been under-explored. Based on the exploratory nature of this st udy, the following two research questions were posed (Figure 2-3 and 2-4 visual ize the research questions): Research Question 1: How can relationshi p maintenance strategies be measured? Research Question 2: To which and what extent is relationship maintenance strategies positively connected with which relationship quality outcomes? Figure 2-3 represents the measurement model of indicators for relationship maintenance strategies—access, positivity, op enness, sharing of tasks, networking, and assurances. Each strategy indicator has multiple items. Figure 2-4 illustrates a conceptual model consisting of multiple indicators of relationship maintenance strate gies and relational outcomes. As explained above, this study examines how the relationship maintenan ce strategies an organization uses affect relational outcomes. Part II: Relationship Quality Outcomes To develop a system of measurement fo r the organization-public relationship, numerous public relations schol ars have adopted perspectives from other disciplines, including the studies of interpersonal and interorganizational re lationships, psychology, psychotherapy, and relationship marketing (e .g., J. E. Grunig & Huang, 2000; Hon & J. E. Grunig, 1999; Huang, 1997, 2001b; Jo, 2003, 2006; Kim, 2001; Ledingham & Bruning, 1998, 2000; Ledingham, Bruning, & Wilson, 1999).

PAGE 48

30 Figure 2-3. Measurement model of re lationship maintenance strategies

PAGE 49

31 Assurances Sharing of Tasks Networking Openness Positivity Access Control Mutuality Satisfaction Trust Commitment Figure 2-4. An initial conceptual model li nking relationship maintenance strategies and relationship quality outcomes After exhaustively reviewing interpersonal li terature, Ferguson (1984) proposed that there are five attributes to a relationship: (1) dynamic versus stat ic; (2) open versus closed; (3) the degree to whic h both the organization and the public are satisfied with the relationship; (4) power distribution in the re lationship; and (5) and mutual understanding, agreement, and consensus. Ledingham, Bruning, Thomlison, and Lesko (1997) suggested openness, trust, involvement, inve stment, and commitment as the dimensions of an organization-public relationship. In a later study, Ledingham and Bruning (1998)

PAGE 50

32 researched the relationship dimensions upon which positive organization-public relationships are initiated, de veloped, and maintained and they discovered that there were five relationship dimensions—trust, openness, involvement, investment, and commitment—different among three group types—st ayers, leavers, and undecideds. The results of the study determined that relations hip dimensions impact a public’s perceptions of overall satisfaction with the organizati on, and they can be more influential in predicting customer behavior than price or product features (Ledingham & Bruning, 1998). Hon and J. E. Grunig (1999) developed measurement indices to evaluate the success of relationship building efforts through reviewing the literatur es on interpersonal relationship and psychology. They concl uded that the organization’s long-term relationships with its key publics could be ev aluated by the following four indicators of relationship quality outcomes: control mutua lity, satisfaction, trust, and commitment. Several scholars have used Hon and Grunig’ s measurement scales and confirmed their reliability and validity (e.g., Huang, 2001b; Jo, 2003, 2006; Kim, 2001). Huang (2001b) and Jo (2006) refined the relationship measur ement scales and improved their reliability and validity. Huang (2001b) measured the re lationship between legislative members and their assistants using Hon and J. E. Gruni g’s relationship measurement index, and she added face and favor as a specific measure of the relationship in Taiwanese culture. Jo (2006) also tested the validity of this or ganization-public relationship measurement and determined that satisfaction, trust, and comm itment are a “global measure” that can be applied to organization-public relationships in any situation or in any environment, and that personal network was a “specific measure” in South Korean culture (p. 151).

PAGE 51

33 Hon and J. E. Grunig (1999) also sugge sted two types of relationships—the exchange relationship and the communal rela tionship—along with four relationship quality outcome indicators. Relationship type s explain what an organization attempts to achieve using public relations programs. According to Hon and J. E. Grunig (1999), in an exchange relationship “one party gives benefits to the ot her only because the other has provided benefits in the past or is expected to do so in the fu ture” (p. 20). In an exchange relationship, the party that gi ves benefits to the other is most likely to get equitable benefits from the other. In a communal rela tionship, parties give benefits to one another since they care about mutual welfare, even if they do not expect to gain anything in return. Based on these existing two relati onship types, Hung (2005) developed several new types of relationships in a study on mu ltinational companies and their publics in China. The new types of relationship she di scovered were mutual communal, covenantal (win-win), contractual, symbiotic, manipulativ e, and exploitive relationships. She also found that several relationship type s can exist at the same time. COVENANTAL (WIN-WIN) RELATIONSHIP. Both parties in the relationship are dedicated to the welfare of each other through their open exchanges and the norm of reciprocity. One party always gives the other an opportunity to “ask for insight, to provide criticism, and to place a clai m upon some of the individual’s time” and “the obligation of the other side is always to listen and provide responses (Benett, 2001, p.89). CONTRACTUAL RELATIONSHIPS. Parties have an agreement about each party’s obligations. It is similar to writing a cont ract at the beginning of a relationship. However, this type of relationship cannot guarantee equal relationships (Hung, 2005). SYMBIOTIC RELATIONSHIPS. As a cooperative relationship, this relationship occurs when organizations are interdependent in a given situation and they work together with a certain public that shares a common interest in order to survive in the situation (Hung, 2005).

PAGE 52

34 MANIPULATIVE RELATIONSHIPS. An organization knows what a public wants and uses asymmetrical or fake-symmetrical a pproaches in communication with a public for its own interest (Hung, 2005). EXPLOITIVE RELATIONSHIPS. One party takes advantage of the other and one does not carry out his/her obligations in the ex change relationship (C lark & Mills, 1993). Relationship Quality Outcome Dimensions for This Study Four relationship indicators culled from the existing literature served as the basic structure from which items were derived during the first stage of developing the organization-public relationships measur ement scale. The four measurement indicators—control mutuality, satisfaction, tr ust, and commitment—were conceptualized as the essence of organization-public relationships. There are several reasons why the four indicators were chosen. First, these indicators have been consid ered key relational features and have appeared as important indicators for divers e disciplines including interpersonal and interorganizational relationships, relationship mark eting, and organization-public relationships. The indicators are contro l mutuality (Burgoon & Hale, 1984, 1987; Canary & Spitzberg, 1989; Canary & Stafford, 1992; Ferguson, 1984; Stafford & Canary, 1991), satisfaction (Ferguson, 1984; L. A. Grunig et al., 1992; Millar & Rogers, 1976; Stafford & Canary, 1991), trust (L. A. Grunig et al., 1992; Morgan & Hunt, 1994), and commitment (Aldrich, 1975, 1979; Burgoon & Hale, 1984, 1987; Canary & Spitznerg, 1989; Canary & Stafford, 1992; Morgan & Hunt, 1994). Additionally, three of the indicators—satisfaction, trust, and commitm ent—have been shown to be critical relationship indicators even in cross-cultural settings su ch as in Taiwan (Huang, 1997; 2001b) and South Korea (Jo, 2006) . Jo (2006) called these three critical relational dimensions a global measure for or ganization-public relationships.

PAGE 53

35 Secondly, the four key relational features typically represen t the essence of organization-public relations hips (J. E. Grunig & Hua ng, 2000; Hon & J. E. Grunig, 1999; Huang, 2001b; Jo, 2003, 2006; Kim, 2001) and several studies have demonstrated them to be valid and reliable (Huang, 1997; Jo, 2003, 2006; Kim, 2001). Control Mutuality Stafford and Canary (1991) defined cont rol mutuality as “the degree to which partners agree about which of them shoul d decide relational goals and behavioral routines” (p. 224). Control mutuality is re lated to the decision making process and the extent to which the opinion of each party is reflected in the final decision. The sense of control mutuality between the parties invol ved in a relationship is significant to interdependence and relational stability (Stafford & Canary , 1991). Therefore, control mutuality could constructively lead the oppos ing public to search for creative and mutually beneficial solutions or to seek assistance from a thir d party to resolve a conflict. For a stable and positiv e relationship, control mutuality among the parties should exist to some degree. The concept of control mutuality is sim ilar to other concepts proposed as being essential to positive relati onships: mutual legitimacy (Bruning & Ledingham, 1999), reciprocity (Aldrich, 1975, 1979), empowerment (Moore, 1986), and power distribution (Ferguson, 1984). Regarding power di stribution, Ferguson (1984) stated: Other variables related to the relationship might be how much control both parties to the relationship believe th ey have, how power is dist ributed in the relationship, whether the parties to the relationship belie ve they share goals, and whether there is mutuality of understanding, agreement, and consensus. (p. 20) Huang (1999) found that control mutuality is one of the two major variables mediating the effects of public relations strategies on conf lict resolution (the other is

PAGE 54

36 trust). She proposed that symmetrical, or two-way communication, is an antecedent of control mutuality in the re lationship (quoted in Huang, 2001b). Thus, the concept of control mutuality is pertinent to ex cellent public relations practices. This study adopts the concep tualization of Hon and J. E. Grunig (1999). They defined control mutuality as “the degree to which parties agree on who has the rightful power to influence one anothe r” (p. 3). They acknowledged that some degree of power imbalance might naturally exist, but, in a stable relationship between publics and an organization, both parties need some extent of control over th e other (Hon & J. E. Grunig, 1999). Satisfaction Although researchers in relati onship studies have consider ed satisfaction a complex concept to measure, it is a commonly accepte d indicator for evaluating relationships (Ferguson, 1984; J. E. Grunig & Huang, 2000; L. A. Grunig et al., 1992; Hon & J. E. Grunig, 1999; Huang, 2000, 2001a, 2001b; Lewis & Spanier, 1979; Stafford & Canary, 1991). In public relations scholar ship, satisfaction is referred to as “the extent to which each party feels favorably toward the othe r because positive exp ectations about the relationship are reinforced” (Hon & J. E. Grunig, 1999, p. 3). Satisfaction occurs when one party perceives that the other behaves to maintain a positive relationship (Hosmer, 1996; Miles, Patrick, & King, 1996). From a so cial exchange perspe ctive, satisfaction often increases with rewards received and decreases with costs incurred by the relationship (Kelly & Thibaut, 1978; Jo, Hon, & Brunner, 2004). Satisfaction is typically calculated by the extent to which the benefits of the relationship exceed the expectations that both parties have and a sa tisfying relationship produces more benefits than costs. In sum, satisfaction is based on the extent of th e discrepancy between the expectations in a

PAGE 55

37 relationship and what is actua lly experienced. Adopting Hon and J. E. Grunig’s (1999) conceptualization, this study defi nes satisfaction as “the exte nt to which each party feels favorably toward the other.” Trust Over the last several decades, a variety of disciplines, incl uding interpersonal, interorganizational co mmunication, and relational marketing, have emphasized that trust is one of the main constructs used to meas ure a successful relationship between parties. Rotter (1967), for example, stated the following: One of the most salient fact ors in the effectiveness of our present complex social organization is the willingness of one or mo re individuals in a social unit to trust others. The efficiency, adjustment, and ev en survival of any social group depends upon the presence or absence of such trust. (p. 651) The classic view of trust espoused by Rotter (1967) is “a generalized expectancy held by an individual that the word of another . can be relied on” (cited in Morgan and Hunt, 1994, p. 23). As a parallel with the classic view of trust, Moorman, Deshpand, and Zaltman (1993) defined trust as “a willingness to rely on an exchange partner in whom one has confidence” (p. 82). One of the co mmonly adopted definitions of trust in relational marketing is that it “exist(s) when one party has confiden ce in an exchange partner’s reliability and integr ity” (Morgan & Hunt, 1994, p. 23). Trust also has been revealed as one of th e primary indicators of relationship quality in organization-public rela tionships (e.g., Hon & J. E. Grunig, 1999; Huang, 2001; Jo, 2003, 2006). In theories of economics and of strategic management, trust of publics such as employees, stockholders, customers, medi a, governments, and communities, allows organizations to exist (Vercic & J. E. Gruni g, 1995). Without trust, employees will leave an organization, customers will not purchase products or services the organization

PAGE 56

38 provides, media will run a negative story about the organization, and government will make regulations or policies against the organization. Hon and J. E. Grunig (1999) conceptualized trust as “one party’s level of confidence in and willingness to open oneself to the other party” (p. 3). The natu re of trust is said to include integrity, dependability, and competence (Barney & Hansen, 1994; Carnevale, 1995; Daley & Vasu, 1995; Whitener, Brodt, Korsgaard, & Werner , 1998). Integrity is the belief that an organization is fair and just. Dependability refers to the belief that an organization will keep promises about what it says it will do. De pendability is a key component of trust, as evidenced by Ledingham and Bruning’s (1998) operationalization of trust as an organization “doing what it says it will do.” Competence is defined as the belief that an organization has the ability to do what it sa ys it will do (Hon & J. E. Grunig, 1999). In summary, trust is a belief by publics that an organization is reliable, honest, and stands by its words as well as accomplishes its promised obligations. Using Hon and J. E. Grunig’s definition of trust, this research defines trus t as “the willingness to rely on the other party in whom one has confidence.” Commitment Commitment has been a focal concept in social exchange literature (Blau, 1964; Thibaut & Kelly, 1959), organization and buyer behavior (Becker, 1960; Reichers, 1985), and relational marketing (Berry & Parasuraman, 1991; Morgan & Hunt, 1994). In management literature, commitment is defined as A psychological state genera ted by an individual’s perceptions, beliefs, and emotions which provoke the willingness or intention of devel oping and maintaining a stable and durable relations hip, because the individual want s it or feels that he/she should make it, and which manifests itse lf in a behavior which bears certain obligation. (Sanchez, & Iniesta, 2004, p. 231)

PAGE 57

39 In the organizational behavioral literature , commitment has been shown to lead to such significant outcomes as decreased tu rnover (Porter, Steers, Mowday, & Boulian, 1974), higher motivation (Farrell & Rusbult, 198 1), increased organi zational citizenship behaviors (Williams & Anderson, 1991), and jo b equity (Williams & Hazer, 1986). In public relations scholarship, commitm ent is also a central concept for organization-public relationships and some sc holars have proposed definitions. Hon and J. E. Grunig (1999) conceptualized commitme nt as “the extent to which each party believes and feels that the relationship is worth spending energy to maintain and promote” (p. 3). According to Meyer and Allen (1984), commitment has two underlying dimensions—continuance and affective. Continuance commitment is “commitment to continue a certain line of action” (Meyer & Allen, 1984, p. 373). The other view of commitment, affective commitment, is “an emotional orientation,” which indicates a psychological attachment between people a nd organizations (Buc hanan, 1974; Kanter, 1968; Porter, Crampon, & Smith, 1976; Porter et al., 1974; Sheldon, 1971; Steers, 1977). Stressing corporate social responsibility as evidence of commitment, Ledingham and Bruning (1998) defined commitme nt as “the organization being committed to the welfare of the community.” This study defines commitment as “the beli ef that an ongoing relationship with the other party is important as to warrant maximum efforts at ma intaining it.” This means that the committed party believes the relationshi p is worth working on to ensure that it endures indefinitely. Despite the research done in the aims of creating a reliable and valid measurement scale for the organization-public relations hip (e.g., Huang, 2001b; Jo, 2003, 2006; Ki &

PAGE 58

40 Hon, 2007; Kim, 2001), a scale is still needed fo r more independent research to refine the existing measurements. In response to this necessity, the current st udy attempts to exact a relationship quality outcome measurement us ing multiple-item development procedures. Therefore, another research question is suggested: Research Question 3: What is a reliabl e and valid relationship quality outcome measurement? Although the four indicators comprise the most widely used scales for relationship measurement, studies have not addressed which of the four indicators captures organization-public relationships most . Thus, the following research question is suggested: Research Question 4: Which of the four indicators represents relationship quality outcomes the most? Linkages among the Four Relational Dimensions Although the significance of the key re lationship quality outcomes—control mutuality, satisfaction, trust, and commi tment—has been well acknowledged in the organization-public relationshi p literature, some important questions concerning the interrelationships among these relationship qua lity outcomes have yet to be addressed properly. Some scholars have argued that some of the relationship indicators are actually antecedents of others (e.g., Ki & Hon, 2005, 2007; Jo, 2003, 2006). Looking at relationship theory in other disciplines, thes e scholars have suggest ed that satisfaction might be a predictor of trust and that trust is an antecedent of commitment. One of the objectives of this study is to explore th e possible longitudinal sequence of the key relationship quality outcomes in organization-public relatio nships. Therefore, the following research question was drawn:

PAGE 59

41 Research Question 5: How do relationship quality outcome indicators affect each other? Satisfaction and trust Relationship studies in mark eting have found that satis faction is an important predictor of trust (Anderson & Narus, 1990; Ganesan, 1994; Garb arino & Johnson, 1999; Selnes, 1998). The parties gain positive expe rience (satisfaction) about the relationship as time goes by and they learn to trust each other. Trust might be formed when the relationship between the two parties develops. A member of the public can hardly deny that a certain number of positive experiences with an organization will at least support the development of trust toward that organiza tion. If members of a public have already decided that an organization is able and w illing to fulfill their needs and demands, and the organization is reliable and predictable, then the public is satisfied and will be likely to trust the organization. Based on the literature, the follo wing hypothesis is posed: HP1: The degree of satisfaction will positively influence the degree of trust. Trust and commitment Diverse disciplines i nvestigating relationships have found that trust has a direct positive impact on commitment or is even a major determinant of commitment (i.e., Achrol, 1991; Anderson & Narus, 1990; Ganesan, 1994; Garbarino & Johnson, 1999; Miettil & Mller, 199 0; Morgan & Hunt, 1994). For ex ample, the business-to-business relationship literature has shown that when a firm trusts its supplier, the firm is more committed to staying in the relationshi p (Anderson & Weitz, 1989; Morgan & Hunt, 1994). International research on business relationships has discovered that trust significantly impacts commitment (Friman, G rling, Millett, Matts son, & Johnston, 2002).

PAGE 60

42 Also, research on supply chain relationships found that trust is a critical factor for fostering commitment (Kwon & Suh, 2004). Tr ust has even been called an essential predictor of commitment (i.e., Dyer, 1996). Pe ople are less likely to be committed to a relationship if trust is not developed b ecause commitment involves vulnerability and sacrifice. As Morgan and Hunt (1994) have suggested, parties tend to seek only trustworthy partners because comm itment encompasses vulnerability. If an organization is not perceived to be honest and trustworthy, a public cannot depend on the organization and thus will not co mmit to the relationship. Applying this logic to organization-public relationships, the followi ng hypothesis can be made: HP2: The degree of trust will positively influence the degree of commitment. Satisfaction Trust Commitment HP2 HP1 Figure 2-5. Proposed model for linkage s among relationship quality outcomes Part III: A Comprehensive Model Linkin g Relationship Quality, Attitude, and Behaviors Organization-Public Rela tionships as Perception A relationship is an abstract and elusive c onstruct because of it s intangibility. The majority of literature dealing with defining the construct has shown that a ‘relationship’ is established by individuals’ perceptions . For instance, social re lationships literature has measured relationships based on the “indi vidual’s viewpoint” (Duck, 1973, p. 147). In negotiation literature, relati onship has been defined as “a subjective experience” (Greenhalgh, 1987, p. 237). One interpersonal relationship study me asured relationship

PAGE 61

43 as “a perception of the partic ipants or as a function of that perception, not as a phenomenon independent of their ob servations” (Kenny & Kashy, 1991, p. 6). Like in other disciplines, the construct of re lationship as conceptualized in the public relations scholarship involves the pe rceived quality of an organization-public relationship. It has been demonstrated that relationship quality outcomes have been well captured by the perceptions of the parties invo lved in the relationships (Dougall, 2006; Ki & Hon, 2005, 2007). Along the same lines, Broom, Casey, and Ritchey (1997) indicated that a relationship is built when the partie s “have perceptions” about the relationship (p. 95). For instance, Broom and Dozier (1990) measured organizationpublic relationships based on the perceptions of organizations and key publics to eval uate the degree of agreement and accuracy. Another study that measured organization-public relationships focused on the public’s perceptions about tr ust, openness, involvement, investment, and commitment with an organization (Ledi ngham & Bruning, 1998) and demonstrated that these variables impacted both the public’s loyalty and suppor tive behavioral intentions toward the organization as long as the member s of the public perceived the organization’s relationship building efforts. Based on Huang’s (1997) study, Hon and J. E. Grunig (1999) identified the four relationship dimensions of control mutual ity, satisfaction, trust, and commitment by measuring the public’s perceptions of the relationship. Bruning and Ledingham (2000) also assessed organization-public relationships by measuring the public’s perceptions of personal,6 professional,7 and community8 relationships. They revealed that the 6 Personal relationship how a member of the public is treated as an individual (Bruning & Ledingham, 2000).

PAGE 62

44 perceptions members of the public had about their personal and prof essional relationship with the organization strongly influenced th e public’s assessment of overall satisfaction with the organization. Additionally, popul arly employed relationship measurement scales have their roots in interpersonal comm unication studies. Most scholars in the field of interpersonal relations con ceptualized relationships either as individuals’ perceptions or a function of their perceptions (Ki & H on, 2007). Thus, it is logical to make an assumption that relationships can be measured by the perceptions held by the members of the public. Attitude Attitude is one of the most extensively used concepts in social science research and has been defined in various but similar ways for several decades. In the early s, Rosenberg and Hovland (1963) conceptualized attitude as “predispositions to respond in a particular way toward a specified class of objects” (p. 1). One of the most popularly used definitions of attitude is “a learned predisposition to respond in a consistently favorable or unfavorable manner with resp ect to a given object” (Fishbein & Ajzen, 1975, p. 6). Mitchell and Olson (1981) demons trated that attitude is “an individual’s internal evaluation of an object” (p. 318). Scholars in psychology described several char acteristics and featur es of attitude. For instance, Fishbein and Ajzen (1975) illustrated th e characteristics of attitude as “a learned predisposition to res pond in a consistently favorable or unfavorable manner with 7 Professional relationship how good of a job an organization performs for a public in providing services or products (Bruning & Ledingham, 2000). 8 Community relationship how an organization is thought of as a corporate citizen or a member of the community (Bruning & Ledingham, 2000).

PAGE 63

45 respect to a given object” (p. 6). This descri ption explains the three fundamental features of attitude: 1) the notion that attitude is lear ned; 2) attitude predisposes action (behavior); and 3) such actions are consistently favorable or unfavorable toward the object (Fishbein & Ajzen, 1975). Attitude is us ually viewed as a latent or underlying variable that is assumed to guide or influence be havior (Fishbein & Ajzen, 1975). In public relations, greater attention ha s been given to evaluating the public’s attitude in order to measure the overall impact or effec tiveness of a public relations program. As Lindenmann (2002) pointed out, attitude research measures not only what the public says about something but also what they feel and how they are inclined to act (their motivational or drive tendencies). Gi ven the importance of a ttitude research in measuring public relations outcomes, the per ceptions of organizati on-public relationships are used here to predict the public’s attitude toward the organization. This research assumes that perceptions precede shifts in attitude because the relationship literature suggests that relationship perceptions are antecedents of suppor tive (or absence of unsupportive) feelings and behaviors among public s toward organizations. In particular, this study defines attitude as “evaluation of an organization by members of a public.” This study assumes the relationship perception that members of a public have impact the way they feel about the organization. Behavior Intentions Behavioral intention is also one of the most popularly used concepts in social behavior research. Scholars have produced va rious definitions of be havioral intention. Behavioral intention is “the immediate determ inant of behavior, and when an appropriate measure of intention is obtained it will prov ide the most accurate prediction of behavior” (Azen & Fishbein, 1980, p. 41). In anothe r definition, Perloff (2003) explained the

PAGE 64

46 conceptualization of behavioral intentions as “the inten tion to perform a particular behavior, a plan to put behavi or into effect” (p. 92). This study adopted behavioral intentions instead of actual behavior for several reasons. First, observing actual behavior is often difficult, particularly at the time that research is conducted, so social scientists fr equently measure behavioral intentions as a predictor of actual behavior (Ki & Hon, 2007). Second, and more importantly, asking people about their behavioral intentions is the most reliable predic tor of behavior, and behavioral intentions tend to be identical to behavior since most so cial behavior is under the individual’s control (Perloff, 2003). Lastl y, behavioral intentions are an intermediate variable between attitude and behavior, acco rding to the theory of reasoned action (Fishbein & Ajzen, 1975) and the theory of planned behavior (Ajzen, 1991). Hierarchy of Effects The current study adopts a hier archy of effects model to explain the sequence of influence among relationship perceptions, attitu de toward the organization, and behavior. The theory explains the long-te rm effect of communication activ ities, which is related to today’s public relations goals which emphasi ze relationship management, especially a stable, long-term relationship with the organi zation’s strategic publics. The effect that the hierarchy of effects theory explains is precisely the type of communication that researchers have considered public relations acti vities to be. Therefore, the theory can be used to measure the effectiveness of pub lic relations depending on each developmental stage. A “hierarchy” refers to “g raded or ranked series” ( Merriam-Webster's collegiate dictionary, 2006, p. 586). As the definition implies, the hierarchy theory explains that communication effects occur th rough a number of stages. This theory can provide a

PAGE 65

47 meaningful description of th e connections between relations hip perception, attitude, and behavior as a series of st ages that members of a public might go through (see Jeffres & Perloff, 1997). While Lavidge and Steiner (1961) esta blished the fundamental theoretical perspective of the hierarchy th eory to measure advertising’ s effectiveness, according to Barry (1987), Palda (1964) was the first to use the term, ‘hierarchy of effects.’9 This theoretical perspective originat es in social learning (Ba ndura, 1986; Lavidge & Steiner, 1961) and the diffusion of innovations theory (Rogers, 1995; Valente, Paredes, & Poppe, 1998), and has functioned as a major theoretic al framework in marketing communication (Ray, 1973; Barry, 1987) and other research areas related to be havioral decision-making. Studies in such mass communi cation areas as advertising (Ray, 1982), public relations, health communication, and political comm unication have actively applied this perspective (Chaffee & Roser, 1986). Three primary classifications—cognition (= perception), affect (=attitude), and behavior—are involved in the hierarchy of eff ects. They are parallel to the three major levels of the typical attitude structure components of cognitive , affective , and conative (Ray, 1973). The cognitive element contains several variables such as attention, awareness, comprehension, and learning. Alth ough some studies cons idered them to be somewhat different variables (e.g., Greeno & Bjork, 1973; Kintsch, 1970), they all have 9 For a detailed description of the historical development of the hierarchy of effects, see Barry, T. E. (1987). The development of the hierarchy of effects: An historical perspective. Current Issues and Research in Advertising, 10 , 251-295. He categorized three development steps of the hierarchy of effects—early development, mode rn development, and challenge and defense (Barry, 1987, p. 252). The current study adopted the framework of the theory for the modern development stage since the hierarchy of effect s model in the early development stage focuses on only immediate outcomes or effects. The fo cus of communication effects in public relations is the long-term.

PAGE 66

48 typically fallen under the cognitive categor y (Ray, 1973). The affective component includes the variables of inte rest, evaluation, attitude, fee ling, conviction, and yielding. The variables intention, behavi or, and action are grouped under the category of ‘conative’ (Ray, 1973). Despite the on-going debate on the seque nce of influence by which the three components are arranged, this theoretical perspective suggests that communication effects occur through a series of steps, including cognition, affect, and behavior (CAB) (Berelson, 1996; Lavidge & Steiner, 1961; McGuire, 1986 ; Severin & Tankard, 2001). Cognitive reaction is thought to precede the effect on attitude, which precedes a behavioral reaction (Ray, 1973). Attitude is a reasoned product of perception and is a reliable predictor of behavior (Chaffee & Roser, 1986). Herein, public relations is considered a communication activity so that the theory of a hierarchy of effects can explain how the effects of public relations can occur. Although scholars in public re lations have not extensivel y applied the theory of hierarchy of effects to relationship cogniti ons, attitude, and behaviors, some have attempted to link relationship perceptions a nd attitude or relati onship attitude and behaviors. For example, Ledingham and Bruni ng (1998) examined the link between five relationship dimensions—trust, openness, involvement, investment, and commitment— and attitude toward an organization. Th eir study revealed that “organizational involvement in, and support of, the community in which it operates can engender loyalty toward an organization among key publics wh en that involvement/support is known by key publics (p. 63).” In another study, Bruning and Ledingham (2000) investigated consumers’ perceptions of th ree types of relationships (p ersonal, professional, and

PAGE 67

49 community relationships) with a bank and f ound that consumers’ perceptions of personal and professional relationships had a signifi cant impact on their evaluations of their overall satisfaction with th e organization. Bruning (2002) , who studied the universitystudent relationship as it relates to student retention, found that rela tionship attitude is different between students who return and thos e who leave, suggesti ng that links between relationship attitude and outcome behaviors exist. None of these studies, however, has suggested a comprehensive model explaining perceptions, attitude, and behavior, nor their linkages. Therefore, this study drew th e following research question. Research Question 6: What is a mode l linking relationship quality perception, attitude, and behavioral intentions? Based on the literature, the two fo llowing hypotheses were drawn. HP3: A public’s perceptions of its relati onship with the organization will influence the public’s attitude to ward the organization. HP4: A public’s attitude toward the organization will influence the public’s behavioral intentions. Relationship Perception Attitude Behavioral Intention HP 3 HP 4 Figure 2-6. Proposed model linking relati onship quality perception, attitude, and behavioral intentions In addition to the traditional seque ntial model explained above (cognition affect behavior), several alternative orders am ong cognition, affect, and behavior have been suggested (Ray, Sawyer, Rothschild, Heel er, Strong, & Reed, 1973) . From available alternative orders, this study adopted a lowinvolvement hierarchy which might usefully

PAGE 68

50 explain the sequential influence of public re lations on the members of a low-involvement public. Therefore, the current study adopted a low-involvement hierar chy to find the best model of sequence for the linkages between relationship perception, attitude, and behavior for low-involved publics. The Low-Involvement Hierarchy The low-involvement hierarchy is also called the cognitive-conative-affective model. Krugman (1965) attacked the standard sequential order in the hierarchy of effects model, which is conceptually similar to Fishbein’s theory of reasoned action’.10 Krugman (1965) developed the low-involveme nt hierarchy to explore why television advertisements had significant combin ation effects on the viewer even though experimental research seemed to reveal onl y minimal effects on an individual’s attitude change. He implied that most televisi on watchers do not pa y attention to the advertisement, meaning there was either no or little perceptual defense against the messages. According to the theory, when individuals are in a low involvement condition, they might proceed directly from their cognitive impact to action without reasoned action, followed by attitude change. In other words, minimal awareness comes first, followed by behavior or trial, and then attitude process. This theory has usually been used to explain situations in which individuals are involved, but alternatives have led to little or no differentiations. 10 The basic concept of the ‘theory of reasoned action’ is the collection of attitude toward behavior and subjective norms that cause behavioral intentions, which drive behavior (Fishbein, 1980). For a detailed description of the theory of reasoned action, see Fishbein, M. (1980). A theory of reasoned action: Some applications and implications. In M. M. Page (Ed.), Nebraska symposium on motivation (Vol. 27). Lincoln, Nebraska: University of Nebraska Press.

PAGE 69

51 The theory of the low-involvement hier archy of effects can be applied to organization-public relationships. Take for example how a member of a public who is only marginally aware of his or her relationshi p with the American Ca ncer Society or of the organization itself will donate money to a breast cancer prevention program run by the American Cancer Society. Although that member of the public did not form any attitude toward the organizati on prior to their do nation (behavior), he or she is more likely to establish a positive attitude toward the organization to justify behavior after conducting that behavior. Applying the theory to an organizationpublic relationship, members of a public, who have just started a relationship with an organization or have been in the relationship for only a short period of time, are most likely to have minima l awareness of the relationship or even the organi zation than those who have been in a relationship with an organization for a long period of time. Alt hough members of the public did not have any attitude toward the organization initially or prior to starting the relationship with the organization, they are more likely to devel op a positive attitude toward the organization in order to support their choice. The behavioral intentions in this study are based on the low-involvement hierarchy theory, which assumes that the relationship perception of a public with low-involvement with an organization (meaning in a relationshi p with that organizati on for a short period of time) might affect their beha vior toward that organization. Also, their attitude toward the organization might be affected by their beha vioral intentions. Based on this logic, the following two hypotheses were drawn: o HP-5: For a public experiencing low-involvement, the public’s relationship perception will positiv ely influence their behavior.

PAGE 70

52 o HP-6: For a public experiencing lowinvolvement, the public’s behaviors will positively influence their at titude toward the organization. Relationship Perception Behavior Attitude HP5 HP6 Figure 2-7. Proposed model li nking relationship quality outco me perception, behaviors, and attitude (based on low-invol vement hierarchy of effects) Comprehensive Model Based on the literature explained throughout this chapter, the suggested comprehensive model links relationship maintenance strategies, relationship quality outcomes, attitude, and behaviors. Figure 2-8 illustrates the relationship maintenance strategies that precede relationship quality outcomes, and positively affect attitude, a predictor of supportive behavi or. Thus, this study sugge sts the following research question: Research Question 7: What is a model li nking relationship maintenance strategies, relationship quality outcomes, attitude, and behavior? Relationship Maintenance Strategies Positive Attitude Supportive Behaviors Relationship Quality Outcome Figure 2-8. The comprehensive model Summary of Research Questions and Hypotheses Based on the nature of this study and the l iterature review, seve n research questions and six hypotheses were drawn.

PAGE 71

53 Research Question 1: How can relational maintenance strategies be measured? Research Question 2: To which and what extent is relationship maintenance strategies positively connected with which relationship quality outcomes? Research Question 3: What is a reliabl e and valid relationship quality outcome measurement? Research Question 4: Which of the four indicators represents relationship quality outcomes the most? Research Question 5: How do relationship quality outcome indicators affect each other? o HP-1: The degree of satisfaction will positively influence the degree of trust. o HP-2: The degree of trust will positively influence the degree of commitment. Research Question 6: What is a mode l linking relationship quality perception, attitude, and behavioral intentions? o HP-3: A public’s perceptions of its re lationship with the organization will influence the public’s attitude toward the organization. o HP-4: A public’s attitude toward the organization will influence the public’s behavioral intentions. o HP-5: For members of a public experiencing low-involvement, their relationship perception will positiv ely influence their behaviors. o HP-6: For members of a public experiencing low-involvement, their behaviors will positively influence thei r attitude toward the organization. Research Question 7: What is a model li nking relationship maintenance strategies, relationship quality percepti on, attitude and behavior?

PAGE 72

54 CHAPTER 3 METHODOLOGY The purpose of this study is to elucidate the concepts of rela tionship maintenance strategies and relationshi p quality outcomes and to test causal linkages among relationship maintenance strategies, rela tionship quality outcomes, attitude, and behavioral intentions. The following models were proposed and tested: Two measurement models: A measurement model of relationship maintenance strategies—access, positivity, openness, sharing of tasks, ne tworking, and assurances A measurement model of relationship quality outcomes—control mutuality, satisfaction, trust, and commitment Five causal models (struc tural equation models): A model to test linkages among relationship quality outcomes A model to test which and to what extent relationship maintenance strategies affect relationship quality outcomes Two models to test linkages between relationship quality outcome perception, attitude, and behavioral intentions o Sequential order 1: rela tionship quality perception attitude behavior o Sequential order 2: rela tionship quality perception behavior attitude A comprehensive model that links relations hip maintenance strategies, relationship quality outcomes, attitude, and be havior toward an organization This chapter first explains the approach to developing measurement scales and the appropriateness of the method for this study a nd goes on to describe the ways in which the data were collected. This chapter then gives an overview of th e operationalization of variables, validity and reliability tests, pretes t procedures/results, and statistical methods.

PAGE 73

55 Multiple regression, principal component anal ysis, confirmatory factor analysis, and structural equation modeling were the main st atistical procedures used for analyzing the collected data. The Process of Developing Measurement Scales The first step necessary for testing the pr oposed models is to develop measurement scales for each variable. This study deve lops a measurement scale for relationship maintenance strategies and refines an existi ng measurement scale for organization-public relationship based on the steps for developing a summated rating scale, which is also known as multivariate measurements, proposed by Spector (1992). Latent constructs such as relationship maintenance strategies and relationship quali ty outcomes require multiple items to more accurately capture the complexity of the constructs. The framework introduced by Spector (1992) is useful for multiple-item scale development and is composed of five major steps. Figur e 3-1 shows the procedur es involved in scale development. 1. Define Construct 2. Design Scale 3. Pilot Test 4. Administration & Item Analysis 5. Validation and Norm Figure 3-1. Major steps for developing a measurement scale (Spector, 1992, p. 8)

PAGE 74

56 During the first step of scale development, the constructs of interest—relationship maintenance strategies and re lationship quality outcomes—s hould be defined as clearly as possible. As described in the previ ous chapter, this study defines relationship maintenance strategies as “any organizational be havioral efforts that attempt to sustain or cultivate relationships with strategic publics” and rela tionship quality outcomes as “factors that determine or ch aracterize successful relationshi ps between an organization and its strategic publics.” The second step involves designing the scale, which is associated with the format of the scale and includes sele cting response choices. In th is study, all of the questions aside from demographic background informati on were asked using a nine-point Likert scale. A nine-point Likert scale was chosen for this st udy based on the guidelines of summated rating scale construc tion. According to Spector (1992), a greater number of choices would be better, because more choice results in greater accuracy. Regarding the number of response choices, it has been suggest ed that between five and nine choices are optimal for most uses (Ebel, 1969; Nunnally, 197 8). The researcher selected the ninepoint scale, which provides respondents with the most choices. Additionally, the ninepoint Likert scale includes options ranging from Strongly Di sagree (1) to Strongly Agree (9) and Neutral (5). However, no verbal la bels for the remaining scale points were employed for this study. This format of agreement choices, which is bipolar and symmetrical around a neutral point, is one of the most appropriate to use for summated scale development (Spector, 1992). In the third step of scale development, th e researcher should conduct a pilot-test for the initial version of measurement items with a handful of participants who were asked to

PAGE 75

57 critique the measurement items. In this step, these participants should identify which items are unclear or confusing as well as which items cannot be rated along the dimension chosen. The measurement items then should be revised based on feedback and comments from the participants (the pilo t test procedure will be explained later in this chapter). The fourth step involves a full administ ration with at least 100 to 200 respondents completing the measurement items and a subseq uent item analysis. The data then are used to determine internal consistency of the measurement items. The coefficient alpha is used to calculate the extent to which the items are internally consistent and reliable. When successful internal consistency of th e measurement items is determined, it is possible to proceed to the final step. Otherw ise, the researcher should go back to an earlier step to revise the scale. A series of validation checks should be conducted to identify whether the items measure as predicted (the validation process is explained in detail in the reliability and validity section). This step tests whethe r the underlying items measure each construct properly. Then, data are collected for verifi cation of the theoretical prediction. On the basis of evidence collected, the scale is eval uated to determine whether it measures the construct as intended. Spector (1992) identified four characteri stics that make a scale a summated rating scale. First, a scale should incorporate mu ltiple items. As the name literally indicates, multiple items should be combined or summ ated. Each variable in this study is represented by at least four measurement items to assess each indicator. Second, each individual item should measure somethi ng that has an unde rlying, quantitative measurement continuum, meaning that the scale measures an attribute of something that

PAGE 76

58 varies quantitatively rather than qualitativ ely. This study quantitatively measures each variable using a nine-point Li kert scale with responses rang ing from 1 (strongly disagree) to 9 (strongly agree). Third, each item has no “right” or “wrong” answer. This characteristic differentiates the summated rating scale from a multiple-choice test. Respondents can independently rate each statement based on their own thoughts and feelings. Lastly, each measurement item shoul d be a statement that participants are asked to rate using several response choices. As mentioned above, this study uses nine-point Likert scales for each statement so that res pondents have several response choices. The researcher selected the procedures of de veloping summated rating scales, because the measurement scales the researcher is developing are well-suited for the four characteristics explained above. Moreover, a well-developed summa ted rating scale can demonstrate reliability and validity. This chapter follows the order of the scale development procedure framework. Population and Samples To select an organization for this stud y, the researcher sent a letter requesting sponsorship for the research to five organi zations listed as places of employment for members of a departmental advisory c ouncil for a public university. The letter highlighted the objectives of the study and the benefits that an organization would receive if it sponsored the research. Two of the or ganizations responded to the letter, and the researcher selected the one that was more res ponsive and willing to s upport the research. The selected organization was the Florida Farm Bureau (FFB hereafter), which is the largest agricultural organization in Florida, a state in which agri culture is the second most important economic activity (“Florida Quick Facts,” 2006). The organization was created in the 1940s in order to handle a broa d range of agricultural issues or concerns

PAGE 77

59 (“The 1940s”, 2006). FFB is an independent , non-governmental, voluntary organization representing its members, who are primarily farmers and ranch families. As of 2006, the organization also represents sixty-four county level farm bureaus in Florida (“Welcome to Florida Farm Bureau,” 2006). The programs and services FFB provides ar e diverse, includi ng those related to insurance,1 labor, and legislation. The organizatio n provides services such as estate planning, safety, research on taxation and environmental rules, legislative lobby, and testimony before various agricu lture hearings on behalf of the Farm Bureau among many others (“The 1960s,” 2006). Since FFB is a grass roots organi zation, communication with its members is vital to th e organization’s operation. As a part of its public relations programs, the organization offers regular communication to its members through such efforts as publishing FloridaAgriculture and producing regula r radio and television programs (“The 1970s,” 2006). In the United States, all 50 states and Pu erto Rico have a state level farm bureau and several county level farm bureaus.2 Individual members belong to a county farm bureau, which is a member of a state farm bureau. State level farm bureaus attempt to enforce policies or rules that their member s suggest and provide programs and services which can improve members’ quality of life and well-being (“We are Farm Bureau,” 2006). The relationship with members is esse ntial to organizations such as the Farm Bureaus, because all of the organizations’ ac tivities are based on their interactions with members. More importantly, the organiza tion’s life and power are dependent on its 1 Insurance that has been provided to its member s is not limited to general causality insurance such as fire, wind, hail, hurricane, etc, but also includes life and auto mobile insurances, etc. 2 All state level farm bureaus in the United States are listed in the following Web site of the American Farm Bureau. http://www.fb.org/state/

PAGE 78

60 membership. The best way for this kind of organization to improve its power and influence is to increase its number of members, which in turn increases the sheer force of its membership (Kile, 1948). If a farm bureau did not make efforts to sustain or cultivate better relationships with its members, members would not likely remain affiliated with the organization. The population of this study is composed of current members of FFB, which now has a membership of over 140,000 (“Welcome to Florida Farm Bureau,” 2006). These members could provide meaningful evaluati ons about the strategies FFB has employed for maintaining relationships with them as well as the quality of their relationships with FFB. It is obvious that the members are a key public and target for FFB’s relationship building efforts. The researcher sent questionnaires to 2,100 of the current FFB members. This sample size was selected based on several crit eria. To test the proposed models using structural equation modeling, at least 200 valid cases are required. The general response rate of mail surveys ranges between 5 and 40% (Babbie, 2001). Us ing invitation, followup, and replacement survey questionnaires will in crease the response rate by at least 5%. To reach the minimum number of valid respons es based on the mail survey literature, it was necessary to send about 2,000 questionnaire s. An additional 100 questionnaires were added for replacement of unusable cases. Th e individuals used in this study were randomly selected from the FFB me mbership directory (N=2,100). Quantitative Research Approach This study employed a survey as the pr imary data collection method. Babbie (2001) referred to survey research as “the ad ministration of questionnaires to a sample of respondents selected from some population” (p. 282). The ma jor concerns of this study

PAGE 79

61 dealt with developing a solid measuremen t scale for the two main constructs— relationship maintenance strate gies and relationship quality outcomes—and exploring the causal links among these constructs. The su rvey method, which is applicable to nonexperimental data, is most suitable for thes e research purposes. According to Weisberg, Krosnick, and Bowen (1996), the survey met hod is appropriate to use when measuring answers to questions concerning attitudes, beliefs, and behaviors. They also suggested that surveys are appropriate for measuring a ttitudes (or preferences), belief (including predictions and assessments of importance ), or facts (including past behavioral experiences). Like any other research method, the survey method has its strengths and weaknesses. The first strength of the met hod is that a survey can be employed to investigate problems in a realistic setting (Wimmer & Dominick, 2000). Second, the cost of a survey is reasonable considering the amount of information it allows a researcher to gather (Wimmer & Dominick, 2000). A res earcher also can control expenses by choosing the most appropriate of the four survey research techniques—telephone, mail, personal interview, or group administration. Th ird, a researcher can easily collect a large amount of data from a variety of people, examine a large number of variables (e.g., demographics, social economic status, motives, attitude, behavioral in tentions, etc.), and use multivariate statistics to analyze the data collected. Despite these strengths, survey research has been criticized for displaying several weaknesses. As Babbie (2001) noted, a surv ey can seldom effectively address the context of social life. Babbie also pointed ou t that survey researchers can rarely develop a “feel” for the total life situation in which respondents are thinking and operating to the same extent as participant observers are able to (p. 237). Similarly, Sypher (1990)

PAGE 80

62 pointed out that traditional quantitative methods cannot “bring to life the nuances of work life and talk” (pp. 3-4). Qua litative methods are needed to often deal with a depth of contextual information (Marshall & Ro ssman, 1998). Despite these weaknesses, quantitative research is still the most useful a pproach for this study. As stated in the first chapter, the primary purpose of this study is not to explore the im pact of the specific contextual variables on public re lations practices or to use th ese variables as the basis for data interpretation. Rather, one of the objectives of this study is to attempt to find causeand-effect models among relationship main tenance strategies, relational outcomes, attitude, and behaviors in reality. The cau se-and-effect models function on verifiable truth, Popper (1965) argued. Therefore, a qua ntitative research method is more suitable for this study than a qualitative one. Pretest Before mailing the survey, a one-time pilot test was conducted to ascertain the quality of the questionnaire ite ms. The primary purpose of the pretest was to make sure that the questionnaire was unde rstandable and that the que stions would produce valid responses. The pilot test was administered online from April 21 to April 29, 2006 with one e-mail reminder. An e-mail requesting par ticipation in the pilot study was sent to 140 current members of FFB. Out of 140, 16 ema ils failed to be delivered. Out of 124 valid samples, 28 members participated in the pilot test, resulting in a 23 percent response rate for the pilot study. Respondents were as ked to identify ambiguous or unclear items and provide suggestions for changes. The pilot test revealed that a couple of questions related to relationship maintenance strategies were not applicable to the members of the public in this study. For example, several members commented that they have never used the organization’s

PAGE 81

63 Web site, so that they could not answer the questions related to Web site. It is reasonable that many current members of FFB may not be familiar with new technology due to their industry (agriculture) characteri stics as well as the fact that the participant demographics were skewed toward older members. As a result, the word ‘Web site’ was dropped from the survey, and Web site relate d questions were modified to become questions related to general communication medium. One of the original questions, ‘The information FFB provides on its Web site for members is of little use to them,’ was changed to ‘The information FFB provides for members is of li ttle use to them.’ The other question was changed from ‘FFB’s Web site provides member s with adequate contact information’ to ‘FFB provides members with ade quate contact information.’ Mail survey A large-scale mail survey was conducted to measure relationship maintenance strategies and organization-public relationships and to test the models proposed. A mail survey was favored over other available surv ey methods such as one-on-one interviews, telephone surveys, etc. Because it is necessary to have at least 200 data points for each variable to test the propos ed models using structural equation modeling, it would not have been feasible to conduct one-on-one inte rviews. In addition, telephone surveys have a high rejection rate due to people’s annoyan ce with telemarketing calls. Furthermore, telephone surveys are useful fo r asking only a short list of questions, while a mail survey can make large samples feasible at a reas onable cost and easily accommodates the long list of questions required for th is study (Babbie, 2001). The mail survey was administered statewid e. Generally, a mail survey achieves a response rate of 5 to 40 percent. However, a low return casts doubt on external validity of the findings. To generalize the results of any study’s findings, it is essential to produce

PAGE 82

64 a response rate that is as high as possible. A response rate of about 50 percent is commonly considered as genera lizable (Babbie, 2001). To increase response rate, this study used three waves of mailing. A brie f pre-notice postcard was sent to each respondent a few days prior to sending the ma in questionnaire. In the second mail, each of the randomly selected indivi duals received a package that contained a cover letter, the survey questionnaire, and a return envelope with paid postage (see Appendix A to F for samples of mail survey items). As the la st wave, a follow-up reminder was mailed one week after the original questionnaire was sent . Additionally, a replac ement survey packet was sent to the participants who had not res ponded two weeks after the initial mailing. Measures Relationship maintenance strategies In this study, relationship maintenance strategies de scribe the organization’s behavioral efforts to maintain or cultivate a relationship with its key publics. Concepts representing six dimensions of relationshi p maintenance strategies—access, positivity, openness, sharing of tasks, networking, and assurances—were adopted from Hon and J. E. Grunig (1999). A nine-point Likert scale, including resp onses ranging from Strongly Disagree (1) to Neutral (5) and Strongly Agr ee (9) with no verbal labels for scale points two through eight, accompanie s each statement. To develop the measurement items for re lationship maintenance strategies, this study adopted the scales by Stafford and Cana ry (1991) and significan tly modified them to be applicable to organization-public relationship settings. For example, one of the items measuring positivity in interpersonal rela tionships originally stated, “S/he is very nice, courteous and polite when we talk” and was changed to “FFB’s communication with members is courteous.” To establish additional measurement items for relationship

PAGE 83

65 maintenance strategies, the researcher asked th ree top-level public re lations professionals who are key decision-makers in FFB’s pub lic relations program to list all the communication strategies and tactics that they believe represent each relationship maintenance strategy. The practitioners were provided with the definition of each maintenance strategy and then were asked for specific examples. Based on their answers, some of the relationship maintenance strategi es were developed. The strategy items were randomly ordered. Table 3-1 shows questionnaire items for each strategy. Table 3-1. Relationship maintenanc e strategies measurement items Access 2. FFB provides members with adequate contact information. 8. FFB provides members with opp ortunities to meet its staff. 26. When members have questions or concerns , FFB is willing to an swer their inquiries. 29. FFB provides members with adequate cont act information for specific staff on specific issues. Positivity 1. Attending FFB’s annual meeting is helpful to members. 7. The member benefits (i.e. insurance se rvices, bank services, etc.) FFB provides are important to members. 13. Receiving regular communications (e.g., FloridaAgriculture ) from FFB is beneficial to members. 15. FFB’s communication with members is courteous. 18. FFB attempts to make its intera ctions with members enjoyable. 27. The information FFB provides members w ith is of little use to them. [R] 28. FFB is cooperative when handli ng disagreements with members. Openness 3. FFB’s Annual Report is a valuable source of inform ation for members about what FFB has done. 12. FFB shares enough information with memb ers about the organization’s governance. 22. FFB’s member meetings are a valuable way for members to communicate their opinions to FFB. 24. The issue briefings FFB provides help members understand the issues. 25. FFB does not provide members with enough information about what FFB does with members’ dues. [R]

PAGE 84

66 Table 3-1. Continued Sharing of Tasks 5. FFB works with members to develop so lutions to problems that benefit members. 6. FFB is involved in managing community issues that members care about (e.g., disaster relie f, environmental protection). 9. FFB works effectively to resolve regulat ory issues its members are facing such as pesticide or food safety issues. 16. FFB and members do not work well toge ther at solving joint problems. [R] Networking 11. FFB effectively builds coalitions with groups (i.e. Suwannee Partnership) that impact members. 17. The coalitions that FFB forms with other agricultural groups benefit FFB members. 20. The Ag Coalition for legisla tive activities that FBB is invol ved in is helpful to FFB members. 21. FFB’s alliances with other like-minde d groups are useless to members. [R] Assurances 4. FFB makes a genuine effort to provide personal responses to members’ concerns. 10. Members do not believe that FFB rea lly cares about their concerns. [R] 14. FFB communicates the im portance of members. 19. FFB’s policy development process allows me mbers adequate opportunity to raise an issue and propose a solution. 23. When members raise concerns, FF B takes these concerns seriously. Note: [R] indicates reverse code. Relationship quality outcomes Items representing the four dimensions of organization-public relationships— control mutuality, satisfaction, trust, and commitment—were adopted from Hon and J. E. Grunig’s (1999) relationship indices. This study adopted the fu ll version of their measurement items, which had never previously been tested so that version needed to be refined. The relationship items were presen ted in random order. Table 3-2 shows the initial version of the organi zation-public relationship measur ement items for the FFB.

PAGE 85

67 Table 3-2. Relationship quality outcome measurement items Control Mutuality 4. FFB believes the opinions of members are legitimate. 5. Florida Farm Bureau (FFB) neglects members. [R] 9. When dealing with members, FFB has a tendency to throw its weight around. [R] 12. FFB really listens to wh at members have to say. 16. FFB seems to ignore members’ opinions in the decisions that affect members. [R] 21. When members interact with FFB, members f eel that they have some sense of control. 24. FFB cooperates with members. 29. Members have influence with the decision makers at FFB. Satisfaction 7. Both FFB and members benefit from their relationship. 11. Members are dissatisfied with their interaction with FFB. [R] 13. Members are happy with FFB. 15. Generally speaking, members are unhappy with the relationship FFB has established with them. [R] 19. Members enjoy dealing with FFB. 26. FFB fails to satisfy members’ needs. [R] 27. Members feel they are important to FFB. 31. In general, nothing of value has been accomplished by FFB for members. [R] Trust 2. FFB treats members fairly and justly. 6. Whenever FFB makes an important deci sion, members know FFB will consider the decision’s impact on members. 10. FFB can be relied on to keep its promises to members. 14. FFB takes the opinions of members into account when making decisions. 17. Members feel very confid ent about FFB’s abilities. 22. Members believe that FFB lacks the abilit y to accomplish what it says it will do. [R] 25. Sound principles guide FFB’s behavior. 29. FFB misleads members. [R] Commitment 1. FFB is trying to maintain a l ong-term commitment to its members. 3. There is only a short-term bond between FFB and its members. [R] 8. FFB wants to maintain a positive relationship with its members. 18. Compared to other farm organizations, members value their relationship with Florida Farm Bureau the most. 20. Members would rather work with FFB than without it. 23. Members want to have a relationship with FFB. 30. Members feel a sense of loyalty to FFB. Note: [R] indicates reverse code.

PAGE 86

68 Although these four relations hip indicators have been examined and commonly used as the measure of organization-public relationships, it is uncertain that these relationship indicators are act ually measuring the relations hip between an organization and its public. Therefore, th is study examines the extent to which these four indicators explain relationship quality outcomes. This study assesses relationship quality by asking respondents to state their overall assessment of the quality of their re lationships with the organization on a nine-point Like rt scale ranging from very negative (1) to very positive (9). Attitude The category of attitude toward the organi zation is defined as the public’s overall evaluation of the organization. Attitude was al so evaluated by using a nine-point scale ranging from Strongly Disagree (1 ) to Strongly Agree (9) and Neutral (5) with no verbal labels for scale points two through eight accompanying each statement. The following statements were used to measure attitude (“ Members’ impression of Florida Farm Bureau is favorable.” “Members’ impression of Florid a Farm Bureau is negative,” “Florida Farm Bureau is useful to members,” and “Members dislike Florida Farm Bureau.”). Most of the studies using this scale have reported a highly reliable coeffici ent of items, ranging from .84 to .97 (Woo, 2001). Behavioral intentions Scholars have often used behavioral inten tions rather than measures of actual behavior, because intentions ar e the most reliable predictor of people’s real behavior. Moreover, behavior intention, behavior , and action are grouped under the same “conative” heading (Ray, 1973). Thus, it was reasonable to use behavioral intentions

PAGE 87

69 instead of behavior for the current study. Also, the current study considers behavioral intentions and behaviors equally. The scales measuring behavioral intenti ons were taken from Zeithaml, Berry, and Parasuraman (1996), although they were signifi cantly modified for the organization and the purposes of this research. Three items we re developed that used the same nine-point Likert scale as the organization-public rela tionship items (Very Unlikely (1) to Very Likely (9) and Neutral (5)). The statements measuring cu rrent members’ behavioral intentions were the following: [“Members would recommend membership in Florida Farm Bureau to other farmers.” “Members would retain their membership in Florida Farm Bureau even if membership in a compar able association were available.” “Members would like to retain their membership with FFB for at least another five years.”]. Involvement Although there are several diffe rent ways of measuring i nvolvement (Day, Stafford, & Camacho, 1995), this study uses the numbe r of years each individual has been a member of the organization to measure involv ement. Length of time is often considered an involvement because a measure of invol vement increases as the length of time increases (Evans, 1993). More often than not, varying the directi on of questions can minimize bias from participants. So roughly a third of the que stionnaire items were negatively worded. Researchers usually add negative words such as no or not to reverse the questionnaire items. However, this approach can increase measurement error because participants easily miss the word or true meaning of questions (Spector, 1992). Adopting the suggestion of Spector (1992), the meanings of a third of the questionnaire items were reversed without using negatives if possibl e. For example, a measurement item for

PAGE 88

70 control mutuality stating, “This organization and people like me are attentive to what each other say,” was changed to “Florida Farm Bureau neglects members” to reverse the meaning of the statement. Demographic information Several demographic items were addre ssed at the end of the questionnaire, including gender, age, level of education, ethnicity, area of the state the respondents came from, and their main commodity. Reliability and validity test In quantitative research, two important concepts should be considered in order to establish the quality of measurement—reliability and validity.3 Using the summated rating-scale format to develop measurement ite ms for relationship maintenance strategies and relationship quality outcome items ensure s reliability and validity. It is often suggested that a well-developed summated ra ting scale can have sound reliability and validity, because the scale usually has good psychometric properties (Spector, 1992). Reliability means the ability of a measur e to provide the same result comparably over time. Reliability is important, because unreliable measures cannot be used to determine the relationships between variables. Of the several types of reliability, this study focuses on internal reliability, which is al so known as internal consistency. Internal consistency reliability is associated with the question of whether various subparts of a test can provide comparable data. By and large, Coefficient alpha (Cronbach, 1951) is used to measure internal consistency. Alpha is an overall measure of how well the items that 3 To see the overview of reliability and validity in de tail, see Carmines, E. G., & Zeller, R. A. (1979). Reliability and validity assessment . Newbury Park, CA: Sage.

PAGE 89

71 measure the same characteristics are correlated w ith each other. The results of reliability for this study are discussed in the following chapter. Reliability is not a sufficient condition for establishing validity. That is, a consistent measure does not guara ntee that it measures what it is designed to measure. Validity is related to whether a scale actually measures what the researcher intended it to measure. Several types of validity have been widely tested in so cial science research— face validity, content validity, and construct va lidity. Face validity is the simplest and most basic kind of validity. This vali dity can be achieved by looking over the questionnaire items to check if, on the surface, they measure what they appear to measure. Face validity is usually accepted based on the credibility of the researcher. Content validity can occur by asking experts in the study area to review the measures. Content validity should investigate two as pects: “the thoroughness with which the construct to be scaled and its domain were e xplicated and the extent to which the scale items represent the construct’s domain” (P arasuraman et al., 1988, p. 28). Since these two validities (face and content) are based on credibility and authority, they were achieved through several scholars’ examination of the measures before the survey was conducted. Before the pretest, the research er consulted with three public relations professors, each of whom had more than five years of experience in conducting empirical research in the area of relationship mana gement, to ensure the quality of the questionnaire. The last validity to be considered for th is study is construct validity, which relates to whether the scores measure the hypothetical construct that the rese archer believes they do (Calder, Phillips, & Tybout, 1983; Kline, 2005). Construct validity is measured indirectly because hypothetical constructs cannot be directly observed. Among numerous

PAGE 90

72 available options for assessing construct valid ity, confirmatory factor analysis is a valuable tool. Therefore, this study uses confirmatory factor analysis to evaluate construct validity of the hypothetical constructs for this study. Data Reduction and Data Analysis Prior to the data analysis procedure, th e scores of negatively worded items, i.e. items 10, 16, 21, 25, 27 in relationship maintenan ce strategies measurement, items 3, 5, 9, 11, 15, 16, 22, 26, 29, 31 in organization-public re lationship measurement, and items 2 and 4 in attitude measurement were reversed to gain a factor score in a positive direction. Items without responses were coded as missing data. Statistical Procedures for Data Analysis The following statistical methods were employed for data reduction and data analysis. For the construct measurements such as relationship mainte nance strategies and relationship quality outcome, this study us ed principal component analysis (PCA)4 to reduce the observed variables to a small set of composite components. In other words, PCA was used to form a composite indicator. It was determined that PCA would be more appropriate to use for this study than exploratory factor analys is, which is also called common factor analysis.5 In order to justify using PCA as a factor analytic method for this study, it is necessary to explain the difference between inductive and deductive approaches and the appropriate factor analysis for each approach. In the deductive approach, which is an exploratory approach, 4 To find the overview of principal component an alysis, the following reference is recommended: Dunteman, G. H. (1989). Principal component analysis . Newbury Park, CA: Sage. 5 To see the detailed comparison of principal compone nt analysis and exploratory factor analysis, the following source is useful: Park, H. S., Dailey, R., & Lemus, D. (2002). The use of exploratory factor analysis and principal co mponents analysis in communication research. Human Communication Research, 28 , 562-577.

PAGE 91

73 items are administrated to subjects first. Th en factor analytic pr ocedures are performed in an attempt to reveal the constructs with in the items. The deductive approach is an exploratory approach, and the conceptual work is focused on result interpretation instead of developing hypotheses. In this case, exploratory factor anal ysis (EFA) is more appropriate because the purpose of EFA is to find a latent structure of observed variables by uncovering common factors that influence the measured variables (Park, Dailey, & Lemus, 2002). A study using an inductive approach starts the defining constructs based on the existing literature (Spector, 1992). Then, the definitions of the constructs guide subsequent measurement items. The theo retical ideas guide th e validation of the developmental work. It is obvious that this res earch used an inductive approach for scale development, because the definition of each c onstruct (i.e., six indicators of relationship maintenance strategies and four indicator s of relationship quality outcomes) was established based on guidelines in the existing literature on the topic. The construct definition guides the subsequent scale deve lopment (subsequent measurement items for each construct). For a study employing an induc tive approach, PCA is more appropriate than EFA because the objective of PCA is to reduce measured variables (observed variables) into a smaller set of composite components. Also, a principal component is an exact mathematical transformation of the orig inal variables (Algina, 2005). This means that the principal components ar e not latent variables (factors), and scores of the principal components can be computed exactly. In order to determine how many factors to extract, the current study used the eigenvalue-greater-than-one rule and the scree test. The items with low factor loadings on intended factors or those highly loaded on unintended factors were eliminated.

PAGE 92

74 After removing those items, a confirmatory factor analysis (CFA) was conducted to confirm the theoretical factor structure and test for the generalizability of the factor structure over different data sets and to obtain answers to the following two research questions: “How can relational maintenance st rategies be measured?” (RQ 1) and “What is a reliable and valid rela tionship quality outcome measurement?” (RQ3). CFA can specify which variables define which factor. It can test construct validity (reliability between items) and discriminant validity (d ifference between factors). Cronbach’s alpha coefficient was employed to test the intern al consistency of the measurement items. Multiple Regression Analysis This study used stepwise regression analysis to obtain answers to the following two research questions: “To which and what exte nt is relationship maintenance strategies positively connected with which relationship qu ality outcomes? (RQ2) and “Which of the four indicators represent rela tionship quality outcomes the mo st?” (RQ4). In research question 2, relationship maintenance strate gies are independent variables, and relationship quality outcomes are dependent variables. Multip le regression analysis is useful to identify which relationship main tenance strategies (i.e., access, positivity, openness, sharing of tasks, networking, and assurances) significantly relate to relationship quality outcomes (i.e. control mutu ality, satisfaction, trust, and commitment). In the fourth research question, each relati onship quality outcome is an independent variable, and the overall evalua tion of the relationship is a dependent variable. It has been suggested that to regress the public ’s perceptions regarding overall relationship quality on each relationship dimensions can be an effective approach for evaluating the relative importance of relationship dimens ions (Huang, 2001b; Parasuraman, Zeithaml, &

PAGE 93

75 Berry, 1988). Therefore, multiple regressi on analysis identifies which relationship indicators explain ove rall relationship quality most accurately. Structural Equation Modeling To test if the proposed models fit the da ta, this study used structural equation modeling (SEM), a multivariate technique combining aspects of multiple regression and factor analysis to simultaneously estimate a se ries of interrelated relationships. Kaplan (2000) also described SEM as “a melding of f actor analysis and path analysis into one comprehensive statistical methodol ogy” (p. 3). One of the greatest advantages of SEM is its ability to accommodate measurement error di rectly into the estimation of a series of dependent relationships (Hair, Anderson, Tatham, & Black, 1998). A two-step procedure of structural equation modeling was used. In the process, the first stage estimates the measurement model, which establishes associat ions between latent (unobserved or factor) variables6 and multiple observed items, and relates to confirmatory factor analysis. The second stage estimates the stru ctural model, which uses the measurement model that was estimated from the first stage (Anderson & Gerbing, 1988; Hair et al., 1998; James, Muliak, & Brett, 1982; Kenny, 1979; Kline, 2005; Muliak, James, Van Alstine, Bennett, Lind, & Stillwell, 1989; Williams & Hazer, 1986). SEM helps to obtain answers to the follo wing research questions and hypotheses: Research Question 5: How do relationship quality outcome indicators affect each other? o HP-1: The degree of satisfaction will positively influence the degree of trust. 6 Latent variables are the underlying constructs that cannot be directly measured by any one set of measures, but they are hypothesized to affect pa rticular observed variables in the measurement model (Holbert & Stephenson, 2002). However, the latent variables are what researchers want to capture, which cannot be measured by any one form of observed variable (Duncan, 1975).

PAGE 94

76 o HP-2: The degree of trust will positively influence the degree of commitment. Research Question 6: What is a mode l linking relationship quality perception, attitude, and behavioral intentions? o HP-3: A public’s perceptions of its re lationship with the organization will influence the public’s attitude toward the organization. o HP-4: A public’s attitude toward the organization will influence the public’s behavioral intentions. o HP-5: For members of a public experiencing low-involvement, their relationship perception will positiv ely influence their behaviors. o HP-6: For members of a public experiencing low-involvement, their behaviors will positively influence thei r attitude toward the organization. Research Question 7: What is a model li nking relationship maintenance strategies, relationship quality perception, att itude and behavioral intentions? In structural equation modeling, sample size plays a significant role in the estimation and interpretation of the results. Although there is no universally accepted single indicator for determini ng a reasonable sample size in SEM, it has been suggested that a minimum of four cases per parameter to be investigated is a desirable sample size (Tanaka, 1987). Also, a minimum sample size of 200 is recommended in order to be suitable (Hoyle & Kenny, 1999). The final sa mple size of this study is 385, which is sufficient for using the SEM to analyze the data.7 Model Fit In order to evaluate the degree to which the proposed models fit the observed data, this study utilized several crite ria, as shown in Table 3-3. The most widely used model estimator is maximum likelihood (ML) (Bo llen, 1989; Chou & Bentle r, 1995). ML is appropriate to use for large samples because it assumes normality of the sample, which 7 The following four factors influence the sample size: 1) the specific model fitted to the analyzed data, 2) model size, 3) normal distribution of observed variables, and 4) statistical method(s) employed for purposes of parameter estimation and model testing (Hair et al., 1998, pp. 604-605).

PAGE 95

77 leads to normal error distribution. In other words, with a larg e sample size, ML is robust to deviation from normal distri bution (Hu, Bentler, & Kano, 1992). First, the 2 goodness of fit statis tic is used as an index of model adequacy with a nonsignificant value indicating a good fit to the data. Inst ead of using the value of 2 itself, this study used the ratio of 2 to the degree of freedom because 2 is quite sensitive to sample size, according to Bollen (1989). A va lue less than five of the ratio generally indicates a good fit (Bollen, 1989). Other common fit indices indicating how well the specified model explains the observed data ar e as follows: comparative fit index (CFI), goodness of fit index (GFI), normed fit inde x (NFI), root mean squared residual approximation (RMSEA), and root mean squa red residual (RMR). For CFI, GFI, and NFI, values range from 0 to 1.00, with higher values indicating better fit; .90 and above is commonly regarded as a good fit. RMSEA values close to .08 or le ss typically indicate good fit (Byrne, 2001; Kline, 2005). RMR should be equal to or smaller than .05. Also, regression coefficients for the hypothesized st ructural relations also were reported with their statistical significance. Signifi cant alpha levels for all tests are .05. Table 3-3. Model fit criteria Models Fit Criteria Chi-square/(df) < 5 Comparative Fit Index (CFI) > .90 Goodness of Fit Index (GFI) > .90 Normed Fit Index (NFI) > .90 Rood Mean Squared Error Residual (RMSEA) < .08 Root Mean square Residual (RMR) < .05

PAGE 96

78 CHAPTER 4 RESULTS This chapter consists of five parts. The first portion offers a description of the data collected, specifying respondents’ demographi c information and response rates. The second section addresses descrip tive statistics while the third part pr ovides an overview of the results of the two measurement models – relationship maintenance strategies and organization-public relationship. Fourth, the results of stepwise regression analysis including correlation analysis are provided follo wed by path analysis. Last, the results of structural equation modeling for all proposed and tested models are explained. Description of Samples A statewide mail survey was distribute d on May 11, 2006. One week before the mail survey packages were sent, an invitati on postcard was sent to each member of the sample group. Twenty-one hundred questionn aires were sent to randomly selected members from the current membership director y of the Florida Farm Bureau. One week later, reminders were sent to each sample member. Two weeks after mailing the original questionnaires, replacement packages were sent to those in the sample group who did not initially respond. Response Rate Of the 2,100 questionnaires mailed, a tota l of 453 were collected. Among the 453 collected surveys, 405 surveys were comple ted, 5 surveys were incomplete, 18 were refused and returned by the respondents, a nd 25 surveys failed to deliver and were returned by the United States Postal Service (USPS). Although response rate is typically

PAGE 97

79 calculated based on the number of complete questionnaires with re porting units divided by the number of usable questionnaires in the sample (Frankel, 1983; Groves, 1989; Massey, 1995), there are numerous ways of determining response rates. Based on the response rate equations suggested by the American Association for Public Opinion Research, the response rates calculated fo r this study range from 19.00 percent to 19.75 percent (“Standard definitions,” 2006) as show n in Table 4-1. On the basis of the most extensively used response rate, RR4, the respon se rate of this study is 19.75 percent. Among the 410 complete or partially comple te surveys, only complete surveys (N = 405) were considered for analysis, thus e liminating the incomple te surveys (N = 5) from the data set. Out of 405 complete surv eys, 20 of the questionnaires which displayed response set (i.e., answers that were the same for all of the questions) were dropped. As a result, only 385 surveys remained for use in the final data analysis. Table 4-1. Response rate calculation RR Type Formula Application Result RR11 Complete / total 405/2100 19.00 RR2 Complete / (total – non-delivered) 405/ (2100-25) 19.51 RR3 (Complete + incomplete) / total (405+5) /2100 19.52 RR4 (complete + incomplete) / (total – nondelivered) (405+5)/(2100-25) 19.75Adopted from American Association for Pub lic Opinion Research (AAPOR, 2006, p.32). Note: RR indicates response rates. Total=the number of surveys mailed Demographics As shown in Table 4-2, among the 381 respondents in the sample, 251 (67.8%) were males, 119 (32.2%) were females, and 11 respondents did not answer the gender question. Comparing the proportion of gender for the pretest (82% for male vs. 28% for female), females demonstrated greater partic ipation in the main survey. However, the composition of gender of the sample showed an explicit difference with that of the 1 RR1 is the minimum response rate.

PAGE 98

80 population2 ( Census of Agriculture State Prof ile”), indicating that the sample was skewed toward male individuals. As Table 4-2 shows, the sample included a notably low number of young members (under 35 years old) (N = 10, 2.8%) and a larg e number of older (65 and older) members (N = 190, 52.4%) compared to the population (8.6% (under 35 years old) vs. 22.1% (65 and older)). The average respondent age was 64, which is older than the average population age (M = 57). Also, as compared to the average age of the pretest participants (M = 49), the main survey partic ipants were significantly older. Therefore, it can be said that the sample was skewed toward older members. With respect to education, approximately 7% of respondents had some schooling, another 29% had received a high school diplom a, about 50 percent at tended college, and another 17% had received a graduate degree.3 In terms of race, respondents indicating a ‘White/Caucasian’ background dominated part icipation in the survey (N = 360, 96.5%), accurately reflecting the ethnic distribution of the population. Only one African American participated (.3%) along with three Latino/ Hispanics (.8%), se ven Native Americans (1.9%), one Anglo (.3%), and one responde nt indicated other (e.g., multiracial). With regard to the geographical background of the respondents, most were from the North Central portion of the state (N = 84, 22.8%) followed by the Southeast (N = 78, 21.2%). Sixty-three respondents were from Northwest (17.1%), 46 resided in the 2 This study uses the demographic data from Unite d States Department of Agriculture Florida Agricultural Statistics Service, because the demographic information about the member population of Florida Farm Bureau is not availabl e. There may be some differences between the demographic information from Florida Agricultura l Statistics Service and the member population of Florida Farm Bureau. However, the demographi c information used here is the most similar available to that of the population of the Florida Farm Bureau members. 3 The demographic information from United States Department of Agriculture does not provide educational information about the population. Therefore, comparison of demographic information in this category could not be provided.

PAGE 99

81 Northeast (12.5%), 36 indicated East Central (9.2%), 30 answ ered West Central (11.5%), and 11 participants were from the Southwest (4.2%). In response to the question concerning th e participants’ repr esentative commodity, the most prevalent commodity indicated wa s livestock (N = 138, 38.7%). Approximately 13% were engaged in work related to forest ry (N = 46), approximately ten percent dealt with citrus (N = 37), about ni ne percent were involved in ornamental horticulture (N = 33), seven percent produced field crops (N = 25) , about six percent grew other fruits and vegetables (N = 21), one percent specialized in dairy (N = 4), and about 15 percent indicated other in response to this question (e.g., gardening, hay, etc). The mean number of years that participants had been members of the organization was 26 years. Compared to the pretest participants (M = 17 years), the main survey participants had been affiliated with the organization for a longer period of time. In summary, the sample of the current study was skewed toward older male White individuals. Table 4-2. Sample de mographic description Variables Category Frequencies Percentage Gender Male 251 65.9 Female 119 32.2 Total 370 100.0 Under 25 years old 2 .6 24-34 8 2.2 35-44 12 3.3 45-54 67 18.5 55-64 83 22.9 65-74 99 27.3 75 years and older 91 25.1 Age Total 362 100.0

PAGE 100

82 Table 4-2. Continued Variables Category Frequencies Percentage Some schooling 25 6.7 High school diploma 108 29.0 Some college 106 28.5 Bachelor’s degree 71 19.1 Some graduate school 23 6.2 Graduate degree 29 7.8 Doctoral degree 10 2.7 Education (Highest level of education) Total 372 100.0 Race White/Caucasian 360 96.5 African American 1 .3 Asian American 0 0 Latino/Hispanic 3 .8 Native American 7 1.8 Anglo 1 .3 Other 1 .3 Total 373 100.0 North Central 84 22.8 Southeast 78 21.2 Northwest 63 17.1 West Central 47 12.8 Northeast 46 12.5 East Central 36 9.8 Southwest 13 3.5 Other 1 .3 Area of the state Total 368 100.0 Livestock 138 38.7 Forestry 46 12.9 Citrus 37 10.4 Ornamental horticulture 33 9.2 Field crops 25 7.0 Other fruits and vegetables 21 5.9 Dairy 4 1.1 Others 53 14.8 Representative Commodity Total 357 100.0 NMinMaxMean SD Years of Membership 32816125.88 13.83 Descriptive Statistics For the descriptive statistics, mean and sta ndard deviation were utilized because all of the measurement variables—relationship ma intenance strategies, relational outcomes, attitude and behavioral intentions—were me tric (=numeric). The means and standard

PAGE 101

83 deviations of all measurement items and cons tructs appear in Tabl e 4-3 for relationship maintenance strategies, Table 4-4 for relati onship quality outcomes, and Table 4-5 for attitude and behavioral intenti ons. As previously stated, nine -point Likert scales ranging from (1) strongly disagree to (9) strongly agree were used in all measures. Relationship Maintenance Strategies The mean score for the total relationshi p maintenance strategies was 7.03. Among the measures of six relationship maintenan ce strategies, the respondents rated positivity highest and openness lowest (M = 7.23 vs. M = 6.75). Apparently, respondents generally believed that the organization made an effo rt to foster a positive relationship with members, but they felt the organization ha s not effectively provided enough information about the nature of the organization and its current activities. Sharing of tasks was ranked as the second highest relationship maintenance strategy (M = 7.17, SD = 1.49) followed by assurances (M = 7.10, SD = 1.47), access (M = 7.05, SD = 1.53), networking (M = 6.89, SD = 1.40), and openness (M = 6.75). With respect to access , the respondents moderately agreed that FFB provided enough contact information to them (M = 7.39) and that the organization was willing to answer their inquiries or con cerns that are raised (M = 7.19) . However, they thought the opportunities of meeting FFB’s staff (M = 6.72) and obtaining contact information for specific staff on specific issue we re more limited (M = 6.87). In regard to positivity , respondents believed that FFB ’s communication with them was courteous (M = 7.89) and that the communi cation materials they received were quite beneficial (M = 7.85) and useful to them (M = 7.33). Also, they moderately agreed that the membership benefits they received were important to them (M = 7.73) and that FFB made efforts to make its interaction with members enjoyable (M = 7.41). The

PAGE 102

84 respondents, however, demonstrated the lowest levels of agreement with respect to the helpfulness of the annual meeting (M = 6.05) and FFB’s cooperation with handling their disagreements (M = 6.44). Specifically, a pproximately 50 percent of the respondents said that they had no positive or negative opinions about the annual meeting (N = 180) while only 17 percent strongly agreed that th e annual meeting was useful to them. In reference to the cooperation of the organization in handling disagreements, more than 40 percent of the respondents had a neutral opi nion (N = 149) while 21 percent strongly agreed with the statement about cooperativeness (N = 77). On the questions regarding openness , the respondents agreed moderately that the annual report is a valuable source of information concerning FFB’s accomplishments and efforts of the year (M = 7.09) and that the me mber meetings provide a valuable venue for them to communicate their opinions to FFB (M = 7.01). Respondents indicated lower levels of agreement regarding the manner in which FFB shares governance of the organization with members (M = 6.63) and the helpfulness of issue briefings in understanding the issues (M = 6.98). They l east agreed that FFB provided members with information about how the organization utiliz es member dues (M = 6.02). Particularly, more than 50 percent of the respondents disa greed with the statement to some degree (N = 189). The respondents’ answers to th e four items dealing with sharing of tasks were similar in ranges with the average values ranging from 7.09 to 7.23, indicating that participants moderately agreed that FFB ha d made significant efforts to work on projects or solve problems of mutual interest between themselves and the organization. They thought that FFB and members worked well toge ther to solve joint problems (M = 7.12) and that the organization worked with them to develop solutions that benefit members (M

PAGE 103

85 = 7.22). Specifically, the respondents seemed to agree that FFB worked effectively to solve regulatory issues they were facing (M = 7.23) and also took care of community issues (M = 7.09). With respect to networking , the respondents believed th at the Ag coalition (M = 7.28) and the coalition that FFB formed with other agricultural groups (M = 7.14) were useful to them. Specifically, more than 55% of the respondents highly agreed with the usefulness of the Ag coalition (N = 203). They agreed to a somewhat lesser extent with the statements specifying that FFB’s alliance with other lik e-minded groups are useful to them (M = 6.82) and that FFB effectively bui lds coalitions with groups that impact members (M = 6.31). With reference to assurances , the respondents displayed the highest levels of agreement with the idea that FFB communicat ed the importance of members (M = 7.34). Specifically, more than 55 percent of them strongly or somewhat agreed with the statement (N = 123). Also, they believed th at FFB cared about their concerns (M = 7.26), that the organization made a genuine effort to provide personal responses to members’ concerns (M = 7.24), and that the organization seriously ad dressed the concerns that members raised (M = 6.99). However, the re spondents agreed least with the statement that FFB’s policy development process provided enough opportunity for members to raise issues and propose solutions (M = 6.66). Particularly, approxi mately 30 percent of the respondents had neutral opinions about the statement (N = 114).

PAGE 104

86 Table 4-3. Descriptive statistics fo r relationship maintenance strategies Variables MeanSD Relationship Maintenance Strategies (N = 365) Access 7.051.53 Q1-2. FFB provides members with ad equate contact information. 7.391.66 Q1-8. FFB provides members with oppor tunities to meet its staff. 6.722.02 Q1.26. When members have questions or concerns, FFB is willing to answer their inquiries. 7.191.91 Q1-29. FFB provides members with ad equate contact information for specific staff on specific issues. 6.871.84 Positivity 7.231.33 Q1-1. Attending FFB’s annual meeting is helpful to members. 6.051.87 Q1-7. The member benefits (i.e. insu rance services, bank services, etc.) FFB provide s are important to members. 7.731.69 Q1-13. Receiving regular communications (e.g., Florida Agriculture) from FFB is beneficial to members. 7.851.47 Q1-15. FFB’s communication with members is courteous. 7.891.60 Q1-18. FFB attempts to make its inter actions with members enjoyable. 7.411.65 Q1-27. The information FFB provides members with is of little use to them. [R] 7.332.04 Q1-28. FFB is cooperative when handling disagreements with members. 6.441.97 Openness 6.751.40 Q1-3. FFB’s Annual Report is a valu able source of information for members about what FFB has done. 7.091.80 Q1-12. FFB shares enough information with members about the organization’s governance. 6.631.76 Q1-22. FFB’s member meetings are a valuable way for members to communi cate their opinions to FFB. 7.011.89 Q1-24. The issue briefings FFB provides help members understand the issues. 6.981.75 Q1-25. FFB does not provide members with enough information about what FFB does with members’ dues. [R] 6.022.48 Sharing of Tasks 7.171.49 Q1-5. FFB works with members to deve lop solutions to problems that benefit members. 7.221.85 Q1-6. FFB is involved in managing co mmunity issues that members care about (e.g., disaster relief, environmental protection). 7.091.40 Q1-9. FFB works effectively to resolve regulatory issues its members are facing such as pesticide or food safety issues. 7.231.79 Q1-16. FFB and members do not work well together at solving joint problems. [R] 7.122.14

PAGE 105

87 Table 4-3. Continued Variables MeanSD Relationship Maintenance Strategies (N = 365) Networking 6.891.40 Q1-11. FFB effectively builds coaliti ons with groups (i.e. Suwannee Partnership) that impact members. 6.311.69 Q1-17. The coalitions that FFB form s with other agri cultural groups benefit FFB members. 7.141.82 Q1-20. The Ag Coalition for legislative activities that FFB is involved in Is helpful to FFB members. 7.281.69 Q1-21. FFB’s alliances with other li ke-minded groups are useless to members. [R] 6.822.41 Assurances 7.101.47 Q1-4. FFB makes a genuine effort to provide personal responses to members’ concerns. 7.241.89 Q1-10. Members do not believe that FFB r eally cares about their concerns. [R] 7.262.21 Q1-14. FFB communicates the importance of members. 7.341.78 Q1-19. FFB’s policy development process allows members adequate opportunity to raise an issue and propose a solution. 6.661.85 Q1-23. When members raise concerns, FFB takes these concerns seriously. 6.991.87Note: N = the sample size after listwise deletion for missing variables. [R] indicates reverse-coding. Relationship Quality Outcomes The mean score for the total relatio nal outcomes was 7.35, which indicated a positive evaluation by the respondents concerning their relationship with FFB. The mean values for the four indicato rs in the relationship qualit y outcomes were as follows: 7.48 for commitment, 7.45 for satisfaction, 7.32 for tr ust, and 7.14 for control mutuality as indicated in Table 4-4. Overall, the respondents most strongly ag reed that something of value had been accomplished by FFB for its members (M = 7.93), which is one of satisfaction items. This response correlates with participants’ high overall satisfaction rating concerning their relationship with the organization. More than 70 percent of the members who participated in this study strongly agreed with the statement (N = 226). However, respondents agreed least with the statement regarding the influence they had with

PAGE 106

88 decision makers at the organization (M = 6.53) , which is one of the control mutuality measurement items. This item seemed to lead to a lower average score in the category of control mutuality (M = 7.14) as compared to the other three relati onship indicators. Broadly, the respondents indicated a modera te degree of agreement on statements related to control mutuality in the relationship. Seventytwo percent of the respondents noted some levels of agreement with the statement, FFB believes the opinions of members are legitimate (M = 7.14, N = 262). Similarly, 80 percent of the respondents agreed that the organization did not have a tendency to throw its weight around when dealing with them4 (M = 7.16, N = 292). The results showed that the respondents possessed similar levels of agreement with the following two statements, FFB seems to ignore members’ opinions in the decisions that affect members5 and FFB cooperates with members (M = 7.31 and M = 7.37, respectivel y). Also, about 50 percent of the respondents demonstrated some level of agreem ent in reference to the statement that FFB listened to what they had to say (M = 6.98), and approximately 60 percent of them agreed with the statement, When memb ers interact with FFB, members feel that they have some sense of control (M = 6.79). Nevertheless, only 46 percent of re spondents agreed that they had influence with the decision ma kers at the organization (M = 6.53). The respondents were highly satisfied with the relationshi p they had with FFB (overall mean of satisfaction = 7.46). Esp ecially, more than 60 percent of the respondents strongly agreed that FFB had ach ieved something valuable for them (M = 7.97, N = 226), and 43 percent of them expre ssed strong agreement that both they and FFB benefited from the relationship (M = 7.75, N = 155). Approximately 40 percent of 4 The statement was reverse-coded. 5 The statement was reverse-coded.

PAGE 107

89 them were highly satisfied with their inte raction with the organization (M = 7.30), and similar portions of the participants indi cated a great deal of happiness with the relationship FFB had established with them (M = 7.41). Likewise, 46 percent strongly believed that FFB successfully satisfied thei r needs (M = 7.51). The remaining three statements had relatively lower levels of agreement than other satisfaction items: Members are happy with FFB (M = 7.22), Members enjoy dealing with FFB and Members feel they are important to FFB (M = 7.24 for the last two). The members in this study generally had established some level of trust with the organization. Particularly, a majority of th e respondents (59%) str ongly believed that the organization did not mislead them (M = 7.82, N = 214). Forty-two percent of them strongly agreed that FFB treated them fairly and justly (M = 7.72, N = 153). Additionally, they believed that sound principles guided the behavior of the organization (M = 7.36). Although many agreed with th e statement, Whenever FFB makes an important decision, members know FFB will c onsider the decision’s impact on members (M = 7.28), about 26 percent expressed some le vel of disagreement that the organization took members’ opinions into account when making decisions (M = 6.96). The remaining three statements related to trust received similar levels of agreement: Members feel very confident about FFB abilities, Members believ e that FFB lacks the ability to accomplish what it says it will do,6 and FFB can be relied on to keep its promises to members (M = 7.26, 7.16, and 7.12, respectively). Overall, the members of the publics who responded to this study expressed a higher level agreement related to commitment in the relationship between themselves and the organization than they did for any other rela tionship indicator. More than 45 percent of 6 The statement was reverse-coded.

PAGE 108

90 the participants in this study expressed strong agreement with both of the following statements: FFB wants to maintain a positiv e relationship with members and Members would rather work with FFB than without it (M = 7.82 and M = 7.74, respectively), and 43 percent strongly disagreed that there wa s only a short-term bond between them and FFB7 (M = 7.16). Forty-four percent of th e respondents strongly agreed that the organization was attempting to maintain a long-term commitment to them (M = 7.71). More than 50 percent of the members of the public who participated in this study wanted to keep a relationship with FFB (M = 7.59) and a similar portion of respondents felt loyalty to the organization (M = 7.37). M eanwhile, approximately 30 percent disagreed or had neutral opinions about the statement th at compared to othe r farm organizations, members value their relationship with th e organization the most (M = 7.04). In summary, the members of the publics who participated in the current study evaluated their relationship with the organization (FFB) fairly positively across all indicators of relationship quality outcomes in cluding control mutuality, satisfaction, trust, and commitment. In particular, they rated commitment at the highest level followed by satisfaction, trust, and control mutuality. Th erefore, the members are fairly committed to their relationship with FFB, however; they believed that comparatively more power balance is needed. Table 4-4. Descriptive statistics of relationship quality outcomes Variables MeanSD Relationship Quality Outcome (N = 366) Control Mutuality 7.141.61 Q2-4. FFB believes the opinions of members are legitimate. 7.201.87 Q2-5. Florida Farm Bureau (FFB) neglects members. [R] 7.771.95 Q2-9. When dealing with members, FFB has a tendency to throw its weight around. [R] 7.162.23 7 The statement was reverse-coded.

PAGE 109

91 Table 4-4. Continued Variables MeanSD Relationship Quality Outcomes (N = 366) Control Mutuality 7.141.61 Q2-12. FFB really listens to wh at members have to say. 6.982.07 Q2-16. FFB seems to ignore members’ opini ons in the decisions that affect members. [R] 7.312.03 Q2-21. When members interact with FFB , members feel that they have some sense of control. 6.791.99 Q2-24. FFB cooperates with members. 7.371.77 Q2-28. Members have influence with the decision makers at FFB. 6.532.02 Satisfaction 7.461.47 Q2-7. Both FFB and members be nefit from their relationship. 7.751.65 Q2-11. Members are dissatisfied with th eir interactions with FFB. [R] 7.302.02 Q2-13. Members are happy with FFB. 7.221.82 Q2-15. Generally speaking, members ar e unhappy with the relationship FFB has established with them. [R] 7.411.94 Q2-19. Members enjoy dealing with FFB. 7.241.77 Q2-26. FFB fails to satisfy members’ needs. [R] 7.512.04 Q2-27. Members feel they are important to FFB. 7.241.91 Q2-31. In general, nothing of value has been accomplished by FFB for Members. [R] 7.971.73 Trust 7.321.53 Q2-2. FFB treats members fairly and justly. 7.621.85 Q2-6. Whenever FFB makes an im portant decision, members know FFB will consider the decision’s impact on members. 7.281.78 Q2-10. FFB can be relied on to keep its promises to members. 7.121.98 Q2-14. FFB takes the opinions of me mbers into account when making decisions. 6.961.98 Q2-17. Members feel very confident about FFB’s abilities. 7.261.74 Q2-22. Members believe that FFB lack s the ability to accomplish what it says it will do. [R] 7.162.12 Q2-25. Sound principles gui de FFB’s behavior. 7.361.86 Q2-29. FFB misleads members. [R] 7.821.88 Commitment 7.491.38 Q2-1. FFB is trying to maintain a long-term commitment to members. 7.711.70 Q2-3. There is only a short-term bond between FFB and members. [R] 7.162.29 Q2-8. FFB wants to maintain a posi tive relationship with members. 7.821.73 Q2-18. Compared to other farm or ganizations, members value their relationship with Florida Farm Bureau the most. 7.041.83 Q2-20. Members would rather work with FFB than without it. 7.741.65 Q2-23. Members want to have a relationship with FFB. 7.591.54 Q2-30. Members feel a sense of loyalty to FFB. 7.371.74Note: N = the sample size after listwise deletion fo r missing variables. [R] indicates reverse-coding.

PAGE 110

92 Attitude and Behavioral Intentions As indicated in Table 4-5, the mean scores for attitude and behavioral intentions were 8.01 and 7.73, respectively. The respondents rated attitude toward the organization somewhat higher than their supportive behavioral inten tions (M = 8.01 vs. M = 7.73). Moreover, the mean score of attitude was rated as highest across all indicators. Specifically, sixty-six percent of the responde nts strongly disagreed with the statement that members dislike Florida Farm Bureau8 (M = 8.25, N = 241). Also, fifty-eight percent of the members who responded to th is study expressed strong disagreement with the item stating that Florida Farm Bureau is negative (M = 8.05, N = 213). Consistent with the results, more than 40 percent of the respondents strongl y agreed with the usefulness (M = 7.94) and favorableness of the organization (M = 7.79). The majority of the respondents had strong be havioral intentions, meaning that they would recommend membership in the organiza tion to other farmers (M = 7.88) and that they would like retain their membership with the organization for at least another five years (M = 7.76). Approximately forty-two pe rcent of the members who participated in the survey strongly believed th at they would retain their membership in FFB even if membership in a comparable asso ciation were available (M = 7.53). In summary, the respondents overall tended to have strong positive attitude toward the organization and were more likely to behave in a supportive manner toward the organization while attitude scores were higher than behavioral intentions. 8 The statement was reverse-coded.

PAGE 111

93 Table 4-5. Descriptive statistics of attitude and behavioral intentions Variables MeanSD Attitude (N = 374) 8.011.36 Q3-1. Florida Farm Bureau is favorable. 7.791.56 Q3-2. Florida Farm Bureau is negative. [R] 8.051.60 Q3-3. Florida Farm Bureau is useful to members. 7.941.46 Q3-4. Members dislike Flor ida Farm Bureau. [R] 8.251.44 Behavioral Intentions (N = 374) 7.731.55 Q4-1. Members would recommend membership in Florida Farm Bureau to other farmers. 7.881.67 Q4-2. Members would retain their membership in FFB even if membership in a co mparable association were available. 7.531.88 Q4-3. Members would like to retain their membership with FFB for at least another five years. 7.761.68Note: N = the sample size after listwise deletion fo r missing variables. [R] indicates reverse-coding. Reliability of Initial Measurement Items Table 4-6 displays the result of Cronbach’s alpha reliability test for the initial measurement items. Although there is no unive rsally accepted standa rd, one of the most widely accepted rules of thumb is that the alpha should be at least .70 for a scale to demonstrate internal consistency (Nunnall y, 1978). It has been suggested that a Cronbach coefficient alpha of approximate ly .90 is excellent, around .80 is very good, and values around .70 are adequate (Hair et al., 1998; Kline, 2005). Based on the suggested criteria of reliabil ity, three relationship quality outcome indicators—control mutuality, satisfaction, and trus t—and attitude were consider ed to have an ‘excellent’ reliability because the alphas for the variab les were larger than .90. Three relational maintenance strategies (access, positivit y, assurances), the relational outcome commitment, and behavioral intentions were deemed to have ‘very good’ reliability because the alphas were larger than .80. All of the remaining variables demonstrate alphas higher than .70, which are perceived to be in the ‘good’ reliability range. Therefore, all of the items examined ha d appropriate levels of reliability.

PAGE 112

94 Table 4-6. Results of reliability tests of initial measurement items Variable # of items # of cases Cronbach’s Cronbach’s (standardized) Relationship Maintenance Strategies Access 4374.84 .84 Positivity 7373.82 .83 Openness 5376.77 .79 Sharing of Tasks 4375.79 .79 Networking 4377.71 .75 Assurances 5375.83 .83 Relationship Quality Outcomes Control Mutuality 8371.93 .93 Satisfaction 8371.91 .92 Trust 8375.92 .93 Commitment 7374.88 .89 Attitude 4375.92 .92 Behavioral Intentions 3376.88 .88 Measurement Models In order to examine the two sets of meas ures of relational maintenance strategies and organization-public relati onship as related to resear ch questions 1 and 3, two statistical analysis procedures – principal factor analysis and confirmatory factor analysis – were used. Factor Analysis As the first step in developing measurem ent models, factor analysis, especially principal component analysis, was conducted. As a general guide for sample size when using factor analysis, there s hould be at least five observati ons per variable (Hair et al., 1998). In this study, 60 variables were observed for consideration in f actor analysis with 381 samples; therefore, the sample size for th e factor analysis is adequate based on the general guideline.

PAGE 113

95 In order to eliminate the items with lo w factor loadings and to check factor loadings of each measurement items on its inte nded factor, factor analysis was conducted. In this step, the primary goal was to assess the dimensionality and th e appropriateness of the measurement variables for each latent vari able. Measurement items were deleted if they 1) were extracted as the second factor of the intended factor , 2) included opposite signs of factor loading coefficients among the other items in the intended factors, and 3) had factor loading values of less than .65 with the othe r items of their respective subscales based on the suggestion. In other words, items having low communalities9 with their intended factor or that produ ced high loading on unintended factors were dropped. The ten latent variables with multiple items were targeted for this analysis. For example, in order to measure relationshi p maintenance strategies, 29 items were employed: four items for access, seven items for positivity, five items for openness, four items for sharing of tasks, four items for ne tworking, and five items for assurances. Also, to measure relationship quality outcomes, a to tal of 31 items were used: eight items for control mutuality, eight items for satisfaction, eight items for trust, and seven items for commitment. This study used composite variables that were produced from principal component analysis for the latent variables with multiple items due to identification10 issues of 9 Communality is the total amount of variance an or iginal variable shares with all other variables included in the analysis (Hair et al., 1998, p. 88) 10 All structural equation models should be over-i dentified (Hoyle, 1991; Raykov & Marcoulides, 2000). Over-identified means that there should be more equations for the model than unknown parameters, which are those for which the SEM produces numerical values. A model is overidentified if it has one or more degree of freedom. If a model has zero degree of freedom, then the model is just-identified. A just-identifie d model always produces perfect fit and cannot be properly assessed (MacCallum, 1995). If a model is under-identified, meaning there is more

PAGE 114

96 confirmatory factor analysis. Additionall y, using composite vari ables is useful for making the proposed causal model more parsimonious and easy for convergence. Relationship maintenance strategies A nine-point Likert scale with responses ranging fr om (1) strongly disagree, (5) neutral to (9) strongly ag ree was employed for relationship maintenance strategies measurement. Respondents were instructed to evaluate what the organization (Florida Farm Bureau) has done in terms of main taining relationships with them. Table 4-7 shows that all initial items of access have fairly high factor loadings ranging from .78 to .87. The extracted components accounted for approximately 70 percent of the variation in the observed va riables with a 2.77 eige nvalue. All initial measurement items were retained for final analysis because they have sufficient factor loadings. Table 4-7. Factor loadings for access Items Loading AC1. When members have questions or concerns, FFB is willing to answer their inquiries. AC2. FFB provides members with adequate information for specific staff on specific issues. AC3. FFB provides members with adequate contact information. AC4. FFB provides members with opportunities to meet its staff. .87 .84 .81 .78 Eigenvalue Percentage variance accounted for Cronbach’s (standardized) 2.77 69.46 .84 With respect to positivity , as provided in Table 4-8, seven items were initially included in the measure, while the followi ng two items with low factor loadings – Attending FFB’s annual meeting is helpful to members and The information FFB provides members with is of little use to them—were eliminated. The remaining items parameter than degree of freedom, programs cannot perform the analysis. Therefore, in order to perform SEM analysis, there should be more equa tions for a model than unknown parameter.

PAGE 115

97 have factor loadings rangi ng from .68 to .82. The extract ed components accounted for approximately 50 percent of the variati on in the observed variables with a 3.49 eigenvalue. Table 4-8. Factor loadings for positivity Items Loading PO1. FFB attempts to make its inte ractions with members enjoyable PO2. FFB’s communication with members is courteous. PO3. Receiving regular communication from FFB is beneficial to members. PO4. The member benefits FFB provides are important to members. PO5. FFB is cooperative when handling disagreements with members PO6. The information FFB provides memb ers with is of little use to them.[R]a PO7. Attending FFB’s annual meeting is helpful to members.a .82 .80 .77 .69 .68 .63 .50 Eigenvalue Percentage variance accounted for Cronbach’s (standardized) 3.49 49.83 .84b Note: a indicates the item which eventually was deleted from the factor. [R] indicates that the item was reverse-coded. b This was calculated after removing the two items with low factor loadings and communality. Table 4-9. Factor loadings for openness Items Loading OP1. The issue briefings FFB provides help members understand the issues. OP2. FFB shares enough information w ith members about the organization’s governance. OP3. FFB’s annual report is a valuable source of information for members about what FFB has done. OP4. FFB’s member meetings are a valuable way for members to communicate their opinions to FFB. OP5. FFB does not provide members with enough information about what FFB does with members’ dues.a [R] .85 .78 .76 .75 .55 Eigenvalue Percentage variance accounted for Cronbach’s (standardized) 2.77 55.35 .81b Note: a indicates the item which eventually was deleted fr om the factor. [R] indicates the item was reversecoded. b This was calculated after removing one item with low factor loadings. Five items were initia lly designed to measure openness as shown in Table 4-9. The fifth statement, FFB does not provide member s with enough information about what FFB does with members’ dues, was discarded due to its low factor loading. The items remaining for the final analysis have relativ ely high factor loadings ranging from .75 to

PAGE 116

98 .85. The extracted components explained about 55 percent of the variation in the observed variables with a 2.77 eigenvalue. Four items were initiall y designed to investigate sharing of tasks as given in Table 4-10. Factor analysis revealed that all these items could be included for the final analysis. Among the four items, the first three items had f actor loadings ranging from .80 to .86. The last item had .66 factor loadings . The extracted components explained about 62 percent of the variation with a 2.49 eigenvalue. Table 4-10. Factor loadings for sharing of tasks Items Loading ST1. FFB is involved in managing community issues that members care about (e.g., disaster relief , environmental issues). ST2. FFB works effectively to resolv e regulatory issues its members are facing such as pesticide. ST3. FFB works with members to devel op solutions to problems that benefit members. ST4. FFB and members do not work well t ogether at solving joint problems. [R] .86 .82 .80 .66 Eigenvalue Percentage variance accounted for Cronbach’s (standardized) 2.49 62.18 .80 Note: a indicates the item which eventually was deleted fr om the factor. [R] indicates that the item was reverse-coded. Table 4-11. Factor lo adings for networking Items Loading NE1. The Ag Coalition for legislative act ivities that FFB is involved in is helpful to its members. NE2. The coalitions that FFB forms with other agricultural groups benefit its members. NE3. FFB effectively builds coalitio ns with groups (i.e. Suwannee Partnership) that impact members. NE4. FFB’s alliances with other like-mi nded groups are useless to members. [R]a .85 .83 .83 .48 Eigenvalue Percentage variance accounted for Cronbach’s (standardized) 2.33 58.19 .81b Note: a indicates the item which eventually was deleted fr om the factor. [R] indicates that the item was reverse-coded. b This was calculated after eliminating one item with low factor loadings.

PAGE 117

99 Four items were constructed in the measure of networking as seen in Table 4-11, while one item with low factor loadings, FFB ’s alliances with other like-minded groups are useless to members, was eliminated. A ll of the three remaining items have similar factor loadings ranging from .83 to .85. The extracted components explained approximately 58 percent of the variati on in the observed variables with a 2.33 eigenvalue. Five items initially we re designed to measure assurances as shown in Table 4-12. Factor analysis demonstrated that the follo wing four items were included in the final analysis, and one item stating that Members do not believe that FFB really cares about their concerns had low factor loadings a nd was removed for the final analysis. The remaining four items had factor loadings ranging from .79 to .86. The extracted components explained approximately 58 per cent of the variatio n in the observed variables with a 2.33 eigenvalue. Table 4-12. Factor loadings for assurances Items Loading AS1. When members raise concerns, FFB takes these concerns seriously. AS2. FFB makes a genuine effort to pr ovide personal responses to members’ concerns. AS3. FFB’s policy development process allows members adequate opportunity to raise an issue a nd propose a solution. AS4. FFB communicates the importance of members. AS5. Members do not believe that FFB r eally cares about th eir concerns. [R]a .86 .81 .79 .79 .62 Eigenvalue Percentage variance accounted for Cronbach’s (standardized) 3.02 60.33 .85b Note: a indicates the item which eventually was deleted fr om the factor. [R] indicates that the item was reverse-coded. b This was calculated after removing one item with low factor loadings. In summary, factor analysis suggest ed measuring relationship maintenance strategies with 24 items, consisting of four items for access, five items for positivity, four items for openness, four items for sharing of tasks, three items fo r networking, and four items for assurances. The composite scores of each indicator from principal component

PAGE 118

100 factor analysis after deleting the items with low factor loadings and communality were used throughout the an alysis in this study. Relationship quality outcomes In order to measure relationship quality outcomes, 31 items were initially used: eight items for control mutuality, eight items for satisfaction, eight items for trust, and seven items for commitment. With respect to control mutuality , the factor analysis revealed that all measurement items were included in the final measures as shown Table 4-13. The highest loadings were found in the following item: FFB coopera tes with members, which demonstrated a .89 factor loading. In terms of eigenvalue and explained variance, more than 66 percent of the variances were explained by the extract ed factor analysis, with an eigenvalue of 5.31. Table 4-13. Factor loadings for control mutuality Items Loading CM1. FFB cooperates with members. CM2. FFB really listens to what members have to say. CM3. FFB believes the opinions of members are legitimate. CM4. FFB seems to ignore members’ opinions in the decisions that affect members. [R] CM5. Members have influence with the decision makers at FFB. CM6. When members interact with FFB, members feel that they have some sense of control. CM7. When dealing with members, FFB has a tendency to throw its weight around. [R] CM8. FFB neglects members. [R] .89 .84 .84 .82 .81 .78 .77 .76 Eigenvalue Percentage variance accounted for Cronbach’s (standardized) 5.31 66.35 .93 Note: [R] indicates that the item was reverse-coded. Similarly, all eight items used for measuring satisfaction were included in the subscale as provided in Table 4-14. The loadings range from .73 to .84. Also, in terms

PAGE 119

101 of eigenvalue and explained variance, about 63 percent of the variance was accounted for by the extracted factor analys is with a 5.03 eigenvalue. Table 4-14. Factor loadings for satisfaction Items Loading S1. Members are happy with FFB. S2. Members feel they are important to FFB. S3. Both FFB and members benefit from their relationship. S4. Members enjoy dealing with FFB. S5. FFB fails to satisfy members’ needs. [R] S6. Generally speaking, members are unha ppy with the relationship FFB has established with them. [R] S7. In general, nothing of value has been accomplished by FFB for members. [R] S8. Members are dissatisfied with their interaction with FFB. [R] .84 .83 .83 .82 .78 .76 .75 .73 Eigenvalue Percentage variance accounted for Cronbach’s (standardized) 5.03 62.88 .92 Note: [R] indicates that the item was reverse-coded. Table 4-15. Factor loadings for trust Items Loading T1. FFB takes the opinions of members into account when making decisions. T2. Members feel very confid ent about FFB’s abilities. T3. FFB can be relied on to keep its promises to members. T4. Whenever FFB makes an important decision, members know it will consider the decision’s impact on members. T5. FFB treats members fairly and justly. T6. FFB misleads members. [R] T7. Sound principles guide FFB’s behavior. T8. Members believe that FFB lacks th e ability to accomplish what it says it will do.a [R] .86 .86 .86 .85 .85 .79 .77 .64 Eigenvalue Percentage variance accounted for Cronbach’s (standardized) 5.29 66.12 .93b Note: a the item eventually was deleted from the factor. [R] indicates that the ite m was reverse-coded. bThis was calculated after removing one item with low factor loadings. As shown in Table 4-15, eight items were initially designed to measure trust . The factor analysis identified seven of these items to be included for final analysis in this study. The factor loadings of the remaining items range from .77 to .86. With respect to eigenvalue and explained va riance, approximately 66 percent of the variance was

PAGE 120

102 accounted for by the extracted factor anal ysis with a 5.29 eigenvalue. One item, Members believe that FFB lacks the ability to accomplish what it says (T8), was eliminated because of its low factor loadings. Regarding commitment , seven items were initiall y constructed to measure commitment as shown in Table 4-16, while two items were removed. The item stating that there is only a short-term bond between FFB and members was deleted due to low factor loading, and the other item indicating that members want to have a relationship with FFB’ was eliminated due to low communa lity. The factor loadings of the five remaining items range from .77 to .85. With re gard to eigenvalue and explained variance, about 60 percent of the variance was accounted fo r by the extracted factor analysis with a 4.16 eigenvalue. Table 4-16. Factor loadings for commitment Items Loading C1. FFB is trying to maintain a long-term commitment to members. C2. FFB wants to maintain a positive relationship with members. C3. Compared to other farm organizati ons, members value their relationship with FFB the most. C4. Members would rather work with FFB than without it. C5. Members feel a sense of loyalty to FFB. C6. There is only a short-term bond between FFB and members.a [R] C7. Members want to have a relationship with FFB.a .85 .83 .78 .83 .77 .63 .67 Eigenvalue Percentage variance accounted for Cronbach’s (standardized) 4.16 59.45 .89b Note: a the item eventually was deleted from the factor. [R] indicates that the ite m was reverse-coded. b This was calculated after deleting two items with low factor loadings. In summary, the factor analysis produced measuring relationship quality outcomes with 28 items, consisting eight items for cont rol mutuality, eight items for satisfaction, seven items for trust, and five items for commitment. After deleting the items with low factor loadings, the composite scores of each indicator of relationship quality outcomes

PAGE 121

103 were calculated through principa l component analysis. The composite scores for each indicator were employed for all an alyses throughout this study. Confirmatory Factor Analysis As the second step, confirmatory factor analysis was conducted in order to assess the adequacy of the factor structure that had been constructed (i.e. relationship maintenance strategies a nd relationship quality outcomes) using AMOS 6.0.11 Confirmatory factor analysis confirms if a hypothesized factor model or measurement model fits the given data and is appropriat e to use following principal component factor analysis for a couple of reasons. First, confir matory factor analysis is appropriate for the proposed structural validity that is derived from questions such as : 1) the number of factors (i.e., latent variables) that underlie responses to items on a test, 2) the associations among those factors, and 3) the contribution of the factors to the items of the test (Hoyle & Smith, 1994). Research questions 1 and 3 in this study are such questions. Second, confirmatory factor analysis can offer a statis tical test of the extent to which a proposed measurement model fits observed, empirically collected data (Hoyle, 1991). More importantly, confirmatory factor analysis is more appropriate for this study than exploratory factor analysis because the hypot hesized factor structure was inductively established. In other words, the structure was developed based on existing literature and theories. To handle missing data,12 a list-wise deletion procedure was used. This 11 AMOS is a statistical software package for ru nning structural equation modeling. In social science research, AMOS, LISREL, and EQS are the most popularly used software packages. Detailed information about how to use LISREL can be found at Jreskog, K., & Srbom, D. (1996). LISREL 8: User's reference guide. Chicago: Scientific Software International. Information about EQS can be found at Bentler, P. M. (1992). EQS: Structural equations program manual . Los Angeles: BMDP Statistical Software; Bentler, P. M., & Wu, E. J. C. (1993). EQS/Windows user's guide . Los Angeles: BMDP Statistical Software. 12 There are several methods of handling missing data in structural equation modeling. Here, the following three most commonly used procedures of handling missing data in SEM are introduced.

PAGE 122

104 reduced the sample from 381 to 365 indivi duals who responded to every item on the questionnaire (N = 365). Table 4-17 provides the results of the c onfirmatory factor analysis for the measurement model of relationship maintenan ce strategies. All of the indicators were significantly and successfully loaded on their designated factors. The magnitude of the factor loadings demonstrated th at all indicators in the relationship maintenance strategies measurement model demonstrated highly st rong loadings with the relationship maintenance strategies. All indicators had higher than .75 st andardized loadings. Among the six indicators, assurances had the highest loading at .95, followed by .91 for positivity, .89 for access, .88 for sharing of tasks, .85 for openness, and .78 for access. All factor loadings in the standardized solutions were statistically significant at p <.001 Table 4-17. Confirmatory factor analysis of relationship maintenance strategies Indicators Relationship Maintenance Strategies Access .89a Positivity .91*** (.04) Openness .85*** (.04) Sharing of tasks .88*** (.04) Networking .78*** (.05) Assurances .95*** (.04) Note : aValues were not calculated because loading was set to 1.0 to fix construct variance. The numbers outside parentheses indicate standardized estimates ( ). The numbers in parentheses indicate standard error. * p< .05; ** p< .01; *** p< .001 The most commonly used procedure is listwise de letion which eliminates any participant from the data set with one or more scores missing on the variables being analyzed. This method is good for goodness-of-fit indices. However, if sample size is small with lots of missing cases, the method is not recommended. The second most fre quently used procedure is full information maximum likelihood. In this procedure, if a par ticipant has scores on a subset of the variables or on all of the variables, the participant’s scores ar e utilized to calculate the likelihood function. Thus, all available data are used in the analysis . However, this procedure often causes a problem in goodness-of-fit indices. The last method is multiple imputation, which eliminates missing scores based on the relationships among the variables. This method is the least frequently used of the three methods due to its complexity of mu ltiple analyses and combining results (Algina, 2005).

PAGE 123

105 Moreover, as shown in Table 4-18, all model fit indices i ndicated that the measurement model reasonably fits the da ta (CFI = .99, NFI = .99, and RMSEA = .07). Therefore, the measurement model of re lationship maintenance strategies had good construct reliability and validity. Figure 4-1 illustrates the measurement model of relationship maintenance strategies. Table 4-18. Fit measures for relationship maintenance strategies measurement model Fit Index Criteria Fit Statistics Chi-square > .005 25.43 ( p< .05; df = 9) Chi-square/df < 3 2.83 Comparative Fit Index (CFI) > .9 .99 Normed Fit Index (NFI) > .9 .99 Root Mean Squared Error Residual (RMSEA) < .08 .07 Relationship Maintenance Strategies Access Positivity Openness Sharing of task Networking Assurances.89 (.03) .91*** (.03) .85*** (.03) .88*** (.03) .78*** (.03) .95*** (.04) Note: The number outside parentheses indicate s standardized coefficient and the number inside parenthesi s indicates standardized error. The solid line indicates significant path. *** p< .001 Figure 4-1. Measurement model of relationship maintenance strategies13 13 This study also analyzed the two measurement models that include multiple items for each indicator. However, the relationship qualit y outcome measurement model, which includes multiple items for each indicator including control mu tuality, satisfaction, trust, and commitment, was not identified. Because the model has more parameter than equations, the analysis could not

PAGE 124

106 Table 4-19 illustrates the re sults of the confirmatory factor analysis for the relationship quality outcome measurement m odel. Like the relationship maintenance strategies measurement model, every indi cator of relationship quality outcomes was significantly and successfully loaded on its corresponding factor. The magnitude of the factor loadings identified that all of the indicators in the rela tionship quality outcome measurement model had higher than .90 standardi zed loadings. All fact or loadings in the standardized solutions were statistically significant at p< .001. Among four indicators of the relationship quality outcomes, tr ust had the highest loading value ( = .97), followed by satisfaction at .96, control mutuality at .95, and commitment at .92. Moreover, as shown in Table 4-20, the measurement model ge nerally indicated appropriate model fits with the exception of RMSEA. Although Chisquare and the ratio of Chi-square/degree of freedom, which are sensitive to sample size, were not satisfactory, other important fit indices demonstrated the measurement model to be desirable (CFI = .99 and NFI = .99). Therefore, it can be said that the measure of the relationship quality outcomes displayed good construct reliability and validity. Figure 4-2 presents a visual of the measurement model of relationship quality outcomes. Table 4-19. Confirmatory factor anal ysis of relationship quality outcomess be performed. For this reason, each indicator was c onsidered an observed variable for the rest of model testing. Indicators Relationship Quality Outcomes Control Mutuality .95a Satisfaction .96*** (.02) Trust .97*** (.02) Commitment .92*** (.03) Note: aValues were not calculated because loading was set to 1.0 to fix construct variance. The numbers outside parentheses indicate standardized estimates ( ). The numbers in parentheses indicate standard error. * p< .05; ** p< .01; *** p< .001

PAGE 125

107 Table 4-20. Fit measures for relations hip quality outcome measurement model Fit Index Criteria Fit Statistics Chi-square > .05 22.46 ( p< .001; df = 2) Chi-square/df < 5 11.23 Comparative Fit Index (CFI) > .9 .99 Normed Fit Index (NFI) > .90 .99 Root Mean Squared Error Residual (RMSEA) < .08 .17 Relationship Quality Outcome Control Mutuality Satisfaction Trust Commitment.95 (.02) .92 (.03) .97 (.02) .96 (.02) Note: The number outside parentheses indicates standardi zed coefficient and the number inside parentheses indicates standardized error. The solid line indicates significant path. p< .001 Figure 4-2. Measurement model of relationship quality outcomes Additionally, since the two initial meas urement models for both relationship maintenance strategies and the relations hip quality outcomes were satisfactory, respecification was not necessary. Therefore, the initial measurement models were used as the final measurement model throughout model testing. Reliability and Validity Test The last step of testing measurement mode ls is the reliability and validity check. This study used internal consistency as a reliability check. Also, this study assessed construct validity of the confir matory factor analysis model on the basis of the average

PAGE 126

108 amount of variance in indicator variables as explained by each factor in the measurement model. Reliability test. In order to examine the intern al consistency of the measurement scales, Cronbach’s alpha was used. The va lues of coefficient alpha for internal consistency were .95 for relationship mainte nance strategies and .97 for relationship quality outcome measurements. Although re lationship maintenance strategies had slightly lower internal consistency, both meas urements demonstrated an ‘excellent’ level of reliability (int ernal consistency) (Hair et al., 1998; Kline, 2005). Reliabilities of each indicator for relationship maintenance stra tegies are as followed: .84 for access and positivity, .81 for openness, .80 for sharing of tasks, .81 for networking, and .85 for assurances. All indicators with multiple it ems for relationship maintenance strategies have acceptable reliability. Internal consis tencies of each indicator for relationship quality outcomes are as follows: .93 for cont rol mutuality, .92 for satisfaction, .93 for trust, and .89 for commitment. All display ex cellent levels of internal consistency aside from commitment, which is close to the excellent level. Validity test. As explained in the previous ch apter, the two funda mental validities, face and content validities, were checked befo re the main data collection. In order to assess the other important validity of the measurement models, the current study used construct validity, which indicates the extent to which the scores measure the hypothetical construct. Confirmatory factor analysis is one of the most widely used procedures for evaluating cons truct validity. Cons truct validity was measured based on the average amount of variance in indicator variables accounted for by each factor in the confirmatory factor analysis. Hancock and Muller (2001) suggested that an appropriate level of construct validity is equal to or higher than .50 fo r the variance extracted. The

PAGE 127

109 factors of both relationship maintenance stra tegies and relationship quality outcomes surpassed the desirable level of construct validity. The extracted variances in the latent variables were .80 for relationship maintenan ce strategies and .91 for relationship quality outcomes. Therefore, the two measurement mode ls met the desirable levels of reliability and validity. In summary, the first and third research questions relating to the measurement models of relationship maintenance strategi es and relationship qua lity outcomes were answered. The results from confirmatory fact or analysis indicated that both measurement models were adequate. Comparing the two measurement models, it seemed that the measurement model of the relationship quality outcomes is better in terms of loading values. Moreover, both measurement models achieved higher levels than acceptable reliability and validity criteria. Correlation Analysis The composite scores obtained from principal component analysis for each of the ten indicators on the two sets of measures —relationship maintenance strategies and relationship quality outcomes—were employe d for correlation analysis. Prior to performing a regression statisti cal analysis, a correlation anal ysis was conducted to check the relationships between independent and dependent variables in addition to identifying any violations of regression assumptions, es pecially multicollinearity. In this part, relationship maintenance strate gies are independent variab les and relationship quality outcomes act as dependent variables. Tabl e 4-21 presents the results of correlation analysis of relationship maintenance stra tegies and relationship quality outcome variables. Overall, all variables of relati onship maintenance strate gies and relationship quality outcomes are significant and demonstr ate high association with each other.

PAGE 128

110 Overall, control mutuality displayed signi ficant associations with all relationship maintenance strategies. In particular, contro l mutuality had the highest correlations with assurances (r = .80). Additionally, control mu tuality is highly associated with access (r = .75), positivity (r = .74), and sh aring of tasks (r = .73). Comp ared to other strategies, control mutuality had relatively low associ ations with openness (r = .69) and networking (r = .62). Satisfaction also had statis tically and signifi cantly strong associ ations across all relationship maintenance strategies. In particular, satisfaction had the strongest relationship with assurances (r = .78). Moreover, control mutuality displayed a significantly high association with positivity (r = .75), access (r = .72), and sharing of tasks (r = .70). Compared to other strategi es, satisfaction had rela tively weak correlations with openness (r = .67) a nd networking (r = .59). Trust had significant associations with al l of the relationship maintenance strategy indicators. Similar to control mutuality a nd satisfaction, trust also had the strongest association with assurances (r = .82). A dditionally, the indica tor was also highly correlated with other relationship maintenan ce strategies including positivity (r = .77), access (r = .75), sharing of tasks (r = .73), and op enness (r = .71). Similar to the results of control mutuality and satisfaction, trust de monstrated the weakes t correlations with networking (r = .63). The last variable of relationship qualit y outcomes, commitment, also displayed strong associations with all st rategy variables. Unlike the results of other relationship quality outcome variables, commitment ha d the strongest associations with both positivity (r = .79) and assurances (r = .79). Additionally, commitment displayed good substantial associati ons with access (r = .71), followed by sharing of tasks (r = .69) and

PAGE 129

111 openness (r = .68). Like the results of th e other relationship quality indicators, commitment also produced the least a ssociation with netw orking (r = .60). In summary, among the relationship mainte nance strategies, assurances displayed the highest associations across all of the relationship quality outcome variables. This might indicate that the degree of the organiza tion’s efforts to communicate to publics that their concerns are reasonable and attende d to represent substantial communication strategies to maintain and sustain better re lationships with public s. Networking had the lowest associations across all of the rela tionship quality outcome variables, possibly indicating that coalitions with groups the publics care about may not be the best communication strategies to fost er relationships with a targ et public. However, due to high correlations between and among the re lationship maintenance strategies and relationship quality outcome variables, this study’s reported a ssociations should be interpreted cautiously. Multiple Regression Analysis Prior to regression analysis, several underlying statistical assumptions were assessed, including normality, linearity, and multic ollinearity. First, normality, which is the most frequently violated assumption (H air et al., 1998) was tested through ShapiroWilks test statistics ( n < 2000) and visual normal QQ plot inspection (Johnson & Wichern, 1992). Second, the linearity was check ed using a partial regression plot, which represented the association between a singl e independent variable and the dependent variable. Last, multicollinearity was check ed among independent variables by using correlation analysis (relationship maintenance strategy variables for the second research question and relationship quality outcome variables for the f ourth research question). None of the assumption was violated.

PAGE 130

112Table 4-21. Correlation matrix between relationship maintenance strategi es and relationship quality outcomes 1 2 3 4 5 6 7 8 9 10 1. Access ----2. Positivity .82**----3. Openness .77**.75**----4. Sharing of tasks .76**.78**.76**----5. Networking .66**.73**.68**.71**----6. Assurances .84**.86**.81**.84**.72**----7. Control mutuality .75**.74**.69**.73**.62**.80** ----8. Satisfaction .72**.75**.67**.70**.59**.78** .92**----9. Trust .75**.77**.71**.73**.63**.82** .93**.93**----10. Commitment .71**.79**.68**.69**.60**.79** .85**.89**.91**----* p< .05; ** p< .01

PAGE 131

113 The correlation analysis revealed the high a ssociations between a ll of the variables in relationship maintenance strategies and relationship quality outcomes. Although none of the variance-inflating fact ors (VIF) for any of the independent variables is over 10, which is a determining criteria for multicoll inearity violation (Guj arati, 2003), it would be more appropriate to conduct path analysis when variable s are highly correlated. Due to the correlation among variables, stepwise regression analysis31 was conducted. The second and fourth research questions were tested through the regression analysis. RQ2: Which and to what extent are rela tionship maintenance strategies positively connected with which rela tionship quality outcomes? RQ4: Which of the four indicators represents relationship quality outcomes the most? The second research question asked to which and what degree relationship maintenance strategies positively affect which relationship quality outcomes. With reference to the relative influe nce of relationship maintenance strategies on the relational outcomes, the relational outcomes were regr essed on the maintenance strategy factors using step-wise regression analysis. Relationship maintenance strategies are independent variables, and each indicator of relationship quality outcomes is a dependent variable for the regression analysis. Therefore, the regression models are defined as follows: (1) Control Mutuality = 1 + 1Access + 2 Positivity + 3Openness + 4Sharing of Tasks + 5Networking + 6Assurances + 1 (2) Satisfaction = 2 + 7Access + 8 Positivity + 9Openness + 10Sharing of Tasks + 11Networking + 12Assurances + 2 31 Stepwise regression estimation is a method of selecting variables for inclusion in the regression model which starts by selecting the best pred ictor of the dependent variable. Additional independent variables are chosen based on the incr emental explanatory power they can add to the regression model. Independent va riables are added until their partial correlation coefficients are statistically significant (Hair et al., 1998). This method is often used as a partial remedy for highly correlated independent variables.

PAGE 132

114 (3) Trust = 3 + 13Access + 14 Positivity + 15Openness + 16Sharing of Tasks + 16Networking + 17Assurances + 3 (4) Commitment = 4 + 18Access + 19 Positivity + 20Openness + 21Sharing of Tasks + 22Networking + 23Assurances + 4 where S=standardized regression coefficients S= error term Each coefficient (= ) in equation 1 through 4 represents the magnitude of the effect of each strategy on each dimension of organization-public relationship. As shown in Table 4-22, the analysis re vealed several findings. First, the regression model of control mutu ality as a dependent variable was significant (F = 243.79, p = .002) and explained about 67 percent of the total variance. Among the six relationship maintenance strategy indicators, three indicators including assurances, access, and sharing of tasks, demonstr ated significant effects on cont rol mutuality. It should be noted that the magnitude effect of assura nces on control mutuality is 1.7 times bigger than that of access and 2.6 times bigger than th at of sharing of tasks. The other three indicators, positivity, openness, and networking, were excluded in the stepwise regression analysis because they do not significantly im pact control mutuality. It should be noted that the magnitude effect of assurances is bigger on control mutual ity than approximately 1.7 times that of access and 2.6 times that of shar ing of tasks. The other three indicators such as positivity, openness, and networking were excluded in the stepwise regression analysis because they do not signifi cantly impact on control mutuality. In addition, the regression analysis of satis faction as a dependent variable was also significant (F = 208.73, p = .025) and accounted for about 63 percent of the total variance.

PAGE 133

115 Out of the six relationship maintenance stra tegy indicators, the th ree indicators of assurances, positivity, and access produced signifi cant effects on satisfac tion. In terms of the magnitude of effects, assurances on satis faction are greater than positivity and access by about 1.5 times 3 times, respectively. Th e remaining three indicators of openness, sharing of tasks, and networking were excluded in the analysis. Third, the stepwise regression analysis of trust as a depe ndent variable was statistically significant (F = 267.36, p = .03) and explained approximately 69 percent of the total variance. Among the six relationshi p maintenance strategy indicators, the results suggested that the three strategy indicators, including assurances, positivity, and access, significantly influenced trust. In terms of the magnitude of the effects, assurances on trust are greater than access and positivity by approximately 3 times and about 4 times, respectively. The remaining three indicators, including openness, sh aring of tasks, and networking, were excluded in the analysis. Finally, the stepwise regression analysis of commitment as a dependent variable was also significant (F = 361.72, p = .00) and accounted for approximately 67 percent of the total variance. Among the six relationshi p maintenance strategy indicators, only two indicators, positivity and assurances, significantly affected commitment, and these showed the similar range of effects size on commitment ( = .43 for positivity and = .42 for assurances). The other four strategy indi cators, including access, openness, sharing of tasks, and networking, were excluded in the analysis. In summary, all tested stepwise regre ssion analysis produced significant results. First, the public’s percepti ons of maintaining a relations hip through assurances was the primary predictor of control mutuality, sa tisfaction, and trust. Second, the public’s

PAGE 134

116 evaluation of positivity best predicted the public’s perception of commitment in the relationship. Third, the public’s evaluation of the organization’s e fforts in maintaining relationships by positivity and access were impor tant (i.e., second or third), and positivity and access influenced the three relationship quality outcome indicators of control mutuality, satisfaction, and trust. The public ’s assessments of openness and networking were not significantly associated with a ny of the relationship quality outcomes. Additionally, the amount of expl ained variance of the relational outcomes is consistent and strong, with adjusted R2 ranging between .63 and .69. Overall, the ways in which the members of the public perceived an organiza tion’s relationship maintenance strategies affected their evaluation of relationship qual ity outcomes of control mutuality, trust, satisfaction, and commitment. Table 4-22. Stepwise regre ssion analysis of relationshi p maintenance strategies on relationship quality outcomes Relationship Quality Outcome Variables Relationship Maintenance Strategies factor Control MutualitySatisfaction Trust Commitment Access .27 (2).14 (3).13 (3) Positivity .28 (2).19 (2) .43 (1) Openness Sharing of Tasks .17 (3) Networking Assurances .45 (1).41 (1).55 (1) .42 (2) F 243.79**208.73*267.36* 361.72** R2 .67.63.69 .67 Adjusted R2 .67.63.69 .67 Note: Numbers outside parentheses indicate standardized . Numbers in parentheses refer to the order of inclusion in step-wise regression equation. Each of the relationship quality outcome indicators of control mutuality, satisfaction, trus t, and commitment were the dependent variables for each regression analysis. * p< .05; ** p< .01 The research question addressing which indi cator best represents the relationship quality outcomes among the four indicators was tested using stepwise regression analysis. In this analysis, the four relationship qua lity outcome indicators were independent

PAGE 135

117 variables, and evaluation of overall relationship quality was a dependent variable. Therefore, the regression mode l is defined as follows: Overall Relationship = + 1Control Mutuality + 2 Satisfaction + 3Trust + 4Commitment As shown in Table 4-23, the regres sion model was significant (F = 340.04, p = .015) and explained about 74 percent of the total variance. Among the four relationship indicators, the three indicators of commitmen t, satisfaction, and trust produced significant effects on th e overall relationship quality while the magnitude of effects was different. Among the three signif icant indicators, commitment produced the highest standardized coefficient ( = .69), followed by satisfac tion and trust, indicating that commitment is the best predictor of overall relationship quality ( = .45 and = .33, respectively). It is noteworthy that the e ffect size of commitment on overall quality relationship is bigger than sa tisfaction by approximately 1.5 times and trust by 2 times. The other indicator, control mutuality, was excluded in the analysis. Table 4-23. Stepwise regre ssion analysis of relationshi p quality outcomes on overall relationship Independent Variables (Relational Outcome Indicator) Standardized Control Mutuality Satisfaction .45 (2) Trust .33 (3) Commitment .69 (1) F 340.04* R2 .74 Adjusted R2 .74 Note: Numbers outside parentheses indicate standardized . Numbers in parentheses refer to the order of inclusion in the step-wise regression equation. Overall, relationship quality was the dependent variable. * p< .05

PAGE 136

118 Testing the Model of Relationship between Relationship Maintenance Strategies and Relationship Quality Outcomes Although the regression analysis provided easily understand able results related to the second and fourth research questions, the current study performed path analysis for the two research questions because of the drawback of regression analysis, multicollinearity. When using regression analysis, it is not easy to deal with multicollinearity problems. As shown in the correlation analysis, the variables used in this study have strong associations. Access Control Mutuality Assurances Networking Sharing of Tasks Openness Positivity Commitment Trust Satisfaction Figure 4-3. Initial model of th e relationship between relations hip maintenance strategies and relationship quality outcomes

PAGE 137

119 Path analysis, however, can be a practical method to reduce the multicollinearity problem since the technique allows in terdependent relationships am ong independent variables. Path analysis, thus, utilizes simple bivariat e correlations by specifyi ng the relationships in a series of regressi on-like equations that can then be estimated by determining the amount of correlation attributable to each eff ect in each equation simultaneously (Hair et al., 1998). To test the model w ith all its paths, the study us es a goodness-of-fit test from a structural equation program. In public relations scholarship, it has b een suggested that the six maintenance strategies—access, positivity, openness, shar ing of tasks, networking, assurances—tend to produce relationship quality outcomes such as control mutuality, sa tisfaction, trust, and commitment (J. E. Grunig & Huang, 2000; Hon & J. E. Grunig, 1999). Based on this suggestion, this study tested to what extent the six identi fied relationship maintenance strategy variables posit ively influenced which relationshi p quality outcomes. In this model testing, relationship maintenance strategies are exogenous variables,32 and relationship quality outcome s are endogenous variables33. Figure 4-3 illustrates the initial model of relationship between relationship maintenan ce strategies and relationship quality outcomes. Table 4-24 shows the results of path mode l analysis for the relationship between each variable of relationship maintenance st rategies and each variable of relationship quality outcomes. Access significantly aff ected control mutuality. Positivity had a significant impact on control mutuality, satis faction, and trust. Sharing of tasks 32 Exogenous variables are those that do no t have causes specified in the model. 33 Endogenous variables are those that have cause s specified in the model. They are dependent variables here.

PAGE 138

120 significantly influenced contro l mutuality and satisfaction. Assurances had a significant impact on all of the relationship quality outco me indicators. Inte restingly, openness and networking did not have any si gnificant association with a ny relationship variables. Table 4-24. Path model of relationship mainte nance strategies and relationship quality outcomes Path Standardized coefficient Standardized Error Access Control mutuality .20***.06 Access Satisfaction .12 .06 Access Trust .10 .06 Access Commitment .01 .06 Positivity Control mutuality .07***.06 Positivity Satisfaction .24* .07 Positivity Trust .14***.06 Positivity Commitment .40 .06 Openness Control mutuality .02 .05 Openness Satisfaction .04 .06 Openness Trust .08 .05 Openness Commitment .07 .05 Sharing of tasks Control mutuality .16** .06 Sharing of tasks Satisfaction .13* .06 Sharing of tasks Trust .09 .06 Sharing of tasks Commitment .02 .06 Networking Control mutuality .00 .05 Networking Satisfaction -.04 .05 Networking Trust -.01 .04 Networking Commitment -.03 .05 Assurances Control mutuality .36***.08 Assurances Satisfaction .49***.07 Assurances Trust .40***.07 Assurances Commitment .42***.07Note: *** p< .001; ** p< .05; * p< .01 Figure 4-4 provides the results of the stat istical tests for the individual paths, including magnitude and signifi cances of the coefficients for the model. Overall, the analysis is consistent with the stepwise regr ession analysis. Access, positivity, sharing of tasks, and assurances significantly affected control mutuality even though the effect sizes are different. The effect size of assurances on control mutuality ( = .36, p<.001) is

PAGE 139

121 bigger than 1.8 times that of access ( = .20, p <.001), about that 2.3 ti mes of sharing of tasks ( = .16, p <.05), and 5.1 times that of positivity ( = .07, p <.001). The remaining maintenance strategies, openness and networ king, did not significantly affect control mutuality. Positivity, sharing of tasks, and assurances presented significant impacts on satisfaction. The effect size of assurances on satisfaction ( = .49, p <.001) is approximately 2 times bigger than that of positivity ( = .24, p <.01) and 4 times larger than that of sharing of tasks ( = .13, p <.01). The remaining maintenance strategies— access, openness, and networking—were not sign ificant predictors of satisfaction. Only two relationship maintenance strate gy variables, positivity and assurances were significant predictors of trust, even t hough the effect size of assurances on trust ( = .40, p<.001) is approximately 3 times more than positivity ( = .14, p <.001). The remaining four maintenance strategies— access, openness, sharing of tasks, and networking—were not significan t predictors of trust. Only assurances showed significant impact on commitment ( = .42, p<.001), and all of the remaining maintenance strategi es were not significant predictors of commitment. The two strategy variables of openness and networking did not significantly influence any relationship quality outcome variable. Interestingly, networking is negatively associated with sa tisfaction, trust, and commitme nt even though the magnitude and significance of this vari able is not substantial. In summary, assurances were found to be most significant among the six relationship maintenance strategies across all relationship quality outcome variables.

PAGE 140

122 Positivity, sharing of tasks, and access had some degree of significant impact on some relationship quality variables. However, openness and networking did not have any significant effect on any relationshi p quality outcome variables. Access Control Mutuality Assurances Networking Sharing of Tasks Openness Positivity Commitment Trust Satisfaction .20 (.06)*** .07 (.64)*** .24 (.06)* .14 (.06)*** .16 (.06)** .13 (.06)* .36 (.08)*** .49 (.07)*** .40 (.07)*** .42 (.07)*** Note : Dotted lines indicate non-significant paths. Solid lines indicate significant paths. The numbers outside parentheses indicate standardized coefficient and those in parentheses indicate standardized error. * p< .05; ** p< .01; *** p< .001 Figure 4-4. Final model linki ng relationship maintenance st rategies and relationship quality outcomes Best Indicator of Overa ll Relationship Quality This study also tested the relative infl uence of relationship quality outcome indicators on overall relations hip quality using path analysis . In this model testing, the

PAGE 141

123 four relationship quality outcome indicator s are exogenous variable s, and the public’s evaluation of the overall re lationship quality with the organization is an endogenous variable. Table 4-25 shows the results of a path analysis with a maximum likelihood estimation for the relationship between four relationship quality out come variables and the overall relationship quality variable. Cons istent with the results of the stepwise regression analysis, path analysis indicated that commitment had the strongest and most significant impact on overall relationship quality ( = .42, p< .001). Also, satisfaction and trust significantly affected overall relationship quality ( = .25 and = .18, respectively). It should be noted that e ffect size of commitment on overall relationship quality is 1.68 times greater th an that of satisfaction and 2.33 times greater than that of trust. However, control mutuality did not show any significant effect on overall relationship quality. Figure 4-5 displa ys the results of the path model. Table 4-25. Path model of relationship quality outcomes on overall relationship quality Path Standardized coefficient Standardized Error Control mutuality Overall Relationship Quality .04 .13 Satisfaction Overall Relationship Quality .25** .14 Trust Overall Relationship Quality .18* .15 Commitment Overall Relationship Quality .42*** .11Note: *** p< .001; ** p< .05; * p< .01

PAGE 142

124 Control Mutuality Satisfaction Trust Commitment .42 (.11)*** .18 (.15)* .25 (.14)** Overall Relationship Quality Note: Dotted lines indicate non-significant paths. Solid lines indicate significant paths. The numbers outside parentheses indicate standardized coefficients and those in parentheses indicate standardized errors. * p < .05; ** p <.01; *** p <.001 Figure 4-5. Path model of relationship qual ity outcome on overall relationship quality Structural Equation Modeling The following research questions and hypothe ses were tested and answered through path analysis and struct ural equation modeling: RQ5: How do relationship quality outco me indicators affect each other? o HP-1: The degree of satisfaction will positively influence the degree of trust. o HP-2: The degree of trust will positively influence the degree of commitment. RQ6: What is a model linking relationsh ip quality perception, attitude, and behavioral intentions? o HP-3: A public’s perceptions of its re lationship with the organization will influence the public’s attitude toward the organization. o HP-4: A public’s attitude toward the organization will influence the public’s behavioral intentions.

PAGE 143

125 o HP-5: For a public experiencing low-involvement, the public’s relationship perception will positiv ely influence their behavioral intentions. o HP-6: For a public experiencing lowinvolvement, the public’s behaviors will positively influen ce their attitude toward the organization. RQ7: What is a comprehensive model lin king relationship maintenance strategies, relationship quality outcomes, a ttitude, and supp ortive behavior? The analysis of the proposed models, incl uding path models and causal models, on principal component scores was conducted. Principal component scores were employed for the 12 factors: six for maintenance strate gies, four for relationship quality outcomes, one for attitude, and one for behavioral inte ntions, meaning that the twelve constructs were treated as twelve observed variables. This approach is helpful for accomplishing model parsimony and ease of convergence. The proposed model was tested through a structural equation model using maximum likelihood in AMOS 6.0 (Byrne, 2001), which also provides a set of modification indices for each possible parameter not specified in the original theoretical model. The maxi mum likelihood (ML) is the most extensively used model estimator (Bollen, 1989; Chou & Be ntler, 1995) and is a ppropriate for large samples, because normality of the sample is its basic assumption, leading to normal error distribution. Relationship among Relationship Quality Outcome Indicators The relationship among relationship quality outcome indicators was tested using path analysis with maximum likelihood estimati on. In this model te sting, satisfaction is an exogenous variable, and trust and commitment are endogenous variables. Table 4-26 shows the results of a path analysis with maximum likelihood estimation for the linkages between the three relationship indicators— satisfaction, trust, and commitment. Both of the hypotheses were strongly supported. For the first

PAGE 144

126 hypothesis, satisfaction was found to be a st rong and significant predictor of trust ( = .92, p <.001). Also, trust strongly and si gnificantly affected commitment ( = .90, p< .001). Table 4-27 demonstrates that the import ant fit indices such as CFI, GFI, NFI, and RMR fit the path model’s excellent leve l of satisfaction. Although a couple of fit indices (e.g., ratio of chi-square/df, RMSEA) di d not meet the criteria, it can be said that the proposed model fits the data based on othe r fit indices. Figure 4-6 shows the results of the path model. Table 4-26. Path model of the relations hip among relationship quality outcome indicators Path Standardized coefficient Standardized Error Conclusion HP-1: Satisfaction Trust .93***.02 Strongly supported HP-2: Trust Commitment .90***.02 Strongly supportedNote: *** p <.001 Table 4-27. Fit measures of the relati onship among relationship quality outcome indicators Fit Index Criteria Fit Statistics Chi-square > .05 26.97 ( p< .001; df = 1) Chi-square/df < 5 26.97 Comparative Fit Index (CFI) > .9 .98 Goodness of Fit Index (GFI) > .90 .95 Normed Fit Index (NFI) > .90 .98 Root Mean Squared Error Residual (RMSEA) < .08 .27 Root Mean square Residual (RMR) < .05 .02 Satisfaction Commitment Trust .93 (.20)***.90 (.23)*** HP-1 HP-2 Note: The solid lines indicates significant paths. Th e numbers outside the parentheses indicate standardized coefficients and those in parenthese s indicate standardized errors. *** p <.001 Figure 4-6. Path model of the relationship among relationship quality outcome indicators

PAGE 145

127 Linkages among Relationship Quality Outcomes, Attitude, and Behavioral Intentions The research question and hypotheses relate d to the linkages among perceptions of organization-public relationship, attitude, and behavioral inte ntions were tested using structural equation modeling with maximum likelihood estimation. The model tested had four exogenous variables—control mutualit y, satisfaction, trust, and commitment—and two endogenous variables—attitude and behavioral intentions. Table 4-28 shows the results of structural equation models for the linkages between relationship quality outcomes, attitude, and behavioral inte ntions (See Figure 4-7 for the initially proposed model). Hypothesis three is related to the link be tween perceptions of the organization-public relations hip and attitude, and hypothesis four is related to the link between attitude and behavioral intentions. Both hypotheses were st rongly supported. Of the four relationship indicators, commitment most strongly and si gnificantly affected attitude, followed by trust ( = .33 and = .28, p <.05). Also, attitude was found to be a very strong predictor of behavioral intentions ( = .79, p <.001). As shown in Table 4-29, according to the given fit indice s, the fit of the model, incl uding CFI, GFI, and NFI, met the satisfaction level. Although it can be said that the proposed m odel fits the data, it was determined that the model should be revised based on the modi fication index and theo retical justification. The modification index suggested that addi ng a path from commitment to behavioral intentions improves the model significantly. More importantly, Garbarino and Johnson (1999) discovered that customers’ perceptions of commitment directly affected future behavioral intentions. Als o, another study suggested that the perception of commitment is most closely related to behavior (Becker, Randall, & Riegel, 1995).

PAGE 146

128 As shown in Table 4-28, the newly added path (commitment behavioral intentions) in the revised model had significant impact ( = .50, p <.001). Moreover, compared to the original model, the fit m easures were greatly improved: from .97 to .99 for CFI, from .93 to .99 for GFI, and from .97 to .99 for NFI as appeared in Table 4-29. Also, all of the non-significant fits in the originally tested model reached significant fit indices as follows: from 22.97 to 2.57 for the ratio of Chi-square to the degree of freedom, from .25 to .07 for RMSEA, and from .06 to .01 for RMR. According to the statistical analysis, the revised model was bette r than the initial model. Additionally, the revised model demonstrates almost perfect f it in terms of fit measures and theoretical support. Figure 4-8 shows the revised model. Table 4-28. Tested model linkages among relati onship quality perception, attitude, and behavioral intentions Path Initial Model Revised Model Control mutuality Attitude .14 (.08).14 (.08) Satisfaction Attitude .14 (.09).14 (.09) Trust Attitude .28** (.09).28** (.09) Commitment Attitude .33***(.07).33***(.07) Attitude Behavioral Intentions .79***(.03).37***(.03) Commitment Behavioral Intentionsa N/A .50***(.05)Note: The numbers outside the parentheses indicate standardized coefficients. The numbers in the parentheses in dicate standardized errors. *** p <.001; ** p <.05; * p <.01 a The new path is added in the revised model. Table 4-29. Fit measures of linkages among relationship quality perception, attitude, and behavioral intentions Fit Index Criteria Fit Statistics (Initial Model) Fit Statistics (Revised Model) Chi-square > .05 91.88 ( p <.001; df = 4)7.70 ( p <.01; df = 3) Chi-square/df < 5 22.972.57 CFI > .9 .97.99 GFI > .90 .93.99 NFI > .90 .97.99 RMSEA < .08 .25.07 RMR < .05 .06.01Note : CFI= Comparative Fit Index; GFI= Goodness of Fit Index; NFI= Normed Fit Index; RMSEA=Root Mean Squared Error Residual; RMR= Root Mean square Residual

PAGE 147

129 Control Mutuality Attitude Commitment Trust Satisfaction Behavioral Intention .28 (.09)** .33 (.06)*** .787 (.03)***Note: Dotted lines indicate non-significant paths. Solid lines indicate significant paths. The numbers outside parentheses indicate standardized coefficients, and those in parentheses indicate standardized errors. ** p <.01; *** p <.001 Figure 4-7. Initial model linking relationship qu ality outcomes, attitude and behavioral intentions Control Mutuality Attitude Commitment Trust Satisfaction Behavioral Intention .28 (.09)** .33 (.06)*** .37 (.03)*** .50 (.05)***Note: Dotted lines indicate non-significant paths. Solid lines indicate significant paths. The numbers outside parentheses indicate standardized coefficients, and those in parentheses indicate standardized errors. ** p <.01; *** p <.001 Figure 4-8. Revised model linki ng relationship quality outcomes, attitude and behavioral intention Linkages among Relationship Quality Perception, Attitude, and Behavioral Intentions of Low-Involved Group This study tested the last two hypothese s regarding the se quential order of relationship perception, attitude , and behavioral intentions on a group of a public who

PAGE 148

130 indicated a relatively short-term relationshi p with the organization, which was defined as low-involvement in this study. Hypothesis five indicates that for a low-involvement public, relationship perceptions will positively affect behavior,34 and hypothesis six indicates that attitude is positively affected by behavior. The model tested had four exogenous variables—control mutuality, satis faction, trust, and commitment—and two endogenous variables—attitude a nd behavioral intentions. Table 4-30 shows the results of structural equation models for the linkages between perceptions of relationship quality outcomes , behavior, and attitudes for a group of lowinvolvement public. The two proposed hypot heses were partially supported. Among four relationship indicators, only perceptions of commitment str ongly and significantly affected behavior ( =.62, p <.001). Also, behavior was show n to be a strong predictor of attitude ( =.81, p <.001). According to the given fit i ndices, Table 4-31 displays that the fit of the model met the satisfaction level, in cluding CFI, GFI, and NFI. Therefore, it can be said that the proposed model adequately fits the data (See Figure 4-9 for a visual illustration of the model). Table 4-30. Tested model linkages among perc eptions of relationship quality perception, attitude, and behavioral inte ntions on low-involvement Path Standardized Coefficient Standardized Error Control Mutuality Behavior .16.135 Satisfaction Behavior .22.156 Trust Behavior -.15.170 Commitment Behavior .62***.111 Behavioral Intentions Attitude .81***.043Note: The number outside the parentheses indicates standardized coefficient and those in parentheses indicate standardized error. *** p <.001; ** p <.05; * p <.01 34 In this study, behavior was measured as behavi oral intention because both are categorized into a conative. Therefore, behavior and beha vioral intention will be treated equally.

PAGE 149

131 Table 4-31. Fit measures of linkages among relationship quality outcome, attitude, and behavioral intentions on low-involved public Fit Index Criteria Fit Statistics Chi-square > .05 61.483 ( p <.001; df=4) Chi-square/df < 5 15.371 Comparative Fit Index (CFI) > .90 .964 Normed Fit Index (NFI) > .90 .962 Root Mean Squared Error Residual (RMSEA) < .08 .286 Root Mean square Residual (RMR) < .05 .086 Control Mutuality Behavior Commitment Trust Satisfaction Attitude .62 (.11)*** .81 (.04)***Note: Dotted lines indicate non-significant paths. Solid lines indicate significant paths. The numbers outside parentheses indicate standardized coefficient and those in parentheses indicate standardized error. ** p <.01; *** p <.001 Figure 4-9. The model tested model linka ges among relationship quality outcomes, attitude, and behavioral inte ntions on low-involvement Comprehensive Model Linking Maintenan ce Strategies, Relationship Quality Outcomes, Attitude and Behavioral Intentions This study also tested the comprehensiv e model linking relationship maintenance strategies, relationship quality outcomes, attitude, and behavioral intentions. In the model, six relationship maintenance strate gy variables (access, positivity, openness, sharing of tasks, networking, and assurances), four relationship quality outcome variables (control mutuality, satisfaction, tr ust, and commitment), as well as attitude and behavioral

PAGE 150

132 intentions are observed exoge nous variables. The unobserved exogenous variable is the relationship maintenance strategy construct, while the relationship quality outcome construct is an unobserved endogenous variab le. Here, the two latent constructs, relationship maintenance strate gies and relationship quality outcomes, are represented by composite indicators (i.e. access, positivity, control mutuality, satisfaction, etc.) that include multiple items with the exception of at titude and behavioral intentions, which are represented by single-item com posite indicators. Bollen (1989) suggested that even though multiple-item indicators are more desirable than singleitem indicators in terms of internal consistency and reliability, single-it em indicators have been commonly used in the structural equation modeling approach and have produced acceptable model fit, because the effects of unreliabi lity on the analysis are addressed using a latent variable model. As shown in Table 4-32, all effects of th e paths linking relationship maintenance strategies, relationship quality outcomes, attit ude, and behavioral intentions were strong and significant. Specifically, the effect of relationship maintenance strategy on relationship quality outco me was the strongest ( = .88, p <.001), followed by relationship quality outcomes to attitude ( = .87, p <.001) and attitude to be havioral intentions ( = .79, p <.001). Moreover, as Table 4-33 shows, the fit of the proposed comprehensive model was generally satisfactory according to the given fit indi ces (the ratio of 2/df, CFI, GFI, NFI, and RMR). Although RMSEA (.10) di d not meet the criteria, it is close to doing so. Therefore, it can be noted that acco rding to the several fit indices, the tested comprehensive model fits the data.

PAGE 151

133 In summary, the direct re lationship between relationshi p maintenance strategy and relationship quality outcome was stronge r than any other relationship in the comprehensive model. Also, all the paths in the model are strong and significant with desirable fit measures for the model. Table 4-32. Testing comprehensive model Path Standardized Coefficient Standardized Error Relationship Maintenance Strategy Access .89*** .03 Relationship Maintenance Strategy Positivity .91*** .03 Relationship Maintenance Strategy Openness .85*** .03 Relationship Maintenance Strategy Sharing of Tasks .87*** .03 Relationship Maintenance Strategy Networking .77*** .04 Relationship Maintenance Strategy Assurances .95a Relationship Maintenance Strategy Relationship Quality Outcomes .88*** .03 Relationship Quality Outcomes Control Mutuality .95a Relationship Quality Outcomes Satisfaction .96*** .03 Relationship Quality Outcomes Trust .97*** .02 Relationship Quality Outcomes Commitment .92*** .03 Relationship Quality Outcomes Attitude .87*** .03 Attitude Behavioral Intentions .79*** .03Note: *** p< .001; ** p< .05; * p< .01 a Values were not calculated because loading was set to 1.0 to fix construct variance. Table 4-33. Fit measures of comprehensive model Fit Index Criteria Fit Statistics Chi-square > .05 261.38 ( p< .001; df = 53) Chi-square/df < 5 4.93 Comparative Fit Index (CFI) > .9 .96 Goodness of Fit Index (GFI) > .90 .90 Normed Fit Index (NFI) > .90 .96 Root Mean Squared Error Residual (RMSEA) < .08 .10 Root Mean square Residual (RMR) < .05 .05

PAGE 152

134 Relationship Maintenance Strate g ies Relationship Quality Outcome Attitude Behavioral Intention Control Mutuality Commitment Trust Satisfaction Access Positivity Openness Sharing of tasks Networking Assurances .89 (.03)***.91 (.03)***.85 (.03)***.87 (.03)***.77(.04)***.95 a .88(.03)***.87(.0)***.79(.03)***.95 a.96 (.03)***.97 (.02)***.92 (.03)***Note: *** p< .001; ** p< .05; * p< .01 a Values were not calculated because loading was set to 1.0 to fix construct variance. The r ectangles indicate observed variables, and the ovals represent latent variables. Figure 4-10. Comprehensive model linking ma intenance strategies, relationship quality outcomes, attitude, and behavioral inten tions

PAGE 153

135 Post Hoc Analysis A couple of post hoc analyses were performed based on the length of the respondents’ relationship with the organization: 1) the model linking relationship quality, attitude and behavioral intentions, 2) the comprehensive model linking relationship maintenance strategies, relationship quality outco mes, attitude, and behavioral intentions. This study asked the respondents how many year s they had had a relationship with the organization. The average length of the respond ents’ relationships is 25 years, which is a relatively long time period. In these additi onal analyses, the respondents were divided into two groups based on the length of th eir relationship with the organization. The respondents who had associations shorter than 25 years were put into the first group (N = 177), and the remaining respondent s with a relationship longer than 25 years were placed into the second categ ory (N = 150). Revised Model Linking Relationship Qua lity Outcomes, Attitude, and Behavioral Intentions The first additional analysis is the revised model linking relationship quality outcomes, attitude, and behavioral intenti ons across the two groups. Table 4-34 shows the results of the two models – Model 1, based on respondents with organizational relationships shorter than 25 years, and Model 2, which deal t with respondents who had associations for longer than 25 years. In the first model, among the four relationship quality outcome indicators, satisfaction is the best predictor of attitude ( = .37, p< .001), followed by commitment ( = .26, p< .05). The remaining tw o indicators, control mutuality and trust, are not significant pred ictors of the attitude. Attitude and commitment strongly and significantly affected behavioral intentions ( = .43 and = .47, respectively, p< .001).

PAGE 154

136 In the second model looking at the longe r relationship group, trust is the best predictor of attitude among the four relati onship quality outcome indicators, followed by commitment ( = .45, p< .05 and = .34, p< .001, respectively). The remaining two relationship indicators, control mutuality a nd satisfaction, did not significantly impact attitude. Both attitude and commitment are strong and significant predictors of behavioral intentions ( = .41 and = .43, p< .001 for both). Commitment was the strongest and most significant predictor of attitude and behavioral intentions across th e two models. Also, attitude produced substantial impacts on behavioral intentions for both models. Table 4-34. Tested model linkages among relati onship quality outcomes, attitude, and behavioral intentions on th e length of the relationship Path Model 1 Model 2 Control mutuality Attitude .10 (.13).21 (.13) Satisfaction Attitude .37** (.15)-.12 (.13) Trust Attitude .13 (.16).45** (.15) Commitment Attitude .26** (.10).34***(.11) Attitude Behavioral Intentions .43***(.07).41***(.08) Commitment Behavioral Intentions .47***(.06).43***(.09)Note: The numbers outside the parentheses indicate standardized coefficients. The numbers in the parenthese s indicate standardized errors. *** p< .001; ** p< .05 Table 4-35. Fit measures of linkages among relationship quality outcomes, attitude, and behavioral intentions on the length of the relationship Fit Index Criteria Fit Statistics (Model 1) Fit Statistics (Model 2) Chi-square > .05 2.09 ( p = .555; df = 3)12.668 ( p< .05; df = 3) Chi-square/df < 5 .704.22 CFI > .9 1.00.99 GFI > .90 .99.97 NFI > .90 .99.99 RMSEA < .08 .00.15 RMR < .05 .00.02Note : CFI= Comparative Fit Index; GFI= Goodness of Fit Index; NFI= Normed Fit Index; RMSEA=Root Mean Squared Error Residual; RMR= Root Mean Square Residual

PAGE 155

137 As shown in Table 4-35, both models ade quately fit the data . Although the number of significant paths in both m odels is the same, the fit measures of the models indicate that the first model demonstrates a better fit. The Comprehensive Model Linking Relationship Maintenance Strategies, Relationship Quality Outcomes, Atti tude, and Behavioral Intentions The second additional analysis, which is the comprehensive model linking relationship maintenance strategies, rela tionship quality outcomes, attitude, and behavioral intentions, was tested across the two groups. As displa yed in Table 4-36, the comprehensive models across the two groups pr oduced similar results in terms of the number of significant paths and fit measur es, although minor differences of magnitude exist. Therefore, it can be said that the te sted comprehensive model is more likely to apply to different groups of publics. Table 4-36. Testing the comprehensive m odel on the length of the relationship Path Model 1 Model 2 Relationship Maintenance Strategy Access .90***(.04) .890***(.05) Relationship Maintenance Strategy Positivity .91***(.04) .898***(.05) Relationship Maintenance Strategy Openness .87***(.05) .823***(.06) Relationship Maintenance Strategy Sharing of Tasks .89***(.04) .859***(.05) Relationship Maintenance Strategy Networking .80***(.05) .742***(.06) Relationship Maintenance Strategy Assurances .96a .946a Relationship Maintenance Strategy Relationship Quality Outcomes .85***(.05) .898***(.06) Relationship Quality Outcomes Control Mutuality .95a .953a Relationship Quality Outcomes Satisfaction .97***(.03) .939***(.04) Relationship Quality Outcomes Trust .98***(.03) .977***(.03) Relationship Quality Outcomes Commitment .93***(.04) .915***(.04) Relationship Quality Outcomes Attitude .85***(.05) .852***(.05) Attitude Behavioral Intentions .81***(.05) .760***(.05)Note: The numbers outside the parentheses indicate standardized coefficients. The numbers in the parentheses indicate standardized errors. a Values were not calculated because loading was set to 1.0 to fix construct variance. *** p< .001; ** p< .05

PAGE 156

138 Table 4-37. Fit measures of the comprehens ive model on the length of the relationship Fit Index Criteria Fit Statistics (Model 1) Fit Statistics (Model 2) Chi-square > .05 213.93 ( p< .001; df = 53)165.08 ( p< .001; df = 53) Chi-square/df < 5 4.043.12 CFI > .9 .95.95 GFI > .90 .84.85 NFI > .90 .93.93 RMSEA < .08 .13.12 RMR < .05 .07.05Note: CFI= Comparative Fit Index; GFI= Goodness of Fit Index; NFI= Normed Fit Index; RMSEA=Root Mean Squared Error Residual; RMR= Root Mean Square Residual

PAGE 157

139 CHAPTER 5 DISCUSSION AND CONCLUSION This study explores two central concepts—r elationship maintenance strategies and relationship quality outcomes—and their measur es. In addition, this study investigates the relationships between or among 1) re lationship maintenance strategies and relationship quality outcomes; 2) the influence of sequence in relationship indicators; 3) the influence of order in rela tionship perceptions, attitude, an d behavioral intentions; and 4) relationship maintenance strategies, relationship quality outcomes, attitude, and behavioral intentions. This chapter consists of three sections. The first section summarizes the findings of the study, followed by a discussion of theoretical and managerial implications. The third section identifies the study's limitations as well as notes possible directions for future research. Summary of Results This section summarizes the findings from statistical analysis, including 1) descriptive statistics, 2) corre lation analysis between relationship maintenance strategies and relationship quality outcomes, 3) measures of relationship maintenance strategies and relationship quality outcomes, 4) causal links between relatio nship maintenance strategies and relationship quality outcomes, 5) best indicators of relationship quality outcomes of overall relationship quality, 6) influential order among relationship quality outcome indicators, 7) linkages among relationship quali ty perception, attitude, and behavior, 8)

PAGE 158

140 the comprehensive model linking relationship ma intenance strategies, relationship quality outcomes, attitude, and behavioral inte ntions, and 9) post hoc analysis. Descriptive Statistics Relationship maintenance strategies The members who participated in this study evaluated the va rious relationship maintenance strategies used by the Florida Farm Bureau (FFB) as daily communication activities. Participants noted that the FFB put the most effort into maintaining quality relationships with publics through courteous communications, but did little to provide information about the organization’s nature and purpose. Members asserted that the FFB’s second greatest effort was in sharing proj ects and problems of mutual interest with their publics. Participants evaluated the FFB’s attempts to communicate with members and determined that the organization’s third gr eatest effort was in ensuring members that they and their concerns were addressed. Me mbers also noted the organization’s attempts to provide channels for communication, such as contact information. Lastly, the members agreed less that the organization’s efforts to build coalitions was a helpful maintenance strategy. Relationship quality outcomes Members of the public participating in this study generally evaluated their relationship with the Florida Farm Bureau positively for all rela tionship indicators, including control mutuality, satisfaction, trust, and commitment. Commitment rated highest in participants’ evaluations, followed by satisfaction, trust, and control mutuality. Because they strongly believed the organization made an effort to maintain positive and lasting relationships with them, the study pa rticipants are likely to commit to the

PAGE 159

141 organization-public relationship. However, th ey believed that comparatively more power balance is necessary. Correlation Analysis Correlation analysis showed that co ntrol mutuality displayed significant associations with all relationship maintenance strategies. In particular, control mutuality had the strongest association wi th assurances. This relations hip can be understood if an organization made a communication effort to as sure the members of its publics that their concerns were being attended to, the publics are more likely to have higher levels of power balance between themselves and the organization. Like control mutuality, satisfaction ha d statistically and significantly high correlations across all relationship maintenance strategies. In partic ular, satisfaction had the strongest relationship with assurances. Th is finding might imply that the degree of an organization’s efforts to assure the members of publics that they a nd their concerns are attended to is related to the degree of memb er satisfaction in their relationship with the organization. Trust had significant associations with al l of the relationship maintenance strategy indicators. Similar to control mutuality a nd satisfaction, trust also had the strongest association with assurances. Additionally, the indicator also was highly correlated with other relationship maintenance strategies including positivity, access, sharing of tasks, and openness. The last variable of relationship qualit y outcomes, commitment, also displayed strong associations with all st rategy variables. Unlike the results of other relationship quality outcome variables, commitment ha d the strongest associations with both positivity and assurances. Both the organizational concerns of making communication

PAGE 160

142 activities enjoyable and attempting to assure th e publics that their concerns are attended to are related to the degree of commitment. All relationship quality outcome factors produced the least associ ation with networking. Measures of Relationship Maintenance Strategies and Relationship Quality Outcomes A primary purpose of this study was to construct the means for measuring both relationship maintenance stra tegies and relationship quality outcomes. Relationship maintenance strategy measures were devel oped based on the guide lines provided by Hon and J. E. Grunig’s (1999) study, while existi ng relationship qualit y outcome measures were refined using multiple-item measurement procedures suggested by Spector (1992). These two constructed measures were furthe r refined using factor analysis. Factor analysis suggested the inclusion of 24 items in the final relationship maintenance strategy measure, consisting of four items for access, five items for positivity, four items for openness, four items for sharing of tasks, th ree items for networking, and four items for assurances. The final measurement items app ear in Table 5-1. Factor analysis also proposed including 28 items in the final re lationship quality outcome measurement, consisting of eight items for control mutualit y, eight items for satisfaction, seven items for trust, and five items for commitment. The final measurement items are provided in Table 5-2. Confirmatory factor analyses we re used to evaluate the hypothesized factor structure and showed that measures for relationship maintenance strategies and relationship quality outcome s both had reliable and va lid factor structures.

PAGE 161

143 Table 5-1. Final measurement items for relationship maintenance strategies Access 1. _____ provides members with adequate contact information. 2. _____ provides members opportunities to meet its staff. 3. When members have questions or concerns, _____ is willing to an swer their inquiries. 4. _____ provide members with adequate contact information for specific staff on specific issues. Positivity 1. The member benefits _____ provide s are important to members. 2. Receiving regular communications from _____ is beneficial to members. 3. _____’s communication with members is courteous. 4. _____ attempts to make its intera ctions with members enjoyable. 5. _____ is cooperative when handli ng disagreements with members. Openness 1. _____’s Annual Report is a valuab le source of information for members about what _____ has done. 2. _____ shares enough information with memb ers about the organization’s governance. 3. _____’s member meetings are a valuable way for members to communicate their opinions to _____. 4. The issue briefings _____ provides help members understand the issues. Sharing of Tasks 1. _____ works with members to develop soluti ons to problems that benefit members. 2. _____ is involved in managing community issues that members care about (e.g., disaster relief, environment protection). 3. _____ works effectively to resolve regulatory issues its members are facing such as Pesticide or food safety issues. 4. _____ and members do not work well togeth er at solving joint problems. [R] Networking 1. _____ effectively builds coalitions with groups that impact members. 2. The coalitions that FFB forms with other agricultural groups be nefit _____ members. 3. The Ag Coalition for legislativ e activities that _____ is involved in is helpful to _____ members. Assurances 1. _____ makes a genuine effort to provide pe rsonal responses to members’ concerns. 2. _____ communicates the importance of members. 3. _____’s policy development process allows memb ers adequate opportunity to raise an issue and propose a solution. 4. When members raise concerns, __ ___ takes these concerns seriously. Note: The blanks need to be replaced with the name of the organization. [R] indicates reverse-coding. Table 5-2. Final measurement items for relationship quality outcomes Control Mutuality 1. _____ believes the opinions of members are legitimate. 2. _____ neglects members. [R] 3. When dealing with members, _____ has a tendency to throw its weight around. [R]

PAGE 162

144 Table 5-2. Continued Control Mutuality 4. _____ really listens to what members have to say. 5. _____ seems to ignore members’ opinions in the decisions that affect members. [R] 6. When members interact with _____, members feel that they have some sense of control. 7. _____ cooperates with members. 8. Members have influence with the decision makers at _____. Satisfaction 1. Both _____ and members benefit from their relationship. 2. Members are dissatisfied with their interaction with _____. [R] 3. Members are happy with _____. 4. Generally speaking, members are unhappy with the relationship _____ has established with them. [R] 5. Members enjoy dealing with _____. 6. _____ fails to satisfy members’ needs. [R] 7. Members feel they are important to _____. 8. In general, nothing of value has been accomplished by _____ for members. [R] Trust 1. _____ treats members fairly and justly. 2. Whenever _____ makes an important decisi on, members know _____ will consider the decision’s impact on members. 3. _____ can be relied on to keep its promises to members. 4. _____ takes the opinions of members into account when making decisions. 5. Members feel very conf ident about _____ abilities. 6. Sound principles guide _____’s behavior. 7. _____ misleads members. [R] Commitment 1. _____ is trying to maintain a long-term commitment to members. 2. _____ wants to maintain a positive relationship with members. 3. Compared to other farm organizations, me mbers value their relationship with _____ the most. 4. Members would rather work with _____ than without it. 5. Members feel a sense of loyalty to _____. Note: The blanks need to be replaced with the name of the organization. [R] indicates reverse-coding. Causal Links between Relationship Main tenance Strategies and Relationship Quality Outcomes Using stepwise regression analysis and path analysis (which produced similar results), this study examined how relations hip maintenance strategies used by an organization affected publics' perceptions of relationship quality outcomes. Links among the six relationship maintenance strategies and relationship qualit y outcomes provide new information concerning the function of main tenance effects. Overall, relationship

PAGE 163

145 maintenance strategies like access, positivity, sharing of tasks, and assurances represent the proactive approaches that organizations may employ to maintain or cultivate positive quality relationship with its target publics. Access Control Mutuality Assurances Sharing of Tasks Positivity Commitment Trust Satisfaction + + + + + + + + ++ M aintenance Strategies R elationship Quality Outcome Note: + (plus symbol) indicates positive impact. Figure 5-1. Relationships among relationship maintenance strategies and relationship quality outcomes First, a public’s perception of an organiza tion’s efforts to maintain a relationship through assurances was the primary predic tor of all relationship quality outcome indicators (i.e., control mutu ality, satisfaction, trust, and commitment). Namely, the use

PAGE 164

146 of assurances as a relationship maintenan ce strategy increased a public’s perceptions about control mutuality, satisfaction, trust, a nd commitment in the relationship with an organization and its public. Second, an organization’s effort to make an organization-public relationship more enjoyable via courteous communication and interaction with the public positively affected power balance (contro l mutuality), satisfaction, and trust in both analyses. Third, both analyses found that an or ganization that provides communication channels to facilitate public feedback and interaction signifi cantly influences a public’s evaluation of power balance. Although the re gression analysis suggested that access had a substantial impact on satisfaction and trust, the path analysis did not reveal the same results. These differing results may be at tributed to the high correlations among the relationship quality outcome va riables. Fourth, only the pa th analysis indicated that sharing of tasks, or an organization’s effo rt to share projects a nd problems of mutual interest to both parties, impacted control mu tuality and satisfaction. Last, both analyses revealed that the relationship maintenance strategies openness and networking were not significant predictors of any relationship quality outcome indicators. Figure 5-1 illustrated the overall results. Best Indicator of Relationship Quality Ou tcomes of Overall Relationship Quality This study asked which indicator of relati onship quality outcomes best assesses the overall quality of a relationship. To answer this question, regressi on and path analyses were conducted. Although the magnitude of impact of each relationship quality outcomes on overall relationship quality is diffe rent between the regr ession analysis and path analysis, both analyses produced the same results in terms of the order of significance. The analyses showed that comm itment is the strongest predictor of overall

PAGE 165

147 relationship quality, followed by satisfacti on and trust. Control mutuality did not significantly impact overall relati onship quality in both analyses. Influential Order among Relationship Indicators This study asked how relationship quality out come indicators affect each other and draws two specific hypotheses based on the pr evious literature. The first hypothesis posits that degree of satisfaction will positivel y affect degree of trust, and the second asserts that the degree of trust will positively influence the degree of commitment. Both hypotheses were strongly supported in this study, as satisfaction proved a strong and positive predictor of trust and trust preceded commitment. Therefore, the influential order of relationship indi cators is as follows: Satisfaction Commitment Trust ++ Figure 5-2. The influential sequence among re lationship quality out come indicators Linkages among Relationship Quality Perception, Attitude, and Behavioral Intentions This study was designed to empirically te st the linkages betw een perceptions of relationship quality, attitude, and behavioral intentions w ith members of a key public based on a standard hierarchy of effects model using structural equation modeling. Although the originally tested model met the criteria, the model was revised based on the modification index and theoretical justifica tion. A new path from commitment of the relationship quality outcomes to behavioral intentions was added. The revised model significantly improved the model fit. The fi nal revised model, which links relationship perceptions, attitude, and behavioral in tentions, is illustrated as follows.

PAGE 166

148 As shown in the revised final model, per ceptions of trust and commitment in the relationship are significant predictors of a ttitude, and attitude is a strong predictor of behavioral intentions. Among the four rela tionship indicators, a public’s perception of commitment in the relationship strongly impacted behavioral intentions. Control Mutuality Attitude Commitment Trust Satisfaction Behavioral Intention + + + + Figure 5-3. Final model linking relationship quality perceptio n, attitude, and behavioral intentions The Comprehensive Model This study was designed to empirically te st a comprehensive model that links relationship maintenance strategies, rela tionship quality outcomes, attitude, and behavioral intentions using structural equa tion modeling. The results illustrated that all effects of the paths linking relationship maintenance stra tegies, relationship quality outcomes, attitude, and behavioral intentions in the tested comprehensive model were significantly strong.

PAGE 167

149Figure 5-4. Comprehensive model linking rela tionship maintenance strategies, relations hip quality outcomes, attitude, and beha vioral intentions Relationship Maintenance Strate g ies Relationship Quality Outcome Attitude Behavioral Intentions++ +

PAGE 168

150 Post Hoc Analysis This study performed the following additiona l analyses based on the length of the relationship members of the public had with the organization: 1) the model which links perceptions of organization-publ ic relationships, attitude, an d behavioral intentions; 2) the comprehensive model which links relations hip maintenance strate gies, perceptions of relationship quality, attitude, a nd behavioral intentions. Linkage among relationship quality perceptions, attitude, and behavioral intentions on relationship length The first additional analysis showed that satisfaction was the best predictor of public attitude toward the organization wh en the organization-public relationship has lasted for a brief period of time. The remain ing two indicators of the relationship, control mutuality and trust, did not affect attit ude toward the organization. In this group, perceptions of relationship commitment are also a good predictor of attitude toward the organization. Furthermore, attitude and perceptions of commitment strongly and significantly affected beha vioral intentions. In the second model, which focused on l onger relationships, perceptions of trust most significantly affected at titude toward the organization, followed by commitment. The remaining two indicators of relationshi p perception were found to be insignificant predictors of attitude. Like the other model, commitment and attitude proved good indicators of a public’s suppor tive behavioral intentions. The comprehensive model on length of relationship The comprehensive model links relationship maintenance strategies, perceptions of organization-public relationshi ps, attitude, and behavioral intentions based on two groups—those who had comparatively short rela tionships with an organization and those

PAGE 169

151 with longer organization-public re lationships. The analysis revealed that the two models indicated no differences for either group. Theoretical and Managerial Implications This research contributes to several substantial topics in public relations scholarship by revealing 1) a new measure of relationship maintena nce strategies; 2) a more refined measure of relationship quality outcomes; 3) the sequential order of the relationship indicators; 4) the be st indicator of overall relation ship quality; 5) the ways in which communication activities as relationshi p maintenance strategi es contribute to relationship quality; 6) the ways in which relationship quality outco me contributes to attitude toward an organizati on and their supportive behavior toward an organization; and 7) a comprehensive model linking relations hip maintenance strategies, relationship quality outcomes, attitude, and behaviors. This section concludes with implications for teaching. New Measure of Relationship Maintenance Strategies Relationship maintenance strategies have gained little attention due to existing scholarship’s emphasis on organization-public re lationships as a means of measuring the effectiveness of public relations. Thus, this is the first study to focus on and develop empirical measures for relationship maintena nce strategies, or the daily communication activities performed by organizations. As such, this study should identify methods of maintaining and improving organization-publ ic relationships. The measures of relationship maintenance strategies developed in this research were established as reliable and valid. These developed measures can help practitioners better understand how to maintain or cultivate relationships with their targ et publics.

PAGE 170

152 The instrument designed provides a basic skeleton for the evaluation format and statements for each of the six relationship main tenance strategy dimensions so that it can be applicable to other types of organizati on-public relationships. However, the skeleton can be adapted when necessary or supplemente d to fit the characteristics or needs of a particular organization. Tabl e 1 can guide relationship maintenance strategies measure applicable to an organization. This instrument would prove valuable when used periodically to track the quality of relationship maintenance strategies, and when used in conjunction with the refined measures of organization-public relationshi ps. Specifically, by administering both instruments of relationship maintenance stra tegies and relationship quality outcomes among a target public, an organization would learn a great deal a bout the strategies available to improve its public relations. In addition, the survey could provide quality information by allowing practitioners to soli cit and analyze the pub lics’ suggestions or complaints. Furthermore, the instrument could evaluate the quality of maintenance strategies that an organization uses along each of th e six strategy dimensions by averaging the different scores on items. It also can a ssess an overall measure of the maintenance strategies in the form of an average score across all six dimensions. However, respondents surveyed should have some relati onship or experience w ith the organization being researched, as meaningful responses to the perception statements can be gained from a public’s experiences. The relationship maintenance strategies instrument can be used for an organization’s multiple target publics to track the quality levels of each strategy used by

PAGE 171

153 the organization. This would allow organizations to identify which stra tegies need to be improved or revised based on the needs of specific publics. More Refined Measure of Relationship Quality Outcomes Like several previous studies, this st udy attempted to refine a measure of relationship quality outcomes (i.e., Hua ng, 2001b; Jo, 2003, 2006; Kim, 2001). This measure achieved desirable levels of reliability and validity and can be used to better understand a public’s perceptions of its rela tionship with organizations, thus helping organizations assess how to improve these relationships. The instrument that measures relationship quality outcomes can be used to improve program management in public relations. For example, if control mutuality was found to be an essential relationship dimension but sc ores for the items are low, an organization would know to consider ways of improving its publics' involvement in the organizational decision-making process. As discussed previously, the refined m easure of relationship quality outcomes would prove useful when used in conjunc tion with the instrument for relationship maintenance strategies. An or ganization that uses the refined measure to identify the most salient relationship quality dimensions for its strategic publics and to compare its strengths and weaknesses in thes e dimensions will get a sense of what its public relations program's priorities should be with regard to relationship maintenance. Sequential Order of the Relationsh ip Quality Outcome Indicators The current study also presents an opport unity to gain a be tter understanding of, and better insights into, the seque ntial order of the th ree indicators (satis faction, trust, and commitment) of relationship quality outcom es through information obtained by path

PAGE 172

154 analysis. The results show that a public ’s perceptions of satisfaction with the organization-public relationship pl ay a significant role in gene rating perceptions of trust. This finding is consistent with studies in relational marketing which support the positive effect of satisfaction on trust (Anderson & Narus, 1990; Ganesan, 1994; Garbarino & Johnson, 1999; Selnes, 1998) and previous s uggestions made by scholars in public relations (Jo, 2003, 2006; Ki & Hon, 2007). In accordance with the ex isting literature in relational marketing (Anderson & Narus, 1990; Ganesan, 1994; Garbarino & Johnson, 1999; Miettil & Mller, 1990; Kwon & Suh, 2004; Morgan & Hunt, 1994), this study also documented that trust is a strong pred ictor of commitment in relationship quality outcomes. The empirically tested sequential order of satisfaction, trust, and commitment will aid organizations seeking to improve relations hips with their target publics. The most effective way to ensure publics of an organization's honesty, competence, and benevolence during the initial stages of the rela tionship is to provide publics with positive experiences and a sense of satisfaction. If the members of a public know that an organization is able and willing to satisfy th eir needs and provide reliable and predictable service, publics are more likely to establish trust in organization-public relationships. Recognizing the importance of satisfaction fo r establishing a relati onship in the initial stages would allow public relations practit ioners to work on developing satisfied relationships with their strategic publics. As Hrebiniak (1974) claimed, trust is an essential factor in every relationship because the parties involved are more likely to commit themselves to such relationships. Once trust is developed in the final stage of relationship building, organizations must

PAGE 173

155 devote their public relations programs to establishing public commitment in the relationship. Public relations programs th at seek to establish commitment should encourage long-term relationships once public perceptions of trust are developed (Dwyer, Schurr, & Oh, 1987; Morgan & Hunt, 1994). Best Indicator of Overa ll Relationship Quality This study determined the relative impor tance of the four dimensions of relationship quality outcomes in affecting pub lics' overall relationship perceptions. As mentioned previously, three relationshi p indicators—satisf action, trust, and commitment—are the significant predictors of overall relationship quality. Consistent with Hon and Brunner's (2002) study, cont rol mutuality was the only insignificant variable. The findings of this study reveal that these three indicators are key to understanding relationship qualit y. Furthermore, this finding is consistent with and confirm Jo’s (2006) study, which suggested th at satisfaction, trust, and commitment are the global measures of relationship quality. Among the three significant indicators, th is study found that commitment is the most critical factor in predicting publics' perceptions of overall organization-public relationships, followed by satisfa ction and trust. The remain ing factor, control mutuality, was the only insignificant predictor of relati onship quality perceptions. The findings of this study are inconsistent with Huang’s ( 2001b) study which identified control mutuality as the best predictor of relationship per ceptions and commitment as the second best predictor of relationship per ceptions. This inconsistency might be attributed to the differences in relationship characteristics between this study and Huang's. Public members who participated in this study ha d relatively long relationships with the

PAGE 174

156 organization, as illustrated by an average relationship length of 25 years. As the sequential order of relationship indicators prove d, commitment forms in the last stage of the relationship. Although control mutuality was the most substan tial factor in Huang’s study, it was the only insignificant factor in predicting ove rall relationship quali ty. The relationship Huang measured was the executive-legislative relationship, which can be characterized as one with high political involvement and f ace-to-face interaction. Unlike the relationship in Huang's study, members of the Florida Farm Bureau are not directly involved in the organization's decision-making process because individual members belong to the county organization rather than the state organization, rendering power balance difficult to evaluate in this context. Additionally, members of the public may have few opportunities to interact with organization managers. Ho wever, as Ki and Hon (2007) noted, control mutuality might be an important factor in predicting overall employee-, stockholder-, or activist-organization re lationships where power balance in the decision making process is considered important. Comparing the results of Huang (2001b), Jo (2006), Ki and Hon (2007), and the current study suggests that sa tisfaction, trust, and commitme nt are essential dimensions for relationships that offer little or no di rect contact or face-t o-face communication with management, such as relationships established with large corporations. The Ways in Which Communication Activi ties as Relationship Maintenance Strategies Contribute to Rela tionship Quality Outcomes This study examined how publics’ percepti ons of an organization’s relationship maintenance strategies (access, positivity, openness, sharing of tasks, networking, and access) affect evaluations of the relati onship quality outcomes (control mutuality,

PAGE 175

157 satisfaction, trust, and commitment). Based on suggestions in the scholarship (i.e., Hon & J. E. Grunig, 1999; J. E. Grunig & Hu ang, 2000), it is assumed that a public’s assessment of relationship maintenance efforts positively affects the relationship quality. The associations between the proposed rela tionship maintenance strategies and the relationship quality outcomes provide new information about the function of the organization’s maintenance efforts. This study found that the maintenance strategies affected different relationship quality outcome s in a variety of ways. The maintenance strategies, especially the four strategies of access, positivity, sharing of tasks, and assurances, represent proactive approaches th at organizations can use to maintain or cultivate relationship with their strategic publics. The strategy of access was found to have a positive impact on control mutuality. Given the characteristics of control mutual ity, which are related to the decision-making process and the extent to which a party’s opi nion is reflected in th e final decision, having accessibility to express one’s opinion is critical . The access factor includes items that are clear ways to get members of publics to jo in in the decision-making process of the organization, such as referring contact inform ation, meeting with the organization’s staff, and being willing to an swer public inquiries. An organization’s use of positivity as a maintenance strategy was found to be a significant factor predicting control mutuality, satisfaction, and trust. That positivity was significant to satisfaction is consistent with studies on interpersonal communication. For instance, Dindia (1989) found that the percep tion of one’s positivity maintenance efforts positively affected the partner’s relationa l satisfaction. Courteous and enjoyable communication with members of publics might encourage publics’ cooperation in the

PAGE 176

158 organization and help the relationship preserve interdependence. The positivity items clearly show the rewards or benefits the orga nization provides to public s. From a social exchange perspective, a clear di splay of rewards and benefits to publics instills them with a positive perception that increases their sati sfaction because as the rewards and benefits increase, publics feel that the costs incurred by the relationship decrease (Kelly & Thibaut, 1978; Jo et al., 2004). These econom ic benefits, which can be great at times, can give publics strong incentive to remain in a relationship with the organization. Sharing of tasks was an important f actor predicting cont rol mutuality and satisfaction. In this study, the concept of ‘sharing of tasks’ is similar to the concept of corporate social responsibilit y. Performing corporate social responsibility activities has frequently appeared as an essential issue in public relations. Such actions as learning what concerns a public and participating in activities such as disaster relief, environmental protection in the community, a nd resolving regulatory issues are necessary to accomplishing the interdependent goals and objectives an organization and its publics have. As Lez-Herrero and Pr att (1996) argued, the sharing of tasks (corporate social responsibility) can be a proactive strategy to respond to the media’s negative coverage during a crisis. The causal relationship whic h the sharing of tasks promotes between control mutuality and satisfacti on confirms the adage that actions speak much louder than words. Assurances were the key strategy for pr oducing all of the relationship quality outcomes. This finding is consistent with studies on interpersonal relationship which have discovered and confirmed that assura nces is the most effective strategy for maintaining relationship commitment between two individuals (Canary & Stafford, 1992,

PAGE 177

159 1993; Stafford & Canary, 1991). Assurances al so have been a signi ficant predictor of trust in several studies on interpersonal re lationships (Canary & Stafford, 1993; Stafford & Canary, 1991). In accordance with a study by Canary and Stafford (1992), the current study revealed that an organi zation’s use of assurances as a maintenance strategy can engender commitment in organization-public rela tionships. Assurances items include the organization’s efforts to provide personal responses to the pub lic’s concerns, to communicate to the members of a public how im portant they consider the public to be, and to allow the members opportunities to ra ise issues and propose solutions during the policy development process. The organization’ s demonstration of a desire to assure the publics that they and their concerns ar e attended to implies the organization’s commitment to a long-term and stable re lationship with the public. Moreover, perceptions of assurances probably lead the public to believe that the organization is willing to invest organizational resources in the relationship to ensure its success. This study’s finding about openness is consiste nt with the interpersonal relationship literature, but not consistent with the public relations literat ure. For example, Stafford and Canary (1991) found that openness was the leas t predictive of relati onal features in a couple’s relationship. However, scholars in public relations have consistently found and suggested that openness is an important pred ictor of relationship quality outcomes. Specifically, Ledingham and Br uning (1998) discovered that openness was a significant predictor of relationa l satisfaction. Also, it has been proposed that openness is a fundamental indicator to evalua te relationship quality with an organization’s target public (L. A. Grunig et al., 1992). Successful relati onships with target publics can indeed be maintained through open communication, but the insignificant finding related to

PAGE 178

160 openness might be due to the fact that openne ss is not mutually exclusive but rather a dimension of all of the relati onship maintenance strategies. This study revealed that an organization’ s use of networking doe s not generate any relationship quality outcomes. Although pers onal networking opport unities have been found to be the most important means of attr acting and retaining members in farm bureau organizations (King & Walker, 1992), the impo rtance of public networks was not shown in this study. This study’s finding about ne tworking also might be due to insufficient number of measurement items (i.e., only 3 items) included for tapping the networking strategy the organization used. The findings indicate that maintenance st rategies vary in their relevance to relational outcomes; therefore, causal linka ges between maintena nce strategies and relationship quality outcomes could provide guidelines fo r how an organization should use each strategy to affect specific rela tionship quality outcomes. Accordingly, organizations may select one maintenance strategy approach over another depending on the relationship quality outcome of concern. For example, an organization can use the assurance strategy to maintain or promote commitment amon g a strategic public. Also, an organization can use access to ensure control mutuality. For more than a decade, the effective management of relationships between an organization and its target publics has become an increasingly important topic for both public relations scholars and practitioners . Quality relationships improve an organization’s effectiveness (Dozier et al., 1995; Hon, 1997; Huang, 1999; L. A. Grunig et al., 2002), help resolve conflicts between the organi zation and its publics (Huang,

PAGE 179

161 1997), and affect publics’ suppor tive attitude and behavior s toward the organization (Bruning, 2002; Ki & Hon, 2007). The results of this study provide some strate gic implications for institutions that are seeking to build their organiza tion-public relationships. In essence, this study found that the ways in which publics perceived an organi zation’s relationship maintenance strategies directly affected their ratings of the relationship quality outcomes of control mutuality, satisfaction, trust, and commitment. Although there are many ways for organiza tions to cultivate or maintain better relationships with their strategic publics, the assurance strategy was the most successful across all relationship quality outcomes. In most conditions, providing publics with benefits and participating in enjoyabl e and courteous communication with them effectively engenders control mutuality, satisfaction, and trus t in the organization-public relationship. For public relati ons professionals or managers of organizations, providing assurances and looking for ways to dem onstrate positivity are crucial for improving relationships with publics. The Ways in Which Relationship Quality Outcomes Contribute to Attitude and Behavior This study was designed to empirically te st a model that posited the linkages between perceptions of the organization-publ ic relationship, attit ude, and behavioral intentions toward an orga nization among members of a ke y public. The test produced results with strong support of some theoreti cal assumptions that unde rlie the relationship management perspective for public relations theo ry and its practice. It has been assumed that positive, long-term relationships with strategic publics can be valuable because quality relationships are more likely to dr ive supportive behaviors such as sales,

PAGE 180

162 donations, favorable legislation, and high empl oyee performance (L. A. Grunig et al., 2002). The linkages tested here substant iated assumptions about the effects of organization-public relationships on publics’ supportive attitude and behaviors toward an organization. These linkages can be used as a toolbox to measure public relations effectiveness. For the four relationship dimensions used in this study—control mutuality, satisfaction, trust, and commitment—part icipants’ perceptions of relationship commitment were key to their positive at titude toward the organization. More importantly, the final model shows that per ceptions of commitment also can engender supportive behavior toward the organizati on among members of a key public. Members of the public studied here with strong perceptions of relati onship commitment were more likely to recommend membership to others and retain their membership even if membership in a comparable association was available. This finding is consistent with those from a study in relational marketing which found that customers’ relationship commitment directly affected their behavi oral intentions (Gar barino & Johnson, 1999). This finding confirms the notion that commitme nt is most related to behavior (Becker et al., 1995). Perceptions of trust also strongly affected member s’ attitude toward the organization. A positive attitude and supportiv e behavior toward the organization were affected by members’ perceptions of whether they felt some degree of confidence in and willingness to rely on the organization. Thes e results suggest that using positivity and assurances, which can engender trust, are key relationship maintenance strategies.

PAGE 181

163 As found in the study by Garbarino a nd Johnson (1999), among the relational dimensions, trust and commitment are key predicto rs of attitude and behavioral intentions for members of a public who have strong relatio nships. In the current study, the average relationship length was approximately 25 y ears, which could be considered a long relationship. For members of a public who ha ve long-term relationships, perceptions of trust and commitment drive supportive attitude s and behaviors toward the organization. Even when the relationship is well-establishe d, an organization might need to continue to use strategies to build trust and commitmen t among members of a public. Therefore, public relations programs focused on managing satisfaction will be more effective for members of a public within a short period tim e. Public relations programs directed toward long-term members of a public shoul d focus on maintaining and building trust and commitment rather than satisfaction. Although the sequential model linking perceptions of organization-public relationships, attitude, and beha vioral intentions shows a good fit with the data, some of the proposed theoretical paths among the individual dimensions were not found in the empirically collected data set. For exam ple, perceptions of control mutuality and satisfaction had no significant eff ects on attitude and behavioral intentions. These results are inconsistent with Ki a nd Hon’s (2007) study which test ed the sequential order of perceptions of organization-public relationships, attitude, and behavioral intentions based on the university-student relationship. In their study, using attitude as a mediating variable, they found that satisfaction and contro l mutuality were the two key predictors of supportive behaviors.

PAGE 182

164 The insignificant findings related to c ontrol mutuality’s effect on attitude and behavioral intentions might be explained by th e characteristics of th e organization-public relationship used in this study. As mentione d previously, since ther e are relatively few opportunities for direct contact between member s of this public and the decision-makers of the organization, control mutuality might not be reflective of members’ attitude and behavioral intentions. This may be especia lly true for this study which focuses on the relationship between state farm bureaus (the organization) and memb ers of a public who belong to a county-level organization. The insignificant findings about satisfaction’s effect on attitude and behavioral intentions might come from the sequential or der of satisfaction, trus t, and commitment. As found in this study, satisfaction is a predicto r of trust and trust generates commitment. Since most of the members already formed perceptions of trust and commitment, satisfaction might be the foundation of the relationship because members of the publics have repeat satisfied experien ces over several years that lead to trust and commitment. A Comprehensive Model Linking Relationship Maintenance Strategies, Relationship Quality Outcom es, Attitude, and Behavior The current research also was designed to empirically test a model that posited the linkages among relationship maintenance stra tegies, relationship quality outcomes, supportive attitude, and beha viors toward an organiza tion among members of a key public. One of the most importa nt contributions this study ma kes to the public relations arena is an explanation of the effects of relationship maintenance strategies on relationship quality outcomes a nd connect those to the public ’s supportive attitude and behavior toward an organiza tion in one comprehensive model (Figure 5-4; the detailed causal linkages are explained th roughout this chapter). Even post hoc analysis, which

PAGE 183

165 analyzes the model across different gr oups, supported the validation of the comprehensive model. These findings indicate that an organization’s effort to build and maintain quality organizationpublic relationships should be the focus of public relations programs that endeavor to eventually achieve the objectives and goals of the organization. As the scholarship shows, an organization’ s efforts and its strategies to maintain and cultivate quality relationships are essent ial for producing quality relationships with members of a key public (J. E. Grunig & Huang, 2000; Hon & J. E. Grunig, 1999). The current study documents that relationship main tenance strategies can effectively generate quality relationships between an organizati on and its publics. Furthermore, a public’s supportive attitude and behavi ors toward the organizati on can occur through positive perceptions of organizati on-public relationships. Since a quality, long-term relationship with an organization’s st rategic publics is not established by a one-shot public relations strategy, an organization should conduct research to obtain information about key publics’ perceptions of the quality of the organization-public relationship. Based on the information, relationship maintenance strategies should be planned and executed with the goal of improving particular perceptions of the rela tionship. For example, if an or ganization wants to increase the public’s perceptions of satisfac tion, it needs to implement positivity since positivity is the most effective strategy for engendering satisfa ction. By doing this, public is more likely then to display supportive att itude and behavioral intentions toward the organization. As shown in Figure 5-5, an organi zation should continue to evaluate each stage and make a continual effort to improve.

PAGE 184

166 The models tested throughout this study are suggested m odels but they are not the only possible models. It is always arguable th at any model can be mi s-specified to some degree (Whang & Hancock, 1997). In addition, th e models in this study are not complete inasmuch as “it is impossible to account for all potential causal elements in a system” (Whang & Hancock, 1997). However, in terms of model respecification, future research should rely on a scholarly, theoretical basis for finding better models for measuring and understanding public relations effectiveness. Relationship Quality Outcome Public’s Supportive Attitude Public’s Supportive Behavior Relationship Maintenance Strategies Figure 5-5. A suggested comprehensive model The findings of this study also suggested so me implications for teaching. First of all, the comprehensive model should be intr oduced into the classroom as a measure of

PAGE 185

167 public relations effectiveness. The model is especially relevant to addressing how organizations’ communication activities affect th e quality of their relationships with their strategic publics. These organization-publ ic relationships are crucial in helping organizations to achieve their goals and object ives, which are associated with the public’s supportive attitudes and behavi ors toward the organizations . Secondly, research method courses need to include teaching of SEM, wh ich would be beneficial in public relations research. The use of SEM, including CFA, might be helpful in finding better causal models to explain the impact of public relations programs on organizations’ planned outcomes. Limitations and Future Research Although this study is original and compelli ng in several ways, it has its limitations which can nonetheless help guide future re search endeavors. To begin with, points related to newly developed measures of ma intenance strategies are worth mentioning. The definitions and measurement tools used in this study were developed based on the literature. However, there is no guarantee th at this newly developed measurement system proves scientifically or practic ally useful, or that it will measure what it intends to measure. Also, there are some shortcomings to the measurements of the relationship maintenance strategies. For example, although the reliabilities of a ll the indicators met their acceptable level, they were somewhat w eak at around .80 compared to the ones used to measure the organization-public relati onship (those exceeded .90). These low reliabilities are probably due to the insuffi cient number of items used for each scale, especially for access, networking, sharing of ta sks, and assurances, all of which have only three or four items. Therefore, future re search should increase the number of items for these scales to better tap each dimension. Mo reover, because of the characteristics of the

PAGE 186

168 public targeted in this study, questions re lated to new technologi es for communication were dropped. However, corporate Web s ites are a commonly used communication channel, suggesting that measurement item s for relationship maintenance strategies through Web sites should be considered in future research. Since the six dimensions of maintenance strategies originated in the field of interpersonal relationship st udies, there might be some gaps between relationship maintenance strategies used in interpers onal relationships and those in organizationpublic relationships since the nature of the two relationshi ps is different even though theories of interpersonal relationships ar e well applicable to organization-public relationships. Future research should consider a qualitative approach to find original measurement items for relationship mainte nance strategies a pplicable to public relationships. Secondly, this study collected data fro m a single organization. Although the random sampling method was used to select the sample, the findings in this study must be cautiously applied to other types of orga nization-public relationships because each organization faces different situations. In orde r to improve external validity of the newly developed measures and proposed models, as well as develop more refined measures of organization-public relationships , independent studies must apply the two measures and the models to diverse types of organizat ions such as for-profit, non-profit, and multinational companies until they achieve consistent findings. Third, the current research treated me mbers of a key public as a homogeneous group. Although it is obvious that all of the respondents in th is study have a relationship with the organization, the leve ls of involvement in the re lationship might vary and the

PAGE 187

169 sequential order of relationship perception, atti tude, and behavior that each person in the group experiences may differ. Therefore, future research should segment members of key publics based on their involvement and othe r variables that might impact relationship perceptions. Fourth, the relationship perspective assu mes a two-way model of public relations so that evaluating both sides’ perceptions of the relationship and determining the effects of the relationship on both part ies is essential (Ki & Hon, 2007). This study evaluated only one side of the relationship, the public side . Therefore, future research might assess the organization’s perception of its relationship with public s and evaluate whether and how perceptions among organizational memb ers are different from the publics’ perceptions. By doing this, an organization can evaluate gaps in the way management and publics perceive the relati onship. As Lindenmann (1999) indicated, this kind of gap analysis can provide strategies for relationship maintenance. Fifth, involvement, which was measured in terms of the length of membership with the organization, might not assess the strength of the relationship the members of a public might have with the organization. Although some members of a public have long-term membership in an organization, they might not be engaged in the relationship. In other words, they just remain as members but refr ain from active participation. On the other hand, although some members of a public have had relationships with the organization for only a short-period of time, they could be actively involved in the relationship with the organization. Therefore, future research might need to differentiate members of a public based on their level of involvement. J. E. Grunig and Hunt suggested three

PAGE 188

170 categories of publics based on situational theory —latent1, aware2, and active3 public— that could be applied in future studies. Using these categories, latent public can be classified as the low-involved public and active public as the high-involved public. Based on the level of public’s involvement in the relationship, future research should explore how these different groups of publics diff er in terms of the perceptions they have of their relationship with the organizati on and how differently or similarly these perceptions drive their attitude and be havioral intentions or behavior. Although the links and eff ects between relationship maintenance strategies, perceptions of relationship quality outcomes, and public ’s supportive attitude and behaviors toward an organiza tion are neither straightforwar d nor simple, more attention should be paid to how the effects of relati onship maintenance strategies are translated into public’s perceptions of relationship qua lity and how the public ’s perceptions of relationship quality are translated into attitude s and behaviors. Theref ore, future research should investigate conditions such as time la g on the effects of relationship maintenance strategies on the public’s pe rceptions of relationship qual ity as well as the publics’ attitudes and behavioral inte ntions that may influence th e relative superiority of the model. 1 Latent public is a group that does not recognize a situation as problematic. This group might have the lowest involvement in the relationship with an organization. 2 Aware public is a group that has moved from a latent stage and recognizes a problem. This group might have higher involvement in the re lationship with an organization but lower involvement than the active public. 3 Active public is a group that is active and organizes to discuss and do something about the situation. This group might have the highest involvement in the relationship among the three groups.

PAGE 189

171 Conclusion Numerous scholarly works have emphas ized the importance of relationship measurement in public relations effectivene ss (e.g., Dozier et al ., 1995; J. E. Grunig & Huang, 2000; L. A. Grunig et al., 2002; Hon, 1997; Hon & J. E. Grunig, 1999; Huang, 1997, 1999, 2001a, 2004; Ki & Hon, 2007; Jo, 2003, 2006; Kim, 2001). However, none of the studies comprehensively documented how organizations’ relationship maintenance efforts produce quality relationships with their strategic publics and how quality relationships between an organization a nd its public drive th e public’s supportive attitudes and behavioral intent ions or behaviors toward an organization. The empirical data gathered in this study through administ ration of a mail survey to members of an organization’s essential public supported the linkages and causal effects among relationship maintenance strategies, rela tionship quality outcomes, the public’s supportive attitude, and behavioral intentions and/or behaviors. In other words, this study documented evidence that organizations’ commun ication activities result in relationship quality outcomes that drive public’s suppor tive attitudes and be haviors toward an organization. Therefore, these findings ar e essential in understa nding and assessing the effectiveness and value of pub lic relations not only for organizations but also for members of their publics as well as society. The researcher belie ves the current study provides significant insight into understandi ng the potential of re lationship management for demonstrating the effectiven ess of public relations and it s contribution to the field.

PAGE 190

172 APPENDIX A SURVEY INVITATION Dear Florida Farm Bureau member: In a few days you will receive a packet from Florida Farm Bureau. In this packet, you will find a survey about the relationships be tween the organization and its members. Having you rate the job it is doing will help Florida Farm Bureau improve its level of service to you in the future. Please look for the questionnaire to arrive in th e next few days. It is extremely important that you complete and return it promptly. If you have any questions, please feel fr ee to contact me at (352) 846-1048 or ejki@jou.ufl.edu Sincerely Eyun-Jung Ki, Ph.D. Candidate College of Journalism and Communications University of Florida

PAGE 191

173 APPENDIX B COVER LETTER Date: Member’s Address: Dear Name of the member: On behalf of Florida Farm Bureau, I am currently conducting a study to examine your perceptions of Florida Farm Bureau and FloridAgriculture magazine. The study will concentrate on what Florida Farm Bureau does to maintain or impr ove its relationships with members and what are the results of those efforts. Also, this research will examine how these efforts affect members’ attitudes and future behavior toward Florida Farm Bureau. You were randomly selected from the current membership directory of Florida Farm Bureau. This survey will take about 15 minutes. Since only a limited number of questionnaires have been distributed, your response is extremely important and valuable to this research. Your answers will be used for statistical purposes only and will remain strictly confidential to the ex tent provided by law. All re sponses are confidential and no individual data will be repor ted. You may return the complete d questionnaire in the selfaddressed stamped envelope. If you have any questions about the proj ect, please feel free to call me at (352) 846-1048 or email me at ejki@jou.ufl.edu . Also, you may keep the attached informed consent form. Thank you for your help in this important endeavor. Sincerely yours, Eyun-Jung Ki Ph.D. candidate College of Journalism and Communications University of Florida, Gainesville

PAGE 192

174 APPENDIX C SURVEY QUESTIONNAIRE SURVEY QUESTIONNAIRE Thank you for taking time to answer the questions in this survey. This research is abou t members’ perceptions of, as well as members’ attitudinal and behavioral intentions to ward Florida Farm Bureau. Please write the number that best describes what you belie ve members in general think and feel. Your answers will be used only for statistical pur poses and will remain strictly confidential to the extent provided by law. Please read th e instructions and questions carefully. Section I Please write the number in the blank that y ou believe best indicates members’ agreement with each item and what Florida Farm Bureau has done. Responses range from strongly disagree to strongly agree. For example, if you strongly disagree with the provided statement, write 1 in the blank. If you strongl y agree with the provided statement, write ” in the blank. If you feel neutral about a statement, write ” in the blank. Strongly Disagree Neutral Strongly Agr ee ______ 1. Attending FFB’s annual meeting is helpful to members. ______ 2. FFB provides members with adequate contact information. ______ 3. FFB’s Annual Report is a valuable source of information for members about what FFB has done. ______ 4. FFB makes a genuine effort to provide personal responses to members’ concerns. ______ 5. FFB works with members to develop so lutions to problems that benefit members. ______ 6. FFB is involved in managing community issues that members care about (e.g., disaster relief, environment protection). ______ 7. The member benefits (i.e. insurance se rvices, bank services, etc.) FFB provides are important to members. ______ 8. FFB provides members op portunities to meet its staff. ______ 9. FFB works effectively to resolve regul atory issues its members are facing such as pesticide or food safety issues. ______10. Members do not believe that FFB really cares about their concerns. ______11. FFB effectively builds coalitions with groups (i.e. Suwannee Partnership) that impact members. ______12. FFB shares enough information with members about the organization’s governance. ______13. Receiving regular communications (e.g., FloridaAgriculture ) from FFB is beneficial to members. ______14. FFB communicates the importance of members.

PAGE 193

175 ______15. FFB’s communication with members is courteous. ______16. FFB and members do not work well together at solving joint problems. ______17. The coalitions that FFB forms with othe r agricultural groups benefit FFB members. ______18. FFB attempts to make its interactions with members enjoyable. ______19. FFB’s policy development process allows members adequate opportunity to raise an issue and propose a solution. ______20. The Ag Coalition for legislative activities that FFB is involved in is helpful to FFB members. ______21. FFB’s alliances with other like-mi nded groups are useless to members. ______22. FFB’s member meetings are a valuable way for members to communicate their opinions to FFB. ______23. When members raise concerns, FFB takes these concerns seriously. ______24. The issue briefings FFB provides help members understand the issues. ______25. FFB does not provide members with enough information about what FFB does with members’ dues. ______26. When members have questions or concerns , FFB is willing to answer their inquiries. ______27. The information FFB provides with members is of little use to them. ______28. FFB is cooperative when handling disagreements with members. ______29. FFB provide members with adequate contact information for specific staff on specific issues. Section II Please evaluate quality of the relationship between members and Florida Farm Bureau. Please write the number that you believe best i ndicates members’ agreement with each item. If you strongly disagree with the provided statement, please write “ 1 ” in the blank. If you strongly agree with the provided statement, please write “ 9 ” in the blank. Strongly Disagree Neutral Strongly Agr ee ______ 1. FFB is trying to maintain a long-term commitment to members. ______ 2. FFB treats members fairly and justly. ______ 3. There is only a short-term bond between FFB and members. ______ 4. FFB believes the opinions of members are legitimate. ______ 5. Florida Farm Bureau (FFB) neglects members. ______ 6. Whenever FFB makes an important decisi on, members know FFB will consider the decision’s impact on members. ______ 7. Both FFB and members benefit from their relationship. ______ 8. FFB wants to maintain a positive relationship with members. ______ 9. When dealing with members, FFB has a tendency to throw its weight around. ______10. FFB can be relied on to keep its promises to members. ______11. Members are dissatisfied with their interaction with FFB. ______12. FFB really listens to what members have to say. ______13. Members are happy with FFB.

PAGE 194

176 ______14. FFB takes the opinions of member s into account when making decisions. ______15. Generally speaking, members are unhappy with the relationship FFB has established with them. ______16. FFB seems to ignore members’ opinions in the decisions that affect members. ______17. Members feel very c onfident about FFB abilities. ______18. Compared to other farm organizations, members value their relationship with Florida Farm Bureau the most. ______19. Members enjoy dealing with FFB. ______20. Members would rather work with FFB than without it. ______21. When members interact with FFB, member s feel that they have some sense of control. ______22. Members believe that FFB lacks the abilit y to accomplish what it says it will do. ______23. Members want to have a relationship with FFB. ______24. FFB cooperates with members. ______25. Sound principles guide FFB’s behavior. ______26. FFB fails to satisfy members’ needs. ______27. Members feel they are important to FFB. ______28. Members have influence with the decision makers at FFB. ______29. FFB misleads members. ______30. Members feel a sense of loyalty to FFB. ______31. In general, nothing of value has been accomplished by FFB for members. ______32. Please rate the overall quality of members' relationship with FFB based on 1 to 9 point scale. (1=very negative, 9= very positive) Section III Please write the number that you believe best indicates members’ agreement with each item and how it describes members’ attitude toward Florida Farm Bureau. If you strongly disagree with the provided statement, write ” in the bl ank. If you strongly agree with the provided statement, write ” in the blank. Strongly Disagree Neutral Strongly Agr ee _______ 1. Members’ impression of Florida Farm Bureau is favorable . _______ 2. Members’ impression of Florida Farm Bureau is negative . _______ 3. Florida Farm Bureau is useful to members. _______ 4. Members dislike Florida Farm Bureau. Section IV Based on your overall experience with Florida Farm Bureau, please indicate how likely or unlikely you believe members are to take the following actions. If you believe they are very unlikely to take the stated action, write “ 1” in the blank. If you believe they are very likely to take the stated action, write “ 9 ” in the blank. Very Unlikely Neutral Ver y Likely

PAGE 195

177 ______ 1. Members would recommend membership in Florida Farm Bureau to other farmers. ______ 2. Members would retain their membership in FFB even if membership in a comparable association were available. ______ 3. Members would like to retain their me mbership with FFB for at least another five years. Section V Demographics 1. Gender Male [ ] Female [ ] 2. What year were you born? [ ] 3. Total number of years as a Florida Farm Bureau member: ____________ years 4. What is your highest level of education? [ ] Some schooling [ ] Bachelor’s degree [ ] High school diploma [ ] Some graduate school [ ] Some college [ ] Graduate degree [ ] Doctoral degree 5. Which of the following best describe your racial or ethnic identification? [ ] White/Caucasian [ ] Latino/Hispanic [ ] African American [ ] Native American [ ] Asian American [ ] Anglo [ ] Other: ______________________ (please specify) 6. What area of the state are you from? [ ] Northwest [ ] West Central [ ] Northeast [ ] Southwest [ ] North Central [ ] Southeast [ ] East Central 7. Pick the best representation of your main commodity: [ ] Livestock [ ] Dairy [ ] Citrus [ ] Other fruits and vegetables [ ] Ornamental horticulture [ ] Field crops [ ] Forestry [ ] Other: ________(please specify) Thank you so much for your participation!

PAGE 196

178 APPENDIX D INFORMED CONSENT FORM STATEMENT OF INFORMED CONSENT Dear Florida Farm Bureau Member: On behalf of Florida Farm Bureau, I am currently conducting a study to examine your perceptions of Florida Farm Bureau and FloridAgriculture magazine. The study will concentrate on what Florida Farm Bureau does to maintain or impr ove its relationships with members and what are the results of those efforts. Also, this research will examine how these efforts affect members’ attitudes and future behavior toward Florida Farm Bureau. You were randomly selected from the current membership directory of Florida Farm Bureau. This survey will take about 15 minutes. Since only a limited number of questionnaires have been distributed, your response is extremely important and valuable to this research. Your answers will be used for statistical purposes only and will remain strictly confidential to the ex tent provided by law. All re sponses are confidential and no individual data will be repor ted. You may return the complete d questionnaire in the selfaddressed stamped envelope. If you have any questions about the proj ect, please feel free to call me at (352) 846-1048 or email me at ejki@jou.ufl.edu . If you have any questions about your rights as a participant, please ca ll the UF Institutional Review Board at (352) 392-0433. Thank you in advance for your participation. “I have read the procedure described above. I voluntarily agree to participate in the procedure and I have received a copy of this description.” Principal Investigator: __________________________ Date: ______________________

PAGE 197

179 APPENDIX E FOLLOW-UP POSTCARD CONTENT Dear Florida Farm Bureau Member: A few days ago you received a survey asking you to evaluate Florida Farm Bureau’s relationship with its members. Florida Farm Bureau wants to know how you think it is doing in order to do a better job of serving Farm Bureau members. If you have already completed the survey and mailed it back, thank you so much. If not, please take the time to complete the questi onnaire and return it in the self-addressed, stamped envelope in the packet. If you have any questions, please feel fr ee to contact me at (352) 846-1048 or ejki@jou.ufl.edu Eyun-Jung Ki, Ph.D. candidate College of Journalism and Communications University of Florida, Gainesville

PAGE 198

180 APPENDIX F COVER LETTER SAMPLE OF REPLACEMENT MAIL Date Address: Dear Name of the member: A short time ago I sent you a survey a bout the relationship between Florida Farm Bureau and its members on behalf of Florida Farm Bureau. Having you rate the job it is doing will help Florida Farm Bureau improve its level of service to you in the future. Your opinion of what Florida Farm Bu reau does and what it should be doing is important. If you have completed the survey questionn aire and returned it to me, thank you! If, however, you have misplaced it, or not qui te found the time to respond, a replacement questionnaire is enclosed in the packet. Pl ease take the 10-15 minutes and return it to me in the enclosed self-add ressed, pre-paid envelope. Thanks again for your help in this important endeavor. Sincerely yours, Eyun-Jung Ki, Ph.D. candidate College of Journalism and Communications University of Florida

PAGE 199

181 LIST OF REFERENCES 2002 Census of agriculture state profile: Florida . (2002). Washington, DC: United States Department of Agriculture. Achrol, R. (1991). Evolution of the ma rketing organization: New forms for turbulent environments. Journal of Marketing, 55 (4), 77. Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179. Ajzen, I., & Fishbein, M. (1980). Understanding attitudes an d predicting social behavior . Englewood Cliffs, NJ: Prentice-Hall. Aldrich, H. E. (1975). An organizationenvironment perspective on cooperation and conflict between organizations in the manpower training system. In A. R. Negandhi (Ed.), Interorganizational theory (pp. 47). Kent, OH: Kent State University Press. Aldrich, H. E. (1979). Organizations and environments . Englewood Cliffs, NJ: Prentice-Hall. Algina, J. (2005). Structural equation modeling [Course material for EDF 7412: Structural Equation Modeling]. Univer sity of Florida, Gainesville. Anderson, J. C., & Gerbing, D. W. (1 988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103 , 411. Anderson, J. C., & Narus, J. A. (1990) . A model of distributor firm and manufacturer firm working partnerships. Journal of Marketing, 54 , 42. Anderson, J. C., & Weitz, B. (1989). Determ inants of continuity in conventional industrial channel dyads. Marketing Science, 8 , 310. Aronson, E. (1969). The theory of cognitiv e dissonance. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 4). New York: Academic Press. Babbie, E. R. (2001). The practice of social research (9th ed.). Belmont, CA: Sage.

PAGE 200

182 Bandura, A. (1986). Social foundations of thought and action . Englewood Cliffs, NJ: Prentice Hall. Barney, J. B., & Hansen, M. H. (1994) . Trustworthiness as a source of competitive advantage. Strategic Management Journal, 15 , 175. Barry, T. E. (1987). The development of th e hierarchy of effects: An historical perspective. Current Issues and Research in Advertising, 10 , 251. Becker, H. S. (1960). Notes on the concept of commitment. American Journal of Sociology, 66 , 32. Becker, T. E., Randall, D. M., & Riegel, C. D. (1995). The multidimensional view of commitment and the theory of reasoned action: A comparative evaluation. Journal of Management, 21 , 617. Bem, D. J. (1972). Self percepti on theory. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 4). New York: Academic Press. Bennett, J. B. (2001). Teaching with hospitality. Teaching and Learning News, 10 (3), 88. Bentler, P. M. (1992). EQS: Structural equations program manual . Los Angeles: BMDP Statistical Software. Bentler, P. M., & Wu, E. J. C. (1993). EQS/Windows user's guide . Los Angeles: BMDP Statistical Software. Berelson, B. (1996). KAP studies on fertilit y. In B. Berelson, R. K. Anderson, O. Harkavy, J. Maier, W. P. Ma uldin & S. J. Segal (Eds.), Family planning and population programs: A review of world developments (pp. 665). Chicago: University of Chicago Press. Berko, R. M., Rosenfeld, L., & Samovar, L. A. (1997). Connecting (2nd Ed.). New York: Harcourt Brace College Publisher. Berry, L. L., & Parasuraman, A. (1991). Marketing services . New York: John Wiley & Sons, Inc. Blau, P. M. (1964). Exchange and power in social life . New York: John Wiley & Sons, Inc. Bok, S. (1989). Secrets-on the ethics of concealment and revelation . New York: Vintage Books. Bollen, K. A. (1989). Structural equations with latent variables . New York: Wiley.

PAGE 201

183 Bridges, J. A., & Nelson, R. A. ( 2000). Issues management: A relational approach. In J. A. Ledingham & S. D. Bruning (Eds.), Public Relations as Relationship Management: A Relational Approach to the Study and Practice of Public Relations (pp. 95). Mahwah, NJ: Lawrence Erlbaum Associates. Broom, G. M., Casey, S., & Ritchey, J. (1997). Toward a Concept and Theory of Organization-Public Relationships. Journal of Public Relations Research, 9 , 83. Broom, G. M., Casey, S., & Ritchey, J. (2000). Toward a con cept and theory of organization-public relationships: An update. In J. A. Ledingham & S. D. Bruning (Eds.), Public relations as relati onship management: A relational approach to public relations (pp. 3). Mahwah, NJ: Lawrence Erlbaum Associates. Broom, G. M., & Dozier, D. M. (1990). Using research in public relations: Applications to program management . Englewood Cliffs, NJ: PrenticeHall. Bruning, S. D. (2002). Relationship buildi ng as a retention strategy: linking relationship attitudes and satisfaction evaluations to behavioral outcomes. Public Relations Review, 28 , 39. Bruning, S. D., & Ledingham, J. A. (1999) . Relationships between organizations and publics: Development of a mu lti-dimensional organization-public relationship scale. Public Relations Review, 25 , 157. Bruning, S. D., & Ledingham, J. A. ( 2000). Perceptions of relationships and evaluations of satisfaction: An exploration of interaction. Public Relations Review, 26 , 85. Buchanan, B. (1974). Building organizati onal commitment: The socialization of managers in work organizations. Administrative Science Quarterly, 19 , 533. Burgoon, J. K., & Hale, J. (1984). Th e fundamental topoi of relational communication. Communication Monographs, 51 , 193. Burgoon, J. K., & Hale, J. L. (1987). Validation and measurement of the fundamental themes of relational communication. Communication Monographs, 54 , 19. Byrne, B. M. (1994). Structural equati on modeling with EQS and EQS/Windows: Basic concepts, applications, and programming . Thousand Oaks, CA: Sage.

PAGE 202

184 Byrne, B. M. (2001). Structural equation modeling with AMOS: Basic concepts, applications, and programming . Mahwah, NJ: Lawrence Erlbaum Associates. Calder, B. J., Phillips, L. W., & Tybout , A. M. (1983). Beyond external validity. Journal of Consumer Research, 10 , 112. Canary, D. J., & Spitzberg, B. H. (1989). A model of the perceived competence of conflict strategies. Human Communication Research, 15 , 630. Canary, D. J., & Stafford, L. (1992). Rela tional maintenance strategies and equity in marriage. Communication Monographs, 59 , 239. Canary, D. J., & Stafford, L. (1993). Pres ervation of relational characteristics: Maintenance strategies, equity, and locu s of control. In P. J. Kalbfleisch (Ed.), Interpersonal communication: Evol ving interpersonal relationships (pp. 237). Hillsdale, NJ: Lawrence Erlbaum. Canary, D. J., & Stafford, L. (1994). Ma intaining relationships through strategic and routine interaction. In D. J. Canary & L. Stafford (Eds.), Communication and relational maintenance (pp. 3). San Diego, California: Academic Press. Carmines, E. G., & Zeller, R. A. (1979). Reliability and validity assessment . Newbury Park, CA: Sage. Carnevale, D. G. (1995). Trustworthy government: Leadership and management strategies for building tr ust and high performance . San Francisco, CA: Jossey-Bass. Chaffee, S. H., & Roser, C. (1986). Involvement and the consistency of knowledge, attitudes, and behaviors. Communication Research, 13 , 373– 399. Chou, C. P., & Bentler, P. M. (1995). Estim ates and tests in structural equation modeling. In R. H. Hoyle (Ed.), Structural equation modeling (pp. 37). Thousand Oaks, CA: Sage. Church, A. H. (1993). Estimating the eff ect of incentives on mail survey response rates: A meta-analysis. Public Opinion Quarterly, 57 , 66. Clark, M. S., & Mills, J. (1993). The di fference between communal and exchange relationships: What it is and is not. Personality and Social Psychology Bulletin, 19 , 684.

PAGE 203

185 Coombs, W. T. (2000). Crisis management advantages of a relational perspective. In J. A. Ledingham & S. D. Bruning (Eds.), Public relations as relationship management: A relational approach to the study and practice of public relations (pp. 73). Mahwah, NJ: Lawrence Erlbaum Associates. Cronbach, L. J. (1951). Coefficient alpha and the internal st ructure of tests. Psychometrika, 16 , 297. Cutlip, S. M., Center, A. H., & Broom, G. M. (2000). Effective public relations (8th ed.). Upper Saddle River, NJ: Prentice Hall. Dainton, M. (1991, May). Relational maintenance revi sited: The addition of physical affection measures to a maintenance typology. Paper presented at the International Communicati on Association, Chicago, IL. Daley, D. M., & Vasu, M. L. (1995). Fo stering organizational trust in North Carolina: The pivotal role of admi nistrators and political leaders. Administration & Society, 30 , 62. Day, E., Stafford, M. R., & Camacho, A. (1995). Research note: Opportunities for involvement research: A scale development approach. Journal of Advertising, 24 , 69. Dimmick, S. L., Bell, T. E., Burgiss, S. G., & Ragsdale, C. (2000). Relationship management: A new professional model. In J. A. Ledingham & S. D. Bruning (Eds.), Public relations as relati onship management: A relational approach to the study and pr actice of public relations (pp. 117). Mahwah, NJ: Lawrence Erlbaum Associates. Dindia, K. (1989). Toward the development of a measure of marital maintenance strategies. Paper presented at the International Communication Association, San Francisco. Dindia, K., & Canary, D. J. (1993). Defi nitions and theoretical perspectives on relational maintenance. Journal of Social and Personal Relationships, 10 , 163. Dougall, E. K. (2006). Tracking organi zation-public relations over time: A framework for longitudinal research. Public Relations Review, 32, 174– 176. Dozier, D. M., with Grunig, L. A., & Grunig, J. E. (1995). Manager's guide to excellence in public relations and communication management . Mahwah, NJ: Lawrence Erlbaum Associates. Duck, S. (1973). Personal relationships and personal constructs . London: Wiley.

PAGE 204

186 Duncan, O. D. (1975). Introduction to structural equation models . New York: Academic Press. Dunteman, G. H. (1989). Principal component analysis . Newbury Park, CA: Sage. Dwyer, F. R., Schurr, P. H., & O h, S. (1987). Developing buyer-seller relationships. Journal of Marketing, 51 (April), 11. Dyer, J. H. (1996). How Chrysler created an American Keiretsu. Harvard Business Review, July/August , 42. Ebel, R. L. (1969). Expected reliabilit y as a function of choices per item. Educational and Psychological Measurement, 29 , 565. Ehling, W. P. (1992). Estimating the value of public relations and communication to an organization. In J. E. Grunig, D. M. Dozier, W. P. Ehling, L. A. Grunig, F. C. Repper & J. White (Eds.), Excellence in public relations and communication management (pp. 616). Hillsdale, NJ: Lawrence Erlbaum Associates. Evans, R. H. (1993). Customers' invo lvement and the replacement cycle. Psychological Reports, 72 , 1087. Farrell, D., & Rusbult, C. (1981). Excha nge variables as predictors of job satisfaction, job commitment, and turnove r: The impact of rewards, costs, alternatives, and investments. Organizational Behavior and Human Performance, 28 , 78. Ferguson, M. A. (1984). Building theory in public relations: Interorganizational relationships. Paper presented at the A ssociation for Education in Journalism and Mass Communication, Gainesville, FL. Fishbein, M. (1980). A theory of reas oned action: Some applications and implications. In M. M. Page (Ed.), Nebraska symposium on motivation (Vol. 27). Lincoln, Nebraska: University of Nebraska Press. Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention and behavior: An introduction to theory and research . Reading, MA: Addison-Wesley. Fisher, R., & Brown, S. (1988). Getting together: Building a relationship that gets to Yes. Boston: Houghton Mifflin. Florida Farm Bureau Federation (2004). Florida quick facts . Retrieved May 20, 2006, from http://www.stateofflorida.com/Port al/DesktopDefault.aspx?tabid=95#2710 5

PAGE 205

187 Florida Farm Bureau Federation (2004). The 1940s: A humble beginning . Retrieved May 3, 2006, from http://www.floridafarmbur eau.org/aboutUs_history1940.aspx Florida Farm Bureau Federation (2004). The 1960s: Expansion and political upheaval . Retrieved May 3, 2006, from http://www.floridafarmburea u.org/aboutUs_history1960s.aspx Florida Farm Bureau Federation (2004). The 1970s: The whirlwind of change . Retrieved May 3, 2006, from http://www.floridafarmburea u.org/aboutUs_history1970s.aspx Florida Farm Bureau Federation (2004). We are Farm Bureau . Retrieved May 3, 2006, from http://www. fb.org/about/thisis/ Florida Farm Bureau Federation (2004). Welcome to Florida Farm Bureau, the voice of the Florida agriculture . Retrieved May 3, 2006, from http://www.floridafarmbureau.org/Default.aspx Frankel, L. R. (1983). The report of the CA SRO task force on response rates. In F. Wiseman (Ed.), Improving data quality in a sample survey . Cambridge, MA: Marketing Science Institute. Friman, M., Grling, T., Millett, B., Mattsson, J., & Johnston, R. (2002). An analysis of internationa l business-to-business relationships based on the commitment-trust theory. Industrial Marketing Management, 31 , 403– 409. Ganesan, S. (1994). Determinants of l ong-term orientation in buyer-seller relationships. Journal of Marketing, 58 (April), 1. Garbarino, E., & Johnson, M. S. (1999). The different roles of satisfaction, trust, and commitment in customer relationships. Journal of Marketing, 63 (April), 70. Gerbing, D. W., & Anderson, J. C. (1993) . Monte Carlo evaluations of goodnessof-fit indices for structur al equation models. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation models (pp. 40). Newbury Park, CA: Sage. Greenhalgh, L. (1987). Rela tionship in negotiations. Negotiation Journal, 3 , 235– 243. Greeno, J. G., & Bjork, R. A. (1973). Math ematical learning theory and the new 'mental forestry'. Annual Review of Psychology, 24 , 81. Groves, R. M. (1989). Survey errors and survey costs . New York: John Wiley & Sons.

PAGE 206

188 Grunig, J. E. (1989). Symmetrical pres uppositions as a framework for public relations theory. In C. Bota n & V. T. Hazelton (Eds.), Public relations theory . Hillsdale, NJ: Lawrence Erlbaum Associates. Grunig, J. E. (1993). Image and substance: From symbolic to behavioral relationships. Public Relations Review, 19 , 121. Grunig, J. E., & Grunig, L. A. (1992) . Models of public relations and communication. In J. E. Grunig (Ed.), Excellence in public relations and communication management (pp. 285). Hillsdale, NJ: Lawrence Erlbaum Associates. Grunig, J. E., & Huang, Y.-H. (2000). Fr om organizational effectiveness to relationship indicators: Antecedents of relationships, public relations strategies, and relationship outcomes. In J. A. Ledingham & S. D. Bruning (Eds.), Public relations as relationship management: A relational approach to the study and pr actice of public relations (pp. 23). Mahwah, NJ: Lawrence Erlbaum Associates. Grunig, J. E., & Hunt, T. (1984). Managing public relations . New York: CBS College Publishing. Grunig, J. E., & White, J. (1992). The e ffect of worldviews on public relations theory and practice. In J. E. Grunig (Ed.), Excellence in public relations and communication management (pp. 31). Hillsdale, NJ: Lawrence Erlbaum Associates. Grunig, L. A., Grunig, J. E., & Dozier, D. M. (2002). Excellent public relations and effective organizations: A study of communication management in three countries . Mahwah, NJ: Lawrence Erlbaum Associates. Grunig, L. A., Grunig, J. E., & Ehling, W. P. (1992). What is an effective organization? In J. E. Grunig, D. M. Dozier, W. P. Ehli ng, L. A. Grunig, F. C. Repper & J. White (Eds.), Excellence in public relations and communication management (pp. 65). Hillsdale, NJ: Lawrence Erlbaum Associates. Guerrero, L. K., Eloy, S. V., & Wabni k, A. L. (1993). Linking maintenance strategies to relationship de velopment and disengagement: A reconceptualization. Journal of Social and Personal Relationships, 10 , 273. Gujarati, D. N. (2003). Basic econometrics (4th ed.). Boston: McGraw Hill. Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate data analysis. Upper Saddle River, NJ: Prentice Hall.

PAGE 207

189 Hancock, G. R., & Mueller, R. M. (2001). Rethinking construct reliability within latent variable systems. In R. C udeck, S. du Toit & D. Sorbom (Eds.), Structural equation modeling: Present and future--A Festschrift in honor of Karl Joreskog (pp. 195). Lincolnwood, IL: Scientific Software International. Heath, R. L. (Ed.). (2001). Handbook of public relations . Thousand Oaks, CA: Sage. Hon, L. C. (1997). What have you done for me lately?: Exploring effectiveness in public relations. Journal of Public Relations Research, 9 , 1. Hon, L. C., & Brunner, B. (2002). Measuri ng public relationships among students and administrators at the University of Florida. Journal of Communication Management, 6 , 227. Hon, L. C., & Grunig, J. E. (1999). Guidelines for measuring relationships in public relations . Gainesville: Instituti on for Public Relations. Hosmer, L. T. (1996). Research notes and communications: Responses to 'Do good ethics always make for good business?' Strategic Management Journal, 17 , 501. Hoyle, R. H. (1991). Evaluating measur ement models in clinical research: Covariance structure analysis of latent variable m odels of self-conception. Journal of Consulting and Clinical Psychology, 59 , 67. Hoyle, R. H., & Kenny, D. A. (1999). Samp le size, reliability, and test of statistical mediation. In R. H. Hoyle (Ed.), Statistical strategies for small sample research (pp. 195). Thousand Oaks, CA: Sage. Hoyle, R. H., & Smith, G. T. (1994). Form ulating clinical research hypotheses as structural equation models: A conceptual overview. Journal of Consulting and Clinical Psychology, 62 (3), 429. Hrebiniak, L. G. (1974). Effects of j ob level and participation on employee attitudes and percepti ons of influence. Academy of Management Journal, 17 , 649. Hu, L. T., Bentler, P. M., & Kano, Y. ( 1992). Can test statistics in covariance structure analysis be trusted? Psychological Bulletin, 112 , 351. Huang, Y.-H. (1997). Public relations strategies , relational outcomes, and conflict management strategies. Unpublished Doctoral Dissertation, University of Maryland, College Park, MD.

PAGE 208

190 Huang, Y.-H. (1999). The effects of public relati ons strategies on conflict management. Paper presented at the International Communication Association, San Francisco. Huang, Y.-H. (2001a). Values of public relations: Effects on organization-public relationships mediati ng conflict resolution. Journal of Public Relations Research, 13 , 265. Huang, Y.-H. (2001b). OPRA: A cross-cultur al, multiple-item scale for measuring organization-public relationships. Journal of Public Relations Research, 13 , 61. Huang, Y.-H. (2004). PRSA: Scale devel opment for exploring the impetus of public relations strategies. Journalism & Mass Communication Quarterly, 81 , 307. Hung, C.-j. F. (2005). Exploring types of organization-public relationships and their implications for relationshi p management in public relations. Journal of Public Relations Research, 17 , 393. Huston, T. L., McHale, S., & Crouter, A. (1986). When the honeymoon's over: Changes in the marriage relationship ove r the first year. In R. Gilmour & S. Duck (Eds.), The emerging field of personal relationships (pp. 109– 132). Hillsdale, NJ: Lawrence Erlbaum Associates. Hutton, J. G., Goodman, M. B., Alexa nder, J. B., & Genest, C. M. (2001). Reputation management: The new f ace of corporate public relations? Public Relations Review, 27 , 247. James, L. R., Muliak, S. A., & Brett, J. M. (1982). Causal analysis: Assumptions, models and data . Beverly Hills, CA: Sage. Jarvis, C. B., Mackenzie, S. B., & Podsa koff, P. M. (2003). A critical review of construct indicators and measur ement model misspecification in marketing and consumer research. Journal of Consumer Research, 30 , 199. Jeffres, L. W., & Perloff, R. M. (1997). Mass media effects (2nd Ed.). Prospect Heights, IL: Waveland. Jo, S. (2003). Measurement of organization-public relationships: Validation of measurement using a manufacturer-retailer relationship . Unpublished doctoral dissertation, University of Florida, Gainesville. Jo, S. (2006). Measurement of organiza tion-public relationships: Validation of measurement using a manufact urer-retailer relationship. Journal of Public Relations Research, 18 , 225.

PAGE 209

191 Jo, S., Hon, L. C., & Brunner, B. R. (2004). Organization-pub lic relationships: Measurement validation in a University setting. Journal of Communication Management, 9 , 14. Johnson, R. A., & Wichern, D. W. (1992). Applied multivariate statistical analysis (3rd ed.). Englewood Cliffs, NJ: Prentice Hall. Jreskog, K., & Srbom, D. (1996). LISREL 8: User's reference guide. Chicago: Scientific Software International. Kanter, R. M. (1971). Commitment a nd social organiza tion: A study of commitment mechanisms in Utopian communities. American Sociological Review, 33 , 499. Kaplan, D. (2000). Structural equation modelin g: Foundations and extensions . Thousand Oaks, CA: Sage. Kelleher, T., & Miller, B. M. (2006). Or ganizational blogs and the human voice: Relational strategies and relational outcomes. Journal of ComputerMediated Communication, 11 , article 1. Kelley, H. H. (1973). The pro cesses of causal attribution. American Psychologist, 28 , 107. Kelley, H. H., & Thibaut, J. W. (1978). Interpersonal relatio ns: A theory of interdependence . New York: Wiley. Kenny, D. A. (1979). Correlation and causation . New York: Wiley. Kenny, D. A., & Kashy, D. A. (1991). Analyz ing interdependency in dyads. In B. M. Montgomery & S. Duck (Eds.), Studying interpersonal interaction (pp. 275). New York: Guilford. Ki, E.-J. (2003). Relationship maintenance strategies on Fortune 500 company Web sites. Unpublished Master's Thesis, Univ ersity of Florida, Gainesville, FL. Ki, E.-J., & Hon, L. C. (2007). Testi ng the linkages among the organizationpublic relationship and attitude and behavioral intentions. Journal of Public Relations Research, 19 (1). Ki, E.-J., & Hon, L. C. (2005). Testing the linkages among the organizationpublic relationship and attitude and behavioral intentions. Paper presented at the Association for Education in Journalism and Mass Communication, San Antonio, TX.

PAGE 210

192 Ki, E.-J., & Hon, L. C. (2006). Relations hip maintenance strategies on Fortune 500 company Web sites. Journal of Communication Management, 10 , 27– 43. Ki, E.-J., & Shin, J.-H. (2006). The stat us of organizationpublic relationship research from an analysis of published articles, 1985-2004. Public Relations Review, 32 , 194. Kile, O. M. (1948). The farm bureau through three decades . Boltimore, MD: The Waverly Press. Kim, Y. (2001). Searching for the organi zation-public relatio nship: A valid and reliable instrument. Journalism & Mass Communication Quarterly, 78 , 799. King, D. C., & Walker, J. L. (1992). The pr ovision of benefits by interest groups in the United States. Journal of Politics, 54 , 394. Kintsch, W. (1970). Learning, memory and conceptual processes . New York: John Wiley. Kline, R. B. (2005). Principles and practice of st ructural equation modeling (2nd ed.). New York: The Guildford Press. Krugman, H. E. (1965). The impact of te levision advertising: Learning without involvement. Public Opinion Quarterly, 29 , 349. Kwon, I.-W. G., & Suh, T. (2004). Factor s affecting the level of trust and commitment in supply chain relationships. Journal of Supply Chain Management, 40 (2), 4. Lavidge, R. J., & Steiner, G. A. (1961). A model for predictive measurements of advertising effectiveness. Journal of Marketing, 25 , 59. Ledingham, J. A. (2003). Explicating Re lationship Management as a General Theory of Public Relations. Journal of Public Relations Research, 15 , 181. Ledingham, J. A., & Bruning, S. D. (1998) . Relationship management in public relations: Dimensions of an organization-public relationship. Public Relations Review, 24 , 55. Ledingham, J. A., & Bruning, S. D. (2000) . A longitudinal study of organizationpublic relationship dimensions: Defi ning the role of communication in practice of relationship management. In J. A. Ledingham & S. D. Bruning (Eds.), Public relations as relationship management: A relational approach to the study and pr actice of public relations (pp. 55). Mahwah, NJ: Lawrence Erlbaum Associates.

PAGE 211

193 Ledingham, J. A., Bruning, S. D., Thomlison, J. A., & Lesko, C. (1997). The applicability of interpersonal relation ship dimensions to an organizational context: Toward a theory of relational loyalty. Academy of Managerial Communications Journal, 1 , 23. Ledingham, J. A., Bruning, S. D., & Wilson, L. J. (1999). Time as an indicator of the perceptions and behavior of me mbers of a key public: Monitoring and predicting organization-public relationships. Journal of Public Relations Research, 11 , 167. Leichty, G. B., & Springston, J. (1993). Reconsidering public relations models. Public Relations Review, 19 , 327. Lez-Herrero, A. G., & Pratt, C. B. (1996). An integrated symmetrical model for crisis-communications management. Paper presented at the Association for Education in Journalism and Mass Communication. Lewis, R. A., & Spanier, G. (1979). Theo rizing about the quality and the stability of marriage. In W. R. Burr, R. Hill, F. I. Nye & I. L. Reiss (Eds.), Contemporary theories about the family (Vol. 1). New York: Free Press. Lindenmann, W. K. (1999). Overview . In L. C. Hon & J. E. Grunig, Guidelines for measuring relationshi ps in public relations . Gainesville: The Institute for Public Relations. Lindenmann, W. K. (2002). Guidelines for measuring the effectiveness of PR programs and activities . Gainesville, FL: Institute for Public Relations. MacCallum, R. C. (1995). Model specificat ion: Procedures, strategies, and related issues. In R. H. Hoyle (Ed.), Structural equation modeling: Concepts, issues, and applications (pp. 16). Thousand Oaks, CA: Sage. Marshall, C., & Rossman, G.B. (1998). Designing qualitative research (3rd ed.). Thousand Oaks: Sage Publications. Massey, J. T. (1995). Estimating the response rate in a telephone survey with screening. Paper presented at the Amer ican Statistical Association, Alexandria, VA. McGuire, W. J. (1986). The myth of massive media impact: Savagings and salvagings. In G. Comstock (Ed.), Public communication and behavior (Vol. 1, pp. 173). Orlando, FL: Academic Press. Meyer, J. P., & Allen, N. (1984). Testing the side-best theory of organizational commitment: Some methodological considerations. Journal of Applied Psychology, 69 (3), 372.

PAGE 212

194 Miettil, A., & Mller, K. (1990). Interaction perspective into professional business service: A conceptual analysis. Paper presented at the Research Development on International Industria l Marketing and Purchasing, Milan. Miles, E. W., Patrick, S. L., & King, W. C. (1996). Job level as a systemic variable in predicting the relations hip between supervisory communication and job satisfaction. Journal of Occupati onal and Organizational Psychology, 69 , 277. Millar, F. E., & Rogers, L. E. (1976). A relational approach to interpersonal communication. In G. R. Miller (Ed.), Explorations in interpersonal communication . Thousand Oaks, CA: Sage. Mitchell, A. A., & Olson, J. C. (1981). Are product attribut e beliefs the only mediator of advertising effects on brand attitude? Journal of Marketing Research, 18 , 318. Moorman, C., Deshpand, R., & Zaltman, G. (1993). Factors affecting trust in marketing research relationships. Journal of Marketing, 57 (1), 81. Moore, C. W. (1986). The mediation process . San Francisco: Jossey-Bass. Morgan, R. M., & Hunt, S. D. (1994). The commitment-trust theory of relationship marketing. Journal of Marketing, 58 (3), 20. Morrill, C., & Thomas, C. K. (1992). Or ganizational conflict management as disputing process: The probl em of social escalation. Human Communication Research, 18 , 400. Muliak, S. A., James, L. R., Van Alstine, J., Bennett, S., Lind, S., & Stillwell, D. C. (1989). An evaluation of goodness of fit indices for structural equation models. Psychological Bulletin, 103 , 430. Nohria, N., & Eccles, R. G. (1992). F ace-to-face: Making network organizations work. In N. Nohria & R. G. Eccles (Eds.), Networks and organizations: Structure, form, and action (pp. 288). Boston, Massachusetts: Harvard Business School Press. Nunnally, J. C. (1978). Psychometric theory (2nd ed.). New York: McGraw-Hill. Palda, K. S. (1964). The measurement of cumu lative advertising effects . Englewood Cliffs, NJ: Prentice-Hall. Parasuraman, A., Zeithaml, V. A., & Berry, L. L. (1988). SERVQUAL: A multiple-item scale for measuring consumer perceptions of service quality. Journal of Retailing, 64 , 12.

PAGE 213

195 Park, H. S., Dailey, R., & Lemus, D. (2002). The use of exploratory factor analysis and principal components an alysis in communication research. Human Communication Research, 28 , 562. Perloff, R. M. (2003). The Dynamics of persuasion: Communication and attitudes in the 21st century (2nd ed). Mahwah, NJ: Lawrence Erlbaum Associates. Plowman, K. D. (1995). Congruence between public relations and conflict resolution: Negotiating in the organization. Unpublished doctoral dissertation, University of Maryland, College Park. Popper, K. R. (1965). Conjectures and refutations: The growth of scientific knowledge . New York: Basic Books. Porter, L. W., Crampon, W. J., & Smith, F. J. (1976). Organizational commitment and managerial turnover. Organizational Behavior and Human Performance, 15 , 87. Porter, L. W., Steers, R. M., Mowd ay, R. T., & Boulian, P. V. (1974). Organizational commitment, job satisfaction, and turnover among psychiatric technicians. Journal of Applied Psychology, 59 , 603. Putman, L. L. (1990). Reframing integr ative and distribut ive bargaining: A process perspective. In B. H. Sheppard, M. H. Bazerman & R. J. Lewicki (Eds.), Research on negotiation in organizations . Greenwich: Jai. Putman, L. L., & Wilson, C. E. ( 1982). Communication strategies in organizational conflict: Reliability and validity of a measurement scale. In M. Burgoon (Ed.), Communication yearbook (Vol. 6). Beverly Hills: Sage. Ray, M. L. (1973). Marketing communication and the hierarchy-of-effects. In P. Clarke (Ed.), N ew models for mass communication research . Beverly Hills, CA: Sage. Ray, M. L. (1982). Advertising and communication management . Englewood Cliffs, NJ: Prentice Hall. Ray, M. L., Sawyer, A. G., Rothschild, M. L., Heeler, R. M., Strong, E. C., & Reed, J. B. (1973). Marketing communicat ions and the hierarchy of effects. In P. Clarke (Ed.), New models for mass communication research (pp. 147). Beverly Hills, CA: Sage. Raykov, T., & Marcoulides, G. A. (2000) . A method for comparing completely standardized solutions in multiple groups. Structural Equation Modeling, 7 , 292. Raykov, T., & Widaman, K. F. (1995). Issues in applied structural equation modeling research. Structural Equation Modeling, 2 , 289.

PAGE 214

196 Reichers, A. E. (1985). A review and reconceptualization of organizational commitment. Academy of Management Review, 10 , 465. Robinson, J. P., Shaver, P. R., & Wright sman, L. S. (1991). Criteria for scale selection and evaluation. In J. P. Robinson, P. R. Shaver & L. S. Wrightsman (Eds.), Measures of personality and social psychological attitudes . San Diego, CA: Academic Press. Rogers, E. M. (1995). Diffusion of innovations (4th ed.). New York: Free Press. Rosenberg, M. J., & Hovland, C. I. (1963) . Cognitive, affective, and behavioral components of attitudes. In M.J. Rose nberg, C.I. Hovland, W.J. McGuire, R.P. Aberlson & J.W. Brehm, (Eds)., Attitude organization and change: An analysis of consistency among attitude components , (pp.1). New Haven: Yale University Press. Rotter, J. B. (1967). A new scale for the measurement of interpersonal trust. Journal of Personality, 35 , 651. Rusbult, C. E. (1983). A longitudinal test of the investment model: The development (and deterioration) of satisfaction and commitment in heterosexual involvements. Journal of Personality and Social Psychology, 45 , 101. Sanchez, M., & Iniesta, M. A. (2004). Th e structure of commitment in consumerretailer relationships: Conceptu alization and measurement. International Journal of Service Industry Management, 15 , 230. Sallot, L. M., Lyon, L. J., Acosta-Alz uru, C., & Jones, K. O. (2003). From Aardvark to Zebra: A new millennium analysis of theory development in public relations academic journals. Journal of Public Relations Research, 15 , 27. Selnes, F. (1998). Antecedents and conse quences of trust and satisfaction in buyer-seller relationships. European Journal of Marketing, 32 , 305. Severin, W. J., & Tankard, J. W., Jr. (2001). Communication theories: Origins, methods, and uses in the mass media . San Francisco: Longman. Sheldon, M. E. (1971). Investments and involvements as mechanisms producing commitment to the organization. Administrative Science Quarterly, 16 , 143. Spector, P. E. (1992). Summated rating scales construction: An introduction . Newbury Park, CA: Sage.

PAGE 215

197 Stafford, L., & Canary, D. J. (1991). Maintenance strategies and romantic relationship type, gender, a nd relational characteristics. Journal of Social and Personal Relationships, 8 , 217. Standard definitions: Final di spositions of case codes and outcome rates for surveys . (2006). The American Associati on for Public Opinion Research. Steers, R. M. (1977). Antecedents and outcomes of organizational commitment. Administrative Science Quarterly, 22 , 46. Stevens, J. M., Beyer, J. M., & Trice, H. M. (1978). Assessing personal, role and organizational predictors of managerial commitment. Academy of Management Journal, 21 , 380. Sypher, B. D. (1990). Case studies in organizational communication . New York: Guilford. Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics (4th ed.). Boston: Allyn and Bacon. Tanaka, J. S. (1987). How big is big enough?: Sample size an d goodness of fit in structural equation models with latent variables. Child Development, 58 , 134. Tanaka, J. S. (1993). Multifaceted conceptions of fit in structural equation models. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation models (pp. 10). Newbury Park, CA: Sage. Thibaut, J. W., & Kelly, H. H. (1959). The social psychology of groups . New York: John Wiley & Sons, Inc. Thomlison, J. A. (2000). An interpersonal primer with implications for public relations. In J. A. Ledingha m & S. D. Bruning (Eds.), Public relations as relationship management: A relational approach to the study and practice of public relations (pp. 177). Mahwah, NJ: Lawrence Erlbaum Associates. Toth, E. L. (2000). From personal influen ce to interpersonal influence: A model for relationship management. In J. A. Ledingham & S. D. Bruning (eds.), Public relations as relationship m anagement: A relational approach to the study and practice of public relations . Mahwah, NJ: Lawrence Erlbaum Associates. Valente, T. W., Paredes, P., & Poppe, P. R. (1998). Matching the message to the process: The relative ordering of know ledge, attitudes, and practices in behavior change research. Human Communication Research, 24 , 366.

PAGE 216

198 Vercic, D., & Grunig, J. E. (1995). The origins of public relations theory in economics and strategic management. Paper presented at the Second International Public Relations Rese arch Symposium, Bled, Slovenia. Webster, M. (2006). Merriam-Webster's co llegiate dictionary . (11th ed.). Springfield, MA: Merriam-Webster. Weisberg, H. F., Krosnick, J. A. & Bowen, B. D. (1996). An introduction to survey research, polling, and data analysis (3rd ed). Thousand Oaks, CA: Sage. Whang, P. A., & Hancock, G. R. (1997). Modeling the mathematics achievement of Asian-American elementary students. Learning and Individual Differences, 9 , 63. White, H. (1980). A heteroscedasticity cons istent covariance matrix estimator and a direct test of heteroscedasticity. Econometrica, 48 , 817. Whitener, E. M., Brodt, S. E., Korsg aard, M. A., & Werner, J. M. (1998). Managers as initiators of trust: An exchange relationship framework for understanding managerial trustworthy behavior. Academy of Management Review, 23 , 513. Williams, L. J., & Anderson, S. E. (1991) . Job satisfaction and organizational commitment as predictors of organi zational citizenship behaviors and inrole behaviors. Journal of Management, 17 , 601. Williams, L. J., & Hazer, J. T. (1986). Antecedents and consequences of satisfaction and commitment in trust models: A reanalysis using latent variable structural equation methods. Journal of Applied Psychology, 71 , 219. Wilmot, W. W., & Sillars, A. L. (1989) . Developmental issues in personal relationships. In J. F. Nussbaum (Ed.), Life-span communication: Normative process (pp. 119). Hillsdale, NJ: Lawrence Erlbaum Associates. Wilson, L. J. (2001). Relationships within communities: Public relations for the new century. In R. L. Heath (Ed.), Handbook of public relations (pp. 521– 526). Thousand Oaks, CA: Sage. Wimmer, R. D., & Dominick, J. R. (2000). Mass media research: An introduction . Belmont, CA: Wadsworth. Woo, C. (2001). Optimal Aad-Ab-PI measures in the journal of advertising: 1991 through 2000. Unpublished master’s thesis , University of Florida, Gainesville.

PAGE 217

199 Wood, J. T. (1995). Relational communication . Belmont, CA: Wadsworth. Zeithaml, V. A., Perry, L. L., & Parasuraman, A. (1996). The behavioral consequences of service quality. Journal of Marketing, 60, 31.

PAGE 218

200 BIOGRAPHICAL SKETCH Eyun-Jung Ki double majored at Sookmyung Women’s Univer sity in Seoul, Korea, earning a B.A. with academic distinction in both English Language & Literature and Mass Communication. She received her mast ers’ degree in Mass Communication with an emphasis in Public Relations from Univer sity of Florida at Gainesville in 2003. She also completed her Ph.D. in Mass Communication focusing on Public Relations from the University of Florida at Gainesville in 2006. During her doctoral studies, Eyun-Jung was awarded the first Al and Effie Flanagan Professor assistantship and worked as a research/teaching assistant. She taught tw o classes, Public Relations Campaigns and Public Relations Visu al Communications. Her research work has appear ed in the following journals: Journal of Communication, Journal of Communication Ma nagement, Journal of Media Economics, and Journal of Public Relations Research, J ournal of Radio Studies, Public Relations Review , and International Journal on Media Management . Eyun-Jung also received several prestigious awards including the ‘ 2006 Graduate Student Research Award’ from the College of Journalism and Communications at the Universi ty of Florida, the ‘Korean Graduate Student Research Award’ from the Un iversity of Florida, the Top Student Paper Award from The International Communicati on Association, and the Third Place Paper Award from The Association for Educati on in Journalism and Mass Communication, among many others.