Citation
Toward a model of leadership perception in small group interactions

Material Information

Title:
Toward a model of leadership perception in small group interactions
Creator:
Burris, Patricia Ann, 1949-
Publication Date:
Language:
English
Physical Description:
xii, 223 leaves : ; 28 cm.

Subjects

Subjects / Keywords:
Adjectives ( jstor )
Buses ( jstor )
College students ( jstor )
Gestures ( jstor )
Leadership ( jstor )
Parking ( jstor )
Personality traits ( jstor )
Ratings ( jstor )
Social interaction ( jstor )
Social psychology ( jstor )
Dissertations, Academic -- Psychology -- UF ( lcsh )
Leadership ( lcsh )
Psychology thesis Ph. D ( lcsh )
Small groups -- Research ( lcsh )
City of Gainesville ( local )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1981.
Bibliography:
Includes bibliographical references (leaves 216-221).
Additional Physical Form:
Also available online.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Patricia Ann Burris.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
07904908 ( OCLC )
0028132594 ( ALEPH )

Downloads

This item has the following downloads:


Full Text












TOWARD A MODEL OF LEADERSHIP PERCEPTION
IN SMALL GROUP INTERACTIONS










BY


PATRICIA ANN BURRIS






















A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY



UNIVERSITY OF FLORIDA 1981
































Copyright 1981 by


Patricia Ann Burris















DEDICATION









To Gill Woodall
my dearest friend, my life partner.

You share every joy
and every pain of my life.
You bring meaning to it all.

Thank you for sharing with me
the toil and frustration
of this dissertation
and the joy of its completion.

Now let's get on to better things
together.















ACKNOWLEDGMENTS


A dissertation is rarely the sole effort of one

individual. Others provide inspiration, ideas, criticisms, and emotional support during the course of doctoral and dissertation work. I therefore wish to offer thanks to the following individuals. To Dr. Marvin Shaw who chaired this research, my deep gratitude. Thank you for your patience and continual support throughout this effort. You were the perfect chairman. My respect for you and your scholarship has increased with each year that I have known you. It has been a pleasure to be your student and I hope that relationship continues for many years to come. To Dr. Norman Markel, special thanks to you for initiating my interest in communicative behaviors as the founding father of the Gainesville Society for the Study of Coverbal Behavior. Thank you for your contributions to my education as a communication researcher and equally, for the political education and friendship you have given me. To Dr. Tom Simon, very special thanks for your insightful questions and suggestions, for your unending enthusiasm and support. Thank you for your friendship and especially, for the model you provide of an academician with broad interests and as one who takes positions and is



iv









willing to fight for them. To Dr. Bob Ziller for your creative and original work and to Dr. Patricia Miller for so kindly agreeing to serve on my committee at the last moment.

I wish to thank my hard-working and under-paid

actors, Bill Beck, Matt Cupach and Paul Hanley. If effort and talent have any reward, you will all have successful futures. Thanks also to my highly capable and dependable research assistants, Linda Helman and Kirk McDonnel. You are both well on your way to becoming excellent social scientists. My gratitude to my subjects who generously gave me an hour of their time. Special thanks to Sandra Coy, my typist, for her excellent and speedy work and her patience.

Finally, for those who contributed to my happiness, sanity, fulfillment, and sense of purpose beyond the narrow bounds of this dissertation, which after all was only a relatively small part of my life these past several years, my gratitude and love. Thank you, my parents, Kiddle, dear Cindy Gallois, my brothers and sisters of the Gainesville Society for the Study of Coverbal Behavior and the Graduate Student Union, my Senseis and fellow karateka of Coung-Nhu Karate and Dragon Gate Dojo, and to my husband, Gill Woodall, who is part of everything I am and do.








v















TABLE OF CONTENTS

Page

ACKNOWLEDGMENTS .. . . . . . . . iv

LIST OF TABLES. ............. . viii

LIST OF FIGURES ................. x

ABSTRACT ....... . ............. . . .. xi


CHAPTER

I INTRODUCTION ....... . .. .. . 1

Conceptions of and Approaches to the
Study of Leadership . . . . . 3 Trait Approach. ... .. ..... . 3
Style Approach .... ... . . 6
Behavioral Approach . . . . . 10 Situational Approach ...... . 17 Interactionist Approach ... .... 23
Leadership as an Influence
Relationship ........ . . . 28
Rationale and Statement of Hypotheses
for Studies One and Two .. ...... 38 Study One ..... ....... . 38
Study Two . . . . . . . 45


II METHOD. .... . . . . . ... 52

Study One . . . . . . . 52
Subjects .. . . . . . . . 52
Procedure . . . . . . . 52
Design . . . . . . . 54
Group Interaction and Videotape
Construction . . . . . . 56
Covariate Measures. .. .. ..... 83
Dependent Measures. .......... 85
Study Two .... .... ... . . 87
Subjects .. ..... . . .. . 87
Procedure . . . . .. . . .. 88
Design .. ............. . 88



vi









Chapter Page

Mode of Presentation-Communicative
Channels .. ............. 89
Dependent and Covariate Measures . 92


III RESULTS ....... ........ . 93

Study One . . . . . . . . 93
Test of Hypotheses for Study One . .. 93 Supplementary Analyses for Study One 105
Study Two .... ..... ..... . 112
Test of Hypotheses for Study Two . . 112 Supplementary Analyses for Study Two 122


IV DISCUSSION ................. 129

Summary of Results for Study One .... 129 Summary of Results for Study Two . . 138 Recommendations for Future Research . 144 APPENDICES

A Bales Content Analysis Categories ..... 150

B Pre-Test Scale for Perceived Naturalness
of Interaction .. .... ....... . 152

C Pre-Test Scales for Perceived Actor
Credibility, Interpersonal Attraction
and Leadership ... ........... 154

D Covariate Scales ............. 159

E Post-Experimental Measures .... ... 162 F Study Two Condition Four Transcript . .. 173
(Low Influence, Low Leader-Like
Behavior Condition)

G Study Two Condition Four Transcript .. 180
(High. Influence, Low Leader-Like
Behavior Condition)

H Verbatim Verbal and Nonverbal Scripts
From Videotaped Interactions ..... 187 BIBLIOGRAPHY ..... ......... . ..... 216

BIOGRAPHICAL SKETCH ............... 222



vii















LIST OF TABLES

Table Page

1. Experimental Conditions for Study One . 57

2. Summary of the Verbal and Nonverbal
Behaviors Used to Operationalize LeaderLike Behavior Factor . . .. . 67

3. Summary of Verbal and Nonverbal Behaviors
Used to Operationalize Influence Factor 68

4. Summary of Frequencies and Percentages for
Behaviors Indicative of Level of LeaderLike Behavior and Level of Influence for
Group Members One, Two and Three for
Condition One ...... .... ... 71

5. Summary of Frequencies and Percentages for
Behaviors Indicative of Level of LeaderLike Behavior and Level of Influence for
Group Members One, Two and Three for
Condition Two ........... . .73

6. Summary of Frequencies and Percentages for
Behaviors Indicative of Level of LeaderLike Behavior and Level of Influence for
Group Members One, Two and Three for
Condition Three ...... . ... . 75

7. Summary of Frequencies and Percentages for
Behaviors Indicative of Level of LeaderLike Behavior and Level of Influence for
Group Members One, Two and Three for
Condition Four ...... ... . . 77

8. ANOVA Summary for Differences Between
Actors Based on Physical Appearance on
Pre-test Scales . . . . . .. 84

9. Experimental Conditions for Study Two . .. 90

10. ANOVA and ANCOVA Summary for Leadership Ratings . . . . . . . . 94




viii









Table Page 11. Summary of Adjusted Means for Leadership Ratings . . . . . . . . 95

12. ANCOVA Summary for Leadership Ratings for Group Members Within Cells Two, Three
and Four . . . . . . . . 100

13. Summary of Adjusted Means for Leadership Ratings for Group Members Within Cells
Two, Three and Four .......... 102

14. ANCOVA Summary for Leadership Ratings for Group Members Within Cell One ...... 103 15. Summary of Adjusted Means for Leadership Ratings for Group Members Within Cell
One ......... ....... . 104

16. ANOVA and ANCOVA Summary for Influence Ratings . .. . . . . .... 107

17. Summary of Adjusted Means for Influence Ratings . . . . . . . . 108

18. ANOVA and ANCOVA Summary for Confidence Ratings . .. . . . . ... . 110

19. Summary of Adjusted Means for Confidence Ratings .. ........ . . ill

20. ANOVA Summary for Leadership Ratings for Group Members Within Conditions . . . 116 21. Summary of Means for Leadership Ratings for Group Members for Each Condition . .. 117 22. ANOVA Summary for Leadership Ratings for Group Member Three . . . . . . 118

23. Summary of Means for Leadership Ratings for Group Member Three . . . . . 120

24. ANOVA Summary for Influence Ratings for Group Member Three ............ 124

25. Summary of Means for Influence Ratings for Group Member Three ......... 125

26. ANOVA Summary for Confidence Ratings. .... 127 27. Summary of Means for Confidence Ratings . 128


ix















LIST OF FIGURES

Figure Page

1. Interaction of Influence and Channel for
Leadership Ratings ....... ...... 121

2. Interaction of Influence and Channel for
Influence Ratings ..... .. . . 126









































x














Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

TOWARD A MODEL OF LEADERSHIP PERCEPTION
IN SMALL GROUP INTERACTIONS

By

Patricia Ann Burris

August 1981

Chairman: Dr. Marvin E. Shaw Major Department: Psychology

In contrast to trait, style, and behavioral views of leadership, recent theorists suggest leadership is more properly conceptualized as an influence, relational phenomenon. The purpose of the present studies was to provide support for the view that perception of small group leadership is associated with the presence of an influence relationship existing among group members and that this relationship is conveyed through verbal and nonverbal channels.

For study one, it was predicted that strongest

perceptions of leadership occur where levels of influence are high (one group member has high degrees of influence over other group members) rather than low (members are equally influential). Further, perception of leadership is more strongly associated with influence levels than levels of leader-like behavior (one group member engages in low or high levels of those behaviors previously listed xi








in the literature as typical of leaders). For study two, it was predicted that influence relationships, thus leadership, would be conveyed along verbal and nonverbal communicative channels.

Study one employed a two (high vs. low levels of influence) by two (high vs. low levels of leader-like behavior) design. Three actors were videotaped engaging in scripted group interactions. Experimental conditions were created by varying verbal and nonverbal behaviors of each actor such that either all group members were equally influential and engaged in equal numbers of leader-like behaviors, or one group member was either more influential than other group members, engaged in more leader-like behaviors, or was both more influential and engaged in more leader-like behavior than other group members. Study two experimental conditions were created by presenting audio-visual, visual-only, audio-only, or written transcript-only information of the interactions videotaped for study one.

Results supported the hypotheses. Perception of

leadership was related to high levels of influence and was more strongly associated with the influence factor than leader-like behavior factor. Influence relationships were conveyed across verbal and nonverbal channels, with exception of the audio-only channel. Results emphasize the importance of viewing leadership as an influence relation existing among group members and conveyed through verbal and nonverbal behaviors.
xii















CHAPTER I


INTRODUCTION


The purpose of each of the present investigations is to provide empirical support for recently acknowledged views that (1) leadership is a relational phenomena (leadership can exist only as part of an interpersonal relationship. Leadership is an attribute of a relationship among people, not an attribute of a single individual). In particular, leadership is properly conceptualized as an influence relationship, and (2) that the most appropriate approach to the study of leadership is one which focuses on the types of interaction patterns occurring among group members. This approach is called the interactionist approach. Each study provides information regarding the types of group interactions leading to the perception of leadership. Study one demonstrates that interaction patterns which are indicative of influence relationships existing among group members are strongly associated with the perception of leadership within the group. Study two demonstrates that group interactions occurring along verbal as well as nonverbal channels convey influence and hence, leadership information.




I





2


Interest in leadership in small groups has been long standing and at times, intense. Certainly this reflects the importance of small group membership and leadership roles in our lives. We interact daily in a variety of small group situations (within the family, in informal social and community organizations, and in more formal business and political groups). Leadership is an important aspect of nearly all of these group memberships. As Hollander (1978) states, "almost any task related to an organized activity involves leadership, or at least is associated with it. There is nothing so central to the functioning of groups or organizations, whether in government, industry, or any other place in society" (p. 3). Others have noted that of all other concepts in small group process and small group structure, leadership has received the most empirical and theoretical attention (Cartwright & Zander, 1968; Lashbrook, 1975; Shaw, 1976).

The following sections of this chapter will provide a review of some of the theoretical and empirical findings which have resulted from years of work in this area. The purpose of this survey is twofold: (1) to provide some review of the knowledge gains made in this area over the years, and (2) to point out some of the failings of this work in an effort to provide a rationale for the theoretical underpinnings of the present investigations: that leadership is a relational phenomenon and in particular an influence relationship, and that the most





3


useful approach to the study of leadership is an interactionist approach. It is suggested here that some of the failings of earlier work center around the inappropriate choice of an approach to the study of leadership (e.g., the trait, style, behavioral, and situational approaches). Further, these approaches were selected either due to the lack of clear agreement regarding the definition of leadership or due to an inaccurate conception of the nature of leadership. With regard to this definitional problem, several researchers have noted that there has been little agreement regarding the nature of leadership (Gibb, 1969; Shaw, 1976). Bass (1960) recorded over 130 definitions prior to 1949. More recently, others have noted that there are practically as many definitions of leadership as there are leadership theories, and nearly as many of those as there are psychologists working in the field (Fiedler, 1971, p. 1). The investigations reported here should serve to clarify the nature of leadership. Conceptions of and Approaches to the Study of Leadership

Trait approach. At least among psychologists, the earliest and most actively pursued approach to the study of leadership involved the search for leadership traits. This approach encompassed hundreds of studies in which leadership theorists, largely in the early part of the century, attempted to identify those personality or other individual attributes which accounted for the leadership






4


of certain individuals (for discussions of this approach see Cartwright & Zander, 1968; Gibb, 1969; Hollander & Julian, 1969). This approach to the study of leadership carried with it an underlying conception of what leadership was. The ability to behave as a leader was assumed to be based upon some underlying, unitary attribute (trait) or pattern of attributes (traits) within the individual. Leadership was conceptualized as an aspect of personality: a quality, characteristic, or predisposition which accounted for why some individuals rose to positions of leadership, which was presumed to distinguish leaders from non-leaders, and which was held in common by all leaders across variations in time and situation (Fiedler & Chemers, 1974; Shaw, 1976).

However, the results of the trait approach have been disappointing. Within the leadership area few consistencies across investigations were found. When consistencies in the traits associated with leadership were found (e.g., traits which were found across a variety of leaders and leadership situations, traits which appeared to distinguish leaders in general from non-leaders) the correlations obtained most often tended to be low (Cartwright & Zander, 1968; Gibb, 1969; Hollander & Julian, 1969; Mann, 1959; Shaw, 1976; and Stogdill, 1948). Included among those traits found to correlate at least to some extent, with leadership were the following: in general the leadership status was more often than not





5


associated with superior intelligence, scholarship, insight and knowledge of how to get things done. In contrast to non-leaders, leaders tended to be more original, they had greater verbal fluency, were more persistent, more self-confident, more responsible, dependable, and cooperative. They were more aggressive, more lively, and had greater social participation and a greater desire to excel (Stogdill, 1948). Other reviews of this literature may be found in Bass (1960), Gibb (1969), and Mann (1959).

Overall, the view that leadership may be defined in terms of personality traits, that there is one basic personality pattern for leaders regardless of the leader or the type of leadership situation, is not well supported. Further, the approach to the study of leadership which involves the search for and measurement of specified personality traits does not appear to be useful. Both the approach and the underlying view of leadership have been largely abandoned by current social scientists. Numerous studies within this tradition failed altogether

to find any consistent pattern of traits which characterized leaders (Gibb, 1969). Those classes of traits which were found to be associated with leaders (1) did not distinguish leaders from followers to any great degree and

(2) had low consistency across situations. However, as is true of many investigative approaches, there were a few gains with this approach in terms of our understanding of leadership. When the situation is taken into account,





6


certain general classes of traits do appear to relate, at least to some extent, to leadership. Group members' personalities do make a difference to group performance and would thus be expected to affect that aspect of group behavior to which the leadership concept applies (Gibb, 1969, p. 227). Following his review of the trait approach to leadership, Gibb concluded that

a person does not become a leader by virtue of his possession of any one particular pattern of personality traits, but the pattern of personal characteristics of the leader must bear relevant relationship to the present characteristics, activities, and goals
of the group of which he is a leader. (p. 226.)

In conclusion, despite the fact that personality traits may play some role in terms of group behavior, including leader behavior, the trait approach does not appear to be a useful one for understanding the nature of leadership.

Style approach. The next approach to the study of leadership involved the identification and study of styles of leadership (such as, autocratic, democratic, laissezfaire; participatory, supervisory; authoritarian, nonauthoritarian; nondirective and directive) and the nature of their effects on groups. This tradition is primarily exemplified by the classic work of Lewin, Lippitt and White (1939) and Lippitt and White (1943). This approach to leadership involved a change in focus in two different ways from the earlier trait approach. First, this approach directed attention away from those individual trait characteristics which predisposed given individuals to





7


leadership and focused attention instead on the types of motivational and behavioral tendencies associated with different types of leaders or styles of leadership. Thus, this approach acknowledged differences in leaders and focused on these differences in terms of both the internal motives and the behavioral tendencies of leaders. Second, attention was focused for the first time on the effects of leadership behavior on group interaction and outcome variables such as measures of group productivity and satisfaction.

One problem within this tradition has been the confusion with regard to the term "leadership style." According to Fiedler (1971), some investigators have used the term leadership style to refer to the leader's typical way of behaving. For example, Hersey and Blanchard (1976) define leadership style in terms of the consistent pattern of behavior exhibited when one attempts to influence the activities of others. Attempts to influence may be typically task-oriented or relationship-oriented behavior (or some combination of both). Such consistency in behavior is developed over time, eventually coming to be known as the leader's style or leadership personality. This type of definition defines leadership style in terms of the consistencies in the behavior of the leader. Other investigators (Fiedler, 1971, for example) define leadership style in terms of both leader and follower behavior and in terms of behavior as well as the motives underlying









that behavior. Thus, leadership style refers to relatively consistent systems of interacting with others in subordinate positions (Fiedler, 1971, p. 10). Leadership style refers to a consistent pattern of behavioral interaction among both leader and non-leaders. Further, Fiedler relates these consistencies of interaction to the motivational system of the individual leader (that is, individuals may tend to be more relationship motivated or more task motivated).

While there remains some disagreement regarding

the term "leadership style," this approach to the study of leadership was not specifically aimed at clarifying the nature of leadership itself. Rather, the assumption appears to have been made that leadership, regardless of the particular leadership style employed, involved either an influence relationship (among the powerful and their subordinates) or attempts at influence. The goal of this research approach was to examine the effects which various leadership styles had on group interaction, productivity, and satisfaction. Frequently, the assumption was made that although leadership could take many styles or forms, there was one ideal style which was most effective across a variety of situations.

Numerous studies have followed the approach of

Lewin et al. (1939) in examining the effects of varieties of leadership style under diverse conditions (see for example, Morse & Reimer, 1956; Rosenfeld & Plax, 1975;





9


Shaw, 1955). Indeed, many of the Lewin et al. (1939) findings have been replicated. Results of leadership style investigations have shown that markedly different patterns of small group interaction and outcome factors emerge as a function of differences in leadership style. Such differences include variations in the levels of group member hostility, aggression, scapegoating and reported satisfaction, as well as various qualitative differences in group productivity resulting as a function of autocratic, democratic, and laissez-faire leadership styles (Lewin, Lippitt, & White, 1939). Other studies have found differences in error rates, the time required for task solution, and group member satisfaction ratings as a function of authoritarian versus non-authoritarian leadership in small groups (Shaw, 1955). Summarizing this research, Shaw (1976) notes that in most investigations, despite the variety of terminology used to describe the leadership style, researchers are dealing essentially with comparisons of directive and non-directive leadership styles. Further, Shaw reports that the results concerning group member reactions to those group leadership styles are quite consistent across the range of situations and groups tested. That is, members of groups with nondirective leaders, as compared to those with directive leaders, tend to react more positively to the group (1976, p. 278). Evidence regarding the group's productivity shows less consistency across situations although it does





10


appear that in general the non-directive led groups are less productive than the directive led groups when productivity differences are obtained (Shaw, 1976, p. 279).

This approach to leadership has and continues to be useful (particularly within applied settings) in terms of delineating certain basic styles of leadership and in finding commonalities across a wide range of situations in terms of style effects on group interaction and outcome variables. Useful as this approach has been and continues to be with regard to identifying the effects of various leadership styles, it has not served to clarify the nature of leadership or identify those factors associated with the perception or emergence of leadership in small group situations. This approach took for granted that certain behaviors were leadership behaviors. Certain of these behaviors were grouped to define a particular leadership style. The focus was then placed on the effects these styles had on outcome variables such as group productivity and satisfaction. No attempt was made to clarify the nature of leadership.

Behavioral approach. The next approach which can be identified in the history of the study of leadership has as its primary goal the identification of leadership behaviors. This particular tradition, if it may be called that, is not as clearly distinguished within the literature as a distinct leadership approach as are the trait and style approaches. The primary intent of this approach








is to determine.which behaviors define leadership, to identify those behaviors (or characteristics of behaviors) which are consistently found in leaders across a variety of situations and which consistently distinguish leaders from non-leaders. The underlying assumption is that there is a set of behaviors which define leadership in the sense that they are exhibited by all leaders, across situations, and are exhibited only by leaders. This approach then is analogous to the earlier trait approach which focused on identification of leader traits. The behavioral approach no doubt arose out of the failure of the trait approach and from the impetus provided by the style approach which had already begun to focus on behavior. Unlike the style approach however, the behavioral approach is directed at clarifying the nature of leadership and at specifying which individuals, by virtue of their behavior, will emerge or be perceived of as leaders.

This approach to leadership rests on the distinct conception of leadership as a set of behaviors engaged in by the leader. The leader is one who behaves like a leader (Fiedler, .1971). Many research theorists have accepted (either theoretically or operationally) this definition of leadership. This behavioral definition is one of the most frequently used definitions. In Carter's (1953) summary of various alternative definitions of leadership, he indicates that he preferred the definition of leader as "one who engages in leadership behaviors."





12


Fiedler (1971) argued that leadership should be defined in terms of acts and behaviors and has used each of the following definitions in his own theoretical and empirical work: the "leader is the person who acts like a leader," the "leader is the individual in the group who has the task of directing and coordinating task-relevant group activities" (p. 2). Hemphill (1952) similarly takes a behavioral definition of leadership: "to lead is to engage in an act which initiates a structure in the interaction of others as part of the process of solving a mutual problem" (cited in Gibb, 1969, pp. 214-215). Leaders are identified as those who engage in leadership behaviors or by the relative frequency with which they engage in such leadership acts. Other researchers using this conception of leadership include those who use Bales' (1951) schema for identifying the leader. Leader behavior is characterized as having high levels of performance output consisting of suggestions, opinions, information, and orienting or integrating statements (Eskilson & Wiley, 1976).

There are, however, serious problems with this type of definition. It in no way approaches a conceptual definition for leadership. It is certainly, an operational definition with obvious empirical usefulness and clarity, but it does not and indeed cannot bring the theoretician any closer to a conceptualization of leadership. At the worst this type of definition runs the risk of circularity "Wlat is a leader? The leader is the person who engages





13


in leadership behavior. What is leadership behavior? It is the behavior of one who leads." At best, this conception of leadership becomes broad to the point of being practically meaningless when leadership is defined in terms of sets or lists of specific leadership behaviors. That is, "this approach leads to a heterogeneous mass of specific acts that supposedly identify leadership in the group. What constitutes leadership depends upon the view of the person who is listing leadership behaviors" (Shaw, 1976, p. 274). Further, it depends upon the behaviors which happen to be exhibited by the particular leader in the particular situation just observed. As Gibb (1969) concludes, "to shift the problem of definition from that of defining the leader to that of defining leader behavior or leadership acts has advantages for particular researchers . but it offers no solution to the definitional problem" (p. 215). In any case, we are left with the problem of deciding what constitutes "leadership behavior" and ultimately this must be decided from within a specified, prior, conceptualization of leadership. The general category must be conceptualized or defined prior to and separately from the identification of the members of that category.

Despite the difficulties of the underlying conceptualization of leadership which this behavioral approach assumes, we may still go on to survey some of the findings this research approach has offered. "The






14


psychological literature contains many analyses which have attempted to describe what it is that leaders actually do" (Gibb, 1969, p. 228), and such studies continue up to the present. Insofar as studies within this tradition attempted to find specific leader defining behaviors, this approach has met with little success. The behavioral approach has largely failed along similar lines as the trait approach. Research again points to the importance of the situation in calling forth different types or sets of leader behavior, especially when the effectiveness of leader behavior is the focus. Leader behavior has been found to vary considerably from situation to situation; behavior which characterized leaders in one task situation does not necessarily characterize leaders in others (Fiedler & Chemers, 1974). As in the search for leader traits, research here indicates the futility of looking for behaviors which would invariably distinguish leaders from followers. In general, the research shows little if any difference in the types of behavior exhibited by leaders and followers, although often there do appear to be real differences between those positions in terms of the frequency with which certain behaviors (such as directing, planning, controlling, interpreting what was taking place, offering ideas, initiating procedures for accomplishing the task, etc.) were performed (Fiedler & Chemers, 1974). Fiedler (1971) concludes that the difference between leaders and followers is considerably less









than we ordinarily assume. Although we can usually identify leaders, there appear to be no specific behaviors in which only leaders engage. Leadership acts are to a greater or lesser degree performed by more group members than just the leader.

Although this approach has not been fruitful in

clarifying the nature of leadership, the approach has been useful in providing lists or descriptions of leader behavior and potential operational definitions of leadership based on those lists. This approach has extended our knowledge of what leader behaviors are or can be, their functions, and allowing us to categorize and dimensionalize leader behaviors while also allowing for the fact that such behaviors will vary across situations and will vary especially in their effectiveness across differing situational contexts. Further, this wealth of useful information comes to us from a wide variety of methodologies and research concerns: from the observational studies which attempted to identify factors of leader behavior (e.g., Carter, 1953, who distinguished three main factors of leader behavior: group goal facilitation, individual performance-prominence, and group sociability) to those attempts to describe leader behavior (e.g., the research done at Ohio State University and a summary of some of that research by Halpin and Winer, 1952, who found four categories which accounted for all leader behavior: consideration, initiating structure, production emphasis and





16


sensitivity or social awareness. The first two were found to account for over 80% of leader behavior). Also included is the work by Bales (1951, 1953) and others who attempted to categorize and delineate leader role behaviors. Research within this tradition may be examined within the review provided by Gibb (1969) and in the work of Bales (1953), Stogdill and Coon (1957), Couch and Carter (1952), among others.

One particularly well-supported and useful conclusion which may be drawn from the research literature in this area is that leader behavior can be conceptualized as having two primary and distinct dimensions: those of consideration and initiation of structure (Gibb, 1969) or the parallel, of socioemotional leadership and task leadership (Bales, 1953). As Fiedler, 1971, states,

there is abundant evidence .. that the consideration and initiation dimensions, or similar factors,
are of overriding importance in most leadership situations. Their identification constitutes one of the
most important achievements of leadership research.
(p. 7.)

The behavioral approach, while it has been useful for providing descriptions of leader behavior and identifying factors common to leadership, remains largely unsatisfactory. This approach does not serve well in terms of clarifying the nature of leadership. If anything, it clouds the issue with its lists and lists of specific leader acts. It fares not much better than the trait approach to the extent that few consistencies in leader





17


behavior are found across situations and few behaviors are found to consistently distinguish leaders from non-leaders. It is suggested here that the major failings of this approach center on the lack of a clear conceptual or theoretical definition of leadership and the misdirected focus of attention on leader behavior instead of on the interaction taking place between leader and led.

Situational approach. One residue from the Lewin et al. (1939) style approach as well as from what this author has called a behavioral approach to leadership is the movement toward viewing the different contexts in which leadership occurs. This ultimately evolved into the situational approach to leadership. This approach took firm hold of the leadership area in the 1950s and continues to some extent into the present (Hollander & Julian, 1969). Research from the earlier trait approach also led to this new tradition as more and more evidence pointed to the importance of situational demands in determining the qualities, characteristics, and skills required of a leader; traits which were positively related to leadership in one situation were often unrelated, if not negatively related, to leadership in another situation (Fiedler, 1971 Fiedler & Chemers, 1974; Shaw, 1976). Thus, this new approach to the study of leadership began with a realization of the importance to leadership of situational variables. The main focus of this situational approach was the study of leaders in different settings where the






18


setting was defined especially in terms of variations in group task and group structure (Hollander & Julian, 1969, p. 387). Variations in setting also included such things as changes in group member characteristics, the environment, and outside influences. The intent was to find "families of situations" within which the leadership role could be seen as relatively consistent (Gibb, 1969). Situational factors were seen to influence not only the specific types of behavior leaders exhibited and how they were valued or reacted to, the types of qualities and skills required of a leader, the effectiveness of the leader, and the perceptions of leadership, but such factors were also found to be influential in the determination of who would emerge as leader (Hollander & Julian, 1969; Shaw, 1976). This new approach to leadership was not so much directed at clarifying the nature of leadership as it was directed at the identification of those situational factors which affected leadership (in the ways described above). That is, this tradition began to take a multifactorial approach to the study of leadership.

One example of an attempt to take a multifactorial, situational approach to the study of leadership can be seen in Fiedler's Contingency Model of Leadership Effectiveness. Research attempts to support and specify this model are many and continue a healthy trend in leadership research today. This particular model has been developed in an attempt to explain and predict leader behavior,





19


primarily leader effectiveness, across situations. The model utilizes factors from both the trait and situational approaches as they interact together. The key personality measure is the LPC score which involves a description of one's least preferred coworker and which is related to a motivation or tendency toward a particular leadership style. The situation is important insofar as it is defined to be favorable or unfavorable to either of the two personality types (high LPC scorers and low LPC scorers). This favorability-unfavorability dimension is determined in terms of three situational factors: (1) the nature of the interpersonal relationship between the leader and the group members, (2) the task structure dimension, and (3) the leader's position power (Fiedler, 1971; Fiedler & Chemers, 1974; Shaw, 1976). Thus, this model of leadership effectiveness suggests that effectiveness (in terms of group performance) is contingent upon both the motivational system of the leader, as measured by the LPC scale, and the situational favorableness (e.g., the degree to which the leader has control and influence in a particular situation).

The relationship between the LPC score and effectiveness as a group leader has been examined in numerous studies and the results from such investigations have indicated that the situation (as defined by the factors mentioned above) is highly important. Fiedler (1967) finds for example that the willingness of group members to






20


be influenced by the leader is conditioned by leader characteristics but that the quality and direction of this influence is contingent on the situational factors defined by group relations and task structure. However, given a specified situation, Fiedler and others have found consistent and significant correlations between the LPC score and type of leader behavior and group performance thus suggesting that LPC is an index of certain behavioral preferences (Fiedler & Chemers, 1974). Generally, the high LPC person has a basic goal to establish good affective ties with the others in the group situation (parallel to what has been termed a socioenotional type of leader) while the low LPC person's primary goal is successful task accomplishment (parallel to a task leader). The low LPC leader is generally more effective in situations which are either highly favorable or unfavorable to the leader, whereas the high LPC leader is generally more effective in situations which are moderately favorable. As Shaw (1976) concludes, although there remain some problems and unanswered questions, Fiedler's model does make a promising beginning toward the integration of leadership styles and situational factors as determinants of group effectiveness (p. 283).

In addition to the work of Fiedler, a great deal of other research has taken this situational or multifactorial approach to the study of leadership. From this body of research the following statements can be made





21


regarding effects of situational variables on leader behavior, leader effectiveness, and leader emergence within small group situations. Leadership performance on one type of task is essentially unrelated to leadership performance on another type of task (Fiedler & Chemers, 1974); thus one cannot really speak of effective or ineffective leaders or leadership styles since effectiveness depends on the situation. In terms of leader emergence, the findings from this research approach "substantially supported the contention that who became a leader depended in some degree upon the nature of the task" (Hollander & Julian, 1969, p. 388). Also, people who are most motivated and visible within the group situation are most likely to emerge as the leader (Fiedler & Chemers, 1974). From a variety of research efforts (e.g., research on reward structures, communication networks, etc.) we see that the increased visibility which often leads to leadership status may be afforded in a variety of ways within the small group situation: holding a central position within a communication network, having a high participation level relative to other group members, having obviously high status which may include being labeled as an expert, having personal attributes which make the individual more visible than other group members, etc. (Bavelas, Hastorf, Gross & Kite, 1965; Fiedler & Chemers, 1974; Gintner & Lindskold. 1975; Riecken, 1958).





22


In summing up the importance of the situation in understanding leadership behavior, Gibb (1969) concluded that "leadership is always relative to the situation" (p. 246) and Fiedler and Chemers (1974) conclude that most differences in the way people act are relatively minor when we consider how much of their behavior is determined by social context (p. 56). "Our data suggest that leader behavior is more strongly determined by the situation than by what the individual would like to do or thinks he ought to do" (Fiedler & Chemers, 1974, p. 96).

Even though this approach was based upon the

recognition that the behavior, effectiveness, and emerqence of leaders were dependent upon the demands of the situation and that the definition of the situation included not only group structure and the nature of the task, but also included the characteristics, perceptions, and behaviors of the group members, "comparatively little attention was directed to the followers, especially in terms of the phenomenon of emergent leadership" (Hollander & Julian, 1969, p. 389). And even though leadership was beginning to be seen as an outcome of a relationship involving the leader, the led, and their shared situation, studies of leadership within this tradition paid little attention to the process aspects of leadership within the group situation. As Hollander and Julian have stated, "the situational view made it appear that the leader and the situation were quite separate" (1969, p. 389). The





23


trait and situational approaches erred in a similar manner i.e., both tried to approach the study of leadership by emphasizing "parts of a process which are by no means separable" (Hollander & Julian, 1969, p. 389). This author would add that the style and behavioral approaches to the study of leadership have erred in a similar manner.

Interactionist approach. The final approach to be discussed here will be called the interactionist approach to leadership. It is not clear that there is a single, primary trend in the area of leadership research today since many of the traditions already discussed continue to enjoy research attention. Additionally, there are other approaches to the study of leadership in the current literature which have not been mentioned such as investigations into leadership training and an entire body of work focusing on leadership effectiveness per se (definitions of effectiveness, measures of, training for greater leader effectiveness, etc.). Further, accompanying the decrease in interest among behavioral scientists within the last decade in small group research, there has been a concomitant decrease in research on small group leadership. Thus, although one cannot say that at present a healthy research trend exists in terms of an interactionist approach to leadership, one certainly could say that much of the preceding research in leadership, coming from the varied traditions mentioned, points to the importance if





24


not the necessity of taking a new, broader, interactionist approach to the study of leadership.

An interactionist approach, at least to this

author, does not imply a single approach to leadership. There appears to be at least two interactionist views. One view is based on the assumption that leadership can best be studied by taking into account the contributions of those factors explored in the previous approaches (e.g., personality traits and attributes of leaders and other group members, group member behavior and leadership styles, and situational variables) as they act alone or interact to affect leadership emergence, leader behavior, leader effectiveness, etc. That is, to study a particular aspect of leadership one would have to view all inputs to the small group situation (at least until a determination was made regarding which inputs were most important to specified aspects of leadership) as they act alone and in concert. This approach may be seen as the logical extension of Fiedler-like approaches whereby instead of looking at the interplay of only two sets of variables, one must look at the interplay of several sets of variables and attempt to determine which interactions are most informative. This constitutes what more properly may be termed a systems or a multifactor approach to leadership. In essence then, interaction here refers to a methodological or statistical interaction among the variety of sets of relevant variables. Other researchers in the field have





25


noted the importance of studying leadership from this perspective and have in fact begun to do such research themselves (e.g., Shaw & Harkey, 1976).

A second approach within an interactionist perspective of leadership, and one which shares many similarities with what Hollander (1978) has recently discussed as a transactional approach to leadership, involves a different methodological focus as well as a new, or elaborated conception of leadership from those approaches previously discussed. Within this approach, in the attempt to study and understand various aspects of leadership (such as emergence, behavior, effectiveness, etc.) focus is directed on the interaction occurring among the group members. Leadership is seen as a type of relationship existing between people which is expressed by the type of interaction taking place. The goals of research coming from this perspective are: (1) to identify the type of interaction patterns that express a perceived leadership situation and (2) to delineate those aspects of the interaction which carry the most information in defining the nature of the leadership relationship (e.g., which aspects of the verbal and nonverbal interactive, communicative behaviors specify leadership relationships). A secondary line of research within this approach is to determine which (and in what manner) personal and situational variables affect interaction variables and-thus indirectly affect leadership relationships.





26


It is this interactionist approach which the

present writer suggests as being the most useful approach to the study of leadership and which provides the perspective for the present investigations. The value of this interactionist approach lies in its most central point: that leadership is a relationship. Therefore, to study leadership one must view the behavior of all individuals (leaders as well as followers) during the process of interaction. This approach places a new emphasis upon the relationships, the interactions, and follower's behavior within the group; all of which have been largely neglected in previous approaches to the study of leadership.

While a research tradition taking this approach is not yet well established by students of leadership, the idea of an interactionist approach to leadership is not at all new. As Gibb (1969) points out, one of the earliest empirical investigations of leadership, done in 1904 by Turman, assumed something of this perspective. -Various studies throughout the leadership literature of the past, especially those using real-life situations, have also pointed out the interactive feature of leadership. For example, Gibb (1969) pointed out that leadership could not rest in the stimulus individuals alone, but necessarily rested as well within the attitudes, beliefs and responses of the others toward the stimulus person (p. 269). In other words, leadership is defined within the interaction by taking into account follower perceptions and behavioral





27


responses to the leader, rather than being defined simply in terms of the behaviors or attributes of the leader. Others who have taken this interactive perspective include role theorists for whom one role is defined in part by its counterpart (Sarbin & Allen, 1968) and who by necessity must look at the role enactment of all group members visa-vis each other. Hollander's (1978) recently elaborated transactional approach similarly emphasizes the relational aspects of leadership. This view of leadership stresses the necessity of studying the transactional process existing among leaders and followers as they exchange benefits and mutually influence one another. "Leadership must include the reactions of followers" (p. 2). Without responsive followers there is no leadership because the concept of leadership is relational. Other contemporary authors (Beckhouse, Tanur, Weiler & Weinstein, 1975; Hollander & Julian, 1969; Lashbrook, 1975) have also pointed to the usefulness of this approach and suggested it as a new perspective for leadership research. Hollander and Julian have pointed to the "greater sensitivity to the social processes of interaction and exchange" within recent leadership research and to the "increasing signs of movement toward a fuller analysis of leadership as a social influence process" (p. 388). They go cn to say that "the tendency now is to attach far greater significance to the interrelationships between the leader, the followers, and the situation" (p. 388). They point out






28


that one consequence of this redirection of focus is the realization that the problem of studying leadership and understanding leadership relationships is much more formidable than was earlier assumed (p. 388).

The arguments in favor of this approach are many

but the central argument suggests that this interactionist approach, of all other approaches mentioned, is best suited to the study of leadership because of the nature of leadership itself. A discussion of the interactionist approach entails a discussion of the conception of leadership.

Leadership as an influence relationship. Although it has been shown earlier that there is little apparent agreement regarding the nature or definition of leadership, certain trends in preference for specific definitions have been noted. Specifically, leadership has been defined in terms of (1) personality traits (within the trait approach), (2) as a set of behaviors engaged in by the leader (within the behavioral approach), and (3) as a type of interaction existing among leaders and their subordinates (Fiedler, 1971), as a set of influence attempts (Hersey & Blanchard, 1976)., as a process of influencing group activities toward goal setting and goal achievement (Stogdill, 1948), and as the exertion of positive (e.g., desired by the leader) influence over the other group members (Shaw, 1976), to name just a few definitions. If one looks closely, it is this third view, leadership as





29


influence, which appears continually throughout the variety of past leadership approaches.

In Carter's (1953) summary of the different definitions of leadership appearing in the literature up to that date, three of the five categories of definitions he distinguished were clearly associated with an influence relationship view of leadership. Carter (1953, pp. 262264) listed the following categories of leadership definitions: (1) The leader is the person who is the focus of group behaviors. The leader receives more communications, has greater influence upon the group's decisions, etc.

(2) The leader is the person who is able to lead the group toward its goals. (3) The leader is defined as the person so named by the members of the group, based on sociometric choice. (4) The leader is defined as the person who has the most demonstrable influence upon group syntality. This refers to Cattell's (1951) theory; the leader is the person who causes syntality change, e.g., a change in the level of group performance, and finally (5) The leader is the person who engages in leadership behaviors. Category three above clearly reflects an operational, not a theoretical, view of leadership. This definition, as Carter points out, merely identifies the occupant of the "leader" position and does not tell one anything about the characteristics of leadership. Further, who is chosen by sociometric choice as leader depends upon what question the researcher asks the group members: e.g., "who do you think





30


was the leader?" "who would you most like to work with again?" "who did you like.the best?" or "who exerted the most influence upon you?" The choice of the question will depend upon the researcher's prior conception of leadership. Thus, this type of definition is only an operationalization of an earlier selected theoretical definition of leadership. Category five above essentially defines the behavioral view of leadership. But in categories one, two and four, it is clear that the central concept is successful influence of the leader over the group. Leadership is either related to the leader's ability to make changes in group member behavior such that group decisions reflect the views of the leader, or the group moves in the direction of goal accomplishment or the level of group performance is altered due to the effect of the leader. While there may be difficulties, as Carter has suggested, with some aspects of each of the above definitions (e.g., the focus of group behavior may be a deviant group member, not the leader; leaders don't always lead groups toward a group defined goal), the influence relationship aspect of leadership appears well accepted by many researchers in the area (Cartwright & Zander, 1968; Fiedler, 1971; Fiedler & Chemers, 1974; Gibb, 1969; Shaw, 1976).

While leadership as influence has obviously been a part of the conception of leadership for some time (Cowley, 1928, stated that the leader was one who succeeded in getting others to follow him) many researchers in the





31


leadership area have only recently begun to acknowledge or emphasize this view. Gibb (1969) concluded his review of the definitions given to leadership over the years with the following comment: "in general, it is an essential feature of the concept of leading that influence is exerted by one individual upon another" (p. 212). Tedeschi and Lindskold (1976) note that "the definition of leadership in terms of influence or power is a relatively recent development" (p. 581). Hollander and Julian (1969) state, "since the beginning of the 1960's there has been a redirection of interest in leadership toward processes such as power and authority relationships" (p. 388). Finally, Fiedler and Chemers (1974) recently suggested that the two most important threads running through nearly all leadership definitions were: (1) that leadership is a relationship between people in which influence and power are unevenly distributed and (2) that there can be no leader in isolation, that followers must either explicitly or implicitly consent to their part in this influence relationship. Thus, leadership implies some sort of exchange between leaders and followers (p. 4). "Above all, leadership is a relationship. More specifically, it is a relationship based on one person's power and influence over others" (Fiedler, 1971, p. 2).

There is thus overwhelming agreement that leadership is an influence relationship and this brings the discussion back to the suggestion of the usefulness of the






32


interactionist approach to the study of leadership. It is the only approach discussed which fccuses on just that aspect of group behavior where influence relationships are visible: the group interaction, the communicative process existing between leader and led. Leadership can not be studied by looking at the leader alone (whether it be in terms of leader traits, behavior or leadership style) nor the situation, if one neglects to look at the behavior of the followers specifically in terms of their responses to leader behavior. It is not at all surprising that other approaches have met with failure with regard to clarifying the nature of leadership. Within an interactionist perspective, one would not necessarily expect leaders to differ from non-leaders (by trait or behavior) or for leaders to behave consistently across different situations Rather, what one would expect, given this view of leadership, is that the relationship between leader and led would remain essentially the same across situations. The responses of non-leaders toward leader behavior would be different from the responses of group members toward nonleader behavior (e.g., the relationships between leader and non-leader would be different from the relationships existing among non-leaders).

This interactionist approach is essentially the same as Hollander's (1978) transactional approach. Hollander argues that this is the approach which leadership researchers should now be taking. Thus, well respected





33


leadership theorists have argued that a new approach should be taken to the study of leadership. Further, failures of previous approaches suggest that new directions to the study of leadership must be taken. And finally, the concept of leadership itself necessitates the choice of an approach which focuses on the group interaction as a whole because it is that interaction which defines the nature or existence of an influence or leadership relationship.

The final section of this chapter will provide a brief survey of studies which have focused on interactional aspects of group behavior. First, with regard to studies of leadership in particular in small group situations, there do not appear to be many studies which have actually been done from the interactionist perspective, and none which have looked at the small group interaction in its entirety as the focus for statements about leadership. There are a few investigations which have pointed to specific aspects of follower behavior in either defining a leadership situation or in leadership emergence. Gibb (1969) cites one study by Krout (1942) who reported the case of an-engineer turned cult leader. Calling himself "The Great I Am," he is said to have marshaled thousands of people into abject worship. Though it is apparent that the man was insane, instead of being labeled a madman he became a leader to large numbers of people. The essential difference between his being labeled a





34


madman, versus a leader, was in the attitudes and responses of the followers toward him. A more recent study similarly points to the role of the follower in terms of emergent leader behavior. Beckhouse et al. (1975) found that leadership behavior was a joint function of situational factors, individual predispositions, and the followers' responsive behaviors. When subjects (previously appointed by the researcher as the group task leader) were pushed to lead by the followers, they led. When they were pushed to follow, they increased behaviors associated with a socioemotional role. Other studies have found that positive and negative reinforcement of the leader's verbal behavior by other group members has the effect of either decreasing or increasing leader behavior (Bavelas, Hastorf, Gross & Kite, 1965; Pepinsky, Hemphill & Shevitz, 1958). Studies which have focused on other aspects of leader and follower behavior include those which have looked at the relative rates of group member participation occurring over the course of the entire interaction. Time and again the emergence or perception of leadership is linked to the individual group member having the highest level of participation (Sorrentino & Boutillier, 1975; Stang, 1973; Stein, Geis & Damarin, 1973). Burke (1974) has linked leadership not just to relative rates of participation but to the control of the communicative interaction in the group. Leaders were more active than followers in taking





35


the speaking floor and they were also given the floor more often than other group members.

Studies in which leadership was not the primary concern, but where attention was focused on the group interaction include the following. Bales' (1951) classic work of interaction process analysis provides a system for categorizing the behavior of all group members vis-a-vis one another, as they interact in a small group situation. This system certainly allows for categorization of leader behavior in isolation but also allows categorization of responses to leader behavior by other group members, thus defining relationships among group members. Bales and Strodtbeck (1951) examined group interactions in terms of an elaboration of distinct phases occurring during group problem solving. Similar work on phases in group development has been done by Bennis and Shepard (1956). Both of these phase development analyses could be re-examined in terms of what they might have to offer regarding leader and follower roles in group decision making and in terms of leader emergence during group development.

Even among communication researchers, research in terms of interaction analysis of communication process is quite recent and is rather sparse. The bulk of this research has occurred only within the last decade (Fisher, 1978). This relatively new approach (called the pragmatic perspective) places emphasis on the ongoing behavioral sequences of interaction which either constrain or serve






36


to define the social situation (Fisher, 1978). In one of the few interactive studies of leadership, Bell (1978) has analyzed sequences of communication behaviors occurring between leader and led in small groups with either taskoriented leaders or relation-oriented leaders. Though not directly related to the study of leadership, communication researchers have developed a variety of category systems (content analysis schemes) for analyzing interactive behaviors, some of which might be seen as being related to leadership. Systems have been developed to look at processes involved in idea development (e.g., Scheidel & Crowell, 1964), decision making (e.g., Fisher, 1970), and conflict management (e.g., Ellis & Fisher, 1975). Category systems have also been developed which allow analysis of various types of relationships created or expressed through interpersonal or group communication. Much of this research (e.g., Rogers & Farace, 1975) focuses on communications which are related to an interactant's attempt to dominate (by means of one-up communication patterns), attempt to be submissive (by using one-down communications), or attempt to be equivalent (by use of one-across communication patterns).

Psychologists as well as communication researchers have recently begun to focus on the nonverbal aspects of interactions. With the exception of Stein's (1975) investigation which found that emergence and perception of leadership in small group settings were conveyed along






37


both verbal and nonverbal channels, few studies deal with leadership per se. Most of the studies which deal with nonverbal behavioral aspects of interaction deal with perceptions of dominance and status. These studies come largely out of the communication field and they may be seen to be related to the concept of leadership insofar as they deal with behavior patterns which are indicative of influence relationships. Perceived dominance has been associated with attempted and successful interruptions (Folger & Sillars, 1977; Henley, 1977), taking relaxed and arms akimbo body postures (Mehrabian, 1969), crowding another's space and pointing (Henley, 1977). High power in group or interpersonal settings has been associated with being able to look less at the other members of the group or interaction. The lesser power individual is presumed to have a greater need to monitor the expressive behavior of the other (the high power individual) in order to gain information regarding the high power individual's reactions and so be able to adjust behavior. Looking also enables the low power person to indicate that she or he is attentive to the higher power person (Exline, 1974). Finally, high power is also associated with having ones gestural behaviors or body postures matched. Persons in agreement and persons of lower power will unconsciously reflect the behaviors of those in power according to Henley (1977)





38


This concludes a review of the leadership literature in terms of the conceptions of and approaches to leadership. A brief review of literature covering the trait, style, behavioral, situational and interactionist approaches to the study of leadership has been provided. The value as well as the problems associated with each of the above approaches have also been discussed. At present it appears that the interactionist approach holds the most promise for future leadership research. The following sections will provide a brief summary and a more specific rationale than that provided in these sections for studies one and two.


Rationale and Statement of Hypotheses for Studies 1 and 2

Study 1. The purpose of Study 1 was to determine to what extent observers of a small group interaction based their perceptions.of group leadership on an influence relationship and to provide an empirical demonstration that perceptions of leadership rest more upon the presence of social influence relationships existing among group member interactions than the mere presentation of what has been previously described as "leader-like" behaviors on the part of one or more group members.

Theorists (Beckhouse, Tanur, Weiler & Weinstein, 1975; Hollander, 1978; Shaw, 1976) in the area of leadership have recently come to suggest that. leadership is properly conceptualized as a social influence relationship






39


among interacting individuals where the behavior of both leader and follower(s) must be taken into account insofar as "leaders" make influence attempts within the interaction and "followers" accept those attempts. Thus, the role of each is in part defined by the role of the other. In this case, the role of leader (offering advice, making suggestions regarding such things as goal choice and procedures for goal accomplishment, offering suggestions for group structure, attempting to control group member behavior, etc.) is fulfilled when followers adopt their role in accepting or agreeing to these influence attempts of the leader. This conception of leadership as influence is in contrast to earlier (though still prevalent) views that leadership could be defined or characterized by referring solely to the behavior of the leader (see Bales, 1951, 1953; Fiedler, 1971; and Hemphill, 1952). That is, according to this earlier conception, leaders are those individuals who engage in such leader-like behavior as: offering ideas and advice, making suggestions, interpreting and offering directions to group behavior, having a high degree of participation and visibility, etc., without regard to any other group member's behavior. In essence, this view is half correct. It does, as well it should, focus some attention on the behavior of leaders. Further, this conception of leadership has been most valuable in providing detailed descriptions of the variety of behaviors in which leaders engage. Valuable as it has been,





40


however, it has not provided a clear, cross-situational view of leadership (Fiedler & Chemers, 1974). It has been plagued by generation of list after list of "leader" behavior since the behavior of leaders can change markedly from one situation to another and from one leader to another. This problem exists because this is an incomplete conception of leadership. It looks at only part of the picture, one-half of the group interaction. The focus is misdirected at the behavior of one individual alone, the leader, where it should be placed on the interaction among leader and follower. Although specific leader behaviors might be expected to change across situations, what does not change across situations is the relationship between group members. That is, the relationship between leader and follower, as defined by the acceptance of followers of leader influence attempts, does not change. Thus, leadership does not change across situations.

Although theoretically sound, the influence conception of leadership has received little empirical attention. One study which has linked influence and leadership is Gibb's (1950) investigation. Gibb formed groups of ten men each and assigned each group a variety of activities to involve them for a series of three, three-hour sessions. Initially, groups were left unstructured or leader-less but they were allowed to select a leader for the second and third sessions. After each session.general sociometric questions were asked of each group member including





41


one question designed to discover each group member's judgment of the behavior influencing the group. Specifically, each member was asked the following question, "some groups are so closely knit that the removal of any one person changes its complexion. For which persons, if any, in this group, would this be the case?" (p. 230). The group member chosen most frequently was seen as being the most influential. Group members were also asked who they would judge to have been leaders in each of the last two sessions (members received no specific definition of leadership). In addition to these group member judgments, independent observers had been instructed to observe each session and code the behaviors of each of the participants. Gibb found that: (1) the participants' responses to the influence question identified the same individuals as those who were behaviorally identified by the outside observers as leaders. The correlation obtained between these two items was .80. (2) Participant and observer judgments of leader also correlated .80. Thus, this is one study which succeeded in making an empirical link between leadership and influence. However, even this study may be considered weak to the extent that the question asked of the participants to determine the most "influential" group member does not necessarily convey that a true influence relationship existed among the participants. That is, selecting the individual group member "whose removal from the group would most change the






42


complexion of the group" does not provide anything but a

very vague notion of influence. There do not appear to be

any other studies in the leadership literature which

specifically attempt to demonstrate that observers (or

group members) rely on influence-relationship information

in making their judgments of perceived group leadership.

The following propositions and hypotheses were

offered for study one.

P1: Leadership is an interactional or relational
phenomena.

P2: The type of relationship is defined by the type of
interaction occurring among group members.

P3: The perception of leadership is dependent upon the
observation of the behavior of both "leaders" and
"followers" independently and especially, vis-avis each other (that is, in interaction).

P4: Leadership is defined by the existence of a particular kind of interaction or relationship among
group members; that being an influence relationship.

P5: The perception of leadership is dependent upon the
observation of "leader" behaviors directed at or
in the presence of other individuals which have an observable, perceived causal (influential) effect
on the behavior of the others (the "followers").

From the propositions stated above, the following

hypotheses were derived.

Hl: Leadership within a group will be perceived when
an influence relationship exists within the group.

Specifically, hypothesis one predicts that observers' perceptions of leadership will be
strongly affected by the presence of an influence
relationship within the group. That is: there
will be a main effect for amount of influence on the leadership ratings for the designated leader
(group member three) such that high levels of influence are associated with high leadership






43


ratings and low levels of influence are associated
with lower leadership ratings.

Proposition three above proposes that both leader behavior and follower behavior are important dimensions in the perception of leadership. This is derived from the body of previous leadership literature (both theoretical and empirical) which has demonstrated the importance of the leader's behavior in and of itself for the perception of leadership in small group interactions. Thus, hypothesis two is proposed to take into account the role of leader alone for the perception of leadership.

H2: The perception of leadership within a group varies
with the perception of leader-like behaviors by at
least one group member occurring within the group
interaction.

Specifically, hypothesis 2 predicts that: there will be a main effect for amount of leader-like
behavior on the leadership ratings for the designated leader (group member three) such that high
levels of leader-like behaviors are associated with high leadership ratings and low levels of
leader-like behaviors are associated with lower
leadership ratings.

However, in a determination of the relative importance of influence versus leader-like behavior for the perception of leadership, clearly, it is the entire interaction among group members conveying the influence relationship between leader and the followers which forms the basis for leadership perception. Thus, hypothesis three states:

H3: The perception of the influence type of interaction among group members is more highly associated
with perception of leadership than the perception of leader-like behaviors by one or more group members within the group interaction. That is, while
both influence and leader-like behavior





44


(non-influential) are associated with the concept of leadership, perceptions of leadership are more
strongly associated with the presence of an influence relationship than the presence of leaderlike behavior which has no influence effect on either the interaction or on other group member
behavior.

Specifically, hypothesis 3 predicts that: the
amount of influence main effect factor will accourt
for a greater percent of the variance than the
leader-like behavior main effect factor.

As has already been stated, both amount of influence and

amount of leader-like behavior affect the perception of

leadership. A main effect for each has already been

predicted (see hypotheses one and two). Above and beyond

these effects, it is expected that in combination, high

levels of these two factors will produce an even stronger

perception of leadership than any of the factors acting

alone. Thus, hypothesis four states that:

H4: There will be an interaction between amount of
influence and amount of leader-like behavior on ratings of leadership for the designated leader
such that the condition containing high influence and high leader-like behavior will produce higher
leadership ratings than in any other condition.

Specifically, the expected ordering of the means for leadership ratings for the designated leader (group member three) will be as follows: lowest
mean for cell one (low influence, low leader-like
behavior) followed by cell three (low influence,
high leader-like behavior), cell two (high influence, low leader-like behavior), with cell four
(high influence, high leader-like behavior) receiving the highest mean leadership rating.

Each of the preceding tests of hypotheses which

examine the leadership ratings of group member three in

determination of effects due to level of influence and

level of leader-like behavior is based on the assumption





45


that it is group member three (the.designated leader) who is in fact chosen as group leader. Thus, hypothesis five is a test of that basic assumption.

H5: The group member designated as leader (group
member three) will receive significantly higher leadership ratings than either of the other two
group members. This effect will be found for each of the cells in which there is a designated leader;
specifically, in cells two, three and four.

Specifically, hypothesis five predicts that a main
effect for group member within cells will be obtained such that group member three receives
higher leadership ratings than group members one
and two in cells two, three, and four.

Study 2. The intention behind both studies one and two was to provide empirical support for the conception of leadership as an influence relationship existing among group members and expressed by group members'. interactive behaviors. By definition, influence can only be examined within the context of a relationship or interaction among individuals. However, the elements of influence and interactive behavior may be stressed separately and this is what studies one and two purport to do. That is, study one emphasizes that leadership exists and will be perceived within the small group interaction when one individual has greater influence over the other group members. Leadership is defined primarily as an influence relationship existing among interacting individuals (although not all influence relationships constitute leadership). Study one attempts to provide empirical support for the proposition that leadership is properly conceptualized as an influence relationship.





46


Study 2, on the other hand, takes influence as a given for the perception of leadership, but attempts to examine more specifically the means by which influence, and hence leadership, is conveyed. The leader may engage in many different types of behavior which have the result of eliciting or causing a change in a variety of follower behavior. In this case, the influence conveying behaviors are classified into "types" according to the communicative channel along which they occur: verbal or nonverbal. Thus,' the purpose of study 2 was to demonstrate that influence, and hence leadership, may be conveyed both verbally and nonverbally within the group interaction. Study 2 attempts to provide leadership researchers with more detailed information than they presently have regarding how the influence relationship is actually conveyed to observers of small group interactions. Specifically, this study deals with the following question: to what extent does each communicative channel (verbal and nonverbal) convey influence relationships necessary for the perception of leadership.

The research direction taken in this study follows recently developing trends in small group and leadership research. These trends include the increasing attention directed to interactive processes such as communicative interactions (rather than interactive outcomes such as measures of group goal accomplishment and group satisfaction) and the increasing attention given to the






47


multichannel nature of human communication (that is focusing on verbal as well as nonverbal communication in various interactive contexts). It is communication researchers for the most part (e.g., Burke, 1974; Gouran, 1973; Lashbrook, 1975; Spillman et al., 1978) who are presently focusing attention on the ongoing, communicative process variables of small group interaction and emergent leadership. However, the classic work of Robert Bales (1951) in terms of content analyses schemes for interaction process is the foundation for much of this research. An example of work in this area includes Bell's (1978) investigation in which communicative sequences between leaders and followers were analyzed. Bell found a pattern within groups with relation-oriented leaders (but no identifiable patterns within groups with task-oriented leaders) such that leader's "tension release" was followed by nonleader member's "tension release." While it does not appear that interaction process analysis of leadership has received much attention in the empirical work of social psychologists, such a focus of attention has been encouraged in recent theoretical reviews of the leadership area (e.g., Hollander & Julian, 1969).

Focusing attention on the interactive process as well as the multichannel nature of human interactions, communication researchers and recently, social psychologists have demonstrated that people can and do communicate a variety of information alone different channels. That





48


is, information is carried by the verbal channel in terms of the content of what is said. Information is also carried by the nonverbal channel which includes vocalics (involving the tone of voice, amount of speech, intonation patterns, speech volume and speech rate, interruptions, etc. Vocalics includes all those features of the voice other than the words themselves) and kinesics (which includes facial expressions, the type and amount of eye contact, head nods and shakes, hand gestural activity, body position orientation and shifts, etc.). Although the importance of verbal behavior is well recognized, this has not always been so with regard to nonverbal behavior, particularly in regard to certain sub-topic areas such as leadership within the context of a small group interaction. Researchers, especially in social psychology, have only just begun to examine the ways in which nonverbal (especially vocalic and kinesic) channels convey information and the types of information such channels are capable of conveying. However, we do know that each of these channels is important in communicative behavior: people both send and receive messages on them (Burgoon & Saine, 1978; Weitz, 1974). Some theorists have even argued that the majority of the information (e.g., over 65%) we send and receive is conveyed on the nonverbal channels, as opposed to the verbal-content channel (Birdwhistell, cited in Knapp, 1972, p. 12).






49


A few investigators within the area of leadership have begun to look at nonverbal behavior and in particular, how leadership is conveyed by the various channels. For example, Stein (1975) has demonstrated that subject-observers were able to accurately identify emergent leadership within a group when provided with the following types of information: (a) full channel information as provided by the presentation of audio-video tapes of small group interaction, (b) verbal content and relative percent participation information provided by written transcripts of the group interaction, (c) kinesic channel information only as provided by the video-only (no-sound) condition, and (d) a filtered speech condition accompanied by the visual channel which provided kinesic channel information in addition to some vocalic information, but no verbal content information, to the observer subjects. Stein concluded that his study provided the first direct evidence that nonverbal behaviors are useful in perceiving leaders. Although Stein did demonstrate that observers can make accurate (accuracy was defined in terms of agreement with the judgments or choices of leadership made by the actual group members) judgments of leadership based upon information conveyed along verbal and nonverbal channels, he does not demonstrate that it is in fact "leadership" that is being dealt with since he gives no evidence of an influence-relationship. The study does not demonstrate what it is that is conveyed along these two





50


channels. His methodology does not provide the reader with information regarding whatspecific behaviors or variations in behavior are occurring along each channel and between "leaders" and "followers." "Leader," in Stein's study, is that person chosen as leader.

Study 2 provides a demonstration of the importance of both verbal and nonverbal channels for the perception of an influence-leadership relationship. In addition, it provides information regarding the specific nature of the behaviors being conveyed along both verbal and nonverbal channels.

The following propositions and hypotheses were offered for Study 2.

P1: Leadership is an interactional or relational
phenomena (See P1 for Study 1).

P2: Leadership is defined in part by the existence of
a particular kind of interaction or relationship
among group members, that being an influence
relationship (See P4 for Study 1).

P3: Leadership (the existence of an influence relationship) is conveyed along verbal as well as
nonverbal communicative channels within the small
group interaction.

Following from these propositions, the following hypotheses were derived.

HI: The group member designated as leader (group
member three) will receive higher leadership
ratings than either of the other two group members regardless of whether the observer subjects
receive information along verbal, nonverbal, or a
combination of verbal plus nonverbal channels.

Specifically, hypothesis one predicts that a main
effect for group member within each experimental
condition will be obtained such that group member








three receives significantly higher leadership
ratings than group members one and two.

This hypothesis follows from proposition three above that

influence relationships and thus leadership, will be

conveyed through both verbal and nonverbal communicative

channels within group interaction. The following hypothesis is proposed following the logic that the more influence conveyed, the stronger the perception of leadership.

H2: A cumulative effect is hypothesized such that
leadership ratings for the designated leader will
be higher for conditions in which the influence
relationship is conveyed along two rather than one
communicative channel.

Specifically, hypothesis two predicts that there
will be a main effect for condition on leadership
ratings for the designated leader (group member
three) such that condition one (in which influence
is conveyed along both verbal and nonverbal channels through audio and visual means) is significantly different than any other experimental
condition (where influence is conveyed by either verbal or nonverbal means but not both). Group
member three will receive significantly higher
mean leadership ratings in condition one than in
conditions two, three or four.















CHAPTER II


METHOD


Study One

Subjects. Subjects were 130 students enrolled in basic speech courses at Cleveland State University and basic psychology courses at John Carroll University. Participation in the experiment was on a credit given basis.

Procedure. Upon arrival at an experimental lab, subjects were assigned in pairs, on a random basis, to experimental testing rooms. As they arrived in the experimental testing room each subject took a seat at a table. The seats were divided by a curtain partition to reduce interaction and eye contact among the students. Subjects were then told that the experimenter was interested in their reactions to a videotaped small group interaction. Subjects.were told that they would observe on a videotape TV monitor a small group interaction which had occurred among three fellow students. Following observation of the interaction, they were told that they would be asked to respond to a brief questionnaire regarding what they had just observed. Subjects then read and filled out informed consent forms.


52






53


Subjects were then requested to pay careful attention to the TV monitor and to the experimenter's instructions. They were also requested not to interact with each other in any way nor to make any audible sounds (aside from requests for clarification of instructions from the experimenter) so as not to interfere with the other subject's observation of the group interaction. Following these instructions, subjects were exposed to a no-sound videotaped still shot of group member number one and they were requested to fill out a series of scales by McCroskey and McCain (1974) designed to measure physical and task attraction. These scales were to be used as covariate measures. After each subject had filled out these scales, the same procedure was employed for group members two and three respectively. Subjects were instructed to fill out the scales based on the still shot information in front of them and they were instructed to rate each individual independently without attempting to compare the group members to each other. Subjects were then told that they were about to observe an interaction which took place among the three students whose still shots they had just viewed. They were requested to pay careful attention to the interaction and they were again reminded not to interact with each other in any way. At this time, subjects were exposed to one of four versions (comprising the four experimental conditions) of a videotaped interaction among group members one, two and three.





54


Subsequent to exposure to the videotaped group

interaction, subjects were administered the post-experimental measures. Following completion of the post-experimental questionnaire, all participants were thanked for their participation and told that they would receive a handout in their classes within two to three weeks informing them of the purposes and results of the study. Following completion of data collection of the study, all students received a written explanation of the purposes, hypotheses and results of the study in which they participated.

Design. Subjects were randomly assigned to one of four experimental conditions (N = 130). In all four conditions subjects observed a videotaped interaction by means of a videotape recorder and television monitor. Within each of the four conditions, level of influence was manipulated on two levels (low and high) and amount of leader-like behavior was manipulated on two levels (low and high). This constitutes a 2 x 2 factorial design. All four experimental conditions employed the same three actors as group members (Cleveland State University theater majors) and across each condition, actor and role assignment remained constant. Thus, actors one and two always played follower roles in all conditions. Actor three always played the leader in the three experimental conditions where a leader was designated.





55


Holding actor and role assignment constant across conditions can provide problems in interpretation of results such that one is unsure whether obtained effects are due to the experimental manipulations (e.g., levels of influence or leader-like behavior) or the result of differences due to uncontrolled-for actor variables (e.g., actor three being perceived as more or less leader-like than actors one and two) based on physical appearance characteristics. This problem may be eliminated in either of two ways. In terms of design, actors could be rotated through the roles such that each actor played each role (or played at least one other role) for each of the experimental conditions (thus blocking across actors). This procedure was deemed impractical due to the time and effort involved in videotaping, the difficulty the actors had in learning their roles, and the number of subjects which would then be required. An alternative procedure involves devising measures for those possible effects due to uncontrolled-for actor variables (these should be primarily due to physical appearance since nearly all behaviors were controlled for by the scripting) and covarying out these effects from the effects due to experimental manipulations. This statistical solution to possible effects due to actor differences was deemed most practical for the present investigations.

Thus, a 2 (low vs. high levels of influence) x 2 (low vs. high levels of leader-like behavior) factorial






56


design with two covariates (measures of physical and task attraction for each actor based on physical appearance alone) was used to analyze for differences due to levels of influence and leader-like behavior. A summary of the design and conditions may be seen in Table 1.

Group interaction and videotape construction. A three member group interaction in which members were attempting to offer solutions for the parking problem at Cleveland State University was used as the stimulus material for this study. This topic was chosen because it involved a problem with which nearly all college students were familiar (both Cleveland State students and John Carroll students had moderate to severe student parking problems).

The verbal script for the small group interaction was carefully constructed through the procedures described below. Two volunteer groups of three individuals each (who did not participate in the later experimental phase of the studies) were asked to participate in a small group interaction. They were asked to offer solutions for the parking problem at Cleveland State University. They were told that the experimenter was interested in the behavior of small group, problem-solving groups and that their entire interaction would be tape recorded. Both groups then discussed the parking problem and were tape recorded. After twenty to thirty minutes the experimenter stopped the discussion, turned off the tape recorder and thanked






57


Table 1

Experimental Conditions for Study One


Level of Leader-Like Behavior Low High Low Condition 1 Condition 3 Level of Influence
High Condition 2 Condition 4





58


the group members for their participation. They were then told that their interaction would be used as the basis for scripts for actors who would be portraying members in a small group interaction for a later study.

Each of the recorded small group interactions was then transcribed. These transcripts of naturally-interacting groups then provided the basis for a working script to be used for the creation of the scripts for each of the four experimental conditions. The working script for a three member group interaction was developed by selecting various verbatim passages (sometimes a single speaker utterance, sometimes a series of interchanges among two or more speakers) from the transcripts. Passages were selected for use when they offered reasonably clear suggestions for the solution of the parking problem or procedural directions to the group, when suggestions were followed by active group discussion, when passages offered reasonably clear responses (either positive or negative) to suggestions made. On occasion, words or whole speaking turns were inserted in order to produce a better flow of conversation. There were more than enough possible passages to select from the two, twenty-minute conversation transcripts. Once a draft of the working script was composed, it was read through and timed, shortened or lengthened, until the resulting interaction lasted from six to eight minutes. A six to eight minute interaction seemed most appropriate for the experimental conditions






59


since it would last long enough for a good deal of group interaction but would not be so long as to lose the attention of the subjects.

This method of developing a basic working script had the advantages of producing a more naturalistic group interaction than the experimenter, alone, could create. Natural verbal content was obtained by using the ideas and words which actual group members had used. Such a method produced more naturalistic phrasing (e.g., the frequently non-grammatical, non-fluent speech) of students engaged in an informal speech situation.

It was from this basic working script that scripts for each of the four experimental conditions were developed. It was intended that scripts for each condition be as nearly identical as possible, within the constraints of the experimental manipulations, in terms of general content and overall length of the interaction. This intention was carried through.

In creating the script for each experimental condition, the verbal behavior of the interactants was dealt with first. Not all aspects of verbal content are necessarily related to leader behavior or to the establishment of an influence relationship among group members in a task oriented group. Those aspects of verbal content which were considered. important (based upon findings from the behavioral approach to leadership) and which operationalized leader-like behavior included the following: (a)






60


comments which give suggestions or directions to others (implying autonomy for others), (b) comments which give opinion, evaluation, analysis or which express feelings or a wish, and (c) comments which give orientation, information, confirmation or clarification. Many of these have been listed by various leadership researchers as being leader-like behavior (Bales, 1951; Carter et al., 1968; Cartwright & Zander, 1968; Eskilson & Wiley, 1976; Morris & Hackman, 1969). Specifically, these are descriptions of behavior which Bales (1951) has categorized in his interaction process analysis content analysis scheme as small group interaction, task oriented behaviors. Those comments listed above in a belong to Bales' category number four, those in b belong to Bales' category five, and those in c belong to Bales' category six. See Appendix A for a complete list of the Bales categories.

Those verbal content behaviors or interactions

which are important to the establishment of an influence relationship are those which signify agreement with or compliance to the expressed suggestions, directions, analyses or feelings of another. Thus, verbal content indicative of an influence relationship was operationalized in terms of a group member's use of Bales' category number three: "agrees, shows passive acceptance, understands, concurs or complies" in response to another member's suggestion, opinion, etc. As Hare (1976) has pointed out, low power individuals including followers,






61


generally pay more attention to and respond more favorably to persons of high power, which includes leaders. An absence of utterances coded as Bales' category three or the presence of utterances coded as belonging to Bales' category ten (disagrees, shows passive rejection, formality, withholds help) or Bales' category twelve (shows antagonism, deflates another's status, defends or asserts self) by a group member in response to another group member's suggestion, opinion, etc., was indicative of a lack of an influence relationship between the group members. Thus, the system used to operationalize verbal content as being leader-like or indicative of an influence relationship was the content analysis system provided by Bales (1951). Specifically, utterances coded as belonging to categories four, five, and six defined leader-like behavior and utterances coded as belonging to categories three, ten and twelve defined the presence or the absence of an influence relationship. The level of leader-like behavior and the level of influence was dependent upon the number of utterances for each group member coded as belonging to these categories.

The Bales' content analysis system has a total of twelve categories of which only six were used systematically in the operationalization of factors in the present investigation. Categories seven, eight and nine all deal with asking questions (seeks information, asks for opinion, asks for suggestion or direction, etc.). In every






62


condition an attempt was made to equalize across all three group members the number of utterances they each had which were coded as belonging to Bales' categories seven, eight or nine. Thus, each group member, in each experimental condition, asked an approximately equal number of questions. This was done because it is not at all clear from the leadership literature who it is, leaders or followers, that asks the most questions. Leaders may demand answers from followers, but followers, respecting the ability of the leader, frequently seek information, the guidance and opinion of their leaders. Bales' categories one (shows solidarity, raises others' status, gives help or reward), two (shows tension release, jokes, laughs, shows satisfaction) and eleven (shows tension, asks for help, withdraws out of field) were not dealt with systematically in the present investigation.

Thus, operationalization by means of verbal content of level of leader-like behavior involved the number of utterances each group member had which were coded according to Bales' interaction process analysis as belonging to categories four, five or six. Operationalization by means of verbal content of level of influence involved the number of utterances belonging to Bales' category three (and the lack of Bales' categories ten and twelve) which each group member received from other group members following their utterance of a Bales' four, five or six type.






63


Nonverbal vocalic leader-like behavior was operationalized along two dimensions: (1) level of participation and (2) attempts to interrupt the speech of another group member in order to gain control of the floor. Each of these dimensions has been linked to leadership. Time and again, researchers have shown that leaders have higher rates of participation than non-leaders and that those individuals with high rates of participation are perceived as being leaders (Knutson & Holdridge, 1975; Morris & Hackman, 1969; Sorrentino & Boutillier, 1975; Stang, 1973). Fiedler and Chemers (1974) have shown that less dominating or lower status individuals are less likely to interrupt or attempt to interrupt the speech of others who are more dominant or are of higher status. Henley (1977) points out that dominant individuals or those in positions of authority may show their dominance or authority by interrupting subordinates. In the present investigation, level of participation for each group member was systematically varied in terms of the number of speaking turns for each group member. A speaking turn has been defined by N. Markel as beginning when "one interlocutor starts solo talking and ends when a different interlocutor starts solo talking" (1975, p. 190). Level of participation was also operationalized in terms of the total number of words spoken by each group member. Attempts to interrupt the speech of another were operationalized as a period of solo speaking by one group member during which another group






64


member began simultaneous speech which was not successful in gaining the conversational floor.

The nonverbal vocalic operationalization of level of influence was in terms of the number of successful interruptions achieved. That is, an individual was influential, along the nonverbal vocalic channel, when he was able to take control of the floor away from another group member by interrupting that group member's speaking turn. Henley (1977) has stated that "a hierarchy of power in a group could be plotted by ordering people according to the number of successful interruptions they achieve" (p. 69).

Nonverbal kinesic channel behaviors were also

dealt with in creating scripts for group member behavior for each experimental condition. Again, only those kinesic behaviors which have been previously linked to leadership were used. Leader-like kinesic behaviors in the present investigation included attempts by one group member to indicate who the next speaker should be by means of hand gesture. It has been described as part of leader behavior to attempt to control or structure the interaction of the group members. Deciding who shall speak when is one aspect of this type of behavior. According to Henley (1977) authority is shown in controlling others from a distance. A dominant person may direct others by gesture and "point at them in a way that will shut them






65


up, stop other action, or evoke attention and submissiveness" (p. 128) and presumably, direct them to speak as well.

Kinesic behaviors which were indicative of an

influence relationship included the following: (1) success at indicating the next speaker by means of gesture. That is, when the person pointed to or gestured toward was in fact the person who spoke next, the turn indicating gesture was deemed influential or successful. (2) Having a gesture matched by at least one other group member or (3) having a body posture shift matched by at least one other group member was also indicative of an influence relationship. Research has indicated that high status or powerful individuals are more likely to be matched by lower status others in terms of both their gestural activity and their body posture (Burgoon & Saine, 1978). The individual is influential if she or he is successful in affecting some of the kinesic behavior of the other group members. A final behavior indicative of an influence relationship in the present investigations was the receipt of a head nod indicating agreement for an utterance of Bales' type four, five or six. Thus, agreement may be sent along the verbal content channel as described earlier or along the nonverbal kinesic channel. The presence of an influence relationship is also indicated by the lack of kinesic head behavior (head shakes) which show disagreement.





66


See Table 2 for a summary of the verbal and nonverbal behaviors used to operationalize leader-like behavior for each of the experimental investigations. See Table 3 for a summary of the verbal and nonverbal behaviors used to operationalize influence for each experimental condition.

The importance of each of these verbal and nonverbal behaviors in the creation of each experimental condition was not in the sheer number of each behavior occurring during the group interaction for any given individual, but in the number performed by each group member relative to the number performed by each other group member. That is, a low level of leader-like behavior was operationalized in terms of there being nearly equal numbers of leader-like behavior (verbal and nonverbal) performed by each of the three group members. A high level of leader-like behavior was operationalized in terms of one group member (always group member three, the designated leader) performing a greater number of leaderlike behaviors (verbal and nonverbal) than either of the other two group members (who engaged in nearly equal numbers of those behaviors when compared to each other). A low level of influence was operationalized in terms of there being nearly equal levels of influence (in terms of the effect any one group member's behavior had on any other group member's behavior) among all three group members. Finally, a high level of influence was






67


Table 2

Summary of the Verbal and Nonverbal Behaviors
Used to Operationalize Leader-Like Behavior Factor


Channel Behavior

Verbal 1. Number of utterances per group member which Content give suggestion, direction, implying autonomy for others (Bales' category four).

2. Number of utterances per member which give
opinion, evaluation, analysis, or express feeling or a wish (Bales' category five).

3. Number of utterances per group member which
give orientation, information, repeats,
clarifies or confirms (Bales' category six).

Nonverbal 4. Number of speaking turns for each group Vocalic member.

5. Number of words (over the entire conversation) spoken by each group member.

6. Number of attempts by each group member to
take the control of the floor away from
another speaker for himself (an interruption attempt).

Nonverbal 7. Number of attempts by each group member to Kinesic indicate the next speaker by means of a
hand gesture.






68


Table 3

Summary of Verbal and Nonverbal Behaviors Used to Operationalize Influence Factor


Channel Behavior

Verbal 1. Number of verbal agreements (indications of Content agreement, passive acceptance, understanding, concurrence, compliance) received by
each group member following their own
utterance of a suggestion, opinion, etc.
(e.g. reception of a Bales' category three
from another group member for a Bales'
category four, five, or six utterance).
The lack of verbal disagreements (Bales'
category ten and twelve) for an utterance
coded as Bales' four, five or six.

Nonverbal 2. Number of successful attempts by each group Vocalic member to interrupt, to take the control of
the floor away from another speaker.

Nonverbal 3. Number of successful attempts by each group Kinesic member to indicate the next speaker by
means of a hand gesture.

4. Number of body posture shifts by each group
member which are matched by one or more
other group members.

5. Number of hand gestures by each group
member which are matched by one or more
other group members.

6. Number of nonverbal agreements (head nods)
received by each group member from other
group members following an utterance coded as Bales' four, five, or six. The lack of
a nonverbal disagreement (head shake) for an utterance coded as a Bales' four, five
or six.





69


operationalized in terms of one group member (always group member three, the designated leader) having more influence over the other two group members (numbers one and two) than either of them have over other group members.

Thus, scripts for each experimental condition were created by taking passages from the working script, but making systematic alterations in the specific verbal or nonverbal behavior, or making changes with regard to which group member engaged in a specific verbal or nonverbal behavior, such that the experimental conditions were created. Thus, condition one interaction (low levels of leader-like behavior and low levels of influence) was scripted such that group members one, two and three each engaged in nearly equal numbers of leader-like behavior and each had nearly equal influence over other group members. This condition then had no designated leader.

The interaction for condition two was scripted

such that all group members engaged in nearly equal numbers of leader-like behavior (low levels of leader-like behavior) but group member three had more (nearly two times more) influence over the other group members, through the scripted verbal and nonverbal behaviors, than group members one and two had (high levels of influence). The condition three interaction (high levels of leaderlike behavior, low levels of influence) was scripted such that group member three engaged in more (nearly two times more) leader-like behavior than did group members one and





70


two. However, all three group members were nearly equal in the influence they exerted over other group members. Finally, the interaction for condition four (high levels of influence and high levels of leader-like behavior) was scripted such that group member three exerted greater influence (nearly two times more) over other group members than either of them exerted and group member three also engaged in more (nearly two times more) leader-like behaviors than did members one and two. See Tables 4 through 7 for a summary of the amounts of leader-like and influence behaviors obtained for each group member for conditions one through four respectively.

Thus, in the creation of each script for the experimental conditions, the script was written and repeatedly edited until group members one, two and three each had the condition-appropriate leader-like behaviors and level of influence relative to each other (along both verbal and nonverbal channels). For reliability, a trained coder as well as the experimenter coded each turn or utterance (on occasion there was more than one utterance or thought within a turn) for each group member, for each condition, according to the Bales system. All instances of disagreement over the categorization of utterances or turns were resolved for the final script. Thus the script represented a high degree of agreement regarding Bales category assignments.






71


Table 4

Summary of Frequencies and Percentages for Behaviors
Indicative of Level of Leader-Like Behavior and
Level of Influence for Group Members One,
Two and Three for Condition One
(low levels of leader-like behavior,
low levels of influence)


Group Member

Factor Behavior 1 2 3

Leader-Like Number of utterances coded Behavior as Bales 4,5,6 26 23 22

% of each group member's
utterances coded as Bales
4,5,6 62% 62% 63%

Number of speaking turns 25 22 23

% of speaking turns for
each group member out of the total number of turns
by all 36% 31% 33%

Number of words spoken 422 439 430

% of words spoken by each
group member out of the
total number of words
spoken by all 33% 34% 33%

Number of interruption
attempts 5 4 4

% of interruption attempts for each group member out of the total number of attempts made by all 38% 31% 31%

Number of turn indicating
gestures 2 2 2

% turn indicating gestures
for each group member out of
the total number of such
gestures by all 33% 33% 33%






72


Table 4--continued.


Group Member

Factor Behavior 1 2 3

Influence Number of Bales' category 3
(agreements) received in
response to making a Bales
4,5,6 utterance 7 6 8

% of agreement (Bales 3)
received when Bales 4,5,6
utterances were made 27% 26% 36%

Number of successful
interruptions 2 2 3

% successful interruptions
for each group member out of
the total number of interruption attempts 40% 50% 75%

Number of head nods (agreements) received in response
to making a Bales 4,5,6
utterance 3 5 8

% of agreements (head nods)
received when Bales 4,5,6
utterances were made 12% 22% 36%

Number of hand gestures
matched 2 1 1

% hand gestures matched for
each group member

Number of body posture
shifts matched 2 2 1

% body posture shifts
matched for each group
member 100% 67% .33%

Number of successes at
indicating next speaker turn
by gesture 1 1 1

% of successes for each individual for indicating next
speaker turn by gesture 50% 50% 50%






73


Table 5

Summary of Frequencies and Percentages for Behaviors
Indicative of Level of Leader-Like Behavior and
Level of Influence for Group Members One,
Two and Three for Condition Two
(low levels of leader-like behavior,
high levels of influence)


Group Member

Factor Behavior 1 2 3

Leader-Like Number of utterances coded Behavior as Bales 4,5,6 24 24 26

% of each group member's utterances coded as Bales
4,5,6 57% 61% 68%

Number of speaking turns 25 21 24

% of speaking turns for each group member out of the total number of turns
by all 36% 30% 34%

Number of words spoken 448 453 442

% of words spoken by each
group member out of the
total number of words
spoken by all 33% 34% 33%

Number of interruption
attempts 5 4 6

% of interruption attempts for each group member out of the total number of attempts made by all 33% 27% 40%

Number of turn indicating
gestures 3 3 3

% turn indicating gestures
for each group member out of
the total number of such
gestures by all 33% 33% 33%






74


Table 5--continued.


Group Member

Factor Behavior 1 2 3

Influence Number of Bales' category 3
(agreements) received in
response to making a Bales
4,5,6 utterance 5 7 15

% of agreement (Bales 3)
received when Bales 4,5,6
utterances were made 21% 29% 58%

Number of successful
interruptions 2 2 5

% successful interruptions
for each group member out of
the total number of interruption attempts 40% 50% 83%

Number of head nods (acreements) received in response
to making a Bales 4,5,6
utterance 4 4 13

% of agreements (head nods)
received when Bales 4,5,6
utterances were made 17% 17% 50%

Number of hand gestures
matched 0 0 2

% hand gestures matched for
each group member

Number of body posture
shifts matched 1 1 3

% body posture shifts
matched for each group
member 50% 33% 100%

Number of successes at
indicating next speaker turn
by gesture 1 1 2

% of successes for each individual for indicating next
speaker turn by gesture 33% 33% 67%






75


Table 6

Summary of Frequencies and Percentages for Behaviors
Indicative of Level of Leader-Like Behavior and
Level of Influence for Group Members One,
Two and Three for Condition Three
(high levels of leader-like behavior,
low levels of influence)


Group Member

Factor Behavior 1 2 3

Leader-Like Number of utterances coded Behavior as Bales 4,5,6 21 14 36

% of each group member's utterances coded as Bales
4,5,6 52% 41% 71%

Number of speaking turns 24 20 28

% of speaking turns for each group member out of the total number of turns
by all 33% 28% 39%

Number of words spoken 323 312 626

% of words spoken by each
group member out of the
total number of words
spoken by all 26% 25% 50%

Number of interruption
attempts 6 4 8

% of interruption attempts for each group member out of the total number of attempts made by all 33% 22% 44%

Number of turn indicating
gestures 0 0 3

% turn indicating gestures
for each group member out of
the total number of such
gestures by all 0% 0% 100%





76


Table 6--continued.


Group Member

Factor Behavior 1 2 3

Influence Number of Bales' category 3
(agreements) received in
response to making a Bales
4,5,6 utterance 10 8 6

% of agreement (Bales 3)
received when Bales 4,5,6
utterances were made 48% 57% 17%

Number of successful
interruptions 5 2 4

% successful interruptions
for each group member out of
the total number of interruption attempts 83% 50% 50%

Number of head nods (agreements) received in response
to making a Bales 4,5,6
utterance 4 5 4

% of agreements (head nods)
received when Bales 4,5,6
utterances were made 19% 36% 11%

Number of hand gestures
matched 1 2 1

% hand gestures matched for
each group member

Number of body posture
shifts matched 1 1 1

% body posture shifts
matched for each group
member 100% 100% 20%

Number of successes at
indicating next speaker turn
by gesture 0 0 0

% of successes for each individual for indicating next
speaker turn by gesture 0% 0% 0%





77


Table 7

Summary of Frequencies and Percentages for Behaviors
Indicative of Level of Leader-Like Behavior and
Level of Influence for Group Members One,
Two and Three for Condition Four
(high levels of leader-like behavior,
high levels of influence)


Group Member

Factor Behavior 1 2 3

Leader-Like Number of utterances coded Behavior as Bales 4,5,6 16 18 33

% of each group member's utterances coded as Bales
4,5,6 43% 69% 70%

Number of speaking turns 25 16 24

% of speaking turns for each group member out of the total number of turns
by all 38% 25% 37%

Number of words spoken 330 294 605

% of words spoken by each
group member out of the
total number of words
spoken by all 27% 24% 49%

Number of interruption
attempts 3 1 8

% of interruption attempts for each group member out
of the total number of attempts made by all 25% 8% 67%

Number of turn indicating
gestures 0 2 4

% turn indicating gestures
for each group member out of
the total number of such
gestures by all 0% 33% 67%






78


Table 7--continued.


Group Member

Factor Behavior 1 2 3

Influence Number of Bales' category 3
(agreements) received in
response to making a Bales
4,5,6 utterance 5 4 16

% of agreement (Bales 3)
received when Bales 4,5,6
utterances were made 31% 22% 48%

Number of successful
interruptions 2 0 6

% successful interruptions
for each group member out of
the total number of interruption attempts 66% 0% 75%

Number of head nods (agreements) received in response
to making a Bales 4,5,6
utterance 6 1 13

% of agreements (head nods)
received when Bales 4,5,6
utterances were made 37% 5% 39%

Number of hand gestures
matched 0 0 4

% hand gestures matched for
each group member

Number of body posture
shifts matched 0 0 4

% body posture shifts
matched for each group
member 0% 0% 100%

Number of successes at
indicating next speaker turn
by gesture 0 1 3

% of successes for each individual for indicating next
speaker turn by gesture 0% 50% 75%





79


After the four scripts were completed, student actors from the theater department at Cleveland State University were recruited. Initially attempts were made to use male actors who resembled each other along several physical characteristics. However, the three actors who were finally used in creation of the four interactions differed a great deal from each other in physical appearance. However, all three were experienced actors. The actors were given a script and a few weeks of rehearsal time in order to learn the scripted verbal and nonverbal behaviors. They were told to learn the interaction exactly as scripted. Once the actors appeared to have an acceptable level of performance in rehearsal each interaction was videotaped in the campus television studio.

Technical aspects of videotape production such as camera position and shots, lighting and audio levels were held constant within and across recordings. Each interaction was filmed against a backdrop of paneling and bookcases. The three actors maintained the same seating positions across all conditions and they were arranged in chairs in a semi-circular fashion around a small round coffee table. Thus, all three were visible in nearly face-on positions. All actors dressed similarly (in a casual manner) and each actor wore the same clothing across all tapings.

Each interaction began with a fade from black and one actor saying, "The problem is parking." For the first






80


thirty seconds of the interaction numbers were imposed on the tape to appear at the bottom of the television monitor screen identifying actors as group members one, two and three. The interaction proceeded until its completion about seven minutes later. Approximately fifteen seconds before the end of the interaction, the numbers were imposed on the tape to reappear on the television screen, again identifying each actor by his number; one, two or three. This was done to aid subjects in identifying each actor.

Imposed on the videotape immediately prior to the start of the group interaction was a series of three still shots, one of each actor (sitting alone with a neutral expression), labeled by the appropriate number at the bottom of the screen. These shots were used for the covariate measures (judgments of task and physical attraction based on physical appearance).

Three to four takes of each condition's interaction were videotaped. Following completion of taping, the experimenter carefully reviewed each take to determine which one best fit the original script. Interactions never conformed exactly to the script. Actors found these roles very difficult to perform as precisely as the experimenter expected and inevitably some scripted behaviors were left out and other, non-scripted behaviors were included. In the process of selecting the best take for each experimental condition, a new, revised transcript was





81


written (scripting what was actually said as opposed to the original plan of what was to be said) and counts of all relevant behavior (turns, words, interruptions, gestures, body shifts, head nods and counts of Bales' categories) for each actor were remade. Tables 4, 5, 6, and 7 reflect the behaviors which actually occurred for each videotaped interaction (for conditions one, two, three and four, respectively) as opposed to what was originally scripted. On the whole, deviations from the original scripts were relatively minor.

At this point, each interaction was pre-tested to determine how natural of an interaction it appeared to be. Pre-test subjects (N = 59) were basic speech and psycho.logy students at Cleveland State University and John Carroll University. Approximately fifteen students each watched one of the four conditions. They were given a three item scale (see Appendix B) measuring perceived naturalness of the interaction. The scale scores could range from a low of three (very unnatural) to a high of twenty-one (very natural). Mean ratings indicated each condition was perceived to be moderately natural (X for condition one = 9.8; X for condition two = 9.1; X for condition three = 11.0; and X for condition four = 9.4). Thus, these conditions fell within an acceptable range of naturalness for use in the present investigations.

Because of the rather large differences in physical appearance among the three actor, group members and






82


the impracticality of controlling the differences due to actor appearances in the design, a second pre-test was performed. This was done in order to confirm that a covariate analysis (covarying out effects due to differences in physical appearance of the actors from any effects due to the experimental factors) was necessary. Thus, another pool of pre-test subjects (N = 58) who were basic speech and psychology students at Cleveland State University and John Carroll University were presented with a series of scales designed to measure each of the following: (1) the extent to which the actor was perceived to be credible based on his physical appearance, (2) the extent to which the actor was judged attractive on social, task and physical dimensions based on his physical appearance, and (3) the extent to which the actor was perceived as being leader-like again based on his physical appearance alone.

Five dimensions of source credibility (competence, character, composure, extroversion, and sociability) were measured by a series of seven-interval, semantic differential type scales developed by McCroskey, Jensen, and Valencia (1973). The measures of task, social, and physical attraction used in the pre-test were developed by McCroskey and McCain (1974). These were a series of twelve (four per dimension) likert type, seven-interval scales. Three items written by the experimenter measured perceived leadership. Two items were seven-interval, likert type scales and the third used a seven-interval






83


semantic differential scale bounded by follower and leader. See copies of all pre-test scales for actor appearance differences in Appendix C.

The following procedures for the pre-test were used. Subjects were shown a still shot of one of the three group members on a videotape television screen. While the picture remained on the screen, pre-test subjects were given the three scale questionnaire packet and instructed to respond.to each item based on the information given them on the television screen.

A one way analysis of variance testing for differences due to actor appearance on each scale dimension was performed. Results indicated that differences between the actor-group members were significant on all dimensions of each of the three scales. See Table 8 for a summary of the analysis of variance results. Thus, differences found in the pre-test analysis suggested that a covariate analysis would have to be performed to control for effects due to differences in the physical appearance of the actors.

Covariate measures. The covariate scales used in the present study were the task and physical attraction scales developed by McCroskey and McCain (1974). These scales were the most theoretically relevant as covariates to the present study since differences in pre-test data were the result of physical appearance differences among the actors. The physical attraction scale most directly measures reactions to physical appearance differences.






84


Table 8

ANOVA Summary for Differences Between Actors
Based on Physical Appearance on Pre-test Scales


Sum of
Factor Dimension Squares df F

Credibility sociability 359.34 2 10.68**
character 84.88 2 3.50*
extroversion 344.33 2 8.98**
composure 185.21 2 4.22*
competence 292.55 2 12.44**


Attraction physical 478.22 2 14.78**
task 179.11 2 9.83** social 177.49 2 5.69**


Leadership item one 16.66 2 4.80*
item two 49.05 2 12.63** item three 27.04 2 11.16**


*p<.05 **<.01






85


The task attraction scale most directly deals with how an actor might be expected to behave in task situations, again based on physical appearance. The task dimension was relevant in this study since the group was a task oriented small group. These covariates were expected to equalize effects due to differences in physical appearance among the actors in the analyses. See Appendix D for a copy of the covariate measures.

Dependent measures. A packet of post-experimental measures was used to assess subjects' reactions to the videotaped group interaction. Subjects' reactions toward the group interaction in terms of their ratings of perceived group leadership were measured by a one item, seven-interval, likert type scale bounded by the descriptions: none/almost no leadership and extremely great amount of leadership. Subjects were requested to rate each of the three group members on leadership on the three identical scales.

Also included in the post-experimental package was a measure of how confident the subject felt regarding each of the leadership ratings they had just made (e.g., how confident the subject felt about their leadership rating for group member one, number two and number three). Thus, following the leadership rating scale for each group member, a one item, five-interval scale measuring confidence was inserted. This likert-type scale was bounded by the descriptions: not at all confident and very confident.





86


In addition to items measuring perceived leadership and the confidence associated with the judgment of leadership, an item measuring perceived influence was included. This item required the subject to rate each group member in terms of how much influence he appeared to have on either of the other group members during the course of the interaction. Thus, three rating scales, one for each group member, following a likert type format with a seven-interval scale was used. The scale was bounded by none/almost no influence and extremely great amount of influence. As with the leadership rating scales, subjects were requested to indicate how confident they felt about each influence rating they made. Confidence was measured here in the same manner as described above.

Two other rating scale items were included in the post-experimental package. These were rating scales for sociability (by means of a five-interval, likert type scale bounded by the descriptions: very little sociability and extremely great amount of sociability) and for pleasantness (again by means of a five-intervaL likert type scale bounded by the descriptions of not at all pleasant and very pleasant). These items were included as extra scales in which to embed the leadership and influence items.

Following the rating scales for leadership, influence, confidence, sociability and pleasantness, subjects were requested to indicate the group member or





87


members to whom they had given the highest leadership rating. Subjects were then presented with a thirty-two item adjective check list and were requested to check the fifteen items which were most central to their leadership rating decisions for the individual most highly rated on leadership. Following this, subjects were instructed to rank order the five most important adjectives of those fifteen just checked. The purpose of these two measures was to determine whether subjects would select adjectives descriptive of individual traits, individual behaviors, or characteristics of the interaction as being the most important in their leadership judgments. These measures would also be used to determine whether different adjective descriptions were selected for different conditions.

The last page of the post-experimental packet

contained a free response question asking the subject's opinion of the purposes of the study and a question asking the subject if they knew any of the three group members, and if so, in what way. Questions regarding the subject's sex, age, race, school of attendance, and whether they were native speakers of English were also asked. See Appendix E for a copy of the post-experimental packet. Study Two

Subjects. Subjects were 264 students enrolled in basic speech and psychology courses at Cleveland State University and John Carroll University. Participation was on a credit given basis.






88


Procedure. The procedure for study two was nearly identical to that described for study one. Subjects were told that they would be exposed to a small group interaction and afterwards they would fill out a questionnaire concerning their reactions to that interaction. Subjects then read and filled out informed consent forms.

Subjects were then exposed to the covariate stimuli and were requested to fill out the covariate scales (the task and physical attraction scales described in study one). This was the case for all subjects except those who were assigned to the transcript-only condition. These subjects received no covariate stimuli (since they would only be reading a transcript of the group interaction rather than viewing or hearing the interaction). The covariate scales were filled out by these subjects as part of the post-experimental packet (exposing them to the same questionnaires as all other subjects).

All subjects were then exposed to some form of the group interaction. Afterwards, subjects received the post-experimental measures. Following completion of the post-experimental questionnaire participants were thanked for their participation and told that they would be informed of the purposes and results of the study at a later date. All subjects received a full explanation of the study at the completion of data collection.

Design. Subjects were randomly assigned to one of eight experimental conditions (N = 264) comprising a 2




Full Text

PAGE 1

TOWARD A MODEL OF LEADERSHIP PERCEPTION IN SMALL GROUP INTERACTIONS BY PATRICIA ANN BURRIS A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 19 81

PAGE 2

Copyright 1981 by Patricia Ann Burris fc i ii gii Mm'

PAGE 3

DEDICATION To Gill Woodall my dearest friend, my life partner. You share every joy and every pain of my life. You bring meaning to it all. Thank you for sharing with me the toil and frustration of this dissertation and the joy of its completion. Now let's get on to better things together „ ir^|.'.r-n>l^,.. '--*n
PAGE 4

ACKNOWLEDGMENTS A dissertation is rarely the sole effort of one individual. Others provide inspiration, ideas, criticisms, and emotional support during the course of doctoral and dissertation work. I therefore wish to offer thanks to the following individuals. To Dr. Marvin Shaw v;ho chaired this research, my deep gratitude. Thank you for your patience and continual support throughout this effort. You were the perfect chairman. My respect for you and your scholarship has increased with each year that I have known you. It has been a pleasure to be your student and I hope that relationship continues for many years to come. To Dr. Norman Markel, special thanks to you for initiating my interest in comjnunicative behaviors as the founding father of the Gainesville Society for the Study of Coverbal Behavior. Thank you for your contributions to my education as a communication researcher and equally, for the political education and friendship you have given me. To Dr. Tom Simon, very special thanks for your insightful questions and suggestions, for your unending enthusiasm and support. Thank you for your friendship and especially, for the model you provide of an academician with broad interests and as one. who takes positions and is XV --,- -.ifiiu^r"(-*j*.*-s ta^)wivijCTiN'>i

PAGE 5

willing to fight for them. To Dr. Bob Ziller for your creative and original work and to Dr. Patricia Miller for so kindly agreeing to serve on my committee at the last moment I wish to thank my hard-working and under-paid actors. Bill Beck, Matt Cupach and Paul Hanley. If effort and talent have any reward, you will all have successful futures. Thanks also to my highly capable and dependable research assistants, Linda Helman and P'irk McDonnel. You are both well on your way to becoming excellent social scientists. My gratitude to my subjects who generously gave m.e an hour of their time. Special thanks to Sandra Coy, my typist, for her excellent and speedy work and her patience. Finally, for those who contributed to my happiness, sanity, fulfillment, and sense of purpose beyond the narrow bounds of this dissertation, which after all was only a relatively small part of my life these past several years, my gratitude and love. Thank you, my parents. Kiddle, dear Cindy Gallois, my brothers and sisters of the I Gainesville Society for the Study of Coverbal Behavior and '• the Graduate Student Union, my Senseis and fellow karateka of Coung-Nhu Karate and Dragon Gate Do jo, and to my husband. Gill Woodall, who is part of everything I am and do. .siwi*wi=Ti?Sr=!TP?'*ria-r-r-^**uii-. .v-i.'J*i1T i ir r^h -rrTiiift?t r i rTd?fri

PAGE 6

TABLE OF CONTENTS Page ACKNOWLEDGMENTS iv LIST OF TABLES viii LIST OF FIGURES x ABSTRACT ........... xi CHAPTER I INTRODUCTION 1 Conceptions of and Approaches to the Study of Leadership 3 Trait Approach 3 Style Approach 5 Behavioral Approach 10 Situational Approach 17 Interactionist Approach ......... 23 Leadership as an Influence Relationship ..... 2S Rationale and Statement of Hypotheses for Studies One and Two ........ 38 Study One 38 Study Two 4 5 II J-ISTHOD 52 Study One 52 Subjects 52 Procedure 52 Design 54 Group Interaction and Videotape Construction 56 Covariate xMeasures 83 Dependent Measures 85 Study Two 87 Subjects 87 Procedure 8 8 Design „ 8 8 Vi

PAGE 7

Chapter Page Mode of Presentation-CoiTiin.unicative Channels 89 Dependent and Covariate Measures • 92 III RESULTS 93 Study One 9 3 Test of Hypotheses for Study One .... 93 Supplementary Analyses for Study One 105 Study Two 112 Test of Hypotheses for Study Tv7o .... 112 Supplementary Analyses for Study Two 122 IV DISCUSSION ..... 129 Summary of Results for Study One ..... 129 Summary of Results for Study Two ..... 138 Recomjnendations for Future Research 144 APPENDICES A Bales Content Analysis Categories 150 B Pre-Test Scale for Perceived Naturalness of Interaction ....... 152 C Pre-Test Scales for Perceived Actor Credibility, Interpersonal Attraction and Leadership 154 D Covariate Scales ... 159 E Post-Experimental Measures ., 162 F Study Tv70 Condition Four Transcript .... 173 (Low Influence, Low Leader-Like Behavior Condition) G Study Two Condition Four Transcript .... 180 (High Influence, Low Leader-Like Behavior Condition) H Verbatim Verbal and nonverbal Scripts From Videotaped Interactions 187 BIBLIOGRAPHY .................... 216 BIOGRAPHICAL SKETCH .. 222 V3.1

PAGE 8

LIST OF TABLES Table Page 1. Experimental Conditions for Study One .... 57 2. Suiranary of the Verbal and Nonverbal Behaviors Used to Operationalize LeaderLike Behavior Factor 67 3. Suininary of Verbal and Nonverbal Behaviors Used to Operationalize Influence Factor 6 8 4. Summary of Frequencies and Percentages for Behaviors Indicative of Level of LeaderLike Behavior and Level of Influence for Group MeiTibers One, Two and Three for Condition One 71 5. Suminary of Frequencies and Percentages for Behaviors Indicative of Level of LeaderLike Behavior and Level of Influence for Group Members One, Tv/o and Three for Condition Two 73 6. Summ.ary of Frequencies and Percenteiges for Behaviors Indicative of Level of LeaderLike Behavior and Level of Influence for Group Members One, Two and Three for Condition Three „...,.,.. 75 7. Summary of Frequencies and Percentages for Behaviors Indicative of Level of LeaderLike Behavior and Level of Influence for Group Members One, Two and Three for Condition Four ........ 77 8. ANOVA Summary for Differences Between Actors Based on Physical Appearance on Pre-test Scales 84 9. Experimental Conditions for Study Two .... 90 10. ANOVA and ANCOVA Summ.ary for Leadership Ratincis 94 VI 11

PAGE 9

Table Page 11. Summary of Adjusted Means for Leadership Ratings = 95 12. ANCOVA Summary for Leadership Ratings for Group Members Within Cells Two, Three and Four 100 13. Summary of Adjusted Means for Leadership Ratings for Group Members VJithin Cells Two, Three and Four 102 14. ANCOVA SumJTiary for Leadership Ratings for Group Members Within Cell One 103 15. Summary of Adjusted Means for Leadership Ratings for Group Members Within Cell One 104 16. ANOVA and ANCOVA Summary for Influence Ratings 107 17. Summary of Adjusted Means for Influence jr Rarings .... ........ 108 18. ANOVA and A1<1C0VA Summary for Confidence Ratings 110 19. Summ.ary of Adjusted Means for Confidence Ratings Ill 20. MJOVA Summiary for Leadership Ratings for Group Members Within Conditions ...... 116 21. Summary of Means for Leadership Ratings for Group Members for Each Condition .... 117 22. A.NOVA Summary for Leadership Ratings for Group Member Three 118 23. Summary of Means for Leadership Ratings for Group Member Three ...... 120 24. ANOVA Summary for Influence Ratings for Group MemjDer Three 124 25. Summary of Means for Influence Ratings for GrouTJ MemJjer Three 125 26. ANOVA SuiraTiary for Confidence Ratings 127 27. Summary of Means for Confidence Ratings ... 128 XK i->'t7||).-<;^ny -E>-'.^^>ra^fSi^s: air^-:if r* s= wll.^;^^t^^ ^— ,.>6j;-^^, t i *>.Mir.fk5>f*Vi'=;ii""'^'r'

PAGE 10

LIST OF FIGURES Figure Page 1. Interaction of Influence and Channel for Leadership Ratings 121 2. Interaction of Influence and Channel for Influence Ratings ........ 126 X fifei^.^^v*v^fiT,p:r**,r?m(fr9rwa>v

PAGE 11

Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy TOWARD A MODEL OF LEADERSHIP PERCEPTION IN SMALL GROUP INTEPJ=iCTIONS By Patricia Ann Burris August 19 81 Chairman : Dr Marvin E Shaw Major Department: Psychology In contrast to trait, style, and behavioral views of leadership, recent theorists suggest leadership is more properly conceptualized as an influence, relational phenomenon. The purpose of the present studies was to provide support for the view that perception of small group leadership is associated with the presence of an influence relationship existing among group m.erribers and that this relationship is conveyed through verbal and nonverbal channels For study one, it was predicted that strongest perceptions of leadership occur where levels of influence are high (one group member has high degrees of influence over other group members) rather than low (m.embers are equally influential). Further, perception of leadership is more strongly associated with influence levels than levels of leaderlike behavior (one group member engages in low or high levels of those behaviors previously listed xi

PAGE 12

in the literature as typical of leaders) For study two, it was predicted that influence relationships, thus leadership, v/ould be conveyed along verbal and nonverbal coiriinunicative channels. Study one employed a two (high vs. low levels of influence) by two (high vs. low levels of leader-like behavior) design. Three actors were videotaped engaging in scripted group interactions. Experimental conditions were created by varying verbal and nonverbal behaviors of each actor such that either all group m.embers were equally influential and engaged in equal numbers of leader-like behaviors, or one group member was either more influential than other group members, engaged in m.cre leader-like behaviors, or was both more influential and engaged in m.ore leader-like behavior than other group members. Study two experimental conditions were created by presenting audio-visual, visual-only, audio-only, or written transcript-only information of the interactions videotaped for study oneResults supported the hypotheses Perception of leadership was related to high levels of influence and was more strongly associated with the influence factor than leader-like behavior factor. Influence relationships were conveyed across verbal and nonverbal channels, with exception of the audio-only channel. Results emphasize the importance of viewing leadership as an influence relation existing among group members and conveyed through verbal and nonverbal behaviors, xii

PAGE 13

CHAPTER I INTRODUCTION The purpose of each of the present investigations is to provide empirical support for recently acknov/ledged views that (1) leadership is a relational phenomena (leadership can exist only as part of an interpersonal relationship. Leadership is an attribute of a relationship among people, not an attribute of a single individual) In particular, leadership is properly conceptualized as an influence relationship, and (2) that the most appropriate approach to the study of leadership is one which focuses on the types of interaction patterns occurring among group members. This approach is called the interactionist approach. Each study provides information regarding the types of group interactions leading to the perception of leadership. Study one demonstrates that interaction patterns which are indicative of influence relationships existing am.ong group members are strongly associated with the perception of leadership within the group. Study two demonstrates that group interactions occurring along verbal as well as nonverbal channels convey influence and hence, leadership information. 4 — |-Vif aiii.VSBIi'O—w*^ — iA*.—

PAGE 14

Interest in leadership in small groups has been long standing and at tiiTi.es, intense. Certainly this reflects the importance of small group membership and leadership roles in our lives. Vie interact daily in a variety of small group situations (within the family, in informal social and community organizations, and in miore formal business and political groups) Leadership is an important aspect of nearly all of these group memberships. As Hollander (1978) states, "almost any task related to an organized activity involves leadership, or at least is associated with it. There is nothing so central to the functioning of groups or organizations, whether in government, industry, or any other place in society" (p. 3) Others have noted that of all other concepts in small group process and small group structure, leadership has received the most empirical and theoretical attention (Cartwright & Zander, 1968; Lashbrook, 1975; Shaw, 1976). The follov/ing sections of this chapter will provide a review of some of the theoretical and empirical findings which have resulted from^ years of work in this area. The purpose of this survey is twofold: (1) to provide som-e review of the knov/ledge gains made in this area over the years, and (2) to point out some of the failings of this work in an effort to provide a rationale for the theoretical underpinnings of the present investigations: that leadership is a relational phenomenon and in particular an influence relationship, and that the m.ost

PAGE 15

useful approach to the study of leadership is an interactionist approach. It is suggested here that some of the failings of earlier work center around the inappropriate choice of an approach to the study of leadership (e.g., the trait, style, behavioral, and situational approaches) Further, these approaches were selected either due to the lack of clear agreement regarding the definition of leadership or due to an inaccurate conception of the nature of leadership. With regard to this definitional problem, several researchers have noted that there has been little agreement regarding the nature of leadership (Gibb, 1969; Shaw, 1976) Bass (1960) recorded over 130 definitions prior to 1949. More recently, others have noted that there are practically as many definitions of leadership as there are leadership theories and nearly as many of those as there are psychologists working in the field (Fiedler, 1971, p. 1) The investigations reported here should serve to clarify the nature of leadership. Conc eptions of and Approaches to the S tudy of Leadership Trait approach At least among psychologists, the earliest and most actively pursued approach to the study of leadership involved the search for leadership traits. This approach encompassed hundreds of studies in which leadership theorists, largely in the early part of the century, attempted to identify those personality or other individual attributes which accounted for the leadership .r^*^t ^i^enta-r*..

PAGE 16

of certain individuals (for discussions of this approach see Cartwright & Zander, 1968; Gibb, 1969; Hollander S Julian, 1969) This approach to the study of leadership carried with it an underlying conception of what leadership was. The ability to behave as a leader was assumed to be based upon some underlying, unitary attribute (trait) or pattern of attributes (traits) within the individual. Leadership was conceptualized as an aspect of personality: a quality, characteristic, or predisposition which accounted for why some individuals rose to positions of leadership, which was presumed to distinguish leaders from non-leaders, and which was held in common by all leaders across variations in time and situation (Fiedler & Chemers, 1974; Shaw, 1976) However, the results of the trait approach have been disappointing. Within the leadership area few consistencies across investigations were found. When consistencies in the traits associated with leadership were found (e.g., traits which were found across a variety of leaders and leadership situations, traits vrhich appeared to distinguish leaders in general from non-leaders) the correlations obtained most often tended to be low (Cartwright & Zander, 1958; Gibb, 1969; Hollander & Julian, 1969; Mann, 1959; Shaw, 1976; and Stogdill, 1948). Included among those traits found to correlate at least to some extent, v/ith leadership were the following: in general the leadership status was more often tlian not ;^i:^^;^£^v<^W^Ukih..-si4^>*>..;^.*MAilirM

PAGE 17

associated with superior intelligence scholarship, insight and knowledge of how to get things done. In contrast to non-leaders, leaders tended to be more original, they had greater verbal fluency, were more persistent, more self-confident, more responsible, dependable, and cooperative. They were more aggressive, more lively, and had greater social participation and a greater desire to excel (Stogdill, 1948) Other reviews of this literature m.ay be found in Bass (1960), Gibb (1969), and Mann (1959). Overall, the view that leadership may be defined in terms of personality traits, that there is one basic personality pattern for leaders regardless of the leader or the type of leadership situation, is not well supported. Further, the approach to the study of leadership which involves the search for and measurement of specified personality traits does not appear to be useful. Both the approach and the underlying viev; of leadership have been largely abandoned by current social scientists, Numicrcus studies within this tradition failed altogether to find any consistent pattern of traits which characterized leaders (Gibb, 1969) Those classes of traits which were found to be associated with leaders (1) did not distinguish leaders from followers to any great degree and (2) had low consistency across situations. However, as is true of many investigative approaches, there were a few gains with this approach in terms of our understanding of leadership. When the situation is taken into account.

PAGE 18

certain general classes of traits do appear to relate, at least to some extent, to leadershipGroup members' personalities do make a difference to group performance and would thus be expected to affect that aspect of group behavior to which the leadership concept applies (Gibb, 1969, p. 227). Following his review of the trait approach to leadership, Gibb concluded that a person does not become a leader by virtue of his possession of any one particular pattern of personality traits, but the pattern of personal characteristics of the leader must bear relevant relationship to the present characterisrics activities, and goals of the group of which he is a leader. (p. 22 5.) In conclusion, despite the fact that personality traits may play some role in terms of group behavior, including leader behavior, the trait approach does not appear to be a useful one for understanding the nature of leadership. Style approach The next approach to the study of leadership involved the identification and study of styles of leadership (such as^ autocratic, democratic, laissezfaire; participatory, supervisory; authoritarian, nonauthoritarian; nondirective and directive) and the nature of their effects on groups. This tradition is prim.arily exemplified by the classic work of Lewin, Lippitt and Wliite (19 39) and Lippitt and White (19 43) This approach to leadership involved a change in focus in two different ways from the earlier trait approach. First, this approach directed attention away from, those individual trait characteristics which predisposed given individuals to

PAGE 19

leadership and focused attention instead on the types of motivational and behavioral tendencies associated with different types of leaders or styles of leadership. Thus, this approach acknowledged differences in leaders and focused on these differences in terms of both the internal motives and the behavioral tendencies of leaders. Second, attention was focused for the first time on the effects of leadership behavior on group interaction and outcome variables such as measures of group productivity and satisfaction. One problem within this tradition has been the confusion with regard to the term "leadership style." According to Fiedler (1971) some investigators have used the term leadership style to refer to the leader's typical way of behaving. For example, Hersey and Blanchard (1976) define leadership style in terms of the consistent pattern of behavior exhibited when one attempts to influence the activities of others. Attempts ro influence may be typically task-oriented or relationship-oriented behavior (or some combination of both) Such consistency in behavior is developed over time, eventually coming to be known as the leader's style or leadership personality. This type of definition defines leadership style in term.s of the consistencies in the behavior of the leader. Other investigators (Fiedler, 1971, for example) define leadership style in terms of both leader and follov/er behavior and in terms of behavior as well as the motives underlying

PAGE 20

that bahavior. Thus, leadership style refers to relatively consistent systems of interacting with others in subordinate positions (Fiedler, 1971, p. 10). Leadership style refers to a consistent pattern of behavioral interaction among both leader and non-leaders. Further, Fiedler relates these consistencies of interaction to the motivational system of the individual leader (that is, individuals may tend to be more relationship motivated or more task motivated) While there remains some disagreement regarding the term "leadership style," this approach to the study of leadership was not specifically aimed at clarifying the nature of leadership itself. Rather, the assumption ap, pears to have been made that leadership, regardless of the particular leadership style em.ployed, involved either an influence relationship (am.ong the powerful and their subordinates) or attempts at influence. The goal of this research approach was to examine the effects which various leadership styles had on group interaction, productivity, and satisfaction. Frequently, the assumption was m.ade that although leadership could take many styles or forms, there was one ideal style which was most effective across a variety of situations. Numerous studies have follov/ed the approach of Lewin et al. (19 39) in examining the effects of varieties of leadership style under diverse conditions (see for example, Morse & Reimer, 1955; Rosenfeld & Plax, 1975;

PAGE 21

Shaw, 1955) Indeed, luany of the Lewin et al (1939) findings have been replicated. Results of leadership style investigations have shown that markedly different patterns of small group interaction and outcom.e factors emerge as a function of differences in leadership style. Such differences include variations in the levels of group member hostility, aggression, scapegoating and reported satisfaction, as well as various qualitative differences in group productivity resulting as a function of autocratic, democratic, and laissez-faire leadership styles (Lewin, Lippitt, & White, 1939) Other studies have found differences in error rates, the time required for task, solution, and group member satisfaction ratings as a function of authoritarian versus non-authoritarian leadership in sm.all groups (Shaw, 1955) Summarizing this research, Shaw (1976) notes that in most investigations, despite the variety of terminology used to describe the leadership style, researchers are dealing essentially with comparisons of directive and non-directive leadership styles. Further, Shaw reports that the results concerning group member reactions to those group leadership styles are quite consistent across the range of situations and groups tested. That is, m.erabers of groups with nondirective leaders as compared to those with directive leaders, tend to react more positively to the group (197 6, p. 278). Evidence regarding the group's productivity shows less consistency across situations although it does

PAGE 22

10 appear that in general the non-directive led groups are less productive than the directive led groups when produc-; tivity differences are obtained (Shaw, 1976, p. 279). This approach to leadership has and continues to be useful (particularly within applied settings) in terms of delineating certain basic styles of leadership and in finding coirancnalities across a v/ide range of situations in terms of style effects on group interaction and outcome variables. Useful as this approach has been and continues to be v/ith regard to identifying the effects of various leadership styles, it has not served to clarify the nature of leadership or identify those factors associated with the perception or emergence of leadership in small group situations This approach took for granted that certain behaviors were leadership behaviors. Certain of these behaviors were grouped to define a particular leadership style. The focus was then placed on the effects these styles had on outcome variables such as group productivity and satisfaction. No attempt v/as mads to clarify the nature of leadership. B ehavioral approach The next approach which can be identified in the history of the study of leadership has as its primary goal the identification of leadership behaviors. This particular tradition, if it may be called that, is not as clearly distinguished within the literature as a distinct leadership approach as are the trait and style approaches. The primary intent of this approach

PAGE 23

11 is to determine which behaviors define leadership, to identify those behaviors (or characteristics of behaviors) which are consistently found in leaders across a variety of situations and which consistently distinguish leaders from nonleaders. The underlying assumption is that there is a set of behaviors which define leadership in the sense that they are exhibited by all leaders, across situations, and are exhibited only by leaders. This approach then is analogous to the earlier trait approach which focused on identification of leader traits. The behavioral approach no doubt arose out of the failure of the trait approach and from the impetus provided by the style approach v;hich had already begun to focus on behaviorUnlike the style approach however, the behavioral approach is directed at clarifying the nature of leadership and at specifying which individuals, by virtue of their behavior, will emerge or be perceived of as leaders. This approach to leadership rests on the distinct conception of leadership as a set of behaviors engaged in by the leader. The leader is one who behaves like a leader (Fiedler, 1971) Many research theorists have accepted (either theoretically or operationally) this definition of leadership. This behavioral definition is one of the most frequently used definitions. In Carter's (1953) summ.ary of various alternative definitions of leadership, he indicates that he preferred the definition of leader as "one who engages in leadership behaviors."

PAGE 24

12 Fiedler (1971) argued that leadership should be defined in terms of acts and behaviors and has used each of the following definitions in his own theoretical and empirical work: the "leader is the person who acts like a leader," the "leader is the individual in the group who has the task of directing and coordinating task-relevant group activities" (p. 2) Hemphill (1952) similarly takes a behavioral definition of leadership: "to lead is to engage in an act which initiates a structure in the interaction of others as part of the process of solving a mutual problem" (cited in Gibb, 1969, pp. 214-215) „ Leaders are identified as those who engage in leadership behaviors or by the relative frequency with which they engage in such leadership acts. Other researchers using this conception of leadership include those who use Bales' (19 51) schema for identifying the leader. Leader behavior is characterized as having high levels of performance output consisting of suggestions, opinions, inform.ation, and orienting or integrating statements (Eskilson & Wiley, 1976) There are, however, serious problems with this type of definition. It in no way approaches a conceptual definition for leadership. It is certainly, an operational definition v/ith obvious empirical usefulness and clarity, but it does not and indeed cannot bring the theoretician any closer to a conceptualization of leadership. At the worst this -cype cf definition runs the risk of circularity: "\-7hat is a leader? The leader is the person who engages ^^*v|i oniti ,a'-^ 'ii'i If .'I** >r^i'#iA.t vsai

PAGE 25

13 in leadership behavior. What is leadership behavior? It is the behavior of one who leads." At best, this conception of leadership becomes broad to the point of being practically meaningless when leadership is defined in terms of sets or lists of specific leadership behaviors. That is, "this approach leads to a heterogeneous mass of specific acts that supposedly identify leadership in the group. V7hat constitutes leadership depends upon the view of the person who is listing leadership behaviors" (Shaw, 1976, p. 274) Further, it depends upon the behaviors which happen to be exhibited by the particular leader in the particular situation just observed. As Gibb (1969) concludes, "to shift the problem of definition from that of defining the leader to that of defining leader behavior or leadership acts has advantages for particular researchers but it offers no solution to the definitional problem." (p. 215) In any case, we are left with the problem of deciding what constitutes "leadership behavior" and ultimately this must be decided from within a specified, prior, conceptualization of leadership. The general category must be conceptualized or defined prior to and separately from the identification of the m^embers of that category. Despite the difficulties of the underlying conceptualization of leadership which this behavioral approach assumes, v/e may still go on to survey some of the findings this research approach has offered. "The mm\\T [I'r^Miir Till ~ "w n i 'iin* '>' — i. r h i g i,|i w i L III' '>'^*'^ w wy<^ fJ '^ '' ''*^** .S*>^**r**-— -^**s.-*-^~j*''^**''^**i&>i^

PAGE 26

14 psychological literature contains many analyses V7hich have attempted to describe what it is that leaders actually do" (Gibb, 1969, p. 228), and such studies continue up to the present. Insofar as studies within this tradition attempted to find specific leader defining behaviors, this approach has met with little success. The behavioral approach has largely failed along similar lines as the trait approach. Research again points to the importance of the situation in calling forth different types or sets of leader behavior, especially when the effectiveness of leader behavior is the focus. Leader behavior has been found to vary considerably from situation to situation; behavior which characterized leaders in one task situation does not necessarily characterize leaders in others (Fiedler & Chem.ers 1974). As in the search for leader traits, research here indicates the futility of looking for behaviors which would invariably distinguish leaders from followers. In general, the research shows little if any difference in the types of behavior exhibited by leaders and followers, although often there do appear to be real differences between those positions in terms of the frequency with which certain behaviors (such as directing, planning, controlling, interpreting what was taking place, offering ideas, initiating procedures for accomplishing the task, etc.) were perform.ed (Fiedler & Chemers, 1974). Fiedler (1971) concludes that the difference between leaders and followers is considerably less rt tf— t™ ^ *i— — — ***ii'— ji I "' ig".^' :


PAGE 27

15 than we ordinarily assume. Although we can usually identify leaders, there appear to be no specific behaviors in which only leaders engage. Leadership acts are to a greater or lesser degree performed by more group members than just the leader. Although this approach has not been fruitful in clarifying the nature of leadership, the approach has been useful in providing lists or descriptions of leader behavior and potential operational definitions of leadership based on those lists. This approach has extended our knowledge of what leader behaviors are or can be, their functions, and allowing us to categorize and dimensionalize leader behaviors while also allowing for the fact that SLich behaviors will vary across situations and will vary especially in their effectiveness across differing situational contexts. Further, this wealth cf useful information comes to us from a wide variety of methodologies and research concerns: from the observational studies v/hich attempted to identify factors of leader behavior (e.g., Carter, 1953, who distinguished three m.ain factors of leader behavior: group goal facilitation, individual performance-prominence, and group sociability) to those attem.pts to describe leader behavior (e.g., the research done at Ohio State University and a summary of some of that research by Halpin and Winer, 1952, who found four categories which accounted for all leader behavior: consideration, initiating structure, production emphasis and B V-, •fc*-l^rt*-i,^

PAGE 28

16 sensitivity or social awareness. The first two were found to account for over 80% of leader behavior) Also included is the work by Bales (1951, 1953) and others who attempted to categorize and delineate leader role behaviors. Research within this tradition may be examined within the review provided by Gibb (19 69) and in the work of Bales (19 53) Stogdill and Coon (19 57) Couch and Carter (1952), aiaong others. One particularly well-supported and useful conclusion which may be drawn from the research literature in this area is that leader behavior can be conceptualized as having two prim.ary and distinct dimensions: those of consideration and initiation of structure (Gibb, 1969) or the parallel, of socioem.otional leadership and task leadership (Bales, 1953). As Fiedler, 1971, states, there is abundant evidence that the consideration and initiation dimensions, or similar factors, are of overriding importance in m.ost leadership situations. Their identification constitutes one of the most important achievements of leadership research. (p. 7.)' The behavioral approach, while it has been useful for providing descriptions of leader behavior and identifying factors coirinon to leadership, remains largely unsatisfactory. This approach does not serve well in terms of clarifying the nature of leadership. If anything, it clouds the issue with its lists and lists of specific leader acts. It fares not much better than the trait approach to the extent that few consistencies in leader

PAGE 29

17 behavior are found across situations and few behaviors are found to consistently distinguish leaders from non-leaders. It is suggested here that the major failings of this approach center on the lack of a clear conceptual or theoretical definition of leadership and the misdirected focus of attention on leader behavior instead of on the interaction taking place between leader and led. Situational approach One residue from the Lewin et al, (19 39) style approach as well as from v/hat this author has called a behavioral approach to leadership is the m.ovement toward viewving the different contexts in which leadership occurs. This ultimately evolved into the situational approach to leadership. This approach took firm hold of the leadership area in the 19 50s and continues to some extent into the present (Hollander & Julian, 1969) Research from the earlier trait approach also led to this new tradition as more and m.ore evidence pointed to the importance of situational demands in determining the qualities, characteristics, and skills required of a leader; traits which were positively related to leadership in one situation were often unrelated, if not negatively related, to leadership in another situation (Fiedler, 19 71; Fiedler & Chem.ers 1974; Shaw, 1976). Thus, this new approach to the study of leadership began with a realization of the importance to leadership of situational variables. The main focus of this situational approach was the studv of leaders in different settings where the ,r— ..-.(-

PAGE 30

setting v;as defined especially in terms cf variations in group task and group structure (Hollander & Julian, 1969, p. 387) Variations in setting also included such things as changes in group member characteristics, the environment, and outside influences. The intent was to find "families of situations" within which the leadership role could be seen as relatively consistent (Gibb, 1969), Situational factors were seen to influence not only the specific types of behavior leaders exhibited and how they were valued or reacted to, the types of qualities and skills required of a leader, the effectiveness of the leader, and the perceptions of leadership, but such factors v;ere also found to be influential in the determination of v7ho -would em.erge as leader (Hollander & Julian, 1969? Shaw, 1976). This new approach to leadership was not so much directed at clarifying the nature of leadership as it was directed at the identification of those situational factors which affected leadership (in the ways described above). That is, this tradition began to take a multifactorial approach to the study of leadership. One exam^ple of an attempt to take a multifactorial, situational approach to the study of leadership can be seen in Fiedler's Contingency Model of Leadership Effectiveness. Research attempts to support and specify this model are many and continue a healthy trend in leadership research today. This particular model has been developed in an attempt to explain and predict leader behavior,

PAGE 31

19 primarily leader effectiveness, across situations. The model utilizes factors from both the trait and situational approaches as they interact together. The key personality measure is the LPC score which involves a description of one's least preferred coworker and which is related to a motivation or tendency toward a particular leadership style. The situation is important insofar as it is defined to be favorable or unfavorable to either of the two personality types (high LPC scorers and low LPC scorers) This favorability-unf avorability dimension is determined in term.s of three situational factors: (1) the nature of the interpersonal relationship between the leader and the group members, (2) the task structure dimension, and (3) the leader's position power (Fiedler, 1971? Fiedler & Chemers, 1974; Shaw, 1976). Thus, this model of leadership effectiveness suggests that effectiveness (in terms of group performance) is contingent upon both the motivational system, of the leader, as m.easured by the LPC scale, and the situational f avorableness (e.g., the degree to which the leader has control and influence in a particular situation) The relationship between the LPC score and effectiveness as a group leader has been exarained in numerous studies and the results from, such investigations have indicated that the situation (as defined by the factors m>enticned above) is highly im.portant. Fiedler (1967) finds for examiple that the willingness of group m^embers to T -n-er. •

PAGE 32

2 be influenced by the leader is conditioned by leader characteristics but that the quality and direction of this influence is contingent on the situational factors defined by group relations and task structure. However, given a specified situation, Fiedler and others have found consistent and significant correlations between the LPC score and type of leader behavior and group performance thus suggesting that LPC is an index of certain behavioral preferences (Fiedler & Chemers 1974). Generally, the high LPC person has a basic goal to establish good affective ties with the others in the group situation (parallel to what has been termed a socioemotional type of leader) while the lov; LPC person's prim.ary goal is successful task accom.pl ishment (parallel to a task leader) The low LPC leader is generally more effective in situations which are either highly favorable or unfavorable to the leader, whereas the high LPC leader is generally more effective in situations which are moderately favorable. As Shaw (1976) concludes, although there remain some problems and unanswered questions, Fiedler's model does make a prcm.ising beginning tov;ard the integration of leadership styles and situational factors as determinants of group effectiveness (p. 283) In addition to the work of Fiedler, a great deal of other research has taken this situational or multifactorial approach to the study of leadership. From this bodv of research the following statem^ents can be made

PAGE 33

21 regcirding effects of situational variables on leader behavior, leader effectiveness, and leader emergence within small group situations. Leadership performance on one type of task is essentially unrelated to leadership performance on another type of task (Fiedler & Chemers, 1974); thus one cannot really speak of effective or ineffective leaders or leadership styles since effectiveness depends on the situation. In terms of leader emergence, the j findings from this research approach "substantially supported the contention that who became a leader depended in som.e degree upon the nature of the task" (Hollander & Julian, 1969, p. 388). Also, people who are most motivated and visible within the group situation are m.ost likely to enterge as the leader (Fiedler & Chemers, 1974) From a variety of research efforts (e.g., research on reward structures, communication networks, etc.) we see that the increased visibility which often leads to leadership status may be afforded in a variety of ways within the small group situation: holding a central position v/ithin a communication network, having a high participation level relative to other group members, having obviously high status v/hich m.ay include being labeled as an expert, having personal attributes which make the individual more visible than other group members, etc. (Bavelas, Hasrorf, Gross & Kite, 1965; Fiedler & Chemers, 1974: Gintner & Lindskold. 1975; Riecken, 1958).

PAGE 34

22 In suiTuning up the importance of the situation in understanding leadership behavior, Gibb (1969) concluded that "leadership is always relative to the situation" (p. 246) and Fiedler and Chemers (1974) conclude that most differences in the way people act are relatively minor when we consider how much of their behavior is determined by social context {p. 56) "Our data suggest that leader behavior is more strongly determined by the situation than by what the individual would like to do or thinks he ought to do" (Fiedler & Chemers, 1974, p. 96). Even though this approach was based upon the recognition that the behavior, effectiveness, and emergence of leaders vzere dependent upon the demands of the situation and that the definition of the situation included not only group structure and the nature of the task, but also included the characteristics, perceptions, and behaviors of the group men)ers, "comparatively little attention was directed to the follov;ers, especially in term.s of the phenomenon of emergent leadership" (Hollander & Julian, 1969, p, 389). And even though leadership was beginning to be seen as an outcome of a relationship involving the leader, the led, and their shared situation, studies of leadership within this tradition paid little attention to the process aspects of leadership within the group situation. As Hollander and Julian have stated, "the sj.tuational view made it appear that the leader and the situation were quite separate" (1969, p. 389). The

PAGE 35

23 trait and situational approaches erred in a similar manner7 i.e., both tried to approach the study of leadership by emphasizing '-"parts of a process which are by no means separable" (Hollander & Julian, 1969, p. 389). This author would add that the style and behavioral approaches to the study of leadership have erred in a similar manner. Interactionist approach The final approach to be discussed here will be called the interactionist approach to leadership. It is not clear that there is a single, primary trend in the area of leadership research today since many of the traditions already discussed continue to enjoy research attention. Additionally, there are other approaches to the study of leadership in the current literature which have not been mentioned such as investigations into leadership training and an entire body of work focusing on leadership effectiveness per se (definitions of effectiveness, measures of, training for greater leader effectiveness, etc.). Further, accompanying the decrease in interest am>ong behavioral scientists v/ithin the last decade in small group research, there has been a concomitant decrease in research on small group leadership. Thus, although. one cannot say that at present a healthy research trend exists in terms of an interactionist approach to leadership, one certainly could say that much of the preceding research in leadership, coming from the varied traditions mentioned, points to the importance if

PAGE 36

24 not the necessity of taking a new, broader, interactionist approach to the study of leadership. An interactionist approach, at least to this author, does not imply a single approach to leadership. There appears to be at least two interactionist views. One view is based on the assumption that leadership can best be studied by taking into account the contributions of those factors explored in the previous approaches (e.g., personality traits and attributes of leaders and other group members, group member behavior and leadership styles, and situational variables) as they act alone or interact to affect leadership emergence, leader behavior, leader effectiveness, etc. That is, to study a particular aspect of leadership one would have to view all inputs to the small group situation (at least until a determination was made regarding which inputs were most important to specified aspects of leadership) as they act alone and in concert. This approach m^ay be seen as the logical extension of Fiedler-like approaches whereby instead of looking at the interplay of only two sets of variables, one must look at the interplay of several sets of variables and attempt to determine which interactions are most informiative. This constitutes what more properly may be termed a systems or a multif actor approach to leadership. In essence then, interaction here refers to a methodological or statistical interaction among the variety of sets of relevant variables. Other researchers in the field have r^.•*ffm*te.> -•,?"• •lHM.~r
PAGE 37

noted the importance of studying leadership from this perspective and have in fact begun to do such research themselves (e.g., Shaw & Harkey, 1976). A second approach within an interactionist perspective of leadership, and one which shares many similarities with what Hollander (1978) has recently discussed as a transactional approach to leadership, involves a different methodological focus as well as a new, or elaborated conception of leadership from those approaches previously discussed. Within this approach, in the attempt to study and understand various aspects of leadership (such as emergence, behavior, effectiveness, etc.) focus is directed on the interaction occurring among the group members. Leadership is seen as a type of relationship existing between people which is expressed by the type of interaction taking place. The goals of research coming from this perspecti'/e are: (1) to identify the type of interaction patterns that express a perceived leadership situation and (2) to delineate those aspects of the interaction which carry the most information in defining the nature of the leadership relationship (e.g., which aspects of the verbal and nonverbal interactive, communicative behaviors specify leadership relationships) A secondary line of research within this approach is to determine which (and in what manner) personal and situational variables affect interaction variables and thus indirectly affect leadership relationships.

PAGE 38

26 It is this i.nteractionist approach which the. present writer suggests as being the most useful approach to the study of leadership and which provides the perspective for the present investigations. The value of this interactionist approach lies in its most central point: that leadership is a relationship. Therefore, to study leadership one must view the behavior of all individuals (leaders as well as followers) during the process of interaction. This approach places a new emphasis upon the relationships, the interactions, and follov;er's behavior within the group; all of which have been largely neglected in previous approaches to the study of leadership. While a research tradition taking this approach is not yet v/ell established by students of leadership, the idea of an interactionist approach to leadership is not at all new. As Gibb (1969) points out, one of the earliest empirical investigations of leadership, done in 19 4 by Turman, assumed something of this perspective. Various studies throughout the leadership literature of the past, especially those using real-life situations, have also pointed out the interactive feature of leadership. For example, Gibb (19 69) pointed out that leadership could not rest in the stimulus individuals alone, but necessarily rested as v/ell within the attitudes, beliefs and responses of the others toward the stimulus person (p. 269) In other words, leadership is defined within the interaction by taking into account follower perceptions and behavioral

PAGE 39

27 responses to the leader, rather than being defined simply in terms of the behaviors or attributes of the leader. Others who have taken this interactive perspective include role theorists for whom one role is defined in part by its counterpart (Sarbin & Allen, 19 68) and who by necessity must look at the role enactment of all group m.embers visa-vis each other, Hollander's (1978) recently elaborated transactional approach sim.ilarly emphasizes the relational aspects of leadership. This view of leadership stresses the necessity of studying the transactional process existing among leaders and followers as they exchange benefits and mutually influence one another. "Leadership must include the reactions of followers" (p. 2). Without responsive followers there is no leadership because the concept of leadership is relational. Other contemporary authors (Deckhouse, Tanur, Weiler & Weinstein, 1975; Hollander & Julian, 1969; Lashbrook, 1975) have also pointed to the usefulness of this approach and suggested it as a new perspective for leadership research. Hollander and Julian have pointed to the "gr-eater sensitivity to the social processes of interaction and exchange" v/ithin recent leadership research and to the "increasing signs of movement toward a fuller analysis of leadership as a social influence process" (p. 383) They go en to say that "the tendency now is to attach far greater significance to the interrelationships between the leader, the followers, and the situation" (p. 388). They point out .Tft(T'*HKiA*<*.:^^5*,^-.fcB-, JI|lt*1l u jC:;-gaBi W .l t^cy^ ? 'g jW> l 'T ^I*

PAGE 40

28 that one consequence of this redirection of focus is the realization that the problem of studying leadership and understanding leadership relationships is much more formidable than was earlier assumed (p. 388). The arguments in favor of this approach are many but the central argument suggests that this interactionist approach-, of all other approaches mentioned, is best suited to the study of leadership because of the nature of leadership itself, A discussion of the interactionist approach entails a discussion of the conception of leadership. Leadership as an influence relationship Although it has been shown earlier that there is little apparent agreement regarding the nature or definition of leadership, certain trends in preference for specific definitions have been noted. Specifically, leadership has been defined in terms of (1) personality traits (within the trait approach) (2) as a set of behaviors engaged in by the leader (within the behavioral approach) and (3) as a type of interaction existing among leaders and their subordinates (Fiedler, 1971) as a set of influence attempts (Hersey & Blanchard, 197 6)., as a process of influencing group activities toward goal setting and goal achievement (Stogdill, 1948), and as the exertion of positive (e.g., desired by the leader) influence over the other group members (Shaw, 1976), to name just a few definitions. If one looks closely, it is this third view, leadership as %'5CO^lg^^WtHx^^iwr****^-*

PAGE 41

29 influence, which appears continually throughout the variety of past leadership approaches. In Carter's (1953) summary of the different definitions of leadership appearing in the literature up to that date, three of the five categories of definitions he distinguished were clearly associated with an influence relationship view of leadership. Carter (1953, pp. 262264) listed the following categories of leadership definitions: (1) The leader is the person who is the focus of group behaviors. The leader receives more communications, has greater influence upon the group's decisions, etc. (2) The leader is the person v/ho is able to lead the group t.ov7ard its goals. (3) The leader is defined as the person so nanied by the members of the group, based on sociometric choice. (4) The leader is defined as the person who has the most dem.onstrable influence upon group syntality. This refers to Cattell's (1951) theory; the leader is the person who causes syntality change, e.g., a change in the level of group perf orm.ance, and finally (5) The leader is the person who engages in leadership behaviors. Category three above clearly reflects an operational, not a theoretical, view of leadership. This definition, as Carter points out, merely identifies the occupant of the "leader" position and does not tell one anything about the characteristics of leadership. Further, who is chosen by sociometric choice as leader depends upon what question the researcher asks the group members: e.g., "who do you think

PAGE 42

30 was the leader?" "who would you most like to work with again?" "who did you like, the best?" or "who exerted the most influence upon you?" The choice of the question will depend upon the researcher's prior conception of leadership. Thus, this type of definition is only an operationalization of an earlier selected theoretical definition of leadership. Category five above essentially defines the behavioral view of leadership. But in categories one, two and four, it is clear that the central concept is successful influence of the leader over the group. Leadership is either related to the leader's ability to make changes in group member behavior such that group decisions reflect the views of the leader, or the group moves in the direction of goal accomplishment or the level of group performance is altered due to the effect of the leader, VJhile there may be difficulties, as Carter has suggested, with some aspects of each of the above definitions (e.g., the focus of group behavior may be a deviant group member, not the leader; leaders don't always lead groups toward a group defined goal) the influence relationship aspect of leadership appears well accepted by many researchers in the area (Cartwright & Zander, 19 6 8; Fiedler, 19 71; Fiedler & Chemers, 1974; Gibb, 1969; Shaw, 1976). While leadership as influence has obviously been a part of the conception of leadership for some time (Cowley, 1928, stated that the leader was one who succeeded in getting others to follow him) many researchers in the

PAGE 43

31 leadership area have only recently begun to acknowledge or emphasize this view. Gibb (1969) concluded his review of the definitions given to leadership over the years with the following comment: "in general, it is an essential feature of the concept of leading that influence is exerted by one individual upon another" (p. 212) Tedeschi and Lindskold (1976) note that "the definition of leadership in term.s of influence or power is a relatively recent development" (p. 581) Hollander and Julian (1969) state, "since the beginning of the 1960 's there has been a redirection of interest in leadership toward processes such as power and authority relationships" (p. 388) Finally, Fiedler and Chemers (197 4) recently suggested that the tv/o m.ost important threads running through nearly all leadership definitions were: (1) that leadership is a relationship between people in which influence and povrer are unevenly distributed and (2) that there can be no leader in isolation, that followers must either explicitly or implicitly consent to their part in this influence relationship. Thus, leadership implies some sort of exchange between leaders and followers (p. 4). "Above all, leadership is a relationship. More specifically, it is a relationship based on one person's power and influence over others" (Fiedler, 1971, p. 2) There is thus overwhelming agreement that leadership is an influence relationship and this brings the discussion back to the suggestion of the usefulness of the Tr^tv3e;^l'1ik'^4r .t. ^ wdmmi vfiir^fimtf m* /ui>b..*l**1inoju

PAGE 44

32 interactionist approach to the study of leadership. It is the only approach discussed whi.ch focuses on just that aspect of group behavior where influence relationships are visible: the group interaction, the cornmunicative process existing between leader and led. Leadership can not be studied by looking at the leader alone (whether it be in terms of leader traits, behavior or leadership style) nor the situation, if one neglects to look at the behavior of the followers specifically in terms of their responses to leader behavior. It is not at all surprising that other approaches have met with failure v/ith regard to clarifying the nature of leadership. Within an interactionist perspective, one Tvould not necessarily expect leaders to differ from non-leaders (by trait or behavior) or for leaders to behave consistently across different situations. Rather, what one v/ould expect, given this view of leadership, is that the relationship between leader and led I would remain essentially the same across situations. The if responses of non-leaders toward leader behavior would be different from the responses of group members toward nonleader behavior (e.g., the relationships betv/een leader and non-leader would be different from^ the relationships existing among non-leaders) This interactionist approach is essentially the same as Hollander's (1978) transactional approach, Hollander argues that this is the approach which leadership researchers should now be taking. Thus, well respected I l ii ail ll i i M IIIMi M l

PAGE 45

33 leadership theorists have argued that a new approach should be taken to the study of leadership. Further, failures of previous approaches suggest that new directions to the study of leadership must be taken. And finally, the concept of leadership itself necessitates the choice of an approach which focuses on the group interaction as a whole because it is that interaction which defines the nature or existence of an influence or leadership relationship. The final section of this chapter will provide a brief survey of studies which have focused on interactional aspects of group behavior. First, with regard to studies of leadership in particular in small group situations, there do not appear to be many studies which have actually been done from the interactionist perspective, and none which have looked at the small group interaction in its entirety as the focus for statements about leadership. There are a few investigations which have pointed to specific aspects of follower behavior in either defining a leadership situation or in leadership em.ergence. Gibb (19 69) cites one study by P^rout (19 42) who reported the case of an engineer turned cult leader. Calling himself "The Great I Am, he is said to have marshaled thousands of people into abject worship. Though it is apparent that the man was insane, instead of being labeled a madman he became a leader to large numbers of people. The essential difference betv/een his being labeled a g i t ^m g. ii i iii ahiw f S ^n w y ifSemmmmmm^ x M^mmm f ttmn t mm ^ i w iiM ii mii HiiM iwwiiw^^ui''i< ta' w ii.— j h —

PAGE 46

34 madman, versus a leader, was in the attitudes and responses of the followers toward him. A more recent study similarly points to the role of the follower in terms of emergent leader behavior. Backhouse et al. (1975) found that leadership behavior was a joint function of situational factors, individual predispositions, and the followers' responsive behaviors. When subjects (previously appointed by the researcher as the group task leader) were pushed to lead by the followers, they led. When they were pushed to follow, they increased behaviors associated with a socioemotional role. Other studies have found that positive and negative reinforcement of the leader's verbal behavior by other group members has the effect of either decreasing or increasing leader behavior (Bavelas, Hastorf, Gross & Kite, 1965; Pepinsky, Hemphill & Shevitz 1958). Studies which have focused on other aspects of leader and follower behavior include those which have looked at the relative rates of group member participation occurring over the course of the entire interaction. Time and again the emergence or perception of leadership is linked to the individual group member having the highest level of participation (Sorrentino & Boutillier, 1975; Stang, 1973; Stein, Geis & Damarin, 1973) Burke (1974) has linked leadership not just to relative rates of participation but to the control of the comm.unicative interaction in the group. Leaders were m.ore active than follov/ers in taking '<* i> 4 Ci? i^l gT ai| '. pi fc? i n i"ii.t iJirTf^iW/'^'!'*a lhh i— mt ia fcr n

PAGE 47

35 the speaking floor and they were also given the floor more often than other group members. Studies in which leadership was not the primary concern, but where attention was focused on the group interaction include the following. Bales' (1951) classic work of interaction process analysis provides a system for categorizing the behavior of all group members vis-a-vis one another, as they interact in a small group situation. This system certainly allov/s for categorization of leader behavior in isolation but also allows categorization of responses to leader behavior by other group members, thus defining relationships among group members. Bales and Strodtbeck (19 51) examined group interactions in terms of an elaboration of distinct phases occurring during group problem solving. Similar work on phases in group developm.ent has been done by Bennis and Shepard (19 56) Both of these phase development analyses could be re-examined in terms of what they might have to offer regarding leader and follower roles in group decision making and in termLS of leader emergence during group development. Even among communication researchers, research in terms of interaction analysis of communication process is quite recent and is rather sparse. The bulk of this research has occurred only within the last decade (Fisher, 1978) This relatively new approach (called the pragm.atic perspective) places em.phasis on the ongoing behavioral sequences of interaction which either constrain or serve

PAGE 48

36 to define the social situation (Fisher, 1978). In one of the few interactive studies of leadership. Bell (1978) has analyzed sequences of communication behaviors occurring between leader and led in small groups with either taskoriented leaders or relation-oriented leaders. Though not directly related to the study of leadership, communication researchers have developed a vairiety of category systems (content analysis schemes) for analyzing interactive behaviors, some of which might be seen as being related to leadership. Systems have been developed to look at processes involved in idea development (e.g., Scheidel & Crowell, 1964), decision m.aking (e.g., Fisher, 1970), and conflict managem^ent (e.g., Ellis & Fisher, 1975). Category systems have also been developed which allow analysis of various types of relationships created or expressed through interpersonal or group com-munication. Much of this research (e.g., Rogers & Farace, 1975) focuses on communications which are related to an interactant s attempt to dominate (by m.eans of one-up communication patterns) attempt to be submissive (by using one-down communications) or attemipt to be equivalent (by use of one-across communication patterns) Psychologists as well as communication researchers have recently begun to focus on the nonverbal aspects of interactions. VJith the exception of Stein's (1975) investigation which found that emergence and perception of leadership in small group settings viera conveyed along ••J ie-*>M>.( i^-^ ,'">,ii:
PAGE 49

37 both verbal and nonverbal channels, few studies deal with leadership per se. Most of the studies which deal with nonverbal behavioral aspects of interaction deal v/ith perceptions of dominance and status. These studies come largely out of the comm^unication field and they m.ay be seen to be related to the concept of leadership insofar as they deal with behavior patterns which are indicative of influence relationships. Perceived dominance has been associated with attempted and successful interruptions (Folger & Sillars, 1977; Henley, 1977) taking relaxed and arms akimbo body postures (Mehrabian, 1969), crowding another's space and pointing (Henley, 1977). High power in group or interpersonal settings has been associated with being able to look less at the other members of the group or interaction. The lesser power individual is presumed to have a greater need to monitor the expressive behavior of the other (the high power individual) in order to gain information regarding the high power individual's reactions and so be able to adjust behavior. Looking also enables the low power person to indicate that she or he is attentive to the higher power person (Exline, 1974). Finally, high power is also associated with having ones gestural behaviors or body postures matched. Persons in agreement and persons of lower power will unconsciously reflect the behaviors of those in power according to Henley (1977) WM^> M iW^:W;fcttji,>*ii— jci tt '' .'^ Twiim ^

PAGE 50

38 This concludes a review of the leadership literature in terms of the conceptions of and approaches to leadership. A brief review of literature covering the trait, style, behavioral, situational and interactionist approaches to the study of leadership has been provided. The value as well as the problems associated with each of the above approaches have also been discussed. At present it appears that the interactionist approach holds the m.ost promise for future leadership research. The following sections will provide a brief summary and a more specific rationale than that provided in these sections for studies one and two. Rationale and Statement of Hypotheses for Studies 1 and 2 Study 1 The purpose of Study 1 was to determine to what extent observers of a small group interaction based their perceptions of group leadership on an influence relationship and to provide an empirical demonstration that perceptions of leadership rest more upon the presence of social influence relationships existing am.ong group member interactions than the mere presentation of what has been previously described as "leader-like" behaviors on the part of one or more group members. Theorists (Beckhouse, Tanur, Weiler & Weinstein, 1975; Hollander, 1978; Shaw, 1976) in the area of leadership have recently come to suggest that. leadership is properly conceptualized as a social influence relationship < ~irrni M m i Kw t*. ^ '^ ^ ^tig' ^Mcam*

PAGE 51

39 among interacting individuals where the behavior of both leader and follower (s) must be taken into account insofar as "leaders" make influence attempts within the interaction and "followers" accept those attempts. Thus, the role of each is in part defined by the role of the other. In this case, the role of leader (offering advice, miaking suggestions regarding such things as goal choice and procedures for goal accomplishment, offering suggestions for group structure, attempting to control group member behavior, etc.) is fulfilled when followers adopt their role in accepting or agreeing to these influence attempts ; of the leader. This conception of leadership as influence is in contrast to earlier (though still prevalent) views that leadership could be defined or characterized by referring solely to the behavior of the leader (see Bales, 1951, 1953; Fiedler, 1971; and Hemphill, 1952) That is, according to this earlier conception, leaders are those individuals who engage in such leader-like behavior as: offering ideas and advice, making suggestions, interpreting and offering directions to group behavior, having a high degree of participation and visibility, etc., without regard to any other group member's behavior. In essence, this view is half correct. It does, as well it should, focus some attention on the behavior of leaders. Further, this conception of leadership has been most valuable in providing detailed descriptions of the variety of behaviors in which leaders engage. Valuable as it has been. Ti.ri[invrMiHi"..'i'~"'Ti ii r—rn i n i --mM i i i i nffiiin i m -' r T ^iTVirfi f ti. n .]ii ji i

PAGE 52

40 however, it has not provided a clear, cross-situational view of leadership (Fiedler & Chemers 1974). It has been plagued by generation of list after list of "leader" behavior since the behavior of leaders can change markedly from one situation to another and from one leader to another. This problem exists because this is an incomplete conception of leadership. It looks at only part of the picture, one-half of the group interaction. The focus is misdirected at the behavior of one individual alone, the leader, where it should be placed on the interaction among leader and follov/er. Although specific leader behaviors might be expected to change across situations, what does not change across situations is the relationship between group members. That is, the relationship between leader and follower, as defined by the acceptance of follov7ers of leader influence attempts, does not change. Thus, leadership does not change across situations. Although theoretically sound, the influence conception of leadership has received little empirical attention. One study which has linked influence and leadership is Gibb s (1950) investigation. Gibb formed groups of ten men each and assigned each group a variety of activities to involve them for a series of three, three-hour sessions. Initially, groups were left unstructured or leader-less but they were allowed to select a leader for the second and third sessions. After each session • general sociometric questions were asked of each group member including

PAGE 53

41 one question designed to discover each group member's judgment of the behavior influencing the group. Specifically, each member was asked the following question, "some groups are so closely knit that the removal of any one person changes its complexion. For which persons, if any, in this group, v/ould this be the case?" (p. 230) The group m.ember chosen m.ost frequently was seen as being the most influential. Group members were also asked who they would judge to have been leaders in each of the last two sessions (members received no specific definition of leadership) In addition to these group member judgments, independent observers had been instructed to observe each session and code the behaviors of each of the participants. Gibb found that: (1) the participants' responses to the influence question identified the same individuals as those who were behaviorally identified by the outside observers as leaders. The correlation obtained between these two items was .80. (2) Participant and observer judgments of leader also correlated .80. Thus, this is one study which succeeded in making an empirical link between leadership and influence. However, even this study may be considered weak to the extent that the question asked of the participants to determine the most "influential" group mem.ber does not necessarily convey that a true influence relationship existed among the participants. That is, selecting the individual group member "whose rem.oval from the group would most change the

PAGE 54

42 complexion of the group" does not provide anything but a very vague notion of influence. There do not appear to be any other studies in the leadership literature which specifically attempt to demonstrate that observers (or group members) rely on influence-relationship information in making their judgments of perceived group leadership. The following propositions and hypotheses were offered for study one. PI: Leadership is an interactional or relational phenom.ena P2: The type of relationship is defined by the type of interaction occurring among group members P3: The perception of leadership is dependent upon the observation of the behavior of both "leaders" and "followers" independently and especially, vis-avis each other (that is, in interaction) P4: Leadership is defined by the existence of a particular kind of interaction or relationship among group members; that being an influence relationship, P5: The perception of leadership is dependent upon the observation of "leader" behaviors directed at or in the presence of other individuals v/hich have an observable, perceived causal (influential) effect OR the behavior of the others (the "f ollov/ers") From, the propositions stated above, the following hypotheses were derived. HI: Leadership within a group will be perceived when an influence relationship exists within the group. Specifically, hypothesis one predicts that observers • perceptions of leadership will be strongly affected by the presence of an influence relationship xvithin the group. That is: there will be a main effect for amount of influenc e on the lead ershi p ratings for the designated leader (group member" three) such that high level3~of influence are associated with high leadership i-to&Gitf'Si^rAioM^^Ursui^ff^^iB:

PAGE 55

43 ratings and low levels of influence are associated with lower leadership ratings. Proposition three above proposes that both leader behavior and follov/er behavior are important dimensions in the perception of leadership. This is derived from the body of previous leadership literature (both theoretical and empirical) which has demonstrated the importance of the leader's behavior in and of itself for the perception of leadership in small group interactions. Thus, hypothesis two is proposed to take into account the role of leader alone for the perception of leadership. H2: The perception of leadership v/ithin a group varies with the perception of leader-like behaviors by at least one group member occurring within the group interaction. Specifically, hypothesis 2 predicts that: there will be a main effect for amount of leader-like behavior on the leadership ratings for the designated leader (group m.ember three) such that high levels of leader-like behaviors are associated with high leadership ratings and low levels of leader-like behaviors are associated with lower leadership ratings. Hov/ever, in a determ.ination of the relative importance of influence versus leader-like behavior for the perception of leadership, clearly, it is the entire interaction among group members conveying the influence relationship between leader and the followers which forms the basis for leadership perception. Thus, hypothesis three states: H3: The perception of the influence type of interaction among group members is more highly associated with perception of leadership than the perception of leader-like behaviors by one or more group members within the group interaction. That is, while both influence and leader-like behavior -*''=" '•— ^" T y^'^l"--"— • I ~" — ^•'~----= — '^ --^^ — — -TT

PAGE 56

44 (non-influential) are associated with the concept of leadership, perceptions of leadership are more strongly associated with the presence of an influence relationship than the presence of leaderlike behavior which has no influence effect on either the interaction or on other group meinhier behavior. Specifically, hypothesis 3 predicts that: the a mount of influence main effect factor will ac court for a greater percent of the variance than the leader-like behavior main effect factor As has already been stated, both amount of influence and amount of leader-like behavior affect the perception of leadership. A main effect for each has already been predicted (see hypotheses one, and two) Above and beyond these effects, it is expected that in combination, high levels of these two factors v/ill produce an even stronger perception of leadership than any of the factors acting alone. Thus, hypothesis four states that: H4 : There will be an interaction between am.ount of influence and amount of leaderlike behavior on ratings of leadership for the designated leader such that the condition containing high influence and high leader-like behavior will produce higher leadership ratings than in any other condition. Specifically, the expected ordering of the means for leadership ratings, for the designated leader (group m.ember three) v^ill be as follows : lowest mean for cell one (low influence, low leader-like behavior) followed by cell three (low influence, high leader-like behavior) cell two (high influence, low leader-like behavior) with cell four (high influence, high leader-like behavior) receiving the highest mean leadership rating. Each of the preceding tests of hypotheses which exam.ine the leadership ratings of group member three in determiination of effects due to level of influence and level of leaderlike behavior is based on the assumption

PAGE 57

45 that it is group member three (the designated leader) who is in fact chosen as group leaderThus, hypothesis five is a test of that basic assumption. H5 : The group member designated as leader (group member three) will receive significantly higher leadership ratings than either of the other two group m.embers. This effect v/ill be found for each of the cells in which there is a designated leader; specifically, in cells two, three and four. Specifically, hypothesis five predicts that a main effect for group m.ember within cells will be obtained such that group member three receives higher leadership ratings than group members one and two in cells two, three, and four. Study 2 The intention behind both studies one and two was to provide em.pirical support for the conception of leadership as an influence relationship existing among group members and expressed by group m.embers V interactive behaviors. By definition, influence can only be examined within the context of a relationship or interaction among individuals. However, the elements of influence and interactive behavior may be stressed separate]y and this is what studies one and tv;o purport to do. That is, study one em.phasizes that leadership exists and will be perceived within the small group interaction when one individual has greater influence over the other group members. Leadership is defined primarily as an influence relationship existing among interacting individuals (although not all influence relationships constitute leadership) Study one attempts to provide empirical support for the proposition that leadership is properly conceptualized as an influence relationship. — X ii I *ii^^ni*Si>lw ^ 'Si^^.i '• <~'| i I II n iri li n I "T i-|i 11 -• mill una" ii iii.i'i rn i i iVii'nifJi

PAGE 58

46 Study 2, on the other hand, takes irxfluence as a given for the perception of leadership, but attempts to examine more specifically the means by which influence, and hence leadership, is conveyed. The leader may engage in many different types of behavior which have the result of eliciting or causing a change in a variety of follower behavior. In this case, the influence conveying behaviors are classified into "types" according to the comjriunicative channel along v/hich they occur: verbal or nonverbal. Thus, the purpose of study 2 was to demonstrate that influence, and hence leadership, m.ay be conveyed both verbally and nonverbally within the group interaction. Study 2 attem.pts to provide leadership researchers with more detailed inform^ation than they presently have regarding how the influence relationship is actually conveyed to observers of sm.all group interactions. Specifically, this study deals with the following question: to what extent does each communicative channel (verbal and nonverbal) convey influence relationships necessary for the perception of leadership. The research direction taken in this study follows recently developing trends in small group and leadership research. These trends include the increasing attention directed to interactive processes such as communicative interactions (rather than interactive outcomes such as m.easures of group goal accomplishment and group satisfaction) and the increasing attention given to the

PAGE 59

4 7 multichannel nature of human communication (that is focusing on verbal as v/ell as nonverbal comm.unication in various interactive contexts) It is communication researchers for the most part (e.g., Burke, 1974; Gouran, 1973; Lashbrook, 1975; Spillman et al 1978) who are presently focusing attention on the ongoing, communicative process variables of small group interaction and emergent leadership. However, the classic work of Robert Bales (1951) in terms of content analyses schem.es for interaction process is the foundation for much of this research. An example of work in this area includes Bell's (1978) investigation in which communicative sequences between leaders and followers were analyzed. Bell found a pattern v/ithin groups with relation-oriented leaders (but no identifiable patterns within groups with task-oriented leaders) such that leader's "tension release" was followed by nonleader member's "tension release." While it does not appear that interaction process analysis of leadership has received much attention in the em.pirical v/ork of social psychologists, such a focus of attention has been encouraged in recent theoretical reviews of the leadership area (e.g., Hollander & Julian, 1969) Focusing attention on the interactive process as well as the m.ultichannel nature of human interactions, ccnim.unication researchers and recently, social psychologists have demonstrated that people can and do communicate a varietv of information along different channels. That

PAGE 60

48 is, information is carried by the verbal channel in terms of the content of what is said. Information is also carried by the nonverbal channel which includes vocalics (involving the tone of voice, amount of speech, intonation patterns, speech volume and speech rate, interruptions, etc. Vocalics includes all those features of the voice other than the words themselves) and kinesics (which includes facial expressions, the type and am.ount of eye contact, head nods and shakes, hand gestural activity, body position orientation and shifts, etc.). Although the importance of verbal behavior is well recognized, this has not always been so with regard to nonverbal behavior, particularly in regard to certain sub-topic areas such as leadership within the context of a small group interactionResearchers, especially in social psychology, have only just begun to examine the ways in which nonverbal (especially vocalic and kinesic) channels convey information and the types of inform.ation such channels are capable of conveying. However, we do know that each of these channels is important in communicative behavior: people both send and receive messages on them (Burgoon & Saine, 19 7 8; Weitz, 1974). Some theorists have even argued that the majority of the information (e.g., over 65%) we send and receive is conveyed on the nonverbal channels, as opposed to the verbal-content channel (Birdwhistell cited in Knapp, 1972, p. 12)

PAGE 61

49 A few investigators within the area of leadership have begun to look at nonverbal behavior and in particular, how leadership is conveyed by the various channels. For example, Stein (1975) has demonstrated that subject-observers were able to accurately identify emergent leadership within a group when provided with the following types of information: (a) full channel information as provided by the presentation of audio-video tapes of small group interaction, (b) verbal content and relative percent participation information provided by written transcripts of the group interaction, (c) kinesic channel inform.ation only as provided by the video-only (no-sound) condition, and (d) a filtered speech condition accompanied by the visual channel which provided kinesic channel information in addition to some vocalic information, but no verbal content information, to the observer subjects. Stein concluded that his study provided the first direct evidence that nonverbal behaviors are useful in perceiving leaders. Although Stein did demonstrate that observers can make accurate (accuracy was defined in term.s of agreement v/ith the judgments or choices of leadership made by the actual group members) judgments of leadership based, upon information conveyed along verbal and nonverbal channels, he does not demonstrate that it is in fact "leadership" that is being dealt with since he gives no evidence of an influence-relationship. The study does not demonstrate v/hat it is that is conveyed along these two ''* ^ ^: ^'* ml< ^
PAGE 62

X 50 channels. His ir.ethodology does not provide the reader with information regarding what specific behaviors or variations in behavior are occurring along each channel and between "leaders" and "followers." "Leader," in Stein's study, is that person chosen as leader. Study 2 provides a demonstration of the importance of both verbal and nonverbal channels for the perception of an influence-leadership relationship. In addition, it provides information regarding the specific nature of the behaviors being conveyed along both verbal and nonverbal channels. The following propositions and hypotheses were offered for Study 2. PI: Leadership is an interactional or relational phenomena (See PI for Study 1) P2: Leadership is defined in part by the existence of a particular kind of interaction or relationship among group members, that being an influence relationship (See P4 for Study 1) P3: Leadership (the existence of an influence relationship) is conveyed along verbal as well as nonverbal communicative channels within the small group interaction. Following from these propositions, the following hypotheses were derived. HI: The group member designated as leader (group member three) will receive higher leadership ratings than either of the other two group members regardless of whether the observer subjects receive information along verbal, nonverbal, or a combination of verbal plus nonverbal channels. Specifically, hypothesis one predicts that a main effect for g roup mLember within each expe ri m.ental cond it ion will be o bta ined such that group member "7)'T9lT.i*,i*iF' MdO-Uil

PAGE 63

51 three receives significantly higher leadership ratings than group members one and two. This hypothesis follows from proposition three above that influence relationships and thus leadership, will be conveyed through both verbal and nonverbal communicative channels within group interaction. The following hypothesis is proposed following the logic that the more influence conveyed, the stronger the perception of leadership, H2 : A cumulative effect is hypothesized such that leadership ratings for the designated leader will be higher for conditions in which the influence relationship is conveyed along two rather than one communicatii^e channel. Specifically, hypothesis two predicts that there will be a main effect for condition en lead ership ratings for the designated le ader (group miember three) such that condition one (in which influence is conveyed along both verbal and nonverbal channels through audio and visual means) is significantly different than any other experimental condition (where influence is conveyed by either verbal or nonverbal means but not both) Group member three v/ill receive significantly higher' mean ^ leadership ratings in condition one than in conditions two, three or four. ^ a iwa B TX^ p-m^ KWiTMi i i lfci l u ii m^ WW ilV.tt* rif,;M'm

PAGE 64

CHAPTER II METHOD Study One Subjects Subjects were 130 students enrolled in basic speech courses at Cleveland State University and basic psychology courses at John Carroll University. Participation in the experiment was on a credit given basis Procedure Upon arrival at an experimental lab, subjects were assigned in pairs, on a random basis, to experimental testing rooms. As they arrived in the experimental testing room each subject took a seat at a table. The seats were divided by a curtain partition to reduce interaction and eye contact among the students. Subjects were then told that the experimenter was interested in their reactions to a videotaped sm.all group interaction. Subjects. were told that they would observe on a videotape TV monitor a small group interaction which had occurred among three fellow students. Following ob, servation of the interaction, they were told that they would be asked to respond to a brief questionnaire regarding V7hat they had just observed. Subjects then read and filled out inform.ed consent f orm.s 52

PAGE 65

53 Subjects were then requested to pay careful attention to the TV monitor and to the experimenter's instructions. They were also requested not to interact with each other in any way nor to make any audible sounds (aside from requests for clarification of instructions from, the experim.enter) so as not to interfere v/ith the other subject's observation of the group interaction. Following these instructions, subjects were exposed to a no-sound videotaped still shot of group member number one and they were requested to fill out a series of scales by McCroskey and McCain (1974) designed to measure physical and task attraction. These scales were to be used as covariate measures. After each subject had filled out these scales, the same procedure v;as employed for group members two and three respectively. Subjects were instructed to fill out the scales based on the still shot information in front of them and they were instructed to rate each individual independently without attempting to compare the group members to each other. Subjects were then told that they were about to observe an interaction which took place among the three students whose still shots they had just viewed. They were requested to pay careful attention to the interaction and they were again reminded not to interact with each other in any way. At this time, subjects were exposed to one of four versions (comprising the four experimental conditions) of a videotaped interaction among group miembers one, two and three. iw iy a^ ii T ipiyrinm r i .ia iiw n_
PAGE 66

54 Subsequent to exposure to the videotaped group interaction, subjects were administered the post-experimental measures. Following completion of the post-experimental questionnaire, all participants were thanked for their participation and told that they would receive a handout in their classes within two to three weeks informing them of the purposes and results of the study. Following completion of data collection of the study, all students received a written explanation of the purposes, hypotheses and results of the study in which they participated. Design Subjects were randomly assigned to one of four experimental conditions (N = 130) In all four conditions subjects observed a videotaped interaction by means of a videotape recorder and television monitor. Within each of the four conditions, level of influence was manipulated on two levels (low and high) and amount of leader-like behavior was manipulated on tv7o levels (low and high). This constitutes a 2 x 2 factorial design. All four experimental conditions employed the sam.e three actors as group members (Cleveland State University theater majors) and across each condition, actor and role assignment remained constant. Thus, actors one and two always played follower roles in all conditions. Actor three always played the leader in the three experimental conditions where a leader was designated. ta Mf a J tr .Utf*.^gg"ii"- CCT^

PAGE 67

55 Holding actor and role assignment constant across conditions can provide problems in interpretation of results such that one is unsure whether obtained effects are due to the experimental manipulations (e.g., levels of influence or leader-like behavior) or the result of differences due to uncontrolled-for actor variables (e.g., actor three being perceived as more or less leader-like than actors one and two) based on physical appearance characteristics. This problem may be eliminated in either of tv70 ways. In terms of design, actors could be rotated through the roles such that each actor played each role (or played at least one other role) for each of the experimental conditions (thus blocking across actors) This procedure "was deemed impractical due to the time and effort involved in videotaping, the difficulty the actors had in learning their roles, and the number of subjects which would then be required. An alternative procedure involves devising measures for those possible effects due to uncontrolled-for actor variables (these should be primarily due to physical appearance since nearly all behaviors were controlled for by the scripting) and covarying out these effects from the effects due to experimental manipulations. This statistical solution to possible effects due to actor differences was deemed most practical for the present investigations. Thus, a 2 (low vs. high levels of influence) x 2 (low vs. high levels of leaderlike behavior) factorial '-*•'• -T— ^".^.— —

PAGE 68

56 design with two covariates (measures of physical and task attraction for each actor based on physical appearance alone) was used to analyze for differences due to levels of influence and leader-like behavior. A suininary of the design and conditions may be seen in Table 1. G roup interaction and videotape construction A three m.ember group interaction in which members were attempting to offer solutions for the parking problem at Cleveland State University was used as the stimulus material for this study. This topic was chosen because, it involved a problem with which nearly all college students were familiar (both Cleveland State students and John Carroll students had moderate to severe student parking problems) The verbal script for the small group interaction was carefully constructed through the procedures described below. Two volunteer groups of three individuals each (who did not participate in the later experimental phase of the studies) were asked to participate in a small group interaction. They were asked to offer solutions for the parking problem at Cleveland State University. They were told that the experimenter was interested in the behavior of small group, problem-solving groups and that their entire interaction would be tape recorded. Both groups then discussed the parking problem and were tape recorded. After twenty to thirty minutes the experimenter stopped the discussion, turned off the tape recorder and thanked al* I TI lii i rh iii ^r_^ -1 1 tii-i 1

PAGE 69

57 Table 1 Experimental Conditions for Study One Level of Leader-Like Behavior Low High Level of Influence Low High Condition 1 Condition 2 Condition 3 Condition 4 mftuf^:^^^:^^ tU

PAGE 70

58 the group members for their participation. They were then told that their interaction would be used as the basis for scripts for actors who would be portraying members in a sm.all group interaction for a later study. Each of the recorded small group interactions was then transcribed. These transcripts of naturally-interacting groups then provided the basis for a working script to be used for the creation of the scripts for each of the four experimental conditions. The working script for a three m.eraber group interaction was developed by selecting various verbatim passages (som.etimes a single speaker utterance, sometim.es a series of interchanges among two or more speakers) from the transcripts. Passages were selected for use when they offered reasonably clear suggestions for the solution of the parking problem or procedural directions to the group, when suggestions were followed by active group discussion, when passages offered reasonably clear responses (either positive or negative) to suggestions made. On occasion, words or whole speaking turns were inserted in order to produce a better flow of conversation. There were more than enough possible passages to select from the two, twenty-minute conversation transcripts. Once a draft of the working script was composed, it was read through and timed, shortened or lengthened, until the resulting interaction lasted from. six to eight m.inutes A six to eight minute interaction seemed m.ost appropriate for the experim.ental conditions

PAGE 71

59 since it would last long enough for a good deal of group interaction but would not be so long as to lose the attention of the subjects. This method of developing a basic working script had the advantages of producing a more naturalistic group interaction than the experimenter, alone, could create. Natural verbal content was obtained by using the ideas and words which actual group members had used. Such a method produced more naturalistic phrasing (e.g., the frequently non-grammatical, non-fluent speech) of students engaged in an informal speech situation. It was from this basic working script that scripts for each of the four experimental conditions were developed. It was intended that scripts for each condition be as nearly identical as possible, within the constraints of the experimental manipulations, in terms of general content and overall length of the interaction. This intention was carried through. In creating the script for each experimental condition, the verbal behavior of the interactants was dealt with first. Not all aspects of verbal content are necessarily related to leader behavior or to the establishment of an influence relationship among group members in a task oriented group. Those aspects of verbal content which were considered im.portant (based upon findings from, the behavioral approach to leadership) and which operationalized leader-like behavior included the followina: (a)

PAGE 72

60 coinments which give suggestions or directions to others (implying autonomy for others) (b) comments which give opinion, evaluation, analysis or which express feelings or a wish, and (c) commcents which give orientation, information, confirmation or clarification. Many of these have been listed by various leadership researchers as being leader-like behavior (Bales, 1951; Carter et al., 1968; Cartwright & Zander, 1968; Eskilson & Wiley, 1976; Morris & Hackman 1969). Specifically, these are descriptions of behavior which Bales (1951) has categorized in his interaction process analysis content analysis scheme as small group interaction, task oriented behaviors. Those comments listed above in a belong to Bales category number four, those in b belong to Bales' category five, and those in c belong to Bales' category six. See Appendix A for a complete list of the Bales categories. Those verbal content behaviors or interactions which are important to the establishment of an influence relationship are those which signify agreement with or compliance to the expressed suggestions, directions, analyses or feelings of another. Thus, verbal content indicative of an influence relationship was operationalized in terms of a group member's use of Bales' category number three: "agrees, shows passive acceptance, understands, concurs or complies" in response to another member's suggestion, opinion, etc. As Hare (1976) has pointed out, low power individuals including followers.

PAGE 73

61 generally pay more attention to and respond more favorably to persons of high power, which includes leaders. An absence of utterances coded as Bales' category three or the presence of utterances coded as belonging to Bales' category ten (disagrees, shows passive rejection, formality, withholds help) or Bales' category twelve (shows antagonism, deflates another's status, defends or asserts self) by a group member in response to another group member's suggestion, opinion, etc., was indicative of a lack of an influence relationship between the group mem^bers Thus, the system used to operationalize verbal content as being leader-like or indicative of an influence relationship was the content analysis system provided by Bales (1951) Specifically, utterances coded as belonging to categories four, five, and six defined leader-like behavior and utterances coded as belonging to categories three, ten and twelve defined the presence or the absence of an influence relationship. The level of leader-like behavior and the level of influence v/as dependent upon the number of utterances for each group member coded as belonging to these categories. The Bales' content analysis system has a total of twelve categories of which only six were used system.atically in the operationalization of factors in the present investigation. Categories seven, eight and nine all deal with asking questions (seeks information, asks for opinion, asks for suggestion or direction, etc.). In everv '^*'-''** [iriwt^n-wrTdjjTiinnwii

PAGE 74

62 condition an attempt was made to equalize across all three group members the number of utterances they each had which were coded as belonging to Bales' categories seven, eight or nine. Thus, each group m^ember, in each experim.ental condition, asked an approximately equal number of questions. This was done because it is not at all clear from' the leadership literature who it is, leaders or followers, that asks the most questions. Leaders may dem.and answers from followers, but followers, respecting the ability of the leader, frequently seek information, the guidance and opinion of their leaders. Bales' categories one (shows solidarity, raises others' status, gives help or reward), two (shows tension release, jokes, laughs, shows satisfaction) and eleven (shows tension, asks for help, withdraws out of field) were not dealt with systematically in the present investigation. Thus, operationalization by means of verbal content of level of leader-like behavior involved the number of utterances each group member had which were coded according to Bales' interaction process analysis as belonging to categories four, five or six. Operationalization by means of verbal content of level of influence involved the numiber of utterances belonging to Bales category three (and the lack of Bales' categories ten and twelve) vrhich each group member received from other group members following their utterance of a Bales' four, five or six tvpe.

PAGE 75

63 Nonverbal vocalic leader-like behavior was operationalized along two dimensions: (1) level of participation and (2) attempts to interrupt the speech of another group member in order to gain control of the floor. Each of these dimensions has been linked to leadership. Time and again, researchers have shown that leaders have higher rates of participation than non-leaders and that those individuals with high rates of participation are perceived as being leaders (Knutson & Holdridge, 1975; Morris & Hackman, 1969; Sorrentino & Boutillier, 1975; Stang, 1973). Fiedler and Chemers (1974) have shown that less dom.inating or lower status individuals are less likely to interrupt or attempt to interrupt the speech of others who are more dominant or are of higher status. Henley (197 7) points out that dominant individuals or those in positions of authority may show their dominance or authority by interrupting subordinates. In the present investigation, level of participation for each group member was systematically varied in terms of the number of speaking turns for each group member o A speaking turn has been defined by N. Markel as beginning when "one interlocutor starts solo talking and ends when a different interlocutor starts solo talking" (1975, p. 190). Level of participation was also operationalized in terms of the total number of words spoken by each group miemjber. Attempts to interrupt the speech of another v/ere operationalized as a period of solo speaking by one group m.em.ber during which another group T;**r:.jinT'al* ii" '>i rM i H iim-t>*'-*" rr*i i>-^ u

PAGE 76

64 member began simultaneous speech which was not successful in gaining the conversational floor. The nonverbal vocalic operationalization of level of influence was in terms of the number of successful interruptions achieved. That is, an individual was influential,along the nonverbal vocalic channel, when he was able to take control of the floor away from another group memJDer by interrupting that group FLember s speaking turn. Henley (1977) has stated that "a hierarchy of power in a group could be plotted by ordering people according to the number of successful interruptions they achieve" (p. 69) Nonverbal kinesic channel behaviors were also dealt with in creating scripts for group member behavior for each experimental condition. Again, only those kinesic behaviors which have been previously linked to leadership were used. Leader-like kinesic behaviors in the present investigation included attempts by one group member to indicate who the next speaker should be by means of hand gesture. It has been described as part of leader behavior to attem.pt to control or structure the interaction of the group members. Deciding who shall speak when is one aspect of this type of behavior. According to Henley (1977) authority is shown in controlling others from a distance. A dom.inant person m.ay direct others by gesture and "point at them, in a way that will shut them.

PAGE 77

65 up, stop other action, or evoke attention and submissiveness" (p. 128) and presumably, direct them to speak as vrell. Kinesic behaviors which were indicative of an influence relationship included the following: (1) success at indicating the next speaker by means of gesture. That is, when the person pointed to or gestured toward was in fact the person who spoke next, the turn indicating gesture V7as deemed influential or successful. (2) Having a gesture matched by at least one other group member or (3) having a body posture shift matched by at least one other group member was also indicative of an influence relationship. Research has indicated that high status or pov/erful individuals are m.ore likely to be matched by lower status others in terms of both their gestural activity and their body posture (Burgeon & Saine, 1978). The individual is influential if she or he is successful in affecting some of the kinesic behavior of the other group mem.bers. A final behavior indicative of an influence relationship in the present investigations was the receipt of a head nod indicating agreement for an uttei-ance of Bales' type four, five or six. Thus, agreement may be sent along the verbal content channel as described earlier or along the nonverbal kinesic channel. The presence of an influence relationship is also indicated by the lack of kinesic head behavior (head shakes) which show disagreement. l ii n ii il l iinl l liil i i iii i )ip wil. ill m ill ii w ii i Mju i l ir m n r r—T t nri i nicirr — i M tfjWwi'a i tf ii i'il i iM SJi l ii jutI i r tjr w T i< r iri lM rt'," ''! u mi I mii W iiii > w i li^i '' — l y^ *Trl n'T ^ii nn Mi— li mnnr i r in ^K-f Mitft iwiri

PAGE 78

66 See Table 2 for a suminary of the verbal and nonverbal behaviors used to operationalize leader-like behavior for each of the experimental investigations. See Table 3 for a summary of the verbal and nonverbal behaviors used to operationalize influence for each experimental condition. The importance of each of these verbal and nonverbal behaviors in the creation of each experimental condition was not in the sheer num.ber of each behavior occurring during the group interaction for any given individual, but in the number performed by each group member relative to the number perform.ed by each other group m.ember. That is, a low level of leader-like behavior was operationalized in terms of there being nearly equal numbers of leader-like behavior (verbal and nonverbal) performed by each of the three group members. A high level of leader-like behavior was operationalized in terms of one group mem.ber (always group member three, the designated leader) performing a greater number of leaderlike behaviors (verbal and nonverbal) than either of the other two group members (who engaged in nearly equal numbers of those behaviors when compared to each other) A low level of influence was operationalized in terms of there being nearly equal levels of influence (in terras of the effect any one group member's behavior had on any other group miember s behavior) among all three group m^embers. Finally, a high level of influence v/as Mtijit^t>;^^i-p,ja.TP.^ i *< ; .i^i._:j;^-jqtCT> > i y"j a ;:^

PAGE 79

67 Table 2 Summary of the Verbal and Nonverbal Behaviors Used to Operationalize Leader-Like Behavior Factor Channel Behavior Verbal 1. Number of utterances per group member v/hich Content give suggestion, direction, implying autonomy for others (Bales' category four) Nonverbal Vocalic 2. Number of utterances per member which give opinion, evaluation, analysis, or express feeling or a wish (Bales' category five) 3. Number of utterances per group member v;hich give orientation, information, repeats, clarifies or confirms (Bales' category s.ix). 4. Number of speaking turns for each group member. 5. Number of words (over the entire conversation) spoken by each group member. 6 Numb>er of attempts by each group member to take the control of the floor away from another speaker for him.self (an interruption attempt) Nonverbal Kinesic Number of attempts by each group miember to indicate the next speaker by m.eans of a hand gesture. 'an^; a >. ; )>*ajiqmv ^W n wiu*ii-a.i'w.

PAGE 80

68 Table 3 Suimnary of Verbal and Nonverbal Behaviors Used to Operationalize Influence Factor Channel Behavior Verbal 1. Number of verbal agreements (indications of Content agreement, passive acceptance, understanding, concurrence, com.pliance) received by each group member following their own utterance of a suggestion, opinion, etc. (e.g. reception of a Bales' category three from another group member for a Bales' category four, five, or six utterance). The lack of verbal disagreemients (Bales category ten and twelve) for an utterance coded as Bales' four, five or six. Nonverbal 2. Number of successful attempts by each group Vocalic m.eiriber to interrupt, to take the control of the floor away from another speaker. Nonverbal 3, Number of successful attempts by each group Kinesic member to indicate the next speaker by means of a hand gesture. 4 Number of body posture shifts by each group member which are m.atched by one or more other group members 5. Number of hand gestures by each group member which are matched by one or m.ore other group members. 6. Numbier of nonverbal agreements (head nods) received by each group m.ember from other group members following an utterance coded as Bales' four, five, or six. The lack of a nonverbal disagreement (head shake) for an utterance coded as a Bales' four, five or six.

PAGE 81

69 operationalized in terms of one group member (always group member three, the designated leader) having more influence over the other two group members (numbers one and two) than either of them have over other group members Thus, scripts for each experimental condition were created by taking passages from the working script, but m.aking systematic alterations in the specific verbal or nonverbal behavior, or making changes with regard to which group member engaged in a specific verbal or nonverbal behavior, such that the experimental conditions were created. Thus, condition one interaction (low levels of leader-like behavior and low levels of influence) was scripted such that group members one two and three each engaged in nearly equal numbers of leader-like behavior and each had nearly equal influence over other group members. This condition then had no designated leader. The interaction for condition tv7 was scripted such that all group members engaged in nearly equal numbers of leader-like behavior (low levels of leader-like behavior) but group member three had more (nearly two times more) influence over the other group m.embers through the scripted verbal and nonverbal behaviors, than group members one and two had (high levels of influence) The condition three interaction (high levels of leaderlike behavior, low levels of influence) was scripted such that group mem.ber three engaged in more (nearly two times more) leaderlike behavior than did group members one and =^5a(=' -j i 1 7j, ^ Ml f ^ ln ^ wi^i je^g?&Lii^a^; > rflt^N t,ih U* s

PAGE 82

70 two. However, all three group members were nearly equal in the influence they exerted over other group members. Finally, the interaction for condition four (high levels of influence and high levels of leader-like behavior) was scripted such that group member three exerted greater influence (nearly two tim.es more) over other group members than either of them exerted and group member three also engaged in more (nearly two times m.ore) leader-like behaviors than did members one and two. See Tables 4 through 7 for a summary of the amounts of leader-like and influence behaviors obtained for each group member for conditions one through four respectively. Thus, in the creation of each script for the experimental conditions, the script was written and repeatedly edited until group members one, two and three each had the condition-appropriate leader-like behaviors and level of influence relative to each other (along both verbal and nonverbal channels) For reliability, a trained coder as well as the experimenter coded each turn or utterance (on occasion there was more than one utterance or thought within a turn) for each group miember for each condition, according to the Bales system. All instances of disagreement over the categorization of utterances or turns were resolved for the final script. Thus the script represented a high degree of agreemient regarding Bales category assignments.

PAGE 83

71 Table 4 Suiranary of Frequencies and Percentages for Behaviors Indicative of Level of Leader-Like Behavior and Level of Influence for Group Members One, Two and Three for Condition One (low levels of leader-like behavior, low levels of influence) Group Member Factor Behavior Leader-Like Number of utterances coded Behavior as Bales 4,5,6 26 23 22 % of each group member's utterances coded as Bales 4,5,6 62% 62% 63% Number of speaking turns 25 22 23 % of speaking turns for each group member out of the total number of turns by all 36% 31% 33% Number of words spoken 422 439 430 % of words spoken by each group member out of the total number of words spoken by all 33% 34% 33% Number of interruption attempts 5 4 4 % of interruption attempts for each group m.ember out of the total number of attempts made by all 38% 31% 31% Number of turn indicating gestures 2 2 2 % turn indicating gestures for each group member out of the total number of such gestures by all 33% 33% 33%

PAGE 84

Table 4--continued, 72 Group Member Factor Behavior Influence Number of Bales category 3 (agreements) received in response to making a Bales 4,5,6 utterance % of agreement (Bales 3) received when Bales 4,5,6 utterances were made 275 26% 36% Number of successful interruptions 2 % successful interruptions for each group member out of the total number of interruption attempts 40% Number of head nods (agreements) received in response to making a Bales 4,5,6 utterance 3 50% 75% % of agreements (head nods) received when Bales 4,5,5 utterances vzere made 125 22% 36% Number of hand gestures matched 1 % hand gestures matched for each group member Number of body shifts matched josture % body posture shifts matched for each group member 100% 67^ 33% Number of successes at indicating next speaker turn by gesture 1 % of successes for each individual for indicating next speaker turn by gesture 50% 50% 50%
PAGE 85

73 Table 5 Summary of Frequencies and Percentages for Behaviors Indicative of Level of Leader-Like Behavior and Level of Influence for Group Members One, Two and Three for Condition Two (low levels of leader-like behavior, high levels of influence) Group Member Factor Behavior Leader-Like Number of utterances coded Behavior as Bales 4,5,6 24 24 26 % of each group member's utterances coded as Bales 4,5,6 57% 61% 68% Number of speaking turns 25 21 24 % of speaking turns for each group member out of the total number of turns by all 36% 30% 34% Number of words spoken 44S 453 442 % of words spoken by each group member out of the total number of words spoken by all 33% 34% 33% Number of interruption attempts 5 4 6 % of interruption attempts for each group member out of the total number of attempts made by all 33% 27% 40% Number of turn indicating gestures 3 3 3 % turn indicating gestures for each group member out of the total number of such gestures by all 33% 33% 33%

PAGE 86

74 Table 5 — continued, Group Member Factor Behavior Influence Nuiriier of Bales' category 3 (agreements) received in response to making a Bales 4,5,6 utterance 5 7 15 % of agreement (Bales 3) received when Bales 4,5,6 utterances were made 21% 29% 58Number of successful interruptions 2 2 5 % successful interruptions for each group member out of the total number of interruption attempts 40% 50% 83 Number of head nods (agreements) received in response to making a Bales 4,5,6 utterance 4 4 13 % of agreements (head nods) received when Bales 4,5,6 utterances were made 17% 17% 50% Number of hand gestures matched 2 % hand gestures matched for each group memLper Number of body posture shifts matched 113 % body posture shifts matched for each group member 50% 33% 100% Number of successes at indicating next speaker turn by gesture 112 % of successes for each individual for indicating next speaker turn by gesture 33% 33% 67% itM^UMSjt^ra'aRi

PAGE 87

75 Table 6 Summary of Frequencies and Percentages for Behaviors Indicative of Level of Leader-Like Behavior and Level of Influence for Group Members One, Two and Three for Condition Three (high levels of leader-like behavior, low levels of influence) Group Member Factor Behavior Leader-Like NumJoer of utterances coded Behavior as Bales 4,5,6 21 14 36 % of each group member's utterances coded as Bales 4,5,6 52% 41% 71% Number of speaking turns 24 20 28 % of speaking turns for each group meirJoer out of the total number of turns by all 33% 28% 39% Number of words spoken 323 312 626 % of words spoken by each group member out of the total numJoer of words spoken by all 26% 25% 50% Number of interruption attempts 6 4 8 % of interruption attempts for each group member out of the total numiber of attempts made by all 33% 22% 44% Number of turn indicating gestures 3 % turn indicating gestures for each group memiber out of the total number of such gestures by all 0% 0% 100% H w y *^M Aij t-,-a>.^tf;K.< W J..

PAGE 88

76 Table 6~-continued Group Member Factor Behavior Influence Nuitiber of Bales' category 3 (agreements) received in response to m.aking a Bales 4,5,6 utterance 10 8 6 % of agreement (Bales 3) received V7hen Bales 4,5,6 utterances were made 48% 57% 17% Number of successful interruptions 5 2 4 % successful interruptions for each group merriber out of the total number of interruption attempts 8 3% 50% 50% Number of head nods (agreements) received in response to making a Bales 4,5,6 utterance 4 5 4 % of agreements (head nods) received when Bales 4,5,6 utterances were made 19% 36% 11% Number of hand gestures matched 12 1 % hand gestures m.atched for each group member — Num.ber of body posture shifts m.atched 111 % body posture shifts matched for each group member 100% 100% 20% NumJDer of successes at indicating next speaker turn by gesture % of successes for each individual for indicating next speaker turn by gesture 0% 0% 0%

PAGE 89

77 Table 7 Summary of Frequencies and Percentages for Behaviors Indicative of Level of Leader-Like Behavior and Level of Influence for Group Members One, Two and Three for Condition Four (high levels of leader-like behavior, high levels of influence) Group Member Factor Behavior Leader-Like Nuinber of utterances coded Behavior as Bales 4,5,6 16 18 33 % of each group member's utterances coded as Bales 4,5,6 43% 69% 70% Number of speaking turns 25 16 24 % of speaking turns for each group member out of the total number of turns by all 38% 25% 37% Number of words spoken 330 294 605 % of words spoken by each group member out of the total number of words spoken by all 27% 24% 49% Number of interruption attempts 3 18 % of interruption attempts for e^ach group member out of the to1:al nur:)er of attempts m.ade by all 25% 8% 67% Numiber of turn indicating gestures 2 4 % turn indicating gestures for each group miember out of the total number of such gestures by all 0% 33% 67% wii M wwi— 1 1 nwiiiwii'^axw w% i ^ 1 1 w— "^--r— -^ -r-.--a..ji^--rK-^ ..^^ >>^. .m--—

PAGE 90

78 Table 7-~continued, Group Member Factor Behavior Influence Number of Bales category 3 (agreements) received in response to making a Bales 4,5,6 utterance 5 4 16 % of agreement (Bales 3) received v/hen Bales 4,5,6 utterances were made 31% 22% 48% Number of successful interruptions 2 6 % successful interruptions for each group member out of the total number of interruption attempts 66% 0% 75% Number of head nods (agreements) received in response to making a Bales 4,5,6 utterance 6 1 13 % of agreements (head nods) received when Bales 4,5,6 utterances were made 37% 5% 39% Number of hand gestures matched 4 % hand gestures matched for each group member — — — Number of body posture shifts matched 4 % body posture shifts matched for each group member 0% 0% 100% Number of successes at indicating next speaker turn by gesture 13 % of successes for each individual for indicating next speaker turn by gesture 0% 50% 75% ••lirMWiiiHit^^uji^Miiiii-tiJtfct

PAGE 91

79 After the four scripts were completed, student actors from the theater department at Cleveland State University were recruited. Initially attempts were made to use male actors who resembled each other along several physical characteristics. However, the three actors who were finally used in creation of the four interactions differed a great deal from each other in physical appearance. However, all three were experienced actors. The actors were given a script and a few weeks of rehearsal time in order to learn the scripted verbal and nonverbal behaviors. They were told to learn the interaction exactly as scripted. Once the actors appeared to have an acceptable level of performance in rehearsal each interaction was videotaped in the campus television studio. Technical aspects of videotape production such as camera position and shots, lighting and audio levels were held constant within and across recordings. Each interaction was filmed against a backdrop of paneling and bookcases. The three actors maintained the same seating positions across all conditions and they were arranged in chairs in a semi-circular fashion around a small round coffee table. Thus, all three were visible in nearly face-on positions. All actors dressed similarly (in a casual manner) and each actor wore the same clothing across all tapings Each interaction began with a fade from black and one actor saying, "The problem is parking." For the first -n-y^'i -^fmivta^mpt

PAGE 92

80 thirty seconds of the interaction numbers were imposed on the tape to appear at the bottom of the television monitor screen identifying actors as group members one, two and three. The interaction proceeded until its completion about seven minutes later. Approximately fifteen seconds before the end of the interaction, the numbers were imposed on the tape to reappear on the television screen, again identifying each actor by his number; one, two or three. This v;as done to aid subjects in identifying each actor. Imposed on the videotape immediately prior to the start of the group interaction was a series of three still shots, one of each actor (sitting alone with a neutral expression) labeled by the appropriate numJaer at the bottom of the screen. These shots were used for the covariate measures (judgments of task and physical attraction based on physical appearance) Three to four takes of each condition's interaction were videotaped. Following completion of taping, the experimenter carefully reviewed each take to determine which one best fit the original script. Interactions never conformed exactly to the script. Actors found these roles very difficult to perform as precisely as the experimenter expected and inevitably some scripted behaviors were left out and other, non-scripted behaviors were included. In the process of selecting the best take for each experimental condition, a new, revised transcript was

PAGE 93

81 written (scripting what was actually said as opposed to the original plan of what was to be said) and counts of all relevant behavior (turns, words, interruptions, gestures, body shifts, head nods and counts of Bales' categories) for each actor were remade. Tables 4, 5, 6, and 7 reflect the behaviors which actually occurred for each videotaped interaction (for conditions one, two, three and four, respectively) as opposed to what was originally scripted. On the whole, deviations from the original scripts were relatively minor. At this point, each interaction was pre-tested to determine how natural of an interaction it appeared to be. Pre-test subjects (N = 59) were basic speech and psychology students at Cleveland State University and John Carroll University. Approximately fifteen students each watched one of the four conditions. They were given a three item, scale (see Appendix B) measuring perceived naturalness of the interaction. The scale scores could range from, a low of three (very unnatural) to a high of twenty-one (very natural) Mean ratings indicated each condition was perceived to be moderately natural (X for condition one = 9.8; X for condition two = 9.1; X for condition three II. 0; and X for condition four 9.4). Thus, these conditions fell within an acceptable range of naturalness for use in the present investigations. Because of the .rather large differences in physical appearance am.ong the three actor, group mem)ers and "',

PAGE 94

82 the impracticality of controlling the differences due to actor appearances in the design, a second pre-test v/as performed. This was done in order to confirm that a covariate analysis (covarying out effects due to differences in physical appearance of the actors from any effects due to the experimental factors) was necessary. Thus, another pool of pre-test subjects (N = 58) who were basic speech and psychology students at Cleveland State University and John Carroll University were presented with a series of scales designed to measure each of the following: (1) the extent to which the actor was perceived to be credible based on his physical appearance, (2) the extent to which the actor was judged attractive on social, task and physical dimensions based on his physical appearance, and (3) the extent to vrhich the actor was perceived as being leader-like again based on his physical appearance alone. Five dimensions of source credibility (competence, character, composure, extroversion, and sociability) were measured by a series of seven-interval, semantic differential type scales developed by McCroskey, Jensen, and Valencia (1973). The measures of task, social, and physical attraction used in the pre-test were developed by McCroskey and McCain (197 4) These v/ere a series of twelve (four per dimension) likert type, seven-interval scales. Three items written by the experimenter measured perceived leadership. Two items were seven-interval, likert type scales and the third used a seven-interval nOt^'^w^ariMI'W^I^MWllR

PAGE 95

83 semantic differential scale bounded by follower and leader. See copies of all pre-test scales for actor appearance differences in Appendix C. The following procedures for the pre-test were used. Subjects were shown a still shot of one of the three group members on a videotape television screen. While the picture remained on the screen, pre-test subjects were given the three scale questionnaire packet and instructed to respond. to each item based on the information given them on the television screen. A one way analysis of variance testing for differences due to actor appearance on each scale dimension was performed. Results indicated that differences between the actor-group m.embers were significant on all dimensions of each of the three scales. See Table 8 for a sumraary of the analysis of variance results. Thus, differences found in the pre-test analysis suggested that a covariate. analysis would have to be performed to control for effects due to differences in the physical appearance of the actors. Covariate measures The covariate scales used in the present study were the task and physical attraction scales developed by McCroskey and McCain (1974) These scales were the most theoretically relevant as covariates to the present study since differences in pre-test data were the result of physical appearance differences among the actors. The physical attraction scale most directly measures reactions to physical appearance differences. gltfi^4* >,t>ttti;tiii wfc -g w c^ wm 1 1 iM ffl i m tXv ^i a >-j ; aia TO ig

PAGE 96

84 Table 8 ANOVA Suiumary for Differences Between Actors Based on Physical Appearance on Pre-test Scales Sum of Factor Dimension Squares df F Credibility sociability 359.34 2 10.68** character 84.88 2 3.50* extroversion 344.33 2 8.98** composure 185.21 2 4.22* competence 292.55 2 12.44** Attraction physical 478.22 2 14.78** task 179.11 2 9.83** social 177.49 2 5.69** Leadership item one 16.66 2 4.80* item two 49.05 2 12.63** item three 27.04 2 11.16** .05 .01 mTn ir f~ i'' a^^T'iii' [MHg ir-^ niy"-a iii—ft*eitaMii r.^*;X^£jcia^ -^~ •--^ -i-r..

PAGE 97

85 The task attraction scale most directly deals with how an actor might be expected to behave in task situations, again based on physical appearance. The task dimension was relevant in this study since the group was a task oriented small group. These covariates were expected to equalize effects due to differences in physical appearance among the actors in the analyses. See Appendix D for a copy of the covariate measures. Dependent m.easures A packet of post-experimental measures was used to assess subjects' reactions to the videotaped group interaction. Subjects' reactions toward the group interaction in terms of their ratings of peri ; ceived group leadership were measured by a one item, seven-interval, likert type scale bounded by the descriptions: none/almost no leadership and extremely great amount of leadership Subjects were requested to rate each of the three group members on leadership on the three identical scales. Also included in the post-experimental package was a measure of how confident the subject felt regarding each of the leadership ratings they had just made (e.g., how confident the subject felt about their leadership rating for group member one, number two and number three) Thus, following the leadership rating scale for each group member, a one item, five-interval scale measuring confidence was inserted. This likert-type scale was bounded by the descriptions: not at all confident and very confident

PAGE 98

86 In addition to items measuring perceived leadership and the confidence associated with the judgment of leadership, an item measuring perceived influence was included. This item required the subject to rate each group member in terms of how much influence he appeared to have on either of the other group members during the course of the interaction. Thus, three rating scales, one for each group member, following a likert type format with a seven-interval scale was used. The scale was bounded by none/almost no influence and extremely great amount o f influence As with the leadership rating scales, subjects were requested to indicate how confident they felt about each influence rating they made. Confidence was measured here in the same manner as described above. Two other rating scale items were included in the post-experimental package. These were rating scales for sociability (by means of a five-interval, likert type scale bounded by the descriptions: very little sociability and extremely great amount of sociability ) and for pleasantness (again by means of a f ive-intervaL likert type scale bounded by the descriptions of not at all pleasant and v ery pleasant ) These items were included as extra scales in v/hich to embed the leadership and influence items Following the rating scales for leadership, influence, confidence, sociability and pleasantness, subjects were requested to indicate the group m.ember or £T.Vt>S3(-Ui ^wA|w

PAGE 99

87 inembers to whom they had given the highest leadership rating. Subjects were then presented with a thirty-two item adjective check list and were requested to check the fifteen items which were most central to their leadership ratina decisions for the individual most highly rated on leadership. Following this, subjects were instructed to rank order the five most important adjectives of those fifteen just checked. The purpose of these two measures was to determine whether subjects would select adjectives descriptive of individual traits, individual behaviors, or characteristics of the interaction as being the most important in their leadership judgments. These measures would also be used to determine whether different adjective descriptions ware selected for different conditions. The last page of the post-experimental packet contained a free response question asking the subject's opinion of the purposes of the study and a question asking the subject if they knew any of the three group members, and if so, in what way. Questions regarding the subject's sex, age, race, school of attendance, and whether they were native speakers of English were also asked. See Appendix E for a copy of the post-experimental packet. Study Two I Subjects Subjects were 264 students enrolled in basic speech and psychology courses at Cleveland State University and John Carroll University. Participation was on a credit given basis.

PAGE 100

88 Procedure The procedure for study two was nearly identical to that described for study one. Subjects were told that they would be exposed to a small group interaction and afterwards they would fill out a questionnaire concerning their reactions to that interaction. Subjects then read and filled out informed consent forms. Subjects were then exposed to the covariate stimuli and were requested to fill out the covariate scales (the task and physical attraction scales described in study one) This was the case for all subjects except those who were assigned to the transcript-only condition. These subjects received no covariate stimuli (since they would only be reading a transcript of the group interaction rather than viewing or hearing the interaction) The covariate scales were filled out by these subjects as part of the post-experimental packet (exposing them to the same questionnaires as all other subjects) All subjects were then exposed to some form of the group interaction. Afterwards, subjects received the post-experimental measures. Following completion of the post-experimental questionnaire participants were thanked for their participation and told that they would be informed of the purposes and results of the study at a later date. All subjects received a full explanation of the study at the completion of data collection. Desi gn. Subjects were randomly assigned to one of eight experimental conditions (N 264) comprising a 2

PAGE 101

89 (low and high levels of influence) x 4 (mode of presentation: audio-visual; visual-only; audio-only; transcriptonly) factorial design. The interactions used in study two were conditions one (low levels of influence; low levels of leader-like behavior) and two (high levels of influence; low levels of leader-like behavior) used in study one. Each of these two conditions were presented to subjects in all channel modes as described above. A summary of the design and conditions may be seen in Table 9 Mode' of presentation-communicative channels The data collected for audio-visual conditions (for both low and high levels of influence) in study two was that previously collected for study one (for conditions one and two of that study) The visualonly mode of presentation was created by exposing subjects to the group interaction on a videotape television screen with the audio portion turned off. The audio-only mode of presentation was created by tape recording onto cassette tapes the two interactions for low and high levels of influence (conditions one and two from study one) Preceding each group interaction on the cassetts tapes were recordings of a segment of each group member's voice, identified as being the voice of either group member one, two or three, followed by a series of six more unidentified voice segments. These voice segments were used in the following way. First, subjects "rT." JIL W| I *' .wi| fa.'ljr>^ i Tr.wt "i.'i ^-^t ;S5 ? j r? -.w -fT rw i 71 m mr B t, =i Mw m d l>, -vm t m -m n > m ^i -i ^iliMu r-i ^ ^ a f g ^ ^ rMr u i ^ i i iiii ^ i iM i mmi >

PAGE 102

90 o Eh O & -P u o 4-1 s:! O H JJ -H c O U -P 0) e (U W O u I I c o •H +J JJ 0) w 0) a. O 0) O S o I p H O c M EH C O I o H c o I H rd m > (d H > I O -H O 1-^ •H M-l U O C > >W (^ H •ilcV<^JI!-< :r <., v*cH, -_r— -—

PAGE 103

91 responded to each of the three identified voice segments in terms of the covariate measures (measures of physical and task attraction) Second, subjects were then told that it would be important for them to be able to distinguish each of the three group member's voices and that before they would hear the actual group interaction they would be given another opportunity to hear and identify each of the three voices. The experimenter then played a segment of group member one's voice and repeated, "this is student number one." It was then replayed and labeled again. The experimenter then continued this process of playing and labeling the segment of speech for group members tv70 and three. Subjects were then given pieces of paper and told to listen carefully to each of the following six, unidentified voice segments. They were requested to write down on the paper, after each segment, whether they thought they had heard the voice of group member one, two or three. Six voice segments were then played and subjects made their choices. The experimenter then listed the correct group member numbers for each of the voices heard. Following this the experimenter re-played the original three voice segments again labeling each voice with the appropriate number. At this point, subjects were exposed to the entire group interaction after which they received the post-experimental measures. The transcriptonly mode of presentation was developed by making a verbatim written transcript of the ."lirr— rii""ii fii'"mrwrtri iiinw^ iniiifi~iii m

PAGE 104

92 verbal content of the low influence and high influence conditions (conditions one and two from study one) These transcripts indicated which group member v/as speaking (numbers referring to the group m.ember one, two or three were placed in the left-hand column preceding the start of each speaking turn) the exact content of what was said (punctuation appropriate for the m.eaning was included) but did not include transcription of the nonverbal behavior of the gi'oup members. See Appendixes F and G for copies of these transcripts (for conditions one and two respectively) Dependent and covariate measures These were identical to those used in study one. [^ •gt l EiyM i5 g ag. r ^*i^ 5jrTfc 1.

PAGE 105

CHAPTER III RESULTS Test of Hypotheses for Study 1 Hypothesis one predicted a main effect for amount of influence on the leadership ratings for the designated group leader (group member three) such that high levels of influence are associated with high leadership ratings and low levels of influence are associated with lov/er leadership ratings. A 2 (low vs. high levels of influence) x 2 (low vs. high levels of leader-like behavior) analysis of covariance for unequal N with two covariates (task attraction and physical attraction ratings for the three group members) was performed to test for a main effect of level of influence on leadership ratings. Summaries for the analysis of variance, analysis of covariance, and the adjusted m.eans are provided in Tables 10 and 11, As predicted, a significant main effect for level of influence was obtained, F(l, 124) = 5.82, p<.02. An examination of the adjusted m.eans reveals that high levels of influence result in higher leadership ratings for group member three than do low levels of influence. (X for low levels of influence = 5.35; X for high levels of influence = 5.92). Thus, hypothesis one is supported. 93

PAGE 106

94 Table 10 ANOVA Summary for Leadership Ratings Source Sum of Squares df Mean Square F Influence 10.177 1 10.18 5.70* Leader-Like Behavior 4.431 1 4.43 2.48 Influence x Leader-Like Behavior 9.268 1 9.27 5.19* Physical Attraction Covariate 0.269 1 0.27 0.15 Task Attraction Covariate 1.998 1 1.99 1.12 Error 221.579 124 1.79 ,1 ANCOVA Summary for Leade rship Ratings Source Sum of Squares df Mean Square F Influence 10.402 1 10.40 5.82* Leader-Like Behavior 4.851 1 4.85 2.71 Influence x Leader-Like ; Behavior 7.107 1 7.11 3.98* 1 Physical Attraction Covariate 0.269 1 0.27 0.15 Task Attraction Covariate 1.428 1 1.43 0.80 Error 221.579 124 1.79 *£<.0 5

PAGE 107

95 Table 11 Summary of Adjusted Means for Leadership Ratings Effect M X Influence low levels 65 5.36 high levels 65 5.92 Leader-Like Behavior low levels 65 5.4 5 high levels 65 5.84 Influence x Leader-Like Behavior Low Levels of High Levels of Leader-Like Leader-Like Behavior Behavior Low Levels of X 4.92 X 5.79 Influence N 32 N 33 High Levels of X 5.97 X 5.88 Influence N 33 N 32 IT" iinnrii — h^t arin ni" 'i n\ vwssvhti •'iu o^^

PAGE 108

96 Hypothesis tv/o predicted a main effect for amount of leader-like behavior on leadership ratings for the designated leader (group member three) such that high levels of leader-like behavior are associated with high leadership ratings and low levels of leader-like behavior are associated with lower leadership ratings. The analysis described above for the test of hypothesis one was used to test hypothesis two as well. The F value for the main effect was nonsignificant, F(l, 124) = 2.71. Thus, this hypothesis was not supported. Table 10 summarizes the analysis of variance and analysis of covariance and Table 11 suimnarizes the adjusted means for this test of main effect for leader-like behavior. Hypothesis three predicted that the amount of influence main effect factor would account for a greater percentage of the variance than the leader-like behavior main effect factor. This hypothesis was tested by a statistic used to estimate the magnitude of experim.ental effects (Winer, 1971, pp. 428-430) The results of this analysis indicated that the hypothesis was supported. The main effect of the influence factor accounted for 3% of the variance while the main effect of the leader-like behavior factor accounted for only 1% of the variance. An interaction between amount of influence and amount of leader-like behavior was predicted by hypothesis four This hypothesis was tested by means of a 2 x 2 analysis of covariance for unequal N (described abov/e in jjiiiaiam 1 iifJ'.j.M'itw i^mgVa i''^itf>ii*inf;^(tftii**.^.

PAGE 109

97 the test for hypothesis one) See Table 10 for summaries of the analysis of variance and analysis of covariance and Table 11 for a surriTiary of the adjusted means on leadership ratings for group member three. As predicted, the interaction between amount of influence factor and amount of leader-like behavior factor was found to be significant, F(l, 124) = 3.98, p<.05. However, the order of the adjusted means deviated slightly from, that predicted. The order predicted was: cell one, cell three, cell two and cell four (listing them in order from, lowest m.ean to highest mean for leadership ratings on the designated leader).. The order obtained was (from lowest to highest mean): cell one (X = 4.92), call three (X 5.79), cell four (X --3.88), and cell two (X = 5.97). Note that all means are adjusted for the covariates. Hypothesis five predicted that a main effect would be obtained for group member such that the designated leader (so designated in cells two, three and four) would receive significantly higher leadership ratings than either of the other two group memJoers For each of the appropriate cells a one way analysis of covariance with two covariates (task attraction and physical attraction ratings for group members) was performed. For all three cells a significant main effect due to group member was obtained on leadership ratings. For cell two, F(2, 94) = 39 01 p< 0001 An inspection of the means adjusted for the covariates Ce=ai -^ iiJB> *J Bi jg'a>jL.'"' -J fm

PAGE 110

98 indicated that group member three, as predicted, received the highest leadership ratings (X for member three = 6,01; X for member two = 3.50; and X for miember one = 3.62) A pre-planned t test indicated differences between the means for members three and one were significant, t(94) = 7.47, and differences between the means for members three and two were significant, t(94) = 7.84, each at p<.01. For cell three, F(2, 94) = 32.47, p<.0001. Examination of the adjusted mieans reveals that it was again member three who received the highest leadership rating: X for member three = 5.79; X for member two = 2.93; X for member one = 4.77. The differences betv/een the means for members three and one and for mem.bers three and two were each significant as determined by a pre-i^lanned t test. For the comparison between members three and one, t{94) = 2.95, p<.01, for the comparison between members three and two, t(94) = 8.28, p<.01. Finally, for cell four, F(2, 91) = 21.76, p<.0001. Inspection of the means again reveals that group mem.ber three, as predicted, received the highest leadership ratings: X for member three = 5.84; X for m.ernJjer two = 3.46; and X for member one = 3.04. The differences, as determined by a pre-planned t test, between the adjusted m.eans for group memibers three and one and three and two, were each significant, t(91) = 6.73, and t(91) = 5.72, p<.01, respectively. Thus, for each of the three cells in which a member was designated as being either more influential -,^^-.^.ifp..;— ^..f.^..rfiij|.^-^ J., ^.— ...-,—~f y..-w^.>.... -..^ — i-r --iii-n-r-T-fwt~
PAGE 111

99 than other group members, engaged in more leader-like behaviors than other group m.embers, or was both more influential and engaged in more leader-like behaviors than other group members, that person was in fact perceived by observers of the group interaction to show significantly more leadership than other group m.embers. Thus, hypothesis five was supported. See Table 12 for a summary of the analysis of covariance performed for cells two, three and four and Table 13 for a summary of the adjusted means for leadership ratings for group members within these • cells. An unexpected significant main effect for group member was obtained in cell one in which no leader was designated, F(2, 91) = 3.55, p<.05. This analysis was not planned but was performed as a supplementary analysis at the time of the analyses for cells two, three and four described above.. Inspection of the adjusted means reveals that group member one received the highest leadership rating (X for member one = 4.9 8), followed by member three (X = 4.83), with member two receiving the lowest leadership rating (X = 4.09). At test for differences between the means revealed significant differences between the adjusted means for group members one and two, t(91) = 2.49, and between group members three and two, t(91) = 2.07, D<.05. Tables 14 and 15 present sunmaries of the covariance analysis and the adjusted means for leadership ratings for group members within cell one. ^^mIIIimT' H' l^ t^s Uii^ tn \m-yi r ^ m ii— rya i r 1 1 | r i^a r-^ nBF=ajC r=nii^i: T^, T f^).j ^.> tViH i r^ l m — M1J1 mp iw f ii i nnimrtn nim^ vnr tmr f.-im> ^ U _mtim r ir.-i f i i .'TT i 1 i rt TH fcrrt

PAGE 112

100 Table 12 ANCOVA Suriimary for Leadership Ratings for Group Members Within Cells 2, 3 and 4 Cell 2 Sum of Mean Source Squares df Square P Group Member 131.767 2 65.88 39.01*** Physical Attraction Covariate 2.287 1 2.28 1.35 Task Attraction Covariate 0.116 1 0.12 0.07 Error 158.738 94 1.69 Cell 3 Sum of Mean Source Squares df Square P Group Member 127.848 2 63.92 32 47*** Physical Attraction Covariate 0.001 1 0.00 .00 Task Attraction Covariate 2.134 1 2.13 1 .08 Error 185.048 94 1.97 •*ftw;^H'*

PAGE 113

101 Table 12 — continued, Cell 4 Source Sum of Squares df Mean Square F Group Member 120.717 2 60.36 21.76*** Physical Attraction Covariate 1.248 1 1.25 0.45 Task Attraction Covariate 2.152 1 2.15 0.78 Error 252.428 91 2.77 ***p<.0001 ViV -'i'tllB3'1-'MNP,4M>B^FliUKr 4 M f t iiin rtiifjii,*a^-i if ^Hi^ }j im^m i ^n. >. l^RfU>Aii nit'^V4llKrA4l"**MRMBj£*^ i*-Mi!l^ •*)'<

PAGE 114

102 Table 13 Summary of Adjusted Means for Leadership Ratings for Group Members Within Cells 2, 3 and 4 Factor N X Cell 2: high influence; low leader-like behavior Group Member 1 33 3.62 2 33 3.50 3 33 6.01 Cell 3: low influence; high leader-like behavior Group Member 1 33 4.77 2 33 2.93 3 33 5.79 Cell 4: high influence; high leader-like behavior Group Member 1 32 .3.04 2 32 3.46 3 32 5.84

PAGE 115

J Table 14 ANCOVA Summary for Leadership Ratings for Group Members Within Cell 1 *p<.0 5 103 Sum. of Mean Source Squares df Square F Group Member 14.479 2 7.24 3.55* Physical Attraction Covariate 1.363 1 1.36 0.42 Task Attraction Covariate 0.362 1 0.36 0.67 Error 185.621 91 2.04 inrMi*'***^— *:'~*--*'-ifc''**.* ••*••<.. -.•._-i.^-N.-i **•.-uirf— i_*t..ii.i-'(ti->^ii.^-.,

PAGE 116

104 Table 15 Summary of Adjusted Means for Leadership Ratings for Group Members Within Cell 1 Factor N Cell 1: low influence; low leader-like behavior Group Member 1 32 4.98 2 32 4.09 3 32 4.83

PAGE 117

105 Suppleiuentary analyses for study 1 A Pearson product moment correlation was used to determine the extent to which the leadership ratings for group mem.ber three and influence ratings for group member three covaried. The analysis revealed a significant correlation, r = .49, p<.01. Thus, a moderate relationship between leadership and influence ratings for the designated leader was obtained. A 2 (high versus low levels of influence) x 2 (high versus low levels of leader-like behavior) analysis of covariance for unequal N with two covariates (ratings of group member task attraction and physical attraction) was perform.ed in order to determ.ine whether main effects (for level of influence and level of leader-like behavior) and an interaction effect for the two factors would be obtained on ratings of influence for the designated leader. The analysis revealed a significant main effect for level of influence, F(l, 124) = 10.78, p<.01 as well as a significant main effect for level of leader-like behavior, F(l, 124) = 7.39, p<.01. An inspection of the adjusted m.eans for influence ratings indicates that high levels of influence are associated with higher influence ratings than are low levels of influence (X for low levels of influence = 5.40; X for high levels of influence = 6.11). Further, high levels of leader-like behavior are associated with higher influence ratings than are low levels of leader-like behavior (X for low levels of leader-like aaaj CTi If' jrfrMr=

PAGE 118

106 behavior = 5.46; X for high levels of leader-like behavior = 6.05). The analysis also revealed a significant interaction effect for level of influence by level of leaderlike behavior, F(l, 124) = 5.95, p<.02. Ordering of the adjusted means for influence ratings (from lowest to highest) is as follov7s: cell 1 (X = 4.84); cell 3 (X = 5.97) ; cell 2 (X = 6.08) ; and cell 4 (X = 6.13) See Table 16 for a summary of the analysis of variance and analysis of covariance and Table 17 for a summary of the adjusted means. After watching the videotaped group interaction, each subject was required to assign a leadership rating to each of the three group membersFollowing each leadership rating, subjects were asked to rate how confident they felt about the preceding leadership rating. In an effort to determine whether there were different degrees of confidence associated with the factors of influence or leader-like behavior, a 2 x 2 analysis of covariance (as described above regarding the influence ratings) was performed. The analysis revealed significant main effects for level of influence, F(l, 124) = 5.62, £<.02, and for level of leader-like behavior, F(l, 124) = 6.95, p<.01. No significant interaction effect was obtained. Inspection of the adjusted means indicated that subjects were more confident of their leadership ratings where influence was high than where it was low (X for low levels of influence ~ 4.17; X for high levels of influence = 4.49) and

PAGE 119

107 Table 16 ANOVA Summary for Influence Ratings ) Source Sum of Squares df Mean Square F Influence 15.956 1 15.95 10.65** Leader-Like Behavior 9.423 1 9.42 6.29* Influence x Leader-Like Behavior 13.236 1 13.23 8.84** Physical Z^ttraction Covariate 4.807 1 4.80 3.21 Task Attraction Covariate 2.479 1 2.47 1.66 Error 185.706 124 1.49 ANCOVA Summary for Influence Ratings Sum of Mean Source Squares df Square F Influence 16.148 1 16.14 10.78** Leader-Like Behavior 11.068 1 11.06 7.39** Influence x Leader-Like Behavior 8.916 1 8.91 5.95* Physical Attraction Covariate 4.807 1 4.80 3.21 Task Attraction Covariate 0.722 1 0.72 0.48 Error 185.706 124 1.49 *p<.05 **D<.01 E^c^r^Tca^n^;^sr>ltf^a. ip-jimi.,.vhw-.j>. .^B=M^^r^nr '--<^ttt(irm:sai£S!tM.9fj*'^,>^u>rf'^*f^f'''=^^ i -:-? J7t^^a
PAGE 120

108 Table 17 Summary of Adjusted Means for Influence Ratings Effect Influence low levels high levels N 65 65 X 5.40 6.11 Leader-Like Behavior low levels high levels 65 65 5.46 6.05 In fluence x Leader-Like Be'havior Low Levels of Influence High Levels of Influence Low Levels of Leader-Like Behavior X 4.84 N 32 X 6.08 N 33 High Levels of Leader-Like Behavior X 5.97 N 33 X 6.13 N 32

PAGE 121

10 9 when leader-like behavior was high (X = 4.51) than when leader-like behavior was low (X = 4.17) Refer to Tables 18 and 19 for summaries of the analysis of covariance and adjusted means for confidence ratings. Subjects were asked to indicate (by means of checking and ranking procedures) which adjectives they felt best identified the reasons behind their leadership rating choices. Subjects were given a list of thirty-two adjectives or descriptions for group member behavior or personality. Ratings for any given item could have ranged from one to seven. However, in no cell did an adjective ever receive a higher mean rating than 3.48. Further, only three adjectives out of the thirty-two received mean ratings of at least 3.0 on at least one occasion (these adjectives were: is self-confident, is dominating, is aggressive). Thus, subjects did not appear to be basing their leadership ratings within any given cell or even across cells on the same factors. Only one adjective received a relatively high mean score in each of the four cells. That description given to the individual receiving the highest leadership rating was: "is self-confident." The mean rating for this adjective in each of the cells is as follows: cell one, X = 2.90; cell two, X = 3.48; cell three, X = 3.12, and cell four, X 3.18. Only two other adjective descriptions, having a mean rating of 2.5 or better, were chosen to account for subjects' leadership ratings in more than one cell. These adjectives were: >
PAGE 122

110 Table 18 ANOVA Summary for Confidence Ratings Sum of Mean Source Squares df Square F Influence 3.173 1 3.17 5.92* Leader-Like Behavior 3.076 1 3.07 5.74* Influence x Leader-Like Behavior 0.084 1 0.08 0.16 Physical Attraction Covariate 3.328 1 3.32 6.20* Task Attraction Covariate 0.928 1 0.92 1.73 Error 66.515 124 0.53 ANCOVA Summary for Confidence Ratings Sum of Mean Source Squares df Square F Influence 3.012 1 3.01 5.62* Leader-Like Behavior 3.726 1 3.72 6.95** Influence x Leader-Like Behavior 0.036 1 0.03 0.07 Physical Attraction Covariate 3.328 1 3.32 6.20* Task Attraction Covariate 2.142 1 2.14 3.99* Error 66.515 124 0.536 *p<.05 **p<.01

PAGE 123

Ill Table 19 Suirnnary of Adjusted Means for Confidence Ratings Effect N X Influence low levels 65 4. 17 high levels 65 4. 49 Leader-Like Behavior low levels high levels In fluence x Leader-Like Behavior Low Levels of Influence High Levels of Influence 65 4.17 65 4.51 Low Levels of High Levels of Leader-Like Leader-Like Behavior Behavior X 3.99 X 4.37 N 32 N 33 X 4.34 X 4.64 N 33 N 32 f^ i j:inrv aigj r at. T .w>^iMi^Mca-t' j ii .^ jlm ^

PAGE 124

112 "is dominating" and "is aggressive." For cell three, "dominating" received a mean rating of 3.42; in cell four that adjective received a m.ean rating of 2.75. For cell three, "aggressive" received a mean rating of 2.57; and for cell four, that adjective received a mean rating of 3.34.. Thus, there is low consistency across subjects (both within cells and across cells) with regard to their reasons for m.aking their leadership ratings at least as far as this adjective check list measure indicates. Subjects were asked to rate each group member on sociability as well as pleasantness rating scales. Ratings for both sociability and pleasantness were obtained only because it was deemed desirable m.ethodologically to eirJoed the dependent measure (leadership rating scales) and the other measures of interest (influence and confidence ratings) in other items. Thus, these ratings were of no interest in this study and consequently no analyses for either factor were performed. Test of Hypotheses for Study 2 Hypothesis 1 predicted a main effect for group member within condition such that group m.ember three would receive higher leadership ratings than group members one and two in each of the experimental conditions. Thus, the designated leader was expected to be identifiable by both verbal and nonverbal cues acting alone or in combination. To test this hypothesis, a one-way analysis of variance ggsnla^^saf mmir,nii—ni-—

PAGE 125

113 was performed comparing leadership ratings for each of the three group members within each condition. Findings for each of the four conditions will be presented separately below. For condition 1 (audio-visual condition providing verbal content as well as nonverbal kinesic and vocalic information to the observer subjects) a significant F value V7as obtained for differences in leadership ratings among the three group merrbers; F(2, 96) 39.39, p<.0001. Examination of the means reveals that as predicted, group member three, the designated leader, received the highest leadership ratings (X for group merrtoer three = 6.00; X for group member two = 3.4 2; X for group member one = 3.69) Differences between the mean ratings for group members one and three and for two and three were each significant, £<.01, as determined by a pre-planned t test, t(96) 7.24, and t(96) = 8.09, respectively. For condition 2 (visual-only, providing nonverbal kinesic information) a significant F value was obtained for differences in leadership ratings for group members, F(2, 93) = 32.43, p<.0001. An examination of the means for leadership ratings of each group member indicates that it was group member three, as predicted, who received the highest leadership ratings (X for group member three = 6.19; X for group member two = 3.50; X for group member one = 4.44). Differences between the mean ratings for group members one and three and two and' three were each significant, p<.01, as determined by a t test, t(93) = -.->-— a.f^r-.w.— —--^---

PAGE 126

114 5.16 for differences between means for members one and three, t(9 3) = 7.9 3 for differences between means for members two and three For condition three, the F value was nonsignificant, F(2, 90) = 1.07. Subjects were not able to significantly distinguish the designated leader (group meml^er three) from the other two group members as indicated by their leadership ratings for the group m.embers (X for group member three = 4.93; X for group mem.ber two = 4.39; and X for group member one = 4.90) for this audio-only condition (providing verbal content plus nonverbal vocalic information) Finally, for condition four (transcript, providing verbal content information) a significant F value was obtained for leadership ratings, F(2, 102) = 20.38, p< .0001. Inspection of the means reveals that group member three (the designated leader) was, as predicted, given the highest leadership ratings (X for group member three = 5.34; X for group member two = 4.37; X for group member ^^3 ^ 3.20). Differences, as determined by a pre-planned t test between the m>ean leadership ratings for group members one and three and two and three were each significant, p<.01, t(102) = 6.37, t(102) 2.89 for differences between one and three and two and three respectively. Thus, in summary, hypothesis one was only partially supported. The hypothesis was supported for conditions one, two and four. The designated group leader ^i^mp>'-^H*^0^,jk^m ^-W— ti^^Wrf tj—jl f

PAGE 127

115 received significantly higher leadership ratings than group members one and two when observers were given audiovisual (full information along both verbal and nonverbal channels) information in condition one; when they were given visual information only in condition two (nonverbalkinesic) ; and when they were given transcript information only (verbal content) in condition four. The hypothesis was not supported for condition three where subjects v;ere given audio-only (verbal content plus nonverbal vocalic) information. See Table 20 for a summary of the analysis of variance for each condition and see Table 21 for a summary of the mean leadership ratings for group memibers for each condition. Hypothesis 2 predicted a main effect for channelcondition on leadership ratings for group member three (the designated leader) specifically predicting that leadership ratings would be higher in condition 1 (audiovisual) than in condition 2 (visual only) condition 3 (audio only) or condition 4 (transcript only) To test this hypothesis, a 2 (low vs. high levels of influence) x 4 (audio-visual vs. visual-only vs. audio-only vs. transcript-only channels of cominunication) analysis of variance for unequal N on leadership ratings for group member three was performed. The F value for a main effect due to channel-condition was nonsignificant, F(3, 256) = 0.69. See Table 22 for a summary of the analysis of variance and

PAGE 128

Table 20 116 ANOVA Summary for Leadership Ratings for Group Members Within Conditions .^ Condition 1 Audio-Visual Source Sum of Squares df Mean Square F Group Member Error 132.1414 161.030 2 96 66.07 1.68 39. 39*** Condition 2 Visua 1 Only Source Sum of Squares df Mean Square F Group Member Error 119.0833 170.7500 2 93 59.54 1.84 32 .43*** Condition 3 Audio Only Source Sum of Squares df Mean Square 1 p Group Member Error 5.8709 245,9354 2 90 2.93 2.73 .07 Condition 4 T ranscr ipt Only Source Sum of Squares df Mean Square F Group Member Error 80.5904 201.6571 2 102 • 40.29 1.98 20 .38*** P. 1 J .I. ^— J^ ir-.^^-m— i.g|HHl*Mj'**.-T't'

PAGE 129

117 Table 21 Summary of Means for Leadership Ratings for Group Members for Each Condition Factor N X Condition 1 Audio-Visual Group Member 1 33 3.69 2 33 3.42 3 33 6.00 Condition 2 Visual Only Group Member 1 32 4.44 2 32 3.50 3 32 6.19 Condition 3 Audio Only Group Member 1 31 4.90 2 31 4.39 3 31 4.93 Condition 4 Transcript Only Group Member 1 35 3.20 2 35 4.37 3 35 5.34 --— JT-' -Tji l h l h Yr — — ^ -^"^ •Tr^ *f."T ; "> -JTT T I-r f \ t — -" '•^-' ^.— ^-t. t:.^-^:-^-..*-^!!!*-

PAGE 130

118 Table 2 2 ANOVA Summary for Leadership Ratings for Group Member Three Sum of Mean Source Squares df Square F Channel-Condition 4.5181 3 1.51 .69 Level of Influence 36.4024 1 36.40 16 .67*** Channel x Level of Influence 50.3201 3 16.77 7 .71*** Error 559.1431 256 2.184 ***£<. 0001 I I — lln'rH linaM~fc< Mill -r -yi I ''' -I..'.*..iT^ — -----l.--=-= tu^pa^a

PAGE 131

119 Table 23 for a suirmary of the means for leadership ratings for group member three for this factor. A significant main effect for level of influence on leadership ratings for group member three was obtained, F(l, 256) = 16.67, p<.0001. Examination of the means indicates that high levels of influence (X = 5.62) led to higher leadership ratings than low levels of influence (X = 4.88). See Table 22 for a sumimary of the analysis of variance and Table 2 3 for a summary of the means for leadership ratings for this factor. Additionally, this analysis of variance produced a significant F value on leadership ratings for a level of influence by channel-condition interaction effect, F(3, 256) = 7.71, p<.0001. See Table 22 for a sumnaary of the' analysis of variance. Level of influence interacted v;ith channel-condition such that the highest leadership means were received for high levels of influence in the audiovisual and visual-only conditions and the lowest leadership ratings were obtained for the low levels of influence in the audio-visual and visual-only channel-conditions. So subjects were more likely to perceive leadership when influence level was high in the audio-visual and visualonly conditions. See Figure 1 for a visual presentation of the interaction and Table 2 3 for a summary of the means for the leadership ratings.

PAGE 132

120 Table 23 Summary of Means for Leadership Ratings for Group Member Three Factor N X Level of Influence low 133 4.88 high 131 5.62 Channel-Condition 1 audio-visual 65 5.45 2 visual-only 66 5.16 3 audio-only 67 5.10 4 transcript-only 66 5.24 Level of Influence x Channel-Condition Level of Influence Channel-Condition 12 3 4 low X 4.90 X 4.20 X 5.25 X 5.13 N 32 N 34 N 36 N 31 high X 6.00 X 6.19 X 4.93 X 5.34 N 33 N 32 N 31 N 35

PAGE 133

121 Mean Leadership Ratings 7 .6 • 5 .. A -l 2 1 • Hiqh Influence Condition Low Influence Condition 1 2 3 AudioVisualAudio visual only only 4 Transcriptonly Channel-Condition Figure 1 Interaction of Influence and Channel for Leadership Ratings

PAGE 134

12: Supplementary analyses for study 2 A Pearson product moment correlation was used to determine the extent to which leadership ratings for group mem.ber three and influence ratings for group member three covaried. The analysis revealed a significant correlation, r = .74, p<.01. Thus, a moderately high relationship between designated leader leadership ratings and influence ratings across all channel-conditions was obtained. A 2 (low vs. high levels of influence) x 4 (audiovisual vs. visual-only vs. audio-only vs. transcript-only channel-conditions) analysis of variance for unequal N was performed in order to determine whether a main effect for channel-condition on influence ratings for the designated leader was obtained. Sim.ilar to the results obtained in the test of hypothesis two on leadership ratings, a nonsianificant effect for channel-condition was obtained for influence ratings, F(3, 255) = .83. A significant m.ain effect for level of influence was obtained, F(l, 256) = 16.50, p<.0001, and a significant interaction between level of influence and channel-condition was also obtained, F(3, 256) = 7.45, p<.0001. For the level of influence main effect, examination of the m.eans for influence ratings indicates that the designated leader received higher influence ratings v/hen level of influence was high (X = 5.67) than when level of influence was low (X = 4.97). With regard to the interaction effect due to level of influence and channel-condition, the highest influence p4—aen.vi. f _!' : ii~"i fc "* — rj^ j TJ w Mg j a w oiw ^eo^wn cTKgTi ":> m

PAGE 135

123 ratings were associated with high levels of influence and channel-conditions one and two (audio-visual and visualonly) and the lowest levels of influence ratings were associated with low levels of influence and channel-conditions one and two. Thus, the perception of influence was dependent upon both the level of influence as well as the channel along which that influence was conveyed. See Table 24 for a summary of the analysis of variance for influence ratings and Table 25 for a summary of the influence rating means. Figure 2 provides a visual presentation of the interaction effect just discussed. Finally, the same analysis as that described above ^as performed in order to see whether significant main effects for channel-condition and level of influence and an interaction effect for channel by level of influence would be obtained for the subjects' ratings of how confident they felt about their leadership ratings for group m.ember three. The primary concern here was v;hether channels differentially affected subjects' confidence for their leadership ratings. The F values for both main effects and the interaction were all nonsignificant. Mean ratings indicated subjects were moderately confident across all levels of influence and across all channelconditions. See Table 26 for a summary of the analysis of variance and Table 27 for a summary of the means for confidence ratings.

PAGE 136

124 Table 24 ANOVA Suramary for Influence Ratings for Group Member Three Sum of Mean Source Squares df Square F Channel-Condition 4.8210 3 1.61 .83 Level of Influence 31.8092 1 31.81 16.50*** Channel x Level of Influence 43.0880 3 14.36 7.45*** Error 493.5558 256 1.93 ***p<.0001 •roi •)>.— .fr"r'-

PAGE 137

125 Table 25 Summary of Means for Influence Ratings for Group Member Three Factor N X Level of Influence 133 4.97 low high 131 5.67 Channel-Condition 1 audio-visual 65 5.49 2 visual-only 66 5.24 3 audio-only 67 5.13 4 transcript-only 66 5.41 Level of Influence x Channel-Condition Level of Influence Channel-Condition 12 3 4 low, X 4.81 X 4.44 X 5.16 X 5.48 N 32 N 34 N 36 N 31 high X 6.15 X 6.09 X 5.09 X 5.34 N 33 N 32 N 31 N 35

PAGE 138

If: Mean Influence Ratings 7 1 Audiovisual Hiqh Influence Condition Low Influence Condition i — visualonly Audioonly Channel-Condition — rTransciriptonly Figure 2 Interaction of Influence and Channel for Influence Ratings

PAGE 139

127 Table 2 6 ANOVA Summary for Confidence Ratings Sum of Mean Source Squares df Square Channel 3.6831 3 1.23 1.37 Level of Influence 0.8668 1 0.87 0.97 Channel x Level of Influence 5.6067 3 1.87 2.09 Error 229.1631 256 0.89

PAGE 140

Table 27 Summary of Means for Confidence Ratings Factor Level of Influence low high N 133 131 4.01 4.11 Channel-Condition 1 audio-visual 2 visual-only 3 audioonly 4 transcript-only 65 66 67 66 4.18 4.12 4.06 3.88 Level of Influence x Channel-Condition Level of Influence low high Channel-Condition 2 3 4 X 4.00 X 4.18 X 4.17 X 3.64 N 32 N 34 N 36 N 31 X 4.36 X 4.06 X 3,93 X 4.08 N 33 N 32 N 31 N 35

PAGE 141

CHAPTER IV DISCUSSION S U Fjnary o f Results for Study One The results of the present study were supportive of the first hypothesis which predicted a main effect for amount of influence on leadership ratings for the designated group leader. As predicted, high levels of influence were associated with high leadership ratings and low levels of influence were associated with lower leadership ratings. Thus, subjects, who were observers of a small group interaction, were apparently able to utilize inform.ation regarding the presence or level of an influence relationship within the group interaction when assigning leadership ratings to group members. This provides support for the notion that leadership or, the oerception of leadership by outside observers, is based at least in part on the existence of an influence relationship between the '' leader" and "led" and that the degrees or levels of influence are directly associated with the degrees or level of leadership assigned to a group memlier. Hypothesis two, which predicted a main effect for amount of leader-lika behavior on leadership ratings for the designated leader -was not supported. When the 129

PAGE 142

130 designated leader had high levels of such leader-like behaviors as: attempts to control the speaking behavior of other group members, offers of numerous suggestions, statements of opinion and information, and having a high rate of participation; subjects did not assign significantly higher leadership ratings than when levels of these behaviors (leader-like behavior) were low. Thus, the presence of even relatively high levels (by one group member vis-a-vis other group members) of leader-like behaviors does not appear to increase the likelihood of receiving high leadership ratings over conditions where there were low levels of leader-like behaviors. The explanation for this lack of a significant difference due to level of leader-like behavior remains unclear. Previous research has linked the perception of leadership to the presence of leader-like behaviors (Fiedler, 1971; Hem.phill, 19 52) However this investigation em:phasizes the importance, not of leader-like behaviors, but of the effect of those behaviors, specifically vvhether they are perceived to affect or change follower behavior. It is possible then, that varying levels of leader-like behavior without evidence of their having influence effects, may not be strongly associated with leadership perception. On the other hand, possible v;eak manipulations of levels of leaderlike behavior or high levels of error variance in the analysis may also have accounted for the nonsign i f i c ant finding.

PAGE 143

131 Results of the present study were supportive of hypothesis three which predicted that a greater percentage of the leadership rating variance would be accounted for by the level of influence factor than the level of leaderlike behavior factor. Hov/ever, the difference betv/een the two was small (only 2%) and the total percent of the variance accounted for by each factor V7as also small (3% for the influence factor and 1% for the leader-like behavior factor) Because only a small amount of the variance was accounted for by these two factors and so much of the variance was due to error, the relationship between these two factors, in terms of which one accounts for a significantly larger percentage of the variance, remains unclear. Thus, although results were supportive of hypothesis three, conclusions should rem.ain tentative and this question should be subjected to further study. With regard to the great amount of error variance found in the study, two primary, contributing factors may be suggested. There were several uncontrolled-f or factors in this study including the use of a very wide variety of subjects. Subjects included males and females, blacks and whites, young adults as v/ell as older adults, urban, state university students as well as private parochial college students. Perhaps with greater subject sim.ilarity, the error variance would have been decreased. Further, although many of the group m.emJsers behaviors were carefully scripted and controlled for, not all behaviors (including naK ..jiwai^aM

PAGE 144

132 such things as facial expression, non-matched or non-turnindicating gestures) were specifically scripted, counted or controlled for. It would be nearly impossible to control for all of the behavior for each of the three group members and still have a reasonably naturalistic group interaction, thus error due to this factor would be difficult to reduce to any significant extent. Hypothesis four predicted a significant interaction effect on leadership ratings for the designated leader between the influence factor and the leader-like behavior factor. Results of the present study were generally supportive of this hypothesis. A significant interaction effect was obtained but the ordering of the means vzas slightly different from, the order predicted. The conception of leadership taken in this study holds that leadership is primarily an influence relationship; ratings of leadership for group member three should thus be minimal for both cells one and three v/here no such relationship exists. As expected, the cell which involved lov7 levels of both influence and leader-like behavior received the lowest leadership ratings. No group memi)er was experimentally designated as leader in this cell (cell 1) Also as expected, cell 3 (low levels of influence and high levels of leader-like behavior) received the next lowest leadership ratings. However, instead of the expected order of the means where cell 2 v/ould have the third highest mean and cell 4 would receive the highest

PAGE 145

133 mean, the order obtained was reversed. Cell 2, in V7hich level of influence was high, but leader-like behavior was low, resulted in the highest leadership rating for the designated leader (even higher than that received for cell 4 where both level of influence and leader-like behavior were high) It is unclear what this reversal in the ordering of the means for leadership ratings means. It may be due entirely to chance factors, the difference between means for cell two (X = 5.97) and cell four (X = 5.87) is nonsignificant. Tukey s HSD test for post hoc comparison of means (Kirk, 1968, p. 77) was performed. I X^-X, = .091KHSD = .8579 and is thus nonsignificant. 2 4 What the results do appear to clearly suggest is that it is the presence of an influence relationship which is most strongly associated v;ith the perception of leadership in the sense that it is in those cells v/ith high levels of influence where the highest leadership ratings for the designated leader were received. A main effect for group member was predicted by hypothesis five such that the designated leader (group member three) would consistently receive higher leadership ratings than nonleaders one and two. This effect v/as predicted to occur in all cells in which a leader was designated (either by giving him more influence over his fellov; group members, having him engage in more leaderlike behaviors, or both) Results of the present study were supportive of this hypothesis. The individual 3>i^is*r'—t>'Ki*i^w w TT 1 'an jii t/ m

PAGE 146

134 experimentally designated as group leader was in fact given significantly higher leadership ratings than other group members. Thus, the manipulations were successful in creating a group leader in the perception of the observersubjects Hov7ever, a related, unhypothesized effect was also found to be significant. A significant main effect for group member was obtained in cell one where no one group member was either more influential than other mem.bers or engaged in m.ore leader-like behaviors than other members. The pattern of the mean leadership ratings for each group meirJoer reveals that it was group m.ember one who received the highest leadership ratings, followed by member three and then two. The most likely explanation is that this effect was due primarily to chance or uncontrolled factors. All group members' behaviors were scripted to be nearly equal in influence and leader-like behavior and effects due to differences in physical appearance were covaried out. Further, the differences between the mean leadership ratings for members one and three was quite small (X for member one 4.9 3 and X for member three was 4.83) and nonsignificant. Overall, the results of the tests of hypotheses for study one support the contention of the present investigation that the perception of leadership is based, at least in part, on the presence of an influence relationship among individuals. Further, the influence II i.Mi i r'''a iii 111iiinr h III

PAGE 147

135 relationship appears to play a greater role in subjects' judgments of leadership than the enactment of certain leaderlike behaviors on the part of at least one group member. Ancillary analyses revealed a number of additional findings. First, leadership ratings and influence ratings were found to be moderately correlated. Second, main effects for both level of influence and level of leaderlike behavior were obtained on influence ratings for the designated leader such that ratings of influence were higher v/here influence was high rather than low and ratings of influence were high when there was a high level of leaderlike behavior versus when levels were low for leader-like behavior. The main effect for level of influence was expected. However, the main effect for level of leader-like behavior is difficult to interpret. One possible explanation is that subjects' perception of influence for a group member is related to the number of leader-like influence attempts which that groiip member makes regardless of whether the attempts are successful or not. Thus, while perceptions of influence may be based on an individual's attempts at influence, perceptions of leadership are based on an influence relationship, i.e., an individual is actually perceived to influence another. A significant interaction effect between level of influence and level of leader-like behavior was also obtained on ratings of influence. Thus, ratings of • v*fli *h g i.^^ s.isg K ypd

PAGE 148

136 influence depended both upon influence as well as leaderlike behavior manipulations, as did ratings of leadership. Finally, subjects reported being more confident of their ratings of leadership for group merri)er three in conditions where influence was high rather than low and where levels of leader-like behavior were high rather than low. Both m.ain effects (for level of influence and level of leader-like behavipr factors) were significant for confidence ratings.' The above results suggest a relationship between influence ratings and leadership ratings in the sense that essentially the same patterns emerge in terms of obtained main effects and interaction effects and in terms of the obtained moderate correlation between the two. Confidence ratings for leadership would be expected to be high in conditions where the leadership manipulations were strong, and this was indeed found to be the case. Subjects were given the opportunity to indicate by means of an adjective check list and by ranking procedures, those adjectives which best described their personal reasons for their leadership ratings. Specifically, they were asked to give the reasons behind their leadership rating for that group member to whom they gave the highest rating (in m.ost cases, for cells two, three and four, this was group member three) The author was curious to see to what extent the subjects indicated that they perceived the experimental m.anipulations (e.g., would the adjectives

PAGE 149

137 chosen vary across conditions reflecting the differences in conditions) and to determine to what extent subjects were consistent in their adjective choices. Overall, there was little consistency among subjects for the adjectives ranked and checked and no patterns emerged across conditions. The three adjectives chosen most frequently indicated that subjects perceived the most leader-like group member to be self-confident, dominating and aggressive. These are rather similar types of descriptions, and ones which are occasionally used to describe leaders, task leaders in particular. For the m.ost part, they do not reflect a direct awareness on the part of the subject, that the group leader was in fact influential or engaged in more leaderlike behavior. The adjectives chosen are more descriptive of personality traits than behaviors and are somev/hat m.ore descriptive of an individual group member's behavior rather than being descriptive of a type of relationship or interaction between group members (e.g., subjects chose adjectives such as, aggressive, dominant, and self-confident to describe the leader. They did not choose descriptions such as, controls the speaking floor, is more influential, is agreed with the most, structures the group's interactions, etc. to describe the leaderfollower interaction). Thus, the presence of an influence relationship, and to some extent, the presence of leaderlike behaviors, directly affected subject-observers Jtin* jji^-BM-gp-iw

PAGE 150

13 8 perceptions of leadership even without their being directly aware of these particular factors. Summary of Results for Study Two The results of the present study were partially supportive of the first hypothesis which predicted that the designated leader would in fact be rated as most leader-like (receive the highest leadership ratings) in each experimental condition: whether subjects were given full audio-visual information along both verbal -content and nonverbal kinesic and vocalic channels, visual (nonverbal kinesic) information only, audio (verbal content plus nonverbal vocalic) informiation only, or transcript (verbal-content) inform.ation only. For all conditions except the audio-only condition, group member three, the designated leader, received significantly higher leadership ratings than group members one and two. Leadership can thus be conveyed through verbal as well as nonverbal m.eans as predicted and as found in previous, similar studies (Stein, 1975) The main effect due to group member was nonsignificant in the audio-only condition. Several explanations for this result are possible. One explanation, of course, is that leadership is not conveyed well through audio-only situations. However, this seems implausible. Previous research (Stein, 1975) has demonstrated that leadership is perceived accurately v/hen subjects are given audio-only information. Further, ^x-ib^go^i l iw u u un w aii^< a w m i m n ti r/a >'mf>'v* 9:i '^m ^j^T*iV.jj w^^-iXi^-^iv^^i^iii^ acgv? c— ". "" *^ ~ .i* -J -* < .:^iit'^>s^:Mib*'<*
PAGE 151

139' audio-only conditions provide leadership information along two channels simultaneously: verbal content as well as nonverbal-vocalic. The verbal content-alone condition (the transcript condition) was successful in conveying accurate leadership information. Thus, it does not make sense that a channel offering verbal-content and nonverbal vocalic information should fail to convey leadership. A more plausible explanation for this result lies in a methodological weakness of the study. Subjects had som.e difficulty distinguishing the voices of the three group m.embers on pre-test trials and in remembering v/hich voice belonged to which numbered group m^ember at the time that they were required to make leadership ratings for group members. Perhaps a better methodology would have involved using m.ore distinguishable voices or em.ployed a system whereby a visual numJaer (one, two or three, corresponding to each voice) appears on a TV m.onitor screen as each voice is played during the group interaction. This method V70uld allow the subject to identify each voice by number as it occurs throughout the interaction and this should aid in both distinguishing and remembering the voices of the group members Overall, the results from the tests of hypothesis one suggest that an influence relationship, and hence leadership, is conveyed by both verbal and nonverbal means. However, because full support for the hypothesis was not obtained (e.g., for all conditions) conclusions should ii> >— IL J M KWO ^LJ iS j aM ff^ *| .T1ri l T ^ a^gl^1 < ^'• ^^^^ n^ w lll^^anl^? ^^ i*J fc^ '' < f t < ^M ^ lli n j w i i iifjm

PAGE 152

remain somewhat tentative and the question of the degree to v/hich different verbal and/or nonverbal coiranunicative channels convey leadership should be subjected to further investigation. Hypothesis two predicted a main effect for channelcondition on leadership ratings for the designated leader. Specifically, it was expected that leadership ratings V70uld be higher in condition one (audio-visual) than in all other conditions since more influence is conveyed in more different ways here than in any other condition. That is, influence in condition one is conveyed through the verbal content (by such m.eans as: leader offers suggestions and foilov/ers verbally accept those suggestions) through the nonverbal-vocalic channel (e.g., leader attem.pts an interruption and succeeds) and through the nonI verbal kinesic channel (e.g., leader indicates by gesture who is to speak next, and he is successful in selecting the speaker, leader changes his body posture orientation Euid is then matched by the followers) All of these behaviors, as in any multichannel interaction, may occur separately or simultaneously. Thus, the expectation was that the more channels along which influence is conveyed, the stronger the perception of leadership. This hypothesis v;as not supported in the results. A main effect for channel-condition was not obtained. An examination of the means reveals that there was little difference between mean leadership ratings for the designated leader across 'i,(i*('*c=*r-'-i"*=-^"^ -^-

PAGE 153

141 channels. Further, the mean ratings were all moderately high. Thus, subjects were able to perceive (as indicated by their ratings of leadership for group member three) leadership accurately in all conditions; all conditions were successful in conveying leadership. Although these results did not support hypothesis two, they do conform to results of other studies in which channel comparisons have been made insofar as other studies (e.g., Stein, 1975) have found that leadership is conveyed accurately across a variety of verbal and nonverbal channels. However, two explanations are offered here for the lack of the predicted cumulative effect for amount of influence (e.g..., influence is conveyed by multiple communicative channels) on leadership ratings. Perhaps the sheer number of influence interactions occurring across nonverbal and verbal channels within a group is not so important to the perception of leadership as the fact that a minimumL level of influence is perceived on at least one channel. That is, leaders are those who are perceived to have influence over other group m.emliers, however, having more influence does not m.ake them, even more leader-like. A second explanation deals with the fact that we do not know what subjects attend to in multi-channel communication interactions (as provided by the audio-visual channel in the present investigation) That is, just because the experimental manipulation offered subjects influence' relationship info.rm.at ion along verbal-content, nonverbal-kinesic and ywc^ff^gx^ itit^tei r i Tr— Mr-f?rL i ^=B <* 'r a a -gj r
PAGE 154

142 nonverbal-vocalic channels, does not mean that subjects in fact attend to all of that information. Subjects may attend primarily to one channel (e.g., kinesic or verbalcontent only) even in the audio-visual condition thus making it functionally equivalent to the other singlechannel conditions (visual-only, audio-only, transcriptonly) If subjects do tend to focus on a sub-set of communicative behaviors rather than on the entire set of behaviors in multichannel situations and they adapt their focus to whatever inform.ation is available to them in single channel situations, then a single versus multichannel comparison would not be expected to produce different levels of leadership ratings. Results from the test for this hypothesis revealed a significant main effect for level of influence. Thus, regardless of the channel along which an influence relationship was portrayed, higher leadership ratings were given to the designated leader when influence levels were high rather than low. A significant interaction effect V7as also obtained between levels of influence and channelcondition such that the highest leadership ratings were associated with high levels of influence for audio-visual and visual-only conditions and the lowest leadership ratings were associated with low levels of influence in these same two channel-conditions. Thus, the audio-visual and visu£il-only channel-conditions appeared to best convey the different levels of influence. Apparently the channel

PAGE 155

14 3 along which information is conveyed is important v^hen the type of information conveyed is taken into account. Differences between high and low levels of influence appear to be best conveyed in audio-visual and visual-only conditions than in the audio-only and transcript-only conditions Ancillary results for study two, as in study one, support the view that leadership ratings and ratings of influence for the designated leader are related. A moderately high correlation was obtained betv/een leadership, and influence ratings. Further, influence ratings shov7 the same patterns as leadership ratings with regard to a main effect for channel-condition (no significant effect was obtained) a m.ain effect due to level of influence (a significant effect was obtained) and the interaction effects between level of influence and channel-condition. The interaction effects obtained for both leadership ratings and influence ratings were significant and the type and degree of interaction v/as nearly identical in each case. Finally, results indicate that subjects felt equally or nearly equally confident of their leadership ratings across all channel-conditions. Thus, at least as far as the subjects were concerned, they apparently felt that they had enough or adequate information provided them, in each experim.enral condition on which to base their leadership ratinQS. tf^^ifV^g^i>^^ nT mi rm ( m*' i* W*^*it' • v^ ^ j t ^J* 'VH'&'Wfc ^nr -**^i-r=-3i*-XMSibjT^iJiWMc^--— ^^ ji^;r*(r=i.^.'-^iai*i^-=t > iJiJ5^.^ -^

PAGE 156

14 4 Overall, results from both study one and two provide support for an interactional conception of leadership and in particular, for the view of leadership as an influence relationship. These studies provide support for models of leadership suggested by this author as well as models proposed in recent theoretical statements by leading leadership researchers and theorists. These studies offer support for the usefulness of the current trend in both social psychological and communication research to focus on interactional or process variables and to examine the contribution of verbal as well as nonverbal communication channels. Rec omne n dations for Future Research Several recomm.endations may be offered for future research and some of these are suggested by the lim.itations of the present investigations. Future studies should continue to investigate the importance of influence relationships to the conception and perception of leadership. Specifically, future investigations should look at a variety of different types of groups and situations, for example, groups which are more social in nature and goal than those which are problem-solving task groups as those examined in the present studies. It would be preferable to examine influence in naturally interacting groups. A disadvantage of the present studies in scripting group m.ember verbal and nonverbal behaviors is that it is ~,^tdotr^'-— ••'•^^'.-aw — ~^— i.^L^j^o-rQ.i -^^.t.t

PAGE 157

145 difficult to create completely naturally appearing groups. f The advantage, of course, is that one is av/are of exactly :i| what behaviors do occur in the group, but a methodology which uses presently available verbal and nonverbal coding schemes for coding the behavior of narurally interacting groups would provide nearly the same advantages. This methodology, comlDined with the type of methodology used by Stein (19 75) in which naturally interacting groups selected their own leaders while subjects observed the group interaction and made their judgments of group leadership would be a good alternative to the present m.ethodology In addition to the study of influence and leadership in naturally interacting groups, future investigations should focus on naturally occurring croups (as opposed to laboratory created groups) of all kinds from family settings to business settings, and dealing with long-term groups as well as short-term groups. Certainly one lim.itation of these studies which suggests future lines of research, is that the com.position of the group consisted of all males. Future research should examine influence and leadership among female as well as mixed-sex groups. Other methodological limitations of the present studies suggest other concerns for future research. In any miethodclogy that involves communicative channel comparisons, care must be taken that subjects are given enough infcrm.ation along each channel in order to make meaningful judgments (e.g., judgments of leadership). For 3i ^^ ^p.^_i..^->w:^

PAGE 158

146 example, the audio-only condition in the present study provided verbal content and vocalic influence cues which were necessary for the perception of leadership. However, other important cues which would have enabled subjects to distinguish between voices and remember who-said-what may have been eliminated as well, thus subjects may have been unable to m.ake meaningful leadership judgments for each of the three group members. An alternative methodology from that used in study two was suggested earlier in this chapter and could be employed in future investigations. Finally, it is useful to devise methodologies v/hich separate out nonverbal cues of various kinds from verbalcontent. The present investigation did not separate verbal content from nonverbal vocalics so that any effects due to vocalic-alone information on leadership could be examined. However, caution should be exercised in these channel com.parisons ho separate the channels in such a way that one does not stray too far from what constitutes natural hiiman interaction since that is precisely what we are attempting to study. Use of voice synthesizers and use of voice filters (as used in Stein's 1975 study) may produce highly artificial effects. Above and beyond the limitations of the present investigations, there are lines of research which should be pursued which this m.odel of leadership suggests. Further investiaations of the relationship of influence relationships to the conception and perception of

PAGE 159

leadership should be carried out. The following issues should be addressed. Are different types of leadership (e.g.f task-oriented vs. relation-oriented, directive vs. non-directive) associated with different types of in. fluence? What would be the critical parameters of influence relationships? Which of the following would be most important to the perception of influence and leadership: the frequency of influence interactions occurring within the group,the duration of those influence periods, whether influence takes place during certain critical periods of group interaction, or over v^hat aspects of group behavior influence is exercised (e.g., influence over group goal choice, influence over group structure, influence over speaking turn allocation, etc.)? What personal and situational variables affect the nature and strength of the influence relationship? Interactive approaches to leadership should be pursued. The question regarding the relative importance of influence relationships versus leader-like behavior to the perception of leadership should be subjected to further research. With regard to investigations of the m.ultichannel communication of leadership, farther research should be pursued to determine the extent to which verbal and nonverbal channels convey influence relationships and leadership. Each of the following questions need further investigation: v.'hich channel (s) do interactants focus upon and which channel (s) do observers of group interactions

PAGE 160

148 focus upon to determine the nature of the relationship among group members? Within m.ultichannel interactions we must determine whether individuals attend to all incoming inform.ation. If they focus on only one subset of the incoming information, what subset is chosen and what determines that choice? Finally, do all communication channels convey all types of information? That is, the present investigation obtained interaction effects which suggest that channel comparison studies must also take into account the nature of the type of information conveyed In summary, future lines of research should continue to investigate the relationship of influence to the conception and perception of leadership in sm.all groups in an effort to build a more solid and stable (cross-situational) model of leadership. Such research by necessity should focus on the interactional or process aspects of group interaction, taking into account the verbal and nonverbal components of that interaction process. wwFJiiviwi'.HMoj Bw iTBTiic'jytfiiFg^tg-^-i^i^^i-fi'T ".: ^--M7'^fir'-Sitfx^-*'rf*iy^'-'"^''^--<^*^^-.''''*'=>*^-

PAGE 161

APPENDIX A BALES CONTENT ANALYSIS CATEGORIES

PAGE 162

BALES CONTENT ANALYSIS CATEGORIES Categories 1 Shov7s solidarity, raise other's status, gives help, reward: 2 Shows tension release, jokes, laughs, shov;s satisfaction: 3 Agrees, shov/s passive acceptance, understands, concurs, complies: 4 Gives suggestion, direction, implying autonomy for others: 5 Gives opinion, evaluation, analysis, expresses feeling, v;ish: 6 Gives orientation, inf orm.ation repeats, clarifies, confirm.s: 7 Asks for orientation, information, repetition, confirmation: 8 Asks for opinion, evaluation, analysis, expression of feeling: 9 Asks for suggestion, direction, possible ways of action: 10 Disagrees, shows passive rejection, formality, v/ithholds help: 11 Shows tension, asks for help, withdraws out of field: 12 Shows antagonism, deflates other's status, defends or asserts self: 150

PAGE 163

APPENDIX B PRE-TEST SCALE FOR PERCEIVED NATURALNESS OF INTERACTION

PAGE 164

PRE-TEST SCALE FOR PERCEIVED NATURALNESS OF INTEPJICTION Condition Male Fenale circle one Belov; are a series of statements about the interaction you have just seen. Please place an "X" in the space which best reflects the degree to which you agree or disagree with the statement. 1. The conversation was similar to those people have in everyday interaction. strongly strongly agree disagree The conversation seemed natural. strongly strongly agree disagree 3. The conversation seem.ed awkward. strongly strongly agree disagree 4. Do you have any particular comments or reactions to the conversation you have just seen? 152 -.— i,. K*iB-. ::.-^*^gf*^
PAGE 165

PRE-TEST SCALES FOR PERCEIVED ACTOR CREDIBILITY, INTERPERSONAL ATTRACTION AND LEADERSHIP

PAGE 166

PRE-TEST SCALE FOR PERCEIVED ACTOR CREDIBILITY Below are a series of adjective pairs. Please indicate your reactions to the communicator by placing a m.ark in the space that best reflects your evaluation of the person. The closer your mark is to one of the adjectives, the more you feel that word represents your attitude toward the communicator. A m.ark in the center scale position indicates a neutral response on that set of v/ords. Competent Sociable Unjust Active Unselfish Nervous Outgoing Honest Tense Qualified Energetic Cheerful Lacking Confidence Informed Illogical Cruel Irritable Anxious Unfriendly Silent Incompetent Unsociable Just Passive Selfish Poised Withdrawn Dishonest Relaxed Unqualified Tired Gloomy Confident Uninformed Logical Kind Good Natured Calm Friendly Talkative 154 -N.'t*i^i'^^u fc T^i^w^' i-j'faBKwttf^ .=^s i^ iv • <8 awi(rL--cwi ^ -5 g>^ r-. 'g=i*r=ijcr;n*P*/ *JM ^s— — v**i-niP*^"

PAGE 167

155 PRE-TEST SCALE FOR PERCEIVED iVCTOR INTERPERSONAL ATTRACTION Below are a series of statements about the person. Please place an "X" in the space which best reflects the degree to v/hich you agree or disagree with the statement. 1. I think he could be a friend of mine Strongly Strongly agree disagree 2. I couldn't get anything accomplished with him. Strongly Strongly acfree disagree 3. It would be difficult to m.eet and talk with himi. Stronaiy Strongly agree disagree 4. He is very sexy looking. Stronaiy Strongly aaree disagree III 5. If I wanted to get things done, I could probably depend upon him. Strongly Strongly agree disagree 6. I find him very attractive physically. Stronglv Strongly agree disagree 7. We would never establish a personal friendship with each other. Strongly Strongly agree disagree

PAGE 168

J.D D He would be a poor problem solver, Strongly aqree ___ Strongly disagree 9. I think he is quite handsome. Strongly agree Strongly disagree 10. I have confidence in his ability to get the job done Strongly agree Strongly disagree 11. I would like to have a friendly chat with him. Strongly agree Strongly disagree 12. I don't like the way he looks Strongly agree Strongly disagree -^. ^,a?^ffi B -f n — p g=is*.--cg= yF '^ *>. =' -MJC-^1tM>^j:^ t *^^i ^) p -.,Kir=' t -.>^i.^ i*-li;

PAGE 169

157 PRE-TEST SCALE FOR PERCEIVED ACTOR LEADERSHIP Answer each of the following items in the same manner as on the previous three pages. 1. I think this person has strong leader-like characteristics. Strongly agree Strongly disagree 2. I think this person is likely to be a leader in small group situations. Strongly agree Strongly disagree 3. I think this person v/ould be concerned with other people's point of view. Strongly aaree Strongly disagree Leader 5 Concerned Follower Unconcerned

PAGE 170

APPENDIX D COVARIATE SCALES > *^fr?-™ J*— J.

PAGE 171

COVARIATE SCALES Subject # Student #_ Below are a series of statements about the person. Please place an "X" in the space which best reflects the deqree to which you agree or disagree with the statement. 1. 1 couldn't get anything accomplished with him. Stronalv Strongly agree disagree 2. He is very sexy looking. Strongly • Strongly aqree disagree 3. If I v/anted to get things done, I could probably depend upon him. Strongly Strongly aaree disagree 4. I find him very attractive physically. Strongly Strongly agree disagree 5. He would be a poor problem solver. Strongly Strongly agree disagree 6. I think he is quite handsome. Strongly Strongly agree disagree 7. I have confidence in his ability to get the job done. Strongly Strongly acree disagree 159 i-<^^^ACJig'Mjtf'gir=i^^---^-*>='-"'--w
PAGE 172

160 I don't like the way he looks, Strongly agree Strongly disagree 1

PAGE 173

APPENDIX E POST-EXPERIMENTAL TREASURES -;^A--'-V''-'-*T**-"^"^'"^T*^

PAGE 174

POST-EXPERIMENTAL MEASURES Subject #_ Condition INSTRUCTIONS: Look carefully at the TV inonitor until you are sure that you remember which student is number 1; which is number 2; and which is number 3. You will need to keep this in m.ind as you fill out the following questionnaire. If you should begin to confuse the students, ask the experimenter and he/she will show you the pictures of the three of them again.* Turn to the next page and begin. *These instructions were modified to be appropriate for audio-only and transcript-only conditions of Study 2. 162

PAGE 175

PLEASE ANSWER E ACH OF THE F0LL0V7ING QUESTIONS VERY CAREFULLY. IT IS PARTICULARLY IMPORTANT THAT YOU ANSWER EVERY QUESTION AND PLEASE ANSWER ONLY ON THE LINE PROVIDED. PLACE AN X ON THE LINE WHICH BEST REFLECTS YOUR OPINION. On the scales provided below, please rate each of the three students (#1; #2; #3) on the amount of leadership you think he exhibited during the interaction you just viev/ed (heard, read) Note: different students (#1; #2; #3) may receive the same leadership rating if you think that they showed equal amounts of leadership during the interaction. A rating of "1" indicates that you think the student exhibited "none or almost no leadership" in the interaction. A rating of "7" indicates you think he exhibited an "extrem.ely great am.ount of leadership" in the interaction. Leadership rating for Student #1 none/ extremely alm.ost no great leadership amount of leadership Now, please indicate on the scale below hov; confident you are of the leadership rating you just gave Student #1. 1 2 3 4 5 not very at all conficonfident dent

PAGE 176

164 Leadership rating for Student #2 none/ extremely almost no great leadership amount of leadership Nov/, please indicate on the scale below, how confident you are of the leadership rating you just gave Student #2. not very at all conficonfident dent Leadership rating for Student #3 none/ extremely almost no great leadership amount of leadership Now, please indicate on the scale below how confident you are of the leadership rating you just gave Student #3. not very at all conficonfident dent -*;-;iC3i^i-jvr*'.^c-sJ.Hgt^^w^-"r'==t— — i r-J -*---_*!•— ifl-'^u*"' -.•fc>0-^''*i-ir^wi.*i^D*™O'*'^T'****s*'"'

PAGE 177

165 On the scales below, please rate each of the three students (#1; #2; #3) on how sociable they seemed on the basis of the interaction you have just seen (heard, read) Note: different students may receive the same rating if you think that they showed equal amounts of sociability in the interaction. A rating of "1" indicates that you think the student showed "none or very little sociability" in the interaction. A rating of "7" indicates that you think the student exhibited an "extremely great amount of sociability" in the interaction. Sociability rating for Stiident #1 7 none/very little sociability extremely great amount of sociability Sociability rating for Student #2 none/very little sociability extremely great amount of sociability Soc lability rating for Student #3 none/very little sociability extrem^ely great amount of sociability a.' -i#*?qti^U!iui -1 1"*^ •#':* 0*t^^^*"^ '' .-^•a^-tue^ j|%ivv^ • r rm M m J 'i ii '^i <^'

PAGE 178

. D tl On the scales below, please rate each of the three students on the amount of influence he appeared to have on either one or both of the other two students during the interaction you have just seen (heard, read) Note ; different students may receive the same rating if you think that they showed equal am.ounts of influence on the other m.embers of the group. A rating of "1" indicates that you think the student showed "none or almost no influence" on either of the other students during the interaction. A rating^^of "7" indicates that you think the student showed "a areat deal of influence" on either one or both of the other two students in the interaction. Influence rating for Student #1 none/' :" almost no influence extremely great amount of influence Now, as you have done before, please indicate hov7 confident you are of the influence rating that you just gave to Student #1. not at all confident very confident Influence rating for Student #2 none/ almost no influence extremely great amount of influence

PAGE 179

167 Now, please indicate ho w confident you are oZ the Influence rating you just gave to Student #2. not at all confident very confident Influence rating for Student #3 none/ almost no influence extreraely great amount of influence Now, please indicate how confident you are of the Influence rating you just gave to Student #3. not at all confident very confident On the scales below, please rate each of the three students on how pleasant each of them seemed, based on the interaction you have just seen (heard, read) ^ Note: different students may receive the same rating If~you think that they appeared to be equally pleasant during the interaction. A rating of "1" indicates that you think the student was "not at ail pleasant" during the interaction. A rating of "7" indicates you think that the student was "very pleasant" during the interaction. '?' ^ W ^ '''I^T~~-n .^-•t k— (!——*^'•' i>4>*j;4* oK-ar-IX

PAGE 180

168 Pleasantness rating for Student #1 not very at all pleasant pleasant Pleasantness rating for Student #2 12 3 4 5 6 7 not very at all pleasant pleasant Pleasantness rating for Student #3 not very at all pleasant pleasant Please go back to Question nuinber 1 on the first two pages of this questionnaire. Note which student (number 1, student number 2, or student nuinber 3) you rated as shewing the gr eatest aiaount of leadership If tv/o students tied v/ith equal ratings for amount of leadership (if you gave them the same leadership rating) then note which two students these were. Similarly, if you rated all three students v;ith exactly the same leadership rating, then note that. In the space below, write in that student (or those student3--in cases of a tie) who you rated as having the greatest am.ount of leadership during the interaction (see your responses to question number 1)

PAGE 181

169 was/were rated Student (s) numbered as having the greatest amount of leadership. Now, I would like you to think back and try to remember what it was about that student (or those students — in cases of a tie) who you saw as having the greatest amount of leadership — that m.ade you rate him so highly. Which of the following adjectives best describes your reasons for rating that student as having the most leadership? Please place an X. in the space provided next to the 15_ (fifteen) adjectives or descriptions listed below which were most central to your thoughts about that student on the leadership rating. is persistent is responsible talks the most is looked at the most _by the others _is enthusiastic rewards the other's ideas is self-confident _interrupts others is egotistical asks questions "is agreed with the raost _is a good listener _is dominating _is logical effectively controls the _conver s ation _is cooperative _gives the most opinions _is aggressive determines who is to speak_ _by gesturing to them initiated changes in body position in group m.embers __is agreeable _is knowledgeable is the focus of the group's behaviors jnakes the decisions _is personable offers suggestions "has greater influence _over the others _summ.arizes points offers the best/good _suggestions is considerate of the _others structures the group s interactions initiates group proce_dures _is insightful receives more of the group's coiPimunications evaluated ideas "mediated disagreements Now, go back to the 15_ adjectives or descriptions that you just checked. Of these 15, I would like you to select the five (5) which are the most important And of those top 5, rank each of them in order of importance, from. "1" to "5." That is, the one adjective or description which was the most important or most central to your thoughts in rating the raost leader-like student (s) — should receive a score of "1." The one

PAGE 182

170 adjective or description which was the second most important or central to your thoughts in rating the most leader-like student ( s) —should receive a score of "2," and so on until you have selected the adjective or description which was fifth m.ost important or central to you in making your leadership rating for the student. Go back to the preceding page, and assign numbers 1, 2, 3, 4, and 5 to the most important five adjectives or descriptions. Each one of these should have already received a check or an X. Mark the numbers on the line to the left of the description. Please make them clearly legible. Please answer each of the following questions by circling the most correct response. It is very important that you answer each question. 1. Your sex is: Male Female 2. Your age is: below 17 30-35 over 50 17-22 36-42 23-29 43-50 3. Your race is: Caucasian Other Black Oriental 4 Tvhat school do you attend? Cleveland State John Carroll University University 5. Are you a native speaker of English? (Was English the first/one of the first languages you spoke?) Yes No *6. Do you know any of the three CSU students who appeared on the videotape? Ye: No

PAGE 183

171 If you answered Yes to question number 6 above, please state which student or which students you know, and how you know him/ them. 8. Do you have a guess about what the experimenter was trying to find out by this study? If so, please state your idea belov/. *These items were omitted for subjects receiving audio-only or transcript-only conditions in Study Two.

PAGE 184

APPENDIX F STUDY TWO CONDITION FOUR TRANSCRIPT (The Low Influence, Lov? Leader-Like Behavior Condition of Study One)

PAGE 185

study 2 transcript condition 1 INSTRUCTIONS: This is a verbatim (word-for-word) transcript of a conversation which occurred last year among three Cleveland State University students. These three male students were asked to offer suggestions to help solve the parking problem, at Cleveland State. Please read the conversation carefully--but you do not need to go back and re-read any portions of the conversation. Note the numibers (#1, #2 or #3) in the lefthand side of each page. These num.bers refer to each of the students. Thus, #1 is student number 1, #2 is student number 2, and #3 is student number 3. This tells you which student is talking at that time. Note that every tim.e the typed lines are double-spaced — you v/ill find that a different student is speaking. Please read the following conversation. You will be asked to fill out a questionnaire about what you have just read V7hen you have finished. If you have any questions please ask the experimenter. 17 3 laa rr iiii g *t=^^Fi-a< r>g' -=-g ; ? *a ^ i i T:: .i 3..^uv

PAGE 186

174 #3 The problem is parking. 1 There are one-half as many spaces as there are cars — 3 And the police — 1 And traffic control has been issuing $20 tickets and towing cars. 2 Yes, that's right. I knov/ because I got my car towed and I got about three parking tickets /laughter/ 1 Nov/, what we need to do is come up with som.e solutions for the problem. 3 Right, Vvell, I have a suggestion here and we all can discuss it. 1 Well, first, I was thinking thar we could, ah 2 Uh, why not use the money from, them parking tickets and towing charges for one of a number of alternatives, either build additional parking somewhere on campus — 1 Uh, there's no more space for any new lots anywhere nearby. 3 Ah, or, they could use that m.oney to create some type of inter-campus system by which people could park lets say up to two miles away from, campus and buses will — will come pick them, up from the parking lots every 10 or 15 m.inutes. What do you think? 1 Do you mean like a comjnuter shuttle service? 3 Yeah, exactly. 1 Oh. 2 Thats a good idea. Buses could probably be rented from RTA^either for a shuttle or just to increase bus service to campus 1 Those are good ideas We could — 2 Buses could 1 We could rent buses on a daily or quarterly basis. "f fl' liW^—^^''^^

PAGE 187

175 3 With a, with a shuttle service, how many different pick up points would we have — 1 Well, well — 3 I mean, you'd have to have extra parking spaces at the coinrnuter points. 2 We can't decide all the details here. We're just supposed to offer suggestions. Right? 1 Yeah, v?e don't have to figure out all the details. Let's leave that for the bureaucrats! 2 Yeah, right, leave it for the bureaucrats! Let's just list as many suggestions as we can and not go into detail on any of them. 3 I think its been so bad lately, since they closed that lot — 1 What about — 3 By the university tower. l" think that they are building a new lot. 1 We could have car pools and, and do what companies and businesses do with car pooling incentives. 2 What kind of incentives would we use? 3 I know in Portland, uh, people who car pool get a, a special sticker or pass for — 1 I've heard, I've heard that in New York that the car poolers have special traffic lanes-3 They leave the sticker in their car and they are allowed to park m.ost anywhere for free. 1 For free? I'm not so sure we would be able to do that. 2 Still, v/e are talking about car pooling for 17 thousand students. I mean, som.eone would have to figure out who is in proximity to who in order for this car pooling to work at all. 3 Yeah, right. But again, we are only supposed to come up with suggestions, we don't have to work out the details.

PAGE 188

176 X I agree with you. I think that we could cone up with a few, new, general suggestions. But first, why don't you write dovm, uh, the suggestions we already have. 2 Ok. 1 You know, I was thinking that we-could-2 It's possible that the university could work something out with the owner of Willard Garage next to City Hall. Thev say that its half empty all of the time anyhow and^ou" could park there and ride the bus to campus. 1 But that garacfe is too expensive. I think that garage costs $2.50 a day and no one xs going to park there for that much. 3 You could get two or three other students to split the costs of the parking. The price would not be so bad if you are splitting the cost two or three ways. Now don't you think that would work? 2 That could work. Or, mavbe the university could work something out with the owner of the garage so that the school pays so much of. the parking ticket and the student pays the rest. 3 Well, we are talking about getting the school, the parking garage operators and the RTA people all together, now that's not very likely. 2 Let's go back to the shuttle idea that we had first-1 We could build a new parking lot. They are building one now but only with a couple hundred new spaces. 2 Well, even with a big, new lot, how long do you think thatal work with increasing enrollments? 3 If you look long term, with increasing enrcpllments in the next couple" of years, how about prohibiting private cars on campus. 1 Oh ah~ 3 How, how does that strike you? 1 Prohibiting private cars on campus would be severe. Some people have no other way to get here. Ah, because of their work schedules, they can't rely on the bus system. ~t'V — *^^ o\^l--i|''*— t-^'). ..-i*.—|f.>o

PAGE 189

177 3 Yes, that's, that's true I guess. Well then, I think that the shuttle bus is, is one possibility. 2 Yes, the shuttle is one possibility. I think its a good alternative. 1 I think the shuttle bus idea is the best idea we've had so far. 3 As for others, outlav/ing parking on campus would be a little severe. 2 Yes, since some students simply have no other way to get down here. 1 Well, you know I think that the only way we can get people to change the parking situation is to offer some reward, some incentive at the end. 3 Incentives or rewards aren't always necessary. I know the — 2 We could have kids get a CSU student bus pass or ticket and if you use the bus so m.any times a week and have a valid driver stamp — or he punches the card — then at the end of the quarter you would turn it back and you would — 3 Yeah, yes, and v/hen you turn the stamped or punched card, you'd get some kind of reward like maybe a dollar or maybe a free lunch for riding the bus — 2 Or they could get — 3 Or how about offering reduced quarterly tuition rates for those students who car pool? 1 Plus, they could have special lots for those people who car pool. And they could have reduced parking rates at those lots. 2 That would be nice. But how would you be able to prove that somebody car pools? 3 All, all you would need, all you would have to have is one guy at the entrance of the lot for the car poolers. You know, if you don't have 4 people in your car — 2 Yeah 3 You don't get in there to park.

PAGE 190

1 1 o. 1 Ah, come on. Now does a car pool mean you h^ive 4 people when you go in and 4 people when you leave? 3 Mo, what — 1 Or do you have 2 people when you go in and 6 people when you leave? And what would happen if your car-pool people are sick someday? 3 Well, then you wouldn't be able to park there. 1 Ah, come on! 2 These are all questions which will have to be worked out later since we are only offering suggestions. I think that we are being too far-sighted here. 3 Ok, so at this point, we have the idea of the shuttle bus from other parking areas off campus. 1 And we have the idea of car pooling with m.aybe special lots or reduced parking rates at those lots. 2 I think that the shuttle bus idea and the car pooling idea are the two best ideas. 3 Yeah. 2 Plus of course like we said earlier, we could have them build new parking lots or parking towers. 3 And possibly they could build more dorms. 1 Yeah, they could build more dorm.s so less people would have to drive into school. I know lots of people who would like to live downtown. 2 Besides, they are paying uh, rent anyhow — they might as well be paying rent down here and saving m.oney on parking and on gas 1 Yeah! 2 Dorms are a Qood idea ^ V>c^OtM>itV ^i0£S^

PAGE 191

APPENDIX G STUDY TWO CONDITION FOUR TRANSCRIPT (The High Influence, Low Leader-Like Behavior Condition of Study One)

PAGE 192

study 2 transcript condition 2 INSTRUCTIONS : This is a verbatim (word-for-word) transcript of a conversation which occurred last year among three Cleveland State University students. These three m.ale students were asked to offer suggestions to help solve the parking problem at Cleveland State. Please read the conversation carefully — but you do not need to ao back and re-read any portions of the conversation. Note the numbers (#1, #2 or #3) in the le.thand side of each page. These numbers refer to each of the students. Thus, #1 is student number 1, #2 is student number 2, and #3 is student number 3. This tells you which student is talking at that time. Note that every time the typed lines are double-spaced--you will find tha^ a different student is speaking. Please read the following conversation. You v;ill be asked to fill out a questionnaire about what you have just read when you have finished. If you have any questions, please ask the experimenter. ISO t ..j^ ff j a 1 aT*
PAGE 193

1 p,l #3 The problem is parking. 1 Yeah. There are one-half as many spaces as there are cars and — the-3 And the police have been giving out tickets. 1 And traffic control has been issuing $20 tickets and tov7ing cars. 2 That's right. I knov/, because I've got roy car tov7ed and I've got about three parking tickets /laughter/ 1 Now, what we need to do is come up with some solutions for the problem. 3 Right. Vvell, I ha\'-e a suggestion here and we all can discuss it. 1 First, I was thinking that we could, uh 2 Uh, why not use the money from them parking tickets and towing charges, which must be a lot of m^cney and use them, for one of a number of alternatives--either build additional parking somewhere on camipus 1 Uh, there's no more space for any new or additional parking lots anywhere nearby. 3 Ah, or they could use that money to create some type of a inter-camtpus system by which people could park lets say up to two miles away from campus and buses will, will com.e pick themi up from the parking lots every 10 or 15 minutes. What do you think? 1 Do you mean like a comm.uter shuttle service? 3 Yeah, exactly. 1 Oh 2 That's a good idea. Buses could probably be rented from RTA either for a shuttle or just to increase bus service to campus. 1 Those, are good ideas. We could-2 Buses could — 1 We could rent buses on a daily or a quarterly basis. iK^-iJrMvMiMKiMitnii^ii^M

PAGE 194

18: 3 With a shuttle service, how many different pick-up points would you have to have? I mean you have — 1 Well, of course — 3 To have extra parking spaces at the commuter points. 2 Well, we can't decide all the details here. We are just supposed to offer suggestions. Right? 1 Yeah. We can't figure out all the details. Let's leave that for the bureaucrats, uhl /laughter/ 2 Right, leave it for the bureaucrats! Let's just list as many suggestions as v/e can and not go into detail on any of them. 3 I think its been so bad here lately, since they closed that lot by university — 1 What about — 3 Tower. I think that, that they are building a new lot1 Yeah, that, closing down that lot has made ir bad. Well, uh, we could have car pools and do v.mat companies and businesses do for car pooling incentives. 2 What kind of incentives would we use? 3 I know that in Portland, people who car pool get a special sticker or pass for — their — 1 I've heard that in, that in New York the car poolers have special traffic lanes that — 3 They leave the sticker in their car and they are allowed to park most anywhere for free. 1 For free? Hum., maybe we would be able to do that here. 2 Still, we are talking about car pooling for 17 thousand students. I mean, som.eone would have to figure out v7ho is in proximity to who in order for this car pooling to work at all. 3 Yeah, right. But again, we are only supposed to come up with suggestions, we don't have to work out the details. •Bl.^fii*Si""*#fc**ilt.--Jlwi'.M t(,^^

PAGE 195

133 1 I agree with you. We are only supposed to conie up with a few, new general suggestions. But uh, why doesn't somebody write down the suggestions we already have? 3 Now, here's some paper. Why don't you write our suggestions down. 2 Ok. That's a good idea. 1 You know, I was thinking that we could — 2 It's possible that the university could work out something with the owner of Willard Garage next to City Hali. They say it's half empty all of the time anyhow and you could park there and ride the bus to campus 1 That garage is too expensive. That garage costs $2.50 a day — no one is going to want to park there for that much 3 You could get two or three other students to split the costs of the parking. You knov;, the price wouldn't be so bad if you' are splitting the cost two or three ways. Don't you think that would work? 2 Yeah, that could work. Or, and, and m.aybe the university could work out something with the owner of the garage so that the school pays so much of the parking ticket and the student pays the rest. 3 Well, we are talking about getting the school, the parking garage operators, and the RTA people all together — You know — that's not very likely. 2 Yeah, that's true. Why don't v/e go back to the shuttle idea we had first — 1 You knovr, we could have a new parking lots. They are building one now but only with a couple of hundred new spaces. 2 VJell, even with a big, new lot, hov/ long do you think that will work with increasing enrollmients? 3 If you look long term, with increasing enrollm.ents in the next couple of years, how about the idea of prohibiting private cars on campus. Kow does that strike you? ;axV:a i i ^ P >.i
PAGE 196

184 1 Well, outlawing private cars on campus would be a little bit severe. Some people have no other way to get here. Because of their work schedules, they can't rely on the bus system, 3 Yes, that's true I guess. Well then, I think that the shuttle bus is one possibility. 2 Yes, the shuttle is one possibility. I think it's a good alternative. 1 I think the shuttle bus idea is the best idea we've had so far. 3 As for others, prohibiting private cars on campus would be a little severe. 2 Yes, since some students sim.ply have no other way to get down here 1 You know, I think that the only way that V7e can get people to change this parking situation is to offer some reward, some incentive at the end. 3 Incentives or rewards aren't always necessary. I know that-2 We could get the kids to get a CSU student bus pass or ticket and if you use the bus so many times a week and have a valid driver stamp, then at the end of the quarter you could bring it back and you would — 3 Yeah and when you turn in the punched or stamped card, you'd get som.e kind of reward like maybe a dollar or a free lunch for riding the bus-2 Or they could — 3 Or how about offering reduced quarterly tuition rates for those students v/ho car pool 1 Those are good ideas. We could have special reserved parking lots for those people who car pool. And they could have a special rates at those lots. 2 I don t know How would you prove that somebody car pools? 3 All we would have to have is one guy at the entrance of the lots for the car poolers. If you don't have 4 people in your car--

PAGE 197

2 Yeah. 3 You don't get in to park there. 1 Does a car pool mean you have 4 people when you go in and 4 people when you leave or do you have — 3 Well, you'd have — 1 Have 2 people when you go in and 6 people when you leave? And what would happen if your car pool people are sick someday? 3 VJell, then you wouldn't be able to park there. 1 Oh, I guess that sounds reasonable. 2 These are questions which will have to be worked out later since we are only making suggestions now. I think we are being too far-sighted. 3 At this point we have the idea of the shuttle service from other parking areas off campus. 1 And v/e have the idea of car pooling with maybe special lots or reduced parking rates at those lots. 2 I think that the car pooling idea and the shuttle busing idea is the two best ideas. 3 Yeah. 2 Plus of course, like we said earlier, they could build nev/ lots or parking towers 3 And possibly they could build m.ore dorms. 1 Yeah! They could build more dorms so less people would have to drive into school. I know lots of kids who would like to live downtown. 2 Besides they are paying rent anyhow, they might as well be down here paying rent and saving money on parking and on gas I 1 You know your idea-2 Yeah your idea — 2 Your idea about the dorms is a good one. 1 Yeah.

PAGE 198

APPENDIX H VERBATIM VERBAL AND NONVERBAL SCRIPTS FROM VIDEOTAPED INTERACTIONS (Study One; Conditions One, Two, Three, and Four)

PAGE 199

Condition 1 {lov7 levels of leader-like behavior, low levels of influence) Role* Verbal Script Nonverbal Script L The problem is parking,''" All start back. f2 There are one-half as many spaces as there are ca(rs-)** L (and) the police2 f2 And traffic control has been issufl nods m agreeing $20 tjckets and towing ment with f 2 3 cars f2 gestures. 4, fl Yes, that's right. I know because I got my car towed and I got about three gesture. parking tickets. /laughter/ 5 f2 Now, v;hat we need to do is come f2 gestures. up with some solutions for the problem. 5 L Right. ^ ^L m,atches f .2 s Weil, I have a suggestion here and gesture, we all can discuss it. f2 Well, first, I was thinking that we could ah 7. fl Uh, V7hy not use the money from fl shifts up. them parking tickets and'^ towing „ charges for one of a number of f2 & L match alternatives,^ either build addishift in posture tional parking somewhere on by fl. campus f2 Uh, there's no more space for any new lets anywhere nearby. L = Designated leader (actor #3) fl = Designated follower (actor #2) f2 = Designated follower (actor #1) **Parentheses enclose simultaneous (interrupted) V, s pee en 187

PAGE 200

188 Role Verbal Script Nonverbal Script f2 L f2 fl f2 fl f2 f2 L fl f2 fl Ah, or, they could use that money to create some type of inter-campus system by which people could park lets say up to two miles away from campus and buses will--v7ill come pick them up from the parking lots every 10 or 15 minutes. What do you think? ^ Do you mean like a commuter shuttle service? Yeah, exactly. Oh. 10 L looks at & gestures to f2 to speak Thats a good idea. Buses could probably be rented from RTA either for a shuttle or just to increase bus service to campus. Those are good ideas. We (could-) (Buses could) (We) could rent buses on a daily or quarterly basis. With a, with a shuttle service, how many different pick up points (would we have-) (Well, well-) I m.ean, you'd have to have extra parking spaces at the commuter points. We can't decide all the details here. We're just supposed., to offer suggestions. Right? "^ Yeah, we don't have to figure out all the details. Let's leave that for the bureaucrats /laughter/ Yeah, right, leave it for the bureaucrats! Let's just list as many suggestions as we can and not go into detail on any of them. 10 L shifts back and fl & f2 match the shift in posture ''fl looks at & gestures to L to speak.

PAGE 201

1S< Role Verbal Script Nonverbal Script L I think its been so bad lately, since they (closed that lot-) f2 (What about-) L by the university tower. I think that they are building a nev7 lot f2 VJe could have car pools and, and do what companies and businesses do with car pooling incentives. fl What kind of incentives would we use? L I knov/ in Portland, uh, people who car pool get a, a special sticker (or pass for-) f2 (I've heard,) I've heard that in Nev? York that the car poolers have special traffic (lanes-) L (They) leave the sticker in their car and they are allowed to park most anywhere for free. f2 For free? I'm not so sure we would be able to do that. fl Still, we are talking about car pooling for 17 thousand students. I m.ean, someone would have to figure out who is in proximity to who in order for this car pooling to work at all. 12 13 L Yeah, right. But again, we are only supposed to come up with suggestions, we don t have to work out the details. 14 f2 I agree with you. I think that we could come up with a few^ new, general suggestions. But first, why don't you write down, uh the suggestions we already have.l^ "^^fl looks at & gestures to L to speak ^ L nods in agreement with fl. fl nods in agreement with L. ^^f2 looks at & gestures to fl to write

PAGE 202

19 Role Verbal Script fl f2 fl f2 L Ok. 16 You know, I was thinking that (we could-) (It's po) ssible that the university could work something out with the owner of Willard Garage next to City Hall. They say that its half empty all of the time anyhow and you could park there and ride the bus to cam.pus.1'7 But that garage is too expensive. I think that garage costs $2.50 a day and no one is going to park there for that much. You could get two or three other students to split the costs of the parking. The price would not be so bad if you are splitting the cost two or three ways. Now don't you think that would work? 18 That could v7ork Or, maybe the university could work something out with the owner of the garage so that the school pays so much of the parking ticket and the student pays the rest. Well, we are talking about getting the school, the parking garage operators and the RTA people all together, now that's not very likely. Nonverbal Script fl nods in agreement with f2 & writes. -'-'^f2 shakes head in disagreement with fl. fl nods m agreement with L. fl Lets go back to the that we had first-) shuttle (idea (We could build a nev/) parking lot.^' They are building one now but only with a couple hundred new spaces. 1 9 ^^f2 shifts forward. iT i — irii' T n-*!^*— .-. 'i'''?Wff'>'g*T'*'
PAGE 203

191 Role Verbal Script Nonverbal Script with f2. 21fl matches f2's shift forward. on on fl Well, even with a big, new lot, fl shakes head how long do you think that'al work in disagreement with increasing enrollments?21 L If you look long term, with increasing enrollm.ents in the next couple of years, how about prohibiting private cars on campus. f2 0(h ah-) 22 22 L (How,) how does that strike you? L gestures to fl to speak. 23 f2 Prohibiting private cars on camf2 shifts back, pus would be-^-^ severe. Some fl matches. people have no other way to get here. Ah, because of their work schedules, they can't rely on the bus system^. 24 L Yes, thats, that's true I guess. "L gestures. Well then, I think that the shuttle bus is, is one possibility. 24 25 fl Yes, the^shuttle is one possifl nods m bility.2 5 i think its a good agreem.ent v/ith L alternative. and matches L's gesture 2 fi f2 I think the shuttle bus idea is f2 nods in the best idea we've had so far. 2 6 agreem.ent with L and fl. 27 L As for others, outlawing parking f2 nods m on cam.pus would be a little agreement with L. severe. 2V 28 fl Yes, since some students simply fl nods m have no other 2 8 vzay to get down agreement with L. here 29 f2 Well, you know I think that the f2 looks at & only way v/e can get people to gestures to fl change the parking situation is to to speak, offer som.e reward, some incentive at the end. 2 9

PAGE 204

19: Role Verbal Script Nonverbal Script L Incentives or rewards aren't always necessary. 30 (i know the-) fl (VJe could) have kids get a CSU student bus pass or ticket and if you use the bus so many tirp.es a week and have a valid driver stamper he punches the cardthen at the end of the quarter you v/ould turn it back (and you would-) 31 L (Yeah, yes, and) when you turn the stamped or punched card, you'd get some kind of reward like maybe a dollar or maybe a free lunch for (ridincT the bus-) 30 L shakes head in disagreement with f2. M -'L nods in agreem.ent with fl. 32 L shifts up, 33 fl nods m agreement with f2. fl (Or they could get-) L (Or how) about offering reduced quarterly tuition-^^ rates for those students who car pool? f2 Plus, they could have special lots for those people who car pool And they could have reduced parking rates at those lots. 33 fl That would be nice. But how would you be able to prove that somebody car pools? L All, all you would need, all you would have to have is one guy at the entrance of the lot for the car poolers. You know, if you don't have 4 people in your carfl Yeah. L You don't get in there to park. L shifts back, f2 Ah, come on. Now does a car pool mean you have 4 people when you go in and 4 people when you leave? L (No, what-) -^f^tTr^iti^&r

PAGE 205

193 Role Verbal Script Nonverbal Script f2 f2 fl f2 fl L fl f2 fl (Or do) you have 2 people when you go in and 6 people when you leave? And what would happen if your carpool people are sick someday? Well, then you v/ouldn't be able to park there. Ah, come on! 35 These are all questions which will have to be worked out later since we are only offering suggestions. I think that we are being too^*^ far-sighted here. Ok, so at this point, we have the idea of the shuttle bus from other parking areas off campus And we have the idea of car pooling v?ith maybe special lots or reduced parking rates at those lots I think that the shuttle bus idea and the car pooling idea are the two best ideas. Yeah 37 3 S f2 shakes head in disagreement with L. ^^fl shifts up and f2 St L match Plus of course like we said earlier, we could have them build new parking lots or parking towers. And possibly they could build more dorms Yeah, they could build more dormis so less people would have to drive into school. I know lots of people who would like to live downtown. Besides, they are paying uh rent anyhow — they might as well be paying rent down here and saving money on parking and on gas. 3 8 37 L nods m agreement with fl. fl gestures -{iaS3B=aj*.-

PAGE 206

194 Role Verbal Script f2 Yeah! 39,40 fl Dorms are a good idea 43 41,42 Nonverbal Script ^ f2 matches fl's gesture. ^f2 nods in agreement with fl. fl nods in aareeinent with L. '^^fl shifts back. f2 nods m agreement with L. .l,^*.'1(ir|i*j^^tM

PAGE 207

Condition 2 (low levels of leader-like behavior, high levels of influence) Role Verbal Script L The problem is parking. f2 Yeah.'^ There are one-half as many spaces as there are cars an(d-the-l Nonverbal Script All start back. f2 nods in agreement with L L (A)nd the police have been giving out tickets f2 And traffic control has been issuing $202 tickets and towing cars. 3 fl That's right. I know, because I've got my car towed and I've got about three parking tickets /laughter/ f2 Now, what v.-e need to do is come up with some solutions for the problem.. L Right. Well, I have a suggestion here and we^ all can discuss it. f2 First, I was thinking that we could, uh fl Uh, why not use the money from them parking tickets and towing charges, which must be a lot of money, and use them for one of a number of alternatives — either build additional parking somewhere on campus f2 Uh, there's no m.ore space for any nev/ or^ additional parking lots anywhere nearby. L Ah, or they could use that mioney to create som.e type of a inter-campus system by which people could park lets say up to two miles away from camipus and buses will, will come pick them up from the parking lots every 10 or 15 minutes, VJhat do you think? 7 "f2 nods m agreem.ent with Ij. fl nods in acrreem.ent with f2. L nods In agreement with f 2 L shifts up and fl, f2 match the posture shift. f2 shakes head in disagreemient with fl. 7 L looks at & gestures to f2 to speak.

PAGE 208

19 6 Role Verbal Script Nonverbal Script f2 Do you mean like a coininuter shuttle service? L Yeah, exactly.^ ^L shifts back and fl, f2 matchf2 Oh.^ ^f2 nods in^ agreemient with L. fl That's a good idea. Buses could probably be rented from RTA either for a shuttle or just to increase bus service to cam.pus. f2 Those are good ideas. We (could-) fl (Buses could-) f2 (We) could rent buses on a daily or a quarterly basis. L With a shuttle service, how many different pick-up points would you have to have? I m.ean you (have-) f2 (Well, of course-) L (To have) extra parking spaces at the conimuter points fl Well, we can't decide all the details fl looks at & here. We are just supposed to offer gestures to L to suggestions. Right? J-0 speak. f2 Yeah. We can't figure out all the f2 looks at & details. Lets leave that for the gestures to L to bureaucrats, uhl^l /laughter/ speak fl Right, """^ leave it for the bureaufl nods in crats! Lets just list as many sugagreement with gestions as v;e can and not. go into f 2 detail on any of them. L I think its been so bad here lately; since they closed that (lot by university-) f2 (What about-)

PAGE 209

197 Role Verbal Script Nonverbal Script tower. I think that, that they are building a new lot. 13 13 f2 Yeah,"^"^ that, closing down that lot "^ f2 nods in has made it bad. Well, uh, we could^^agreement with L. have car pools and do what companies -^^ and businesses do for car pooling incentives. f2 shifts up. fl What kind of incentives would we use?-5 L I know that in Portland, people who car pool get a special sticker or pass (for-their~) f2 (I've heard) that in, that in New York the car poolers have special traffic la(nes that-) L (They) leave the sticker in their car and they are allowed to park most anywhere for free. f2 For free? Hum,, maybe we would be able to do that here. fl Still, we are talking about car pooling for 17^^ thousand students. I mean, someone would have to figure out who is in proximity to who in order for this car pooling to work at all. 17 L Yeah, right. But again, v^e are only supposed to come up with suggestions, we don't have to work out the details. 1 8 f2 I agree V7ith you. We are only supposed to come up with a few, new general suggestions. But uh, why doesn't somebody vjrite down the suggestions we already have?19 L Now, here's some paper. VJhy don't you write our suggestions aown. 20 fl looks at & gestures to L to speak. """^fl shifts up and L miatches. 17 L noas m agreement with fl. f2 nods m agreement with L. "'"^f2 shifts back and fl & L match20 L looks at & gestures to fl to write.

PAGE 210

198 Role Verbal Script fl Ok, Thats a good idea. 21 f2 You know, I was thinking that we cou(ld-) Nonverbal Script 21 fl nods m agreement v/ith L. fl writes. fl f2 L fl f2 fl (It')s possible that the university could v7ork out something with the owner of Willard Garage next to City Hall. They say its half em.pty all of the time anyhow and you could park there and ride the bus to campus 22 That garage is too expensive. That garage costs $2.50 a day--no one is going to want to park there for that m.uch. You could get two or three other students to split the costs of the parking. You know, the price v/ouldn't be so bad if you are splitting the cost two or three ways. Don't you think that would 'work?^-^ Yeah, that could work. Or, and, and maybe the university could work out som.ething with the owner of the garage so that the school pays so much of the parking ticket and the student pays the rest. Well, we are talking about getting the school, the parking garage operators, and the RTA people all gather. You know--that's not likely. 24 25 ., ._ 26 tovery Yeah,""' that's true. Why don t v.-e go back idea (we had first-) to the shuttle (You know, we) could have a new parking lots. They are building one now but only v/ith a couple of hundred nev; spaces. Well, even with a big, nev/ lot, how long do you think2 7 that v/ill work with increasing enrollm.en (ts?) ^ f2 shakes head in di:agreement with fl. L looks at & gestures to fl to speak. 24 L gestures 25 fl matches L's gesture ^^fl nods in agreement with L. 07 f 1 shakes head in disagreement with f 2 ^ri,*'.3ss>i5ete**ncW'a*3aaBir?Q*"c*r^.*ii' t/'^"*5**wvr"*^'*

PAGE 211

.99 Role Verbal Script Nonverbal Script f2 (If) you look long term, with increasing enrollments in the next couple of years, hov/ about the idea of prohibiting private cars on campus. How does that strike you? Well, outlawing private cars on campus would be a little bit severe. Some people have no other v/ay to get here. Because of their v/ork schedule they can't rely on the bus system. Yes, thats true I guess. Well then, T think that the shuttle bus is one possibility 28 79 • Yes, the shuttle is one possibility. I think it's a good alternative f2 I think the shuttle bus idea is the best idea we've had so far.-^^ L As for others, prohibiting private cars on cam^pus would be a little severe. fl Yes, since som.e students sim.ply have no other way to get dov/n here. f2 You know, I think that the only way that we can get people to change this parking situation is to offer some reward, some incentive at the end. ^1 L Incentives or rewards aren't always necessary. I know (that-) fl (We could) get the kids to get a CSU student bus pass or32 ticket and if you use the bus so many times a week and have a valid driver stamp, then at the end of the quarter you could bring it back and (you would-) 28 L gestures. ^^fl matches L's gesture and nods in agreement with L ^f2 nods in agreem.enr with fl and L. ^'f2 looks at & gestures to fl to speak. 32 fl shifts up. =:"'|r-'T-

PAGE 212

2 00 Role Verbal Script Nonverbal Script L (Yeah and when} you turn in the punched or stamped card, you'd get some kind of reward like maybe a dollar or a free lunch for riding the busfl (Or they could-) "^^ ^^fl shifts back. L (Or) how about offering reduced quarterly tuition rates for those students who car pool. f2 Those are good ideas. We could have special reserved parking lots for those people who car pool. And they could have a, special rates at those lots. fl I don't kncw.^^ fl shakes head Hov/ v/ould you prove that somebody in disagreement car pools? with f 2 L All we v;ould have to have is one guy at the entrance of the lots for the car poolers. If you don't have 4 people in your car35 35-, fl Yeah. fl nods m agreement with L. L You don't get in to park there. f2' Does a car pool mean you have 4 people when you go in and 4 people when you leave or do you (have-) L (Well, you'd have-) f2 (have) 2 people when you go in and 5 f2 looks at & people when you leave? And what gestures to L to would happen if your car pool people speak. are sick some (day?) -^"^ L (Wel)l, then you wouldn't be able to park there. 37 f2 Oh, I guess that sounds reasonf2 nods head able. 37 in agreement with L.

PAGE 213

;ci Role Verbal Script Nonverbal Script fl looks at & qestures to f2 f2 fl These are questions which will have to be worked out later since we are only making suggestions nov/. ^I think to speak, we are being too far-sighted ^^ At this point we have the idea of the shuttle service from other park• ing areas off campus. And we have the idea of car pooling with maybe special lots or reduced parking rates at those lots. I think that the car pooling idea and the shuttle busing idea is the two best ideas. Yeah 39 fl f2 f2 fl f 1 Plus of course, like we said earlier, they could build new lots or parking towers. And possibly they could build more dorms '^C' Yeah! They could build more dorms so less people would have to drive into school. I knov.? lots of kids v7ho v;ould like to live downtown. 41 Besides they are payxng rent anyhow, they m.ight as v/ell be down here paying rent and saving m.oney on parking and on gas (You know your) idea(Yeah your idea,-) Your idea about the dorms is a good 43 4 7 one. -" f2 Yeah. 39, ... L noas xn agreement v/ith fl. "^^L shifts forward & fl, f2 match. fl nods m agreement with f2. 42^, fl nods m agreement with L. 43 f2 nods m agreement with L (& fl) i.*'r>~"n.-

PAGE 214

202 Condition 3 (high levels of leader-like behavior, low levels of influence) Role Verbal Script Nonverbal Script L The problem is parking. 'All start back. f2 Yeah, there are one-half as many spaces as there are (cars-) L (-and the) police (are just-) f2 (and) traffic control has been issuing $20 tickets and towing cars. fl Yeah, thats right. I know because I've got my car towed and I've got about three parking tickets. /laughter/ 2 f2 Well what we have to do here is come f2 shifts forup with som.e solutions 2 for the ward & fl & L problem. match posture shift. L Right. Well, I have a suggestion here and we all can discuss it. fl I was thinking that we could ahL -Uh why not use that money from those tickets and towing charges uhwhich must be a lot of moneyfor one of a num.ber of alternatives: Either build additional parking som.ewhere on campus 3 3 f2 Uh uh there's no more room. f2 shakes head in disagreement with L. 4 L Ahor use that money to create som.e L looks at & type of uh, inter-campus system by gestures to fl which people can park lets say up to to speak. two m.iles away from campus and buses willwill come pick them up from, the lots every 10 or 15 minutes. What do you think? ^ f2 Like a conim.uter shuttle se(rvice?) w^ll— ^A^5k"

PAGE 215

203 Role Verbal Script Nonverbal Script L (Yes,) exactly I fl Buses could probably be rented from RTA either for a shuttle or just to increase bus service to campus. f2 VJellthose are good ideas, (we could-) L (buses could be) f2 We could rent buses on a daily basis. fl With a shuttle serx'ice, how many different pick-up points would you have to have? I mean, you have to have them out (far enough) L (well of cour-) -and also have at all the corrjnuter points.^' ^ g 5 fl -and also have extra parking spaces fl gestures. | f2 Right fl shifts back and f2 & L match, "7 '8 "^f2 matches fl's gesture. 8 f2 nods in agreement with fl. 9 L Well, we can't decide all the deL gestures. tails here. Vie 're just supposed to offer some suggestions, right?^ f2 Yeah.^ -^^£2 matches L's Lets uh, lets leave that for the gesture, bureaucrats L Ok, ok let's leave it for the burL nods in eaucratslll Let's just try and list agreement with as many suggestions as possible and f 2 not go into details on any of them. f2 You know, I think its been so bad latelythis quartersince they closed that lot over by the university tower. (I think that-) i

PAGE 216

:o4 Role Verbal Script Nonverbal Script L (Hey how about-) f2 (that) they are building a new lot. 12 fl (yeah) f2 (but) I'm not sure. fl Yeah. L We could have car pools and do what companies and businesses do for car pooling incentives, fl What kind of incentives would we use? L Well, I know in Portland, people who car pool get a a special sticker or (pass-) f2 (Yeah) in New York they use a system where uhL "They leave the sticker in their car and they are allowed to park most anywhere for free. fl For free? IJ f2 Oh, still we are talking about car pooling for 17 thousand students. Now someones gonna have to figure out who is in proximity to who. 1'^ 15 fl Yeah right, but again, we are only supposed to come up with suggestions. VJe don't have to work out details. 12 fl nods m agreement with f2. fl shakes head in disagreement with L 14 f2 gestures, fl matches f 2 s gesture f2 We could have a ride board, 16 L Well, I think I agree v/ith you. I think we should try to come up with some other general suggestions, Uhbut firstuh why don't you write our suggestions down and we will see what else we can come (up with-) 17 16^ J L nods xn agreement with fl. 17 L looks at & gestures to fl to write.

PAGE 217

) n Role Verbal Scriot Nonverbal Script fl (It's possible) that the university could work something out with the owner of Willard Garage next to city hall. They say its half empty and you could (park there-) f2 (But that) garage is too expensive. L We could have the students ride the bus to campus for free. f2 But that garage costs $2.50 a day and no one's gonna park there for that much. L You could get two or three other L looks at & students to split the costs of the gestures to f2 parking. to speak. You know, the cost wouldn't be so bad if you're splitting the price two or three ways Yeah, don't you think that would work?18 fl Or maybe the university could work something out with the owner of the garage so that the school pays so much of the parking ticket and the student pays the rest. 19 L Well, v/e are talking about getting L shifts forthe uh school, the parking garage v/ard and fl, f2 operators, the RTA people all tomatch. gether;19 thats not very likely. fl Lets go back to the shuttle idea that (we had first-) f2 (We could) build a new parking lot. They are building one now but only with a couple hundred new spaces. 20 -, fl Well even with a big nevj parking, fl gestureb. lot, how long do you think that will work with increasing enrollments?--" 9 1 L Yeah,^"'" if you look long term, uh"L matches f 1 s with increasing enrollments in the gesture, next couple of years. how does the idea of prohibiting private cars ori campusyeahhow does that strike you?

PAGE 218

Role Verbal Script Nonverbal Script f2 Ohprohibiting private cars on campus could be a little bit severe. I mean, some people simply have no other way of getting here. 9 9 22 • L Yes, thats true.' '^ shifts back, I think then that the shuttle bus is one possibility. fl The shuttle is one possibility. It's a good alternative. f2 I think it's our best idea so far. L As for others, prohibiting private cars on campus v;ould be a little severe. f2 Yes, I mean, since some people simply have no other V7ay of getting here. 23 L You know, I think the only way that L shifts up. we can get people to change the parking situation is offer them, som.e kind uh of reward, som.e kind of incentive at the end. 2^ f2 Incentives or rewards aren't always necessary. (Now I know that in-) fl (Like we could) have kids get a CSU bus ticket and if they use the bus so many times a week and they have a valid driver stamp, or they get the card punched, then at the end of the quarter they would turn (it in and-) L (Yes, and when) you turn in the stamped or punched card you'd get some kind of reward like maybe a dollar or maybe a (free lunch) for riding the bus. fl (or they could get-) L Or how about offering reduced quarterly tuition rates. Or at least reduced parking rates.

PAGE 219

207 Role Verbal Script Nonverbal Script f2 fl fl L f2 I. f2 fl L Hey, they could have special lots reserved for just those people who car pool. That would be alright. But how would you prove that somebody car pools? V7ell, there could be lots used just for those people who car pool only. All it would take would be one guy at the, uhentrance. If you don't have 4 people in your carUh hum -You don't get in there to park. 24 Hey now, does a car pool mean that you have 4 people when you go in and 4 people when you leave. (Yeah well-) (Or do you) have 2 people when you go in and 6 people when you leave? And what would happen if somebody was sick? Well youthen you don't get in to park there. 2 5 Hoohrip off! These are things that will be, have to be worked out later since we are only offering suggestions now. Ok, I think v/e are being a little too far-sighted here. f2 L fl f2 fl I don't know, I mean26 Ok. At this point we got the idea of the shuttle bus from, other areas off campus. And we have the suggestion of building a new27 Urn ( hum) (I me) an, to have spe(cial-) 24 fl nods m agreement with L. 25 L shifts back. ^^f2 shakes head in disagreement with L. 27 f2 noas m agreement with fl.

PAGE 220

208 Role Verbal Script Nonverbal Script L (Plus) we got the idea of car pooling with maybe uh-28 special lots or reduced parking rates for them. fl L 28 L shifts up. I think that the shuttle bus idea and the car pooling idea are the two best ideas. 29 Yeah, plus like you said earlier, we could build new lots or parking tow(ers-) f2 (That) v/ouldn't work. But we could build dorms. 30 Yeah. And we could even build more dorm.s so less people v/ould have to drive into schoolYou knov;, I know a lot of kids who would like living downtovm.31 They are paying rent anyv/ay so they might as v/ell be down here paying rent and saving on parking. fl On parking and on gas! 32 f2 Yeah. 33 7Q f2 nods m agreem.ent with fl. fl nods in agreement with f2. 31 fl nods m agreement with L. ^ f2 nods in agreemient with fl. "fl nods m agreement with f2 & L.

PAGE 221

209 Condition 4 (high levels of leader-like behavior, high levels of influence) Role Verbal Script L Ok. The problem is parking, fl Yeah. f2 There are one-half as many spaces as there are (cars for the spaces-) L (And the) university traffic control has been issuing $20 tickets and towing cars, 2 f2 Yeah, that's right. I know because I've gotten my car tov;ed and I've gotten about three parking tickets. /laughs/ fl Now, what we need to do is come up with some solutions for the problem. L Right. -^ Well, I have a suggestion here and we all can discuss it. fl Good.^ L Why not take that money from those tickets and towing charges and use it for one of a number of alternatives, either build additional parking5 -somewhere on campusf2 Uh huh.^ Ah, or, use that money to create some kind of uhinter-campus system by which people can park let's say up to two miles away from campus and buses will, will come pick them up every 10, 15 minutes. What do you think? 7 f2 Like a comiTiUter shuttle service? 8 Nonverbal Script ''All start forward. f2 shifts back, 3 L head nods in agreement to fl. fl nods in agreement to L. fl head nods in agreement to L. f2 nods in agreement to L. L looks at & gestures to f2 to speak. o f2 makes a shift forward. ^.-.^V.--*-— .-.i-W^^JlV • W*JH— sw^**!*,!*-*f^Ti^'V •W."^**^'"'''*-^"' •'^

PAGE 222

210 Role Verbal Script L Yeah. f2 Oh, good idea. You know, I think we could probably get the buses from the uh the (transit auth-) L (RTA) on a, on a daily or quarterly basis f2 Yeah 10 fl How nany different pick up points would you have to have? (I mean-) L (Well, you'd have-) fyou) have to have them far enough out and also have extra parking spaces at all the points fl f2 fl f2 And who would drive the buses? 11 Nonverbal Script ^L nods to f2 in agreement. -'-f2 nods to L in agreement. ^'fl looks at f2 to answer Well, we can't decide all the details here. We're just supposed to offer some suggestions. Right? Yeah. -, J.T, I mean, we can't figure out all the details. Let's leave that for the bureaucrats. /laughs/ So"'-^ what we need to do then is just to list as many suggestions as v/e can and not go into detail on any of them. You know, I think it's been so bad l^telvthis quartersince they closed that lot over by the university tower. I think they are (building a new-) (What about car) pooling. We could have uhcar pools and we could do what businesses and companies do for car pooling incentives. '-^fl nods to f2 in agreement.

PAGE 223

-7 1 1 Role Verbal Script Nonverbal Script f2 Yeah/^ well there would have to be ^ t2 nods to L incentives because you know, nobody in agreement. is just going to say, "I brought ten kids with me this week" and everything will work (out-) L (Yeah, I) know, in Portland, see, people who car pool get uh a special sticker or (pass-) f2 (uh huh, we could-) L 14 (that) they leave on their car, in L gestures their car, 'and people are allowed to park most anr-^here for free.^'f2 Yeah,^^ we could use, uh^"f2 matches L s gesture fl Still, you are talking about 17 thousand people here. Someone would have to figure out who is in proximity with who. f2 Yes, but again, we are only supposed to figure out the solutions. We don't have to work out the details L I think V7e should try an com-e up with L shifts back =;ome other general suggestions so ah, and fl & i-2 well first, -'^'-'-'^ why don't you write match. our suggestions down. 17 And we will just see what we can L ges^u es po • -u fl to write. com.e up wiph. -^-^ ^^ Q^ 18 19 ^^fl shifts up. Mavbe the university could work 2.9 something out with the owner of -1 writes. Willard Garage next to City Hall. 20^ ^^ v, 1 They say its half empty all of the '"fi snifts back, time anyhow and^vou could park (there and ah-) ^^ L (hov7 v;ould-) fl -and ride the bus for free to campus. f2 That won't work because that garage costs $2.50 a day and nobodys gonna pay that much to park. qiatS B T 'Wv-" — ? w ^!*rH^tfr^"-

PAGE 224

Role Verbal Script Nonverbal Script L Uh, you cculd uh, get two or three other uh people to split the cosrs of the parking. The cost V70uldn't be so bad if you are splitting the price two or three ways .21 fl Maybe the university could v/ork something out with the garage owner so that the school would pay so much of the parking ticket and the student would pay the rest. 2 2 L Yeah, but you are talking uh, about aetting the school, the parking garage operators, the RTA people all togetherthat's not very likely. fl Unless the school purchased their ov7n buses. f2 I think we already have some buses. fl But not enough to han(dle-) f2 (Let's) go back to the idea of the shuttle 23 fl VJe could build new parking lots. Now, they are building one now but only with a few hundred spaces. f2 But even v;ith a big new parking lot with increased enrollments and all next year, you know that its not going to be enough. L If you look long term, with increasing enrollments in the24 next ^couple of years, how does the idea of prohibiting private cars on campus strike you? 2 5 fl Hmjnm. Outlawing parking on campus could get a little severe. Uh some students simply have no other ways of getting down, here. f2 Yeah, and besides you know we can't work this thing out overnight. L looks at & gestures to fl to spea.k fl looks at & gestures to L to respond. '^fl shifts up. 74 25 fl shifts backL looks at & gestures to fl to speak.

PAGE 225

213 Role Verbal Script Nonverbcil Script 26 L I think then, that the shuttle bus is one possibility. fl ^^The shuttle bus is a good alternative f2 It's the best one we've had so far, L As for others, prohibiting private cars on campus would be a little severe Some people simply have no other way to get here. f2 f2 fl Yeah, I mean with their work schedules they can't depend on the bus system. ^^I think the only way that v^e can get people to change the parking situation is give them some kind of uh, reward, you know som.e kind of incentive at the end. People won't change their behavior without an incentive. Yeah, incentives are necessary 30 '^^fl nods to f2 in agreement ^"^L hand gestures. ^^fl matches L's gesture. 99 rL shifts up and fl & f2 match. ^f2 nods to L in agreement. Like we could get the students or the kids to get a University uh bus ticket and if they use the bus so manv times a v/eek and have a valid driver stamp, then at the end of the quarter they would turn it back (and-)' (and) when you turn in the stamped or punched card you'd get som.e kind of reward like a, yeah maybe a dollar you know, or m.aybe a (free-) fl (or-) -lunch for riding the bus. Or, how about offering reduced quarterly tuition rates for people v7ho car pool. Or at least reduced parking rates. V/hat do you think?

PAGE 226

214 Role Verbal Script 31 fl They could do that. And they could have specially reserved lots just for car poolers. f2 How are you going to prove that somebody car pools? L There could be special lots for those that car pool only. You know, all it would take would be one guy at the entrance. You know, if you don't have four people in your car32 -you don't get in there to park. f2 Does car pooling mean you have 4 people when you go in and 4 people when you leave? Or does it mean you have 2 when you go in and 6 when you leave? L These are questions v/hich v/ill have to be v7orked out later since we are only offering suggestions now.-^-^ Yeah, I think we are being too farsighted here. f2 Yeah, maybe so. L ^^Ok, at this point we got the idea of the shuttle bus from other areas off campus. 3 5 fl Yeah,'^ (and we have-) L (Plus) we've got the idea of car pooling with maybe special lots or at least reduced parking rates for them f2 I think those two are our best ideas so far. 37 38 L Yeah, plus like you said, they could build new parking lots or even uhparking towers (and ev-) Nonverbal Script ^^fl nods in agreement with L. ^^fl nods to L in aareement. ^^fl nods to L and then f2 in aareement. "L shifts back & fl & f2 match. 35. 36 gestures fl matches L's gesture, ^'^f2 nods to L in agreement. ^^fl nods to f2 in agreement.

PAGE 227

Role Verbal Script f2 39 f2 (We) could build new dorms. L Yeah, and we could even build more dorms so less people would have to drive into school. You know, I know lots of kids vjho like to be downtown living. You know, they are paying rent anyway so they might as well be down here paying rent and saving on parking and gas. Huh? 1^1 '^^''^^That would work, 44: Nonverbal Script ^^fl nods to L in agreement 40 L shirts up & fl & f2 match. 41 L gestures ^"^fl matches L's gesture "^'£2 nods to L in agreement. "^^fl nods to f2 then L in agreement

PAGE 228

BIBLIOGRAPHY Bales, R. F. Interaction process analysis: A method for the study of small groups Cambridge, Mass.: Addi son-Wesley, 19 51. Bales, R. F. The equilibrium problem, in small groups. In T. Parsons, R. F. Bales, and E. A, Shils (Eds.), Working papers in the theory of action Glencoe, 111.: Free Press, 1953. Bales, R. F., and Strodtbeck, F. L. Phases in group problem-solving. Journal of Abnormal and Social Psychology 1951, £6, 485-495. Bass, B. M. Leadershi p, psychology and organizational behavior New York: Harper, 1960. Bavelas, A., Hastorf, A. H. Gross, A. E., and Kite, W. R. Experiments on the alteration of group structure. Journal of Experimental Social Psychology 1965, 1, 55-70. Beckhouse, L., Tanur, J., Weiler, J., and Weinstein, E. "... And some men have leadership thrust upon them Journal of Personality and Social Psychology 1975, 31(3) 557-566. Bell, M. A. An analysis of the communication of groups with emergent taskand relation-oriented leaders. Paper presented at the Annual Convention of the Speech Communication Association, Minneapolis, Novemiber, 197 8. Bennis, V7. G., and Shepard, H. A. A theory of group development. Human Relations 1956, 9_, 415-437. Burgoon, J. K. and Saine, T. J. The unspoken dial ogue: An introduction to nonverbal communication Boston: Floughton Mifflin, 197 8. Burke, P. J. Participation and leadership in small groups. American Socioloaical Reviev/, 1974, 39, 832-843. 216

PAGE 229

21Carter, L. F. Leadership and small group behavior. In M. Sharif and M. 0. Wilson (Eds.), Group relations at the crossroads New York: Harper, 19 53. Carter, L. F., Haythorn, W. Shriver, B., and Lanzetta, J. The behavior of leaders and other group members. In D. Cartwriqht and A. Zander (Eds.), Group d ynamics: Research and theory (3rd ed.) New York: Harper & Row, 1968. Cartwright, D. and Zander, A. (Eds.). Group dynamics: Research and theory (3rd ed.). New York: Harper & Row, 19 68. Cattell, R. B. New concepts for measuring leadership in terms of group syntality. Hum.an Relations 19 51, 4, 161-184. Couch, A. S., and Carter, L. A factorial study of the rated behavior of group members. Paper presented at the Eastern Psychological Association, Atlantic City, April, 1952. Cowley, VI. H. Three distinctions in the study of leaders. Journal of Abnormal and Social Psychology 1928, 2_3, 144-157. Ellis, D. G., and Fisher, B. A. Phases of conflict in small group development. Human Communication Research 1975, 1, 195-212. Eskilson, A., and Wiley, M. G. Sex composition and leadership in small groups. Sociometry 1976, 39_ (3) 183-194. Exline, R. V. Visual interaction: The glances of power and preference. In S. Weitz (Ed.), Nonverbal commu nication: Readings with comjuentary New York: Oxford University Press, 1974. Fiedler, F. E. A theory of leadership effectivene ss. New York: McGraw-Hill, 19 67. Fiedler, F. E. Leadership Morristown, New Jersey: General Learning Press, 1971. Fiedler, F. E., and Chem.ers, M. M. Leadership and effect ive management Glenview, 111.: Scott, Foresman, 1974. Fisher, B. A. Decision emergence: Phases in group decision making. Speech Monographs 1970, 3_1' 53-66. wrE**CI|75efMMi***|!>WTrn.Tt?-'**

PAGE 230

2 1 ,q Fisher, B, A. Perspectives on human communication New York: Macmillan, 197 8. Folger, J. P., and Sillars, A. L. Relational coding and perceptions of dominance. Paper presented at the Annual Convention of the Speech Communication Association, Washington, D.C., December, 1977. Gibb, C. A. The sociometry of leadership in temporary groups. Sociometry 1950, 13_, 226-243. Gibb, C. A. Leadership. In G. Lindzey, and E. Aronson (Eds.), The handbook of socia l psychology (2nd ed., Vol. 4). Reading,' Mass. : Addison-VJesley 1969. Gintner, G. and Lindskold, S. Rate of participation and expertise as factors influencing leader choice. Journal of Personality and Social Psychology 19 7 5, 32(6) 1085-1089. Gouran, D. S. Group communication: Perspectives and priorities for future research. Qu arterly Journal of Speech 1973, 59, 22-29. Kalpin, A. W. and Winer, B. J. The leadership behavi or o f the airplane comm.ander Columbus: Ohio State University Research Foundation, 19 52. Hare, A. P. Handbook of small group researc h (2nd ed ) New York: The 'Free Press, 1976. Hemphill, J. K. Theory of leadership. Unpublished staff report, Ohio State University Personnel Research Board, 19 52. Cited in C. A. Gibb, Leadership. In G. Lindzey and E. Aronson (Eds.), The handbook of so cial psychology (2nd ed Vol. 4). Reading, Mass.: Addison-Wesley 1969. Henley, N. M. Body Polit ics. Englewood Cliffs, NewJersey: Prentice Hall, 1977. Hersey, P., and Blanchard, K. H. Leader effectiveness and adaptability description (LEAD) In J. W. Pfeiffer and J. E. Jones (Eds.), The annual handbook for gro up facilitators La Jolla, Calif.: University Associates, 1976. Hollander, E. P. Leadership dynamdcs: A practical guide to effective relationships New York: Free Press, 1978. Hollander, E. P., and Julian, J. W. Contemporary trends in the analysis of leadership processes. Psychological Bulletin, 1969, 71(5), 387-397..

PAGE 231

:i9 Kirk, R. E. Experiraental design: Procedures for the behavioral sciences Belmont, Calif.: Brooks/Cole, 1968. Knapp, M. L. Nonverbal coir.munication in human interaction. New York: Holt, Rinehart & Winston, 1972. Knutson, T. J., and Holdridge, W. E. Orientation behavior, leadership and consensus: A possible functional relationship. Speech Monographs 1975, 42_, 107114. Krout, M. H. Introduction to social psychology Nevv York: Harper, 1942. Lashbrook, V. J. Leadership emergence and source valence: Concepts in support of interaction theory and^ measurement. Human Communication Research 19 75, 1(4), 308-315. Lewin, K,, Lippitt, R. and White, R. K. Patterns of aggressive behavior in experim.entally created "social climates," Journal of Social Psychology 1939, 10, 271-299. Lippitt, R,, and White, R. K. The "social climate" of children's groups. In R. G. Barker, J. Kounin, and H. Wright (Eds.), Child behavior and development New York: McGraw-Hill, 1943. Mann, R. D. A review of the relationships betv/een personalitv and performance in small groups. Psychological Bulletin 1959, 56_(4), 241-270. Markel, N. N. Coverbal behavior associated with conversation turns. In A. Kendon, R. Harris, and M. Key (Eds.), Organization of behavior in face-to-face interaction The Hague: Mouton, 19 75. McCroskey, J., Jensen, T. and Valencia, C. Measurement of the credibility of peers and spouses. Paper presented at the International Communication Association Convention, Montreal, April, 1973. McCroskey, J., and McCain, T, The measurement of interpersonal attraction. Speech Monographs 1974, 41, 261-266. Mehrabian, A. Significance of posture and position in the comiriunication of attitude and status relationships. Psycholog ical Bulletin 1969, 7J.(5), 359-372.

PAGE 232

220 Morris, C. G. and Hackm.an, J. R. Behavioral correlates of perceived leadership. Journal of Personality and Social Psychology 1969, 13(4), 350-361. Morse, N. C, and Reimer, E. The experimental change of a major organizational variable. Journal of Abnormal and Social Psychology 1956, 52_, 120-129. z Pepinsky, P., Hemphill, J. K., and Shevitz R. N. Attempts to lead, group productivity, and morale under conditions of acceptance and rejection. Journal of Abnormal and Social Psychology 1958, 51_, 47-54. Riecken, H. vl The effect of talkativeness on ability to influence group solutions of problem.s. Sociometry, 1958, 2^, 309-321. Rogers, L. E., and Farace, R. V. Analysis of relational communication in dyads. Human Communication Research 1975, 1, 222-239. Rosenfeld, L. B., and Plax, T. G. Personality determinants of autocratic and democratic leadership. Speech Monographs 1975, 4^, 203-208, Sarbin, T. R. and Allen, V. L. Role theory. In G. Lindzey and E. Aronson (Eds.), The handbook of social psychology (2nd ed., Vol. 1). Reading, Mass.: Addison-Wesley 1958. Scheidel, T. M. and Crowell, L. Idea development in small discussion groups. Quarterly Journal of Speech 1964, 50, 140-145. Shaw, M. E. A comparison of two types of leadership in various comjTiunication nets. J ournal of Abnormal and Social Psychology 1955, 50_, 127-134. Shaw, M. E. Gr oup dynamics: The psychology of small g roup behavior (2nd ed.). New York: McGrax^z-Hill 1976. Shaw, M. E., and Harkey, B. Some effects of congruency of member characteristics and group structure upon ^ group behavior. Journal of Personality and Social P sycholog y, 1976,' 34(3), 412-418. Sorrentino, R. M. and Boutillier, R. G. The effect of quantity and quality of verbal interaction in ratings of leadership ability. J ournal of Experimentai Social Psychology, 1975, "11, 403-411.

PAGE 233

221 Spillman, B., Saillr^-an, R. and Bezclek, J. An interactive, perceptual and shared perspective on leadership: Development of a fuzzy measurement procedure. Paper presented at the Annual Convention of the Speech Com.munication Association, Minneapolis, November, 197 8. Stang, D. J. Effect of interaction rate on ratings of leadership and liking. Journal of Personality and Social Psychology 1973, 27(3), 405-408. Stein, R. T. Identifying emergent leaders from verbal and nonverbal communication. Journ al of Personality a nd Social Psychology 1575, 32.(1)/ 12 5-13 5. Stein, R. T., Geis, F. L., and Damarin, F. Perception of emergent leadership hierarchies in task groups. Journal of Personality an d Social Psychology 19 7 3, 28_(1), 77-87. Stogdill, R. M. Personal factors associated v/ith leadership: A survey of the literature. Journal of P sychol ogy, 1948, 2^, 35-71. Stogdill, R. M. and Coon, A. E. Leader behavior: Its description and measurem.ent Research Monograph No. 8 8^ Columbus: Ohio State University, 1957. Tedeschi, J. T., and Lindskold, S. Socia l psychology: ^^_erdepe_ndenc e, interac tio n and influence New York: John Wiley, 1976. Term.an, L. M. A preliminary study in the psychology and pedagogy of leadership. Pedagogical Seminary 1904, 2, 413-451. Weitz, S. (Ed.). Nonverbal coirnnunication : Readings with comm.entary New York: Oxford University Press, 1974. Winer, B. J. Statistical princ iples i n experim ental_ des ign (2nd ed.).' New"York: McGraw-Hill, 1971.

PAGE 234

BIOGRAPHICAL SKETCH Patricia Ann Burris was born in Denver, Colorado, March 19, 1949. She lived with mother Marian, father Bill, brother, Tom and m.ost beloved dog Scotty Zvnn, in Denver, Pueblo and Lakewood, loving all the while the magnificent mountains, desert prairies, and clear, blue skies of beautiful Colorado. Patricia graduated from Alameda Sr. High in 1967, and attended the University of Colorado at Boulder until graduation in 1971, with a Bachelor of Arts degree in psychology. After having worked in the real-world for two years, Patricia left Colorado and entered the graduate program in social psychology at the University of Florida in Gainesville, in 1973. During her years in Gainesville, she developed an interest in sociolinguistics largely as a result of having the good fortune of joining the com.radship of the Gainesville Society for the Study of Coverbal Behavior. She also developed a life-long interest in the m.artial arts, met and married her life-partner. Gill, and was adopted by the embodimtent of perfection. Kiddle Ann the cat, who is now an equal member of the family. Patricia completed her Master of Arts in psychology in 1975, and continued on in the program as a doctoral student. In 22;

PAGE 235

223 1978, she:; moved with Gill and Kiddle to Cleveland. She served as an instructor in the Cominuriication Department at Cleveland State University and as an assistant professor in Psychology at John Carroll University. Patricia completed her Doctorate of Philosophy in psychology at the University of Florida in August, 19 81. She intends to eventually resettle with Gill and Kiddle in the west and to attem.pt to use whatever skills and education she has gained in the first 32 years of her life to work for the survival of the whales, seals, elephants, wolves, grizzlies, and all other anim.als great and small that mankind is so determinedly setting out to destroy and to help save and preserve their habitats that stupidity, arrogance, greed and the profit motive now plunder. ^ — ^ — ft'-^' r--^'^=';*] ~_ r J I— ^ fc.e^*.rtj^Lt..jfWM i>^i]iii>suB f

PAGE 236

my opm present as a di I ce ion i ation rtify t con and ation that I have read this study and that in forms to acceptable standards of scholar! is fully adequate, in scope and quality, for the degree of Doctor of Philosophy. \ '/ Marvin E, Prof esso'i S h aw f C ha i rip-an ^j -f Q €=:^-\7 nh O i ^'''-^' ^ as I certify tha opin ion it con foriTi sent ati on ana is f a di sse rrarion for t I have read this study and that s to acceptable standards of schol ally adequate, the degree of scope Doctor Gi: in arly V. Robert" C^ZjJ.Prof esso'r^of Psychoiooy I certify that I have read this study and that in flv opinion it conforms to acceptable st.andards ox scholar-Ly crosentation and is fallv adequate, in scope and quality, as a dissertation for the degree of Docror of Philosophy. ( ./? \ 'Nox"nian IM Msirkel Professor of Speech, An t h r o p o I o g y 'sycnoxogy

PAGE 237

H'. and that in I certify that I ha-ze read tnis stu n\y opinion it confoms to acceptable standards oc scho i-h.q'rion and is fullv adequate, in sc:ope and quali rsresentacxc as a dissertation the deqree ot Doctor c of Phi j-osophy Ali3£W ^->^ ^ 3 \Cl^:^Lt::— Thomas W. Simon Associate Professor of tniirosopn^ 1 nave rea-.i uhis st\;dy and tnat m to £iccGDtabj.e standards cd;' scholarly I certify tnat i!?.y opinion it confoms presentation and is fully adequate, in scope and quality as a dissertation for the degree of Doctor of Philosophy ) u PatriGi. Associci a Hi 'Millej: of Ps y'cno.Loqy ;r tat ion vzas submitted to the Gradual tv of the DapartiLierit o^: Psycho l(.;ay .I-.' 1 -I xov. J This dxs Liberal Arts and Sciences and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the deqree of Doctor of Philosophy, Aug; 19 Dean, Graduate


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ERXG4CDGN_E6ULPI INGEST_TIME 2015-01-14T20:21:00Z PACKAGE AA00026558_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES