 

 

it:
IS

.3‘ .112

 

 

NNERSITY LIBRARIES

rm; illit’lllliﬁli; llllil‘lll

3 1293 01015 4155

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

This is to certify that the

dissertation entitled
Role formation within hierarchical decision making

teams with distributed expertise: A role expansion
model

presented by

Douglas J. Sego

has been accepted towards fulﬁllment
of the requirements for

Ph.D.

d . Organizational Behavior
egreeln

 

 

c. (”f/MM

Major professor

Date 7////9‘%

MS U is an Afﬁrmative Action /Eq ual Opportunity Institution 0-12771

 

LIBRARY
Mlchlgan State
University

 

 

 

PLACE N RETURN BOX to monthl- chockout from your record.
1'0 AVOID FINES rotum on or baton date duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU IsAn Afﬂnnativo MUM“ Opportmlty Initiation
mm:

_..._.—__—_.- _ ,_

ROLE FORMATION WITHIN HIERARCHICAL DECISION MAKING
TEAMS WITH DISTRIBUTED EXPERTISE:
A ROLE EXPANSION MODEL
By

Douglas James Sego

A DISSERTATION

Submitted to
Michigan State University
in partial ﬁIlﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Management

1 994

ABSTRACT
ROLE FORMATION WITHIN HIERARCHICAL DECISION MAKING
TEAMS WITH DISTRIBUTED EXPERTISE:
A ROLE EXPANSION MODEL
By

Douglas James Sego

The purpose of this study was to examine the determinants of role expansion.
Three groups of factors were studied, leadership (consideration and initiating structure),
individual (experience) and situational (feedback). In addition, role expansion was also
examined as a predictor of individual decision making accuracy. Two additional
predictors of decision making accuracy were Stress and job knowledge.

Subjects (n = 240) were studied within four person teams using the TIDE2
computer simulation and assigned to one of the four experimental conditions. The two by
two design include experience (operationalized as experiencing one or three jobs within
the team) and type of feedback (operationalized as outcome feedback only or outcome
and cognitive feedback). In addition, the effects of job knowledge and stress on decision
making accuracy was examined.

Experience was found to have a Signiﬁcant negative impact on role expansion
while the leader behavior of consideration had a signiﬁcant positive impact. Both stress
and experience had Signiﬁcant negative effects on individual decision making accuracy.

Implications for future research are discussed.

© Copyright by
Douglas James Sego

l 994

Dedicated to my children Heather,

Joshua and Clarence

iv

ACKNOWLEDGMENTS

There are a number of individuals who helped me to this point, where graduate
school ends and my third professional life begins. The individuals mentioned below never
gave up on me even when I thought I would never ﬁnish. Their support, guidance,
and friendship will always be remembered. First, I would like to thank all of the graduate
students from the Organizational Behavior and Industrial/Organizational programs.
Particularly, those who worked in the Team Effectiveness Lab with me, Jen Hedlund,
Jean Phillips, Chuck Capps, Dave Waldschmidt, Dale Tuttle, Spence Tower, Linda Barrett
and especially Debra Major. They made those long hours of data collection not only
bearable, but frequently quite ﬁJn. Collectively, there were the best and brightest
colleagues and friends one could hope for. Special thanks to Anne O'Leary-Kelly who
always was there when I needed someone to believe in me. Finally, thanks to Samatha
Wolf, a very Special friend, who's will to survive will always help me keep my life in
perspective.

Of the faculty members who were a part of my education, I want to mention ﬁve.
My ﬁrst research project was with Steve Kozlowski. His guidance and help when I was
completely lost was never forgotten. Alison Barber, who agreed to be on my committee

V

at the 13th hour. I wish she had joined the faculty earlier so I may have had an
opportunity to work with her. Daniel Ilgen, whom I have the greatest respect for. As a
scholar and researcher his work speaks for its self. Most importantly, he always gave me
the most honest feedback about my strengths and weaknesses. Kevin Ford, who over the
years has been a teacher, mentor, critic, card partner, best man, and friend. His dedication
to his profession will always inspire me. John Hollenbeck, who brought me into the
program and was a mentor and friend at every stage right up to the end, including being
my committee chair. He did not give up on me and in the end was instrumental in my
ﬁnishing my program. I can not hope to match his achievements, but his example will
always be my goal.

Most importantly I would like to thank my wife, Virginia Unkefer. Her
encouragement lead me to apply to graduate school and her faith in my abilities helped me
to ﬁnish. Together we have faced the hardest time of our lives. I only hope I can be as

supportive for her.

TABLE OF CONTENTS

LIST OF TABLES ................................... ix
LIST OF FIGURES ................................... x
Chapter One: Introduction ............................... 1
Chapter Two: Roles ................................... 5
Overview .................................... 5
History of Role Theory ............................. 6
Classical Role Theory .............................. 8

NPI Role Theory ................................ 11
Vertical Dyadic Linkage Theory ......................... 12
Job/Role Differentiation Theory ......................... 15

Role Expansion ................................. 19
Leadership and Substitutes for Leaders ..................... 23
Research on Task Experience .......................... 26
Research on Feedback .............................. 28
Theoretical Model ................................ 30

Model of Role Expansion ............................ 31
Summary .................................... 47
Chapter Three: Method Section ............................. 49
Subjects ..................................... 49

TIDE2 Simulation ................................ 49
Research design ................................. 53
Manipulations and measures ........................... 54
Experience ................................ 54

Feedback ................................ 55

Leader characteristics .......................... 61

Role Expansion ............................. 62

Stress .................................. 66

Role incumbent knowledge ....................... 66

Team decision making accuracy ..................... 66

Procedures ................................ 66

Data Analysis .................................. 69

vii

Primary data analysis ........................... 69

Secondary data analysis ......................... 69

Power Analysis ............................. 69

Chapter Four: Results .................................. 73
Analysis of the Role Expansion Hypotheses ................... 73
Hypothesis 1 ............................... 73

Hypothesis 2 ............................... 73

Hypothesis 3 ............................... 73

Hypothesis 4 ............................... 78

Hypothesis 5 ............................... 78

Hypothesis 6 ............................... 78

Hypothesis 7 ............................... 78

Hypothesis 8 ............................... 78

Analysis of the Decision Making Accuracy Hypotheses ............. 80
Hypothesis 9 ............................... 80

Hypothesis 10 .............................. 84

Post Hoc Analysis of Decision Making Accuracy Data ............. 84
Chapter Five: Discussion ................................ 87
Role Expansion ................................. 87
Unsupported Hypotheses ........................ 88

Supported Hypothesis .......................... 91

Signiﬁcant results in the opposite direction ................ 91

Decision making accuracy ........................... 92
Unsupported hypotheses ......................... 93

Signiﬁcant ﬁndings ............................ 94

Limitations ................................... 94

Future Research ................................. 96
Conclusion ................................... 97
APPENDIX A: Coding instructions for leader/incumbent communications ....... 98
APPENDIX B: Handbook for command and control simulation ............ 99
APPENDD( C: Job Knowledge Test ......................... 105
APPENDIX D: Interactive Training Script ...................... 117
APPENDIX E: Debrieﬁng Sheet: Experiment 7 ................... 121
BIBLIOGRAPHY .................................. 122

viii

LIST OF TABLES

Table 1: Inter-rater reliability for the leader consideration and initiating structure

scales and text message coding ........................ 63
Table 2: Internal consistency for the task knowledge test across all four conditions. . .67
Table 3: The timing and order of procedures used in the experiment .......... 68
Table 4: Data analysis plan for the regression equation where role expansion

is the dependent variable ............................ 70

Table 5: Data analysis plan for the regression equation where team decision making

performance is the dependent variable ..................... 71
Table 6: Means and standard deviations of variables in the study ............ 74
Table 7: Intercorrelations among the variables in the study ............... 75
Table 8: Regression results with role expansion as the dependent variable ....... 76

Table 9: Regression results with decision making accuracy as the dependent variable . .81
Table 10: Post hoc regression results with decision making accuracy as the

dependent variable .............................. 85

ix

LIST OF FIGURES

Figure 1: Classical role theory model (Katz & Kahn, 1978) ............... 9
Figure 2: Vertical dyadic linkage role model (Graen & Scandura, 1987) ........ 13
Figure 3: Job/Role differentiation theory model (Ilgen & Hollenbeck, 1992) ...... 17
Figure 4: Text message content coding scheme ..................... 24
Figure 5: Role expansion model ............................ 32

Figure 6: The moderating effects of experience on the leader's behavior of
consideration to incumbent role expansion relationship ............ 36
Figure 7: The moderating effects of experience on the leader's behavior of
initiating structure to incumbent role expansion relationship .......... 38
Figure 8: The moderating effects of feedbaCk on the leader's behavior of
consideration to incumbent role expansion relationship ............ 40
Figure 9: The moderating effects of feedback on the leader's behavior of
initiating structure to incumbent role expansion relationship ......... 41
Figure 10: Incumbent and task characteristics as moderators of the role
expansion to task performance relationship ................. 43
Figure 11: The moderating effects of task knowledge on the role expansion

to task performance relationship ...................... 45

Figure 12:

Figure 13:
Figure 14:
Figure 15:
Figure 16:
Figure 17:
Figure 18:
Figure 19:
Figure 20:

Figure 21:

Figure 22:

Figure 23:

The moderating effects of stress on the role expansion to

task performance relationship .....................
Weights and values settings for the TIDE2 simulation (screen 1) .....
Weights and values settings for the TIDE" simulation (screen 2) . . . .
Process feedback form for the carrier position ..............
Process feedback form for the CAD position ...............
Process feedback form for the AWACS position .............
Process feedback form for the cruiser position ..............
Main effect of the leader's behavior of consideration on role expansion . .

Main effect of job experience on role expansion ..............

Interaction between the leader's level of initiating structure and the

type of feedback received on role expansion ...............
Main effect for stress on decision making inaccuracy performance . . . .

Main effect for experience on decision making inaccuracy performance. .

. .46

.51

. .52

.57

.58

.59

.60

.77

.79

.82

.83

.86

Chapter One: Introduction

Research on small groups has a history dating back before World War H. Over
the decades the focus of small group research has often changed. The change is due in
part to less research conducted by social psychologists and more research conducted by
organizational psychologists (Levine and Moreland, 1990). Organizational psychologists
are for the most part reacting to the realities in the work place. The popular press is ﬁll]
of stories of organizations downsizing, reductions in middle management, and the
restructuring work around work groups or self regulating teams.

The last decade has seen an increase in the amount of research conducted on teams
(Ilgen, Major, Hollenbeck, & Sego, 1993) and in particular on decision making teams
(Davis, Holt, Spitzer, & Stasser, 1981; Sniezek & Henry, 1989). Two reasons contribute
to this increased interest. First, there are a number of dramatic examples of decision
making team failures. In the case of the USS Stark (1987), the decision making team
failed to take action that would have placed the ship in a stronger position to defend itself
from the Iraqi war plane. In the case of the USS Vincense (1988), the decision making
team failed to interpret the available data correctly and shot down an Iranian civilian
airliner. Finally, in the case of the Bureau of Alcohol, Tobacco and Fire Arms (ATF) in
Waco, Texas (1993), the decision making team relied on the judgments of experts in the
areas of law, religion, psychology and military tactics. The result was a decision to take
the Branch Davidian compound by force, a decision that will be questioned for a long
time.

A second reason for the increased interest in teams has to do with the challenges

1

2

facing businesses today. The business environment is becoming ever more global and, as a
result, organizations are having to deal with increasing complexity in the workplace and
greater global competition. This has resulted in organizations needing to ﬁnd ways to
decrease Operating costs while, at the same time, increasing productivity and effectiveness.
Teams are often implemented in organizations to achieve these goals.

To decrease costs, many organizations are becoming ﬂatter in their corporate
structures. As levels of management and the amount of supervision are being reduced,
workers are gaining greater control over their own work (Raynor, 1992) often as members
of teams (Solomon, 1991; Jacob, 1992). Teams can be found in many diﬂ‘erent types of
industries, such as the airline (Foushee, 1984), the military (Rouse, Cannon-Bowers, &
Salas, 1991), the automobile (Fortune, 1992; Moskal, 1992), and the consumer goods
industries (Jacob, 1992).

Current research on teams looks into several issues related to teams including
decision making accuracy (Guzzo, 1986; Stevenson, Busemeyer, & Naylor, 1990),
training (Cannon-Bowers & Salas, 1990), and communication (Kleinman, 1990). One
area not currently receiving much attention is how decision making team members deﬁne
and develop their individual roles.

Understanding how roles develop in decision making teams is important for two
reasons. First, incumbents of decision making teams are different from the type of
incumbents that are Often studied in role research. Instead of the role being deﬁned as a
set of observable behaviors that incumbents engage in to perform a task, incumbents in
decision making teams use information to perform their tasks. Although there may be

little diﬂ‘erence between these two groups, there is no research on roles in decision making

teams to base a comparison on.

Second, incumbents in decision making teams are often highly skilled in Speciﬁc
content areas. This high level of specialized competency is important to the team's overall
performance in making decisions. However, specialized competency can create situations
where team members have very different interpretations of a single event or situation.
Resolving this conﬂict between team members may create new problems. One possible
solution is for the team members to learn portions of each other's role so that they can
more clearly see how each piece of information ﬁts together to create a whole. However,
how much sharing of individual team member's roles can take place before the advantages
of specialized expertise are compromised? Another question is: what are the
consequences on performance when speciﬁc roles are shared by more than one team
member? To answer these questions it is important to understand what factors effect role
development, and then to look at how the role development of individual team members
effects their decision making performance.

Although several different types of teams have emerged, one of the most
ubiquitous types is the hierarchical decision making team with distributed expertise. The
central deﬁning features of this type of team are the following. First, the primary purpose
of this team is to collect and share information, and to make judgments or decisions based
on an analysis of the information. Decision making team members come together either
face-to-face or through technology mediated communication to perform their jobs.
Second, the leader is accountable for the decisions made by this type of team.
Occasionally, such as in juries, anonymity within the decision making process is important.

However, within most situations in organizations, accountability is desirable. Third, team

4

members have distributed expertise to help cope with the increased complexity of work.
Flexible manufacturing maintenance teams (Bergstrom, 1992), multi-functional design
teams (Jacob, 1992) and hostage rescue teams (Ling, May 3, 1993) are examples of teams
that require individuals with highly specialized expertise to come together for a common
purpose.

The purpose of this research was to test antecedents of the role development
process, in hierarchical decision making teams with distributed expertise. This study
looked at characteristics of the leader, and the role incumbent, as well as situational
factors and how they effect the way incumbents develop their roles. In addition, factors of
the incumbent and situation are posited as substitutes for behaviors of the leader in the
role deﬁnition process. A secondary purpose of the study was to look at the impact of
role development on decision making performance. Two predicted moderators of this

relationship are incumbent job knowledge and task related Stress.

Chapter Two: Roles
Overview

The purpose of this chapter will be to review several literatures relevant to role
theory and the model of the role development process presented later. The ﬁrst section
provides a historical overview of role research across several different research domains.
The second section reviews classical role theory as developed by Katz and Kahn (1978).
The focus here will be on the role sender inﬂuence in the role deﬁnition process. The third
section reviews Naylor, Pritchard, and Ilgen's (1980) theory of roles. The focus of
this theory is on the role incumbent's inﬂuence in the role development process.

Both Katz and Kahn (1978) and Naylor, Pritchard and Ilgen (1980) suggest that
the leader is inﬂuential in the developing of subordinate's roles. Thus several theories of
leadership that are important to this investigation of roles are reviewed. The fourth
section reviews the vertical dyadic linkage (VDL) literature. VDL is framed as a leader-
incumbent negotiation model. It describes the negotiation process where roles are
deﬁned. The ﬁfth section reviews the recent attempt to integrate the job and role
literature by Ilgen and Hollenbeck (1991). Their job role differentiation theory extends
VDL by focusing on what is negotiated and why. The sixth section takes a closer look at
the construct of role expansion which was introduced in the preceding section. Role
expansion will be discussed within the framework of the incumbent, whose job is to make
decisions. The seventh section brieﬂy reviews the leadership and substitutes for leadership
literatures. Most role theories place the supervisor or leader in a central position in the
role deﬁnition process.

Consequently, possible alternatives for the leader's behavior are proposed within

5

6

the role expansion process. These alternatives include previous experiences and the type
of task feedback. The eighth section reviews the experience literature. This literature
Often views experience as a Single dimension. Experience will be shown to be
multidimensional and these other dimensions are important when looking at experience's
impact on role development. The ninth section reviews the feedback literature. The focus
of this review will be on the differences between outcome feedback and cognitive
feedback. The ﬁnal section develops a theoretical model of the role expansion process.
This will include predictions that incumbent, situational and leader characteristics have
direct effects on the roles expansion process. The model also predicts a direct eﬁ‘ect
between incumbent role expansion and task performance moderated by incumbent and
Situational characteristics.
History of Role Theory

Role theory research has a long history reaching back to 19403 and 19503 (Biddle,
1979). As with most ﬁelds of study, much of the early theoretical work attempted to
deﬁne what roles were and to describe the boundary conditions that existed when deﬁning
roles. Biddle states that some theorists conceived roles as primarily overt behaviors
(Davis, 1949) or covert processes (N ewcomb, 1950). On the other hand, roles have also
been theorized as learning behaviors and thus controlled by internal variables (Child, 1963)
or controlled by others (Cattell, 1963). Most of this early work on roles was done by
sociologists and social psychologists.

In a recent review of the role literature, Biddle (1986) listed ﬁve perspectives of
role theory. The ﬁrst is functional role theory which believes roles are shared norms that

are found in stable social systems. Two shortcomings of this theory are the requirements

7

of shared norms and stable social systems. These two requirements lack credible
evidence, according to Biddle.

The second theory is symbolic interactionist role theory that places the focus on
individual actors. Here roles are thought to represent attitudes, contextual demands and
norms as understood by the actor or role incumbent. Although this perspective also
includes the concept of norms, it does so while recognizing that it is the actor's
interpretation and not the norm that is important.

The third role theory is structural role theory. Structural role theory focuses on
stable organizations of sets of persons sharing the same set Of behaviors. This theory
places the importance of understanding roles on the social environment instead of on the
individual. By placing little emphasis on understanding individual differences, structural
role theory has a modest following.

The fourth theory iS cognitive role theory that for the most part focuses on role
expectations. Areas of research interest include the conditions that lead to expectations,
the measurement of expectations, and the effect of expectations on behavior. Although
cognitive role theory has been criticized for relying too much on contemporary American
culture, it does have a large empirical research base.

The ﬁfth theory is organizational role theory. The focus here is on preplanned
social systems that are task oriented and hierarchically organized. Roles, in organizational
role theory, are social positions with normative expectations. Organizational role theory
includes the concept that norms can vary between individuals holding the same social
position and that the inﬂuences on individuals include ofﬁcial policies as well as peer

pressure. The relationship between role expectations and behaviors is also important in

these theories.

Organizational role theory will be the framework used here to develop a model of
role expansion. The focus is on how individuals expand their roles in hierarchically
arranged decision making teams where team members have distributed expertise. The
following is a brief review of four theories of roles in organizations that were used as the
foundation for the predictions of role expansion made in this study.

Classical Role Theory

Classical role theory (Katz and Kahn, 197 8) is an example of an organization role
theory. Classical role theory is primarily a two-person theory of roles. The two
participants are the role sender and the role receiver. The role sender refers to someone
who has expectations about what behaviors are appropriate for the role receiver. These
expectations include formalized standards for the social position held, such as a job
description, and personal expectations the role sender has for the role receiver beyond the
job description. The role sender is often a supervisor but can be a co-worker or even a
subordinate. The role receiver refers to the person who occupies the organizational
position of interest. I will refer to this person as the role incumbent.

In Katz and Kahn's role theory, the role deﬁnition process starts with the role
sender communicating his/her role expectations to the role incumbent. This is represented
in Figure 1 by the top line between the role sender and the role incumbent. The sent role
is then received and decoded by the role incumbent. In the next step, the role incumbent
engages in some behavior that he/she believes meets the role sender's expectation. Finally,

the role incumbent's behavior is observed by the role sender who decides if the behavior

9

Figure 1: Classical role theory model (Katz & Kahn, 1978).

 

Role sender

 

Focal person

 

Role Sent
expectation role

 

 

 

 

Feedback

 

 

Perceived Role

role

 

behavior

 

 

 

 

10

matches his/her expectations. If it does not, this process is repeated until either the role
incumbent does what is expected or the role sender changes his/her role expectations.

Katz and Kahn (1978) state that they:

consider role-sending as a continuing cyclical process by means of which
each person is socialized into a particular organizational role, informed
about the acceptability of his/her behavior in relation to the requirements of

that role, and corrected as necessary. (p. 187)

Katz and Kahn (1978) note that the role incumbent is not a completely passive
participant in the role deﬁnition process. When the role incumbent receives the sent role
the individual decodes the message and decides how to respond. How the incumbent
interprets the message and chooses to respond through behaviors is largely under this
person's control. This choice is likely to be effected by the role incumbent's experience,
ability, and job knowledge. These act as a ﬁlter to the received role expectations and
inﬂuence the direction of his/her behaviors.

Classical role theory is a closed loop system involving two people. Role
information is sent, received and acted on. The observable behaviors are then used as
feedback and analyzed to determine if the role incumbent ﬁllﬁlled the role sender's
expectations. This process continues until there is no discrepancy between the role
sender's expectations and the incumbent's behaviors. This perspective of role sending and
role taking places most of the responsibility for deﬁning roles on the role sender, who

initiates the role deﬁnition process. Consequently, it would be expected that certain leader

11

behaviors would be very inﬂuential in the role deﬁnition process.
NPI Role Theory

Another theory of roles in organizations was developed by Naylor, Pritchard and
Ilgen (1980), and is typically referred to as NPI theory. Consistent with organizational
role theory, these authors describe roles as stable patterns of interdependent behavior
from individuals who occupy social positions within an organization. The patterns of
behaviors enacted by the role incumbent are central in their theory. Acts are the basic
processes of behaviors and are conceptualized as having the dimensions of time and
effort. These two dimensions are not easily observed or measured. Consequently,
outcome products are used as evidence of role behavior.

Two additional requirements are needed for behaviors to be role related according
to NPI theory. First, the behavior must be relevant to the role. Top management
personnel Often get involved in charity work, such as serving on a local art counsel.
However, this may not be relevant to their jobs and thus is not part of their role. Second,
the product must be important enough to be evaluated by someone. For instance, a
personnel ofﬁce worker may post the company's stock quote each day so that workers
who are vested in company stock can track their investments. However, if this is not part
of the worker's job and no one evaluates him/her for doing this, then, according to NPI
theory, it is not part of his/her role. Thus if the role behavior outcomes are not relevant to
the role or important enough to be evaluated by others, then NPI theory states that the
outcomes are not role related.

A signiﬁcant difference exists between NPI theory and classical role theory (Katz

and Kahn, 1978). Classical role theory posits that the role sender initiates the role

12

deﬁnition process, while NPI theory posits the role incumbent initiates this process. This
is an important distinction. NPI theory allows the role incumbent to take a more proactive
stance in deﬁning his/her role. In classical role theory, the role incumbent is generally
reacting, although with volition, to expectations of a role sender.

These two theories offer examples where both the role sender and the role
incumbent initiate the role deﬁnition process. A commonality is found in that both
theories include a negotiation process, where roles are deﬁned during the negotiations that
go on between the role sender and role incumbent. The following theory takes a closer
look at this negotiation process.

Vertﬁl Dyadic Linkage Theory

The Work by Graen and Scandura (Graen, 1976; Graen & Scandura, 1987) looks
at the process of deﬁning roles within organizations. Their focus is on dyadic relationships
and the negotiation processes that go on between the supervisor and the subordinate (role
incumbent). They refer to this process as the vertical dyadic linkage (VDL). VDL is
primarily relevant for unstructured tasks that are not easily included in a formal job
description. These tasks require that the supervisor and role incumbent negotiate the
incumbent's role using a three-stage process that is depicted in Figure 2.

The ﬁrst stage in this process, shown in Figure 2a, is the sampling phase and is
very similar to the Kahn, Wolfe, Quinn, Snoek, & Rosenthal (1964) role episode model.
The supervisor initiates the process by sending the role incumbent his/her expectations.
The role incumbent receives and then acts, responding to supervisor's expectations. The
role incumbent's behaviors are then used by the supervisor as feedback and are used to

determine if they meet his/her expectations. Graen and Scandura (1987) state that in this

12

deﬁnition process, while NPI theory posits the role incumbent initiates this process. This
is an important distinction. NPI theory allows the role incumbent to take a more proactive
stance in deﬁning his/her role. In classical role theory, the role incumbent is generally
reacting, although with volition, to expectations of a role sender.

These two theories offer examples where both the role sender and the role
incumbent initiate the role deﬁnition process. A commonality is found in that both
theories include a negotiation process, where roles are deﬁned during the negotiations that
go on between the role sender and role incumbent. The following theory takes a closer
look at this negotiation process.

Vertical Dyadic Linkage Theory

The Work by Graen and Scandura (Graen, 1976; Graen & Scandura, 1987) looks
at the process of deﬁning roles within organizations. Their focus is on dyadic relationships
and the negotiation processes that go on between the supervisor and the subordinate (role
incumbent). They refer to this process as the vertical dyadic linkage (VDL). VDL is
primarily relevant for unstructured tasks that are not easily included in a formal job
description. These tasks require that the supervisor and role incumbent negotiate the
incumbent's role using a three-stage process that is depicted in Figure 2.

The ﬁrst stage in this process, shown in Figure 2a, is the sampling phase and is
very Similar to the Kahn, Wolfe, Quinn, Snoek, & Rosenthal (1964) role episode model.
The supervisor initiates the process by sending the role incumbent his/her expectations.
The role incumbent receives and then acts, responding to supervisor's expectations. The
role incumbent's behaviors are then used by the supervisor as feedback and are used to

determine if they meet his/her expectations. Graen and Scandura (1987) state that in this

13

Figure 2: Vertical dyadic linkage role model (Graen & Scandura, 1987).

ENVIRONMENT

 

 

  
     

   

 

 

 

 
    
   

 

 

     

 

 

 

 

 

Superior 3. Member
STRUCTURE
Sent / \
LEADER
Role Received
EXPECTATIONS A Role
Received
Response Response
\ BEHAVIOR J
WORKING
RELATIONSHIP
Opportunity RESOURCES ”CW“
OIIOI % GHQ!
Perceived ACTIVITIES 3°“
Role 4f Role
MEMBER
'\ EXPECTATIONS /
DCR

 

/_

Behavior

 
   

Perceived

 

 

   

POTCO IVOd

NOI’I‘TIS

   
 
   

 

ACTIVITIES
Behavior

 

NOfﬂ'IS

   
 

 

DCR

Role Taking

Role
Making

Role
Routlnization

14

phase the supervisor acts and the role incumbent reacts.

The second phase in Grain and Scandura's role deﬁnition model is the role
development phase. During this phase the supervisor and role incumbent test out different
dyadic interdependencies. This is depicted in Figure 2b. The third phase, the commitment
phase, is where the role is routinized (Figure 2c). According to Graen and his colleagues,
over time the dyadic relationship between the supervisor and role incumbent is established
and becomes relatively stable.

Many studies have looked at VDL as a leadership theory in order to explain the
relationship between the supervisor and incumbent (Dansereau, Cashman, & Graen, 1973;
Vecchio & Gobdel, 1984). The quality of the dyadic relationship reﬂects whether the
incumbent is a member of the supervisor's in-group or out-group. In—group members are
generally given more desirable jobs and more latitude in doing their jobs than are out-
group members. AS an outcome of the role development phase (where roles are
negotiated), some incumbents are allowed to add to their roles while other are not.

VDL is important in understanding the process by which incumbents deﬁne their
roles. First, it Shows how roles are negotiated through the dyadic interactions between a
supervisor and a role incumbent. One result of the negotiation process is members being
considered in- or out-group members. Second, it demonstrates how different incumbents
holding the same job develop different roles. Those incumbents that are in the supervisor's
in-group are more likely to develop greater role responsibilities then those in the
supervisor's out-group. This results in incumbents with the same job having much
diﬁ‘erent roles.

The prominence of the supervisor and incumbent dyad has some limitations,

15

however, when it comes to explaining the role deﬁnition process. Most other role theories
posit that individuals who are not the role incumbent's supervisor can still be inﬂuential in
deﬁning the incumbent's role (Biddle, 1986; Katz & Kahn, 1978; Naylor et al., 1980).
Naylor et al. (1980) refer to this group of individuals collectively as the role incumbent's
role set. Katz and Kahn (1978) posit that peers and subordinates can be role senders.

Seers (1989) attempts to address the exclusion of non-supervisors in the dyadic
negotiation process though the concept of team member exchanges (TMX). TMX is still
a role theory that centers around a dyadic relationship. However, TMX allows for
individuals other than the supervisor, such as the incumbent's peers, to inﬂuence the role
development process. In addition, TMX studies the effects of multiple dyadic
relationships on each incumbent.

This review of the role literature has pointed out important constructs and issues
related to the role development process. There still remains an important issue when
considering roles and role development. That issue focuses on the distinction between
jobs and roles. In the job and role literatures there is ambiguity about the actual
boundaries between jobs and roles. The following section will review recent theoretical
work that attempts to clear up this ambiguity, ﬁrst by clearly deﬁning a job and a role and
then by stating where the similarities between the two end.

Job/Role Differentiation Theory

Each person in an organization is responsible for performing a set of duties or
tasks that are sometimes called task elements. The task elements deﬁne a job and the
behaviors that are expected of individuals in a job. Task elements are formalized within

the organization in the form of a job description, worker handbook or the intrinsic

16

demands of the tasks themselves. Ilgen and Hollenbeck (1991) deﬁne a job as:

a set of task elements grouped together under one job title and designed to

be performed by a single individual (p. 173).

It is important to note that the job is deﬁned by the group of task elements and not by the
individuals who occupy the job.

There is some confusion between the concepts of jobs and roles. This confusion
results in part because a job is at the core of a role. The distinction between a job and a
role is that a role is an individual's expansion of his/her job. Figure 3 shows the
relationship between a job and role as explained by Ilgen and Hollenbeck's (1991) job/role
differentiation theory. The center box represents the task elements that make up the job.
In Ilgen and Hollenbeck's view jobs have four basic characteristics: 1) they are created by
prime beneﬁciaries, 2) they are objective, 3) they are bureaucratic and 4) they are quasi
static. These characteristics clearly suggest that jobs exist independent of the individuals
who occupy them.

The deﬁnition of a role includes the above deﬁnition of a job with the addition of
those things that an individual either brings to the job or negotiates to have become part
of the job. Ilgen and Hollenbeck (1991) call these emergent role elements. In Figure 3
the role is represented by the outer box. This shows the role encompassing the job. The
area between the borders of the two boxes represents the emergent role elements. It is in
this area of the role that explains how two incumbents can have the same job but can have

considerably different roles.

17

Figure 3: Job/Role differentiation theory model (Ilgen & Hollenbeck, 1992).

 

 

 

 

Role
Role
Emergent Role
Job Job

 

 

 

 

 

 

 

 

 

Elements

 

 

 

18

Job role differentiation theory is useﬁJl in clarifying two issues. First, it clearly
deﬁnes the distinction between jobs and roles. The role is not unique in comparison to
the job but rather is an extension of the job. Second, it expands VDL by explaining two
factors. The ﬁrst factor is what is negotiated in the role deﬁnition process (i.e., emergent
role elements). The second factor is why emergent role elements need to be negotiated
(i.e., to expand the basic job to include those things the organization needs to have
accomplished that are not in the actual job description).

Job elements are stable over time and are the same across incumbents within a job.
Emergent role elements are negotiated between the supervisor and the incumbent
according to most role theories. Emergent role elements are necessary for performing a
job because of the limitation in deﬁning a job using a job descriptions. In attempting to
be precise and parsimonious, a job description cannot be all-inclusive. An example of
this problem can be seen during management and labor contract negotiations. Sometimes
union ofﬁcials will threaten to tell their members to do only what is required of them
according to their job description. This is a powerﬁrl threat because so much of what
workers do and what the organization needs to have done is not spelled out speciﬁcally in
the job description but is negotiated and subsequently becomes part of the workers' roles
in the form of emergent role elements. Consequently, the workers' roles are expanded
beyond their basic jobs to meet the needs of the organization. Role expansion is the
process of adding emergent role elements to a set of task elements that make up and

deﬁne a job. The ﬁnal result of this process is the worker's role.

19

Role Expansion

If jobs are expanded to form roles and two workers are likely to expand their roles
in unique ways, then it is important to understand the antecedents of role expansion.

What are the antecedents and consequences to the role expansion process? The two parts
of this question are explored in this research. There is little or no empirical research in the
Industrial/Organizational literature on the role expansion process, with the exception of
VDL. Instead, the empirical research on roles has focused on issues such as the impact of
conﬂicting expectation of an incumbent's role (i.e., role conﬂict) or the uncertainty over
exactly what the incumbent's role is (i.e., role ambiguity).

Yet there are reasons to look at the process itself. If role expansion can be
predicted, then there could be implications for job design. It is possible that different ways
of designing jobs could result in incumbents developing different roles. There are also
implications for other functions within the organization. When a worker retires, does an
organization recruit someone to ﬁll the job as described in the job description or the role
that was developed by the newly retired worker? If the new employee does not have the
skills or background to ﬁll the emergent task elements of the retiree's role, then how does
the organization go about ﬁlling this void? Do other workers expand their roles to ﬁll the
void or is it the expectation that the new worker will? If the latter is the case, then it will
be important for the organization to create the appropriate conditions to help insure that
the new employee expands his/her role in a way that meets the organization's needs.

Job re-design can also be effected by role expansion. An organization may make
changes to a job to increase autonomy (Hackman & Oldham, 1975) because they believe

that this will lead to greater job satisfaction. The outcome might be disappointing if no

20

consideration is given to how job incumbents have expanded their roles. It is possible that
some or all of the incumbents have created roles that already give them greater autonomy.
The result of job re-design might be no increase in autonomy and hence job satisfaction.
Consequently, understanding role expansion should be of interest to an organization.

What exactly is role expansion? In its simplest form role expansion is the addition
of emergent task elements to a person's role. An example of role expansion is an office
clerk who learns how to use a new computer graphics software package even though there
is no mention of computer graphics in his/her job description. This Skill has utility to the
organization because reports can be made with interesting graphics. Using computer
graphics is not formally a part of the office worker's job because of the newness of the
technology but its use could become part of his/her role. Because most job descriptions
do not change with each new change in technology, workers' roles are expanded to
include the changes. For the purposes of this study, role expansion is deﬁned as workers
adding job relevant task elements to their jobs that go beyond the formal description of
their jobs.

This deﬁnition might suggest that role expansion is a uni-dimensional construct,
the simple adding of task elements. However, few constructs are that simple and role
expansion is no exception. There are many different types of task elements found in any
given job in organizations. Depending on the mix of task elements that are added to a job,
the resulting roles within any given job could be quite different. This study will look at
one type of job, one with decision making that involves multiple interdependent
participants. The decision making job will be the boundary for the following discussion on

multiple dimensions of role expansion.

21

The starting point for looking at role expansion is the job the worker holds. If the
job is one of making decisions, then the task elements have to do with the gathering,
sharing and processing of information. Because processing information is largely a
cognitive event, it will only be considered here in the context of the outcome of the
cognitive process. The outcome for this type of task is a judgment or decision.

The information exchange process takes place through statements, both verbal and
written. Statements can be classiﬁed as either transmitting information (sharing) or
attempting to collect information (gathering). Sharing in the sense of role expansion is the
act of transmitting information to others that exceeds the communication job requirements
of the person transmitting the information. Gathering in the sense of role expansion is the
collecting of information that is not directly part of a person's job. Sharing and gathering
information is considered role expansion to the extent that these acts go beyond the
requirements of the job and are job relevant.

Both sharing or gathering information statements can be further divided into those
that are either job related or non-job related. Job related statements include information
that is useful to the decision task. A statement like "I need the current interest rate on 4
year certiﬁcates of deposit (CDs)" is an example of a job related information gathering
statement in an investment decision making task. A statement like "You can only invest
4% of your gross wages" is an example of a job related information sharing statement in
the same context. Non-job related statements include socializing comments or comments
of encouragement or frustration. For example, a statement like, "don't worry about that
last loan decision, most new loan ofﬁcers make that same mistake, " is meant to console a

team member and is not directly related to the task. These types of statements do not add

22

to the team's ability to make decisions and, according to Naylor, et al. (1980), would not
be considered role relevant. Non-job related statements will not be considered role
relevant in this study.

Another form of job related statements would be information that is intended to
develop a strategy among team members. These statements are job related because the
purpose of the statement is to improve the team's decision making performance by
coordinating team members' actions. For example, a statement like, "send me information
on the interest rate of 4 year CDs every time you get updates" is a job related strategy
forming statement. The purpose of this statement is to reduce the need to ask if there has
been an update on interest.

Finally, if a statement is job related then it can be in one of two forms, raw or
translated. Raw information has not been translated into terms that can be understood by
non-experts. No reﬁnement or value judgments have been added to the original
information. For example, a report that states that a company had $15 million in sales last
year would be an example of raw information. From the information provided it would be
difﬁcult for a novice to evaluate what this sales ﬁgure means. However, if the report also
stated that the company made $2.5 million in proﬁt, and that this level of proﬁt was
outstanding for a ﬁrm of this size in this market, then a novice would now be able to
understand the meaning of the information provided. These examples demonstrate that
there are several ways that informational statements can be classiﬁed when considered
from a role expansion perspective.

In this study, all forms of communication that take place between team members

will be added together to create a global measure of role expansion. Due to the lack of

23

theoretical work in the area of role expansion in decision making teams this uni-
dimensional deﬁnition will be used when formulating hypotheses. In addition, written
communication will be content coded using the coding scheme Shown in Figure 4. This
will allow for an exploratory look into the possible multi-dimensionality of role expansion
in decision making teams.

Thus far, all of the theories discussed involve dyadic interactions, often between a
leader or supervisor and a role incumbent. Because of the importance the
leader/supervisor has in these theories, two leadership theories will be reviewed. The
purpose is to identify characteristics of the leader that will effect the role expansion
process. In addition, substitutes to those leader characteristics are examined.

Leadership and Substitutes for Leaders

Like the role literature, the leadership literature also has a long history of studying
hierarchical dyadic relationships. There are many different theories of leadership and can
be differentiated by their focus on the characteristics of the leader, the follower, and/or the
situation. One theory of leadership comes from the Ohio State University studies and
looks at two behavioral characteristics of the leader, consideration and initiating structure
(Stogdill, 1974).

Consideration is the degree to which a leader acts in a supportive manner and
Shows concern for subordinates (Yukl, 1989). Examples of consideration are asking for
ideas from subordinates and having a willingness to accept their suggestions. Initiating
structure is the degree to which a leader deﬁnes and Structures his/her own role and the
roles of subordinates (Yukl, 1989). Examples of initiating structure are offering new

approaches to problems and coordinating the activities of subordinates.

24

Figure 4: Text message content coding scheme.

/ Task \

Social Strategy
(1) (8)
Seek or
Provide
Seey “Vida
Role/Rule Role/Rule
Target Target
[72/ \Target / \rget
TransmIt Transmit
Non-trans R/R Non-trans.

/\“’ /\m

(3) MT) (6) (7)

25

Initially, it was thought that a good leader was a person who had rated high on both
factors. However, Kerr and his associates (Kerr & Jermier, 1978; Howell, Dorfrnan &
Kerr, 1986) have expanded on the original Ohio State University studies by showing that
characteristics of the subordinate and/or situation can often substitute for consideration
and initiating structure behaviors by the leader. For instance, a group that is highly
cohesive could substitute for a leader that has a high rating on consideration. Group
members would receive the same supportive behavior ﬁom the group as they would from
the leader (Howell & Dorfman, 1981; Howell, Bowen, Dorfman, Kerr and Podsakoff,
1990). This is a situational factor that acts as a substitute for leadership. Another
example is a subordinate who is highly professional and takes pride in his/her work and
accomplishments (Childers, Dubinsky & Skinner, 1990). Here there would be little need
for a supervisor who had a high rating on initiating structure. This is a subordinate factor
that acts as a substitute for leadership.

It is important to consider characteristics of the leader in conjunction with possible
substitutes for leadership. There are two possible substitutes for leadership that seem
most relevant to this discussion of role expansion. They are an incumbent's level of
experience and the type of feedback that the task provides.

Both Katz and Kahn (1978) and Naylor et al. (1980) recognized the importance of
these two factors. Katz and Kahn suggested that a role incumbent's past experiences will
affect how he/She responds to the sent role message. In addition, they theorized that an
individual can send a role to him/herself. In this case the incumbent will use past
experiences to help determine what role expectations will be sent. In developing their self

role model, Naylor et al. (1980) also stated that a role incumbent's past experiences are

26

relevant to the role development process.

Thus a role incumbent's past experiences can substitute for the leader's behaviors
of consideration and initiating structure. To the extent that the past experiences are
relevant to the current task, they may help the role incumbents structure their own tasks.
In this case there would be little need for the role incumbent to depend on the leader to
provide a supportive environment or to structure the task. Although experience has been
tested as a moderator between leader characteristics and general satisfaction and
organizational commitment (Howell & Dorfman, 1981, 1986; Childers, Dubinsky, &
Skinner, 1990), role expansion has not been examined as the dependant variable.

Another possible substitute for leadership is the type of feedback that is provided
by the task. Part of what the leader or role sender does is Observe the behaviors or
behavioral outcomes and provide the role incumbent with feedback. The feedback is then
used by the role incumbent either to reinforce current behaviors or to modify them.
Occasionally a task can provide this feedback directly. Past research has looked at task
feedback as a moderator (Howell & Dorfman, 1981, 1986), however, this has not been
done with role expansion as the criterion.

Experiences and feedback provided by the tasks might also have a direct effect on
role expansion. For these reasons the task experience and feedback literatures are
reviewed in the following two sections.

Research omTﬂErgoerience

Experience is often used as a variable in organizational behavior research. Perhaps

because of its wide use in common discourse, there has been little rigorous effort made to

conceptualize and operationalize experience clearly as a construct. Often when

27

researchers talk about experience on the job, they operationalize it as the amount of time
on the job (tenure). Schmidt, Hunter and Outerbridge (1986) did diﬁ‘erentiate between
organizational experience and job experience as the time in the organization and the time
on a particular job within the organization. In both cases experience is operationalized as
tenure.

Experience has also been conceptualized as the number of different jobs and
different departments within an organization where the incumbent has worked (Bons, Bass
& Komorita, 1970; Alvares & Hulin, 1973; Burroughs, Rollins & Hopkins, 1973). This
conceptualization of experience is based on the type of experience one has and not
on tenure. An example is a worker who is rotated from one job to another as part of a
program of job enrichment (Hackman & Oldham, 1975) or a manager who is rotated from
department to department to gain experience in several ﬁrnctional areas within an
organization (Burroughs et al., 1973). In these studies, the amount of time spent in any
one job or department is generally not important. What is important is that the individual
worked in several different jobs or departments. This kind of experience will be called
multiple job experience.

Experience has also been operationalized as the number of times a task is
performed (Spiker, Harper, & Hayes, 1985). Experience was conceptualized in terms of
the frequency of performing a task. This is referred to as task experience and is important
because of another assumption that is often made when experience is operationalized as
time. That is, more experience equals higher proﬁciency.

However, if time is used to operationalize experience, the implication is that two

individuals with the same tenure will have similar experiences (i.e., they will perform the

28

same set of tasks and they will perform the tasks the same number of times). It has been
shown by Ford, Quinones, Sego and Sora (1992), Schmitt and Cohen (1989) and Vance,
Coovert, MacCallun, and Hedge (1989) that this is not a tenable assumption. Ford et al.
have shown that workers with the same tenure in the same job varied greatly on the tasks
they performed as well as the number of times each task was performed. In this study, the
focus will be on the tasks that are performed

There are three ways to conceptualize experience: the amount of time on the job
(tenure), the different types of positions held in the past (multiple job experience), and the
frequency with which the task was performed (task experience). Multiple job experience
seems most relevant as a substitute for leadership. The multiple job experience is
important because this experience might lead a role incumbent to add to his/her current
role. In short, the greater the diversity in experiences, the greater the number of role
elements that are likely to be relevant and thus included in the incumbent's role.
Research on Feedbagk

Feedback plays a very important role in many areas in organizational behavior and
industrial psychology. Feedback is found in goal setting (Locke & Latham 1990), control
theory (Carver & Scheier, 1982; Campion & Lord, 1982), performance appraisal (Pearce
& Porter, 1986; Greller & Parsons, 1992) and learning (Schmitt, Coyle, & King, 1976). A
review by Ilgen, Fisher, & Taylor (1979) has shown the importance of when feedback
is given, whether the feedback is positive or negative, and the importance of the credibility
of the feedback source. Although there has been a great deal of research on feedback,
most of the studies reviewed by Ilgen et al. dealt with outcome feedback.

Outcome feedback is deﬁned as information on the overall result of performance.

29

An example of outcome feedback is a supervisor telling a worker that the last product
he/she produced did not meet quality standards. Outcome feedback does not provide
information Speciﬁc enough to isolate problems. More speciﬁc information is critical if
feedback is going to be useful in making changes that will lead to improved performance.

Recently a new line of inquiry into feedback has examined cognitive feedback
(Balzer, Doherty & O'Conner, 1989). Cognitive feedback refers to information about
relations rather than outcomes (Todd & Hammond, 1965). In a decision making task, this
might mean feedback on how different pieces of information are related, rather than
information about the correctness or accuracy of the decision.

There are three types of cognitive feedback: task information, cognitive
information, and functional validity information (Balzer et al. (1989). Task information
is the most relevant in developing a model of role expansion. Within the theory of

cognitive feedback, task information:

refers to relations between the cues and the criterion (Ye), information
about the criterion or the cues themselves, or both. (Balzer et al., 1989, p.

412)

The following example clariﬁes this deﬁnition and illustrates the difference
between outcome and task information feedback. A bank manager tells a loan ofﬁce
employee only that her last loan decision was bad for the bank. This is an example of
outcome feedback. If, instead, the bank manager had told the loan officer that her loan

decision was bad because the person seeking the loan did not have enough collateral for

30

the amount requested in the loan, then this would be an example of task information
feedback. The bank manager connected task information, the amount Of collateral, with
the criterion, whether or not to grant the loan request.

The discussion about role expansion and the factors effecting role expansion
include several references to tasks, task experience, task information and task elements.
To avoid any ﬁrrther confusion I will use the above deﬁnition of task information within
the theory of cognitive feedback but will refer to it as process feedback.

Theoretical Model

The model developed is created to explain how leaders, individuals and tasks can
inﬂuence the role expansion process within decision making teams where incumbents have
individual areas of expertise and are interdependent. This model tries to explain how
individual and task or situational factors contribute to the role expansion process.

The model makes the following assumptions. First, the model assumes that the
task is one that requires some interdependence with other people. This is because the two
main categories of emergent role elements in a complex decision making task are
gathering and sharing of information. A decision whether or not to close a factory is an
example of a complex decision task. This decision requires input ﬁom many people
including the organization's legal staff, its manufacturing department, labor unions and the
local government. Though each group is expert in its particular ﬁeld, it can supply only a
small part of the total information needed when considering a factory closing option.
Consequently, interdependency results from the situation where no one individual or
group could possess all the information and expertise required to make a complex

decision.

31

Second, the model assumes that the acts of gathering and sharing speciﬁc pieces
of information are examples of role elements. This is true as long as the information
gathered or shared is relevant to the role.

Finally, the model assumes that whether roles expand or not, the criteria for
evaluating performance remains the same. This is an important point because frequently
as a role changes, so do the standards for evaluating that role. However, this is not
necessarily the casein decision making tasks. The evaluation criterion remains the team
member's decision making accuracy. That is, individual team members are evaluated on
their contribution to the overall team decision. Thus, in decision making teams with
distributed expertise the team members are expected to process the information on which
they are the expert and make their recommendation based on that information. Regardless
of whether a team member expands his/her role to include other pieces of information or
not, the evaluation of his/her performance is still based on the quality of the judgment
he/She makes.

Model of Role Expansion

The role expansion model presented here de-emphasizes the dyadic negotiation in
the role expansion process. Instead this model allows for situational factors and
characteristics of the role incumbent to inﬂuence the role expansion process. Based on the
substitute for leadership literature it is also posited that the incumbent and situational
factors will also moderate the effect of the leader's behavior and role expansion.

The role expansion model is presented in Figure 5. Unlike much of the research on
roles that focuses on outcome variables such as role ambiguity and role conﬂict (Caplan &

Jones, 197 5; Dailey, Ickinger & Coote, 1986), the central purpose of this model is to

32

Figure 5: Role expansion model.

 

 

Incumbent
Characteristics

 

Expenence

 

 

 

 

 

 

Leader
Characteristics

 

Consideration

 

Initiating Struct.

 

 

 

 

 

Situational
Characteristics

 

Feedback

 

 

 

 

 

 

Role
Expansion

 

 

 

33

describe the factors that affect the role expansion process. This model tries to explain
why some role incumbents take on emergent role elements, thus expanding their roles
while others limit their roles to correspond more closely to the original job description.

The role expansion model attempts to build on VDL research (Graen, Orris, &
Johnson, 1973; Graen, Novak & Sommerkamp, 1982; Wakabayashi and Graen, 1984) by
looking beyond the leader incumbent dyad to include Situational factors. In addition, this
model leaves open the possibility that some Situational or incumbent factors might
substitute for the leader and consequently decrease the importance of the dyadic
relationship in certain situations.

The model also suggests that there are outcome consequences to the role
expansion process. Performance is posited as one outcome consequence of the role
expansion process. Performance is affected either positively or negatively by role
expansion depending on certain characteristics of the role incumbent and the task. The
role expansion to performance relationship is of secondary importance in this study. The
primary purpose of this study is to understand what factors effect role expansion behaviors
by the role incumbent.

The starting point of the role expansion model is the leader. Past research has
been mixed in determining the effects of initiating structure and consideration on such
outcomes as work group productivity, incumbent satisfaction, and turnover (see Stogdill,
1974). Despite this, one could make several predictions about the relationship between
these behaviors and incumbent role expansion.

When the leader's level of consideration is high, this should lead to greater

participation in the negotiating process because the role incumbent is receiving

34

encouragement and is being listened to by the leader. In general, it is predicted that
greater role expansion will occur when the leader's level of consideration is high than
when the leader's level of consideration is low. This prediction lead to the following

hypothesis:

Hyp 1: Higher consideration by the leader will result in greater role

expansion by the team members.

The effects of initiating structure will be quite different. When a leader's level of
initiating structure is low, there will be a greater amount of both sharing and collecting of
information in all categories by team members. This is similar to the case when
consideration is high. There will be less sharing and collecting of information in general
then when a leader's level of initiating structure is high. High levels of initiating structure
by leaders will create a more autocratic environment resulting in less information sharing

and gathering opportunities. This prediction lead to the following hypothesis:

Hyp 2: Lower initiating structure by the leader will result in greater role

expansion by the team members.

There are two other sources that are predicted to inﬂuence the role expansion
process of team members. They are divided into two categories as shown in Figure 5.
One category includes characteristics of the incumbent while the other category includes

characteristics of the task. Experience is the incumbent characteristic while the type of

35

task feedback is the task characteristic.

Experience is considered high when the role incumbent has held a variety of job
positions that are in the current task. If a role incumbent has only experienced one role
position then he/she will not be likely to expand his/her own role without inﬂuence from
the leader. If, however, the role incumbent has experienced several relevant role positions
in the past, then he/she will be more likely to incorporate some of those past role elements
into the current role. It is predicted that high experience will in general be related to
greater role expansion. Role incumbents are likely to attempt to incorporate their past
experiences into their role. This will lead to increases in all forms of information gathering

and sharing behaviors. These predictions lead to the following hypothesis:

Hyp 3: Higher amounts of past experiences on the part of the role

incumbent will be associated with greater role expansion.

The substitute for leadership literature suggests that there are times when
characteristics of the task or the incumbent make the leader's behavior irrelevant. For
example, when the role incumbent is high on past experience it is predicted that this will
make the leader's standing on consideration irrelevant. Past experience will act as a
substitute for consideration. When experience is high, it does not matter whether the
leader's level Of consideration is low or high. This relationship is shown in Figure 6 and

leads to the following hypothesis:

36

Figure 6: The moderating effects of experience on the leader's behavior of consideration
to incumbent role expansion relationship.

 

 

 

.c
2"
I
O
o
c
8 t
o
8 \ Vic-2 S
._ \ IDXO z
‘50.; \ 5'5“
(U .9 \ v.8
adj]: \\ 85
\ 48
\ 0
\
3
\ O
_I
.c
.9 8
I ...l
c
.9
m
8
%e
EU]

37

Hyp 4: Consideration and task experience will interact to affect role
expansion. The nature of this interaction is such that the relationship
between the leader behavior of consideration and role expansion is only
negative when experience is low. If experience is high it substitutes for the
leader's behavior, so that role expansion occurs regardless of the level of

consideration.

Another prediction has to do with the interaction between experience and initiating
structure. Once again experience is predicted to act as a substitute for a leader's behavior
(initiating structure). When there is a high amount of past experience it will make the
leader's standing on initiating structure irrelevant. The nature of the interaction is Shown

in Figure 7. The prediction leads to the following hypothesis:

Hyp 5: Initiating structure and task experience will interact to affect role
expansion. The nature of this interaction is such that the relationship
between the leader's behavior of initiating structure and role expansion is
only positive when experience is high. If experience is low it substitutes
for the leader behavior, so that role expansion is limited regardless of the

level of initiating structure.

A characteristic of the task that is of interest in this study is the type of feedback
the role incumbent receives. As mentioned before, the two types of feedback are outcome

and process feedback. Outcome feedback tells the role incumbent whether or not his/her

38

Figure 7: The moderating effects of experience on the leader's behavior of initiating
structure to incumbent role expansion relationship.

 

High

   

Expeﬁence
High

  
 

Past

InItIatIng
Structure

Past Experience Low

Low

 

 

High
Low

Role
Expansion

39

decision was correct. If it was incorrect, then the direction and magnitude of the error is
also contained in the feedback information. Process feedback gives the role incumbent
information about the speciﬁc pieces of information used in the decision task as well as the
relationship between these pieces of information. Process feedback should lead to
increased proﬁciency on the task. As proﬁciency increases, incumbents Should attempt to

expand their roles. Consequently the following prediction is made:

Hyp 6: Role incumbent's receiving process feedback will be associated

with greater role expansion.

Process feedback should also act as a substitute for the leader's behavior. The
leader's standing on consideration and initiating structure is made irrelevant when the role
incumbent is receiving process feedback from the task. The nature of the interaction is

shown in Figures 8 and 9 and in the following two predictions:

Hyp 7: Consideration and feedback will interact to effect role expansion.
The nature of this interaction is such that the relationship between the
leader's level of consideration and role expansion is only negative for role
incumbents receiving outcome feedback. If role incumbents only received
process feedback it substitutes for the leader behavior, so that role

expansion occurs regardless of the level of consideration.

Hyp 8: Initiating structure and feedback will interact to affect role

40

Figure 8: The moderating effects of feedback on the leader's behavior of consideration to
incumbent role expansion relationship.

 

 

 

.C
.9
\ I
\
\\ $36
\ Em
\ o-O
\ 0'0 C
x \ 5e .9
(no \ 0 e--
a)“, \OLI. he
0% \ m0
8o \ 13:9
ho \ (Um
O.I.L \ ‘1’:
\ 48
\
\
\
\
\
\\ 3
\ O
_l
.c 3
.9 o
I .J
C
.9
(D
8
-go.
mil

41

Figure 9: The moderating effects of feedback on the leader's behavior of initiating
structure to incumbent role expansion relationship.

High

   
 

Process
Feedback

    

Outcome Feed back
InItIatIng
Structure

Low

 

 

High
Low

Role
Expansion

 

42

expansion. The nature of this interaction is such that the relationship
between the leader behavior of initiating structure and role expansion is
only positive when role incumbents receive process feedback. When role
incumbents receive outcome feedback it substitutes for the leader behavior,
so that role expansion is limited regardless of the level of initiating

structure.

Understanding the factors that affect role expansion is important and is the primary
purpose of this study. It is also important to understand the link between role expansion
and performance. For example, if an incumbent's role is expanded and his/her
performance is poor, then the organization is likely to see role expansion as negative. On
the other hand, if performance is good, then role expansion will be seen as positive.

The second part the of the study examines the consequence of role expansion on
decision making performance. This study looks at two factors that are hypothesized to
interact with role expansion to effect the role incumbent's decision making performance.
The ﬁrst, task knowledge, is a characteristic of the incumbent. The second, the stress
involved in performing the task, is a characteristic of the situation. These are shown in
Figure 10.

In this study, task knowledge is ﬁrndamentally important to performance. Sharing
and gathering information will not lead to high performance unless the incumbent has the
task knowledge necessary to use the information. This is especially true when the

information that is being received is in an uninterpreted form. In this case the nature of

 

43

Figure 10: Incumbent and task characteristics as moderators of the role expansion to task
performance relationship.

Incumbent Characteristics

 

 

Task
Knowledge

 

 

 

 

 

 

 

 

 

1

ﬂ Task
Performance

 

Role
Expansion

 

 

 

 

 

 

 

 

 

Stress

 

 

 

 

Task Characteristics

44

the interaction between role expansion and task knowledge is described in Figure 11 and

the following hypothesis:

Hyp 9: Role expansion and task knowledge will interact to effect decision
making performance. The nature of this interaction is such that the
relationship between role expansion and decision making performance is
highest when task knowledge is high. This relationship is lowest when task

knowledge is low and role expansion is high.

There are several characteristics of the task that can have an effect on
performance. One of these is the amount of stress that is involved in performing the task.
Stress can be created from several sources including time constraints, the physical
environment where the task is performed and physical demands of the tasks (Ivancevich
& Matteson, 1980). One way to lessen the stress iS to share some of the work load with
others. This is what happens in hierarchical decision making teams with distributed
expertise. This is not to say that in these teams time constraints cannot create Stress, but
only that the nature of these types of teams help to reduce the effect of the stress.

Role expansion, however, could have a negative impact on team decision making
performance. If, over time, each team member takes on an increasing amount of the work
load in the form of emergent role elements, this would produce a role that is too large,
complex and difficult to execute under conditions of high stress. This effect is shown in

Figure 12 and lead to the following hypothesis:

45

Figure l 1: The moderating effects of task knowledge on the role expansion to task
performance relationship.

 

 

0
g) 8:
33 3
3 ‘3‘
2 2
X s:
21.: x
a? 88
i—I I—_I
.C
I .9?
, :I:
/
/
/
/ C
, .9
(D
/ s
/ %O_
,’ mil
/
/
I
/ 3
/ O
.J
‘1:
.9 8
I _J

Performance

Task

46

Figure 12: The moderating effects of stress on the role expansion to task performance

 

 

 

 

relationship.
VI m
m 8 .c
9 3 u 9
*" 0 a I
m _I [
.C
x .9
\ :I:
\
\
\
C
‘ .9
\ 2
\ (I) w
‘ '5 3
\ tr LLI
\
\
\
\
\ 3
\ _l
.C
.9 E
:I: .1

Performance

Task

47

Hyp 10: Role expansion and stress will interact to affect decision making
performance. The nature of this interaction is such that the relationship
between role expansion and decision making performance will be highest

when role expansion is high and stress is low. When role expansion is high

and stress is high the result will be the lowest decision making

performance.

Summa_ry

The primary focus of this study is to examine the antecedents of role expansion
with the secondary focus being on decision making accuracy as a consequence of role
expansion. Based on the theoretical work by Ilgen and Hollenbeck (1991) it was
predicted that the subjects in this study would add role elements to the jobs they were
assigned. Furthermore, it was hypothesized that role expansion would be predictable
based on characteristics of the leader, the situation, the individual, and their interactions.
Finally, it was hypothesized that role expansion would interact with job knowledge and
stress to effect decision making accuracy.

The ﬁrst part of the study deals with four factors that are predicted to effect role
expansion. First, it is posited that leaders that display high amounts of consideration or
low amounts of initiating structure behaviors will lead to increased role expansion by
subordinate team members. Second, it is posited that incumbents who experience more
than one job within the team will be more likely to expand their roles than incumbents
who experience only one job within the team. Third, it is posited that incumbents who
receive process feedback (task level cognitive feedback) will be more likely to expand

their role than incumbents who received only outcome feedback. Fourth, it is posited that

48

experience and feedback will act as substitutes for the leader's behavior of consideration
and initiating structure.

The second part of the study deals with the three factors that are predicted to
effect decision making accuracy. It is posited that under high stress conditions, role
expansion will lead to poorer performance, while under low stress conditions it will lead to
higher performance. Also it is posited that an incumbent with high job knowledge and
high role expansion will make the most accurate decisions and an incumbent with low job

knowledge and high role expansion will make the least accurate decisions.

Chapter Three: Method Section

Subjects

The subjects used in this study were students in an introductory management class
at a large Midwestern university. All students in the class were given an opportunity to
participate by signing up for a single, three-hour time slot. When they arrived at the
Team Effectiveness Research Lab, four subjects were randomly selected to participate in
the experiment. The remaining subjects, if any, were directed to another experiment that
was running concurrently in the lab. Each four-person team was then randomly assigned
to one of four experimental conditions. Each team participated in a Team Interactive
Decision Exercise for Teams Incorporating Distributed Expertise (TIDEZ) computer
Simulation (Hollenbeck, Sego, Ilgen, & Major, l99l).
TIDE2 Simulation

The TIDE2 simulation was set up for a naval command and control center
conﬁguration. The simulation program parameters were set up for a four person team that
had access to a total of nine information cues (target attributes). The four stations
included an aircraft carrier (the team decision maker and leader), a coastal air defense
(CAD) station, an advance warning airborne command system (AWACS) aircraft, and a
naval cruiser. The last three stations will be referred to as the outer stations and these are
the jobs that will be examined for role expansion in this study. The carrier will be referred
to as the leader.

The Object of the simulation was to determine the level of threat that an
unidentiﬁed aircraft represented to the team. Each team member gathered information
about the target that was thought to be necessary for making a decision. Speciﬁcally, the

49

50

outer stations made recommendations or judgments to the team leader, who in turn made
the overall team decision. Figures 13 and 14 Show the parameter settings for the TIDE2
Simulation. These are ﬁIrther described below.

All stations had access to three pieces of information (cues) about the target. The
cues had the following characteristics. First, there were three levels for each cue, non-
threatening, somewhat threatening, and very threatening. A cue that was non-threatening
was assigned a value of zero. A somewhat threatening cue was assigned a value of one.
A very threatening cue was assigned a value of two. Second, each cue was weighted for
its main effect. All cues had a main effect of zero except for Identiﬁcation Friend or Foe
(IFF) which had a cue weight of two. Thus, IFF could range from zero to four (i.e., if the
cue value was two the cue weight was two, 2 x 2 = 4).

The TIDE2 simulation allows for up to four interactions. In this study there were
four two-way interactions. Each interaction had a weight of one. This allows the
interactions also to take on values of zero, one, two, or four. As a result the one main
effect and four two-way interactions were weighted equally.

Each outer station had access to at least one interaction. The team leader had
access to the one main effect cue. This left one interaction that was not directly accessible
by any one outer station. This created a Situation where two stations had to communicate
with each other in order to gather and share information necessary to determine the value
of the fourth interaction.

The Simulation created two characteristics that were important for this study. The
ﬁrst is that there was distributed expertise among team members. Each outer station had

exclusive access to information and knowledge to translate the raw information into the

51

Figure 13: Weights and values settings for the TIDE2 simulation (screen 1).

mmvoz muﬂuz OO 3 HO .mwmcmno Hmucmo Ou O .omconu Ou H**H
+5500 How m .+mwuzmcmz ROM < .meOH >mapoz\3oﬂ> Ou H "umucm

Om Om me OH HmHHzxzxzxzxz H H H H HmHHuemHmz "mcoHeumRmucH

OO O OO om OON OOH E O emcee O O

O O O O m H mmmHO H emcee O O

OH Hm ON HH OH O use HE H maneum Reoo O O

OH O Om HH om Hm Ode H eoHeomuHo O O
O.H m.H m.H O. O. m. an: H mmH m O
O. OH- O O. OH O Owe m eHmca O O

OH OH OO Hm me He E m eNHm O m
OOOOH OOOO OOOmm HOOmH ooomm HOOmm OH H epsuHuHc O m
OOO Hmm Omm HOO OOH OOH eds H cemdm O H
IIIIIII HHHOIIIIII- HOHH HOB HO_ HOE *

m u OEHO uﬂm3 wmcmmmz Hog q n mGOOH # mm. m u moEOouso * HvH

m n mucmempsh * Hma e u mCOHuOonudH # Hm“ m mousmmoz * HHH

monam> a munmﬂwz 603.mem >MHUOZ

52

Figure 14: Weights and values settings for the TIDE2 simulation (screen 2).

Omvoz OOOHS O» 3 HO .mmmcmno Hoodoo Ou O .mmcmnu Ou Htwa
+8500 How m .+mwuzmmwz How d .mcooH %MOCOZ\BOO> O» H "Hmucm

O w w > "one on cue HONH w w O O O O w w w O O O H9H. axes one HON.
» O O O "moo 0» cue HONL O O O w w w O O w w w 9 ”9H. msHm> one HONH
O m m H O w w w w w w w w w w 9 H9O Russo Hmma
m m H O m H O m H O m m "OE
sonb HONIONH new mom O O O O m m m m m m m m H H H H HEoum
HHana 4 O m m H a O m m H < O m m H < O m m H O0 Hoebpmmm
m>ocm HNNL Rom Hem O O O O O m m m m m m m m m m H H H H H HHccHERmE
z w z z z z z w z z z z z w z w w w w w "xocncemm 3mH> Hmmm
O "O m\o HHNO m "Hm\o\mv some xoeppomm HON_
m- mmemmmHo
HI OzmoHOzH ON pseudo O HouHeoz Omm ummHsuu
O mmHz OH eonooo O 3mH>wm OOO 0<3<
H OOH: mmmz OH Somme m eeocOH OmH nae
N BHm HO CAMS o HOMO Oz mmm HOOHHOU
HOH_ .OHL HOHH HOH_ HOHH

mOSHm> O munmﬂmz BO3.OQXO %MOCOZ

53
appropriate threat level. Second, there was interdependence, in that members had to
share both information and expertise. There are seven possible responses to each target in
this conﬁguration of TIDE2 used in this study (ignore, review, monitor, warn, ready, lock-
on, or defend). Each response has a target value associated with it, as shown in Figure 15.
Thus, there is a true score or correct decision for each target. The outer stations'
judgments are compared with the correct decision to determine their judgment accuracy.
Research desiga

In this study there were seven independent variables and two dependent variables.
The independent variables included: leader consideration, leader initiating structure, job
knowledge, experience, feedback, stress, and role expansion. The dependent variables
included: role expansion and decision making performance. Note that role expansion was
used as both an independent and dependent variable. Role expansion was the dependent
variable in the ﬁrst analysis and an independent variable in the second analysis.

This study was divided into two parts. The ﬁrst part of the study had four
independent variables, two of which were manipulated and two of which were measured.
The two manipulated variables in this experiment were experience and feedback.

Subjects in the "low experience" condition remained in a single job for the entire
experiment. Subjects in the "high experience" condition were rotated between all three of
the outer station jobs during the experiment.

There were two types of performance information created by the feedback
manipulation. Subjects in the "outcome feedback" condition received feedback on how
accurate their judgments were compared with the correct judgment. In addition, they

were aware of the direction of their error (i.e., whether they were over aggressive or

54

under aggressive). Subjects in the "process feedback" condition received feedback on
each cue and interaction that they could access. This was in addition to outcome
feedback.

The two independent variables that were measured but not manipulated were the
leader's behavioral characteristics of consideration and initiating structure. The dependant
variable in the ﬁrst part of the study was the amount of role expansion that took place for
each outer station team member.

The second part of the study had three independent variables, one of which was
manipulated and two of which were measured. The manipulated variable was stress.
Targets in the last 16 trials were randomly assigned to either a "high" or "low" stress
condition. In the high stress condition subjects were given only 90 seconds to make their
judgments about the targets. In the low stress condition subjects were given 160 seconds
to make their judgments about the targets.

The two independent variables that were measured but not manipulated were job
knowledge and role expansion. Job knowledge was a measure of how well a subject
understood the information that made up their speciﬁc job. Role expansion was a measure
of the number of role elements that each subject added to their in the ﬁrst part of the
study. The dependent variable in the second part of the experiment was the decision
making accuracy of each station.

Manipulations amd measures

Experience. High experience was operationalized by rotating all outer station

team members in this condition though all of the outer station positions. In the low

experience condition the subjects worked at only one of the outer stations (CAD,

55

AWACS, or Cruiser). At the beginning of the experiment each subject was trained on the
job Speciﬁcs of one station in the high experience condition. Subjects had two practice
trials followed by six more trials. At the end of the eighth trial the outer station subjects
were given a new assignment (different computer station) followed by another Six trials.
The subjects were again rotated to new assignment after the 14th trial. In the low
experience condition subjects were assigned to one station and remained there for the
entire experiment.

Feedback. The feedback provided by the task was one of two types. The ﬁrst
was outcome feedback and consisted of the feedback on the accuracy of each station's
decision on each trial. This allowed each team member to determine the accuracy of
his/her judgment. In cases where the judgment was incorrect, the feedback also provided
the direction of the error. That is, whether the judgment was over aggressive or under
aggressive.

The second type of feedback was process feedback. In the process feedback
condition each team member received the same performance information for each trial as
that provided in the outcome feedback condition. In addition, the process feedback
condition supplied each team member with speciﬁc feedback on each information cue on
which he/she had experience. This included a representation of the raw data, the correct
translation of each piece of data and, in the case of interactions, the correct translation of
the interaction. For example, if the speed of the target was 600 mph, then the process
feedback repeated the raw value of the information (600mph) and the translated value,
which, in this case, is very threatening. Because speed interacts with direction, the result

of the interaction is also provided in the process feedback condition. In one example, the

56

direction was 5 degrees, which was also very threatening and the interaction between
speed and direction was thus very dangerous. Figures 15 through 18 show the process
feedback that each station received after the ﬁrst practice trial. Note that although two
stations had access to information on speed, only one station had access to information on
both speed and direction.

The two feedback conditions thus vary on both the level of speciﬁcity and in the
shear amount of information provided. The two feedback conditions receive the same
outcome feedback. In addition the process feedback condition receives information about
each cue that can be accessed directly. This information is more Speciﬁc than the outcome
feedback information. It could be argued that the feedback manipulation results in two
conditions with one condition simply receiving more feedback information than the other.
However, in this case, the nature of the added feedback is more important than the
volume. In the process feedback condition subjects receive information that should help
them understand where they made errors in interpreting the individual importance of each
cue. This should lead to greater learning of each cues and should result in subjects
in this condition out performing subjects in the outcome feedback condition.

Process feedback naturally provides a greater amount of information because it
provides more detail. However, it is the speciﬁcity of information and not the amount of
information that is critical to role expansion. Outcome feedback does not help isolate the
cause of errors in decision making. It only informs the decision maker that an error has
occurred. Decision making performance improvements would largely be a result of trial
and error learning under this condition, but this is difﬁcult given the complexity of the task

in terms of the cues to criterion relationships. Process feedback, however, does provide

57

Figure 15: Process feedback form for the Carrier position.

Altitude = 26,367 ft ----- non-threatening

IFF = .5 Mhz ------------- non-threatening

Radar = Class 9 ---------- very threatening

58

Figure 16: Process feedback form for the CAD position.

Speed = 325 mph ---------- non-threatening
----- non-dangerous

Direction = 27 dgs ------- non-threatening

Size = 27 meters --------- very threatening

59

Figure 17: Process feedback form for the AWACS position.

Angle = +15 dgs ------- very threatening

----- non-dangerous

Range = 200 miles ----- non-threatening

Radar = Class 9 ------- very threatening

60

Figure 18: Process feedback form for the Cruiser position.

Altitude = 26,367 ft ----- non-threatening
----- non-dangerous

Corridor Status = 5 mi --- non-threatening

Size = 27 meters --------- very threatening

61

information that is speciﬁc enough to isolate the errors in decision making. It thus allows
for greater Speed in taking corrective action that will lead to improved decision making
performance.

The method used to supply the process feedback to subjects after each trial was as
follows: at the end of each trial the team's feedback summary appeared on the screen. The
summary screen contained outcome feedback. A number varying from 1 to 99 appeared
at the bottom of the screen next to the word "CODE". This number referred to a speciﬁc
page in the feedback manual that each subject had at the computer stations. By matching
the number on the computer screen with the correct page in the feedback manual, subjects
received the process feedback for the trial they just completed.

Although the subjects had all of the feedback sheets available for viewing during
the simulation, it was impossible to use the feedback books to cheat. The order of the
feedback sheets was randomly determined and a large number of false feedback sheets
were included. Indeed, there were two pages of false feedback for every one page of
valid feedback. Any attempt to accelerate learning or enhance performance by looking at
the information in the feedback manuals was a waste of time. This was explained to the
subjects at the beginning of the experiment. A review of the text messages suggests that
only one team attempted to use the feedback manuals to help them make their decisions.
Further analysis showed that this did not help or greatly hurt their performance and after
six trials the team abandoned this Strategy.

Lgader characteristics. Characteristics of the leader were measured by content
coding the text messages for each leader/outer station dyad. This allowed for the leader to

be coded independently with respect to each member of the team. The leaders were rated

62

on two leader behavior characteristics, consideration and initiating structure. The scale
used and the instruction given to those coding the text messages are shown in Appendix
A.

Consideration and initiating structure were examined at the dyadic level. That is,
for each leader/subordinate pair (there were three pairs for each team) a separate set of
ratings were made. Two raters who were blind to the purpose of the experiment were
used to code all 240 leader/subordinate pairs. The inter-rater reliability was .88 for
consideration and .92 for initiating structure as shown in Table 1.

This method was chosen over using the leader behavior description questionnaire
(LBDQ) for two reasons. Fiist, most of the items of the leader behavior description
questionnaire are not relevant to the types of interactions that took place while
participating in the TIDE2 simulation. For example, the following is an items from the
LBDQ, "He does personal favors for group members" (Schriesheim & Stogdill, 1975). In
the TIDE2 simulation there were no opportunities to display a behavior like the one
described in the above LBDQ item. Second, by content coding the text messages, each
leader/subordinate pair could be coded. This also meant that the resulting ratings were
not self reports, but were the judgements of two independent raters.

Role Expamsion. The primary dependant variable for the ﬁrst part of the study
was role expansion. Role expansion refers to emergent role elements that are added to a
team member's job. Each subject was given a speciﬁc job at the beginning of the
experiment. These jobs require that each outer station member measure three attributes
of the unknown target. Two of the three attributes interact with each other and it is the

responsibility of each outer station member to measure and interpret this information. The

63

Table l: Inter-rater reliability for the leader consideration and initiating structure scales
and text message coding.

 

 

Number of
Scale Alpha Raters
Consideration .88 2
Initiating Structure .92 2

Text messages .94 2

 

64

ﬁnal interaction required the CAD to transmit size to the AWACS. Finally, each outer
station was required to make a recommendation to the team leader on how to respond to
each target. This constitutes the job of each outer station. There were no additional
requirement of the job for team members to communicate with one another, although the
ability to communication with one another was available to each team member.

Role expansion occurred when a team member engaged in behaviors that are not
explicitly deﬁned as part of their job. For example, if a team member Shared information
with another team member this would be considered role expansion because it goes
beyond that member's job description. Requesting information from other team members
would also be considered role expansion. In Short, any Sharing of information or any type
of communication not demanded by the formal job description was considered role
expansion in this study.

The dimensionality of role expansion was tested by conducting an exploratory
factor analysis of communication messages. No a priori factors were predicted because
there is currently no literature to support such predictions. The ﬁrst step was to code all
text messages using the coding scheme shown in Figure 4.

This coding scheme resulted in eight categories of communication. The text
message categories are: 1) social communication, 2) requests for information about
someone else's job, such as the range and boundaries of the information a team member
can measure, 3) requests for information that could be transmitted by an alternative
method, 4) requests for information that could not be transmitted by an alternative
method, 5) the sharing with others information about one's job, such as the range and

boundaries of information that can be measured, 6) the sharing with others information

65

that could be transmitted by an alternative method, 7) the sharing with others information
that could not be transmitted by an alternative method, and 8) communication used in an
attempt to develop a strategy for use during the simulation.

Three people were used to code these messages. Two raters coded the entire set
of communications while the third rater coded one ﬁfth of the communications. The inter-
rater reliability for the coding of the text messages into categories was .94 (see Table 1).

The results from the factor analysis revealed a two factor solution that included
ﬁve items in the ﬁrst factor and four in the second. Social communications did not
adequately load on either factor. The coefﬁcient alphas were .57 for the ﬁrst factors and
.37 for the second. Because of the low reliability on these two factors an additional factor
analysis was conducted which forced a single factor solution. The reliability of this factor
was an alpha of .53. Given the low reliability of all factor solutions, i made a decision to
choose the single factor solution for two reasons. First, all of the hypotheses are based on
a uni-dimensional deﬁnition of the role expansion construct. Without compelling
statistical evidence of a more complex multi—dimensional structure, the more parsimonious
single factor of role expansion dimension was used to test the hypotheses. Second, the
single factor solution had only a Slightly lower reliability than did the ﬁrst factor in the two
factor solution. Although neither reliability is particularly good, the lack of substantial
difference makes the Single factor a viable option.

The TIDE2 computer simulation task consisted of 36 trials. Role expansion was
measured by averaging the number of emergent task elements for each team member

during trials 3-20.

66

_S_tg§§. Stress was manipulated in the last 16 trials of the experiment. The stress
manipulation was accomplished by varying the amount of time available to the subjects
for making a decision. The high stress condition included trials with only 90 seconds per
target. The low stress condition included trials with 160 seconds per target. High and
low stress trials were randomly distributed.

Role incumbent knowledga. The subjects' knowledge of the information cues
and their ability to interpret the raw information that made up their jobs was assessed with
a test. The test was given at the beginning of the simulation to measure each subject's task
knowledge. The internal consistency for the job knowledge test was determined by
calculating the coefficient alpha. Table 2 Shows that the coefficient alpha for the job
knowledge test was an acceptable .75 (Nunnally, 1978).

:1‘_e_am decision making accuracy. The second dependent
variable was decision making accuracy. Decision making accuracy was the absolute
difference between the judgment of each subject and the correct decision for each target.
The direction of the error was not critical to this study. This measure was automatically
recorded at the end of each trial as part of the data that the TIDE2 simulation collected.
Procedures. An overview of the timing of the procedures is shown in Table 3. The
subjects began the experiment by reading a handbook that described the simulation, the
nine target attributes, and their Speciﬁc roles in their teams (see Appendix B). This was
followed by the ﬁrst administration of the task knowledge test (see Appendix C). At this
point the subjects began the computer simulation with a short training session that
introduced the subjects to the computer simulation environment (see Appendix D). In the

high experience condition subjects rotated every 6 trials. The next 16 trials were used to

67

Table 2: Internal consistency for the task knowledge test across all four conditions.

 

Number of
Station Alpha Items
Q
Total .75 15
CAD .79 1 5
AWACS .69 1 5

Cruiser .77 1 5

 

68

Table 3: The timing and order of procedures used in the experiment.

 

1. Training session
a. Review task description document
b. View video (5 minute version)
c. Task knowledge test
d Interactive training

2. First 20 trials

a. Teams in the high experience condition changed positions every 6 trials.
They were given 8 minutes to study their new roles.
b. All teams were allowed to communicate among team members during the 8

minutes. The leaders were told that they could use this time any way they
wanted to including discussing Strategy. Outer station members were not
instructed as to what to communicate only that the time was available.
3. Next 16 trials
a. Each team member remained at the same station they were at when
ﬁnishing the previous section. During this portion of the experiment Stress
was manipulated by varying the time for each trial from 90 to 160 seconds.
4. Post session
a. Team members ﬁlled out a second task knowledge test.
b. Subjects were debriefed as to the purpose of the experiment and dismissed.

 

69

ﬁlled out a post-experiment questionnaire that included the second task knowledge test
and role expansion measure. Before the subjects left, they were given a debrieﬁng sheet
(see Appendix E). Any questions the subjects had were answered at that time.

Data Analysis

Prima_ry data analysis. This study is divided into two parts. The purpose of the
ﬁrst part was to determine the antecedents to role expansion. Hierarchical regression
analysis was used to test Hypotheses 1 through 8 as shown in Table 4. The ﬁrst step
included characteristics of the leader, testing the inﬂuence of the leader's behavior on role
expansion. The second step included characteristics of the Situation, testing the inﬂuence
of the type of feedback and experience on role expansion. The third step included the
interactions between leader behavior and the situational factors, testing the substitute for
leader behavior hypotheses.

Secondary dat_a analysis. The purpose of the second part of the study was to look
at the effects of role expansion, Stress and job knowledge on decision making
performance. Hierarchical regression analysis was used again, this time to test hypotheses
9 and 10 as shown in Table 5. The ﬁrst step included role expansion which was an
independent variable in the second part of the study. The second step included job
knowledge, stress and their interaction. The third step included the interactions of role
expansion with job knowledge and stress testing hypotheses 9 and 10.

Power Analysis. As recommended by Cohen and Cohen (1983), a power analysis
was conducted to determine the sample size required to achieve the appropriate power
level for the statistical analyses used to test the hypotheses presented in the previous

chapter. The analysis requiring the greatest statistical power is that which tests the

70

Table 4: Data analysis plan for the regression equation where role expansion is the
dependent variable.

 

 

 

Step Independent Variable Hypothesis Tested
1 Consideration (C) 1
Initiating Structure (IS)

C x IS -
2 Experience (EXP) 3
Type of Feedback (FDBK) 4
EXP x FDBK -
3 C x EXP 5
C x FDBK 6
IS x EXP 7
IS x F DBK 8

 

71

Table 5: Data analysis plan for the regression equation where team decision making
performance is the dependent variable.

 

 

 

Step Independent ﬁriable Hypothesis Tested
1 Role Expansion (RE) -
2 Stress -
Job Knowledge (JK) -
STR_JK -
3 RE x JK 9

RE x Stress 10

 

72

hypotheses suggesting moderated effects. Ifthe entire regression equation explains at
least 16% of the variance (R2 = .16), 84 teams would provide a power of .80 at the .05
one tail test level for a regression equation with seven variables.

Although this study focused on individuals within a team, the experimental design
will allow for ﬁiture analysis at the team level. As a result, the power analysis was
calculated to determine the number of teams that would be required. This results in a
sample size that should be more than adequate for the analyses at the individual level.

The ﬁnal number of teams that participated in the study was 80. During the data
collection period there were 96 teams scheduled. One team's data was lost when the
simulation experienced technical difficulties and the simulation was improperly restarted.
15 other teams were not run due to the lack of subjects. Six subjects were recruited for
each team. However, for any number of reasons, including sickness, mid-term
examinations, or social commitments, three or fewer subjects often showed up. Because
a minimum of four subjects are required for the simulation, those who showed up were

either rescheduled for another time or asked to participate in another study.

Chapter Four: Results
The means and standard deviations for all variables are presented in Table 6 and
the intercorrelations among the variables are presented in Table 7. Moderator regression
analysis was used to test hypotheses in this study. The ﬁrst regression equation tested all
hypotheses relating to role expansion. The second regression equation tested all
hypotheses relating to decision making accuracy.
Analysis of the Role Expansion Hypotheses

Hypothesis 1. Hypothesis I predicted a main effect for the leader's level of
consideration on role expansion. The prediction was that the higher leaders were rated on
consideration the greater the amount of role expansion. Table 8 Shows the results of the
moderator regression analyses used to test this hypothesis. Consideration accounted for
6.1% of the variance role expansion (R2 = .061, p < .01). Figure 19 Shows that the effect
was in the predicted direction, the more consideration by the leader resulted in greater
amounts of role expansion by the subordinate.

Hypothesis 2. Hypothesis 2 predicted a main effect for the leader's level of
initiating structure on role expansion. The speciﬁc prediction made by Hypothesis 2 was
that when leaders were rated low on initiating structure there would be greater role
expansion. Initiating structure accounted for only .4% of the variance (R2 = .004, not
signiﬁcant (ns) at the p < .5). The interaction between the leader's level of consideration
and initiating structure was also tested and was not signiﬁcant.

Hypothesis 3. Hypothesis 3 predicted a signiﬁcant main effect for experience in
predicting role expansion. It was predicted that when subjects experienced several jobs
within the team, this would lead to greater role expansion. Although experience had a

73

74

Table 6: Means and standard deviations of variables in the study.

 

 

Standard

Variable Mean Deviation E
Consideration 2.49 1 . 10 240
Initiating Structure" 2.0] 1.00 240
Pre Task Knowledge Test 9.80 3.28 240
Role Expansion 22.92 15.00 240
Decision Making

Performance 1 .36 .54 240

 

* Square root transformation was used to normalized the skewed distribution and the
transformed values were used in all analyses involving initiating structure.

75

Table 7: Intercorrelations among the variables in the study.

 

1 2 3 4 5 6
l-C
2-IS .16**
3-PTKT -.01 -.01
4-RE .25“ -.02 .00
5-DMP .02 .09* .02 -.O8
6-EXP -.l7** .02 .33“ -.31** .20**
7-FDBK -.04 .14” -.06 -.08 .01 .00
8-STR .00 .00 .00 .00 .17** .00 .00

 

C = Consideration

IS = Initiating Structure

PTKT = Pre-Task Knowledge Test

RE = Role Expansion

DMP = Decision Making Performance

EXP = Experience (1=high, 0=low)

FDBK = Feedback (1=process, O=outcome)
STR = Stress (1=high, 0=low)

* < .05
** < .01

76

Table 8: Regression results with role expansion as the dependant variable.

 

 

ﬂap Variable B R R} R2 Change
1. Consideration 3.38 .247 .061 .061“
2. Initiating

Structure -2.67 .255 .065 .004
3. ConsidXIS 1.32 .257 .066 .001
4. Experience -9.18 .395 .156 .090”
5. Feedback -1.63 .398 .159 .003
6. EprFeedback 4.04 .404 .163 .004
7. EprConsid .02 .404 .163 .000
8. FeedbackXConsid -.72 .404 .163 .000
9. EprIS -.40 .404 .163 .000
10. FeedbackXIS -16.24 .441 .195 .032"
*p<.05

**p<.01

77

Figure 19: Main effect of the leader's behavior of consideration on role expansion.

 

29: """"" q “““““““““““““““““““““

27

25

Role Expansion
to
0)

21

17

 

 

 

 

LOW Leader Consideration High

78

main effect on the amount of role expansion accounting for 9% of the variance (R2 = .09,
p < .01), it was not in the predicted direction (see Figure 20). Those subjects who
experienced several jobs expanded their roles signiﬁcantly less than did subjects who
experienced only one job throughout the experiment.

Hypothesis 4. Hypothesis 4 predicted a signiﬁcant main effect for the type of
feedback that role incumbents received on role expansion. Speciﬁcally, it was predicted
that role incumbents who received process feedback would expand their roles more than
subjects who received only outcome feedback. The type of feedback was not signiﬁcant,
accounting for only .3% of the variance in role expansion (R2 = .003, ns p < .05). The
interaction between experience and the type of feedback was tested and was not
signiﬁcant.

Hypothesis 5. Hypothesis 5 predicted a signiﬁcant interaction between the
amount of experience a role incumbent had and the leader's level of consideration in the
prediction of role expansion. This interaction was not signiﬁcant.

Hypothesis 6. Hypothesis 6 predicted a signiﬁcant interaction between the type
of feedback a role incumbent received and the leader's level on consideration in the
prediction of role expansion. This interaction was not signiﬁcant.

Hypothesis 7. Hypothesis 7 predicted a signiﬁcant interaction between the
amount of experience role incumbents received and the leader's level of initiating structure
in the prediction of role expansion. This interaction was not signiﬁcant.

Hypothesis 8. Hypothesis 8 predicted a Signiﬁcant interaction between the type
of feedback role incumbents received and the leader's level of initiating structure in the

prediction of role expansion. This interaction was signiﬁcant accounting for 2.6% of the

79

Figure 20: Main effect of job experience on role expansion.

 

29

 

25: .f

    
 

Role Expansion
h)
0

lo
.a

 

 

 

 

LOW Job Experience High

80

variance (R2 = .026, p < .01). However, Figure 21 shows that feedback did not make the
leader's level of initiating structure irrelevant. Process feedback resulted in less role
expansion when the leader's level of initiating structure was high. It was predicted that
high initiating structure would not have an effect in the process feedback condition (see
Figure 8).

Analysis of the Decision Making Accuracwypotheses

The next two hypotheses refer to the second part of the study which deals with
decision making accuracy. In the ﬁrst part the focus was on role expansion; in the second
part the focus is on how role expansion effects decision making accuracy. The moderator
regression analysis used to test the last two hypotheses included trials in the last half of
the experiment (trials 21 - 36).

The ﬁrst four variables entered into the regression equation were role expansion,
job knowledge, stress and the interaction between job knowledge and stress. There were
no predictions made about the effect of these variables on decision making performance.
Table 9 shows that only stress had a main effect on decision making performance. Stress
accounted for 2.9% of the variance (R2 = .029, p < .01). Figure 22 Shows that when
stress was high decision making performance was signiﬁcantly worse than when Stress was
low. Role expansion, feedback and the interaction between incumbent job knowledge and
stress were not signiﬁcant.

Hypothesis 9. Hypothesis 9 predicted a signiﬁcant interaction between role
expansion and job knowledge in predicting decision making accuracy. The nature of the
interaction is such that when role expansion is high and task knowledge is low, decision

making accuracy will be especially low. Also, when role expansion is high and task

81

Table 9: Regression results with decision making accuracy as the dependant variable.

 

 

Step Variable R R R} R2 Change
1. Role Expansion (RE) -.003 .076 .006 .006+

2. Job Knowledge (JK) .004 .079 .006 .001

3. Stress (Str) .186 .190 .036 .030"

4. JK_Str .013 .194 .038 .002

5. RE_JK -.005 .196 .038 .001

6. RE_STR -.027 .197 .039 .000
+p<JO

* p < .05

**p<.01

82

Figure 21: Interaction between the leader's level of initiating structure and the type of
feedback received on role expansion.

 

29 ....... ....................... r.._..........,; ..... . .....

25';

Role Expansion
re
a

N
.e

 

 

 

Low High

 

 

83

Figure 22: Main effects for stress on decision making inaccuracy performance.

1.8

 

 

 

  

—L
.h

—L
[0

Decision Making Performance

High 1

 

 

0.8 Lev: ~ ' ~ - » * '”"H:;h

Stress

 

 

 

84

knowledge is high, decision making accuracy will be especially high. This interaction
was not signiﬁcant.

Hypothesis 10. Hypothesis 10 predicted a signiﬁcant interaction between the
amount of role expansion and the level of stress on decision making accuracy. The nature
of the interaction is such that decision making accuracy will be especially low when role
expansion of any type is high and stress is high. Decision making accuracy will be
especially high when role expansion of any type is high and stress is low. This interaction
was not signiﬁcant.

Post Hoc Analysis of Decision Makimﬂccuracy Dat_a

Because the amount of variance in decision making accuracy was quite low (4%
overall), a second regression equation was run that included experience and feedback.
This was done to explore the possibility of direct effects for these variables on decision
making accuracy. The order variables were entered into the equation and the results are
reported in Table 10. There was a signiﬁcant main effect for experience (3.8% of the
variance in decision making accuracy, R2 = .03 8, p < .01). Figure 23 shows that decision
making performance was lower in the high experience condition than it was in the low
experience condition. It Should be noted that the addition of feedback and experience did
not change the relationship between stress and decision making accuracy, however, the

role expansion by job knowledge interaction was no longer marginally signiﬁcant.

85

Table 10: Regression results with decision making accuracy as the dependant variable.

 

 

_Stgp Variable 1_3_ R R? R2 Change
1. Role Expansion (RE) -.003 .076 .006 .006+
2. Job Knowledge (JK) .004 .079 .006 .001

3. Stress (Str) .186 .190 .036 .030"
4. Experience .234 .271 .074 .038"
5. Feedback .008 .271 .074 .000
4. JK_Str .013 .274 .075 .001
5. RE_JK -.002 .275 .075 .000
6. RE_STR -.027 .276 .076 .001
+p<l0

"‘ p < .05

** p< .01

86

Figure 23: Main effect for experience on decision making inaccuracy performance.

1.8 ..

Low

Decision Making Performance

High 1

1.6 f; :..,.I?.;' _, _:;_,,;i15_

_a
.h.

.4.
N

0.8

 

 

 

 

 

Experience

Chapter Five: Discussion

Because different sets of hypotheses were tested by each regression equation, the
discussion that follows will be separated into two main sections, role expansion and
decision making accuracy. The primary purpose of this study was to test for antecedents
and consequences of role expansion. The antecedents included characteristics of the
leader and the situation. The major consequence of role expansion studied here was
decision making accuracy. Two regression equations were run to analyze these data. The
dependent variable for the ﬁrst regression equation was role expansion, while the
dependent variable for the second regression equation was decision making accuracy.
The two main sections will be further divided into sub sections. This is because the
results of this study were mixed. Although some ﬁndings were signiﬁcant, they were not
always in the predicted direction. Other times, predicted effects were not found.
Consequently, each main section of the discussion will be separated into categories that
include unsupported hypotheses, supported hypotheses, signiﬁcant ﬁndings in the wrong
direction, and signiﬁcant ﬁnding that were not predicted. These sub-sections will be used
where appropriate.
Role Expansion

There are many models of roles. Most of these models were developed in an
effort to explain the process by which role elements were added to a job. They focus on
the leader/subordinate dyad and the negotiation process. In these models role expansion
was assumed, and the focus was on the process by which role elements are added to roles.
I have proposed that knowing the antecedents of the role expansion process and the
outcomes of that process are equally important. A greater understanding of how leader

87

88

and situational characteristics effect role expansion might make it possible to create
Situations that promote and facilitate role expansion when appropriate or inhibit role
expansion when it is not appropriate. The ﬁrst part of this study attempted to predict role
expansion based on characteristics of the leader and the Situation.

Unsupported Hypotheses. Initiating structure was predicted to have a main effect
on role expansion. I predicted that leaders with high ratings on initiating structure would
be more controlling of the team members' actions and that this would lead to less role
expansion than when leaders had low ratings on initiating structure. No support was
found for this main effect.

There are several possible explanations for this and other negative ﬁndings in this
study. The ﬁrst possible explanation is to consider how initiating structure was measured.
The leader behavior description questionnaire (LBDQ) was not used in this study because
most of the items were not appropriate. The nature of the computer simulation and the
time constraint of the experiment did not create a Situation that resulted in
leader/subordinate interactions for which the LBDQ items are designed.

The distribution of ratings for initiating structure was skewed to the low end of the
scale. One likely factor was the time constraints in this experiment. The time constraints
limited the total amount of communication between the leader and subordinates. In
addition, leaders were only able to send one message to one other team member at a time.
These two situations may have combined to limit the number of messages being sent that
were related to initiating structure. The computer simulation was also likely a limiting
factor. Leaders may not have felt the need to control the efforts of team members for two

reasons. First, the job descriptions clearly deﬁned each member's job and responsibilities.

89

Second, the teams were composed of strangers with the leader having no more experience
performing the simulation than anyone else. Had the leaders had prior experience, they
may have been more proactive.

The second possible explanation is that time was a constraint in a way that limited
the interaction between leaders and other team members. In this experiment the leaders
were constantly engaged in the simulation with very little time between trials. Leaders
had little time to reﬂect on what they were doing or how to do it differently. To
overcome this limitation, two seven minute breaks were introduced after the eighth and
fourteenth trials. These breaks were added as an attempt to create time for reﬂection and
strategy development. However, leaders were still learning the task and getting used to
working with the other team members. It is possible that they did not feel conﬁdent
enough about their own ability in the simulation to initiate structure for the other team
members. In the ﬁrture, the use of leaders who already have experience on the Simulation
rrright increase their conﬁdence and their willingness to display consideration and initiating
structure behaviors.

The lack of time seems to have been a serious shortcoming. Had there been a
second session several days later, the leaders would have had time to reﬂect on the task
and how they were approaching it. In the second session, leaders with a propensity
toward initiating structure would have been more likely to engage in those behaviors.

The predictions that experience and feedback would act as substitutes for
leadership were based on the belief that incumbents largely decide how they are going to
expand their jobs. This would seem most likely when leaders had high ratings on

consideration. Under the high consideration condition it was predicted that incumbents

CC

ti.

[in
10

int

Ac
lea
Th
ad:
set

Int

and
She

lear

90

would be given the most latitude to expand their roles. It was predicted that in this
condition past experience and the speciﬁcity of the task feedback received would inﬂuence
the subject's job in a way that resulted in greater role expansion.

The feedback manipulation involved varying the amount and speciﬁcity of the
feedback information provided to all team members. The outcome feedback condition
limited feedback to a comparison of the team's decision and each team member's decisions
to the predetermined correct decision. The process feedback condition added cue level
information on how each cue and relevant interactions should have been interpreted. The
main effect for this type of feedback was not signiﬁcant.

One potential explanation for this ﬁnding is found in the work by Kanfer and
Ackerman (1989). In their study Kanfer and Ackerman found that during the declarative
learning phase, goal setting did not have the expected beneﬁcial effect on performance.
They suggested that during the declarative learning phase, goal setting represented an
additional cognitive requirement which led to lower performance than subjects without
set goals. In this study feedback may have had a similar effect. If the subjects were still
in the declarative learning phase, processing feedback information would have represented
a cognitive requirement that the subjects were incapable of handling. Subjects might have
ignored this additional feedback information provided in the process feedback condition
and concentrated on the simpler outcome feedback.

Additional support for the above explanation comes ﬁ'om the work of Krumboltz
and Weisman (1962) and Schulz and Runquist (1960). In these studies the authors
showed that there was little added utility in providing feedback when the subjects were

learning a task. In their studies providing feedback to subjects did not increase

per.

his

\—

I01.

tea

inf.

110

8

int

10‘

CC

91
performance and in one case performance suffered.

Supported Hypothesis. There was one supported hypothesis in this study, that
higher levels of leader consideration was positively correlated with greater amounts of
role expansion by team members. Role expansion was greatest when the leader provided
team members support and encouragement. Team members in this condition Shared more
information and also requested more information.

This ﬁnding could be of interest to managers. In cases where role expansion is
not only desirable but necessary for continued job effectiveness, consideration by the
leader could be a very important factor. For example, jobs that require creativity and
innovation would beneﬁt from a leader who displays a large amount of consideration
toward subordinates. On the other hand, in jobs that are rigidly designed, role expansion
could have a negative effect on performance. In this case a leader who displays high
levels of consideration would need to include some amount of initiating structure to
insure that workers understand the limitations and problems associated with role
expansion. By understanding the effects that leadership style has on subordinate role
expansion, we can not only predict when role expansion is likely to take place but also
understand potential problem sources when role expansion happens and it is not desirable.

All of this must be tempered with the fact that this was the only signiﬁcant result
from the eight hypotheses that were predicted. The ﬁnding will need to be replicated with
another sample before recommendations based on these ﬁndings can be made.

Rigniﬁcant results in the opposite direction. The strongest effect in this study
came ﬁ'om the experience manipulation. It was also the most surprising. It was posited

that subjects who experienced several jobs would be more likely to expand their roles than

\l

92

would the subjects who only experienced a single job. I thought that subjects who
experienced several jobs would want to expand their roles to include their other jobs
since all jobs were relevant to the team decision making task. However, just the opposite
was true, role expansion was greatest for subjects who had the same job throughout the
experiment.

Once again a possible contributing factor was the constraint on the amount of time
that was available in this experiment. Although it is clear that after four or ﬁve trials the
subjects understood how to manipulate the computer during the simulation, learning to
interpret and use the information provided apparently took much longer. In the high
experience condition, subjects were rotated from one job to the next every six trials. Here
again the work by Kanfer and Ackerman (1989) applies. In the high experience condition
subjects were in a perpetual declarative learning phase, limiting their ability to take on
additional cognitive demands that include expanding their roles to incorporate elements
from past jobs. In retrospect there was not enough time for the subjects to learn their new
jobs ﬁrlly and then to incorporate parts of their past experiences.

Decision makmwcuracy

The ﬁrst part of the study looked at role expansion as a consequence of other
factors. In addition, role expansion was studied as an antecedent to decision making
accuracy. In this study I was interested in the effect role expansion had on the decisions
that the subjects were required to make. Speciﬁcally, I predicted that role expansion
would moderate the effects of two other factors on decision making performance. The

other factors were job knowledge and stress.

 

Ul'lSl

wer

W35

ope

eh

m

Iol

Ull

T65

SU

1“

f0

pe

 

93

Unsupported hypotheses. Both hypotheses in this part of the study were
unsupported. The interactions between role expansion and job knowledge and stress
were not signiﬁcant. In addition, there was also no main effect for role expansion which
was entered ﬁrst in the regression equation.

One potential explanation for this may lie in how role expansion was
operationalized. As reported earlier, role expansion was not reliably measured. This lack
of reliability could be masking the relationship.

Another potential reason for not ﬁnding the predicted relationships is because role
expansion was viewed as a uni-dimensional construct. A uni-dimensional deﬁnition of
role expansion was used because multi-dimensional factors had lower reliability than the
uni-dimensional factor. It is possible that reliable multi-dimensional factors would have
resulted in signiﬁcant ﬁndings for both role expansion and the two interactions.

For example, assume that receiving only raw information is counter productive
when faced with a StI’CSSfUl situation. On the other hand, receiving messages that
summarizes the raw information is useful in the same stressﬁrl situation. It is possible for
two subjects receive an equal number of messages, however, the ﬁrst subject receives
messages in the form of raw information and the second subject receives messages in the
form of summarized information. The former situation will result in poor decision
making accuracy, while the latter situation would result in good decision making
performance. However, this effect would be lost when both cases fall under a single

deﬁnition of role expansion.

 

Stu

\V a

for

C0

an

th

ef

94

Rigniﬁcant ﬁnmga. There were two signiﬁcant ﬁndings in the second part of this
study that were not part of the hypotheses that were tested. The ﬁrst signiﬁcant ﬁnding
was that as stress increased, decision making performance decreased. Stress was in the
form of unpredictable changes in the time available to perform the task. This ﬁnding is
consistent with the stress literature. I had predicted that the role expansion would
moderate the negative effects of stress on performance.

The second signiﬁcant result was found in a post hoc analysis where experience
and feedback were added to the regression equation. Because of the lack of ﬁndings in
the original regression equation, I checked to see if experience and feedback had direct
effects in addition to any the indirect effect through role expansion.

Because feedback was not signiﬁcant in the ﬁrst part of the study, I did not expect
a signiﬁcant main effect on decision making performance and none was found. However,
experience was signiﬁcant in the ﬁrst part and I expected it to have a signiﬁcant negative
effect on decision making performance and it did. Subjects in the high experience
condition had lower decision making performance than did subjects in the low experience
condition. This ﬁnding would tend to support the idea that subjects in the high experience
condition were still learning their jobs in the later parts of the simulation. Switching from
one job to another did not allow them to reach a proﬁciency level that would be reﬂected
in decision making performance or that would be conducive to role expansion.
Limitations

There were limitations to this study that may have decreased my ability to ﬁnd
some signiﬁcant results. The ﬁrst deal with the limited number of trials. For logistical

reasons, the study was limited to one three hour session per team. Because of the

com

11188

Ina}

the}

int

the

sub

95

complexity of the task, one hour was dedicated to training and the collection of pre-
measures. In the remaining two hours subjects completed 36 trials, with the ﬁrst two
trials involving interactive training with one of the lab assistants.

Many more trials may be needed for feedback and experience conditions to have
their predicted effects. If it is the case that during the ﬁrst 15 or 20 trials subjects are still
in the declarative knowledge learning phase, then it would have been better to administer
the treatments after that point. The experience condition would also have beneﬁtted from
subjects encountering many more trials at each station. By allowing more time at each
job, the subjects would have had more time to automatize their jobs, thus freeing up
cognitive resources needed to expand their roles. To lengthen the experiment to
accommodate these changes would have required each team to return to the lab two
additional times for a total of seven to nine hours.

A second limitation of this study had to do with the job of the leader. The leader's
job in this computer simulation may not have lent itself naturally to leader/subordinate
interactions. The scoring of the leader's behaviors of consideration and initiating structure
were based on the dyadic interactions between the leaders and their subordinates. In some
cases there were only a few communications on which to base these ratings.

The requirements of the simulation may not have created a strong enough need for
leaders to supervise team members. I attempted to overcome this weakness by giving the
leaders two opportunities to communicate with team members to create strategies or
coordinate actions. These opportunities were in the form of seven nrinute breaks between
trials. The leaders were instructed that they could use this time in any way they wanted,

including communicating with team members and analysis showed that leaders did use this

96

time this way. However, it is possible that stronger manipulation instructions would
have lead to clearer and stronger evidence of consideration and initiating structure by the
leaders.

One additional point of concern has to do with the role expansion construct. The
role expansion construct was operationalized as a uni-dimensional construct because
there was no literature or theory to suggest otherwise. It was thought that the data
analysis would uncover a more complex structure. An exploratory factor analysis was
run and resulted in a two factor solution. However, each factor had only four and ﬁve
items with reliabilities of .57 and .37. Because of the low reliabilities of the factors it was
decided to test the hypotheses using the parsimonious uni-dimensional operationalization
of role expansion.

Future Research

Despite these limitations the ﬁndings suggest several areas where ﬁrture research
is warranted. One research direction would speciﬁcally look at the effect of diﬁ‘erent
types and amounts of feedback on performance during the different stages of learning.
Too much feedback too early may slow down the learning process and negatively effect
performance, much like what Kanfer and Ackerman (1989) found with goal setting. Role
expansion could also be studied to see if it takes place only after the declarative
knowledge phase of learning.

Experience is another potentially fruitful direction for ﬁiture research. I still
believe that more variety in experiences will lead to greater role expansion. Incumbents
are likely to incorporate relevant elements of past experiences into their current jobs. If

certain job elements are more likely than others to be included in new jobs, this could have

I
l

imp
diff:

take

leve

role

moc

 

 

 

97

implications on how mentors structure and sequence a worker's experiences. However, a
different research design will likely be needed -- one that is longitudinal in design and that
takes into account the stages of learning.

Future research Should look at the impact of individual role expansion and team
level outcomes. This study looked at individuals within teams with the focus on individual
role expansion and individual decision making performance. What is the impact of
individuals expanding their roles on team decision making accuracy? Can the concept of
role expansion by an individual be aggregated to the team level to consider team role
expansion? This current study has the data that will be required to answer these
questions, however before I can begin with any analyses, I will need to create theoretical
models and hypotheses for team level constructs of role expansion.
2mm

In conclusion, this study demonstrated that role expansion takes place and that
role expansion differs across individuals. In addition, there are factors external to the
traditional supervisor/subordinate dyad that inﬂuence the role expansion process. Finally,
this study attempted to show that there are consequences of role expansion beyond the
amount of role conﬂict and role ambiguity. Research in the future should look at the role
expansion process over a greater length of time and within organizational settings where

possible.

APPENDICES

 

OilSl
I

met"

Ibor

{ch
hou
iOIlS'.

lniti.
role
to p‘
P90}

he
she
18

If I'.
. I
1m;

W0
fat:

 

APPENDIX A

Coding instructions for leader/incumbent communications
CONSIDERATION

Consideration is the degree to which a leader acts in a cooperative manner and shows
concern for subordinates. Examples of consideration are asking for inputs from
subordinates and having a willingness to except their suggestions.

1 2 3 4 5
Never Seldom Occasionally Often Always
shows shows shows Show shows
consid. consid. consid. consid. consid.

IN ITIATING STRUCTURE (IS)

Initiating structure is the degree to which a leader deﬁnes and structures his or her own
role and the role of others. Examples on initiating structure are offering new approaches
to problems and coordinating the activities of subordinates. Also communication telling
people what to do.

1 2 3 4 5
Never Seldom Occasionally Often Always
Shows shows shows shows shows
IS IS IS IS IS

If there is not a lot of communication then the following may be true. First, it will be
impossible to determine negotiating latitude so the rating would be a 1. Second, it will be
difficult to determine consideration and initiating structure. In these cases the ratings
would also be a 1. If there is only a little communication but you are able to make a
rating on one or two of the scales then make the appropriate rating.

98

APPENDIX B

Handbook for command and control simulation
NAVAL COMMAND AND CONTROL TEAM SIMULATION

INTRODUCTION

The year is 1996 and you are a part of a US. naval carrier task force command and
control team stationed in the Middle East. A regional conﬂict between two nations in this
area has recently broken out, and your mission is to protect seagoing commercial trafﬁc in
the area from accidental or intentional attacks.

THE TASK FORCE

A naval carrier task force is an awesome array of Ships and support units. As shown on
the cover of this handbook, it has a concentric ring of missile ﬁring warships which
protect the aircraft carrier at its center. The aircraft carrier, in return provides an overall
umbrella of air protection for the entire task force. The carrier's 9O planes can unleash air
strikes against targets at land, sea and even under water. A carrier task force is also
supported by AWACS reconnaissance planes and a land based Coastal Air Defense
(CAD) unit. Although the carrier itself is equipped with some air patrol capacities, the
cruiser, AWACS and CAD units provide the bulk of the air traﬂic patrol.

TEAM MISSION

The team of which you are a part, will role play the Commanding Ofﬁcers of various
units in the carrier task force. Your mission is to monitor the air space surrounding the
carrier task force, making sure that neutral ships are not attacked. In performing this role,
you must make certain that you do not allow loss of life resulting from accidental or
intentional attacks on ships in the task force. At the same time, it is also of paramount
importance that you do not inadvertently Shoot down ﬁiendly military aircraft or any
civilian aircraft.

OVERVIEW OF ROLES

There are four roles in this Simulation, one for each member of a four person team. The
leader is the Commanding Ofﬁcer (C0) of the Aircraft Carrier. The other team members
include the Co of an AWACS air reconnaissance plane; the C0 of an Aegis Cruiser, and
the C0 of a Coastal Air Defense Unit. The team's task is to decide what response the
carrier task force should make toward incoming air targets. Teams base their decisions on
data they collect by measuring characteristics of aircraft that enter the task force's
airspace.

99

100

POSSIBLE DECISIONS:

(1) IGNORE: This means that the carrier task force should devote no further attention to
the target and instead focus on other possible targets in the area.

(2) REVIEW: This means to leave this target momentarily, so that the team can monitor
other targets, but to return to this target after a short period of time to update its status.

(3) MONITOR: This means that the carrier task force should continuously track the
target on radar.

(4) WARN: This means that the carrier task force sends a message to the target
identifying the task force and alerting the target to steer clear.

(5) READY: This means to steer the ship into a defensive posture and to set defensive
weapons on automatic. A hip in a readied position is rarely vulnerable to attack.

(6) LOCK-ON: This synchronizes the ship's radar and attack weapons so that the
weapons ﬁx themselves on the target.

(7) DEFEND: This is "weapons away" and means to attack the target with Tomahawk
cruise air-to—air missiles.

CHARACTERISTICS OF TARGETS

The incoming air targets can be measured on nine attributes. These are listed below along
with the ranges of possible values on the attributes:

Speed how fast it is ﬂying in miles per hour (mph)

AME how high above sea level it is ﬂying in feet (ft)

_Siie How long it is from nose to tail in meters (mtr)

A_ngl_e the extent it is ascending/descending in degrees

ﬂ radio signal indicating civilian/military status in Mega-Hertz (MHZ)

Direction extent it is ﬂying straight in or passing wide in degrees (dgs)

Corridor distance from center of commercial airline Status corridor in miles (mi)
Radar type of radar ranging from weather to weapons Type arranged in classes

Range how far away it is from the carrier in miles (mi)

GU 1L

Y RT.

TTIA._Sirr\eI.tIIIil.

h
1171
A
if
R

101

Your Speciﬁc Role

Role of Lgader (Cagner). The carrier is the leader who makes the team's ﬁnal decision.
The carrier can only measure and interpret three things: 1) Direction, 2) Range, and 3)
IFF. The range of values and degree of threat associated with each are shown below.

Degree of Threat
Non- Somewhat Very
Threatening Threatening Threatening
Altitude 35,000 - 23,000 - 13,000 -
27,000ft 17,000ft 5,000ft
IFF .2-.6mhz .9- 1.1mhz 1.4- 18th
Radar Class 1 & 2 Class 5 Class 8 & 9

The leader is unique in knowing what the other team members are experts in, however.
The areas of other's expertise are listed below:

Attributes CAD AWAC S Cruiser
Speed X

Altitude X

Size X X
Angle X

IFF

Direction X

Corridor status X
Radar type X

Range X

DETERMINING THE LEVEL OF THREAT

In general, the degree to which an incoming target is threatening depends on its standing
on these nine attributes. There are ﬁve rules for determining the danger associated with
any target and you are tasked with remembering one. The other team members are
responsible for the other four rules. Your rule is: "all else equal, in terms of IFF, high
frequencies are more threatening than low frequencies".

The most threatening target is one where the target is very threatening on all ﬁve rules.
The least threatening target is one where the target is non-threatening on all ﬁve rules.

Six

in
cr.

ll

102

Your Speciﬁc Role

Role of CAD. The CAD collects and shares information with other team members. The
CAD also sends a judgment, based on his or her interpretation of the information, to the
carrier. The CAD can only measure and interpret three things: 1) Speed, 2) Direction and
3) Size. The range of values and degree of threat associated with each are shown below.

Degree of Threat
Non- Somewhat Very
Threatening Threatening Threatening
Speed 100 - 275mph 325 - SOOmph 600 - 800mph
Direction 30 - 22 dgs 18 - 12 dgs 8 - 0 dgs
Size 65 - 43 mtr 37 - 23 mtr 17 - 10 mtr

In addition, your role requires you to get RADAR information from the Cruiser. The
cruiser will have the knowledge for interpreting RADAR information.

DETERIVIINING THE LEVEL OF THREAT

In general, the degree to which an incoming target is threatening depends on its standing
on these nine attributes. There are ﬁve rules for determining the danger associated with
any target and you are tasked with remembering two. The other team members are
responsible for the other three rules. Your rule is:

SPEED and DIRECTION go together, so that fast targets coming straight in are most
threatening. Speed alone and direction alone mean nothing. There is nothing to fear if
fast targets are not headed toward the task force or from objects headed directly for the
task force that are moving slowly.

SIZE and RADAR go together, so that small objects with weapons radar are especially
threatening. Size and radar alone mean nothing. There is nothing to fear from small
targets with weather radar or from large targets with weapons radar.

The most threatening target is one where the target is very threatening on all ﬁve rules.
The least threatening target is one where the target is non-threatening on all ﬁve rules.

103

Your Speciﬁc Role

Role of AWACS. The AWACS collects and shares information with other team members.
The AWACS also sends a judgment, based on his or her interpretation of the information,
to the carrier. The AWACS can only measure and interpret three things: 1) Angle, 2)
Range, and 3) Radar. The range of values and degree of threat associated with each are
shown below.

Degree of Threat
Non- Somewhat Very
Threatening Threatening Threatening
Angle +15 - +8 dgs +3 - -3 dgs -8 - -15 dgs
Range 200- 110 mi 90-60 mi 40-1 mi
Radar Class 1 & 2 Class 5 Class 8 & 9

In addition, your role requires you to send RADAR information from the CAD. The CAD
will not have the knowledge for interpreting RADAR information.

DETERMINING THE LEVEL OF THREAT

In general, the degree to which an incoming target is threatening depends on its standing
on these nine attributes. There are ﬁve rules for determining the danger associated with
any target and you are tasked with remembering two. The other team members are
responsible for the other three rules. Your rule is:

ANGLE and RANGE so together, so that descending targets that are
close are especially threatening. Angle and range alone mean nothing.
Descending targets that are far away or close targets that are on the way up
are not threatening.

The most threatening target is one where the target is very threatening on all ﬁve rules.
The least threatening target is one where the target is non-threatening on all ﬁve rules.

104

Your Speciﬁc Role

Role of Cruiser. The Cruiser collects and shares information with other team members.
The Cruiser also sends a judgment, based on his or her interpretation of the information,
to the carrier. The Cruiser can only measure and interpret three things: 1) Altitude, 2)
Corridor Status, and 3) Size. The range of values and degree of threat associated with
each are shown below.

Degree of Threat

Non- Somewhat Very

Threatening Threatening Threatening
Altitude 35,000 - 23,000 - 13,000 -

27,000 ft 17,000 ft 5,000 ft
Corridor O-8mi 12- 18 mi 22-30mi
Status
Size 65 - 43 mtr 37 - 23 mtr l7 - 10 mtr

DETERMINING THE LEVEL OF THREAT

In general, the degree to which an incoming target is threatening depends on its standing
on these nine attributes. There are ﬁve rules for determining the danger associated with
any target and you are tasked with remembering two. The other team members are
responsible for the other three rules. Your rule is:

ALTITUDE and CORRIDOR STATUS go together, so that low flying
targets that are way outside the corridor are especially threatening.
Altitude alone and corridor status alone mean nothing. There is nothing to
fear from high ﬂying targets well outside the corridor or low ﬂying targets
in the middle of the corridor.

The most threatening target is one where the target is very threatening on all ﬁve rules.
The least threatening target is one where the target is non-threatening on all ﬁve rules.

APPENDIX C

Job Knowledge Test

Carrier

05 degrees of direction represents which of the following?

(DO-00‘8”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that iS out of the possible range.

55 miles of range represents which of the following?

OQOC‘O’

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

2.0 Mhz represents which of the following?

0040C?”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

35 degrees of direction represents which of the following?

(DO—GUN

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

1.5 Mhz represents which of the following?

(DO-00"”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

105

10

11

106

250 miles of range represents which of the following?

OQOC‘W

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

.2 Mhz represents which of the following?

(DO-OCT”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

13 degrees of direction represents which of the following?

(DO-00"”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

150 miles of range represents which of the following?

“Cl-00‘”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

24 degrees of direction represents which of the following?

(DO—00“”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

1.1 Mhz represents which of the following?

(DO-00"”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

12

13

14

15

107

65 miles of range represents which of the following?

(DO-00"”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

16 degrees of direction represents which of the following?

09.0793

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

15 miles of range represents which of the following?

(DO-00‘!”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

.7 Mhz represents which of the following?

(DO-00“”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

108

CAD

600 miles per hour represents which of the following?

(DO-00‘”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

24 degrees of direction represents which of the following?

(DO-00‘”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

50 meters in size represents which of the following?

GQOO‘N

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

35 degrees of direction represents which of the following?

(DO-GUN

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

300 miles per hour represents which of the following?

(DO-00“”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

15 meters in size represents which of the following?

(DO-00"”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

10

11

12

109

330 miles per hour represents which of the following?
A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

(DO-00"”

13 degrees of direction represents which of the following?
A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.

A target that is out of the possible range.

(DO—OCT!”

20 meters in size represents which of the following?
A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

GQOO‘N

250 miles per hour represents which of the following?
A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

(DO-00"”

16 degrees of direction represents which of the following?
A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.

A target that is out of the possible range.

(DO-00"”

5 meters in size represents which of the following?

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

00.069?

13

14

15

110

05 degrees of direction represents which of the following?
A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.

A target that is out of the possible range.

(DO—00"”

850 miles per hour represents which of the following?
A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that iS out of the possible range.

(DO-00"!”

30 meters in size represents which of the following?
A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

(DO-00"”

111

AWACS

-8 degrees of angle represents which of the following?

OQOO‘N

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

55 miles of range represents which of the following?

(DO-00"”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

Class 2 radar type represents which of the following?

(DO—00"”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

+20 degrees of angle represents which of the following?

(DO—OCT”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

Class 7 radar type represents which of the following?

(DO-00"!”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

65 miles of range represents which of the following?

(DO-.073?

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

10

ll

12

112

Class 5 radar type represents which of the following?

(DO—GUN

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

+12 degrees of angle represents which of the following?

(DD—GUN

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

250 miles of range represents which of the following?

(DO-00"”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

-2 degrees of angle represents which of the following?

(DO-00"”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

Class 9 radar type represents which of the following?

(DO-00"!”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

-5 degrees of angle represents which of the following?

(DO-00"!”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

13

14

15

113

150 miles of range represents which of the following?

(DO-OCT”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

Class 0 radar type represents which of the following?

(DO-00"”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

15 miles of range represents which of the following?

0900‘”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

114

Cruiser

31,000 foot altitude represents which of the following?

(DO-00"”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

22 miles out side the corridor represents which of the following?

(DO-06"”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

.2 Mhz represents which of the following?

(DO-00‘8”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

25,000 foot altitude represents which of the following?

(DO-00"”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

5 miles out side the corridor represents which of the following?

(DO-00“”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

2.0 Mhz represents which of the following?

(DO-00"”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

10

11

12

115

21 miles out side the corridor represents which of the following?

(DOOOO‘N

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

1.5 Mhz represents which of the following?

CDQOO‘W

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

11,000 foot altitude represents which of the following?

(DO-00“”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

1.1 Mhz represents which of the following?

(004079?

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

2,500 foot altitude represents which of the following?

(DO-00"”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

16 miles out Side the corridor represents which of the following?

(DD-00"”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

13

14

15

116

22,500 foot altitude represents which of the following?

(DO-.069)

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

.7 Mhz represents which of the following?

(DD-00"”

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

32 miles out side the corridor represents which of the following?

(DQOOO‘W

A non-threatening target.

A somewhat threatening target.

A very threatening target.

A target on the border between 2 threat levels.
A target that is out of the possible range.

APPENDIX D

Interactive Training Script

[To the experimenter: Everything that is in brackets [ ] is for you and does not need to be
read aloud to the subjects. Everything that starts with a number i_s_ to be read to the
subjects. Please try and stay as close to the script as possible to insure that all teams get
the same training. If subjects have a question go ahead and answer it. If it pertains to
something that is coming up in the training tell them that and if there is still ambiguity have
them ask the question again]

[The interactive training involves the ﬁrst two games in each session. The ﬁrst game is
seven minutes long and this is enough time to accomplish all of the objectives. First ask
the group not to do anything until you tell them to. This will make sure that you have
time to cover everything. You and not the subjects should call up the ﬁrst target by
pressing the return key on the carrier computer]

1 I am going to walk you through this ﬁrst target to get you used to working the
computers. We are going to do all of the things you will need to do to perform
you job. Please do not get ahead of me and do something before I tell you to even
if you think you know what is conring up next. There are a couple of places where
if you get ahead of me we will have to back up and that will take time and we want
to get through the complete training on this target. I will now call up the targets
and make sure that everyone is logged on.

2 The ﬁrst thing we are going to do is measure two attributes. To do this ﬁrst press
the ALT key to highlight the menu bar. Remember, you may need to hit the ALT
a few times _to highlight the menu bar. Once the menu bar is highlighted press
return or the down arrow to open this box. Notice that only the attributes that you
can measure are listed in this box. Using the arrow key highlight the ﬁrst one on
the list and then hit the return key.

3 At this point a box will open up in the lower left hand corner that contains the
information for that attribute. This box will stay open for 5 seconds or until you
press return.

4 Go ahead and open the menu box again and measure the second attribute and then
stop. [If someone needs help tell them again how to use the ALT key to highlight
the menu bar and then open up the box]

5 Next I want you to hit the F2 key. This will open up the summary box, in the

lower left hand comer of the screen, and this box lists everything you currently
know about the target. This box will stay open for 10 seconds or until you press

117

10

ll

12

118

the return key.

Now I want you to open up the Query box. Again you have to use the ALT key
to highlight the menu bar and using the arrow key move to the right until Query is
highlight and then use the arrow key or the return key to open up the box. Once
the box is open STOP. First notice that all nine attributes are listed. Using the
arrow key highlight one that you can not measure and STOP. Do not hit the
return key. Now I want the Carrier to Query the CAD; the CAD to Query the
Cruiser; the Cruiser to Query the AWACS; and the AWACS to Query the Carrier.
[Repeat who Queries whom]. Now hit the return key. A box will open up in the
lower right hand corner of the screen. Using the arrow key select the correct
station and when that is done hit return. And hit return again.

Everyone should have a QRY in the lower left hand corner of their screen. [This
may take a couple of seconds to Show up on some screens. If someone made a
mistake go to a station with QRY on the screen and send a query to the station
that does not have one] This means that someone has sent you a Query.

To receive the Query open up the Query box. Notice that there is a Q which
means you have a query and who it is from. Using the arrow key highlight this and
hit the return key. A box will open up in the lower left hand corner and tells you
who is asking for what information. This box will automatically close in 5

seconds.

Now we are going to send someone information. First open up the Transmit box

' and STOP. Notice that the only things listed in this box are attributes that you

have already measured. You can not send information that you have not already
measured. Using the arrow key highlight the ﬁrst one and stop. [It is not
important that they are not sending the information that was requested. This is just
a demonstration] Now we are going to send this information to the same places.
The Carrier is going to transmit to the CAD; the CAD to the Cruiser; the Cruiser
to the AWACS; and the AWACS to the Carrier. [Repeat only if necessary]

Notice that there is a TRN acronym in the lower left hand corner of the screen. To
view what was transmitted open up the receive box. This time you see the letter

T. Select this the same way you did the Query. A box in the lower left corner
opens up and tells you what was sent and by whom.

Now hit F2 again. Remember this is your summary box and tells you what you
know about the target. Notice this time that this box includes not only what you
measured but what was sent to you by transmission.

We are going to open the transmit box again but this time highlight the text option.
Select the same destination you have been using. [Repeat if necessary] A box in
will open up in the middle of the screen. Write a quick message and then hit

13

14

15

119

return. This is how you communicate during the experiment. There is no talking
in this room. If you need to tell someone something use this option.

Everyone should have a MSG acronym in the lower left hand corner of the screen.
Open the receive box and look at this message. When the message box opens you
will have the option to log the message or not to log it. Ifyou log the message it
will be available to you during the entire experiment. Please not that whether you
log the message or not the computer will save all messages and these will be view
by the experimenters.

At this point we have seen how to measure attributes. We know two ways to ask
for information, using the Query or Transmit a Message. We also know two ways
to send information, Transmit data or Transmit a Message. [If there is a lot of
time let the subject continue by measuring additional attributes and send each other
information. At this point tell them the following] This is for the outer stations,
CAD, Cruiser and AWACS, the computer will start to beep when there is 30
seconds left in this trial. At this point you need to make your decision. You need
to send your judgment to the carrier with enough time to receive and analysis your
judgments and then make the team decision. If you what to long then the Carrier
will have to make a decision without your judgment. So it is very important that
when you hear the beeps you need to start to make you judgment. [with 45
seconds left have all stations send in a DEFEND judgment. Now go to the Carrier
and tell him or her that when they are waiting for the others' judgments they must
not have a box open or the menu highlighted. If either is the case then the
judgments will not get in. Also, tell him or her that with about 10 seconds left they
will have to send in the team decision even if there is no judgment ﬁ'om one or
more of the stations. At some point remind everyone:] Now when the Carrier
makes a decision this will end the trial for everyone. It is very important that the
outer stations; CAD, AWACS, Cruiser; not have a box open or the menu bar
highlighted as time runs out. If you do then you may not get the feedback screen.
As the count down clock nears 5 to 10 seconds make sure no boxes are open or
that the menu is not highlighted. If you need to get to the basic sea screen quickly,
hit the Escape Key in the upper left hand corner of the key board.

This is the feedback screen. AS you can see the correct decision is located in the
top blue band. Below it is the carrier's decision which is the team's decision. In
this case the correct station is DEFEND and the team's decision is DEFEND which
result in a HIT. Along the left hand side list each stations judgment. Along the
right side is your team's history which adds the total number of hits, near misses,
misses, incidents, and disasters. Along the bottom blue line is the total number of
points your team has earned and how many points your team will have if you
continue during as well as you have in the past.

[The following is for the process feedback condition only]

16

17

18

19

120

In the bottom of the screen is a two digit number. This number refers to a page in
the feedback book that is located at each station. Look this number up in the book
now. The feedback sheet Shows you the all of the attribute that you were able to
measure and the value of those attributes. In addition, the sheets tell you subject
value of each attribute. For example SPEED was 700mph and that was
dangerous. Next, there are lines connecting the two attributes that make up the
interaction and again it gives you the subject value of that interaction. [FF is not
part of an interaction and thus has only the ﬁrst subjective value.

The next target will come up automatically. There will be a series of beeps that
will announce the next target. This time I am going to let you do everything on
your own. I will answer any questions you have. I will also repeat important
information. Go ahead and determine how dangerous this target is and what
judgments or decision you are going to make.

[Remind the subjects the importance of knowing how much time is left. Remind
them that there is no talking in the room and that you are listening to them on an
intercom. If they have any trouble during the experiment just call out and an
experimenter will come and help. Remind the outer stations that they need to start
sending their judgments when the computers start to beep. This will give the
carrier enough time to receive their judgments and make the team's decision. After
this trial ends ask if there are any last minute questions. Tell them that on the next
trial they are on their own and again remind them that there is no talking. When
the computers start beeping announcing the next trial tell them:]

The next target is coming up. You are on your own. Good luck.

APPENDIX E

Debrieﬁng Sheet: Experiment 7

The purpose of this study was to examine members roles in decision making teams.
The focus was on how much additional responsibility team members take on. This
additional responsibility is in the form of gathering and using information that was not
required according to the handbook you read at the beginning of the experiment.

An important part of the study was looking at different factors that might lead or
inhibit team members from taking on additional responsibility. These included
characteristics of the carrier's leadership, the type of feedback each team member received
and the amount of experience team members had at the different stations.

In addition, team performance will be evaluated. Of interest here is the affect of
taking on additional responsibility on individual and team performance. Two factors will
be looked at as possibly affecting performance. They the degree to which each team
member knew how to evaluate each cues (job knowledge) and the amount of stress
created by the task (in the form of time constraints).

Finally, team performance will be evaluated to determine the bonus pay off. Each
team is in one of four conditions and will only be evaluated against other teams in the
same condition. The four conditions are: 1) task level feedback and multiple station
experience, 2) task level feedback and experience at only one station, 3) no task level
feedback and multiple station experience, 4) no task level feedback and experience at only
one station.

If you have any questions, call the Team Effectiveness Lab at 336-2224 or call Doug Sego
at 336-3520. Thank you.

121

BIBLIOGRAPHY

BIBLIOGRAPHY

Alvares, K., & Hulin, C. (1973). An experimental evaluation of a temporal decay in the
prediction of performance. Organizational Behavior and Human Performance. 2,
p. 169-185.

Anonymous, (1992). GM'S leaders go on the record. Fortune, pp. 50-60.Balzer, W. K.,

Doherty, M. E., & O'Conner, R. (1989). Effects of cognitive feedback on performance.
Psychology Bulletin, 106. 410-433.

Bergstrom, R. P. (1992). The team's the thing. Production. 104. 46-51.Biddle, B. J.
(1979). Role theory: Expectations. identities, and behaviors. New York: Academic
Press.

Biddle, B. J. (1986). Recent developments in role theory. Annual Review of Sociolagy,
1a, 67-92.

Bons P., Bass, A., & Komorita, S. (1970). Changes in leadership style as a ﬁmction of
military experience and type of command. Persoral Psychology, 2, p. 551-568.

Burroughs, W., Rollins, J ., & Hopkins, J. (1973). The effect of age, departmental
experience and prior rater experience on performance in assessment center
exercises. Academy of Management Journal. E, p. 335-339.

Campion, M. A., & Lord, R. G. (1982). A control systems conceptualization of the goal-
setting and changing process. OrganizationABeghavior and Human Performance,
ﬂ, 265-287.

Cannon-Bowers J. A., & Salas, E. (1990). Cognitive psychology and team training:
Shared mental models in complex systems. Presented at the annual meeting of the
Society of industrial and organizational psychology.

Caplan, R. D., & Jones, K. W. (197 5). Effects of work load, role ambiguity, and Type A

personality on anxiety, depression, and heart rate. Journal of Applied Psychology.
_6_O, 713-719.

122

123

Carver, C. S., & Scheier, M. F. (1982). Control theory: A useful conceptual ﬁ'amework
for personality-social, clinical, and health psychology. Psychological Bulletin, 2;,
11 1-135.

Cattell, R. B. (1963). Personality, role, mood, and Situation-perception: A unifying
theory of modulators. Psychological Review. 70: 1-18.

Child, I. L. (1963). Problems of personality and some relations to anthropology and
sociology. In S. Koch (Ed), Paycholoiyz A study ofaacience (Vol. 5). New
York: McGraw-Hill,

Childers, T. L., Dubinsky, A. J ., & Skinner, S. J. (1990). Leadership substitutes as
moderators of sales supervisory behavior. Journal of Business Research, 2_1, 363-
382.

Cohen, J ., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the
behavioral sciences. London: Lawrence Erlbaum.

Dailey, R. C., Ickinger, W., & Coote, C. L. (1986). Personality and role variables as
predictors of tension discharge rate in three samples. Human Rehations. Q, 991-
1003.

Dansereau, F ., Cashman, J ., & Graen, G. (1973). Instrumentality theory and equity theory
as complementary approaches in predicting the relationship of leadership and
turnover among managers. Olgrmgatiorgtl Behavior and Human Perfomnce. 19,
184-200.

Davis K. (1949). Human Society. New York: MacMillan.

Davis, J. H., Holt, R. W., Spitzer, C. E., & Stasser, G. (1981). The effects of consensus
requirements and multiple decisions on mock juror verdict preferences. Journal of
Experimental Social Psychology, ﬂ, 1-15.

Ford, J. K., Quinones, M. A., Sego, D. J., & Sora, J. A. (1992). Factors affecting the
opportunity to perform trained tasks on the job. Personnel Psychology, 1992.

F oushee, H. C. (1984). Dyads and triads at 35,000 feet. American Psymhologiat, 32, 323-
345.

Graen, G. B., Orris, D., & Johnson, T. (1973). Role assimilation processes in a complex
organization. Journﬂf Vocational Behavior, _3_, 395-420.

Graen, G. B. (1976). Role-making processes within complex organizations. In MD.
Dunnette (Ed.), Handbook of industrial and organizational psychology. Chicago:
Rand McNally.

124

Graen, G. B., Novak, M., & Sommerkamp, P. (1982). The effects of leader-member
exchange and job design on productivity and job satisfaction: Testing a dual
attachment model. Organizational Behgior and Human Performance. £1, 109-
131.

Graen, G. B., & Scandura, T. A. (1987). Toward a psychology of dyadic organizing.
Research in Organizational Behavior. 9, 175-208.

Greller, M. M., & Parsons, C. K. (1992). Feedback and feedback inconsistency as sources
of strain and self-evaluation. Humjaa Relations. ﬂ, 601-620.

Guzzo, R. A. (1986). Group decision making and group effectiveness in organizations. In
P. D. Goodman and associates (Ed), Designing effective work ggoups. San
Francisconossey-Bass.

Hackman, J. R., & Oldham, G. R. (1975). Development of the job diagnostic survey.
Journal of Applied Psychology, 69, 159-170.

 

Hollenbeck, J. R., Sego, D. J ., Ilgen, D. R., & Major, D. (in press). Team decision
making accuracy under difﬁcult conditions: Construct validation of potential
manipulations and measures using the THDE2 simulation. In M. Brannick (Ed)
Advances in team performance measurement. San Francisco: Jossey-Bass.

Hollenbeck, J. R., Sego, D. J., Ilgen, D. R., & Major, D. (1991). Team interactive
decision exercise for teams incorporating distributed expertise (TIDEZ): A program
and paradigm for team research. Technical Report 91-1, East Lansing, Michigan:
Michigan State University.

Howell, J. P., & Dorﬁnan, P. W. (1981). Substitutes for leadership: Test of a construct.
Academy of Management Journal. 23, 714-728.

Howell, J. P., & Dorfman, P. W. (1986). Leadership and substitutes for leadership among
professional and nonprofessional workers. Jourﬂof Applied Behavioral Science,
Q, 29-46.

Howell , J. P., Dorfman, P. W., & Kerr, S. (1986). Moderator variables in leadership
research. Academy of Management Review. E, 88-102.

Howell, J. P., Bowen, D. E, Dorfman, P. W., Kerr, S., & Podsakoﬁ‘, RM. (1990).
Substitutes for leadership: Effective alternatives to ineffective leadership.
Organizational Ilvnamics. l_9, 20-38.

Ilgen, D. R., Fisher, C. D., & Taylor, M. S. (1979). Consequences of individual feedback
on behavior in organizations. Journal of Applied Psychology, Q, 349-371.

125

Ilgen, D. R., & Hollenbeck, J. R. 1991. The structure of work: job design and roles. In
M.D. Dunnette & L.M. Hough (Eds), Handbook of industrial and organizational
psychology. Palto Alto Ca.: Consulting Psychologists Press.

Ilgen, D. R., Major, D, Hollenbeck, J. R., & Sego, D. J. (1993). Team research in the
19905. In M. M. Chemers and R. Ayrnan (Eds), Leadership theory and research:
Perspectives and direction. New York: Academic Press.

 

Ivancevich, J. M., & Matteson, M. T. (1980). Stress and worl_<: A managerial perspective.
Glenview, IL: Scott F oresman.

Jacob, R. (1992). The search for the organization of tomorrow. Fortune pp. 92-98.

 

Kahn, R. L., Wolfe, D. M., Quinn, R. P., Snoek, J. D., & Rosenthal, R. A. (1964).
Organizational stress: studies in and ambiguity. New York: Wiley.

Kanfer, R., & Ackerman, P. L. (1989). Motivation and cognitive abilities: An
integrative/aptitude-treatment interaction approach to skill acquisition. Journal of
Applied Psychology, 74 657-690.

—’

Katz, D., & Kahn, R. L. (1978). The social psychology of organizations: Second edition.
New York: Wiley.

Kerr, S., & Jermier, J. J. (1978). Substitutes for leadership: Their meaning and
measurement. gganizatimal behavior and hurrym performance. 2_2_, 375-403.

Kleinman, D. L. & Song, A. (1990). A research paradigm for studying team decision
making and coordination. Proceediags of JDL Symposium on Command and
Control Research.

Krumboltz, J. D., & Weisman, R. G. (1962). The effect of intermittent conﬁrmation in
programmed instruction. Journal of Educational Psychology, i1, 250-253.

Levine, J. M., & Moreland, R. L. (1990). Progress in small group research. Annual
Review of Psychology, ﬂ, 585-634.

Locke, E. A., & Latham, G. P. (1990). A Theoryof Goal Setting & Task Performance.
Englewood Cliffs, N.J.: Prentice Hall.

Moskal, B. S. (1992). Chrysler polishes the creative wheel. Industry Week, 241. 40-42.

Naylor, J. C., Pritchard, R. D., & Ilgen, D. R. 1980. A theory of behavior in
organizations. New York: Academic Press.

Newcomb, T. M. (1950). Social Psychology. New York: Dryden.

126

Nunnally, J. C. (1978). Pavchometric theory. New York: McGraw-Hill.

Pearce, J. L., & Porter, L. W. (1986). Employee responses to formal performance
appraisal feedback. Journal of Applied Psychology, ﬂ, 211-218.

Raynor, B. (1992). Trial-by-ﬁre transformation: An interview with globe metallurgical's
Arden C. Sims. Harvard Business Review. 29, 117-129.

 

Rouse, W. B., Cannon-Bowers, J. A., & Salas, E. (1991). The role of mental modeﬂ
team performance in complex systems. Unpublished manuscript.

Schmidt, F. L., Hunter, J. E., & Outerbridge, A. N. (1986). Impact of job experience and
ability on job knowledge, work sample performance, and supervisory ratings of job
performance. Journal of Applied Psychology, ]_1, 432-439.

Schmitt, N., Coyle, B. W., & King, L. (1976). Feedback and task predictability as
determinants of performance in multiple-cue probability learning tasks.
Organizationzi Behavior and Human Performance. 1_6, 388-402.

Schmitt, N., & Cohen, S. (1989). Internal analyses of task ratings by job incumbents.
Journal of Applied Psychology, 74 96-104.

—,

Schriesheim, C. A. & Stogdill, R. M. (1975). Differences in factor structures across three
versions of the Ohio State leadership scales. Personnel Psychology. 3, 189-206.

Schultz, R. W., & Runquist, W. N. (1960). Learning and retention of paired adjectives as
a function of percentage occurrence of response members. Journal of
Experimental Psychology, 52, 409-413.

Seers, A. (1989). Team-member exchange quality: A new construct for role-making
research. Organizational Behavior and Humﬁ Decision Processes. _4_3, 118-135.

Sniezek, J. A., & Henry R. A. (1989). Accuracy and conﬁdence in group judgment.
Organizational Behavior and Human Decision Processea, _45, 66-84.

Solomon, C. M. (1991). Behind the wheel at Saturn. Personnel Joumal, 19, 72-74.

Spiker, V. A., Harper, W. R., & Hayes, J. F. (1985). The effect of job experience on the
maintenance proﬁciency of Army automotive mechanics. Human Factors, 2], 301-
311.

Stevenson, M. K., Busemeyer, J. R., & Naylor, J. C. (1990). Judgment and decision-
making theory. In M. D. Dunnette & L. M. Hough (Ed), Handbook of Industrial
& Organizational Psychology. Palto Alto Ca.: Consulting Psychologists Press.

127

Stogdill, R. M. (1974). Handbook of Leadership. New York: The Free Press.

Todd, F. J ., & Hammond, K. R. (1965). Differential effects in two multiple-cue probability
learning tasks. Behﬂioral Science. _1_0, 429-435.

Vance, R. J ., Coovert, M. D., MacCallum, R. C., & Hedge, J. W. (1989). Construct
models of task performance. Journal of Applied Psychology, ’15, 447-455.

Vecchio, R. P., & Gobdel, B. C. (1984). The vertical dyad linage model of leadership:
Problems and prospects. Organizatioaal Behavior aid Human Performance. ﬂ, 5-
20.

Wasabayashi, M., & Graen, G. B. (1984). The Japanese career progress study: A seven-
year follow up. Journal of Applied Psychology, _6_9, 603-614.

Yukl G. A. (1989). Leadership in organizations. Englewood Cliffs, N.J.: Prentice Hall.