AN INTEGRATIVE MODEL OF TEAM ADAPTATION
By

Jeffery A. LePine

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Management

1998

ABSTRACT
AN INTEGRATIVE MODEL OF TEAM ADAPTATION
By

Jeffery A. LePine

As in all organizational systems, teams inherit or develop routines that guide their
behavior. Because routines are programmed sequences of behavior, they do not require
active management, and therefore, are efficient means of accomplishing tasks in stable
conditions. Among the critical aspects of teams, however, is the occasional need to
change the way they go about doing their task in order to respond to changes in the work
environment. However, while costly disasters have been linked with teams’ failure to
adapt to such changes, not much research has been aimed at this issue. The purpose of
this dissertation, therefore, is to develop and test a model that investigates the factors
involved in team adaptation (defined as the reactive and non-scripted adjustments to a
team’s role structure that result in a better fit given the new situation). Information
processing theory serves as the underlying foundation for this model. Team composition
and factors suggested by control theory serve as the model’s other key elements.

The proposed model centers on three points. First, while routine activity does not
require team members to actively process information, the type of activity required for
team adaptation does. Second, the factors that influence (a) members’ capacity to process
information actively and (b) the capability to “switch-on” active cognitive processing
when the situation warrants it are critical for team adaptation. Finally, among

characteristics of the team that influence these factors, team composition (team member

characteristics including general cognitive ability, conscientiousness, and openness to
experience), the nature of team goals, and information from the environment regarding
progress towards goals are perhaps most important. Such information depends on the
quality of feedback as well as the nature of the environmental change.

’ Hypotheses were tested in a laboratory setting using a computerized decision
making task and 141 three-person teams. Consistent with past descriptive accounts,
failing to adapt was both prevalent and costly (in terms of team performance). As
expected, teams in abruptly changing conditions or composed of members with high
levels of cognitive ability and openness to experience were more likely to adapt than
teams in gradually changing conditions or when composed of members with low levels of
cognitive ability and openness to experience. Also as expected, the effect of team
members’ cognitive ability on team adaptation was stronger for teams with members who
had high levels of conscientiousness. Unexpectedly, however, there were no main effects
on adaptation for goal difficulty, quality of feedback or members’ conscientiousness.
There was also an unexpected finding in that members’ cognitive ability appeared to be
more highly related to adaptation for teams with easy goals. Findings from post hoc
analyses are consistent with literature suggesting that difficult goals create performance
pressures that cause members to focus their attention on the team’s standing in terms of

their performance instead of on learning how to do the task in the changing environment.

ACKNOWLEDGMENTS

I would like to start off by saying that it would have been impossible to have a
better graduate school experience. I learned a great deal and had a lot of fun doing it. John
Hollenbeck (my chair) was an outstanding mentor and role model and provided me with
the opportunity to be successful. I thank Dan Ilgen and Neal Schmitt for all their incite
and helpful advice. I also thank Linn Van Dyne for always pushing me into considering
the “bigger” picture. The learning and fun I had in graduate school can also be attributed
to a great group of graduate students--especially Jason Colquitt, Marcia Simmering,
Becky Luce, and Mike Wesson. I would also like to thank Alex Ellis, Henry Moon, Lori
Sheppard, Amy Pepper and Christine Jackson for their help during the experiment.
Finally, I would like to thank Mimi for putting up with my absence (as well as my

presence) during these last five years.

iv

TABLE OF CONTENTS

List of Tables ..................................................................................... x
List of Figures .................................................................................... xi
INTRODUCTION .............................................................................. 1
Definition of Teams ........................................................................ 3
The Nature of Routine ..................................................................... 5
The Dilemma ............................................................................... 6
Specialist Decision Making Teams ...................................................... 16
Specific Examples of Specialist Decision Making Teams ............................ 19
Purpose ....................................................................................... 23
CHAPTER 1
LITERATURE REVIEW ...................................................................... 25
Information Processing as a Theoretical Foundation .................................. 25
Problem sensing ....................................................................... 26
Groups as information processors ................................................... 27
The Basis for Routine and Adaptation ................................................... 29
Fitts theory ofmotor skill development..................................... ....... 3O
Anderson’s ACT Theory ............................................................. 31
Routine: Distributed procedural knowledge ....................................... 34
Adaptation: Distributed declarative knowledge ................................. 37
Cognitive Resources ....................................................................... 40
Individual differences in cognitive resources ..................................... 42

Team differences in cognitive resources ........................................... 44

Triggering Declarative Knowledge Based Information Processing .................. 49
Content of information processing control ........................................ 49
Process of information processing control ......................................... 50
Summary ............................................................................... 51

Control Theory .............................................................................. 51
Application to a wide variety of systems .......................................... 52
Hierarchical structure of control loops ............................................. 54
Inappropriate routine as control system failure .................................... 56

Goals ......................................................................................... 57

Feedback .................................................................................... 64

Nature of Environmental Disturbance ................................................... 68

Alternative Reactions to Disturbances ................................................... 71
Team conscientiousness resources .................................................. 72
Team openness resources ............................................................ 76

Summary and Hypotheses ................................................................. 78

CHAPTER 2
METHOD AND ANALYSES ................................................................. 83

Choice of Setting ........................................................................... 83

Planning for Participants: The Power Analysis ........................................ 84.

Study Participants .......................................................................... 86

Task .......................................................................................... 87

Procedure .................................................................................... 93

vi

Overall Design .............................................................................. 96

The Pilot Study ............................................................................. 97
Evidence of routine ................................................................... 97
Necessity for Adaptation ............................................................. 99
Establishing easy and difficult goals ................................................ 99
Manipulation check ................................................................... 100

The Pilot Study: Results .................................................................. 100
Evidence of routine ................................................................... 100
Necessity for adaptation .............................................................. 102
Easy and difficult goals ............................................................... 106
Manipulation check ................................................................... 107
Summary of pilot results ............................................................. 107

The Primary Study: Variables ............................................................ 108
Adaptation .............................................................................. 109
Goal difficulty ......................................................................... 112
Quality offeedback ....... 112
Abruptness of environmental disturbance .......................................... 113
Team cognitive resources ............................................................ 116
Team conscientiousness resources .................................................. 116
Team openness resources ............................................................ 1 17
Team decision accuracy .............................................................. 117

The Primary Study: Analyses ............................................................ 118

vii

CHAPTER 3

RESULTS OF PRIMARY STUDY ........................................................... 119
Overview of Results ........................................................................ 119
Individual-level Learning and Controlled Cognition .................................. 119

The cognitive resources-learning relationship .................................... 120
The manipulations-controlled cognition relationship ............................ 121
Summary ............................................................................... 122
Descriptive Statistics of Team-Level Variables ........................................ 123
Team resources ........................................................................ 123
Manipulations ......................................................................... 123
Team performance .................................................................... 125
Adaptation .............................................................................. 126
Analytic Strategy ........................................................................... 127
Assessment of Hypotheses Using Logistic Regression .............................. 132
Team Resources ....................................................................... 134
Manipulations ......................................................................... 134
Interactions ............................................................................. 134
Exploration of Alternative Operationalizations ........................................ 138
Conjunctive aggregation ............................................................. 139
Disjunctive aggregation .............................................................. 139
Predicting Speed of Adaptation .......................................................... 142
CHAPTER 4
DISCUSSION .................................................................................... 145

viii

Overview of Problem and Theory ........................................................ 145

Overview of Pilot Study Results ......................................................... 146
Overview of Primary Study Results ..................................................... 147
Individual-level elements ............................................................. 147

Team cognitive resources ............................................................ 148
Triggers of controlled information processing .................................... 148

Factors that influence mode of discrepancy reduction ........................... 149
Understanding the Moderating Effect of Goal Difficulty ............................. 150
Absolute goal levels versus performance-goal discrepancies ................... 150

Results of post-hoe analysis ......................................................... 152
Implications of Results .................................................................... 155
Implications to theory ................................................................ 155
Implications to practice ............................................................... 155
Limitations .................................................................................. 159
Overall Conclusion ........................................................................ 161
APPENDIX A: Team Decision Making Consent Form (initial) .......................... 164
APPENDIX B: Team Decision Making Consent F orrn (experiment) .................... 166
APPENDIX C: General Overview Handbook ............................................... 168
APPENDIX D: Role Instructions ............................................................. 174
APPENDIX E: Experiment Protocol ......................................................... 178
APPENDIX F: Task Knowledge Test ........................................................ 192
APPENDIX G: Post Experiment Questionnaire ............................................ 196
LIST OF REFERENCES ....................................................................... 198

ix

LIST OF TABLES

Table 1 - Comparisons of Time to Receive Information .................................... 101
Table 2 - Comparisons of Reliability in Receiving Information ........................... 102
Table 3 - Descriptive Statistics for Teams .................................................... 124
Table 4 - Correlations of Study Variables with Dichotomous Adaptation ................ 128
Table 5 - Logistic Regression of Additive Independent Variables on Adaptation ....... 133

Table 6 - Logistic Regression of Conjunctive Independent Variables on Adaptation... 140
Table 7 - Logistic Regression of Disjunctive Independent Variables on Adaptation... 140
Table 8 - Relationships with Speed of Adaptation ........................................... 142

Table 9 - Logistic Regression of Discrepancy Variable on Adaptation ................... 151

LIST OF FIGURES

Figure 1 - An Integrative Model of Team Adaptation ....................................... 24
Figure 2 - A Routine Among Four Actors ..................................................... 34
Figure 3 - Hierarchical Goal Structure in Teams ............................................ 61
Figure 4 - Role Requirements .................................................................. 88
Figure 5 - Measurable Attributes ............................................................... 90
Figure 6 - Pre-disruption Role Structure ...................................................... 90
Figure 7 - Ideal Adapted Role Structure ....................................................... 110
Figure 8 - Plot of Team Cognitive Resources by Goal Difficulty Interaction ............ 135

Figure 9 - Plot of Team Cognitive Resources by Team Conscientiousness
Resources Interaction ................................................................ 136

Figure 10 - Plot of Team Cognitive Resources by Team Openness
Resources Interaction ............................................................... 137

Figure 11 - Plot of Team Cognitive Resources by Goal Difficulty by Quality of
Feedback Interaction ................................................................. 138

Figure 12 - Stern and Leaf Plot of Adapting Teams’ Speed of Adaptation ................ 143

Figure 13 - Plot of Team Cognitive Resources by Performance-Goal
Discrepancy Interaction ............................................................ 152

Figure 14 - Plot of Performance-Goal Discrepancy by Quality of
Feedback Interaction ................................................................ 154

xi

INTRODUCTION

Although not mutually exclusive to one another, interest in small groups and
teams in organizations can best be described in terms of three somewhat overlapping and
parallel waves. The first wave can be attributed to the Hawthorn studies (Roethlisberger
& Dickson, 1939) which began in the 1920’s. These studies suggested that informal
groups influence the behavior of individuals in work settings. Whyte’s (1955) research is
an example of work in this stream. By showing that group unity was more important than
maximization of personal income, Whyte illustrated the importance of the group on
individuals. Out of this tradition grew the Socio-technical movement (e. g., Trist &
Bamforth, 1951) that assumed that organizational effectiveness demands joint
optimization of both the technical and social systems (informal groups). The importance
of the social aspect of this equation was new and thus triggered interest in groups as units
of interest in their own right. Research in this stream during the 1960’s through 1970’s
generally focused on increasing member satisfaction with the group experience. This
research, often referred to as “team-building,” attempted to improve the nature of
collaboration within the group or relationships among group members (Beckhard, 1989).

The second wave of interest in teams began in the mid to late 1960’s and early
1970’s as American firms began to recognize Japan as a competitive threat. Japan had
used team structures as a way of involving workers in order to increase quality of
industrial processes and end-products to world class levels. Throughout the 1970’s,
project teams, task forces, autonomous work groups, quality circles and new product

development teams were being implemented to such an extent that group structures began

to dominate industrial processes (Thurow, 1983). By the late 1970’s, groups and teams
were increasingly being referred to as basic organizational units (Davis, 1977) or building
blocks (Peters & Waterman, 1982). Indeed, this trend has been recognized as a
revolutionary shift in managerial thinking (Kanter, 1983; Ketchum, 1984; Peters, 1988;
Reich, 1987; Walton, 1985). Whereas interest in the first wave centered on the
importance of the group to the individual, the second wave viewed group structures as
being a potential mechanism for gaining an advantage over competitors.

Today, increasing global competitiveness and information proliferation have
created a great deal of uncertainty for organizations. As a result, the importance of
flexibility or adaptability to organizations has never been more important. Increasing
global competition makes it imperative that organizations can grow or contract to meet
changes in demand. The rapid pace of technology transfer demands that organizations '
quickly adopt innovation. As a response to these demands, organizations have de-coupled
specific tasks from individuals and assigned units of work to teams (Hoerr, 1989).
Flexibility is enhanced in such structures because teams can be composed (or
responsibilities allocated among members) to take advantage of specialized skills and
resources as the demands of the situation change. These types of teams have been
credited with successes at Chrysler (Byme, 1993; Treece, 1992; Woodruff, 1991, 1993)
and thousands of other organizations over the last ten years (Hammer & Charnpy, 1993).

Thus, in many respects, what distinguishes the third wave of interest in small
groups and teams from the second, is that teams are now viewed as more of a necessity
than ever before. That is, whereas in the past teams were implemented in an effort to gain

competitive advantage, now the use of teams is necessary just to compete. However,

while a great deal of research has begun in order to better understand, predict and control
important team processes and performance, reviews of this research suggest that we still
have much to learn (Bettenhausen, 1991; Guzzo & Dickson, 1996; Guzzo & Shea, 1992;
Ilgen, Major, Hollenbeck, & Sego, 1994; Levine & Moreland, 1990; Sunstrom, De
Meuse, & F utrell, 1990). The purpose of this dissertation is to address one specific and
important team problem--how teams adapt to changing environmental conditions. Before
I go further, however, it is necessary to define explicitly what is meant by the word team.
Definition of Teams

Consistent with Katz and Kahn (1978) and other open system theorists (e. g.,
Boulding, 1956; J. Miller 1955, 1965a, 1965b; von Bertalanffy 1968), I will assume that
individuals, groups, organizations, and so on, are partially nested open systems. 116%
refers to the idea that individuals behave in groups which are embedded in organizations.
9&1 refers to the idea that each focal system is dependent on higher level systems for the
resources it needs to survive. Survival requires that each focal system takes resources and
transforms them into outputs. Finally, gym refers to the cyclical patterns of activity
which transform resources to outputs. System boundaries, therefore, are defined in terms
of the set of events that “return upon themselves to complete and renew a cycle of
activities” (Katz & Kahn, 1978, p. 24). For example, a simple group-level social system
consisting of actor A, B, and C might exist if actor A elicits a response from actor B, and
this response, in turn, stimulates C, who, in turn, stimulates A. In this dissertation, I focus
on group-level systems of a specific type-~the team.

Although there are numerous definitions of teams, I use one that synthesizes most

to date (Ilgen et al., 1994). The first three elements of their definition of teams are

common to small groups and teams and were described by Steiner (1986) in his review of
the small groups literature. First, a team must consist of at least two individuals. Second,
these individuals must interact. Third, not only must the group interact, but there must be
interdependence between the actors. Finally, teams exist for some task related purpose.
Members interact to achieve task related goals which members are generally aware of and
to a great extent share. It is this shared purpose element that most clearly differentiates a
team from a group of individuals. For example, six individuals who meet on Thursday
nights to discuss gardening are clearly a small group, yet they would not normally be
considered a team. While such a group is composed of multiple individuals who interact
and who are interdependent (to the extent their discussion depends on everyone’s
contribution), they do not share a task related common goal (e.g., the individuals are not
trying to grow a certain amount of corn). In the context of this dissertation, a team’s
purpose is generally implicitly or explicitly determined by others (such as those in
management). That is, this dissertation will not focus on the types of teams which define
their own purpose.

As Guzzo and Dickson (1996) point out, however, the difference between groups
and teams is in “degrees of difference” rather than “fundamental divergence”. And
although the word “team” is the current jargon in the organizational behavior, human
resources, industrial and organizational psychology, and popular management literatures,
there is along history of research on small groups that share the same elements of the
above definition. Thus, although the concern in this dissertation is with teams as defined

above, this dissertation will draw liberally from the literature on groups.

The Nature of Routine

In many teams, the organization of work is rationally designed to meet the
demands of the task. The designers of such teams work outside the team, typically as
upper level managers or perhaps as consultants. Typically, designers of such teams try to
specify individual tasks and how these tasks are to be integrated in order to arrive at team
products. This approach to work organization includes socio-technical designs which
emphasize social as well as task considerations (Cummings, 1978).

In other instances, work in teams is organized through compliance with
institutionalized norms (Zucker, 1977). For instance, individual specialists may come
together to perform a task for which they each have a clear understanding of their role
‘and what they can expect from other team members. While rational design may play a
role in how these tearns are structured by providing a general organizing fiamework,
actual responsibilities are governed by a set of “taken for granted” norms. Cockpit flight
crews are a good example of this type of team (Ginnette, 1990).

Finally, work in teams can be organized as the result of some developmental
(Tuckrnan, 1965; Gersick, 1989) or role assimilation process (Hollenbeck, LePine, &
Ilgen, 1996). Depending on the specific research perspective, team members are seen as
either reactive (Katz & Kahn, 1978) or proactive (Graen, 1976) in developing roles,
which when viewed as a system, eventually stabilize. Work in self-managed teams is
organized in this manner.

While all three organization mechanisms are conceptually distinct, they are
similar because they lead to the establishment of standard operating procedures, job

descriptions, role expectations, or routines (Cyert & March, 1992; Gersick & Hackman,

1990; March & Simon, 1992; Weick, 1979). Such outcomes, henceforth called routines,
are natural in systems where there is a drive to reduce uncertainty (Thompson, 1967). The
prevalence of routines is evidenced by their existence across many types of groups
(Argyris, 1969) and govern most behavior in organizations (March & Simon, 1992; 141-
142). The creation of routines is generally attributed to “repeated success with particular
behaviors in relatively constant problem environments” (Weiss & Ilgen, 1985, p. 59). To
Gersick and Hackman (1990, p. 69) “a habitual routine exists when a group repeatedly
exhibits a functionally similar pattern of behavior in a given stimulus situation without
explicitly selecting it over alternative ways of behaving”. According to this definition,
routines are (l) observable (patterns of behavior), (2) mindless (only superficial if any
conscious consideration of alternatives), and can (3) vary in strength (how ofien the
similar pattern is triggered). Over the last few decades researchers have invested a great
deal of effort describing routines and standard operating procedures in groups and
organizations and this research suggests that routines can be both beneficial and
detrimental.
The Dilemma

Routine behavior has a number of firnctional consequences which have been well
documented throughout the past half-century. These consequences are not trivial, and as
many have noted, are necessary for the survival of systems at both group and
organizational levels of analysis.

First, because group members are interdependent and must coordinate with each
other, routines provide a mechanism whereby members can anticipate each others’

actions (Argote & McGrath, 1993). Being able to anticipate others’ actions is critical in

situations where members depend on others to carry through with their responsibilities.
As an extreme example, routines are highly beneficial to trapeze artists who must, with
great confidence and without hesitation, twirl through the air into the waiting arms of
another.

Routines also allow increased efficiency (i.e., save time and energy) because they
need not be actively managed (Gersick & Hackman, 1990). Because repeated situations
are handled the same way automatically, complex and time consuming planning and
strategizing is not necessary. In simple terms, routines eliminate wasted resources caused
by constantly having to “reinvent the wheel”. In addition, because routines invoke the
same solution on a problem, system actors become well practiced in the behaviors
specified in the routine. This is especially beneficial where many actors must coordinate
non-redundant or specialized behaviors.

Routines also increase reliability in performance by stamping out deviations from
“normal” behavior (Levinthal & March, 1993). This is most beneficial in systems where
avoiding failures is more important than reaching potential upper limits in terms of
performance and in stable situations with knowledge of past successful behavior. As
March and Simon (1992) suggest, this type of satisficing behavior is inherent in
organizational systems with bounded rationality.

Finally, routines reduce the uncertainty members have regarding their role
responsibilities (Gersick & Hackman, 1990). In this way, routines increase individual
actors’ level of comfort and confidence. In addition, because routines are accepted

automatically, overt competition and disagreement between individuals in groups may be

' avoided. In this way, routines are beneficial to the cohesion, and perhaps long terms
survival, of groups.

Given the benefits of routines outlined in these paragraphs, it is understandable
why routines are prevalent in group and organizations (Gersick, 1988; Gersick &
Hackman, 1990; Hackman, & Morris, 1975). Indeed, Cohen and Bacdayan (1996)
showed that routinized behavior is probably common even in simple systems with
relatively short life-spans. In their study, routines were developed in the concocted
laboratory dyads using a simple task (card game) over a relatively short time period (40
hands in 40 minutes). Thus, it appears that routines are developed even in simple systems
where the importance of developing such patterns would seem to be minimal.

Among the most critical aspects of teams, however, is the occasional need to
change the way they go about doing their task in order to respond to changes in their
context or environment which makes their routine inappropriate. Teams must be able to
deal with unanticipated factors and modify their actions accordingly or team effectiveness
will suffer (Argote & McGrath, 1993). Routines are particularly problematic in this
regard because they are triggered mindlessly and thus can be transferred onto
inappropriate situations (Cohen & Bacdayan, 1996). As Gersick and Hackman (1990, p.
72) point out, routines invite miscoding of stimuli because groups are likely to see what
they know or what they expect, and thus, may fail to recognize that the context has
changed or that they are in a new context entirely. Indeed, research concludes that once
“the” way of going about doing things becomes established, group members tend not to
think about or discuss their strategy (Hackman, 1987, p. 328) never mind change it

(Hackman, Brousseau, & Weiss, 1976; March & Simon, 1992, chapter 6).

Allison’s (1971, p. 109) humorous description of civilian-clad Soviet troops
marching off a dock after they secretly arrived in Cuba illustrates the mindless activation
of routines. Gersick and Hackman’s (1990) account of the Air Florida Flight 90 crash
points out that the inappropriate transfer of routine behavior can have less humorous
consequences, however. In this example, a flight crew that normally flew in warm
weather did not take appropriate action during operations in freezing weather. The captain
mindlessly responded “off’ after the first officer read “anti-ice” from a checklist and the
first officer mindlessly went on reading. As a result of the failure to use the de-icing
equipment, the plane crashed killing 74 people.

Thus, recognizing that the establishment of routines is often desirable and
probably inevitable (Gersick, 1988; Gersick & Hackman, 1990; Hackman, & Morris,
1975), the critical question becomes one of developing an understanding of how teams
abandon or change these “habits” when they become inappropriate. Researchers have
examined this dilemma for decades. For instance, Lewin (1951) recognized that change in
organizations first requires “unfi'eezing” the old situation. Starbuck and Hedburg (1977)
noted that learning in organizations first requires “unlearning”. This research, however,
has been almost exclusively conceptual, normative, and quite abstract (e.g. Argyris, 1976;
Argyris & Schon, 1978; Fiol & Lyles, 1985). Thus, in order to provide some precision to

the topic being studied here, I will focus on “team adaptation”, which I formally define

as; reactive and non-scripted adjustments to a team’s role structure that result in a better

fit given the new situation.

Decomposing this definition shows how adaptation is different from related

concepts such as learning and innovation. The first element in the definition of team

adaptation is the word reactive. Reactive implies response to something, typically a

 

problem, error, or discrepancy. Thus, this element stands in contrast to the concept of
innovation which is typically thought of in terms of proactive change in order to take
advantage of opportunity.

The second element, W, refers to the fact that adaptation does not refer
to reactions to situations (problems) by implementing intact procedures or routines which
were learned in the past. That is, adaptation is not just the reactivation or reinitializaiton
of old schemes (Eckblad, 1981), scripts (Abelson, 1981; Gioia & Manz, 1985; Gioia &
Poole, 1984; Lord & Keman, 1987; or stimulus-response associations. This is different
from notions of learning which are based on the development of proficiency regarding
when to implement lessons from the past (i.e., transfer). Adaptation refers to active and
controlled search for, experimentation with, and implementation of new ways of doing
things in light of a new situation. This is not to say that the adaptation comes from
completely “new” knowledge, but that team members’ knowledge of facts, rules, and
principles is recast in light of the new situational context. That is, old knowledge of team
members is used as a basis for developing a new, more appropriate system of behavior.
The importance of this point will become clear later.

The next element in the definition of adaptation is the word adj ustment.
Adjustment highlights that adaptation implies activity or behavior. Thus, adaptation is
distinct from organizational learning which can mean a change in knowledge without any
change in activity or behavior. Hedburg (1981) stresses that it is important to distinguish
between processes which affect (a) shared interpretation of events, (b) the development of

shared understanding and conceptual schemes, and (c) responses and behaviors, because

10

they are different phenomena. As F ioil and Lyles (1985) note, changes to cognition (a and
b) or behavior (e) do not necessarily reflect changes in the other. Indeed, Hutchins (1996)
has recently demonstrated that changes in the environment can cause team behavior to
change without any shared interpretation of events among team members. Thus, the
definition of adaptation being developed here is closer to conceptualizations proposed by
Fiol and Lyles (1985) and Miller and F riesen (1980) who stress behavior change than to
conceptualizations proposed by Meyer (1982) and Chakravarthy (1982) who stress
cognitive change.

Together, reactive non-scripted adjustments imply behavior change that takes
place in the course of activity as opposed to behavior change that takes place in order to
do the project. Argote and McGrath (1993) make this distinction when they distinguish
teams’ reconstruction processes (modifications of the system as a consequence of doing a
task) from teams’ construction processes (modifications of the system in order to do a
task). The latter process has been the subject of a great deal of research (e.g., Barley,
1986; Poole, Holmes, & Desanctis, 1991; Poole & Roth, 1989a; Poole & Roth, 1989b;
Poole, Seibold, & McPhee, 1985).

The element m in the definition stresses that the focus is on changes in
behavior at the team-level (as defined earlier). That is, adaptation is conceptualized as
changes in the cyclical patterns of activity among those individuals who are part of a
team (as defined earlier) in the course of transforming resources to outputs within the
boundary of the set of events which return upon themselves to complete a new cycle of

activity. While concepts of learning and innovation may occur at different levels, it is

11

best to explicitly ground theory within a level so that biases and fallacies are avoided
(Rousseau, 1985).

By using role structure, this definition also includes the behavioral focus of
change. Using Katz and Kahn (1978, p. 189) as a basis, role structure can be defined as
the pattern of recurring actions of individuals interrelated with the recurring actions of
others. In essence, role structure refers to the cyclical pattern of activity among team
members discussed earlier and in the previous paragraph. This conceptualization is
consistent with Katz and Kahn who view organizations as “open systems of roles” (p.
187)

Finally, the definition includes the requirement that adaptation result in a system
of behavior fit better fits the new situation. This somewhat normative component to the
definition is necessary because adaptation implies increased fit, consistency, congruity or
correspondence between team structure and the situation. That is, from an objective
standpoint, changes in a system of behavior must be more effective in the new situation
than the old system was (or would have been). Activity which simply results in a
different system of behavior cannot be said to be adaptive because it does not bring the
team in closer correspondence with the demands of its situation. This requirement is
necessary to distinguish adaptation from “forgetting” or “change or change for the sake of
change”. I should also note that in this element of the definition (i.e., movement towards
better fit), it is implied that adaptation is a response to negative as opposed to positive
discrepancies between perceived and desired states.

Despite the importance of adaptation to teams (and thus to the organizations

teams are nested in), there has been very little theoretically driven empirical research on

12

this topic. Indeed, this was noted 30 years ago by researchers who stated that “[o]ne of
the least understood phenomena in task-oriented team performance is the manner of
adjustment to unprogrammed changes” (Behling, Coady, & Hopple, 1967). What
research exists, focuses almost exclusively on describing how routines form and why they
are difficult to change (Allison, 1971; M. Cohen, 1991; M. Cohen & Bacdayan, 1994;
Cyert & March, 1992; Gersick & Hackman, 1990; Levitt & March, 1988; March &
Simon, 1992; Nelson & Winter, 1982; Weiss & Ilgen, 1985) and thus only highlights
importance of team adaptation.

Perhaps the lack of research on team adaptation lies in the complexity involved in
conducting it. Indeed, team adaptation not only involves multiple actors who may have
only a partial understanding of an entire system, but adaptation is also emergent. That is,
it is difficult to specify, a priori, what a new system of team-level behavior should entail.
In fact, if a priori specification of a new system were possible, then adaptation would be
unnecessary-~planning would be more efficient. Thus, given the nature of adaptation, it is
understandable why theoretical research on adaptation has not progressed much further
than description. However, given the increasing prevalence of teams that operate in an
increasingly dynamic environment, the need to increase our understanding of team
adaptation has never been greater.

One stream of literature which would seem to be highly relevant to our
understanding of team adaptation is the research on individuals’ ability to perform in new
or changing situations. For instance, attributes such as change-orientation (Saville &
Holdsworth, 1990), experience-seeking (Hogan & Hogan, 1992), flexibility (Gough,

1987; Paulhaus & Martin, 1988), social intelligence (Cantor & Kihlstrom, 1987; Zacan'o,

13

Gilbert, Thor, & Mumford, 1991), and openness to experience (Costa & McCrae, 1992)
all would seem to be relevant to the topic of this dissertation. There are, however, at least
three problems with using this literature as a starting point for developing an
understanding of team adaptation.

First, it is possible that a person on a team with certain characteristics may attempt
to change his or her role to meet the demands of a changing situation, and if successful,
change the role structure of the team. However, the other members of the team would
have to “allow” this to occur and various lines of research would predict against this .
happening. As Hackman (1992, p. 248) summarizes: “Uniformity, conformity to norms,
and adherence to one’s role is the rule. When someone deviates, other members provide
potent doses of discretionary stimuli intended to persuade or coerce the person to get back
into line.” Obviously there are situations which would mitigate against this happening
(e. g., the deviant has power over the other team members, or the deviant has a suggestion
which is clear and unambiguously correct), however, there is no theory which integrates
these factors with deviant and team member characteristics.

Second, it is difficult to choose the best variables from among the list of possible
alternatives. Obviously, one could use brute force and try them all. However, there are a
number of flaws with not using theory guided variable selection, not the least of which
includes being left with having to draw conclusions based solely on empirical results.
This is a problem because without theoretical understanding, it would be difficult to
specify the boundaries of generalizability. In order to overcome this problem, one would
have to examine each variable in different settings and then infer which variable worked

best in each setting. While such an approach would increase our ability to predict

14

adaptation, it would do so in a very inefficient manner. In addition, it does nothing to
increase our understanding of adaptation in terms of underlying psychological
mechanisms. Furthermore, since many of these concepts are redundant with one another,
it is not likely that such an approach would lead to a very parsimonious literature base.

A third problem is that it might be misguided to assume that the relationship
between adaptability (originally defined in terms of some individual-level characteristic)
and adaptation is homologous across the individual- and team-levels. The questions that
must be asked in this regard relate to the extent to which individual-level scores on
adaptability mean the same thing as when these individual-level scores are aggregated to
represent the team. Such questions are critical since assuming homology can lead to a
number of biases and fallacies (Rousseau, 1985). Addressing this problem requires the
development and testing of theory which relates the individual- and team-level concepts
(i.e., composition theories). However, since this stream of literature simply relates trait
concepts to measures of learning and performance without much theory or consideration
of the role of contextual factors (the nature of the work, the work setting, others in the
group, etc.), it may be very difficult to specify adequate composition theory. In addition,
since there are so many possible attributes to examine, it is doubtful that this approach
would lead to a parsimonious theory of adaptation. I

Overall, while it is possible to approach the issue of team adaptation from using
the empirical base relating attributes of individuals to their adaptability, such an approach
is not likely to be effective in increasing our understanding of team adaptation. This is not
to suggest that the characteristics of team members are unimportant in team adaptation.

Indeed, this dissertation views the composition of the team in terms of individual

15

attributes to be central to team adaptation. Instead, what I suggest is that research needs to
be grounded in theory which explicitly considers team- or group-level phenomena. This
dissertation takes such an approach. However, the first step in this process must be to
explicitly state the domain of interest because progress in understanding can be made best
when theories focus on a particular subdomain of phenomena rather than when theories

attempt to be all-encompassing.

Sgcialist Decgsion Making Teams

 

While there are many different type of teams that do work in organizations
(McGrath, 1984; Sundstrom et al., 1990), teams have begun to take on a number of
defining characteristics and do not fit neatly in any current theoretical taxonomy or
categorization scheme. First, because the proportion of service jobs has increased
(relative to the proportion of manufacturing jobs) and because manufacturing is
increasingly becoming computerized, teams are increasingly doing cognitive as opposed
to physical work (Beyerlein, Johnson, & Beyerlein, 1995). That is, teams are increasingly
doing their work by applying knowledge and making decisions (e. g., what type of
products to develop, how to set up a cells in a flexible manufacturing systems) rather than
performing psycho-motor tasks (e.g., mining).

Second, while teams can be found at all levels of organizations, they are
increasingly being used at lower levels where the basic work is done. That is, while there
are top management teams, review panels, boards, and advisory councils which are
involved in problem definition (establishing the type of work to be done in the
organization), there are also an increasing number of teams that work on established and

well defined problems (e. g., flexible manufacturing systems, product development).

16

Work in such teams consists of making a recurring set of fairly structured decisions
which can be evaluated in terms of being right or wrong. Feedback indicating the
correctness of these teams’ decisions varies in terms of both preciseness and time-lag. For
instance, while surgical teams get fairly immediate and precise feedback after a major
operation (e.g., the patient either lives or dies), product development teams receive
feedback which is more delayed and is stated in terms of somewhat more ambiguous free-
market criterion. In some of these teams the recurrence of similarly structured decisions
occurs over a relatively short period of time while in others the recurrence of decisions is
more spread out. For instance, a team responsible for configuring computerized
equipment for small batch production runs will make more decisions over a period of
time (e.g., two reconfigurations per week) than a product development team (e. g., two
product introductions a year). Regardless of the timing, however, these teams make
decisions which are recurring, and therefore should be more susceptible to problems
associated with routine behavior than teams responsible for one-time or unique decisions.
Third, because of rapid advances in technology, especially information
technology, teams are increasingly becoming structured with professionals or specialists
who have unique knowledge and skills (Tjosvold & Tjosvold, 1995). Because it is
necessary to integrate multiple areas of expertise in order to arrive at effective team
decisions, these teams can be characterized by their “cooperative interdependence”
between members (Argote & McGrath, 1993). This means that team members exchange
information and resources and coordinate activities with each other cooperatively. Such
teams are distinct from those where individual members’ contributions are relatively

independent or collaborative (e.g., brainstorming) or where members are in competition

17

with one another (e.g., negotiating or resolving conflicts of interest) (Argote & McGrath,
1993). Cooperation is critical because team members are reciprocally interdependent.
That is, each member depends on the other members for informational and/or resource
inputs (Thompson, 1967; Van de Ven, Delbecq, & Koenig, 1976).

Finally, given the points above, the tasks facing many teams in organizations can
be characterized as having a moderate degree of complexity. According to Wood (1986,
pp. 66-71) the three dimensions of task complexity include: W (i.e.,
number of direct acts that need to be executed or the number of information cues that
must be processed in the performance of those acts), coordinative complexig (form and
strength of the relationship between information cues, acts, and products, as well as the
sequencing of inputs), and dynamic complexig (the frequency of which individuals must
adapt to changes in the causeueffect chain or.means--end hierarchy). Because teams are
responsible for entire units or work, it is likely that in most situations performance will
involve at least a moderate amount of component complexity. Because teams are
composed of specialists who are reciprocally interdependent, there is likely to be at least
a moderate amount of coordinative complexity. Finally, because the competitive
environment is dynamic (Howard, 1995) recurring work cycles tend to be short which
means at least moderate levels of dynamic complexity.

As stated in a previous paragraph, the types of teams described above do not fit
neatly in any single scheme. For instance, Sundstrom & Altman (1989) categorized work
groups into four types as defined by four characteristics: work team differentiation
(redundancy of member roles), external integgtion (synchronization with constituents

external to the group), work cycles (the length and uniqueness of a performance event),

18

and apical outputs (physical versus non-physical outcomes). In terms of these
characteristics, the types of teams described above are most clear in terms of having a
high degree of work-team differentiation. However, these teams may have varying
degrees of external integration and length of work cycles. In addition, while the process
of these teams’ work consists of making decisions (the correctness of which can be

considered non-physical), the ultimate outcomes of those decisions may be physical (a

 

new product, a downed aircraft). Thus, for the lack of a better term, I will use the term
Smgialist decision making team to refer to the type of team I described in the last few
paragraphs. This dissertation focuses on these teams because of their increasing
prevalence, importance, and susceptibility to problems with routine (i.e., they make
decisions which are recurring). I should note that these teams can be found in a wide
variety of contexts including business, education, military, and medical. What follows are
a few specific examples.
Specific Examples

In the context of business, a manufacturing organization located in the Midwest
uses specialist decision making teams in the mass production of glass bottles (Pagell,
1997). In this organization, teams are composed of four to five individuals who set-up
production rims and monitor the production process. Set-up responsibilities include
configuring a computerized forming machine which holds up to 24 molds. Configuration
includes mounting the molds in the forming machine and setting initial production
parameters (e.g., temperature, process speed, fill-rate, pressure, etc.) according to the type
of bottle to be produced (size, material, color, finish, etc.). Specialization in these teams

consist of differences in responsibilities rather than in specific areas of expertise as

19

defined by specialized training or education. Recurrence or set-up decisions correspond to
production runs which last between two and three days. During production, team
members monitor sensors on the forming machine and make adjustments to production
parameters such as “temperature”, “speed”, “flow”, and “pressure” in order to keep the
production process going. Team members are interdependent because a change in one
parameter (e.g., temperature) impacts the others (e. g., required pressure, speed, flow).
Any team member can make minor adj ustrnents to parameters, however, when major
adjustments are necessary, or when there are disagreements, a team leader will consider
team members’ opinions and make the final decision. Cooperation is the norm in these
teams because breakdowns are very costly in terms of lost production time and clean—up.
In fact, because teams work in close proximity to molten glass, breakdowns can be very
dangerous. Feedback regarding the correctness of decisions, therefore, is rather
immediate.

Another business example includes the use of teams in the machining of tool parts
(Pagell, 1997). In one such organization, teams are composed with a lead operator (team
leader), three operators, a tool setter and a computer programmer. These teams are
responsible for the setting-up of machines and the production of tool parts using five
configurable computer assisted machines. When a new order for a part comes in, the team
gets together and uses its distributed expertise to make decisions regarding what
machines to use, what “cuts” to make, the order of cuts, and how to “fixture” the
machines (how to configure the machine to make the part). Once the configuration of the
machines is set, the team begins producing prototype parts. During production, each

member is responsible for monitoring the process for indications of problems which

20

result from programming or settings of the machine, materials, or the cutting tools
themselves. Members can make minor adjustments to deal with problems, however, the
lead operator has final authority over changes since he or she is responsible for the team’s
performance.

A third business example of specialist decision making teams includes Chrysler’s
”platform teams”. These teams are composed of designers, engineers, suppliers,
manufacturing personnel, and marketing experts, all of whom work together on a project
at the same time (Byme, 1993). Whereas in the past, this set of functionally differentiated
individuals would have worked independently and sequentially, now they work
interdependently at all stages of the product development process. This structure allows
for problems to be noticed earlier when they are easier to rectify and has been credited for
the success of Chrysler’s minivans, the Grand Cherokee, the Viper, and the Neon (Treece,
1992; Woodruff, 1991). As opposed to the two previous examples, the recurrence of
similar structured decisions in platform teams is more spread out (i.e., the development
cycle) and the feedback regarding correctness of decisions is delayed (e.g., perhaps after
the product is launched) and perhaps more ambiguous (e.g., sales relative to the
competition).

Multidisciplinary Education Teams (METs) which make decisions regarding
placement of students into special education programs are an example of specialist
decision making teams in the context of education. Although composition varies
depending on the child and school district, METs may consist of speech and language
specialists, social workers, school counselors, school nurses, occupational or physical

therapists, adaptive physical education teachers, vocational rehabilitation counselors,

21

juvenile court authorities, physicians, general education teachers, and special education
teachers (Hardman, Drew & Egan, 1996). Members of METs undertake independent
assessments within their specialty, however, placement and program decisions are made
by the team after integrating their knowledge to ensure that the student is viewed from
multiple perspectives. The frequency of recurring team decisions corresponds to the
number of students who are initially assessed for special education programs and the
number of special education students who are re-evaluated (generally after every three
years). In the 1992-1993 school year almost ten percent of all school children in the US
between the ages of six and seventeen (4,893,865) received special education (US
Department of Education, 1994).

Specialist decision making teams can also be found in military contexts. In
command and control teams, for example, individual team members integrate unique
areas of expertise so that they can manage military operations within an assigned area.
Team roles may include surveillance officers who use special rules to identify radar blips
in assigned airspace, weapons officers who guide fighter aircraft towards blips the
surveillance officers cannot identify or identify as hostile, and asenior director who
supervises the surveillance and weapons officers and who also coordinates with outside
agencies (e.g., other command and control turits, fighter aircraft squadrons, missile
batteries, civilian agencies, etc.). Decisions made by these teams are the result of
integrating each team member’s knowledge. For instance, while the surveillance officer
may provide an initial target identification based on a set of rules (e. g., location, speed,
altitude, heading) the other members of the team gather information which also must be

considered in making final target identifications. For instance, the weapons officer

22

gathers information from the pilot of the fighter aircraft he or she controls (type of target
aircraft, national markings, flight characteristics) and the senior directors gather
information from other sources (e. g., intelligence reports). During the course of daily
operations, a command and control team can make tens, if not hundreds of similarly
structured decisions. Feedback regarding the correctness of those decisions is
unambiguous and immediate. As evidenced by highly publicized disasters, most recently
the accidental downing of a friendly helicopter over the no-fly zone in Northern Iraq,
poor decision making can result in severe consequences.

Finally, specialist decision making teams exist in medical contexts. For instance,
surgeons, pathologists, anesthesiologists and nurses integrate their knowledge and make a
series of decisions in order to safely perform a surgical procedure. Cooperative
interdependence between team members is the norm during procedures since success
depends on the joint contribution of each specialist. Recurrence of decisions depends on
the frequency of procedures and feedback regarding the correctness of decisions is often
immediate.
hm:

The purpose of this dissertation is to develop and test a theoretical model of
adaptation in specialist decision making teams. This model will draw primarily from
three literatures. As a brief overview, I first conceptualize routine and adaptation in terms
of literature which views ”groups as information processors” (e.g., Cohen & Bacdayan, ‘
1996; Hinsz, Tindale, & Vollrath, 1997). Based on this literature, I characterize the types
of information processing which occur during routine and adaptive behavior (automatic

and controlled). I then use the literature on individual differences to identify the

23

individual characteristic most likely to be important in the type of information processing
used by team members during team adaptation (general cognitive ability). Finally, I use
control theory as a framework for describing the mechanism by which team members
switch between modes of information processing. The difficulty of goals, the quality of
feedback and the salience of the change in the environment (abruptness of the
environmental change) are identified as critical elements of this mechanism. Other
individual differences (conscientiousness and openness to experience), however, are
hypothesized to play an important role in determining whether or not active or controlled
responses to problems are triggered. Figure 1 illustrates the proposed model that is
developed in the next chapter. Chapter 2 describes a laboratory study to test the
hypotheses generated from this model. Chapter 3 discusses the results of the study.

Chapter 4 summarizes the results and discusses implications to theory and practice.

 

Factors That Influence Mode of Information Processing
(Goals, Feedback, Nature of Environmental Disturbance)

 

 

 

 

 

Ability to Engage in Controlled
Information Processing Team
(Team Members’ General Adaptation

Cognitive Ability) /

Factors That Influence Responses to Perceived
Environmental Disturbances
(Team Members’ Conscientiousnes and Openness)

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 1 - An Integrative Model of Team Adaptation

24

 

Chapter 1

LITERATURE REVIEW

Information Processing as a Theoretical Foundation

Although research on human cognition is diverse in terms of what specific
processes are emphasized, this research shares the view that humans’ think and these
thoughts play a large role in their behavior (Ilgen & Klein, 1988). As Ilgen and Klein
point out, cognitive processes not only influence how objective information or stimuli
from the environment is construed and responded to, but also play a major role in
constructing the realities which are perceived as stimuli (Weick, 1979). This section
describes the role information processing plays in routine and adaptation in teams.

The most familiar view of human information processing consists of a sequence
of operations. These operations generally consist of taking information from sensors (i.e.,
vision, hearing) or long-term memory, transforming this information into forms which
can be interpreted by short term memory or working memory, and then storing this new
information back in long term memory (Lord & Maher, 1991). This is the computer
analogy model of information processing and is only one of many possible
conceptualizations. It has, however, served as a useful heuristic to organize thinking
about specific cognitive sub-processes and how these sub-processes interrelate to one
another. While much of the research concerned with human cognition focused on the
processes described above as subjects of interest in their own right, research in the
organizational sciences has often used these processes to understand organizational

phenomena. Although this work has been reviewed before (e.g., Ilgen & Klein, 1988;

25

Weick, 1979), one stream which would seem to be relevant to this dissertation is the work
on managerial problem sensing.

Problem sensing. In a conceptual paper, Keisler and Sproull (1982) focus on
problem sensing because, in their view, it is a necessary precondition for adaptive
activity, and thus, a crucial managerial activity in a dynamic environment. They define
problem sensing as the cognitive process associated with noticing and constructing
meaning about the environment. Keisler and Sproull’s aim in this paper was to outline
how the very nature of human cognition makes problem sensing difficult. Errors come
from the manner in which information (i.e., cause and effect relationships) is constructed
(e.g., augmentation, discounting, illusory correlations, illusory causation), how
information is organized (associative or schema based knowledge structures in long term
memory), and from biases resulting from human drives and needs (e.g., dissonance,
consistency, and equity theories).

Keisler and Sproull’s view of problem sensing was a departure from past
organizational research which viewed problem solving as a much more rational decision
making process (objectively scanning the environment for opportunities and threats).
However, their research was similar to past research because the focus was on individual
information processing. That is, although individual’s information processing might
focus on social objects or phenomenon (i.e., the content of cognitions), the processes of
interest remain within the individual. Recently, however, this focus has begun to change.

Qppups as infomgtipn processors. There is an emerging view of groups as
information processors. This view holds promise as a meta-theoretical foundation for

explaining many group phenomena (Hinsz et al., 1997) and is useful for understanding

26

group processes for at least three reasons. The first two arguments relate to the general
applicability of the information processing perspective to groups and teams in general
while the third argument suggests that the information processing perspective is
particularly suitable for the study of routines and adaptation in teams.

First, conceptualizing groups as information processors allows researchers to draw
upon the vast accumulation of theoretical and empirical work on the nature of human
information processing. While this literature primarily focuses on how information is
processed within individuals (Weiss, 1990), it is increasingly being extended to include .
how information is processed between individuals resulting in group-level cognitive
outputs (e.g., Bazerrnan, Mannix & Thompson, 1988; M. Cohen, 1991; M. Cohen &
Bacdayan, 1994; Hastie, 1986; Hinsz, Vollrath, Nagao, & Davis, 1988; Hinsz et al., 1997;
Hirokawa, 1990; Ickes & Gonzalez, 1994; Larson & Christensen, 1993; Levine, Resnick,
& Higgins, 1993; McGrath & Hollingshead, 1994; Sniezek, 1992; Streufert & Nogami,
1992; Tindale, 1989; Von Cranach, Ochsenbein, & Valach, 1986; Wegner, 1987). This
literature is an abrupt departure from the past because the focus is not on cognition about
social (i.e. group or team level) phenomena (e.g., F iske & Taylor, 1991), but on how
cognition is accomplished socially (e.g., Hinsz et al., 1997). In other words, there has
been increased interest in, and recognition of, groups as important information processing
systems in and of themselves.

As Larson and Christensen (1993) suggest, the analogy of the computer model of
information processing may be applied to information processing by groups just as well
as it is applied to information processing by individuals. Such processes refer to those

that “relate to the acquisition, storage, transmission, manipulation and the use of

27

information for the purpose of creating a group-level intellective product” (Larson &
Christensen, 1993). Because the primary determinant of the researcher’s focus is the level
of the outcome of which the researcher is interested in (Freeman, 1980), conceptualizing
the information processing system as a group should be especially useful when the
outcome of interest (team adaptation) is also at the group-level.

This is not to say that cognition at the individual-level and group-level is the same
thing. Indeed, this would be like saying an individual’s recollection of an item is the same
thing as someone bringing up an item in a group discussion (Larson & Christensen,
1993). While functionally similar, the nature of these processes is clearly different at the
difi’erent levels of analysis. Just as computer information processing functions as an
analogy which is helpful in developing a better understanding of individual-level
information processing, individual-level information processing serves as an analogy
which is helpful for developing a better understanding of group-level information
processing.

A second reason why an information processing perspective is useful in
understanding group-level phenomena, as mentioned before, is because the work groups
and teams do is increasingly intellectual or cognitive in nature (Galegher, Kraut, & Egido,
1990; Salas, Dickinson, Converse, & Tannenbaum, 1992; Walsh & Ungson, 1991; Weick
& Roberts, 1993). However, even when the ultimate outcomes of interest are not
cognitive in nature (e.g., the production of glass bottles), the processes leading to those
outcomes are becoming increasingly cognitive (e.g., making a series of decisions
regarding how to adjust parameters to keep the process moving). Research in small

groups and teams has begun to emphasize these types of tasks (e.g., Brehmer & Hagafors,

28

 

1986; Hinsz, 1990; Hollenbeck, Ilgen, Sego, Hedlund, Major, & Phillips, 1995; Sniezek
& Henry, 1989; Stasser, Taylor, & Hanna, 1989; Tindale, 1989; Vollrath, Sheppard,
Hinsz, & Davis, 1989).

Finally, the information processing perspective is useful for understanding
routines and adaptation because the literature suggests that the very nature of routine and
adaptation correspond with distinct modes of information processing thought to underlie
human behavior (M. Cohen, 1991; M. Cohen & Bacdayan, 1996). In the next section I
use this literature to explain how modes of information processing relate to routine and
adaptation and how this connection can be used as a basis for developing a model of
adaptation in teams.

The Basis for Routine and Adaptation

In a conceptual paper, Louis and Sutton (1991) used the phrase “switching
cognitive gears” to suggest that the effectiveness of individuals and groups depends upon
the capacity to shift between modes of information processing as dictated by necessities
of the situation. Specifically, they proposed that the effectiveness of systems that process
information depends on the ability to switch back and forth between a mode of
information processing which is relatively unconscious and automatic and a mode of
information processing which is more conscious and controlled. While there have been
many attempts to characterize these modes of information processing at both fi__rm- (e.g.,
Fiol & Lyles, 1985) and individual-levels (e.g., Schneider & Shiffrin, 1977; Shiffrin &
Schneider, 1977), this dissertation will use theory originating from the individual-level
skill acquisition and learning theory research. Among the most important reasons for this

focus is the desire to build on the work of others who have begun to use this literature as

29

a basis for developing theory aimed at understanding routine behavior in groups (e. g., M.
Cohen, 1991; M. Cohen & Bacdayan, 1996). Additionally, however, the literature on
individual-level skill acquisition tends to specify and focus on the process of moving
from one mode of information processing to the other. This emphasis, therefore, has the
potential to elucidate the factors involved in the triggering of shifts between modes of
information processing.

What follows is a brief review of the literature on skill acquisition which
recognizes distinctions between modes of information processing. From this literature, I
conceptualize routine and adaptation in terms of modes of distributed information
processing.

Fitts’ theory of motor skill development. Researchers have long been interested in
the transition from slow and halting performance of the novice to fast and smooth
performance of the expert (e.g., Bryan & Harter, 1897, 1899). In one example of this
research, Fitts (1964) developed a three phase theory of motor skill development to
explain the phenomenon. The first phase in this model is the cognitive stage in which the
actor relies on the use of instructions, facts, and other information, which have to be used
in the working, or short term memory. This phase involves the initial encoding of skill
and requires verbalization or other rehearsal in order to execute a desired behavior. The
second phase, or the associative phase, was seen as a smoothing out period and a
transition to the autonomous phase where components of the task become more automatic
and less subject to cognitive control and interference from other activities or from factors
in the environment. Thus, according to Fitts, the difference between novice and expert

motor performance boiled down to different modes of cognition which guide behavior.

30

While Fitts observations have not been the subject of a great deal of theoretical or
empirical analysis, his ideas have provided the foundation for a much more elaborated
and assessed theoretical framework.

MKS ACT“ Theory. While Fitts (1964) was concerned with skill
acquisition primarily in terms of motor skills, Anderson’s Adaptive Control of Thought
(ACT‘) theory (1982, 1987) is broader and deals with the acquisition of cognitive skills.

Corresponding to the cognitive stage in Fitts’ theory is Anderson’s declarative
ptpgp. In this stage individuals use knowledge about things, facts and rules to solve
problems. In Anderson’s second stage of skill development, knowledge compilation,
practice converts knowledge about things, facts, and rules into a procedural form “in
which it is directly applied without the intercession of other interpretive procedures”
(Anderson, 1982, p. 370). This process happens over time and corresponds roughly to
F itts’ associative stage. Finally, Anderson’s procedural st_age corresponds with F itts’
autonomous stage. Here performance is smoothed out as procedures developed earlier are
combined into larger, more domain specific procedures. In Anderson’s theory, a key
distinction is made between two discrete symbolic memory structures. Declarative
m is knowledge of things, facts, rules, etc. (content) and is obtained from
observation, reading, instruction, and so on. Procedural memog, on the other hand, is
knowledge about how to do things (process) and is gained through experience with the
task or problem.

The underlying foundation for Anderson’s theory lies with the notion that
problem solving is hierarchical (if A do B, if C results--then D, or else do E, etc.). With

experience in successfully applying declarative knowledge to such problems, efficient

31

domain-specific pgpdppm are encoded into procedural memory W).
These productions are simply sequences of cognitions or behavior which, once triggered,
are executed automatically. Using productions to control behavior instead of interpreting
knowledge in declarative form is much more efficient because reliance on declarative
knowledge would require the representation and interpretation of many individual
production steps in short-term or working memory. This would place a heavy burden on
the capacity of this system and would cause errors resulting from missing information
(Anderson, 1982). Over time, these domain-specific productions are collapsed further
into single, even more specific productions which are stored in procedural memory
(compilation). As Anderson notes, while the specialized productions will be used in the
domain in which they evolved, the original declarative memory (e.g., knowledge of the
fact or rule) and less specific productions (i.e., the domain-specific productions which
served as the basis for the more highly specific productions) remain in declarative and
procedural memory. That is, according to Anderson, the foundations for the highly
specialized productions remain in long-term memory.

According to Anderson’s theory, the primary element of an information
processing system is working memory. Working memory contains the currently activated
portion of declarative memory, declarative structures activated by procedural memory,
goals that serve as sources of activation, and finally, information which is encoded from
the outside world or from motor feedback. The information processing system operates as
elements in declarative memory or productions in procedural memory become
sufficiently activated in which case they are “poppe ” into consciousness or working

memory. Once activated, these elements then interact with goals and other information in

32

working memory to further activate elements from declarative and procedural memory.
Thus, Anderson views human cognition as the activation of nodes of memory and
therefore is somewhat of a departure from other models of cognition which follow a strict
computer analogy (proceeding from one step of information processing to the next in a
logical sequence) more literally. While Anderson’s theory was clearly intended to explain
cognition and the acquisition of skill at the individual level of analysis, concepts from his
theory have recently been extended to understand routines at the group-level of analysis
(M. Cohen, 1991, M. Cohen & Bacdayan, 1996).

While there has been research which borrows concepts from cognitive psychology
to understand behavior in groups, generally the cognitive concepts and theories are
applied loosely resulting in little testable theory (e.g., Louis & Sutton, 1991). Often,
higher level systems are anthropomorphisized (i.e., groups think) and individual level
concepts are applied assuming that these concepts have construct validity at this new
level of theory (Weick & Roberts, 1993). As Rousseau (1985) suggests, this may be a
dubious assumption. M. Cohen and Bacdayan (1996) make a significant contribution in
that their use of cognitive concepts remains consistent with the level of theory from
which these concepts originated.

Routine: Distributed procedural knowledge. As Gersick and Hackman (1990, p.
69) suggest, routines in groups exist ”when a group repeatedly exhibits a functionally
similar pattern of behavior in a given stimulus situation without explicitly selecting it
over alternative ways of behavior”. This description corresponds very closely with that
given earlier for proceduralized information processing in individuals. Just as recognized

inputs activate corresponding productions in individuals, perceived stimuli trigger

33

functionally similar patterns of behavior among groups of individuals. Just as information
processing and behavior triggered by productions proceed without much conscious
control in individuals, patterns of behavior among individuals are exhibited in given
stimulus situations without selecting them over other alternatives. Finally, just as learned
productions may be difficult to suppress, modify, or ignore in individuals, habitual
routines in groups are often triggered even when the situation would otherwise suggest
that they are inappropriate.

M. Cohen and Bacdayan (1996) argue that the similarities between proceduralized
behavior in individuals and routine behavior in groups is not coincidental. “As
individuals become skilled in their portions of a routine the actions become stored as
procedural memories and can later be triggered as substantial chunks of behavior. The
routine of a group can be viewed as the concatenation of such procedurally stored actions,
each primed by and priming the actions of others (Tulving & Schacter, 1990)” (Cohen &

Bacdayan, 1996, p. 410). Figure 2 illustrates an example.

Action 2 Action 3
Information Action 1 Action 2 _

fromthe -b.——>|BL 'lD]

Environment ‘ Action 3

Figure 2 - A Routine Among Four Actors

In this example, it is assumed that the four actors in a system already have
developed procedural knowledge through experience. Actor A encodes information from

the outside world into working memory. This information then is matched with a

34

production in actor A’s procedural memory. A corresponding production is then triggered
which controls actor A’s cognition and behavior (action 1). This behavior results in
information (i.e., an altered situation) which actor B encodes in working memory. This
information then triggers a corresponding production which controls actor B’s cognition
and behavior (action 2). The same process, in turn, also serves as a basis for actors C and
D’s behavior. There may also be non-recursive effects as well. For example, D’s
behavior, might influence cognitions and behavior of actors C and D (action 3).

As a more concrete example, assume that the actors in the system in Figure 2 are
members of a military command and control team similar to the one described earlier.
Actor A is the surveillance officer who sees a blip on the radar screen. Based on
experience using a set of decision rules, he or she then assigns the blip an identification of
“unknown” by placing a computer symbol (i.e., U) over the blip on the radar screen. The
surveillance officer also provides initial heading and speed using computer symbols (a
line from the u symbol pointing in the direction the blip is traveling with a length
indicating relative speed, U ------ >). After the surveillance officer notifies the senior
director (actor B) by voice that there is a new unknown on the radar, the senior director
notes that the “unknown” is traveling over a certain speed and in a certain direction, and
thus needs to be identified. Since the unknown is in Actor C’s (a weapons controller)
sector of responsibility, the senior director asks weapons controller C to direct a fighter
aircraft towards the unknown (on an intercept heading) for a visual identification. The
senior director also notifies Actor D (another weapons controller) to be ready to take over
the intercept with the fighter aircraft she has under her control because the unknown may

turn and head towards her sector. Using well practiced procedures, weapons controller C

35

immediately begins conducting the intercept, however, it becomes apparent to weapons
controller C that the Fighter cannot complete the intercept without exiting the C’s
assigned sector. Weapons controller C immediately notifies the senior director of this
occurrence. Weapons controller C also notifies weapons controller D that the unknown is
headed toward her sector and also passes along all known information (e.g., exact
heading, speed, altitude). Based on this information, Controller D directs the fighter
under her control toward the unknown. Once in visual range of the unknown, the pilot of
the fighter under weapons controller D’s direction reports an identification (e.g., Hughes
helicopter, tail number HGlO72) to weapons controller D. Weapons controller D then
reports this information to the surveillance officer, who updates the unknown computer
symbol to a helicopter symbol (i.e., H ------ >).

The identification of an unknown in this example is a team-level outcome because
between team variation in the effectiveness of identifying unknowns is dependent on the
nature of the team (e.g., their pool of skills, the quality of their interaction, etc.), no
individual member could produce the outcome by him or herself, and outcomes which
occur all at once, depend on the entire system of activity. The view of routine as a system
of reciprocally triggering behavior is similar to that described by other researchers (e.g.,
Dewey, 1922; Nelson & Winter, 1982; Weick, 1979), however, M. Cohen and ‘
Bacdayan’s (1996) conceptualization suggests a specific psychological mechanism that
underlies the phenomenon. This is a significant contribution because it allows

understanding and development of propositions regarding related phenomenon

(adaptation).

36

Adaptation: Distributed declarative knowledge. Just as routine corresponds to

team members’ information processing and behavior that is relatively automatic,
adaptation corresponds to team members’ information processing and behavior that is
more controlled. A situation which does not correspond to a known set of responses
needs to spur an active search for and experimentation with new responses. Team
adaptation requires that members use what they have learned in the past as a basis for
developing an appropriate way of dealing with the new situation. Unsatisfactory partial
solutions are abandoned while those which are successful are implemented. Over time,
and assuming the environmental disturbance which necessitates the adaptation stabilizes,
successful partial solutions become integrated and team-level activity becomes smooth.
Using the model suggested by M. Cohen and Bacdayan, however, one can go
further than simply characterizing the mode of information processing and behavior
during adaptation. That is, by integrating the work of Anderson’s theory together with
that of M. Cohen and Bacdayan, team adaptation can be understood as requiring team
members’ use of declarative knowledge and lower level productions in information
processing. That is, if information provided by the situation as encoded by team members
does not result in the triggering of productions, lower level productions and/or an
interpretation of rules specified in declarative knowledge will be used to specify
appropriate action. In fact, because there is interdependence between team members, it is
likely that even those who are not directly impacted by an environmental disturbance will.
come to rely on their declarative knowledge and lower order productions to specify
appropriate role behavior. Members directly impacted by disturbances may have to

respond or behave in ways that are unfamiliar to others. Since those not originally

37

exposed to the environmental disturbance would then perceive a situation which would
not correspond to stored productions, they would have to rely on their declarative
knowledge and lower level productions to specify their role behavior. As an example, let
me return to the command and control scenario.

Command and control teams may have to identify many unknowns during the
course of a work shift and the pattern of activity among team members becomes routine.
However, many types of situations can occur which would make the pattern of activity
problematic. Perhaps the simplest example is a communications failure (either voice or
data) between team members. For instance, if for some reason weapons controller C was
unable to communicate with weapons controller D, transferring intercepts between the
two would be more difficult because last known heading, altitude, and speed of the
unknown would be unknown to weapons controller D (weapons controller D would not
know where to send her fighter aircraft). In this situation, weapons controller C might
recall from past experience that the senior director is too busy communicating with
outside agencies to act as an intermediate communications link, so weapons controller C
contacts the surveillance officer and directly asks him to pass along information regarding
the unknown to weapons controller D. Noting that the surveillance officer is acting as a
communications link, the senior director recalls that in similar situations in the past the
surveillance officer became overloaded and could not effectively keep up with the task of
assigning computer symbology to radar blips. The senior director considers the situation
briefly and then decides to monitor the radar display more closely so that the surveillance
oflicer will no longer have to pass verbal messages when new unknowns enter the team’s

sector. The result is a new pattern of interaction among team members. While some of

38

this new pattern is dictated by the disturbance (i.e., the communications breakdown),
team members used what they knew from past experience to develop a new system
behavior that could cope with the new situation.

Among the most important implications of this conceptualization of adaptation is
that while individuals’ reliance on interpretations of declarative knowledge is flexible in
that it can be applied in novel situations, it has “serious costs in terms of time and
working memory space” (Anderson, 1982, p. 311). During adaptation, team member
behavior is governed by many small productions and knowledge from declarative
memory. Resulting team behavior is much slower and halting than when larger
specialized productions are triggered. Returning to and extending the command and
control example once again should make this point clear. Until the members of this
command and control team proceduralize their new role requirements, the members’
actions will demand their active attention. The senior director will have to remember to
monitor the radar display more closely and learn how to smoothly integrate this task with
other elements of his task. Weapons controller C will have to remember not to bother
trying to communicate with weapons controller D, but that it is necessary to communicate
with the surveillance officer. The surveillance officer has to work out a procedure for how
to receive and re-send information in a timely manner without altering it and without
interfering with other important responsibilities. Finally, since the target information
must be sent through an intermediary, weapons controller D will have to learn how to
interpret information which is more outdated (i.e., the unknown may have changed its

heading, altitude or speed) than if the information were passed to her directly.

39

As the conceptualization outlined above suggests, one primary determinant of
team adaptation may be the capacity of team members to actively process information.
Teams with members who do not possess adequate capacity may not be able to develop
systems which can effectively cope with changing situations. This implies that team
members’ working memory is a critical resource for teams because the declarative
knowledge and lower order productions which are used during active information
processing must be represented there. In the next section, I discuss the notion of working
memory and extend this concept to the team level.

Cogpitive Resources

Traditional models of human information processing posit that working memory
consists of a space or area where cognitive work takes place. Other models of human
information processing posit that working memory consists of that portion of long term
memory which is currently activated. Generally speaking, both these conceptualizations
assume that working memory is unitary in that all cognitive work takes place there and
that once the limit of working memory capacity is reached, processing of other
information (e.g., problem solving) will be problematic because some pieces of
information will become lost. These models are referred to as “unitary fixed capacity”
models of working memory.

Researchers in cognitive psychology began to criticize unitary fixed capacity
models in the 1970’s and 1980’s (Lord & Maher, 1991). These concerns first began
surfacing as researchers in the dual task paradigm found that individuals could perform
two tasks (e.g., responding to a beep while simultaneously attending to a visual display)

without mutual interference. These findings suggested that perhaps human memory was

40

not of fixed capacity. Subsequent research found that while individuals could perform
two tasks concurrently without much disruption when the tasks were widely different
(e.g., verbal and visual), individuals’ performance was severely hampered when two tasks
were highly similar (e.g., both visual) (Baddely & Hitch, 1974; Pashler, 1991, 1994). As
a result of these and other findings, researchers in cognitive psychology began to develop
models of human information processing systems with several relatively independent
working memory subsystems (e. g., a visuospatial scratch pad and an articulatory loop),
each dedicated to different cognitive sub-functions (Lord & Maher, 1991).

While this “multiprocessor” view of memory may be better able to explain human
cognition under certain circumstances (i.e., processing information from discrete sensory
subsystems) than traditional unitary fixed capacity views, it is still useful to conceptualize
working memory as having relatively fixed capacity for purposes of this dissertation. The
primary reason for this statement is that one of the most “fundamental and stable”
properties of human cognition is that the memog span of individuals is limited in
capacity and fairly stable over a wide range of material (Brown, 1958; Peterson &
Peterson, 1959) and that “[t]his limit places severe constraints on people’s ability to
process information and solve problems (G. Miller, 1956; Newell & Simon, 1972)”
(Chase & Ericsson, 1981, p. 141). Since this dissertation is focusing on adaptation, which
is a problem solving activity (March & Simon, 1992), the limit in terms of capacity to
process problem relevant information is the issue, not whether there is a unitary
subsystem for processing unrelated non-problem related activities. Furthermore, the
multiprocessor view implicitly assumes that each separate processing subsystem has a

fixed capacity and is non-compensatory (e. g., if capacity for processing information

41

visually is consumed one cannot process the information auditorially). Thus, since the
nature of a task determines which specific processing subsystem(s) is (are) required, it is
sufficient to say memory capacity is limited when one is considering information
processing in a specific task (e.g., problem solving). Indeed, over the last decade a
number of organizational behavior and industrial and organizational psychology
researchers have successfully applied this model in explaining individuals’ behavior (e. g.,
Barber, Hollenbeck, Tower & Phillips, 1994; Eyring, Johnson, & Francis, 1993; Kanfer
& Ackerman, 1989; Kanfer, Ackerman, Murtha, Dugdale, & Nelson, 1994).

In summary, although there is some debate as to the exact structure of working
memory, it is reasonable to assume that in terms of problem solving, humans possess “at
any moment a finite amount of processing facilities” (Navon & Goher, 1979). This finite
amount of working memory generally relates positively to an individual’s ability to
process information, solve problems (G. Miller, 1956; Newell & Simon, 1972), and
perform in tasks (Navon & Goher, 1979). While terminology for this capacity differs
somewhat across researchers (Kalmeman, 1973; Moray, 1967; Norman & Bobrow, 1975;
Shiffiin, 1976), I use the term cogm'tive resources. The choice of this term is somewhat
arbitrary, however, it is probably the most familiar to those interested in organizational
behavior and industrial and organizational psychology.

Individual differences in cogpitive resources. In the organizational behavior and
industrial and organizational psychology literatures, individual differences in cognitive
resources have been described in terms of general intellectual ability, general cognitive
ability or (g) (Ackerman, 1986, 1987; Kanfer & Ackerman, 1989). Support for the

existence of g comes from the fact that a single factor exists and underlies performance

42

on almost all tests which measure cognitive abilities (Hunter, 1986; Jensen, 1986; Kass,
Mitchell, Grafton & Wing, 1983; Ree & Earles, 1991; Welsh, Watson & Ree, 1990).
Support for the link between g and cognitive resources capacity comes from a long
history of research on two fronts.

First, psychologists studying the nature of intelligence report large positive
relationships (reports generally range from about r=.60 to r=.80) between tests of memory
span and scores on tests which measure g (e.g., Bachelder & Denny, 1977a, 1977b;
Jensen, 1970). Bachelder and Denny (1977a, 1977b) proposed that this span is equivalent
to the ability to respond appropriately when several cues (some of which are irrelevant)
are simultaneously presented. This explanation is consistent with the long held view that
individuals with larger spans or capacities can attend to more stimuli resulting in more
complete and coherent units of perception than individuals with smaller memory spans or
capacities (Lemming, 1922).

Second, industrial and organizational psychologists report positive relationships
between tasks which require active information processing and scores on tests that
measure g (e.g., Hartigan & Wigdor, 1989; Hunter & Hunter, 1984). For instance, Ree
and his colleagues (e.g., Ree & Earles, 1991) have conducted numerous studies that
suggest that individuals’ g is critical in training success. Further evidence comes from
reviews which report higher relationships between g and performance in complex jobs
than between g and performance in simple jobs (e.g., Hartigan & Wigdor, 1989; Hunter
and Hunter, 1984). Finally the literature on skill acquisition is consistent in showing that
g is particularly important in early stages of task performance where a controlled mode of

information processing is necessary (e.g., Ackerman, 1987).

43

In summary, because individuals with higher levels of cognitive ability are more
capable of attending to and making sense of stimuli in novel situations (i.e., engaging in
controlled information processing), it is certainly reasonable to expect that team
members’ cognitive ability will be positively associated with levels of declarative
knowledge gained through instruction and experience (i.e., learning). That is, team
members with high levels of cognitive ability should have more knowledge about
important aspects of the task (e.g., facts and rules) and the situation than team members
with low levels of cognitive ability.

Team differences in cogpitive resources. While the literature cited above focuses
on between-person differences in the capacity to process information, the underlying
principles are not anchored exclusively to systems at this level. Consistent with general
systems theory, Norman and Bowbrow (1975) suggested that information processing
principles transcend levels of analysis: “[t]he processing resources for any system are
limited, and when several processes compete for the same resources, eventually there will
be a deterioration of performance” (p. 44).

Systems such as groups and teams engage in a number of sub-tasks or processes
which consume members’ information processing resources. Groups and teams must not
only (1) develop and maintain themselves but must also (2) do their required work, (3)
modify themselves as a consequence of doing work (adapt), and finally, (4) manage
relations with the organizational and environmental context (Argote & McGrath, 1993).
Given that team members’ cognitive resources must be allocated among these different
sub-tasks (since each team member should be more or less involved in each of the

subtasks), it is reasonable to expect that the larger the pool of cognitive resources on a

44

team, the more likely the team should have adequate cognitive resources to allocate
among these different sub-tasks. That is, teams with members who have high levels of
cognitive resources should be more effective at these subtasks than teams with members
who have low levels of cognitive resources. In this regard, the pool of the team members’
cognitive resources, henceforth team cognitive resources, can be conceptualized as a
characteristic of the team, and therefore can be considered a team-level construct. Since
adaptation is among those sub-tasks in which team members must engage a mode of
information processing which relies heavily on the their capacity in terms of cognitive
resources (i.e., developing and then using declarative knowledge and lower level
productions), it is reasonable to expect a positive relationship between team cognitive
resources and team adaptation.

The notion of team cognitive resources does not suggest that aggregated scores on
measures designed to tap individual’s intelligence can be used to characterize a team as
being intelligent. This would necessitate consideration of group dynamics and the
development of measures which reliably capture these dynamics. Instead, I suggest that
the pool of team members’ cognitive resources can be thought of as a resource which
influences team processes and outcomes and that aggregation is necessary simply to
account for this resource.

I note here that the appropriateness of the specific means of aggregation depends
upon joint consideration of the attribute in question as well as the nature of the task
(LePine, Hollenbeck, Ilgen, & Hedlund, 1997; Steiner, 1972). That is, the task not only
determines which attributes are important for effective team processes and outcomes, but

also how each attribute needs to be distributed among team members. If a task requires

45

every member to possess adequate resources in terms on an attribute, the member who is
least adequate in terms of that attribute will determine the level of task performance
(conjunctive task). Task performance will depend on the member who is most adequate in
terms of a critical resource as long as the task is structured such that one person with
adequate resources could perform effectively alone (disjunctive task). Finally, if each
group member contributes to task performance in proportion to their resource level, team
performance will be a function of the sum or average of team members’ resources
(conjunctive task).

Because there are very few studies on the nature of team adaptation, it is difficult
to specify a precise means of aggregating individual attribute based predictors such a
cognitive ability (and later conscientiousness and openness to experience). For instance,
one could certainly argue that each member of a specialist decision making team might
have non-substitutable expertise (due to high levels of specialization), and therefore, team
adaptation might be conjunctive in terms of important team member resources that leads
to that expertise. On the other hand, perhaps it is sufficient to have only one individual
who is capable of devising a means of coping with unforeseen environmental
contingencies. Thus, the task might be disjunctive in terms of critical team member
resources. However, I would argue that the task a specialist decision making team faces is
neither purely conjunctive nor disjunctive. I suggest that because there are good reasons
to expect that adaptation has both conjunctive and disjunctive elements, perhaps the best
way to represent the team in terms of member characteristics is with the additive model.

This model captures scores for the highest and lowest members and has also been used in

46

the majority of studies on team composition. Thus, the additive model will be used as the
primary means of aggregation in this study.

However, given our current knowledge of adaptation, it is premature to rule out
the alternative operationalizations. Therefore, while the primary method of aggregation
will be additive (i.e., a mean), the viability of alternative means of aggregating individual
characteristics in predicting team adaptation will be examined as a research question.

Thus, setting the issue of means of aggregation aside for the moment, the first

hypothesis is:

Hmthesis 1 (H1): Team cognitive resources and team adaptation will be
positively related such that teams with high levels of cognitive resources will have

a higher likelihood of adapting than teams with low levels of cognitive resources.

While the information processing framework proposed earlier leads one to expect
a positive relationship between team cognitive resources and team adaptation, other
literature might suggest otherwise. As noted previously, routines are often transferred to
inappropriate situations because groups may fail to recognize changes to a familiar
context or that they are in a new context entirely. This problem, however, may be
particularly problematic for successful groups because they develop high levels of
confidence in routines which have worked fine in the past (Gersick & Hackman, 1990).
To the extent that higher levels of cognitive resources in a group leads to higher levels of
performance (LePine, Hollenbeck, Ilgen & Hedlund, 1997; O’Brien & Owens, 1969;

Tziner & Eden, 1985), routines that are perceived to have led to that success are

47

reinforced. In fact, a negative relationship is somewhat consistent with Anderson’s (1982,
1987) own suggestion that as productions are successfully applied, they gain strength. As
individuals acquire confidence in their behavioral patterns, they tend to lower cognitive
activity related to information search and receptivity (Weiss & Ilgen, 1985), thus
lowering the likelihood of triggering controlled information processing (which is
necessary for adaptation in teams). Following this logic, therefore, teams with high levels
of team cognitive resources may be less likely to adapt than teams with low levels of
team cognitive resources because contextual changes which necessitate adaptation are
less likely to overcome the strength of the routine and trigger a more controlled mode of
information processing.

Granted, this line of reasoning does not seem consistent with the common sense
notion that teams with more intelligent members should be more vigilant to changes in
the environment than teams with less intelligent members. However, I could find no
conceptual or empirical support for such a relationship in the literature. Instead, the
information processing perspective suggests that the relationship between a team’s pool
of cognitive resources and adaptation may be moderated by factors that influence the
extent to which active or controlled information processing in team members is triggered.
Thus, the critical question becomes what factors enhance the likelihood of triggering
declarative knowledge based processing so that team cognitive resources can be
marshaled for adaptation? The next section of this dissertation proposal presents a

framework which should be useful in identifying these important factors.

48

Triggering Declarative Knowledge Based Information Processing

Most models of human information processing generally describe how this system
is controlled. These models vary widely in terms of underlying assumptions, but the
primary distinction between them is that they tend to focus on either content (identifying
specific factors which control attention) or process (how humans switches from one mode
of information processing to another).

Content of information processing control. Based on a review and integration of
several literatures (primarily visual and auditory perception and speeded performance)
Kahneman (1973, p. 42) concluded that the allocation of attention (the focus of
information processing) is determined by “momentary task intentions of voluntary
attention” (e.g., preference for attending to something as opposed to something else) and
also by a “more enduring disposition which controls involuntary attention” (e.g., novel
stimuli automatically attract attention). Thus, from this perspective, attention
(information processing) towards a stimulus is viewed as being either on or off, and the
control of attention (i.e., where information processing is directed) is determined by
momentary intentions (allocation policies) and the nature of available information
(novelty or salience).

These two determinants of attention are not inconsistent with those suggested by
most generic information processing models that do not distinguish between the modes of
information processing discussed earlier. For instance, Hinsz et a1. (1997) reviewed the
literature on information processing in groups and categorized the various factors that
have been found to direct attention and subsequent information processing. These authors

suggested that these factors either (a) suggest which types of information are relevant for

49

processing (e. g., goals, roles, norms, etc.) or (b) relate to the nature of available
information itself (e.g., salience, novelty, etc.).

WM As opposed to the models described
above which focus on the content of information processing control, other researchers
have focused on the process of control between different modes of information
processing. Schneider and Shiffrin’s dual-mode theory of cognition is a good example
(Schneider & Shiffrin, 1977, Shiffiin & Schneider, 1977) and is quite similar to the mode
of control specified in Anderson’s ACT“ theory of skill acquisition.

Schneider and Shiffrin proposed that there are two modes of information
processing. Briefly, the controlled mode corresponds roughly to that in Fitts’ (1964)
cognitive stage and Anderson’s (1982) declarative stage, while the automatic mode
corresponds roughly to that in F itts’ autonomous stage and Anderson’s procedural stage.
In this model, control between modes is determined by the familiarity of stimuli. Through
practice, stimulus-responses associations are learned. In the presence of these stimuli
responses are triggered automatically. Controlled processing, therefore, can occur only
when stimuli do not activate these automatic responses.

Anderson’s ACT" theory uses a similar conceptualization of control between
modes of information processing. Essentially, declarative knowledge and lower order
productions are assumed to be used only in situations where the compiled productions
cannot be applied (Anderson, 1987, p. 197). That is, unfamiliar situations will not trigger
productions, and therefore, people will have to rely on declarative knowledge and low

level productions in order to perform in that domain.

50

m While the literatures cited above clearly suggests content and processes
associated with determining a system’s mode of information processing, there is no
clearly specified theory that integrates these factors. Clearly, it would be preferable to use
an articulated theory to ground understanding and guide research and practice rather than
to simply pick and choose concepts from these diverse literatures. Theory guided research
should allow for the development of a systematic view of the phenomenon under
question. Such an approach should reduce wasted effort and confusion caused by
researchers who have studied similar phenomena and have come to similar conclusions
but use different jargon. Using a theory to guide research should be beneficial because it
increases the probability that other researchers will be able to build upon the present
work.

In the following section, I use control theory as a basis for developing a model
that explicates factors involved in the switching between modes of information
processing. As others have suggested, control theory is useful as a meta-theoretical or
integrative framework (e.g., Hyland, 1988; Klein, 1989) for developing understanding of
system behavior at many different levels of analysis (e.g., von Bertalanffy, 1968). The
use of control theory, therefore, may allow for a parsimonious synthesis of the many
factors that could be expected to influence the mode of information processing in teams.
Contrpl Theog

Although originally a branch of engineering, psychologists have used control
theory in attempts to understand human behavior for almost a half-century (e.g., Ashby,
1952; Weiner, 1948). Because of this long history, it is not surprising that control theory

comes in many difi‘erent variations (e.g., Campion & Lord, 1982; Carver & Scheier,

51

1981; Hollenbeck, 1989; Hyland,.l988; Klein, 1989; Powers, 1973; Taylor, Fisher &
Ilgen, 1984). Despite these variations, however, the underlying foundation of control
theory is the negative feedback loop.

In its most basic form, control theory proposes that behavior is regulated by a
negative feedback loop. The term negative is used in reference to this loop because the
control system is intended to negate or reduce perceived deviations from a standard, and
not because the discrepancy is negative in a normative sense of the word (i.e., bad)
(Carver & Scheier, 1982). Inputs that are sensed from the environment (which itself may
have been influenced by the system’s behavior as well as its own influence separately
from the system’s behavior) are compared to a referent standard or goal. If this
comparison process detects a discrepancy or error, it signals the system to take action in
order to reduce the discrepancy or error. The system acts on this signal and then monitors
the impact of this action from information gathered fiom the environment. If that
discrepancy remains, the system sends another error signal, the system then takes action,
and so on. In most conceptualizations of control theory, this sensing--comparing--acting
process continues until the discrepancy or error is eliminated or reduced significantly.

Applicatipn to a wide variety of systems. As von Bertalanffy (1968) points out,
negative feedback schemes exist in electro-mechanical as well as in organic systems. The
most popular example of a mechanical system is the thermostat controlling the
temperature of a room (e.g., Carver & Scheier, 1981; Klein, 1989). A desired temperature
is set and a sensor monitors the room’s ambient temperature. If the room’s temperature
deviates from that which is set, a furnace or air conditioner will activate and operate until

the desired temperature is reached. An organic example includes thermoregulation in

52

warm blooded animals (Carver & Scheier, 1981; von Bertalanffy, 1968). If the
temperature of blood in warm blooded animals drops below a certain temperature, the
brain activates heat producing mechanisms in the body (e.g., shivering) until normal
temperature is reached.

Of course, in the area of applied psychology, control theory has been used to
explain observable human behavior with an emphasis on motivational processes (e. g.,
Campion & Lord, 1982; Hollenbeck, 1989; Hollenbeck & Williams, 1987; Keman &
Lord, 1990; Taylor et al., 1984). As a generic example, if a salesperson senses that there
is a discrepancy between his or her sales performance and quota, he or she may engage in
some sort of behavior (e.g., work harder to achieve quota, argue that the quota is too
high) or cognition (e.g., reject the feedback, change the goal) in order to reduce the
perceived discrepancy.

Most recently, control theory has been used in the applied psychological and
organizational literatures as a meta-theory that integrates other theories of motivation
(e.g., Hyland, 1988; Klein, 1989). Hyland (1988) used control theory to compare
elements fi'om a number of basic motivational theories (i.e., Atkinson, 1957; Deci &
Ryan, 1985; Locke, 1981; Murray, 1938; Weiner, 1980). Klein (1989) took a similar, but
more applied approach by integrating work done by researchers primarily in the domains
of organizational behavior and industrial and organizational psychology (e.g.,
Hollenbeck, 1989; Hollenbeck & Brief, 1988; Lord & Hanges, 1987; Taylor et al., 1984).
Both Hyland and Klein reached the conclusion that most theories of motivation focus on

different aspects of the underlying control process.

53

Although not the focus of current research in the organizational or applied
psychological literatures, it is also reasonable to assume that control systems operate in
higher-order systems such as groups, teams and even organizations. Indeed, Cyert &
March’s (1992) view of the firm (as preferring some states to others and using decision
rules to bring itself closer to those states when it senses discrepancies between those
preferred states and the situation), closely parallels most conceptualizations of control
theory focusing on behavior in lower order organic and non-organic systems.
Furthermore, since control theory specifies the mechanism which links behavior at
multiple hierarchical levels, it is highly useful as a framework for studying behavior in
group and organizational systems.

Hierarchical structure of control loops. A long line of research recognizes that
purposes, standards, or goals are hierarchical-~that is, purposes, standards, or goals are
often means of attaining other higher order purposes, standards, or goals (Barnard, 1938;
March & Simon, 1992; Murray, 1938). Barnard (1938, p. 232) for instance, recognized
this when he stated that one function of executives is to redefine and modify purposes
“level after level”. Control theory acknowledges this structure by maintaining that the
means of reducing discrepancies in higher order feedback loops become standards in
lower-level loops (Carver & Scheier, 1982; Hyland, 1988; Klein, 1989; Powers, 1973).

I should note that the majority of research on control theory focuses on negative
discrepancies. Indeed, this focus can probably be blamed for the fallacious stereotype that
control systems either proceed toward an endpoint or function simply to maintain a
steady state (Carver & Scheier, 1982). In fact, however, the specification of hierarchical

control loops allows for explanation of reactions to positive discrepancies as well as

54

negative discrepancies. Specifically, responses to achievement or over-achievement of
lower order goals depend on the nature of the higher order goal from which the lower
order goal originated. For instance, if a salesperson has a higher order goal of “striving
for excellence”, he or she may set personal sales goals which are higher than the assigned
quota after exceeding this quota. The issue of positive discrepancies, however, is not
highly relevant to this dissertation because team adaptation, as defined earlier, is
conceptualized as a response to a negative discrepancy (i.e., the perceived lack of fit
between the team’s structure and the situation). While it is worthwhile to study the
problem of how to make successful teams continuously strive for excellence through
positive discrepancy creation, this is a different issue--one that will not be addressed in
this dissertation.

One important aspect of control theory is that discrepancy reduction does not
require active or controlled information processing at all levels of the control-loop
hierarchy. Indeed, it is widely believed that individuals have a set of well learned
“programs” (lower order negative feedback loops) that they automatically invoke when
faced with familiar types of discrepancies. As long as perceived discrepancies are such
that they evoke these programs, controlled information processing is not necessary. For
instance, a salesperson may automatically nod with agreement and say “uh-huh” when
interacting face to face with important potential customers.

However, when discrepancies are not associated with a learned response, or when
learned responses no longer decrease discrepancies, active attention focuses on devising
means of addressing the discrepancy. For instance, if the potential customer rolls her eyes

every time the salesperson says “uh-huh”, the salesperson may notice and sense the need

55

to devote some active attention towards devising an alternative strategy (assuming saying
“uh-huh is the only strategy the salesperson knows). Whatever the alternative strategy
entails, he or she will then attend to information provided by feedback to assess the extent
to which the strategy is succeeding. Until this new strategy becomes proceduralized (as
happens when strategies are repeatedly successful), controlled information processing
will be necessary to self-regulate in this specific control loop. This conceptualization is
very similar to Wood and Locke’s (1990) notion that individuals will attempt to develop
new task specific plans (through study, research creative problem solving, trial and error,
etc.) when stored universal plans (directing increased effort or persistence towards the
task) or stored task specific plans (task specific skills) are not applicable to a task/goal
situation.

Inappropriate routine as control system failure. As stated previously, routine
behavior in groups can be characterized by mindlessly repeating patterns of behavior in a
given stimulus situation without consciously considering alternatives (Gersick &
Hackman, 1990). Given this description, therefore, it would seem that the inappropriate
engagement of routine behavior (i.e., failure to adapt) can be conceptualized in terms of
failure in a negative feedback system to detect situations that should not be associated
with programs, scripts or proceduralized responses. Assuming this to be a reasonable
assumption, the question now becomes, what aspects of a control system would facilitate
such a failure? Alternatively, what factors evoke or activate controlled information
processing? In Wood and Locke’s (1990) terms, what triggers the development of new
specific task plans when stored universal plans or stored task specific plans are no longer

appropriate?

56

A variety of researchers, some squarely in the domain of control theory, have
speculated on the determinants of controlled information processing (e. g., Carver and
Scheier, 1981; Lord & Hanges, 1987; Taylor et al., 1984). While these authors differ
somewhat in their primary focus, in the aggregate they suggest at least three factors are
involved; (a) the nature of the standard or goal, (b) the nature of the feedback which is
used in the comparison with the standard or goal, and finally (c) the nature of information
provided by the problem which makes adaptation necessary in the first place. The next
three sections of this proposal will briefly review these factors.

Ms

As various researchers have pointed out, reference standards include goals (Carver
& Scheier, 1981; Klein, 1989; Powers, 1973) that may relate to end states, rates of
progress towards an end state, a non-terminal activity (e. g., self-improvement), or
emotion (Hyland, 1988). Although distinct fiom one another, each serves as the criterion
against which perceptual inputs are compared, and as a group play an important role in
the performance effectiveness of individuals (Locke, Shaw, Saari, & Latham, 1981).

In essence, goals direct attention. “[P]eople are bombarded with information of
every sort, but they act only in response to a small segment of it, namely that segment
which they decide is relevant to their own life interests and goals” (Latharn & Locke,
1991, p. 225). In information processing terms, specific and difficult goals serve as
objectives that influence the focus of information processing. In Kahneman’s (1973)
terms, specific and difficult goals influence momentary intentions which, in turn,

determine how cognitive resources are allocated.

57

:tpn
“racer
I

knit

~

Wart

AI
‘8

a"

hen

Goals have a major influence on the mode of information processing because
discrepancies may necessitate the development of strategies in order to narrow the gap
between the goal and present standing (Campbell & Gingrich, 1986; Earley, Northcraft,
Lee & Litcuhy, 1990; Earley & Perry, 1987; Early, Wojnaroski, & Prest, 1987; Latham &
Saari, 1982; Locke, Frederick, Buckner, & Bobko, 1984; Mitchell & Silver, 1990; Wood
& Locke, 1990). That is, when increasing effort or concentration (stored universal plans)
in well learned activity (stored task specific plans) does not result in error/discrepancy
reduction, attention shifts towards learning how to achieve goals in light of the situation
for which there are no scripted responses (Kluger & DeNisi, 1996). This type of learning
consists of controlled search and experimentation and thus draws on declarative
knowledge and lower order productions (new task specific plans).

A great deal of research supports the conclusion that given adequate ability and
commitment, specific and difficult goals lead to higher individual-level performance than
easy or vague goals. This conclusion appears to be especially valid for performance in
well learned (Kanfer & Ackerman, 1989) and less complex (Wood & Locke, 1990) tasks
than for performance during skill acquisition and in highly complex tasks. Goal
specificity is important because vague goals are compatible with too many outcomes and
thus lead to ambiguity concerning what constitutes effective performance (Latham, &
Locke, 1991). As Klein (1989, p. 154) suggests, “[v]ague goals make poor referent
standards because there are many situations in which no discrepancy would be indicated
and, therefore, there would be no need for corrective action (Campion & Lord, 1982)”.

Goal difficulty is important because individuals adjust their effort to the difficulty

of the task (Latham & Locke, 1991 ). In control theory terms, the more difficult the goal,

58

the more effort is required to avoid errors (Klein, 1989; Lord & Hanges, 1987). In
addition, the more difficult the goal, the more likely discrepancies will be large (Campion
& Lord, 1982). This should not only make discrepancies more salient, and thus more
likely to be perceived, but should also make it less likely that discrepancies will be
dismissed as insignificant.

Most research on goal setting is concerned with individual-level relationships
(Locke & Latham, 1990), however, there is also research which has examined the
relationship between goals and group-level performance (e.g., O’Leary-Kelly,
Martocchio, & F rink, 1994; Pritchard, Jones, Roth, Stuebing & Ekeberg, 1988; Saavdera,
Barley & Van Dyne, 1993; Weingart, 1992; Weingart & Weldon, 1991; Weldon, Jehn &
Pradhan, 1991; Weldon & Weingart, 1988). Conceptually, group-level research on goals
is more complex than individual-level research because, as Zander (1980) notes, there are
different levels of goals within a group (e. g., the goal for the group, each member’s own
goal as a group member). However, this apparent complexity is reduced if one accepts the
hierarchical structure specified by control theory. Specific sub-goals which regulate
individual team members’ behavior (e. g., being responsive to requests from team mates,
making useful recommendations to leaders, etc.) result from their acceptance of a team-
level goal as one of their higher order goals (e. g., producing a product safely). This effect
was confirmed in a laboratory study using an idea generation task (Weingart & Weldon,
1991). In this study, individual members were more likely to set individual-level goals
when they were assigned group goals. To an individual team member, a team-level goal
is just like any other higher order goal which regulates aspects of their behavior—the only

difference being that this goal is more or less shared with the other members of the team.

59

This conceptualization is consistent with Zander (1971) who suggests that self-set
individual goals mediate the relationship between group-level goals and group
performance effects. Zander suggests that such effects should be especially true when
team members perceive that their “personal activity is more relevant to the work of the
group” (p. 198). Thus, this type of individual-level goal setting should occur in specialist
decision making teams because these teams are characterized as having cooperative
interdependence. Since, team members’ roles are integrated into a system of activity
which results in team-level outcomes, improvement in individual-level role performance
should translate into improvements in team-level processes and outcomes (Weingart &
Weldon, 1991; Weldon et al., 1991). For instance, in a laboratory experiment using a
production task, Weldon, et al., (1991) found that when groups were assigned a difficult
goal, individual members developed more efficient strategies for their individual role than
members in teams assigned easier goals. These authors also found that this increased
efficiency resulted in improved group performance. Figure 3 illustrates this
conceptualization of the effect of group-level goals.

At the top of this figure is the team goal. This goal may be explicit (e.g., management
assigning the team a specific and difficult performance goal) or implicit (e.g., assuming a
goal based on the team’s purpose), however, it is this goal that members more or less
share (depicted by the boxes labeled Member X Team Goal). From these goals and a
basic understanding of their task, team members establish sub-loops related to their roles

as they understand them (e.g., acquiring and exchanging information in a timely manner).

60

/

 

 

 

 

Memher A
Team Goal

 

 

 

 

 

 

Comparator

 

 

 

Member A
Subgoals

 

 

Team Goal

 

or Purpose

 

 

 

Member B

 

 

 

 

Comparator

 

 

 

 

 

The primary implication of this conceptualization is that the nature of team-level

 

 

 

Team Goal

 

 

 

 

 

Member B
Subgoals

 

\

 

 

 

 

Memher C
Team Goal

 

 

 

 

 

Comparator

 

 

 

 

 

 

 

 

 

Member C
Subgoals

 

 

 

 

 

 

 

 

Environment

 

 

 

Figure 3 - Hierarchical Goal Structure In Teams

61

goals ultimately influence the nature of sub-goals that guide team members’ behavior and
information processing. That is, difficult team-level goals should translate into sub-goals
that are difficult and easy team-level goals should translate into sub-goals that are easy.
Thus, when teams have difficult goals, environmental disturbances that disrupt the
effectiveness of team members’ role performance should be more likely to result in
perceived discrepancies in the comparator process related to the sub-loop than when
teams have easy goals. Again, given a specific performance level, difficult goals
necessarily lead to a higher probability of negative performance-goal discrepancies than

easy goals. A negative discrepancy in the comparator process, in turn, should trigger the

prob
L11"

5“

men:

search for a cause and solution to reduce the discrepancy. Individuals in teams with
difficult goals, therefore, should be more likely to recognize an environmental
disturbance as a problem that must be dealt with than individuals in teams with easy
goals. When no scripted responses or programs are available to deal with such problems,
the type of controlled information processing necessary for adaptation should be
triggered. Accordingly, I expect that team adaptation will be higher in teams that have
difficult goals than in teams that have easy goals.

I should note two caveats to this discussion. First, it is unlikely that adaptation
will increase in teams where goals are too difficult. This expectation is consistent with the
literature that suggests that if individuals’ expectancies regarding their ability to
overcome discrepancies is too low, they may withdraw from the situation instead of
engaging in discrepancy reduction attempts (Carver & Scheier, 1982). In most goal
setting interventions, however, the intent is to set goals that are difficult but achievable,
and thus, in most real world situations, severe expectancy reduction should not be a
problem. This follows fiom the notion that since goal setting interventions are normally
intended to be motivational, it is unlikely that managers or policy makers will
intentionally set goals that are de-motivating. Of course, events not under team members’
control may occur which may make goal achievement extremely difficult. However, such
situations should be obvious to everyone involved and would seem to result in the goal of
doing one’s best to achieve the stated goal in light of the difficult situation. This
dissertation focuses on difficult goals which are achievable, even after an unexpected
event makes the goal more difficult, and therefore avoids having to deal with specifying

possible curvilinear relationships.

62

Second, goals which are too specific may cause individuals to focus their attention
on aspects of the task situation which, almost necessarily, perpetuate routine. Goals that
focus on processes (e.g., gathering and making decisions based on information A, B and
C) instead of the ultimate criterion (e. g., a correct decision) are be particularly
problematic in this regard in unstable environments (e. g., due to changes in technology
the correct decision becomes a function of A, B and D). This should be particularly true
for the types of teams studied in this dissertation. That is, when teams do not play a role
in the establishment of their own purpose, their being given a goal, in a sense, becomes
their purpose. Since, by definition, these teams cannot change their purpose, it is
problematic to expect they will change it. More colloquially, imposition of a purpose
related to “how something should be done” will result in the team members striving to do
that something “how it should be done”. This dissertation, therefore, will focus on
performance goals related to the team’s purpose (for being) and not on goals related to
process. For example, if a team is created and exists to make machine toolparts, their
performance goal will be stated in terms of some criterion related to making tool parts.

Therefore, with these two caveats in mind, I expect that:

Hymthesis 2a (32a): Goal difficulty and team adaptation will be positively
related such that the likelihood of adaptation will be higher in teams that have

difficult performance goals than in teams that have easy performance goals.

In addition to this hypothesized influence of goals on team adaptation, I expect

that the nature of team goals should also moderate the relationship between team

63

cognitive resources and team adaptation. That is, a team’s pool of cognitive resources
should relate more positively to adaptation in teams with difficult goals than in teams
with do your best goals because the use of declarative knowledge and lower order
productions should result in higher requirements in terms of this resource. More

specifically:

Hymthesis 2b (H_2b): The relationship between the team cognitive resources and
adaptation will be stronger in teams that have difficult performance goals than in

teams that have easy performance goals.

Feedback

Task feedback can be defined as information concerning task performance
received from different sources including others, the task environment, or even from
within the individual him or herself (Ilgen, Fisher, & Taylor, 1979). Research interest in
the relationship between task feedback and performance has been extensive, however,
there has been little theory linking feedback to psychological processes (Ilgen, et al.,
1979). The most recent review and theoretical effort sheds some light on this issue and
proposes that feedback affects the locus of attention and is critical in control processes
which involve learning, motivation and meta-tasks 0(luger & DeNisi, 1996).

Returning to the control loop illustrated in Figure 2, it is apparent that the
comparator process necessitates information that can be related to a referent standard or
goal. Indeed, the notion that feedback is necessary in the evaluation of performance

relative to standards is also consistent with the views of social cognitive theory (Bandura,

64

1991), learned helplessness theory (Mikulincer, 1994), and goal-setting theory (Locke &
Latham, 1990). Among the most important characteristics of feedback is its informational
value. The informational value of feedback can be referred to as feedback quality and
refers to the extent to which it provides meaningful information about the “correctness,
accuracy or adequacy of the response” (Ilgen, et al., 1979). Without feedback that
contains such information, individuals may be biased towards believing that their state is
consistent with expectations (Taylor, et al., 1984). Feedback with high quality should be
ti_mgy and be smpjfip so as to be unambiguous in meaning. If feedback is m and
summarized, the information provided may be ambiguous, and thus, cognitive resources
(that might otherwise be allocated towards adaptation) must be allocated towards “sorting
things out”. In addition, low quality feedback might perpetuate the triggering of scripted
behavior in team members (and perpetuate routine in teams) because there is less
indication that programmed responses are inappropriate or that changes to established
routines are worth making.

Returning to figure 4, when a team member’s comparator process notes a
discrepancy in the sub-goal control loop for which there is no programmed response, the
member will refer to his or her standing in regards to the higher order goal related to
overall team performance. This implies that individuals in teams with a high quality of
performance feedback should be more likely to recognize environmental disturbances as
problems that need to be addressed than individuals in teams with low quality '
performance feedback. If high quality performance feedback is available and

discrepancies are noted or anticipated, controlled information processing may be

65

triggered in order to develop means of decreasing this discrepancy. Therefore, since the

triggering of controlled information processing is an important element of adaptation:

Hymthesis 3a (H3a): The quality of performance feedback (timeliness and
specificity) and team adaptation will be positively related such that teams with
high quality feedback (timely and specific) should be more likely to adapt than

teams with low quality feedback (delayed and summarized).

In addition to the hypothesized influence of performance feedback on team
adaptation, I expect that the quality of performance feedback should also moderate the
relationship between team cognitive resources and team adaptation. Because the presence
of high quality feedback should be more likely to trigger controlled (information, a team’s
pool of cognitive resources should relate more positively to adaptation in teams that
receive timely and specific performance feedback than in teams that receive delayed and

summarized performance feedback

Hyppthesis 3b (33b): The relationship between the team cognitive resources and
team adaptation will be stronger in teams that receive high quality performance

feedback than in teams that receive low quality performance feedback.

Numerous researchers have concluded that both goals and feedback related to
those goals are necessary for system effectiveness (e.g., Bandura & Cervone, 1983;

Becker, 1978; Klein, 1989; Taylor et al., 1984). Stated simply, feedback has more

66

meaning when it can be compared against specific goals and vice versa. Thus, while there
is some evidence concluding that feedback in the absence of goals may cue spontaneous
goal setting (Ammons, 1956) and that goals in the absence of feedback may cue active
feedback seeking (Ashford & Cummings, 1983), it is likely that triggering of controlled
cognition, and thus adaptation, will be highest when difficult goals are coupled with

timely and specific feedback related to those goals.

Hymthesis 4a (H4a): Team adaptation is an interactive function of the difficulty
of team performance goals and the quality of performance feedback such that
adaptation will be more likely in teams that have difficult performance goals
along with high quality performance feedback related to those goals than in teams

with easy performance goals and low quality performance feedback.

In addition to the hypothesized joint influence of team performance goals and
feedback on team adaptation, I expect that these factors should also moderate the
relationship between team cognitive resources and team adaptation. Because the joint
presence of difficult team performance goals and high quality feedback should be
especially likely to trigger a mode of information processing in team members that relies
on their declarative knowledge and lower order productions, a team’s pool of cognitive
resources should relate more positively to adaptation in teams which receive difficult
team performance goals and high quality feedback than in teams that do not receive

diflicult team performance goals and/or quality feedback related to those goals.

67

Hymthesis 4b (fl4b): Team adaptation is an interactive function of the difficulty
of team goals, quality of feedback, and team cognitive resources such that the
likelihood of adaptation will be highest in teams with high levels of cognitive

resources, difficult goals, and high quality feedback.

Na_ture of Environmthal Disturbance

In the previous two sections I used control theory to describe how the nature of
team goals and feedback influence the extent to which controlled cognition in individual
team members is triggered and how these factors impact team adaptation. In addition,
however, the nature of the environmental disturbance that makes a routine inappropriate
also plays an important role in determining the extent to which controlled cognition in
individual team members is triggered in an attempt to overcome it.

In terms of triggering controlled cognition in individuals, outcomes of the self-
regulatory comparison process (between goals and feedback) must not only suggest that
there is an error or discrepancy for which there is no programmed response, but also that
the outcome of the comparison must also suggest that the error or discrepancy is
something that must be actively addressed. That is, in order to trigger the search for and
implementation of new ways of doing things once a routine is established, there must not
only be an indication of a problem, but the problem has to stand out against the backdrop
of other information that may suggest that things are all right or will be all right in the
near term (Weiss & Ilgen, 1985).

Individuals in teams may well expect occasional errors or discrepancies due to

bad luck or simple mistakes. Because perfect reliability or stability is not expected, not all

68

problems will cause increased cognitive activity directed towards search for causes and
solutions. In these situations, errors or discrepancies will be attributed to unstable causes,
and therefore, programmed responses will be deemed appropriate. In a sense, the
response of doing nothing in such situations is simply another well learned or
proceduralized program that is evoked in situations where there are occasional
disturbances or errors for which no other programmed response is available.

The implication of this idea is that if problems are to attract attention, they must
suggest unambiguously that active attention is necessary (i.e., the environment has
changed) rather than just something temporary that will work itself out on its own!
Among the most important aspects of problems in regards to the likelihood that they will
trigger controlled information processing, therefore, is the extent to which change appears
abruptly rather than gradually.

Mum changes to the situation that make routines inappropriate may be
characterized in a number of different ways. For instance, change can occur as consistent
minor inconsistencies gradually build up (e.g., seemingly unrelated minor improvements
to technology which are later integrated into a revolutionary new product). Change to the
situation can also occur as intermittent disruptions (e. g., power fluctuations in a critical
piece of command and control communications equipment that result in messages that
occasionally unreadable) eventually stabilize. In the former case, the change may be too
minor to be perceived. In the latter case, the change may be noticed but it may be
perceived as too random to address using anything but scripted responses or programs
(e.g., ignoring it). Thus, in both cases, the likelihood of triggering controlled cognition

aimed at addressing the problem is reduced. In both situations the environmental change

69

does not dominate their attention over other aspects of the situation that suggests the
situation is all right. In other words, the change in the situation is not salient to team
members.

Abrupt changes to the situation not followed by signs that the situation will return
to normal (e.g., the introduction of a new technology that will obviously revolutionize an
industry, a critical piece of command and control communications equipment catching on
fire) should increase the likelihood that team members will perceive that the situation has
changed and something needs to be done. Accordingly, team members should apply their
declarative knowledge and lower order productions in order to address the errors or
discrepancies caused by the change. Since this type of information processing in team

members is necessary for team adaptation:

Hyppthesis 5a (flSa): The nature of the environmental disturbance necessitating
adaptation will be related to team adaptation such that teams will be more likely
to adapt when the environmental disturbance is abrupt rather than when the

environmental disturbance is gradual

In addition to the hypothesized influence of the nature of the environmental
disturbance (that necessitates adaptation) on team adaptation, I expect that the nature of
the environmental disturbance will also moderate the relationship between team cognitive
resources and team adaptation. A team’s pool of cognitive resources should relate more
positively to adaptation in teams that experience abrupt environmental disruptions than in

teams that experience gradual environmental disruptions because abrupt changes to the

70

situation should be more likely to trigger a mode of information processing in team

members that relies on their declarative knowledge and lower order productions:

Hymthesis 5b (35b): The relationship between the team cognitive resources and
adaptation will be stronger in teams where there is an abrupt environmental

disturbance than in teams where there is a gradual environmental disturbance.

The predictions regarding the effect of goals, feedback, and the nature of
environmental disturbances on team adaptation come from fairly straightforward
applications of control theory. Essentially, these three factors are all hypothesized to
influence the extent to which outcomes of individual team members’ comparator process
trigger the search for a new way of doing the task based on declarative knowledge and
lower order productions versus simply evoking programs, scripts, or proceduralized
responses. In the following section, however, a complicating issue is introduced--
alternative reactions to salient problems.

Alternative Reactions to Disturbances

One important issue that complicates the fiamework outlined above is the fact that
team members may rely on cognitive modes of discrepancy reduction (i.e., abandoning
the standard, changing the standard, or rejecting the feedback message) instead of the
behavioral modes (actively engaging in new behaviors to address the negative
discrepancy) necessary for adaptation (Kluger & DeNisi, 1996). Among factors that may
influence this process is team members’ personality. Personality traits can be thought of

as labels for characteristic tendencies, and research suggests that “conscientiousness” and

71

“openness to experience” influence how individuals respond to performance-goal
discrepancies. In this section, I describe these characteristics and the argue that the
composition of teams in terms of these characteristics is an important factor for team
adaptation.

Team conscientiousness resources. Among personality variables,
conscientiousness has received the most support as a predictor of performance across a
wide variety of settings. At the individual level of analysis, Barrick and Mount’s (1991)
meta-analysis found that conscientiousness related to ratings of job and training
proficiency as well as personnel data across 5 occupational groups with an estimated E of
.22. Perhaps more relevant to the present research is Hough’s (1992) large scale study
which found that two traits associated with conscientiousness (achievement and
dependability) were related to ratings of teamwork (; = .14 and .17 respectively).

At the group level of analysis, there has been far less research on
conscientiousness, and what exists is fairly mixed. Barry and Stewert (1997), for instance,
found that the average level of conscientiousness in 63 self managed (four or five person)
MBA work groups did not relate to any group-level dependent variable (i.e., task focus,
open communications, group cohesion, and performance). LePine et al., 1997), however,
found that conscientious was important (explaining twenty-one percent of the variance) in
the performance of laboratory teams performing a decision making task. The differences
in these findings may be due to a number of factors including the nature of the teams
(intact self-managed MBA teams versus concocted laboratory teams), their task (three
analytic tasks with no correct answers versus a decision making task with correct

answers), the scales used to measure conscientiousness (Goldberg’s (1992) loo-adjective

72

unipolar five factor instrument versus the Costa and McCrae’s (1992) NBC PI-R), how
the composition variables were created (proportion of group members scoring high on the
trait versus using the lowest team members’ score), and the criterion (instructor ratings of
performance on three analytic problem solving tasks with no correct answer versus
objective measure of decision accuracy operationalized as mean absolute error). Given
the differences between these studies it is difficult to draw many conclusions regarding
the role of conscientiousness on team processes and outcomes. However, this does not
mean that conscientiousness is unimportant in teams or that research addressing the role
of conscientiousness in teams should be avoided. Instead, I argue that the role of
conscientiousness in teams should be investigated, but that instead of specifying
relationships at the team level based solely on generalizations fiom the individual-level,
specification should come from carefully considering the nature of the trait and the
criterion under consideration.

The essence of conscientiousness includes dependability (being careful, thorough,
responsible, organized and planful) and volition (hardworking, achievement-oriented, and
persevering) (Barrick and Mount, 1991). Conscientiousness individuals are proactive and
striving, they generally have a high need for achievement, and they also tend to be
committed to their work (Costa, McCrae, and Dye, 1991). Conscientiousness relates to
whether or not individuals tend to set goals for themselves (Jackson, 1974), the difficulty
of goals set for themselves (Hollenbeck and Williams, 1987; Steers, 1975), and the extent
to which individuals tend to be committed to their goals (Hollenbeck, Klein, O’Leary,

and Wright, 1989).

73

Based on the description of conscientiousness in the previous paragraph it is
reasonable to expect that conscientious individuals should tend to stay committed to goals
in the face of discrepancies resulting in the comparator process, and therefore they should
be more likely to engage in active behavioral modes of discrepancy reduction (as opposed
to cognitive modes of behavioral reduction), than individuals who are less conscientious.
Thus, as with cognitive ability, team members’ conscientiousness may be thought of as a
resource that is beneficial to teams in terms of their tendency to adapt to changing
conditions. In this regard, the pool of the team members’ conscientiousness, henceforth,
team conscientiousness resources, can be conceptualized as a characteristic of the team,
and therefore, may be considered a team level construct. Specifically, teams with
conscientious members should tend to react to problems, errors, or discrepancies
behaviorally instead of cognitively. Since such an approach to discrepancy reduction is
necessary for adaptation, teams with high conscientiousness resources should be more
likely to adapt than teams with low conscientiousness resources.

Again, as with team cognitive resources, I do not mean to suggest that aggregated
scores of some measure designed to tap conscientiousness can be used to characterize
teams as conscientious in the sense that the team is dependable or motivated. Instead, I
suggest that teams can be characterized in terms of the level of their members’
conscientiousness and that between team differences in this resource have important
implications for team level adaptation (because team conscientiousness resources
influence the extent to which behavioral as opposed to cognitive responses to

discrepancies, problems, or errors are triggered).

74

As with team cognitive resources, I offer hypotheses relating team
conscientiousness resources to team adaptation assuming an additive model but
acknowledging that other Operationalizations may be appropriate. Thus, the issue of

aggregation is posed as a research question.

Hymthesis 6a (H63): Team conscientiousness resources and adaptation will be
positively related such that teams with high levels of conscientiousness resources
will be more likely to adapt than teams with low levels of conscientiousness

resources.

In addition to this hypothesized relationship, however, I also expect that team
conscientiousness resources will also moderate the relationship between team cognitive
resources and team adaptation. This expectation follows from the notion that while a
team’s pool of cognitive resources may be critical in the search for and implementation of
alternative means of accomplishing the team’s purpose, this pool of resources will be
relatively unimportant if team members invoke cognitive means reducing discrepancies
(e. g., lowering the standard, withdrawing from the situation, etc.). More specifically, a
team’s pool of cognitive resources should relate more positively to adaptation in teams
with high levels of team conscientiousness resources than in teams with low levels of
team conscientiousness resources because behavioral means of discrepancy reduction
should be more cognitive resource dependent than cognitive means of discrepancy

reduction. Therefore:

75

Hypothesis 6b (I_-_I6b): The relationship between team cognitive resources and
adaptation will be stronger for teams with high levels of team conscientiousness

resources than for teams with low levels of team conscientiousness resources.

Team omnness resources. Relative to conscientiousness, openness to experience
is less well understood (Barrick & Mount, 1991). This dimension, also referred to as
intellect (Digrnan, 1988; Peabody & Goldberg, 1989), inquiring intellect (D. Fiske,
1949), culture (Norman, 1963), and intelligence (Borgatta, 1964; Cattell, 1957; Hogan,
1986). Openness to experience includes traits such as imagination, culture, curiosity,
originality, broad-mindedness, intelligence and artistic sensitivity (Barrick & Mount,
1991). At first glance, Openness to experience might sound suspiciously like cognitive
resources described earlier. Research, however, has found only moderate relationships
(correlations ranging from .20 to .30) between scales that measure these two concepts
(McCrae & Costa, 1987).

Although openness to experience has not been found to relate positively to
individual-level performance, there is a strong indication that it is an important predictor
of training proficiency (Barrick & Mount, 1991). This effect has been attributed to the
fact that people who are open to experience are more likely to “have positive attitudes
toward learning experiences in general” (Barrick & Mount, 1991, p. 19). That is, these
individuals may be more willing to engage in the type of self-monitoring and assessment
that is necessary in learning situations. Mche (1987) suggests that openness to

experience is strongly related to divergent thinking and creativity. Individuals who are

76

open to experience are generally more willing to explore alternatives which diverge from
the normal routine.

Errors, discrepancies, or problems should be interpreted by individuals who are
open to experience as opportunities to learn or be creative, and not something that is
necessarily dreadful or threatening to the self. These individuals, therefore, should be less
likely to engage in cognitive responses (i.e., changing the standard, abandoning the goal,
rejecting feedback) than those who are not open to experience. Since behavioral
approaches to discrepancy reduction are necessary for adaptation in teams, teams with
members who have high levels of openness to experience should be more likely to adapt
than teams with members who have low levels of openness to experience. In this regard,
the pool of the team members’ openness to experience, henceforth, team openness
resources, can be conceptualized as a characteristic of the team, and therefore, is
considered a team level construct.

As with team cognitive resources and team conscientiousness resources, I do not
mean to suggest that aggregated scores of the individual concept mean the same thing as
the disaggregated scores. Instead, I suggest that teams can be characterized in terms of
their members’ openness to experience and that between team differences in this resource
has important implications for team adaptation. That is, team openness resources
influence the extent to which behavioral as opposed to cognitive responses to
discrepancies, problems, or errors are invoked. Also, as with team cognitive and
conscientiousness resources, I offer hypotheses regarding the relationship between

openness resources and adaptation, but I examine means of aggregation as a research

question.

77

Hymthesis 7a (H7a): Team openness resources and team adaptation will be
positively related such that teams with high levels of openness resources will be

more likely to adapt than teams with low levels of openness resources.

As with team conscientiousness resources, I also expect that team openness
resources will also moderate the relationship between team cognitive resources and team
adaptation. This expectation follows from the notion that while team cognitive resources
may be critical in undertaking behavioral means of addressing discrepancies, these
resources will be relatively unimportant if team members invoke cognitive means

reducing discrepancies. Therefore:

Hymthesis 7b (H7b): The relationship between team cognitive resources and
adaptation will be stronger for teams with high levels of team openness resources

than for teams with low levels of team openness resources.

SM and Hypptheses

The use of specialist decision making teams in organizations is increasing.
Because these teams do their work by making a stream of recurring decisions, they are
particularly susceptible to the problem of maintaining a routine in the face of some
change in the environment which makes the routine inappropriate. To date, however,
there has been little research on how teams adapt to such change. The purpose of this

dissertation, therefore, is to develop a model of some factors that influence team

78

adaptation (defined as reactive and non-scripted adjustments to a team’s role structure
that result in a better fit given the new situation). Consistent with information processing
as an underlying theoretical framework and a recognition that this framework can be
applied to groups and teams, routine and adaptation can be understood in terms of the
type of knowledge that controls information processing and behavior. Consistent with
Cohen and Bacdayan (1996), routines are conceptualized in terms of a system of
reciprocally triggered procedural knowledge among team members. Adaptation, on the
other hand, requires the application of team members’ declarative knowledge and lower
order productions. Among the most important implications of this framework is the
notion that the latter form of information processing relies heavily on team members’
cognitive resources. A second implication, however, is that in order for team adaptation
to occur, team members’ information processing must switch from a procedural
knowledge based mode to a mode based on declarative knowledge and lower order
productions. Control theory is used as a basis for identifying three factors that should
facilitate triggering the second information processing mode. Specifically, I hypothesized
that specific and difficult team performance goals, high quality performance feedback,
and abrupt environmental disturbances make discrepancies, errors, or problems salient,
and thus, increase the likelihood of triggering declarative knowledge or lower order
production based information processing in team members. One complicating factor is
that while behavioral change is necessary for team adaptation, members may respond to '
discrepancies, errors, or problems by cognitive means (e. g., abandoning or lowering
goals, withdrawing from the situation). Among critical factors in avoiding such cognitive

responses, however, are team members’ commitment to their goals and their willingness

79

to explore alternatives which diverge from the norm. Team members’ conscientiousness
and openness to experience, therefore, should be important in ensuring teams respond to
environmental disruptions through adaptation. In summary, I present the following

hypotheses:

Hyppthesis 1 (H1): Team cognitive resources and team adaptation will be
positively related such that teams with high levels of cognitive resources will have

a higher likelihood of adapting than teams with low levels of cognitive resources.

Hymthesis 2a (H2a): Goal difficulty and team adaptation will be positively
related such that the likelihood of adaptation will be higher in teams that have

difficult performance goals than in teams that have easy performance goals.

Hymthesis 2b (H2b): The relationship between the team cognitive resources and
adaptation will be stronger in teams that have difficult performance goals than in

teams that have easy performance goals.

Hyppthesis 3a (33a): The quality of performance feedback (timeliness and
specificity) and team adaptation will be positively related such that teams with
high quality feedback (timely and specific) should be more likely to adapt than

teams with low quality feedback (delayed and summarized).

80

Hyppthesis 3b (H3b): The relationship between the team cognitive resources and
team adaptation will be stronger in teams that receive high quality performance

feedback than in teams that receive low quality performance feedback.

Hypothesis 4a (H4a): Team adaptation is an interactive function of the difficulty
of team performance goals and the quality of performance feedback such that
adaptation will be more likely in teams that have difficult performance goals
along with high quality performance feedback related to those goals than in teams

with easy performance goals and low quality performance feedback.

Hyppthesis 4b (34b): Team adaptation is an interactive function of the difficulty
of team goals, quality of feedback, and team cognitive resources such that the
likelihood of adaptation will be highest in teams with high levels of cognitive

resources, difficult goals, and high quality feedback.

Hyppthesis 5a (HSa): The nature of the environmental disturbance necessitating
adaptation will be related to team adaptation such that teams will be more likely
to adapt when the environmental disturbance is abrupt rather than when the

environmental disturbance is gradual

Hyp_othesi§ 5b (HSb): The relationship between the team cognitive resources and
adaptation will be stronger in teams where there is an abrupt environmental

disturbance than in teams where there is a gradual environmental disturbance.

81

Hyp_othesis 6a (H6a): Team conscientiousness resources and adaptation will be
positively related such that teams with high levels of conscientiousness resources

will be more likely to adapt than teams with low levels of conscientiousness.

Hymthesis 6b (H6b): The relationship between team cognitive resources and
adaptation will be stronger for teams with high levels of team conscientiousness

resources than for teams with low levels of team conscientiousness resources.

Hyppthesis 7a (fl7a): Team openness resources and team adaptation will be
positively related such that teams with high levels of openness resources will be

more likely to adapt than teams with low levels of openness resources.

Hyppthesis 7b (H7b): The relationship between team cognitive resources and
adaptation will be stronger for teams with high levels of team openness resources

than for teams with low levels of team openness resources.

In addition to these hypotheses, I will examine three different means of
aggregating individuals’ cognitive ability, conscientiousness, and openness to experience.
Specifically, I will attempt to determine which mode1(s) (additive, disjunctive or
conjunctive) best predicts adaptation. The next chapter describes a proposed laboratory

study to test these hypotheses and research questions.

82

Chapter 2

METHOD

Most studies of adaptation have been based on post-hoc qualitative analyses of
single events in the field. Such an approach to studying adaptation is reasonable because
it is not only difficult to be able to anticipate or manipulate an environmental disruption,
but even if this were possible, it is difficult to parsimoniously describe a new pattern of
behavior in terms of either objective or perceptual psychometric variables. Studying
adaptation is difficult because adaptation, by definition, is emergent and complex. Indeed,
qualitative retrospective reports, even if serendipitous and based on a sample of one, can
be enlightening and are important to scientific discovery (McCall & Bobko, 1990).
However, in order to increase understanding of adaptation further, it is necessary to begin
conducting research in settings where a_pp'o_ri hypotheses have a chance of being
disconfirmed. The present research is a step in this direction in that not only are
hypotheses offered, but they will be tested quantitatively in a controlled setting. This
chapter describes the research method and analyses used to assess the hypothesized
relationships. First, I discuss the choice of study setting, sample, task, design, and data
collection techniques. I then describe the results from the pilot and main studies.

Choice of Setting

The research questions raised in this dissertation were assessed in a laboratory
study. The choice of a laboratory setting over the field was based on two factors. First,
the nature of the research question almost necessitates laboratory experimentation. The

laboratory allows for (a) the observation of many teams performing the same task in the

83

same type of environmental conditions (needed for adequate statistical power and to rule
out artifacts such as selection and history), (b) the presentation of an adequate stream of
similarly structured decisions (enough so that routines can be developed and adaptation
can occur), and (c) the manipulation of team goals, feedback, and the nature of the
environmental disruption and random assignment of teams to conditions based on these
manipulations (ensures that there are teams in all possible cells and helps to ensure
equality of teams prior to experimental manipulation). For obvious practical reasons,
these conditions would be extremely difficult to find in a natural field setting.

Second, the hypotheses and their theoretical support do not imply boundary
conditions that make a laboratory setting inappropriate for testing the hypotheses.
Requirements for this study include teams of two or more individuals who interact, are
interdependent, and share a purpose. In addition, these teams must be composed of
individuals who are specialized and who make a stream of recurring structured decisions.
Finally, the task must be such that it is moderately complex, and is long enough to allow
for the development of routinized behavior. As explained in the sections entitled ”Task”
and “Procedures”, these requirements are fulfilled in this study.

Planning for Particimts: The Power Analysis

In planning this study, I conducted a power analysis to determine an appropriate
sample size for the main study (J. Cohen, 197 7; J. Cohen & P. Cohen, 1983). Statistical
power is defined as the probability of correctly rejecting the null hypothesis (i.e.,
rejecting the null hypothesis when the null hypothesis is false). Thus, planning for power
is desirable because it serves as a control for a type 11 error, or the probability of failing to

reject the null hypothesis when it is false.

84

The strategy of planning for statistical power depends on a consideration of four
factors, (1) the power of a statistical test, (2) the specified probability of rejecting the null
hypothesis when it is true (or), (3) the sample size (n), (4) and the magnitude of the effect
size. Once any three factors are known or estimated, the fourth can determined. For
planning purposes in assessing hypotheses, or was set according to convention at .05.
According to convention, I set desired power to be .80. Since there are no studies on team
adaptation (or on adaptation at any level of analysis) which use the variables under
consideration in this study, it is impossible to get estimated effect sizes from the
literature. Establishing a minimum effect size based on theoretical or practical
significance would also be problematic since there are so few studies on this topic. Thus,
the approach used here was to choose an effect size somewhere in-between what Cohen
considers small (r=.10) and medium (r=.30). However, since in some domains (e.g.,
between participation and performance) correlations around .20 have been questioned in
terms of practical significance (Wagner, 1994), I leaned towards the medium end of this
range and chose a correlation of .25 (i.e., R2 =.0625) for the minimum effect size.

I had assumed that data collected would conform to the assumptions necessary for
multiple linear regression so I planned to use this approach to assess my hypotheses. In
order to find the appropriate sample size for the regression analyses, the following
formula is appropriate:

11 = L/F2 + K + 1,
where F 2 = srz/l-Rz, L comes from Table E2 in J. Cohen and J. Cohen (1983) where KB =

1, sr2 is the semi-partial effect size estimate, and K = number of independent variables.

85

First, using this formula for the main effect hypotheses together with the values;
sr2 = .0625, R2 = .30 (6 independent variables X .0625 = .375 X .80 correction factor for
correlated predictors), a .05, and power = .80, L = 7.85 and the appropriate minimum
sample size works out to be 95. Using these same assumptions, the required sample
minimum sample size to detect a seventh term (i.e., an interaction) with power of .80 is
90. Therefore, given the assumptions in this analysis, a minimum sample size of 90 teams
was though to be adequate to test the hypotheses.

Man—ts

Research participants included 489 college juniors and seniors enrolled in a
management course during the fall semester at a large Midwestern university. Sixty of
these students participated in a pilot study and the rest (429) participated in the main
study. The mean age for these participants was 21 (SD = 2.29), their self-reported GPA
was 3.06 out of 4.00 (SD = .43), and half were male. Participation in this study was
voluntary and participants were informed that they could withdraw from the study at any
time. Participants who complete the study received participation credit for their time (3
hours). All teams had the opportunity to earn bonuses paid to the team on the basis of
team performance. The top performing teams in each condition earned $60, the next best
team in each condition earned $45, and the third best team in each condition earned $30.
Individuals who chose not to participate or withdrew from the study were able to obtain
equivalent course credit by completing an assignment that required an equivalent number

of hours of work (3 hrs).

86

 

In the pilot and main studies, research participants worked at a team-based version
of the Team Interactive Decision Exercise for Teams Incorporating Distributed Expertise
computer simulation task (TIDEZ). A brief description of this task is provided below. The
interested reader is referred to Hollenbeck et al. (1995) for a more complete description.

TIDE2 is a software program for a decision task simulation that presents
participants with values on a ntunber of attributes of a problem or object. Research
participants are asked to make decisions about the state of that object (e.g., the
desirability of a job candidate, the value of a particular piece of merchandise, the severity
of a specific injury, etc.) based on these attributes. In this study, TIDE2 was programmed
to simulate a military command and control team. When programmed to simulate a
command and control team, the structure of the experimental task has a high level of
mundane realism (Berkowitz & Donnerstein, 1982). The participants in this study were
stationed at networked computer terminals and were responsible for communicating over
this network in order to make classification decisions about targets within an airspace. In
this way, the task corresponds very closely to what real command and control teams do
on a day to day basis. Past research using TIDE2 programmed as a command and control
scenario has also demonstrated that participants generally enjoy the task and want to do
well (Hollenbeck et al., 1995; LePine, et al., 1997). These types of reactions reduce
potential problems associated with participants being bored and not caring about their
performance. The surface features of the task, however, do not define the boundaries of
this study to military command and control team contexts. Indeed, command and control

teams are but one example of specialist decision making teams in general. Therefore, to

87

the extent that the task models a command and control context, it should provide a
reasonable venue for assessing hypotheses regarding specialist decision making teams.
The team's mission in this study was to monitor an airspace. When a target aircraft
enters this airspace, each team member needed to gather some information about
particular attributes of the aircraft (e. g., its speed, altitude, range, etc.), and then arrive at
a judgment regarding the appropriate response to make toward the aircraft. Each team
member was assigned the role of an officer in one of three command and control stations;
Alpha, Bravo or Charlie. Each team member was trained so as to possess unique expertise
regarding aspects of the task. Alpha was trained how to interpret information related to
location (number of targets, range, and position relative to commercial air corridors) of
the target aircraft (the location rule). Bravo was trained how to interpret information
regarding the motion (heading crossing angle, altitude, and speed) of the target aircraft
(the motion rule). Finally, Charlie was trained how to interpret information regarding the
category (electronic security measure, radar cross section, and rate change altitude) of the

target aircraft (the category rule). These responsibilities are illustrated in Figure 4.

 

 

 

 

 

 

Rule Alpha Station Bravo Station Charlie Station
Number of Targets
Location Range
Corridor Status
Heading Crossing Angle
Motion Altitude
Speed
Electronic Security Measure

Category Radar Cross Section

Rate Change Altitude

 

Figure 4 - Role Requirements

88

 

The instructions were such that participants knew that their area of expertise
included only one decision rule and that making a recommendation based on their rule

required knowledge of all three attributes. This is because a value of non-threatening for

 

any attribute in a rule makes correct decision for the rule non-threatening (ignore). Alpha,
who was responsible for making the final decision, was taught that the final correct
decision for each target aircraft is an additive combination of the three rules. In
monitoring the assigned airspace, teams had to assess a series of air targets in terms of
their level of threat. Judgments on rules and final decisions were rendered on a seven-

point continuum anchored by Ignore (lowest level of threat) to Defend (highest level of

 

threat). Intermediate responses on this scale in increasing level of threat included; (2)

review, (3) monitor, (4) warn, (5) ready, and (6) lock-on.

 

 

The task was programmed so that there was interdependence among staff
members in the way information could be accessed. Although the three rules involved
combinations of three distinct cues, no one team member was able to measure all of the
necessary components of any one rule. For instance, Alpha was able to directly access
“number of targets” and “range” but not corridor-status”. This piece of information had
to come from Charlie. Figure 5 illustrates who could measure which attributes. In this
Structure, Alpha needed to transmit “speed” to Bravo, Bravo needed to transmit “rate
Change altitude” (RCA) to Charlie, and Charlie needs to transmit “corridor status” to

Alpha. This structure is illustrated in Figure 6.

89

 

 

 

 

 

Rule Alpha Station Bravo Station Charlie Station
Number of Targets

Location Range

Corridor Status

Heading Crossing Angle
Motion Altitude
Speed
Electronic Security Measure
Category Radar Cross Section

 

Rate Change Altitude

 

Figure 5 - Measurable Attributes

 

Alpha

 

Corridor Status Speed

 

 

Charlie Bravo

 

 

 

 

 

 

 

RCA

Figure 6 - Pre-disruption Role Structure

90

 

As stated earlier, adaptation is defined as reactive adjustment to a role structure
resulting in a role structure that makes sense given the new situation. The environmental
disruption or situation change in this study was the failure of the transmit mechanism
between Bravo and Charlie. Bravo was able to execute the transmit “RCA” command,
however, the attribute “RCA” did not reach the Charlie.

A clock on the screen counted down the time before the team needed to make a
decision. The computer began to beep when there is 30 seconds remaining. This gave
participants a clear impression that the time available for making a decision was running
out. Recommendations regarding each decision object were forwarded from Bravo and
Charlie to Alpha who then considered his or her own information along with these
recommendations and made a final decision for the team before time runs out.

In the course of the entire experiment (which lasted 3 hrs), participants made
decisions on a series of 83 (including 3 for training and 2 for practice) targets. Following
decisions, team members received feedback indicating the accuracy of each staff
member’s judgment as well as accuracy of the team’s decision. However, only the team's
performance mattered in terms of winning bonus money; hence team members had a
common fate. Team performance was defined in terms of decision accuracy, which in
turn was defined as the absolute difference between the decision object’s “true score” and
the team’s decision on that decision object. Decision objects’ true scores were based on
the rules that comprise participant’s role training. Decision accuracy could range from 0
(a perfect match) to 6 where a low score indicated higher accuracy. Only five outcomes
were fed back to participants, however. A _Ip'_t_ indicated no difference between the true

score and the team’s decision. A near miss indicated a difference of one. A miss indicated

91

a difference of two. An incident indicated a difference of three. Finally, a disaster

 

 

indicated four or more points off. Associated with each outcome were points; 2 for a hit,
1 for a near miss, 0 for a miss, -1 for an incident, finally —2 for a disaster.

The task as just described included several elements which made it a team task.
First, it is clear that more than one individual was involved in the decision making and
these individuals were interdependent. Each team member was dependent on others to
provide them with critical pieces of information (task interdependence). Since there was
only one final decision per object, and since team members were rewarded based on team
performance, members were also dependent on one another in terms of their making
accurate recommendations and judgments regarding target classification (outcome
interdependence). Finally, team members had a common purpose (to make accurate
decisions) for which members shared a common fate (team based performance rewards).

In addition to these characteristics, the teams had several attributes which made
them examples of specialist decision making teams. First, the teams were responsible for
processing and sharing information in order to make decisions and therefore the task is
primarily cognitive as opposed to psychomotor. Second, the teams were responsible for
doing work related to a defined purpose (i.e., making a stream of decisions about targets
that come into their assigned area) and not for defining that purpose. Third, team
members were trained about unique aspects of the team’s task and therefore they were
Specialized. Finally, the task itself was moderately complex. In the task each team
member needed to (a) perform a number of distinct tasks (e. g., measure attributes,
communicate with team members, consider attribute values and make recommendations

about targets’ level of threat), (b) learn who had which attributes and the relative weights

92

of those attributes for predicting the criterion, and (c) be involved in developing a way to
do the task once the situation changed.
Procedure

Research participants were scheduled at the beginning of the semester for their
three—hour experimental session. Research assistants scheduled participants during the
first recitation meeting of the semester. Because appointment slots were alternated from
one section to another, and because most students did not know one another early in the
semester, students did not sign up as intact or familiar groups. At this time, participants
were given general information concerning the purpose of the experiment (“a study of
decision making in teams using a computer simulation”). After participants signed-up,
they were instructed to fill out a consent form (Appendix A). The participants were then
given an abbreviated NEO-PI-R personality inventory (Costa & McCrae, 1992) and the
Wonderlic Personnel Test Form IV (Wonderlic & Associates, 1983) in order to measure
individuals’ conscientiousness, openness to experience, and general cognitive ability.
Teaching assistants made periodic announcements throughout the semester in order to
inform students who missed the first recitation session to make arrangements with the
laboratory to sign-up for the experiment and complete the instruments.

Prior to the experimental sessions, teams were randomly assigned to experimental
conditions. In order to ensure enough participants were available to compose teams,
participants were overbooked by one. Upon arrival to the laboratory, participants filled
out another consent form (Appendix B). After 3 participants arrived, they were taken into
a Separate room and instructed to take a seat at any of the computer consoles. Because

participants chose their own seat, there was random assignment of position. If a fourth

93

participant showed-up for the experiment, he or she was assigned to an alternative study
that required the same amount of time and effort. Once participants were seated, the
experimenter welcomed the participants and gave them a general task overview booklet
(Appendix C) and an individual role responsibility sheet (Appendix D). The experimenter
then overviewed the material and gave participants 8 minutes to read through it
themselves. After 8 minutes the experimenter returned and answered questions. He or she
A then proceeded with the hands on portion of the training.

The hands on portion of the training primarily involved the first (of five) training
trials and is described in the experimenter’s protocol (Appendix E). During this first trial,
research participants were instructed about the mechanics involved in gathering and
sharing information about target attributes. These actions included (a) measuring attribute
values, (b) Mg others for attribute values, (c) directly transmitting attribute values to
others (only permitted on values that could be measured by the station), and (d)
communicating via sentence-long free form text-messages (not permitted between Bravo
and Charlie). The task training also involved instruction on individual role
responsibilities. Role information was provided in a role responsibility sheet which
participants kept throughout the experiment. Role expertise information included: (1) the
specific attributes the participants needed for their role, (2) how to translate raw data on
targets into judgments about how threatening the target is likely to be on a specific
attribute, and also (3) how to combine information on attributes into judgments about
rules. The participant filling the role of Alpha was also instructed how to combine

judgments on rules into a final decision for the team. During the first training trial,

94

participants were also encouraged to practice communicating with each other using all
modes of communication.

Following the first trial, participants were given the opportunity for hands-on
practice on two additional trials. During these two trials, participants were encouraged to
practice communicating with one another in order to obtain the information they needed
for their role. The research assistant stood by to answer questions and clarify obvious
misunderstandings. Prior to the fourth trial, the experimenter instructed participants on
how to exchange information most efficiently (using the structure illustrated in Figure 7).
During the fourth and fifth trials, the participants practiced using this structure.

After the fifth trial, the simulation was paused and any final questions were
answered. Participants were told that in order to make the simulation more realistic and
interesting there may be some communications jamming and that information sent by one
member might not go through to the other member. Participants were also told to do their
best to work around this problem and that it was important to obtain all the information
for their role. This instruction was necessary so that participants would not stop engaging
in the experiment because they thought that there was something wrong with the
equipment that could be fixed by an experimenter (e.g., a computer virus). This
instruction was also important to prevent participant frustration and anger due to their
feeling disadvantaged relative to other teams who may not have experienced such a
problem with their equipment.

At this time participants were also reminded that all communications must
transpire over the computer and that while they could send free-form text messages to one

another at any time during the remaining trials, there would be seven opportunities to

95

send these messages during feedback displays between trials. In contrast to the normal
feedback displays lasting 7 seconds, these opportunities lasted 60 seconds. These
opportunities gave even the slowest typist an opportunity to communicate with the other
members of the team. These opportunities also allowed for communications in a
somewhat less stressful environment. I should note, however, that these periods did not
allow for extensive communications and planning. Following this instruction, participants
were given a 6 item task knowledge test (Appendix F). After participants completed this
test, the experimenter wished the team “good-luck” and re-started the simulation. The
experimenter then left the room but was able to monitor (listen to) the team to ensure
there was no talking during the experiment.

Following the experiment, participants were given an instrument (APPENDIX G)
that included questions about the reliability of communications between members during
the experiment (items 11-34). Some of these items were used as a manipulation check for
the abruptness disruption (depending on the condition). The entire subscale was also used
as a component of a measure of situational awareness (described more fully later).
Appendix G also included manipulation checks items related to perceptions of goal
difficulty (items 1-2), value of feedback (items 3-4), amount of effort (items 7-10), and
perceived importance of the disturbance (items 5-6). After participants finished the
instrument, they were debriefed and released.

Overall Desigp

This main study uses 6 independent variables in a between groups design. Goals

difficulty (difficult versus easy), quality of feedback (timely and specific versus delayed

and summarized), and the nature of the environmental disruption (gradual vs. abrupt)

96

were manipulated. Team cognitive resources, team conscientiousness resources, and team
openness resources were not manipulated in this study and were operationalized as
continuous variables. Prior to the main study, a pilot study was conducted to ensure the
task, scenario, and manipulations were working as intended.

The Pilot Study

Prior to the primary study, I conducted a pilot study using sixty participants.
These participants were arrayed into 20 three-person teams and performed the task as
described above. The pilot study had four specific purposes: (1) to ensure teams would
establish and maintain routines prior to the communications breakdown, (2) to ensure that
adapting the communications structure was necessary to promote high performance after
the communications breakdown, (3) to establish easy and difficult team performance
goals, and finally, (4) to ensure the communications breakdown manipulation captured
the participants’ attention.

The pilot teams also served a number of other important functions. For instance,
initial teams in the pilot were used in the training of five research assistant experimenters.
Additionally, bugs in the computer program were identified and fixed during the first few
pilot runs. Finally, based on pilot participant debriefings, I was able to refine the
experimental procedures and instructions in order to reduce ambiguity and confusion
during the training.

Evidence of routine. One of the most important purposes of the pilot study was to
ensure the task allowed teams to establish and maintain a routine means of exchanging
information prior to the communications disruption. M. Cohen and Bacdayan (1996)

suggested three factors that indicate the presence of a routine: (I) evidence of repeated

97

action sequences, (2) increased speed in performing the task, and (3) increased reliability
in performing a task.

To ensure all teams used the W, participants
were told how to exchange information the most efficient way possible (illustrated in
Figure 7) after the third training trial. To assess the extent to which team members
maintained this structure, I examined the communications data for the teams. The data
should show that for each trial Alpha transmits “Speed” to Bravo, Bravo transmits “Rate
Change Altitude” to Charlie, Charlie transmits “Corridor Status” to Alpha, and finally,
that Alpha, Bravo, and Charlie receive the information that has been transmitted to them.

To assess the increase in smed in p_erforming the task, I compared the average
speed of receiving information across four blocks of ten trials (the first ten trials after
training, trials 14 through 23, trials 24 through 33, and trials 34 through 43). Although
team members’ communications required different types of activities (i.e., measuring
attributes, transmitting information, receiving information), receiving information from
other team members represented the final step in obtaining needed information and
therefore captured prior activity. I expected that there would be a significant decrease in
the amount of time it took for members to receive their information between the first and
second block of trials and that this increase in speed would stabilize between the third and
fourth blocks as their routine becomes established.

To assess the extent to which there was increased reliabilig in worming the
g, I compared the standard deviation of the speed of receiving information across the
same four blocks of ten trials (the first ten trials after training, trials 14 through 23, trials

24 through 33, and trials 34 through 43). I expected that there would be a significant

98

decrease in the standard deviation of time it took the team members to receive their
information between the first and second block of trials. I also expected that the decrease
in the standard deviation would stabilize between the third and fourth blocks as the
routine becomes established.

Necessity for adaptation. The second purpose of the pilot study was to ensure that
at least some teams adapted, and also to make sure adaptation was necessary in terms of
promoting higher performance after the environmental disruption. In order to ensure
some teams adapted, I examined the extent to which they developed role structures that
were beneficial in terms of promoting higher performance after the communications
breakdown. First, I compared the performance of 5 teams that were trained in the ideal
way to exchange information after the disruption made the original routine inappropriate
(this structure is described more fully in the next section) to teams that were not trained in
the ideal adapted structure. This comparison will be based on the average performance (in
terms of decision accuracy) for however many trials on average it takes for the teams to
begin to find new ways to give Charlie the value for "Rate Change Altitude". For
instance, if untrained teams found a way to give Charlie his or her information on trial 60,
the comparison will be on trials 54 (the trial where the disruption first occurs) through 60.
As a second check on the importance of adapting, I compared the average decision
accuracy across an equivalent number or trials prior to and after teams found a way to
deal with the communications breakdown.

Establishing easy and difficult goal levels. The third purpose of the pilot was to
establish difficult and easy goal levels in terms of team level decision accuracy. I

examined the pilot teams’ performance and set the difficult and easy goal levels at

99

approximately 1 standard deviation above and 1 standard deviation below the mean of the
pilot teams’ performance level.

Manipulation check. Finally, I examined the post-experiment manipulation check
questions to ensure that participants noticed the communications disruption.

The Pilot Study: Results

Evidence of routine. As described next, the pilot teams showed evidence of
routine based on three criterion suggested by M. Cohen and Bacdayan (1996).

First, based on a qualitative examination of the communications data, all 20 pilot
teams appeared to use the same pattern of exchanging information they were encouraged
to use during training. Team members transmitted the attribute they needed to transmit
and received the information transmitted to them. I also computed the average number of
times each member of the team transmitted and received all the appropriate information.
Over the 20 trials prior to the communications breakdown (trials 24-43), individuals in
teams received the information they needed for their role 98.5% of the time (sd = .02).

Second, team members appeared to increase the speed in performing their portion
of the team’s task early in the simulation. The mean number of seconds to receive needed
information (with standard deviations in parentheses) for trial blocks 1 through 4 were
22.19 (5.66), 17.40 (3.52), 16.39 (4.32), and 15.89 (3.67), respectively. A One-way
AN OVA revealed that at least one of these blocks of trials differed from another, F(3,
236) = 26.10, p < .001. I conducted 3 post hoc multiple comparison tests (Tukey’s
Honestly Significant Difference, Bonferroni, and Scheffe) in order to find which blocks

differed. The results of this test are shown in Table 1.

100

Table l - Comparisons of Time to Receive Information

 

Significant Differences (p < .05)

Trial Block Difference in Mean Tukey Bonferroni Scheffe
(i) (j) Time to Receive HSD

 

(ii)
1 2 4.79 .00 .00 .00
3 5.80 .00 .00 .00
4 6.31 .00 .00 .00
2 1 -4.79 .00 .00 .00
3 1.01 .58 1.00 .66
4 1.51 .23 .36 .31
3 1 -5.80 .00 .00 .00
2 -l .01 .59 1.00 .66
4 .50 .92 1.00 .94
4 l -6.31 .00 .00 .00
2 -1.51 .23 .36 .31
3 -.50 .92 1.00 .94

 

The Bonferroni test, Tukey’s honestly significant difference test, and the more
conservative Scheffe test all revealed that the mean time to receive information in the first
block of trials was significantly greater than in blocks 2 through 4 which did not differ
significantly from one another. Thus, the pattern of these data suggest that members
increased the speed in performing their portion of the team’s task during the first two
blocks of trials but the increase in speed leveled off during the third and fourth blocks.

Finally, the pilot data also suggested that the reliability of team members’ role
performance increased over time. The standard deviations in seconds of the time it took
members to receive their information (with standard deviations of these standard
deviations in parentheses) were 7.94 (5.26), 4.10 (3.33), 3.90 (3.44), and 3.67 (3.19) in

blocks 1 through 4 respectively. A One-way ANOVA revealed that at least one of these

101

blocks of trials differed from another, E (3, 236) = 12.61, p < .001. Results from a
Bonferroni test, Tukey’s honestly significant difference test, and the Scheffe test
(illustrated in Table 2) revealed that the standard deviation in seconds to receive
information in the first block of trials was significantly greater than in blocks 2 through 4
which did not differ significantly from one another. The pattern of these data suggest that
members increased the reliability in performing their task during the first two blocks of

trials but this increase in reliability leveled off during the third and fourth blocks.

Table 2 - Comparisons of Reliability in Receiving Information

 

Significant Differences (p < .05)

Trial Block Difference in Standard Tukey Bonferroni Scheffe
(i) (j) Deviations to Receive HSD

 

(ii)
1 2 3.84 .00 .00 .00
3 4.04 .00 .00 .00
4 4.27 .00 .00 .OO
2 l -3.84 .00 .00 .00
3 .21 .99 1.00 1.00
4 .43 .95 1.00 .96
3 l -4.04 .00 .00 .00
2 -.21 .99 1.00 1.00
4 .23 .99 1.00 .99
4 1 -4.27 .00 .00 .00
2 -.43 .95 1.00 .96
3 -.23 .99 1.00 .99

 

Necessity for adaptation. As defined earlier, team adaptation implies the
development of a more appropriate role structure in light of a new situation. Therefore,

before I could confidently say that some teams adapted to the communications

102

breakdown, it was important to ensure that at some teams responded by developing new,
more appropriate role structures.

It would be reasonable to say that a W role structure should
ultimately result in higher levels of team performance than other behaviors. Thus, in the
context of this study, it was important to assess the extent to which informing Charlie
(getting "Rate Change Altitude" to Charlie) resulted in higher team-level performance
than not informing Charlie. Of the 15 pilot teams that were not trained in the ideal
adapted structure, 3 never found a way to get Charlie the information he or she needed
after the communications disruption. It took the remaining 12 pilot team an average of 11
(sd = 6) trials to find a way to get “Rate Change Altitude” to Charlie. That is, while the
link between Bravo and Charlie was severed on trial 54, on average these teams began to
inform Charlie on trial 65. Across these first 11 trials, the average performance in terms
of decision accuracy or TDA, (the average of the absolute difference between team
decisions and the correct decisions) of the 15 teams that were not trained how to
exchange information was 2.00. For teams that were trained in the ideal structure,
however, TDA across these 11 trials was .69. That is, teams that were trained how to
inform Charlie made decisions that were, on average, 1.31 points better than teams that
were not trained, E(l,18)= 31.30, p < .001.

Next, I compared the performance of the 12 teams that found a way to inform
Charlie on the first 11 trials after the communications disruption to the performance of
these teams on the last 11 trials in the scenario. This is just a simple comparison of teams'

performance before and after they learned how to inform Charlie. As expected, TDA on

the last 11 post—disruption trials was significantly higher than that on the first 11 post-

103

disruption trials (1.08 versus 2.02), _t_ (11) = 5.48, p < .001. Thus, overall, it appears that
informing Charlie in this task should be beneficial in terms of promoting higher levels of
team performance.

Based on the analyses outlined in the two previous paragraphs, I was fairly
confident that any appropriate role structure should include behavior that results in
getting Charlie the "Rate Change Altitude". However, just because Charlie gets informed,
does not necessarily imply that the team has adapted because the definition of adaptation
implies that the team establishes a new role structure.

As mentioned earlier, a role structure can be defined as the pattern of recurring
actions of individuals interrelated with the recurring actions of others (Katz and Kahn,
1978, pp. 189). In order to say that a team has established a new role structure, therefore,
evidence should suggest that a new pattern of activity or behavior among team members
recurs. Although somewhat arbitrary, I felt that the most reasonable indicator of "a
recurring pattern of activity or behavior" (in the context of this study) would be whether
or not the same pattern of communications actions among team members (queries,
transmits, receives, text messages) was repeated for 3 consecutive trials after which the
majority of remaining trials used the same pattern.

Of the 12 pilot teams that ultimately got information to Charlie after the
disruption, only 7 showed evidence of developing new role structures. That is, it appears
that about half the teams developed coherent role structures after the communications
breakdown. Furthermore, it took these 7 teams an additional 8 trials (sd = 9) on average

to establish a new pattern of activity after they first figured out how to get Charlie his or

104

her information. Stated somewhat differently, it took approximately 20 trials on average
for teams to develop new role structures.

Of course, this leaves five teams that did not seem to establish a new role structure
(as I have defined it here) despite finding ways to inform Charlie. After examining these
teams' communications more closely, it appears that two patterns emerged in these teams.
First, there may have not been enough time for some teams to develop a new structure
during the scenario. Two teams (teams 015 and 019) found a way to inform Charlie rather
late in the scenario (trials 74 and 73 respectively). Because new role structure emergence
took 9 trials on average after teams first found a way to inform Charlie, it is possible that
these two teams would have developed new role structures had they had a few more
opportunities.

The second pattern that emerged was that, for some reason or another, some teams
used a slightly different pattern of activity from trial to trial. Members of two teams (team
013 and 014) apparently did not have a clear picture of the extent of the communications
breakdowns throughout the latter half of the scenario. Team 013 engaged in a great deal
of task related communication aimed at addressing the problem but never settled into a
stable new pattern of activity. In Team 014, members seemed to send information to one
another almost at random in hopes of the right team members getting the right
information after the disruption. Members of another team (team 016) engaged in a great
deal of social communication using text messages during the scenario and never settled
into a routine. This team figured out how to inform Charlie early (only after 3 disrupted
trials) and continued to do so throughout the remainder of the scenario (19 out of 30

trial), however, there was a great deal of seemingly random social communication that

105

interfered with the team developing a new role structure. I should also note that this social
communication appeared to be highly deleterious in terms of team performance. This
team was the third best in term of the number of times Charlie received his or her
information after the communications disruption, yet the team was also the third worst in
terms of decision accuracy.

Finally, teams that established a new role structure including a means of
informing Charlie performed better in terms of decision accuracy (TDA = .94) than teams
that did not establish such a role structure (1.30), p < .10, one tailed test.

Easy and difficult goals. I calculated the mean and standard deviation for the pilot
team’s decision accuracy (the absolute difference between the correct decision — Alpha’s
decision) for the trials preceding communications disruptions (trials 1 through 53). The
mean accuracy of team decisions for the twenty pilot teams was 1.03. That is, the pilot
teams made decisions that were a bit over one decision-level off on average. The standard
deviation for these decision were .33, and therefore the top 16% of the teams should have
decision accuracies of less than or equal to .70 (1.03 - .33). The bottom 16% of the pilot
teams should have decision accuracies of greater or equal to 1.36 (1.03 + .33). Because
participants received feedback according to the scoring_outlined earlier (2 points for O-off
accuracy, 1 points for 1-off accuracy, 0 points for 2-off accuracy, etc), the mean + lsd
goal would have been 1.40 for the difficult goal condition (2 - .70) and .64 (2 - 1.36) for
the easy goal condition. However, these goal levels were adjusted slightly upward for two
reasons. First, because the pilot teams were used for training experimenters and also for
working out problems (i.e., reducing ambiguity) in the training protocol, I felt that the

decision accuracy for teams in the primary experiment would be slightly higher than for

106

teams in the pilot study. Second, I felt that participants would be more likely to
understand and remember their goals if they were at “natural” _points on the scoring
continuum. Therefore, I chose 1.50 and .75 as levels representing difficult and easy
performance goals respectively.

Manipulation check. Each participant in the pilot study was given a “problem log
sheet” that had 2 columns. Trial numbers (1-83) were listed in column 1. Column 2 was
labeled “Problems with Communications” and included blank spaces for participant
notes. Participants were instructed to note the type of problem they were having with
communications in the blank corresponding to the trial when problems occurred. An
examination of these sheets revealed that Charlie station in each of the 15 pilot teams
(i.e., those individuals in teams not trained to adapt) noted a problem with the
communications from Bravo beginning on the 53" trial. Alpha and Bravo, however, did
not become aware of the problem until (a) Charlie either explicitly indicated he or she
was not receiving information from Bravo, or (b) the actions of Charlie caused Alpha or
Bravo to question Charlie about his or her status. These actions included Charlie waiting
longer to send a judgment to Alpha or Charlie not sending in a judgment at all.
Sum—mag of Pilot Study Results

Overall, the results of the pilot study indicated that teams did establish and
maintain a routine way of going about the experimental task. As team members gained
experience with the task, they accomplished the same pattern of activity more quickly and
more reliably. After a certain level of experience, the rate and reliability of task activity

stabilized.

107

The pilot study also supported the notion that some teams adapted and that
adaptation was beneficial in terms of promoting higher levels of team decision making
accuracy. First, the pilot data suggested that an appropriate means of dealing with the
communications breakdown should include finding a way to inform Charlie. Teams that
were trained in means of informing Charlie outperformed teams that did not find a way to
inform Charlie. Teams also performed significantly better once they found a way to
inform Charlie after the commrmications breakdown. Second, almost half the pilot teams
showed evidence of developing new role structures including a means of informing
Charlie in response to the communications breakdowns. Teams developing new role
structures performed better than those teams not developing new role structure. Among
factors that prevented teams from adapting seemed to be a general lack of awareness
about the situation, mindlessly or randomly trying new ways of doing the task, and high
levels of social communications.

Finally, the pilot study supported the use of communications disruption
manipulation. Charlie always noticed the communications disruption but the disruption
did not become apparent to the other team members until later.

The Primary Study: Variables

The primary study originally consisted of 489 participants arrayed in 143 3-person
teams. However, hardware failures during the experiment resulted in unusable data for
two teams. In one team, Charlie’s keyboard and mouse stopped working on trial 11, and
thinking this was part of the experiment, did not tell the experimenter. In the team, a

network crash late in the game (trial 74) resulted in the loss of data prior to that point in

108

the experiment. These two teams were removed from further analysis resulting in a
sample of 141 teams with no missing data.

This study used 6 independent variables in a between groups design. Goal
difficulty (easy vs. difficult), quality of feedback (timely and specific versus delayed and
summarized), and the nature of the environmental disruption (abrupt vs. gradual) were
manipulated. Team cognitive resources, team conscientiousness resources, and team
openness resources were operationalized as continuous variables. Teams were composed
randomly and as a result there were mixed as well as single gender teams. Before I
describe these independent variables, I will describe how the primary outcome variable,
adaptation, was operationalized.

Adaptation. As mentioned in the procedures section, teams were shown how to
exchange information after initial training so that each team member received all the
information they needed to accomplish their role responsibilities. This communications
structure is depicted in Figure 7. After the communications disruption, however, team
members had to use a different structure to get Charlie his or her information (“RCA”).
As the pilot study demonstrated, any new role structure should include informing Charlie.
Given that the task was structured such that Bravo and Charlie could not communicate
using free form text messages, and since only the station that originally measured the
attribute could use the transmit fimction to send the attribute, the most efficient adapted
structure is illustrated in Figure 7. This is the structure the five "trained" pilot teams used.

after the communications breakdown.

109

 

Alpha

 

 

 

 

Corridor Status peed
RCA RCA
Charlie (text (transmit) Bravo
message)

 

 

 

 

 

Figure 7 - Ideal Adapted Role Structure

In this structure Bravo transmits “RCA” to the Alpha, who in turn, sends a free
form text message with the “RCA” values to Charlie. This structure has the fewest
inefficiencies in terms of extra key-strokes. For instance, if Bravo used the text message
function instead of the transmit function to send the “RCA” value to Alpha, he or she
would have to do more keystrokes. In order to use the text ftmction Bravo would have to
select the text function, select Alpha as the station to send the message, type characters
for the “RC8” value, and finally, select enter. On the other hand, if Bravo used the
transmit function, he or she would only have to select the transmit function, select Alpha
as the station to send the message, and finally, select enter. In addition to saving
keystrokes, the transmit function does not require the actual attribute values (e. g.,
“RCA”=5,725) to be held in Bravo’s short term memory.

However, while the experimental task allowed for the ideal adapted structure (in
terms of keystroke efficiency) to be known in advance, there was an issue as to whether
deviations from this structure actually detracted from teams’ performance. If deviations

from the ideal structure did not negatively influence team performance, then I could not

110

say that the structure was ideal. This, of course, would imply that teams adapt when they
simply establish a role structure that includes means of informing Charlie.

In order to assess the extent to which the structure in Figure 7 is ideal, I created a
variable called Adapted Inefficiency and correlated it with the accuracy of teams’
decisions after the communications disruption (WW
Accuracy). I created the Adapted Inefficiency variable by first examining the pattern of
communications used in new role structures (i.e., when the same pattern of
communications actions among team members was repeated for 3 consecutive trials after
which the majority of remaining trials used the same pattern) after the communications
breakdown. Adapted Inefficiency was then calculated for each team as the number of
extra actions (transmits, receives, queries, and text messages) in the team’s structure
above that required to perform the task as illustrated in Figure 8.

For the 75 teams that eventually developed a new role structure, Adapted
Inefficiency was 1.03 on average sd = 1.56). That is, teams in the experiment developed
new role structures that were inefficient by a little over 1 action. The correlation between
Adapted Inefficiency and Post-disruption Team Decision Accuracy was not significant, p
= -.02, p = .84, however. This suggested that inefficient role structures did not detract
fi'om team performance in this task. That is, effectively adapting in this task, did not
necessitate using the ideal structure illustrated in Figure 7. As suggested by the analysis
and the results from the pilot study, adaptation simply required members to establish a
new role structure that included a means of informing Charlie. Accordingly, adaptation
was initially operationalized as the number of trials in the disrupted situations where

teams used a new structure to inform Charlie.

111

As mentioned earlier, evidence of a new role structure should suggest that a new
pattern of activity or behavior among team members recurs. I felt it was reasonable to
characterize a recurring pattern of activity as a pattern of communications actions among
team members (queries, transmits, receives, text messages) that was repeated for 3
consecutive trials after which the majority of remaining trials used the same pattern. I
examined the pattern of each team’s communications after the breakdown to identify
when a new pattern began. Adaptation was calculated as the number of remaining trials.

Goal difficulty. The goal specificity manipulation followed the first training trial.
Research participants in the difficult goal condition were told: “Even though you all have
different roles, you should work as a team to achieve a goal of at least 1.50 as an average
score”. This goal level was determined in the pilot study described above. Research
participants in the easy goal condition were told: “Even though you all have different
roles, you should work as a team to achieve a goal of .75 as an average score.” These
instructions were repeated once more directly following the final training trial. After the
experiment, participants were given two questions (Appendix G items 1 &2) about the
difficulty of their assigned goal (1=very easy to 5=very difficult) to serve as a
manipulation check for goal difficulty. The reliability of this scale was .73, and as
expected, individuals in the difficult goal condition though their goal was more difficult
M = 3.74, SD = .92) than individuals in the easy goal condition (M = 2.88, SD = 1.00),
t(421) = 9.25, p < .001.

Mia of Feedback. Prior to target 41, all teams received the information on the
outcome of the trial as well as the average team-level performance immediately following

each trial. This information was displayed on each members’ monitor. The average team-

112

level performance was displayed next to the team’s goal. After target 40, teams in the low
quality feedback condition received only average (all preceding trials) team level
performance feedback every 9 trials. Participants in this condition were told during
training that “as in many situations in real life, you will not always get immediate
feedback on how well your team is doing. After trial 40 you will get feedback after every
9 trials”. Teams in the immediate feedback condition continued to receive information on
the outcome of each trial as well as average team-level performance. After the
experiment, participants were given two manipulation check questions (Appendix G
items 3 & 4) asking how much performance feedback they received during the task
(1=very little to 5=very much). The reliability of this scale was .66, and as expected,
individuals in the immediate feedback condition though they were more informed about
their standing (M = 3.41, E = .96) than individuals in the delayed feedback condition
(M = 2.79, $ = .96), t(421) = 6.63, p < .001.

Abruptness of environmental disruption. As explained earlier, the environmental
disruption was accomplished through changes in the manner in which information could
be transmitted. The first 43 trials were the same for all teams. All of these trials had the
full communications compliment. For teams in the abrupt disruption condition, the next
10 decisions also had the full communications compliment. However, on the 54th trial,
the communications breakdown occurred (the removal of the transmit link between Bravo
and Charlie) and did not change back for the remaining trials. The teams in the gradual
disruption condition, however, experienced periodic breakdowns beginning on trial 44.
The pattern of these breakdowns were such that they occurred with increasing frequency

(trials 49, 50, 54, 55, 56, 59, 60, 61, 62, and 64-83). That is, after decision 43, there were

113

fewer and fewer decisions in which the team had its full communications complement (4
decisions between 44 and 49, 3 decisions between 50 and 55, 2 decisions between 56 and
59, and 1 decision between 62 and 64). Beginning on decision 64, the disruption occurred
on every decision.

The structure of this manipulation was dictated by a number of considerations.
First, there needed to be adequate time for teams in both conditions to develop routines.
This was facilitated by training teams in their initial role structure and giving them forty
trials in which this structure was appropriate. M. Cohen and Bacdayan (1996) found that
evidence that routines (e. g., increased reliability, speed, repeated action sequences, sub-
optimality in different circumstances) could be developed within 40 hands in a 40 minute
card game. This evidence, together with the results from the pilot study described above,
suggested that 40 trials was enough to develop routines.

Second, the environmental change in both conditions needed to commence with a
significant amount of time remaining in the scenario. From a practical standpoint, teams
might not invest effort in adaptation if the shift occurred too late in the scenario simply
because such effort would have lower utility (i.e., after developing a new system of
activity there will be fewer trials in which the new system can be used). I felt that the
points that could accrue over 30 trials would be fairly significant to teams.

I should note that there was the possibility that teams in the abrupt condition
might develop “stronger” routines (and thus be less prone to adapting) because they had
more experience with the task before any communication breakdowns occur. However,
three factors mitigate this concern. First, as suggested in the pilot study, teams should

approach asymptotic levels of routinization prior to the communication breakdown.

114

Second, there were only 3 trials (trials 44, 49, 50) in which breakdowns occur in the
gradual condition prior to any breakdowns in the abrupt condition (trial 54). Finally, in
order to allay concerns about this issue, I ran twenty teams in a condition where there was
an abrupt communications breakdown beginning on trial 44 and ending on trial 73. There
were no statistically significant difference in Adaptation or Post Communications
Disruption TDA between these twenty teams and the other teams in the abrupt
conditions. Thus, it is not likely that differences between teams in the gradual and abrupt
conditions can be attributed to differences in the strength of their routines prior to the
communications disruption.

After the experiment, participants were asked to rate the unreliability of
communications (0-100%) between stations during different phases of the experiment on
a 7-point scale (Appendix G items 11-34). I acknowledge that the items did not include
the words “abruptness”, and thus, the items and the label I chose for the construct seem to
be a mismatch. However, I felt that participants experiencing the manipulation would be
better able to respond to their experience in terms of “unreliability of communications”
rather than “abruptness of communications disruption”. That is, I felt that it would be
better to match items to what the participants experienced than to a label that may be
imperfect to the extent that it did not perfectly match the manipulation. I felt that these
questions about the unreliability of communications during the period at which the
change was occurring would serve as a reasonable indicator of individuals’ perception
about the abruptness of the communications disruption manipulation. The more abrupt
the disruption, the more the participants should perceive that their communications were

unreliable. Individuals in the 20 teams experiencing an abrupt disruption beginning on

115

target 43 thought that communications between Bravo and Charlie across trials 44-53
were less reliable M = 4.03, E = 2.14) than individuals in the teams in the gradual
condition M = 3.48, S_D = 1.93), t(251) = 1.87, p < .10. Individuals in the teams in the
condition where the abrupt communications disruption began on trial 54 also perceived
that the communications between Bravo and Charlie across trials 54-63 were less reliable
M = 4.79, S1; = 2.07) than individuals in the teams in the gradual condition M = 4.08,
S_ = 2.02), t(361) = 3.28, p < .001.

Tea_rp cpgpitive resources. As stated earlier, cognitive resources can be thought of
in terms of individual differences in general cognitive ability or g. In this study, general
cognitive ability were measured by scores on the Wonderlic Personnel Test (Form IV).
This test has a great deal of support in the educational and psychological literatures as a
reliable (Dodrill, 1983; McKelvie, 1989; Weeless & Serpento, 1982) measure of the same
verbal, quantitative and spatial abilities that indicate g (Hunter, 1986; Jensen, 1977).
Internal consistency reliabilities have ranged from .88 to .94 across forms of the
Wonderlic (Wonderlic & Associates, 1983). Since consistency in terms of cognitive
resources was not expected or specified within groups, no measure of team-level
reliability is appropriate. Consistent with the rationale outlined earlier, team cogg'tive
resources were operationalized three ways: as an average of team member scores, as the
lowest team member’s score, and finally, as the highest team member’s score. These
Operationalizations correspond to the additive, conjunctive and disjunctive models.

Tm amigttiousness resources. Individual conscientiousness scores came from

the long version of the Revised NEO Personality Inventory (NEO PI-R) (Costa &

McCrae, 1992) which was administered when participants initially signed-up for the

116

study. This particular conscientiousness scale includes 48 items capturing six sub-factors
thought to underlie the construct (competence, order, dutifulness, achievement striving,
self-discipline, and deliberation). Internal consistency reliabilities for this measure of
conscientiousness have generally been in the low.90’s (McCrae, 1987). This study was no
exception (Cronbach’s alpha = .90). Since consistency in terms of team conscientiousness
resources was not expected or specified within groups, no measure of team-level
reliability is appropriate. Consistent with the rationale outlined earlier, te__a_m_
copgientiousness resoggps was operationalized using the additive, conjunctive, and
disjunctive models--that is, as an average of team member scores, as the lowest team
member’s score, and finally as the highest team member’s score.

Team omnness resources. Individual openness to experience scores also came
from the Revised ‘NEO Personality Inventory (NEO PI-R) (Costa & McCrae, 1992). This
scale includes 48 items capturing six sub-factors thought to underlie the construct
(fantasy, aesthetics, feelings, actions, ideas, and values). Internal consistency reliability
have ranged from the mid .80’s to the mid .90’s (Mch 1987). Cronbach’s alpha in this
study was .88. Since consistency in terms of team conscientiousness resources was not
expected or specified within groups, no measure of team-level reliability is appropriate.
As with team cognitive resources and team conscientiousness resources, team oxnness
resources was operationalized three ways: as an average of team member scores, as the
lowest team member’s score, and finally as the highest team member’s score.

Team Decision Accuracy. Although not stated as a formal hypothesis, it is
important to assess the extent to which adaptation leads to higher levels of team

performance after the communications breakdown. In addition, it would be informative to

117

assess the extent to which the independent variables relate to team performance prior to
and after the communications breakdown. Accordingly, I calculated two decision
accuracy variables. Pre-disruption Team Decision Accuracy (Pre-TDA) is the average
absolute difference between the teams’ decisions and the correct decisions for the non-
training targets prior to target 44. Post-disruption Team Decision Accuracy (Post-TDA)
is the average absolute difference between teams’ decisions and the correct decisions for
the targets where communications were disrupted.
Th; Primgy Study: Analyses

Analyses were conducted in three stages. In the first stage, I assessed (a) the
extent to which high levels of cognitive resources resulted in higher levels of knowledge
about the task and situation, and also (b) the extent to which goals, feedback, and the
nature of the communications disruption influenced the allocation of cognitive effort
toward the task after the disruption. In the second stage, I examined descriptive statistics
of the study variables with particular emphasis on the distribution of adaptation. Finally, I

conducted analyses aimed at addressing the hypotheses and research questions.

118

Chapter 3

RESULTS OF PRIMARY STUDY

Overview of Rethp

The results of the primary study are presented as follows. First, I present an
analysis on the relationship between individuals’ cognitive resources and their learning.
Second, I present descriptive statistics on the team-level variables after which I describe
the analytic strategy. Fourth, I describe the results of the hypotheses organized in terms of
those relevant to (a) team resources (cognitive ability, conscientiousness, and openness)
(b) the manipulations (goal difficulty, quality of feedback, abruptness of the disruption),
and finally, (c) the interactions. Fifth, I explore the effect of alternative
Operationalizations of the compositional variables. Finally, I conclude with an exploration
of another dimension of adaptation.
Individual-level Learning and Controlled Cogg'tion

Because the literature on human information processing, skill acquisition,
individual differences, and control theory supports the underlying process assumed to be
involved in the proposed model at the individual-level, I did not present aptjpp'
hypotheses in regards to the effects of individual differences and the manipulations on
those processes (i.e., individual-level learning and the triggering of controlled
information processing). However, because these processes are the underpinnings of the
model, I felt that it was important to ensure that they were at work in this particular
research setting. In this section, I present data addressing two questions. First, did

individuals with higher gleam more about the task itself and have more awareness of the

119

team’s situation after the communications breakdown? Second, did the manipulations
(diffith goals, full feedback, abrupt disruptions) trigger higher levels of active thought
about the task? An individual-level focus is appropriate for these analyses because
learning that results in a change in the pattern of activity among team members takes
place in the minds of the individuals who compose the team. Such learning does not have
to be shared among members in order for adaptation to occur (Fiol & Lyles, 1985;
Hedburg, 1981; Hutchins, 1996).

The com’tive resources-learning relationship. In order to assess the extent to
which high levels of cognitive resources resulted in a larger pool of knowledge regarding
the task and situation (assumed to precede the behavioral changes associated with team
adaptation), I examined the relationship between individuals’ cognitive resources (as
measured by individuals’ score on the Wonderlic Personnel Test) and declarative
knowledge acquired during training. Task knowledge was measured by a 6-item
instrument developed for this study and was administered immediately after training. As
expected, there was a positive relationship between an individuals’ score on the
Wonderlic and the task knowledge test, I = .16, p < .01 (p = 423). This statistically
significant relationship suggests that those with higher levels of cognitive resources
learned more during training and their initial experience with the task than those with
lower levels of cognitive resources.

It was also important to examine the extent to which higher levels of cognitive
resources enabled individuals to have more knowledge about the team’s situation during
the experiment (situational awareness). The most relevant aspect of the situation was the

reliability of the communications between each dyad (i.e., Alpha-Bravo, Alpha-Charlie,

120

Bravo-Charlie), therefore, the measure of situational awareness was based on this aspect
of the situation. After the experiment, individuals were asked to rate the unreliability of
communications during different phases of the scenario using items developed for this
specific study (Questions 11-34, Appendix G). While these self-reports suffer from the
weaknesses of being retrospective, asking participants during the simulation would have
created a priming effect. I should also note here that some of these items were used as a
abruptness manipulation check. However, the measure of situational awareness included
only some items related to the communications between Bravo and Charlie. In addition,
these items are only one component of the situational awareness measure. Specifically,
the measure of situational awareness was created by subtracting individuals’ ratings of
communications unreliability for each series of targets from the true unreliability of that
series of targets, and then averaging across each series (series 1= targets 43-53, series 2 =
targets 54-63, series 3 = 64-73, series 4 = targets 74-83). Creating such an index was
necessary because individuals in different conditions experienced disruptions at different
points in time. As expected, individuals with higher levels of cognitive ability had more
situational awareness, I = -.11, p < .05 (lower scores on the situational awareness variable
indicate higher accuracy in terms of knowledge of the situation) than those with lower
levels of cognitive ability (p = 423).

The manipulations-controlled cogpt'tion relationships. In order to ensure that the
manipulations triggered controlled cognition intended to overcome the communications
breakdown, I examined the relationship between scores on a self-report cognitive effort
scale (items 9 and 10 in Appendix G, Cronbach’s alpha = .66) and the goal, feedback, and

environmental disruption manipulations. As with the measure of situational awareness, I

121

felt that the potential problems caused by priming effects outweighed the weaknesses of a
self-report retrospective measure. After controlling for the effects of the individual
differences (cognitive resources, conscientiousness, and openness to experience), two of
the three manipulations resulted in higher levels of self-reported cognitive effort.
Significant (one tailed) partial correlations indicated that the goal as well as the
abruptness manipulation resulted in higher levels of self reported cognitive effort (t = .08,
p < .05 and I = .11, p < .05 [n = 423], respectively). There was not a statistically
significant relationship between quality of feedback and cognitive effort, however.
Summgg. Overall, the results these analyses were fairly supportive of the
assumed effects of the independent variables on individuals’ acquisition of declarative
knowledge and triggering of active or controlled information processing. The only
relationship that failed to indicate support for the implied process was that between the
feedback manipulation and cognitive effort. That is, there was not a significant difference
in the amount of self-reported cognitive effort between individuals who received timely
and specific feedback from those who received delayed and summarized feedback. As
pointed out earlier, individuals in the high quality feedback condition did perceive that
they were more informed about their standing than individuals in the delayed feedback
condition. One explanation for this unexpected finding is that individuals in the low
quality feedback condition had to expend relatively more cognitive effort in order to self-
regulate in the new environment (monitor and evaluate their standing). That is, while
there was no difference in the amount of cognitive effort expended, the focus of the effort
was directed differently for those in the different feedback conditions. If this is the case,

then individuals in the high quality feedback condition may have expended more

122

cognitive effort trying to devise means of overcoming the breakdown, while individuals
in the low quality feedback condition expended more cognitive effort trying to figure out
where they stood. An alternative explanation for the lack of a feedback effect is that while
there were differences in perceived amount of information provided by feedback, both
conditions provided enough information to trigger active thinking.
Descriptive Statistics of Team Level Vagabjgs

Descriptive statistics for the study variables at the team level of analysis (n = 141)
are presented in Table 3. The first two columns of data list means and standard deviations
of the team-level variables. The columns numbered 1-15 list zero order correlations.

Team Resources. The continuously measured individual difference variables are
listed in the first 9 rows of the matrix (three characteristics aggregated three different
ways). Across characteristics, the means for the conjunctive operationalizations are
approximately one standard deviation less than the additive and two standard deviations
less than the disjunctive operationalizations. Alternative operationalizations of the same
characteristic are more highly related to one another than to the other characteristics.
Correlations between the additive and other operationalizations (conjunctive and
disjunctive) were high (.70-.80) while the correlations between the conjunctive and
disjunctive operationalizations were all more moderate (.24-.30).

Manipulations. Correlations between the three manipulations are all near zero.
Only one of the twenty-seven correlations between the manipulations and the measured
characteristics reached even a marginal level of significance and the average correlation
was small (r = .02). Overall, therefore, it appears that the random assignment to

conditions worked as intended.

123

2.828.832":

 

 

.8. .8.- 8. .8- 8.- .2- .2.- .2.- 2. 8.- 8. 8.- .8. .8- 8. 82 <E-aoa a:
-- .8- 8.- 2.- 8. 8.- 8. 8. 8. 8. 8. 8.- .8.- .8.- 8. 8. «.88-25 9.:
-- .8. 8. 8.- 8. S. 8. 8.- 8. 8.- .8. .8. .8. 8.8 8.8 82.88... 8:
-- 8.- 8.- 8.- 8. 8. 8.- 8.- 8.- 8.- 2. 2. 8. 8. 883.288.8588.... a:
-- 8. 2. 8. .2. 8.- 8. 8. 2.- 2. 8.- 8. 8. 88888830 2:
-- 8. 2. S. 8. 8. 8. 8.- 8.- 8.- on. 8. 8385 woo 8:
-- .8. .8. 8. :.- 8.- 8. .2. .2. R. 8.... .8580 88.3.8.5 av
- .8. 8.- 8. 8.- .2.- 8.- :.- 8. 2.... .8880 38.5.88 88
-- 8.- 8.- 8.- 8.- 8. 8.- 8. :8 .8880 2.83. 6

80550356260
-- .8. .8. 8.- .3- .z- 8. 8.8 83.2339 88

mmocmflomucomoma—OU
-- .8. 8.- 8. 8.- 8. 2..“ 2.8.3.88 5

mmoemsoueomomeoo
- 8.- 8.- 8.- 8. 88 03.83. E
-- .8. .8. 8.8 8.8 382589 5
-- .8. 8.... 8.8 3805.85 E
- 88 8.8 3883. E

8: A28 8: :3 8: 8 as E 88 A2 A8 5 5 8 am 5...:

 

880,—. a8 murmuflm Dermataom - m 033-

124

Mm Although I made no explicit hypotheses regarding effects of
study variables on team performance (i.e., TDA), effects of some variables have been
found in past research (e.g., Barry & Stewart, 1997; LePine, et al., 1997). In addition, the
definition of adaptation requires that it be beneficial in terms of promoting higher
performance. Therefore, I calculated two measures of team performance and included
them in Table 1: Decision accuracy prior to and after the communications breakdown
(pre-TDA and post-TDA).

As expected, teams performed significantly worse after the communications
breakdown (pm-disruption TDA = .63 versus post-disruption TDA = 1.08). Also, as
expected, the two performance measures were significantly correlated (r = .55, p < .05).
However, the moderate size of this relationship suggests that the teams doing well (or
poorly) prior to the disruption were not necessarily the same teams doing well (or poorly)
after the disruption. Thus, performing well in stable conditions did not guarantee
performing well in changing conditions. Table 3 also suggests that teams performing well
tended to adapt (r = -.27, p < .05) and that adaptive teams had higher performance after
the communications breakdown (r = -.47, p < .05). Overall these results suggest that
adaptation is not the same thing as performance and that adaptation may promote higher
team performance in changing environmental conditions.

There were also significant correlations between the additive and conjunctive
operationalizations of Team cognitive resources (g) and TDA, both prior to and after the
communications breakdown. Teams composed of members who had high g (on average
or in a conjunctive sense) made better decisions on average than teams composed of

members with low g. All three operationalizations of team openness resources had

125

marginally significant relationships with decision accuracy after the communications
breakdown. Teams apparently made more accurate decisions when they were composed
of members who were open to experience. Although there was not a significant
relationship between the feedback intervention and adaptation, teams in the timely
feedback condition made more accurate decisions after the communications breakdown (r
= -.l8, p < .05).

Adaptation. Of the independent variables in this study, only team cognitive
resources (all three operationalizations) and abruptness related significantly to adaptation.
Teams composed of individuals with higher g, had higher levels of adaptation than teams
composed of individuals with lower g. Teams experiencing abrupt communications
breakdowns had higher levels of adaptation than teams experiencing gradual
communications breakdowns.

Based on an examination of histograms, all variables appearing in Table 3 with
the exception of adaptation had distributions that appeared normal. Adaptation had a
strong positive skew. Skewness is an index of the tendency of deviations from the mean
to be concentrated in one direction or the other. Values of 0 indicate normality and
positive (negative) values indicate a positive (negative) skew. The skewness statistic for
adaptation was .81, se = .20. The underlying reason for this skewness is that 66 teams
(47%) had adaptation scores of 0. That is, about half the teams in this study failed to
adapt. This finding, together with the fact that adaptation is not truly continuous (can only
take on integer values of 0 and 3-30), is of significance because it influences the choices
of data analytic strategies. In particular, modeling this variable using linear regression

will violate the assumption of heteroscedasticity. The expected values or predictors and

126

the absolute values of the residuals will be related to one another, thereby distorting the
estimated standard errors of the regression coefficients, and thus, distorting the t tests for
those coefficients (Fox, 1991). In addition, the use of linear regression with such data
often generates a substantial number of out of range predicted values (i.e., values that are
less than 0 when the minimum possible value for the criterion is 0) (Fox, 1991). This, of
course, will undermine confidence in the direction and size of estimated parameters.
Analytic Strategy

After reconsidering the definition of adaptation proposed earlier (reactive and -
non-scripted adjustments to a team’s role structure that result in a better fit given the new
situation) it became apparent that it would be more appropriate to operationalize team
adaptation as a dichotomous variable (either a team adapts or it does not adapt) instead of
a count reflecting how long it took teams to adapt. The reason for this is because there is
nothing in the definition of adaptation that implies speed or quickness to adapt. That is,
adaptation operationalized as a count confounds adapting with smed of adapting. This is
not to imply that the latter concept is unimportant, but that the primary focus of this
dissertation is on the former. Thus, I created a new dichotomous variable for adaptation.
Teams that did not establish an adaptive role structure after the communications
disruption (establishing a way to get Charlie RCA using the same pattern of
communications after which the majority of remaining trials used the same pattern) were
scored 0 while those establishing a new role structure after the disruption were scored 1.

Table 4 shows correlations between the study variables and this new
operationalization. Column 1 lists correlations with the new dichotomous

operationalization and column 2 lists relationships with the original count variable.

127

Table 4 - Correlations of Study Variables with Dichotomous Adaptation

 

 

Adaptation Adaptation

(dichotomy) (count)
Adaptation (count) .81 "'
Additive g .36“ .39*
Conjunctive g .30“ .32"
Disjunctive g .18“ .22‘
Additive Conscientiousness -.05 -.02
Conjunctive Conscientiousness .04 .05
Disjunctive Conscientiousness -.1 l -.09
Additive Openness .14 .08
Conjunctive Openness .09 .10
Disjunctive Openness .18 * .09
Goal Difficulty -.05 -.03
Quality of Feedback .00 .02
Abruptness of Disruption ' .32" .28“
Pre-TDA -.28* -.27"‘
Post-TDA -.39"' -.47"‘

 

n=141;*p<.05

Not unexpectedly, the relationship between the two operationalizations is high (r
= .81, p < .001). I also note that the relationships between the dichotomy and study
variables are very similar to those obtained for the count variable. Thus, while the two
operationalizations are not perfectly related to one another, they do seem to be capturing
much of the same phenomenon. However, it is important to note that mathematically,
relationships with a dichotomous variable should be lower than with the same variable
operationalized continuously. In fact, the magnitude of a relationship with a dichotomous
variable can only be 80% of the magnitude of a relationship with a continuous variable.

However, for several of the correlations in Table 4 this does not seem to be the case. The

128

most obvious example of this is that the correlation between the disjunctive
operationalization of openness and the dichotomy is twice as high as with the continuous
variable. What this data suggests, therefore, is that while the two operationalizations of
adaptation are capturing something similar, they are also capturing something unique.
This provides some empirical support for the use of the dichotomy.

Given that adaptation is now operationalized as a dichotomy, the appropriate
method of analysis is logistic regression. The use of linear regression is problematic in
terms of analyzing dichotomous dependent variables because of inherently
heteroskedastic error variances and the lack of constraints preventing predicted values
from falling outside of O and 1 values (DeMaris, 1995; Wright, 1995). Discriminant
analysis stands as a possible technique for analyzing dichotomous outcomes (Press &
Wilson, 1978), however, it is sensitive to violations of multivariate normality of
independent variables and unequal variance-covariance matrices in the two groups.
Because several of the independent variables are categorical, the data in this study will
not be multivariate normal.

As opposed to linear regression where one attempts to predict scores on a
continuous dependent measure, logistic regression attempts to predict the probabilities
that observations belong to mutually exclusive groups. As in linear regression values are
predicted:

g = be + h. (8).
however, these predicted values are inserted into an equation that gives the predicted

probability of group membership (5):

129

.1! = 95 / (1 + 25)-

An observation is assigned to the target group if the predicted probability is greater than
or equal to .50. Because least squares solutions will not generate solutions for this
nonlinear model, maximum likelihood is used to yield estimates for unknown parameters
that maximize the probability of obtaining the observed data. As in linear regression,
logistic regression output includes b weights, standard errors, and information to assess
overall model fit.

Whereas in linear regression where b’s are changes in X for a unit change in x,
b’s in logistic regression are changes in the natural logarithm of the odds ratio for a unit
change in _X_. The odds ratio indexes the change in the odds of membership in the target
group for a 1 unit change in X. Thus, odds ratios can be expressed as the natural log of b,
that is, 93. For example, in a study where b = .8611, the odds ratio equals g“” or 2.3657.
This in turn can be interpreted as “the odds of X are 2.37 greater for 291 than for X.” It
should be noted that odds are interpreted as the probability of membership in one group
divided by the probability of membership in the other group, and thus range from 0 to
infinity (probabilities range from 0 to 1). The sign of the b’s indicates whether odds
increase (positive 12’s) or decrease (negative b’s) as the predictor increases. Dividing the
b’s by their standard errors yields a z that can be used to test the statistical significance of
individual parameters.

Overall model fit can be assessed a number of ways. One indicator of model fit is.
the percentage of cases that were predicted correctly from the model. This number can be

compared to the number of cases that can be predicted by chance (i.e., by using the mean

130

to estimate the probability of being in the target group). Because classification decisions
are made dichotomously (based on probabilities 2 or < .5), this index does not reveal the
distribution of estimated probabilities. That is, there is no indication of the estimated
probabilities of individual cases (i.e., those assigned to group 0 should have a small
estimated probability of being a 1 and those assigned to group 1 should have a large
estimated probability of being a 1), nor the overall sample.

A second indicator of model fit is based on the probability of observed results,
given the parameter estimates. This likelihood is generally less than one so custom
dictates the use of -2 times the log of the likelihood as a measure of how well the model
fits the data (-2LL). Differences between the -2LL with the constant only and with the
predictors indicate how useful the set of variables are in predicting group membership.
This difference is distributed as E with degrees of freedom equal to the number of
predictors. This improvement statistic is comparable to the E-change test in multiple
regression. Differences in blocks of variables can also examined using differences in -
2LL’s just as in hierarchical regression analysis. The degrees of freedom in this analysis
equals the difference in the number of parameters included in the model.

Given the hypotheses outlined earlier, the data were analyzed hierarchically and
the results of this analysis are shown in Table 5. In step 1, the model includes the
aggregated (additively) individual differences (team cognitive resources, team
conscientiousness, team openness). In step 2, I included the manipulations (goal
difficulty, quality of feedback, and the nature of disruption). Because of the large number
of possible interactions, each of the 6 2-way interactions was assessed in step 3 in a

separate run. Finally, the three way interaction was entered in a model including only the

131

necessary 2-way interactions. The continuous variables were standardized to reduce the
multicollinearity between the predictors and their product terms.
Assessment of Hymtheses Using Logistic Regression

Table 5 shows that depending on the predictors included in the model, between
65.25% and 74.47% of the cases were predicted correctly as belonging to either the
adaptor or non-adaptor groups. Using no predictors, 50.20% of the cases would have been
predicted correctly (the intercept only model). Table 5 also shows that the overall model
at each hierarchical step was significant. That is, the inclusion of the blocks of predictors
increased the fit of the model to the observed data. First, I present results related to the
hypotheses for the main effects of team resources (H1, H6a. H7a). Second, I present
results related to the hypotheses for the main effects of the manipulations (H2a, H3a,
H5a). Finally, I present results related to the hypotheses for the interactions (I-IZb, H3b,
H4a, H4b, H5b, H6b, H7b).

Team resources. As a block, the aggregated (additive) individual difference
variables significantly increased the predictability of team adaptation, X2 = 23.09 (3dr). In
support of H1 , the Team Cognitive Resources predictor was significant, ; = 4.10, p < .05.
The odds ratio indicates that for each standard deviation increase in team cognitive
resources, the odds of adapting are 2.37 times greater (i.e., 9'“). By multiplying the
regression coefficient by a desired level one can also calculate the odds ratio for a
different size increase. For example, the odds of a team adapting with a score for team
cognitive resources of +1 standard deviation are 5.60 times higher than a team with a

score of -1 standard deviation (92 " ‘36).

132

Table 5 - Logistic Regression of Additive Independent
Variables on Adaptation

 

% Model X2 Block X2

 

 

Step Variable Correct (df) (df) b SE 2 eb
(0) Constant Only 50.20
(1) Team Cognitive 65.25 23.09(3)* 23.09(3)* .86 .21 4.10"“ 2.37
Resources (Tg)
Team .00 .19 .00 1.00
Conscientiousness (Tc)
Team Openness (To) .35 .19 1.82' 1.42
(2) Goal Difficulty (GD) 70.92 35.13(6)* l2.05(3)* -.16 .39 -.44 .84
Quality of -.02 .39 -.05 .98
Feedback (F B)
Nature of 1.29 .39 3.31 * 3.63
Disruption (ND)
(3a)“ Tg x GD 72.34 42.02(7)* 6.88(1)* -1.24 .51 -2.43* .29
(3b) Tg x FB 70.21 3588(7)" .75(1) -.41 .48 -.85 .66
(3c) GD x FB 70.21 35.54(7)* .41(1) .50 .78 .64 1.65
(3d) Tg x ND 71.63 35.99(7)* .85(1) -.45 .50 -.90 .64
(3e) Tg X To 73.05 42.34(7)* 7.21(1)* .69 .27 2.56" 1.99
(3f) Tg x To 73.76 42.33(7)* 7.19(1)* -.61 .26 -2.38"I .54
(4)b Tg x GD x FB 74.47 47.78(10)* 4.99(1)* -2.53 1.22 -2.07* .08
n = 141

' Step 3 results are from separate runs.

" Only interactions 3a, 3b, 3c entered prior to step 4.
"' p < .05

‘ p < .10

133

The data in Table 5 do not indicate support for H6a, however. That is, there was
not support for the notion that teams with high levels of conscientiousness resources
would be more likely to adapt than teams with low levels of conscientiousness resources.
Finally, there was marginal support for Hypothesis 7a, teams with higher levels of
openness resources were more likely to adapt than teams with low levels of openness
resources, g = 1.82, p < .10. The odds ratio indicates that the for each standard deviation
increase in openness resources, the odds of adapting are 1.42 times greater. The odds of a
team adapting with a score for openness resources of +1 standard deviation are 2.01 times
higher than a team with a score for openness of -1 standard deviation.

Manipulations. Table 5 illustrates that as a block, the manipulations significantly
increased the predictability of adaptation over and above the effects of the aggregated
individual difference variables, , X} = 12.05 (3df), however, only the nature of the
disruption had a significant effect, , z = 3.31, p < .05. That is, while the goal difficulty
and quality of feedback manipulations did not have significant effects on team adaptation
(HZa and H3a respectively), the nature of the environmental disturbance disruption did
(Hypothesis 5a). The odds of teams in the abrupt condition adapting were 3.63 times
higher than teams in the gradual condition.

Interactions. The significance of an interaction term provides an indication that
the effect of one predictor on adaptation depends on the level of another. However,
statistically significant interactions must be plotted to ensure the pattern of the
relationships is consistent with that hypothesized. Table 5 shows that each of the models
including 2-way interactions was significant, however, half of the hypothesized 2-way

interactions were not supported (H3b, H4a, H4b, H5b). Significant interaction terms were

134

plotted (using median splits for the continuous variables) on the percentage of cases
adapting in each cell.

The team cognitive resources by goal difficulty interaction was significant, 2 =
2.43, p < .05. However, as shown in Figure 8, the pattern of the data was inconsistent
with H2b. That is, the relationship between team cognitive resources and adaptation was

higher in teams that had easy goals rather than in teams that had difficult goals.

 

f llEasyGod 5
l.DifficdtGoal;

 

 

 

 

Team Cog'ltlve' '
Resources (Additive)

Figure 8 - Plot of Team Cognitive Resources by Goal Difficulty Interaction

One way to index the magnitude of the relationships over levels of the moderator
is to examine the two-way frequency tables (e.g., frequency of adapting by levels of team
cognitive resources) and calculate phi coefficients. The relationship between team
cognitive resources (dichotomized) and team adaptation for teams in the low goal

condition was phi =.55 (p < .05), whereas that in the high goal condition was phi = .07,

135

ns. It is noteworthy to mention that having low goal levels was especially helpful to
teams having high levels of cognitive resources (82% of the teams with high cognitive
resources and easy goals adapted).

The team cognitive resources by team conscientiousness resources interaction was
also significant, ; = 2.56, p < .05. As illustrated in Figure 9, the pattern of the data were

consistent with Hypothesis 6b.

flLOW I

Conscientiousness i
' I High ‘3
Conscientiousness i

Teams 0.4.
Adapting 0.3 '
0.2x

0.1

o_

 

 

 

 

Team Cognitive
Resources
(Additive)

Figure 9 - Plot of Team Cognitive Resources by Team Conscientiousness
Resources Interaction
The relationship between team cognitive resources and adaptation was higher in
teams that had high rather than low levels of team conscientiousness resources (phi = .50,
p < .05 versus phi = .03, ns). From Figure 9, it appears that having high levels of

conscientiousness resources was especially helpful for teams with high levels of cognitive

136

resources (80% adapted). That is, adaptation was highest in teams with high levels of
team conscientiousness resources and team cognitive

Table 5 also shows that the team cognitive resources by team openness interaction
was significant, 5 = -2.38, p < .05. Figure 10 shows that the pattern of the data were not
consistent with Hypothesis 7b, however. The relationship between team cognitive
resources and adaptation was higher in teams with low openness resources (phi = .42, p <
.05) than for teams with high openness resources (phi = .16, ns). Low openness resources

was especially debilitating to teams with low cognitive resources (29% adapted).

 

‘lLow Openness ‘
leigh Openness i

 

 

 

Low High

Team Cognitive
Resources
(Additive)

Figure 10 - Plot of Team Cognitive Resources by Team Openness Resources
Interaction
Table 5 shows that the 3-way interaction between team cognitive resources, goal '
difficulty, and feedback reached significance. The pattern of data illustrated in Figure 11,
however, did not support H4b. Instead of finding that the relationship between team

cognitive resources and adaptation would be highest when there were difficult goals and

137

high quality feedback. The relationship between team cognitive resources and adaptation
was lowest when there were difficult goals and high quality feedback ( phi = -.09, ns).

Adaptation was highest when high cognitive resource teams were assigned easy goals.

1 ,

0.8

%ofTeams 0-6'
Adapting Q4
0.2,

  
  

' High Team 9
Low Team 9

LFB = low quality feedback;
HFB = high quality feedback;
EG = easy goals;

06 = Difficult goals

Figure 11 -Plot of Team Cognitive Resources by Goal Difficulty by
Quality of Feedback Interaction

Exploration of Alternative Operationalizations

The analyses presented above used the additive model of aggregating the
characteristics of individuals to represent the team. However, it is possible that alternative
models of aggregation are also viable. To investigate this possibility, I repeated the

logistic regression analysis using conjunctive and disjunctive models of aggregation.

138

Coniunctive W Table 6 shows the results for the analyses using the
conjunctive individual differences variables. As with the additive model, both the team
cognitive resources and abruptness variables were significant. Due to the means of
aggregation, however, the interpretation of the parameter is a bit different. Teams with
relatively intelligent least intelligent members had higher odds of adapting (1.97) than
teams with relatively unintelligent least intelligent members. Also like the additive
model, the three way interaction was significant and the plot of the interaction revealed
an identical pattern to that in the additive analysis. Unlike the additive model there was
not a relationship between openness resources and adaptation, and in addition, there were
no significant 2-way interactions.

Disjunctive aggregation. Table 7 shows the results for the analyses using the
disjunctive operationalizations of the individual differences variables. As with the
additive model, the team cognitive resources, team openness resources, and abruptness of
disruption parameters were significant. Teams with highly intelligent most intelligent
members had higher odds of adapting (2.01) than teams with less intelligent most
intelligent members. Teams with highly open most open members had higher odds of
adapting (2.07) than teams with less open most open members. Also like the additive
model, the team cognitive resources x goal difficulty interaction, the team cognitive
resources x team conscientiousness resources interaction, and the team cognitive
resources x goal difficulty x quality of feedback interactions were significant. In addition,
the pattern of the plots of these interactions were identical to the additive model. Only the
team cognitive resources x team openness resources interaction failed to replicate using

the disjunctive operationalization.

139

Table 6 - Logistic Regression of Conjunctive Independent

 

 

Variables on Adaptation
% Model X2 Block X2

Step Variable Correct (df) (df) b SE 2 eb
(0) Constant Only 50.20
(1) Team Cognitive 65.96 15.l3(3)"' 15.13(3)* .68 .20 3.45* 1.97

Resources (Tg)

Team .04 .18 .21 1.04

Conscientiousness (Tc)

Team Openness (To) .23 .18 1.29 1.26
(2) Goal Difficulty (GD) 68.09 27.17(6)* 12.04(3)* -.22 .38 -.58 .80

Quality of -.18 .38 -.46 .84

Feedback (FB)

Nature of 1.23 .37 3.28* 3.41

Disruption (ND)
(3a)‘ Tg x GD 66.67 27.38(7)* .22(1) -.19 .41 -.46 .83
(3b) Tg x F B 67.38 2736(7)" .19( 1) -.18 .42 -.44 1.20
(3c) GD x FB 68.79 27.74(7)"' .57(1) .58 .76 .76 1.78
(3d) Tg x ND 69.50 27.76(7)* .60(1) -.32 .42 -.77 .73
(3e) Tg x Tc 68.09 28.46(7)"' 129(1) .23 .20 1.16 1.26
(3f) Tg x To 70.21 29.85(7)* 268(1) -.37 .23 1.61 .69
(4)b Tg x GD x FB 70.92 34.43(10)* 6.06(1)"' -2.30 .98 -2.33* .10

 

n=141

' Step 3 results are from separate runs.
" Only interactions 3a, 3b, 3c entered prior to step 4.

*g<.05
‘p<.10

140

Table 7 - Logistic Regression of Disjunctive Independent

 

 

 

Variables on Adaptation
Step Variable % Model X2 Block X2 b SE 2 eb
Correct (df) (df)
(0) Constant Only 50.20
(1) Team Cognitive 65.25 11.17(3)* 11.72(3)* .37 .18 2.01* 1.45
Resources (Tg)
Team -.25 . 18 -1.41 .78
Conscientiousness (Tc)
Team Openness (To) ' .38 .18 2.07* 1.46
(2) Goal Difficulty (GD) 68.09 25.32(6)* 14.15(3)* —.13 .37 -.34 .88
Quality of .04 .38 -.12 1.04
Feedback (F B)
Nature of 1.36 .38 3.61* 3.90
Disruption (ND)
(3a)a Tg x GD 70.21 33.14(7)* 7.82(1)* -1.02 .46 -2.63"‘ .30
(3b) Tg x FB 68.09 25.43(7)* .12(1) -.14 .40 -.34 .87
(3c) GD x FB 68.09 25.42(7)* .10(1) .24 .75 .32 1.27
(3d) Tg x ND 66.67 25.43(7)* .1 1(1) .14 .41 -.33 1.15
(3e) Tg X To 73.05 31.64(7)"' 6.32(1)"' .61 .26 2.38“ 1.84
(30 Tg x To 68.09 25.40(7)* .08(1) .05 .18 .29 1.05
(4)b Tg x GD x FB 68.79 36.21(10)* 2.94(1)‘ -l.70 1.02 -1.67‘ .18
n = 141

‘ Step 3 results are from separate runs.

b Only interactions 3a, 3b, 3c entered prior to step 4.
* p < .05

'p < .10

141

Predicting Smed of Adaptation

The analyses presented above attempted to identify the variables that predict the
likelihood that teams adapt. The data collected can address an additional question: Which
variables predict how quickly teams adapt? In order to address this question, I performed
analyses on scores on the original operationalization of adaptation for the 75 teams that
adapteduthe more trials the team uses an adapted structure, the more quickly a team can
be said to adapt. This count variable will now be referred to as SMd of adaptation.

On average, the 75 teams that adapted used their new role structure for 13.04 trials
(sd = 6.52). Thus, on average, it took these teams almost 17 trials to establish a new role

structure. Table 8 presents correlations with the study variables as defined earlier.

Table 8 - Relationships with Speed of Adaptation

 

 

Speed of Adaptation
Pre-TDA -.10
Post-TDA -.46"‘
Additive g .23*
Conjunctive g .17
Disjunctive g .16
Additive Conscientiousness .04
Conjunctive Conscientiousness .06
Disjunctive Conscientiousness .01
Additive Openness -.07
Conjunctive Openness .05
Disjunctive Openness -.13
Goal Difficulty .02
Feedback Timeliness .05
Abruptness of Disruption .07

 

n=75;*p<.05

142

The significant correlation with post-TDA indicates that the quicker adapting
teams made better decisions on average than teams adapting more slowly. Of the twelve
predictors, only the additive operationalization of team cognitive resources related
significantly to speed of adaptation. Teams with members with higher levels of g on
average tended to adapt more quickly than teams members with lower levels of g.

When only the adapting teams are considered, the original operationalization of
adaptation (i.e., speed of adaptation) appears to be approximately normally distributed as

shown in the stem and leaf plot in Figure 12.

 

Frequency Stem & Leaf

6.00 0. 344444
24.00 0. 555666666677778889999999
11.00 1 . 00000033444
15.00 1 . 555555557788899
15.00 2. 000000000000014
4.00 2. 5559

Stem width: 10.00

Each leaf: 1 case(s)

 

 

 

Figure 12 - Stern and Leaf Plot of Adapting Teams’ Speed of Adaptation

While this data does have a slight positive skew (skewness = .28, se = .28), linear
regression is still most likely the best means of assessing predictive models of speed of
adaptation. Accordingly, I repeated the analysis presented in Tables 5-7 using hierarchical
linear regression. None of the regression runs resulted in a significant overall model.
Thus, with the exception of a main effect for additive team cognitive resources, the

variables measured in this dissertation did not appear predict how quickly teams adapted.

143

Of course, an alternative explanation for the lack of effects is that the analysis had
relatively low power. Using the same assumptions as outlined when describing how the
study was originally planned, the power of this analysis with an n of 75 is approximately

.70. However, the power to detect a model with smaller effect sizes is much lower.

144

Chapter 4

DISCUSSION

This chapter summarizes results from the pilot and primary studies and discusses
the theoretical and practical implications. This chapter also discusses several limitations
of this study. Future research needs are highlighted throughout this discussion.

Overview of Problem and Theog;

Although work is increasingly being structured around teams of interdependent
individuals, writers in both academia (e.g., Gersick & Hackman,, 1990; Hollenbeck, et
al., 1996; Ilgen, LePine, & Hollenbeck, 1997; Janis, 1972; Latane, Williams, & Harkins,
1979; Stasser & Titus, 1985, 1987; Steiner, 1972) and in the popular press (e.g.,
Dumaine, 1994) point out that the use of teams to accomplish work has a number of
limitations. While some recognized small group and team problems (e.g., group-think,
social-loafing, process-loss) have been the subject of much research attention (e.g. Janis,
1972; Latane, et a1, 1979; Steiner, 1972), the problem of adaptation has not. In fact, there
have been no published empirical tests of theory related to adaptation in teams. This is a
significant shortcoming since failure to adapt to changing situations can be blamed for
costly team decision making errors.

This dissertation attempted to address this shortcoming by developing and testing
a model of team adaptation based on a group-level information processing framework.
This model proposes that adaptation is a firnction of the team’s pool of cognitive
resources as well as those factors that control the mode of information processing.

Consistent with a control theory fiamework, I proposed that switching between routine

145

and adaptive behavior necessitates noticing a discrepancy between the desired and present
situation. Accordingly, I proposed that factors such as goals, feedback, and the nature of
the change in the environment should all influence the extent to which discrepancies
become apparent. I suggested that noticing a discrepancy is not enough to guarantee
behavioral responses to the discrepancy, however. Thus, factors that influence
individuals’ commitment to goals and their willingness to explore alternative systems of
behavior should be critical in triggering behavioral (as opposed to cognitive) responses to
discrepancies. Accordingly, I hypothesized that team members’ conscientiousness and
their openness to experience would be important predictors of team adaptation.

Overview of Pilot Study Results

The pilot study was conducted primarily to ensure that the experimental task
allowed for the development of routine team-level behavior and adaptation once the task
environment changed. Overall, the evidence was fairly strong that the task did allow for
the development of routine behavior. With some experience with the task, the pilot teams
accomplished the same pattern of behavior more quickly and reliably up to a point where
no further gains were evident. The evidence was also fairly strong that the task allowed
some teams to adapt, and just as important, that adaptation improved team-level
performance.

The findings of the pilot study are particularly noteworthy because researchers
have suggested that the biggest impediment to research on team performance in the face
of non-programmed environmental change (team adaptation) has been the lack of an
experimental paradigm in which to study the phenomenon (Behling, et a1, 1967). Whether

or not the lack of an experimental paradigm can be blamed for the lack of published

146

empirical research on team adaptation over the last thirty years is unclear. Nevertheless,
in addition to any finding from the primary study, this dissertation makes a contribution
in that a task is developed in which adaptation can be studied by other researchers.
Overview of the Prim Study Results

Overall, the results from the primary study provide mixed support for the
integrative model of team adaptation. The results of this study are reviewed in terms of
individual level findings and the three elements underlying the pr0posed model (depicted
in Figure 1).

Individual-level elements. Individual-level cognitive processes seemed fairly
consistent with expectations underlying the team-level hypotheses. First, there were
significant relationships between individuals’ cognitive resources and declarative
knowledge and situational awareness. Those with higher levels of cognitive resources
gained more declarative knowledge about the task during the training and also had more
accurate perceptions regarding the reliability of communications during the experiment.
Second, afier controlling for individual differences, two of the three manipulations (goal
difficulty and the nature of the environmental disruption) were related to self—reports of
cognitive effort directed towards overcoming difficulties posed by the changing task
environment. Difficult goals and abrupt changes triggered more self-reported cognitive
effort than easy goals or gradual changes. Unexpectedly, however, the quality of feedback
did not seem to trigger additional cognitive effort. As outlined earlier, it may have been
the case that the high and low quality feedback triggered cognitive effort aimed at
alternative ends. While high quality feedback may have triggered cognitive effort aimed

at addressing perceived or expected performance discrepancies, delayed and summarized

147

feedback may have triggered cognitive effort aimed at developing an understanding of
where the team stood in terms of its performance.

At the team-level, on the other hand, three of the thirteen hypotheses were
supported (H1, H4a, H6b) with another received only marginal support (H7a). Of the nine
hypotheses that failed to receive support, three relationships were exactly opposite as that
expected (H3a, H4b, and H7b). These results certainly call into question the validity of
the integrative model as presented in this dissertation. However, it is worthwhile to
review where the model seemed to be supported and where it did not.

Team cogm'tive resources. Perhaps the most consistent finding in this study is the
relationship between team cognitive resources and team adaptation. Whether
operationalized as a mean, as the highest members’ score, or as the lowest members’
score, the level of team cognitive resources was positively related to team adaptation such
that higher levels of resources increased the likelihood of team adaptation. In addition,
team cognitive resources were more highly related to team performance during the period
when adaptation was necessary (post-TDA) than when the working environment was
stable (pre-TDA). Thus, the conceptualization of adaptation as requiring a mode of
thinking that consumes cognitive resources (i.e., active or controlled information
processing) seems to be supported in these data.

Triggers of conglled information processing. The mechanism hypothesized to
play a role in triggering the use of controlled information processing, however, received
less support. While there was a significant relationship between the nature of the
disruption and team adaptation that was consistent with H5a (teams were more likely to

adapt if they were in the abrupt condition rather than the gradual condition), neither the

148

difficulty of the team goal nor the quality of feedback related to that goal had direct
effects on adaptation (H2a, H3a). Calling further doubt onto this conceptualization is the
fact that the moderating effect of goal difficulty was exactly opposite as expected (H2b).
Instead of the cognitive resources--adaptation relationship being stronger for teams with
difficult goals, the relationship was stronger for teams with easy goals. Also, in contrast
to H4b, the three way interaction showed that the relationship between cognitive
resources and adaptation was lowest in teams with difficult goals and high quality
feedback.

Factors that influence mode of discrepancy reduction. Finally, there was mixed
support for the factors hypothesized to influence adaptation through an effect on mode of
discrepancy reduction. There was not a main effect for team conscientiousness resources
on adaptation operationalized additively, conjunctively, or disjunctively (H6a), however,
consistent with expectations, there was a significant two-way interaction (H6b). The
relationship between team cognitive resources and adaptation was higher in teams with
high levels of additive or disjunctive conscientiousness resources than in teams with low
levels of additive or disjunctive conscientiousness resources. The pattern of this
interaction suggested that it was necessary to have both high levels of team cognitive
resources and high levels of team conscientiousness resources to promote the highest
levels of team adaptation.

As for team openness resources, there was a marginally significant main effect
such that highly open teams (either in an additive or disjunctive sense) were more likely
to adapt than teams that were less open (H7a). Additive openness also significantly

moderated the relationship between cognitive resources and adaptation. In contrast to

149

H7b, however, there was a stronger relationship between cognitive resources and
adaptation for teams that were less open. The pattern of this data suggests that to some
extent cognitive resources and openness can compensate for one another. High levels of
team cognitive resources compensated for low levels of team openness resources and
high levels of openness resources compensated (to a lesser extent) for low levels of team
cognitive resources.

Wding the Modera_ti_nLEffect of Goal Difficulty

Perhaps the most important unexpected finding in the dissertation was the nature
of the moderating effect of goal difficulty on the relationship between team cognitive
resources and adaptation. Whereas I expected a more highly positive relationship between
team cognitive resources and team adaptation in the difficult goal condition, the
relationship appeared more highly positive in the easy goal condition. This finding is
particularly noteworthy given that discrepancies between goals and performance levels
are the foundation for the hypothesized mechanism for triggering the type of information
processing necessary for adaptation. In this section, I present an analysis that more
directly assesses the control theory framework outlined earlier.

Absolute goal levels versus pgrformance-goal discrepancies. An underlying
assumption in the use of the goal difficulty variable was that teams with more difficult
goals would be more likely to experience negative discrepancies between their
performance levels and their goal and that teams with easier goals would be more likely
to experience positive discrepancies between their performance levels and their goals.
One problem with this assumption is that the absolute goal level is only one component

of any perceived discrepancy. Because variance in performance may be attributable to a

150

host of factors (e.g., team cognitive resources), teams with easy goals may experience the

same type of performance-goal discrepancies I had assumed to be characteristic of teams

with difficult goals. Thus, because the theory used to develop the hypotheses really

implies effects due to responses to the discrepancy between a team’s standing and their

goal, not just the team’s absolute goal level, the use of an absolute goal level variable

may have been inappropriate. Accordingly, I conducted an analysis using a discrepancy

variableupre-TDA minus the team goal (reflected in terms of decision accuracy, i.e., low

scores imply higher performance) in place of goal difficulty in a logistic regression on

team adaptation. Results from this analysis are summarized in Table 9.

Table 9 - Logistic Regression of Discrepancy Variable on Adaptation

 

Model X2

Block X2

 

Step Variable % z e
Correct (d0 (d0

( 1) Team Cognitive 65.96 19.58(1)* 19.58(1)* 4.01 * 2.29
Resources (Tg)

(2) Performance-Goal 64.54 20.09(3)* .51(2) . -.70 .88
Discrepancy (PGD)
Quality of Feedback (FB) .13 1.05

(3) Tg x PGD 68.09 31.57(6)* 11.47(3)* -2.12* .63
PB x PGD 2.07“ 2.21
FBng -1.03 .61-

(4) Tg x PGD x FB 70.92 35.08(7)* 3.51(l)' -l.75‘ .38

 

n=l4l
*p<.05
tp<.10

151

Results of post-hoe analysis. Consistent with the effects for the goal difficulty
variable, there were no main effects for the performance-goal discrepancy variable when
entered after team cognitive resources. However, the team cognitive resources by
performance-goal discrepancy interaction entered at step three did reach conventional
levels of significance. As Figure 13 shows, the pattern of this interaction is not consistent
with the integrative model because team cognitive resources are more positively related
to team adaptation when there are positive performance-goal discrepancies. That is,
consistent with the analysis using absolute goal level as a predictor, negative
performance-goal discrepancies do not appear to be a trigger of team adaptation as
hypothesized. Apparently, and in contrast to expectations, the best adapters were those
least needing to adapt--high ability teams that were meeting or exceeding their goals

when the environmental disruption first occurred.

 

Teams 0.4
Adapting 0.3

0.2 )
0.1.x” .‘
o

:gNegative

. Discrepancy
lIPositive i
j Discrepancy.

 

 

 

Low High

Team Cognitive
Resources (additive)

Figure 13 - Plot of Team Cognitive Resources by
Performance-Goal Discrepancy Interaction

152

One plausible explanation for this pattern of data lies with the notion that ongoing
task performance and adaptation are distinct activities that compete for a limited supply
of member cognitive resources. When teams are not achieving their goals, members’
cognitive resources become focused on their goal and their level of task performance.
That is, members focus their attention on monitoring their performance, making
behavioral, cognitive, and affective adjustments, and evaluating their performance
relative to their goals (self regulation). Because these activities consume conscious
attention (i.e., cognitive resources), the type of reflection and learning necessary for
adaptation cannot occur. According to this conceptualization, therefore, there is not likely
to be a relationship between team cognitive resources and adaptation when there are
salient negative performance-goal discrepancies because cognitive resources are allocated
towards task focused self-regulation. This explanation is certainly consistent with
research that suggests that the motivational aspects of negative performance-goal
discrepancies can detract from task learning. Kanfer and Ackerrnan (1989), for example,
showed that the assignment of goals stimulated self-regulatory activities that impeded
task learning for individuals with high as well as low levels of cognitive resources.

Members in teams meeting their goals, on the other hand, allocate their cognitive
resources towards learning how to act in the new situation. Individuals in these teams
may feel they have “slack”, and therefore, they may be more willing to sacrifice
immediate performance gains for the type of thoughtful effort (e. g., experimentation and
trial and error learning) adaptation requires. Thus, because adaptation depends on this
type of effort, there should be a positive relationship between team cognitive resources

and adaptation for teams with positive performance-goal discrepancies.

153

As Table 9 indicates, the performance-goal discrepancy by quality of feedback
interaction was statistically significant. As plotted in Figure 14, this interaction is
consistent with the explanation outlined in the previous two paragraphs--adaptation was
highest in teams for teams with positive performance-goal discrepancies and lower
quality information regarding their standing relative to their goal. That is, the likelihood
of team adaptation was the highest when chances of stimulating task focused self-
regulation were the lowest. Although not illustrated here, the plot of the marginal three
way interaction also supports this explanation. The plot indicated that the likelihood of
team adaptation was highest for high cognitive resource teams with positive performance-
goal discrepancies with low quality feedback. In addition, the plot indicated that there
were positive relationships between team cognitive resources and adaptation only when

teams had positive performance-goal discrepancies.

 

 

Low th
Quality of Feedback

Figure 14 - Plot of Performance-Goal Discrepancy by
Quality of Feedback Interaction

154

Implications of Results

While the study described in this dissertation did not support for the model as
proposed, there were several findings that have implications for both theory and practice.

muons to thCOI'L From a theoretical standpoint, this study should make a
contribution because it is the first to present and _t_h__e_n_ test a model of team adaptation.
This is a significant departure from past small group or team research that has generally
been aimed at describing problems associated with routines and offering normative
recommendations for managing routine abandonment (e.g., Gersick & Hackman, 1990).
Although some research has been aimed at adaptation specifically, what exists tends to be
based on retrospective qualitative analyses of a single occurrences (e.g., Hutchins, 1996).
The current effort not only presents a priori hypotheses, but attempts to operationalize
adaptation quantitatively. In addition, this is the first effort to study team adaptation in a
controlled laboratory setting using a large number of teams.

In terms of the integrative model itself, the conceptualization of adaptation as
requiring distributed controlled information processing received supported. Teams with
higher levels of cognitive resources had higher odds of adapting than teams with lower
levels of cognitive resources. While this finding is significant in terms of providing
support for the conceptualization of adaptation proposed in this dissertation, it is not
surprising given research that demonstrates the importance of cognitive resources in
novel or complex learning situations (Kanfer & Ackerman, 1989;Hunter, 1986).

The triggering of controlled cognition, however, appears to be more complicated

than I had originally proposed. At the individual level, goal difficulty and the abruptness

155

disruption triggered more cognitive effort, however, the effort triggered by difficult goals
did not necessarily translate into an increased likelihood of adaptation. In fact, the
likelihood of adaptation for high cognitive resources teams appeared lower when there
were difficult goals. A more direct assessment of the control-theory based triggering
mechanism suggested an explanation consistent with Kanfer and Ackerrnan’s (1989)
integrative model of skill acquisition. That is, because both task related self-regulation
and learning (adaptation) consume cognitive resources from a finite pool (the team),
elements in the environment (e. g., goal difficulty, quality of goal relevant feedback) that
focus attention on performance shortcomings necessarily inhibit learning (adaptation). In
terms of the integrative model, therefore, controlled cognition allocated towards
adaptation is not likely to be triggered as a result of factors that focus members attention
towards performance or the need for performance improvements (e. g., difficult
performance goals, high quality feedback).

I should note that another hypothesized trigger of controlled information
processing, abrupt environmental disturbances, did lead to higher levels of team
adaptation. Thus, at first glance the rationale for the unexpected effects outlined in the
previous paragraphs seems suspect. I suggest, however, that the nature of the
environmental disturbance influences the extent to which team members focus their
attention (cognitive resources) on the nature of the problem. Thus, perhaps instead of a
global triggering mechanism, cognitive resources are allocated accordingly in terms of
factors that direct team members attention either towards the task or towards the problem.
Difficult performance goals and perhaps high quality feedback fall in the former category

while abrupt environmental change falls in the latter.

156

The effects of two individual difference based variables on adaptation received
mixed support. Teams with high levels of cognitive resources were more likely to allocate
cognitive resources towards adaptation if the team also had high levels of
conscientiousness resources (operationalized additively or disjunctively). When one
considers that conscientiousness indicates the tendency to be motivated, the pattern of the
data is consistent with the notion that performance is a function of the classic ability by
motivation interaction.

Teams with higher levels of additive or disjunctive openness resources also
allocated more resources towards adaptation, and to some extent, high levels of these
resources (operationalized additively) compensated for low levels of cognitive resources.

Implications to practice. Consistent with past descriptive accounts, failing to
break out of routines in changing situations was both prevalent and costly. This study
demonstrated that 66 out of 141 teams failed to adapt to an environmental change, and
these non-adaptive teams performed significantly worse (post-TDA = 1.22, sd = .3 7) than
those that adapted (post-TDA = .96, sd = .27), E = 24.27, p < .05. This study was
intended to uncover factors that influence whether or not teams adapt.

Although the proposed theory was not entirely supported, this dissertation makes
a practical contribution because it suggests means by which mangers can enhance
adaptive potential of specialist decision making teams. Generally speaking, these levers
can be grouped in terms of three specific types--team composition, team structure, and
the nature of the environment.

Team commsition refers to the nature of the individuals who work in a team.

From a practical standpoint, research on team composition is important because findings

157

suggest staffing interventions. This dissertation suggests that staffing a team with
individuals who are intelligent and open should facilitate team adaptation. Team
members’ conscientiousness may also be important because it influences the relationship
between intelligence and adaptation. Indeed, the adaptation rates for teams that were high
on these three characteristics was almost twice as high as for teams that were low on
these three characteristics (75% adapted versus 40%).

While somewhat exploratory, I also assessed the extent to which adaptation might
be predicted by alternative operationalizations of the individual characteristics. The
results suggested the most support for the additive and disjunctive models. Thus, it
appears that the members can compensate for one another in terms of the resources they
bring to the task of adapting. Specifically, it is likely that members with high cognitive
resources and openness can compensate to a large extent for members with low ability
and openness in terms of the contributions they make towards their team adapting.

Thus, while this dissertation suggests that team composition plays an important
role in team adaptation, there are many open issues that need to be addressed in future
research. For example, researchers should examine (a) the effects of the distribution of
attributes within a team, (b) the effects of additional types of attributes, and finally, (0)
how different attributes interact with one another. Because there are many ways to create
compositional variables, research should be guided by well developed theory and tested
using designs tailored to assess the validity of specific operationalizations. .

Generally speaking, team structure can be defined in terms of how members are
related to one another and to their task. The nature of goals and feedback are structural in

the sense that they provide information that influences the allocation of member’s

158

attention--either towards or away from the task. Although unexpected, the findings in this
dissertation suggest that difficult goals may be deleterious to teams in terms of
adaptation. This finding is significant because it suggests that performance pressures do
not lead to a search and experimentation for a new and better way to do thing in light of a
changing situation. Furthermore, increasing the quality of information about the situation
do not improve matters, and as suggested by the three-way interaction, may make matters
worse. These findings, therefore, suggest the need to identify factors that might reduce
performance pressures in changing situations. Designing roles such that some members
are primarily responsible for identifying and dealing with environmental change stands as
one possibility. Another possibility might be to “suspend” performance goals in times of
environmental change so that members can allocate their full attention to adapting.

Finally, this study suggests that the nature of the environmental change may play
a role in team adaptation. While controlling the nature of the environmental change may
be extremely difficult, developing mechanisms that enhance teams’ awareness and
understanding of environmental change might be possible. For example, it may be
possible to train teams to monitor their environment for change.
Limitations

Obviously, the primary limitation of the study is its generalizability in terms of
the task, context and sample. That is, because the study was (a) conducted in a laboratory
using a computer game with (b) weak consequences relative to the real world tasks and
with (c) undergraduates composing temporary teams as participates, the results may not
hold up in other settings and/or with other individuals. Indeed, these issues are certainly

important to the extent to which one wants to generalize findings. However, the intent of

159

this study is not to estimate effect sizes and then generalize them to specific populations,
but to test theoretical expectations in such a manner in which they stand the possibility of
being disproven. Because the sample, task, and context fit within the boundaries specified
by the theory (i.e., there is nothing in the theory which states that it would not work in the
study context) they provide a fair venue for testing the theory. I should also note that just
because a study is conducted in a field does not mean it is any more generalizable than
one conducted in the laboratory (Dipboye & Flanagan, 1979). In order for a single study
to be truly generalizable, one would have to sample randomly fiom subjects, tasks and
times (Cook & Campbell, 1979). For practical reasons, this is almost always an
impossibility. Instead, Cook & Campbell (1979) suggest that generalizability can be
enhanced by many heterogeneous small studies. I view this study as the first among such
studies.

A second limitation of this study has to do with the nature of the manipulations.
Specifically, this study focused on a specific dimension of goals (difficult goal
outcomes), feedback (quality), and environmental change (abrupt versus gradual) to the
exclusion of others (e. g., process goals, process feedback, and clear versus ambiguous
change). While results of the study may have been different had I considered additional
variables, practical and statistical concerns limit the number of variables considered in
any one study. Nevertheless, conclusions as to the validity of the theory as presented need
to be tempered in that results may have been different had I used different variables or
different dimensions of the variables I did consider.

A third limitation of this study is the focus on a relatively narrow type of team--

specialist decision making teams that must continue performing in a changing

160

environment. The reason for this focus is that I believe these types of teams are
particularly susceptible to the inappropriate application of routine behavior in changing
situations. The primary reason for this belief is that these types of teams must adapt in the
course of performing their task. The reason why this focus may be a limitation is that
perhaps results are probably quite task specific. The most obvious example would be the
explanation for the unexpected moderating effect of goal difficulty. Certainly, there are
many types of team situations that afford members more of an opportunity for reflection
and rational planning, and thus, allocation of cognitive resources towards the task would
be less of an issue. For example, team members could use their cognitive resources
towards the task until they feel it is necessary to take a break (e. g., stop production) so
that they can allocate their full attention to thinking of ways to behave in the new
situation.
Overall Conclusion

Nearly half of the teams in the experiment failed to adapt and failing to adapt was
costly in terms of decision making performance. The findings in this study suggest that
for teams that need to continue making decisions in an changing environment, the
composition of the team in terms of member characteristics, team structure, and the
nature of the environmental change, all influence the extent to which adaptation to the
change is likely. Indeed, given the large differences in odds of adapting between teams in
the different experimental conditions, additional research aimed at understanding and
ultimately identifying interventions to promote team adaptation may be worthwhile. It is

hoped this dissertation serves as the impetus for such research.

161

APPENDICES

162

APPENDIX A

163

APPENDIX A

Team Decision Making Consent Form (initial)

This study is designed to investigate team decision making effectiveness. If you choose to participate in
this study, you will be asked to learn a computer simulated target identification task, operate the simulation
task with other individuals, and complete a series of questionnaire items. In addition, if you choose to
participate in this study, you authorize the researchers to have access to the questionnaire that you
completed in your Management 302 recitation section at the beginning of the semester, your Management
302 scores on individual and group assignments, and your SAT/ACT scores.

Your participation in the simulation should take approximately three hours. In exchange for your
participation in this study, you will receive miscellaneous credit for your Management 302 class
requirement to participate in a research project. Other research projects or alternatives are available from
your instructor if you decide not to participate in this study. If you do participate, you have the opportunity
to earn a cash bonus. The top teams in each condition will earn a cash prize on the following basis: First
place $20/person; Second place SIS/person; Third place $lO/person. You may also have an opportunity to
take home a questionnaire to fill out, for which you will receive 810.

Your participation in this research is completely voluntary. You are free to decline to answer any questions
or to terminate your participation at any time. Your participation in this study will be totally confidential.
Your data will be included in a summary report along with the data from others. The report will not
include any information that will allow anyone to identify any of your individual responses.

If you have any questions or concerns regarding this study, you may contact Jeff LePine in the
Management Department at 353-7229.

Participants Statement

1 agree to participate in the Team Decision Making Study. I understand that I will learn to operate a
computer simulation, perform the simulation with other individuals, and fill out a series of questionnaires.
I authorize the researchers to use my SAT/ACT scores, the questionnaire that I already completed in the
Management 302 recitation section, and my scores on individual and group assignments in Management
302. It is my understanding that these materials will be strictly confidential and will not be seen by anyone
other than the research team. I consent to having these materials used for research purposes.

I understand that the top three teams in each condition will be eligible for cash prizes on the following
basis: First place $20/person; Second place SIS/person; Third place SIG/person. I may also be able to fill
out an additional questionnaire for $10.

I understand that my participation is voluntary, that I may discontinue participation at any time without
penalty, that all of my individual responses will be kept strictly confidential, and that I will not be identified
in any report of this study.

 

 

 

 

Signature Date
Printed Name PID (MSU ID Number) (for data matching purposes)
My MG’l‘ 302 TA's name MGT 302 Section number (if unsure see board)

 

164

APPENDIX B

165

APPENDIX B

Team Decision Making Consent Form (experiment)

This study is designed to investigate team decision making effectiveness. If you choose to participate in
this study, you will be asked to learn a computer simulated target identification task, operate the simulation
task with other individuals, and complete a series of questionnaire items. In addition, if you choose to
participate in this study, you authorize the researchers to have access to the questionnaire that you
completed in your Management 302 recitation section at the beginning of the semester, your Management
302 scores on individual and group assignments, and your SAT/ACT scores.

Your participation in the simulation should take approximately three hours. In exchange for your
participation in this study, you will receive miscellaneous credit for your Management 302 class
requirement to participate in a research project. Other research projects or alternatives are available from
your instructor if you decide not to participate in this study. If you do participate, you have the opportunity
to earn a cash bonus. The top teams in each condition will earn a cash prize on the following basis: First
place $20/person; Second place SIS/person; Third place SIG/person. You may also have an opportunity to
take home a questionnaire to fill out, for which you will receive $10.

Your participation in this research is completely voluntary. You are free to decline to answer any questions
or to terminate your participation at any time. Your participation in this study will be totally confidential.
Your data will be included in a summary report along with the data from others. The report will not
include any information that will allow anyone to identify any of your individual responses.

If you have any questions or concerns regarding this study, you may contact Jeff LePine in the
Management Department at 353-7229.

Participants Statement

I agree to participate in the Team Decision Making Study. I understand that I will learn to operate a
computer simulation, perform the simulation with other individuals, and fill out a series of questionnaires.
I authorize the researchers to use my SAT/ACT scores, the questionnaire that I already completed in the
Management 302 recitation section, and my scores on individual and group assignments in Management
302. It is my understanding that these materials will be strictly confidential and will not be seen by anyone
other than the research team. 1 consent to having these materials used for research purposes.

I understand that the top three teams in each condition will be eligible for cash prizes on the following
basis: First place $20/person; Second place 31 5/person; Third place SIG/person. I may also be able to fill
out an additional questionnaire for $10.

I understand that my participation is voluntary, that 1 may discontinue participation at any time without
penalty, that all of my individual responses will be kept strictly confidential, and that I will not be identified
in any report of this study.

 

 

 

 

Signature Date
Printed Name PID (MSU ID Number) (for data matching purposes)
My MGT 302 TA's name My MGT 302 Section number (if unsure see board)

 

166

APPENDIX C

167

APPENDIX C

General Overview Handbook

INTRODUCTION

The year is 1997 and you are a part of a US. Air 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 US. assets in the area from
accidental or intentional attacks. As history indicates, this is a highly sensitive task. For
example, in 1987, failure by a command and control team to quickly and accurately
identify a plane as threatening, allowed an Iraqi jet to accidentally fire two Exocet
missiles into the Frigate U.S.S. Stark, killing 37 American serviceman and crippling the
vessel. One year later, a command and control team error resulted in the USS. Cruiser
Vincennes accidentally shooting down an Iranian passenger plane killing 290 innocent
civilians. In 1994 two US. F-15 fighters shot down two friendly helicopters in Northern
Iraq killing 26 people. Another occurrence such as this, or of the USS. Stark or U.S.S.
Vincennes variety, will diminish public support for the current mission, and in turn,
jeopardize peace in this region. It is certain any repeat of mistakes of this kind will lead to
escalated hostility and long-term conflict.

THE TASK FORCE

As a member of a three-person command and control team, you will be linked to
your fellow team members through an electronic data network that can supply bits and
pieces of critical information concerning possible enemy targets. Your mission is to
communicate and coordinate your information, so that the team commander ends up seeing
an accurate overall “big-picture”. Each team member is a specialist in interpreting particular
bits of information. Each person’s specialty must be considered as a whole in order to get
an estimate of the overall threat of a particular target. This is necessary so that the
commander can make appropriate decisions concerning possible enemy targets.

TEAM MISSION: Monitoring Air Space

The team that you are a member of will monitor the airspace surrounding an aircraft
carrier group, making sure that your ships are not attacked. In performing this role, you
must make certain that you do not allow loss of life resulting from enemy attacks on
ally ships in the fleet. At the same time, it is also of paramount importance that you do
not inadvertently shoot down friendly military aircraft or any civilian aircraft. Many
passenger flights move in and out of the region, and fiiendly military aircraft fi'om nations
not involved in the conflict also patrol the area

168

OVERVIEW OF ROLES—There are three roles in this simulation, referred to as ALPHA,
BRAVO, and CHARLIE. The team’s task is to decide what response should be made
toward incoming air targets. Team members will assess the target in accordance to their
specific expertise, and make recommendations to ALPHA who will then make the final
decision for the team. Team members base their decisions on data they collect by measuring
characteristics of targets that enter the task force’s airspace. These measures are obtained
from sophisticated electronic devices. There are seven posgble choices to make for each
incoming target. These responses are ordered in terms of their aggressiveness and there is
one and only one correct response for each aircraft. Each of these possible responses is
described below, moving from least to most aggressive.

SEVEN POSSIBLE DECISIONS

 

(1) IGNORE: This means that no further attention should be devoted to the target. Instead,
focus should be directed on other possible targets in the area. Never ignore a target that
might possibly attack.

(2) REVIEW: This means attention can be shifted away fiom this target momentarily.
After a short period of time this target should be returned to in order to update its status.

(3) MONITOR: This means that the target should be continuously tracked.

(4) WARN: This means that a message is sent to the target ordering it to turn away.
Warning targets that should be ignored detracts fi'om the importance of legitimate warnings.
Warning targets that intend to attack is also had, since the warning makes it easier for an
attacker to locate target the base.

(5) READY: This means to get into a defensive posture and to set defensive weapons on
automatic. A facility in a readied position is rarely vulnerable to attack. This stance should
not be taken to non-threatening targets since weapons set to automatic can fire mistakenly at
innocent targets that fly too close.

(6) LOCK-ON: This synchronizes radar and attack weapons so that the weapons fix
themselves on the target. A ship at Lock-On position can take offensive action at a
moments notice. The capacity to track other targets is severely constrained once there is
Lock-On to a single target, however. Thus, this should be reserved for targets that are
almost certain to be threatening.

(7) DEFEND: This is “weapons away” and means to attack the target with missiles or

depth charges. A defend decision cannot be aborted once initiated and thus must only be
used when enemy attack is imminent.

169

CHARACTERISTICS OF AIRBORNE TARGETS

Airborne targets can be measured on nine attributes. These are listed below along with the
range of possible values for each of the attributes:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Target cues Definition Range
Distance from base operations. In general,
RANGE targets that are closer are more threatening. 0 - 600
Number of feet target is above the ground.
ALTITUDE In general, targets that are low in altitude are 100 -
more threatening. 99,000
RADAR CROSS Estimated size of the target. In general, smaller
SECTION targets are more threatening. 0 - 12
Miles from center of civilian corridor. In
CORRIDOR STATUS general, targets far outside the civilian 0 - 25
corridor are more threatening.
ELECT—'RONIC Indicates threat of radar signals. In general,
SECURITY MEASURE targets with high ESM values are more 0 - 999
(ESM) threatening.
# OF AFR-CRAFT Each target may consist of more than one
WITHIN aircraft, flying close together. In general, a 1 - 20
THE TARGET higher number of aircraft is more
threatening.
Indicates direction of target. In general, the
HEADING CROSSING higher the HCA, the more directly the target is 0 - 180
ANGLE (HCA) headed toward the base, which is more
threatening.
RATE OF ALTITUDE Number of feet/minute ascending or
CHANGE descending. In general, the higher the rate of 0 - 10,000
(RATEAALT') altitude change the more threatening.
Miles per hour. In general, the faster the target
SPEED the more threatening. 0 - 800

 

IT IS IMPORTANT T0 NOTE THAT FOR SOME ATTRIBUTES LOW VALUES ARE
THREATENING WHILE FOR OTHER ATTRIBUTES HIGH VALUES ARE THREATENING—

i.e., don’t assume that high values are threatening across attributes!

170

 

 

DETERMINING THE LEVEL OF THREAT FOR A DECISION RULE

These nine attributes combine according to three rules which are used to determine the
level of threat associated with any target. Each team member has expertise for one of the
three rules -- the one bearing the member’s station name. Alpha is also responsible for
combining recommendations of the team members into a correct overall team decision.

RULE ALPHA: RANGE, CORRIDOR STATUS, # TARGETS

These three values should be considered together to determine the threat level for
this aspect of the target. If the target is threatening on all of these variables, then the
threat level for this aspect is very high. However, if any one of these three values is non-

threatening, the threat level for this asgct is low. Thus, a target can be very threatening
for two out of three of these values, and still be considered safe or non-threatening in

terms of this rule. Therefore it is important to make your judgment on this rule on all
three of the attributes listed above.

RULE BRAVO: ALTITUDE, HEADING CROSSING ANGLE, SPEED

As with Rule Alpha, these three values should be considered together to
determine the threat level for this aspect of the target. If the target is threatening on all of
these variables, then the threat level for this aspect is very high. However, if any one of
these three vglues is non-threatening, the threat level for this aspect is low. Thus, a
target can be very threatening for two out of three of these values, and still be considered
safe or non-threatening in terms of this rule. Therefore it is important to make your
judgment on all three of the attributes listed above.

 

RULE CHARLIE: RADAR CROSS SECTION, ESM, RATE‘ALTITUDE

As with Rules Alpha and Bravo, these three values should be considered together
to determine the threat level for this aspect of the target. If the target is threatening on all
of these variables, then the threat level for this aspect is very high. However, if any one
of these three values is non-threatening, the threat level for this aspect is low. Thus, a
target can be very threatening for two out of three of these values, and still be considered
safe or non-threatening in terms of this rule. Therefore it is important to make your
judgment on all three of the attributes listed above.

HOW ALPHA COMBINES RECOMMENDATIONS FOR TEAM JUDGMENTS

The three rules, ALPHA, BRAVO, AND CHARLIE (assigned to Alpha, Bravo,
and Charlie), combine to determine the overall threat represented by the target. For
example, if team member ALPHA recommends a DEFEND, and BRAVO recommends
a DEFEND, and CHARLIE recommends a MONITOR, and these recommendation
were correctly made, the overall judgement is probably LOCK-ON.

171

OUTCOMES OF DECISIONS

 

Your decisions regarding each target are to be made based upon the rules discussed
on the previous page. Once Alpha makes his or her decision, there are five possible
outcomes associated with the accuracy or that decision (scoring is done automatically by the
computer). The five possible outcomes include:

 

 

 

 

 

 

 

 

 

 

either monitor,
review, or ignore.

 

OUTCOME DEFINITION EXAMPLE POINTS
The decision was exactly You said defend,
(1) HIT correct. correct answer was +2
defend.
The decision was off by one You said defend,
(2) NEAR MISS level. correct answer was +1
lock-on.
The decision was off by You said defend,
(3) MISS two levels. correct answer was 0
ready.
The decision was off by You said defend,
(4) INCIDENT three levels. correct answer was -1
warn.
The decision was off by You said defend,
(S) DISASTER more than three levels. correct answer was -2

 

172

 

APPENDIX D

173

APPENDIX D

Role Instructions

INSTRUCTIONS FOR ALPHA

ALPHA has the responsibilities of both specific expertise and the coordination of information and
recommendations fi'om BRAVO and CHARLIE. Each team member has unique knowledge used to assess
the threat level of an airborne target. This unique knowledge is in the form of information concerning how
to measure and interpret specific pieces of information, and how to use that information to make
judgements regarding a target’s threat level. Each team member needs to get certain pieces of information
from other team members. The role of the team members is to acquire needed information from other team
members and then make a recommendation regarding the target. ALPHA uses these recommendations, as
well as his/her own knowledge, to arrive at the overall assessment of each target.

YOUR SPECIFIC ROLE

ALPHA you are responsible for the team’s final decision. You must combine knowledge from your area of
expertise with the recommendations of other team members. You are uniguely resmnsible for the following
information:

between between
TARGET CUE SAFE safe and MODERATE moderate and THREAT
moderate threatening

 

 

 

 

 

RANGE . 600 -400 399 - 301 300-200 199 -101 100 -0
CORRIDOR srfifis
0-3 4-9 10-15 16-20 21-25
# or AIRCRAFT wr' THIN—
THET‘ARGET 1-2 3-4 5-7 8-11 12-20

 

 

 

 

REMEMBER, THESE THREE CUES INTERACT TO GENERATE AN OVERALL “ALPHA”
ASSESSMENT:

IF ONE OF THESE THREE CUES IS SAFE,
THE OVERALL “ALPHA”ASSESSMENT IS SAFE.

- For example, a target may have threatening range and corridor status, but if the number of targets is
safe, the Alpha assessment for the target is safe.

0 As another example, the target may have threatening corridor status and number of targets, but if the
range is safe, the Alpha assessment for the target is safe.

ALSO REMEMBER:

o If the target is safe on one, two, or all three cues, the Alpha assessment is safe.

0 If the target is moderate on all three cues, the Alpha assessment will also be moderate.

o If the target is threatening on all three cues, the Alpha assessment will be very threatening.

174

INSTRUCTIONS FOR BRAVO

As BRAVO, you have unique knowledge that will allow you to gain a picture of one aspect of an aircraft’s
threat level. This unique knowledge is in the form of information concerning how to measure and interpret
specific pieces of information, and how to use that information to make judgments regarding a target’s
threat level. Each team member needs to get certain pieces of information from other team members. The
role of the team members (including ALPHA) is to acquire needed information from other team members
and then make a recommendation regarding the target. ALPHA uses these recommendations, as well as
his/her own knowledge, to arrive at the overall assessment of each target.

YOUR SPECIFIC ROLE
As BRAVO, you are uniquely resmnsible for the following information:

       
     

 
 

 
 

between between
TARGET CUE SAFE safe and MODERATE moderate and THREAT
moderate threatening

 

 

ALTITUDE 99,000-45,000 44,999-35,001 35,000-10,000 9,999-1,001 1,000 - 100
HEADING CROSSING
ANGLE (HCA) 0-60 61 -89 90- 120 121-149 150-180
SPEED 0- 100 101-199 200-300 301 —399 400-800

 

 

 

 

REMEMBER, THESE THREE CUES INTERACT TO GENERATE AN OVERALL “BRAVO”
ASSESSMENT:

IF ONE OF THESE THREE CUES ARE SAFE,

THE OVERALL “BRAVO” ASSESSMENT IS SAFE.

0 For example, a target may have threatening altitude and HCA, but if the speed is safe, the Bravo
assessment for the target is safe.

0 As another example, the target may have threatening HCA and speed, but if the altitude is safe, the
Bravo assessment for the target is safe.

ALSO REMEMBER:

o If the target is safe on one, two, or all three cues, the Bravo assessment is safe.

0 If the target is moderate on all three cues, the Bravo assessment will also be moderate.

o If the target is threatening on all three cues, the Bravo assessment will be very threatening.

\

175

INSTRUCTIONS FOR CHARLIE

As CHARLIE, you have unique knowledge that will allow you to gain a picture of one aspect of an
aircraft’s threat level. This unique knowledge is in the form of information concerning how to measure and
interpret specific pieces of information, and how to use that information to make judgments regarding a
target’s threat level. Each team member needs to get certain pieces of information from other team
members. The role of the team members (including ALPHA) is to acquire needed information from other
team members and then make a recommendation regarding the target. ALPHA uses these
recommendations, as well as his/her own knowledge, to arrive-at the overall assessment of each target.

YOUR SPECIFIC ROLE
As CHARLIE, you are _u_niquely responsible for the following information:

  

between between
TARGET CUE SAFE safe and MODERATE moderate and THREAT
moderate threatening

RADAR
CROSS 12-10 9-8 7-5 4-3 2-0
SECTION (RCS)

 

 

ELECTRONIC SECURITY
MEASURE (ESM) 0 - 175 176 - 224 225 - 400 401 - 599 600 - 999

 

RATE CHANGE
ALTITUDE 0-2,000 2,001 -2,999 3,000-5,000 5,001 -5,999 6,000-10,000
(RATEAALT')

 

 

 

 

REMEMBER, THESE THREE CUES INTERACT TO GENERATE AN OVERALL “CHARLIE”
ASSESSMENT:

IF ONE OF THESE THREE CUES ARE SAFE,

THE OVERALL “CHARLIE” ASSESSMENT IS SAFE.

o For example, a target may have threatening RCS AND ESM, but if the RATEAALT is safe, the
Charlie assessment for the target is safe.

0 As another example, the target may have threatening ESM and RAT'EAALT, but if the RCS is safe,
the Charlie assessment for the target is safe.

ALSO REMEMBER:

- If the target is safe on one, two, or all three cues, the Charlie assessment is safe.

0 If the target is moderate on all three cues, the Charlie assessment will also be moderate.

o If the target is threatening on all three cues, the Charlie assessment will be very threatening.

176

APPENDIX E

177

APPENDIX E

EXPERIMENT PROTOCOL

I. PRE EHERIIWENT SE T -UP

From the Condition Log in binder figure out what condition you are going to run.
Then note the group number within that condition.

To set up the computer for the this study:
(1) change directories into c:\team
(2) at the Alpha (formerly Carrier) station, type “begint”
(3) enter weight file: jeffO3
(4) enter group number from Condition Log.
(5) enter session ntunber (from condition log):
ABRUPT
81 (easy goal-delayed feedback)
82 (easy goal-full feedback)
83 (difficult goalndelayed feedback)
84 (difficult goal-full feedback)
GRADUAL -
85 (easy goal-delayed feedback)
86 (easy goal-full feedback)
87 (difficult goaludelayed feedback)
88 (diffith goalufull feedback)
(6) enter experiment number: 1
(7) “enter” on Alpha’s computer
(8) type “begin” at the Bravo and Charlie stations and hit Return on the Bravo and
Charlie stations. NOT ON THE ALPHA STATION--THIS WILL BEGIN THE GAME

II. CRASH PROTOCOL

In the event of an Alpha crash, restart everyone (type Shift-Fl to stop Bravo and Charlie)
and type in the above information, but type in a new experiment number. ENTER
WHATEVER YOU DID THE FIRST TIME, BUT REPLACE THE FINAL DIGIT
WITH THE NEXT NUMBER. So if you ran a game with experiment number = 1 and it
crashed halfway through, you would then type in 2 as the experiment number. If more
crashes occur, replace the “2” with an “3”, and so on.

If the Bravo or Charlie machines crash simply type begin and they should come back on
line (but check their clock to note any discrepancies in time remaining).

178

It is critical that you log all crashes in the Teamlab Log Sheet at the end of this binder,
especially Alpha crashes. Note the trial # of the crash and possible reasons for it.
III. TRAINING INSTRUCTIONS

IMPORTANT: fl indicates condition specific instruction

PHASE I - INTRODUCTION AND ADMINISTRATION

1. When subjects come in they should:
(1) sign in (after they do so check to make sure they are on the list for completing
the questionnaires administered during sign-ups).
(2) sign the consent form (CHECK TO MAKE SURE THEY FILL OUT ALL.
PARTS)
(4) As they are doing this, mark their green scantrons.
Team number goes in the SECTION fields.
Position goes in the second column of the FORM field (A=Alpha, B=Bravo,
C=Charlie).
(5) Mark their gender. (They will mark their PID’s when they fill out The
questionnaires)
(6) Mark the team off on the log sheet at the back of this binder and on the consent
form.

Ask the participants (when they are in the lobby) whether they completed the
questionnaire at the beginning of the semester in their recitation or at another
time by appointment. If they did not do it, they will need to schedule a time with
someone from the lab to complete the survey. Please check on the “SURVEY
ROSTER” to make sure all team members completed the survey even if they said
they did it.

2. Once three subjects are available for the study, take them into the room and sit them
at the Alpha, Bravo, and Charlie stations (which correspond to the traditional
Carrier, CAD, and AWACS stations respectively).

Welcome them:

“Welcome to the Team Effectiveness Lab. For the next 3 hours you will be
helping as study how teams make decisions and how we can improve the
efl'ectiveness of those decisions.”

“Today you will be participating in a study that will earn you 302 experiment '
credit. As an added bonus, depending on how well you perform, you will be
eligible for cash prizes.”

First Place = $60 per team ($20 per person)

Second Place = $45 perteam ($15 perperson)
Third Place = $30 perteam ($10 per person)

179

3.

PHASE II - INTRODUCTION TO TASK/TRAINING

Hand out the W. Take them through the
handbook page by page, briefly explaining each section. They will have time to read
it thoroughly, this is just to give them some context.

Page 1 -— Introduction and explanation of your mission. Basically the task is an
Air Force Command and Control simulation, you will each be members of a
Command and Control team. During each trial you will be presented with an
unknown aircraft. There are nine pieces of information that will indicate how
threatening this unknown target aircraft is. Each of you will be trained on how
to interpret 3 of those pieces of information. You will measure this information
as a team, share that information among yourselves, and decide just how
threatening that aircraft is. You are each assigned an individual role (tell them
who they are — Alpha, Bravo, and Charlie) with specific expertise regarding the
team’s decision. In addition Alpha will have the final say in all team decisions.
Page 2 - Overview of the seven possible decisions. These decisions are a
continuum of agressiveness, that is the responses range in agressiveness from
IGNORE to DEFEND with IGNORE being the least agressive decision and
DEFEND being the most agressive decion.

Page 3 - Characteristics of the targets that will be measured.

Page 4 - How the team members determine actual levels of threat. How
ALPHA (team leader) combines rules to determine team judgments.

Page 5 - How the outcomes of the decisions are calculated. In other words it
explains how points are calculated for correct or incorrect answers.

Tell the participants that the booklets will NOT be available to subjects during the
real targets, but they will be able to keep a summary sheet of what they need the
most!

Hand out the INDIVIDUALIZED INSTRUCTION SHEETS for Alpha, Bravo,
and Charlie. Ensure that the correct instruction sheet goes to the correct station.
Explain that the instruction sheet explains their specific roles more fully. Tell
them that this information is their particular responsibility, no one else will have this
information. It is their area of expertise. They can however share this expertise once
the task is underway, by communicating through the computer.

Give them 8 minutes to read the handbook and instruction sheet. Remind them that
they KEEP the INDIVIDUAL SHEETS but not the General Overview booklet.

Ask them: “Are there any questions concerning anything else presented so
far?” Answer any questions that the subjects have.

SET THE TIMER FOR 8 MINUTES & RETURN TO BEGIN PHASE II]

AFTER 8 MINUTES

WHEN YOU RETURN ASK IF THEY HAVE ANY QUESTIONS

REVIEW THE INDIVIDUALIZED INSTRUCTION SHEET. TELL THEM
THAT: EACH TEAM MEMBER NEEDS TO MAKE A JUDGMENT ON

THEIR RULE BY COMPARING THE ATTRIBUTE VALUES TO THE

180

VALUES LISTED IN THE THREE ZONES ON THEIR INSTRUCTION
SHEET (S, M, I}. GIVE THE FOLLOWING EXAMPLES USING PAGE 2 OF
THE INSTRUCTIONS:

(A) MODERATE, MODERATE, MODERATE = moderately threat so WARN

(B) THREAT, THREAT, THREAT= very threatening so DEFEND

(C) MODERATE, MODERATE, THREAT= inbetween moderate and threat but
closer to moderate so READY

(D) MODERATE, THREAT, THREAT=inbetween moderate and threat but close to
threat so LOCK-ON, THEN POINT OUT THATzANY SAFE PIECE OF
INFORMATION WITHIN A RULE MAKES THE RULE SAFE. NOT
NECESSARILY THE OVERALL TARGET BECAUSE ALPHA NEEDS TO
CONSIDER WHAT EVERYONE SAYS ABOUT THEIR RULES. FOR
EXAMPLE IF BRAVO HAS ATTRIBUTES THAT ARE ALL
THREATENTNG HE/SHE WOULD RECOMMEND DEFEND. IF
CHARLIE HAS TWO ATTRIBUTES THAT ARE THREATENING AND
ONE SAFE HIS/HER RECOMMENDATION WOULD BE IGNORE.
ALPHA IF YOUR INFORMATION WAS ALL IN THE MODERATE ZONE
YOUR RULE SAYS WARN. SO NOW AL_PHA HAS THE FOLLOWING
INFORMATION A DEFEND FROM BRAVO AN IGNORE FROM
CHARLIE AND ALPHA’S RULE SAYS WARN. ALPHA NOW
CONSIDERS THIS INFORMATION AND SHOULD PROBABLY CHOOSE
WARN. (GIVE ANOTHER EXAMPLE WHERE BRAVO SAYS WARN,
CHARLIE SAYS DEFEND, & ALPHA’S INFORMATION SAYS DEFEND
(CORRECT ANSWER IS LOCK-ON)

 

 

 

 

 

 

PHASE III - INTERACTIVE COMPUTER TRAINING & TASK PERFORMANCE

There are m practice targets. Briefly explain that there are 5 practice trials and ......

«The first one lasts a long time (16 minutes) and is just for showing off all the features of
the game. Most of the actual games will last 60 secs although there are some longer ones.
«The second is designed to practice all the features of the game and to clarify any
misunderstandings. This target is also long, 5 minutes, although not as long as the first.
«The third practice target will also allow you to practice the features of the game. This game
is 3 minutes long.

--Prior to the fourth practice target, I will show you the most efficient way of exchanging
information. This target is designed to practice this method. This target is also 3 minutes.
-Finally, during the fifth target, you will practice the method of exchanging information I
taught you during the last target This target lasts a minute and a half.

Explain to the subjects: “I will be available during these practice targets to answer
questions and clear up any misunderstandings that you may have.”

“Follow along closely and don’t get ahead. This will ensure that we have time to cover
everything”. m and not the subject should call up the first target by pressing “enter” on
Alpha’s console (you do not need to enter the game number unless you’re starting some
place other than game number 1 in which case you would hit g then the game # then enter.)

1. Point out and explain the icons, game #, time clock, and pull-down menu bar.

181

First of all you are going to learn how to measure the characteristics of the aircraft
that you will encounter. You Measure attributes by using the mouse to open the
Measure menu. To measure, use the mouse to move the cursor to the menu bar
across the top of your screen and click once on “Measure” - this opens the Measure
box. Inside this box you see the THREE attributes, out of a total of NINE, that you
can measure. Use the mouse to point to one of the attributes and click on it. A
gray box will appear on the lower left portion of the screen showing the attribute
value. It will disappear in 3 seconds. You can make it go away faster by pressing
enter or clicking on “OK.” Now measure the rest of the attributes.

 

 

Afier you have measured the other attributes click on Measure again and then click
on the Measure Summary option at the bottom of the menu. This is a summary
box that will display any attributes that you have measured so far. Nothing will show
up in that box until you measure it. You can also hit F2 to look at the Measure
Summary box, which may be faster. MENTION TO THEM THAT THE F2 BOX
F REEZES THE CLOCK FOR 6 SECONDS.

Now you’ll notice that while you can measure three attributes, only ONE of these
pertains to your specific expertise. In other words, you need some information which
you cannot measure. Initially, you’re going to-have to ask your teammates for it --
and the easiest way to do this is by using the Query function. The Query function
allows you to ask others for specific target attributes that you may need or want.
Select the Query option and open this menu by clicking once, but proceed no
further. Please stay with me and don’t get ahead because I want you to send a
query to a particular team member so each of you will receive one.

Notice that the menu box contains all possible target attributes. Thus you can ask
any team member about any possible piece of information. DO NOT click on any of
them yet, I want to explain a little more. As soon as you click on any of them, you
will be asked who you want to Query, in other words, who do you want to ask for
the information you just selected, but you will only have 3 seconds to select the
person you want to ask. You need to quickly select who you want to query by
clicking on the station. Do NOT do so yet, until I tell you who to query.
Since this is the information you will need for the real targets: (DO THIS STATION
BY STATION)

Alpha should query Charlie for Corridor Status

Bravo should query Alpha for Speed

Charlie should query Bravo for Rate Change Altitude
Team members should quickly select the appropriate attribute by clicking on it and
then send that query to the above mentioned stations by clicking on the station in the
box on the lower right comer of the screen. If you don’t click on the station you
want it sent to quickly, after about 5 seconds, by default it will go to Alpha. If
Alpha doesn’t click soon enough it will default to the Bravo (Make sure everyone
has sent the Query to the correct person).

182

Now you’ll notice a er abbreviation at the bottom, left-hand comer of your screen.
That er stands for Query and tells you that someone sent you a Query, someone
is asking you for a piece of information. This QRY comes up, not because of what
you did, but because of what someone else did, that is, someone is querying YOU.
WHENEVER you see something at the bottom left-hand comer, in this case a
Query, you need to receive whatever is down there--someone is trying to
communicate with you. You respond to these indications by using the Receive
firnction. Everyone click on Receive and open up that box. Now inside you see a
RED Q, which stands for Query, then you see who the Query came from. Once you
click, the computer will show you what that person asked for. Make sure you pay
attention to who sent the Query & that they asked for because once that box
disappears, it can't be recalled.

Once someone asks you for a piece of information using Query, you’ll probably
want to reply to that Query by sending them what they asked for. You can do this
using the Transmit function. Select the Transmit option and open the menu. Notice
that the only attributes shown in this box are the attributes you have measured.
You CANNOT send your teammates attributes you cannot measure or haven't
measured yet. What I want you to do now is select an attribute and send it to the
same person that queried you. You should have the information requested, if you
don't remember who queried you, here's who you should transmit to: (DO THIS
STATION BY STATION). Alpha should transmit speed to Bravo; Bravo should
transmit Rate Change Altitude to Charlie; Charlie should transmit Corridor
Status to Alpha

Notice the Trn abbreviation in the lower left part of the screen. Now remember that
you use the Receive function anytime you see a message on the lower left portion of
the screen. Everyone go ahead and Receive this Transmit. After you receive the
Transmit, you can go to the measure menu and open up your Measure Summary
Box. You may also press the F2 key as a faster way to open your summary box. You
see that this transmitted information is now in your summary box. ONE THING TO
REMEMBER IS THAT YOU CAN NOT RETRANSMIT THIS INFORMATION
TO SOMEONE ELSE IF THEY ASK YOU FOR IT. HOWEVER YOU CAN
SEND IT TO THEM USING THE NEXT FUNCTION I’M GOING TO TELL
YOU ABOUT: THE TEXT MESSAGE FUNCTION.

Another important thing we need to discuss is how you, as a team, will communicate
during the course of the game. You can’t communicate verbally because we need to
be able to keep track of the processes you go through in making a decision - we’re
not going to tape record or videotape you, or anything like that. So, in order for us to
be able to keep track of your communication, we will allow you to communicate via
text messages sent through the computer.

ONE CAVEAT TO THIS IS THAT ALPHA IS THE TEXT COMMUNICATIONS
HUB, WHAT THIS MEANS IS THAT CHARLIE AND BRAVO CAN ONLY
PASS TEXT MESSAGES TO AND FROM ALPHA. IF CHARLIE AND BRAVO

183

WANT TO TALK TO EACH OTHER USING TEXT MAESSAGES THEY
HAVE TO DO IT THROUGH ALPHA. THIS ONLY GOES FOR TEXT
MESSAGES-NOT FOR QUERIES OR TRANSMITS.
Everyone open your Transmit menu, but do not click on anything inside. Notice the
Text option at the bottom. If you were to click on this option the computer would
ask you which station you wanted to send a text message to.
For the purposes of this demonstration, I’m going to tell you who I want you to send
one to.
Have the students introduce themselves by typing their names to the below
people: (also tell them not to backspace {spelling does not count} and don’t type
messages that go past the astrics or the game will crash and this will slow them
down from finishing the experiment on time). (AGAIN DO THIS STATION
BY STATION).

Alpha should send a text message to Bravo and then to Charlie

Bravo should send a text message to Alpha

Charlie should send a text message to Alpha

One thing I should tell you is that if time is running ou_t_, the prong stops the
message with 10 seconds left, so you need to start your message with enough
time time remaining in the game. If the computer sends the messagg, it does
NOT go through. If sending a message is real imp_ortang, you may want to do
this early in the trial.

Notice the Msg abbreviation in the lower left of the screen. This means someone has
sent you a message, so you use the Receive function to receive that message.
Everyone go ahead and receive this message. Once you click on the M for message
the computer will give you two options: Log or Don’t Log. By default the computer
will log the message, meaning it will get saved to a file similar to your Measure
Summary File. If everyone would click on the Measure menu I will show you that
file. At the bottom you see Log File. Go ahead and click on that. Now you see the
message that you’ve just received has shown up inside. At any point you can delete
messages in your log file using the delete button on the keypad. Also, F3 is a
shortcut to get to your log file - everyone press F3.

In addition to sending a text message to other people, you have the option of sending
one to yourself. Let’s say you wanted to write yourself a reminder or write down
something you wanted to remember later. For instance, Bravo and Charlie might
want to put in a reminder that they can only send text messages to Alpha, or Alpha
might want to remind him/herself that Bravo and Charlie can’t commrmicate with
each other using text messages. You do this by again clicking on Transmit, then on
Text, but then clicking Log File all the way to the right.

Have them again write such a message. Alpha write “bravo and charlie can’t
send text diretly to each other”. Bravo write “I have to pass text to Charlie
through Alpha” Charlie write “I have to pass text to Bravo through Charlie”.

Now if each of you presses your F3 key, you’ll note that what you just typed has
appeared inside.

184

IO.

11.

12.

13.

So far you’ve learned how to Measure attributes using the Measure fimction, how
to request information from someone else using the Query function, how to send
specific pieces of information to each other using the Transmit function, and how to
write to others and ourselves using Text Messages. The last thing to cover is how to
make Judgments.

At 30 seconds left, the clock will start to beep. That indicates that, if you haven’t
done so by now, Bravo and Charlie should think about sending in their judgments
in to Alpha. Your judgments must reach Alpha with enough time left to make a
team decision. Everyone except Alpha open the Judgment box. I’m not having
Alpha opens up the judgment box because once that happens anything that Bravo or
Charlie send in is locked out. That is another reason to send your judgments in early
-- if you wait you may get locked out and will not have input into the team choice.

To make a judgment you would NORMALLY click on the decision you believe is
correct. You would use the F2 to look at the attributes you are responsible for, refer
to the individual instruction sheet for values, and then decide on a judgment.
Remember one key is that if any of your cues is safe, it makes your rule safe. Not
necessarily the overall target, because Alpha must combine the judgments on the
rules to make a team decision.

For now I want you to CHOOSE DEFEND!!!! If you select a judgment and don’t
do anything else, the judgment will register in about 3 seconds. If you made a
mistake and want to change your judgment, you can click on Cancel and then
resubrrrit your judgment. Once the judgment registers, however, you cannot change
it even if there is time remaining. Note that the Icons indicate the judgment you are
making.

You can submit your judgment at any time during the trial, but if you fail to send a
judgment by the time the clock runs out your decision will be registered as a No Call
which equals an Ignore. And again, if an outlying station sends in a decision when
Alpha has the judgment menu box open, the decision will register as a No Call.
Therefore send in decisions with plenty of time left.

Have Bravo and Charlie look at Alpha’s screen, Alpha’s ICONS turn red and
indicate the recommendations. Alpha should consider this information along with
the information he or she has and then make a decision for the team.

Now its time for the Alpha to CHOOSE DEFEND!!!

The Alpha should now make a judgment. Once this happens you will see
another screen. (don’t say this but the next screen will be different for teams in
each experimental condition). I should note that the length of thse screens is
generally about 5 seconds and then the program will automatically go to the next
target. The feedback screens during the practice trials are longer so that I can explain

things to you.

185

14.

The feedback screen tells you what the team decision was, this is the decision Alpha
made. This decision is compared to the correct decision and gives you an outcome.
In this case your feedback tells you that you got a (HIT). Right below this
inforrntion are Bravo and Charlie’s recommendations. WHAT IS SAID NEXT
DEPENDS ON THE CONDITION.

ABRUPT CHANGE/GRADUAL CHANGE

##

li

CONDITION 81/85 (EASY GOAL - DELAYED FEEDBACK! - If you recall

from the handbook, your score for each target could range between a -2 and a 2
and even though you all have different roles, you should work hard as a team to
achieve a goal of AT LEAST .75 as an average score. This screen reminds you
of this goal. The goal is written out in text on the bottom of the screen to the
left. You can tell how well your team is doing relative to this goal by looking at
the average performance number on the bottom of the screen to the right.
Directly above this you get a summary of how well you did on the previous
trials. As in many situations in real life, however, you will not always get
immediate feedback on how well your team is doing. After trial 40 you will get
feedback after every 9 trials or so, and this feedback will only include average
performance feedback and the trial summaries. You will not get information on
each specific trial outcome, however.

CONDITION 82/86 (EASY GOAL - FULL FEEDBACK) - If you recall from

the handbook, your score for each target could range between a -2 and a 2 and
even though you all have different roles, you should work hard as a team to
achieve a goal of AT LEAST .75 as an average score. This screen reminds you
of this goal. The goal is written out in text on the bottom of the screen to the
left. You can tell how well your team is doing relative to this goal by looking at
the average performance number on the bottom of the screen to the right.
Directly above this you get a summary of how well you did on the previous
trials.

CONDITION 83/87 (QIFFICULT GOAL - DELAYED FEEDBACK) - Ifyou

recall from the handbook, your score for each target could range between a -2
and a 2 and even though you all have different roles, you should work hard as a
team to achieve a goal of AT LEAST 1.50 as an average score. This screen
reminds you of this goal. The goal is written out in text on the bottom of the
screen to the left. You can tell how well your team is doing relative to this goal
by looking at the average performance number on the bottom of the screen to '
the right. Directly above this you get a summary of how well you did on the
previous trials. As in many situations in real life, however, you will not always
get immediate feedback on how well your team is doing. After trial 40 you will
get feedback after every 9 trials or so, and this feedback will only include
average performance feedback and the trial summaries. You will not get
information on each specific trial outcome, however.

186

###

15.

16.

CONDITION 84/88 (QIFFICULT GOAL -- FULL F EEDBACIQ - If you

recall from the handbook, your score for each target could range between a -2
and a 2 and even though you all have different roles, you should work hard as a
team to achieve a goal of AT LEAST 1.50 as an average score. This screen
reminds you of this goal. The goal is written out in text on the bottom of the
screen to the left. You can tell how well your team is doing relative to this goal
by looking at the average performance number on the bottom of the screen to
the right. Directly above this you get a summary of how well you did on the
previous trials.

before we go on to the next target let me stress three important points:

(1) With 30 seconds it is very important that Bravo and Charlie make judgments
quickly so as to leave time for the Alpha to make the team decision.

(2) the computer stops text messages at ten seconds so get these done early in the
trial

(3) When the time is running down and stations have their judgments in, keep your
menu bar clear and be out of any boxes in order to procwd to feedback and the next
target.Also keep your menu bar clear and boxes closed in the SHORT feedbacks. To
make sure everything's clear you can simply hit the ESC or Return key. If you
don’t do this your computer will fieeze up and we may have to re-boot the system
which will slow you down. If a station notices negative time on the clock or you

notice that you are still in feedback when others proceed to the next game, you need
to contact a researcher immediately.

The next (the second) target will automatically appear on their screens. Tell the
subjects to practice measureing, querying, transmitting, receiving things and sending
text messages. (go around and make sure everyone practices and answer any
questions. GO AROUND TO EACH STATION AND WALK THROUGH
SENDING A QUERY, RECEIVING A QUERY, SENDING A TRANSMIT,
RECEIVING A TRANSMIT, AND TEXT MESSAGES. DO THIS SO THAT
THEY UNDERSTAND HOW TO GET INFORMATION THEY NEED. MAKE
SURE ALL PARTICIPANTS HAVE THE INFORMATION THEY NEED IN
THEIR ROLE. After the target, clear up any misunderstandings. Do this again for

the third practice target. It is critical that they know how to use the guen,
transmit and text functions and that if bravo and alpha need to talk to each
other thgy need to do it through alpha. Thgy also must understand how to
make judgments and decisions. The third tapget is an IGNORE because Qch

station has one safe cue. Maks sure thg all understand this. Stop the game if
necessagy in feedback to drive home this point. THE OUTCOME FOR THE

THIRD DECISION SHOULD BE IGNORE. THE REASON WHY IS BECAUSE
EACH STATION HAD ONE PIECE OF INFORMATION THAT WAS SAFE.
MAKE SURE THEY ALL KNOW THAT IF ONE PIECE OF INFORMATION IS
SAFE, IT MAKES THEIR RULE SAFE.

187

l7.

18.

In the feedback after the third practice trial, explain that there is one best way to go
about exchanging information. (YOU MAY HAVE TO HIT F9 TO PAUSE THE
FEEDBACK SO THE GAME DOESN’T BEGIN (cntl F 9)).

“One key to doing well in this game is to exchange information efficiently. The
reason why is because if you exchange information quicikly, you will have more
time to look at the attribute values and think about the decisions you should
make. What I’m going to do now is show you how to exchange information the
most efficient way possible. First, everyone should immediately measure
everything you have, that is measure all three pieces of information in you
measure box. Alpha, you should use the TRANSMIT function to send SPEED
to Bravo. Bravo, you should use the TRANSMIT function to send RATE
CHANGE ALTITUDE to Charlie. Charlie, you should use the TRANSMIT
function to send CORRIDOR STATUS to Alpha.” Once you receive this
information, you will have everything you need to make a recommendation in
your measure summary box (F2). Try and do this from now on beginning with
the next trial. I should point out that it is critical that everyone get all three
pieces of informatin for your role. Every piece of information is important
because one safe piece of information can make a rule safe.

"MAKE SURE PARTICIPANTS USE THIS WAY OF EXCHANING
INFORMATION DURING TARGETS 4 AND 5. This is critical. Make sure
subjects know how to exchange information using this structure, CORRECT THEM
IF NECESSARY DURING TARGETS 4 AND 5.

After the next two training targets are done the game will shift into the real targets.
Before this happens hit Control F9 to pause the game and tell them:

To make the simulation more realistic and interesting, there may be some
communications jamming in which information you send will not go through to
you team mates. If this happens to you, you must continue to do the best you
can to work around the problem. The jamming may mean that you may not be
able to use the most efficient procedure that I told you about. However, you still
should be able to exchange information using a different combination of
queries, transmits and text messages. Remember, it is critical that you get all
three pieces of your rule. If you don’t get all three pieces, it is veg; dangeroup to
make a recommendation because the one missing piece may be safe, and this
would change your recommendation drasical_ly if your other pieces were
threatening.

 

While there should be adequate time to exchange text messages during the actual
targets, sometimes you will want more time to communicate with each other.
Therefore, at certain points of the game you will have a long feedback screens for
this purpose. These feedback screens will last about a numute and a half and during
this time you will be able to use the text-message function to communciate with one
another. There is a sheet posted on the back of your carroll that indicates after

188

###

E

19.

20.

which targets the long feedback occurs. It is a good idea to use this time to
communicate with one another about problems you are having if you don’t
have enough time to do so in the actual games. If you forget to look at the sheet,
you’ll know there is a long feedback for communications if the next game
doesn’t appear after about 5 seconds.

Also tell them:

CONDITION 81/85 (EASY GOAL — DELAYED FEEDBACK) -- I just want

to remind you that even though you all have different roles, you should work
hard to achieve your goal of AT LEAST .75. You can keep track of how well
you are doing by comparing this goal to the periodic feedback you receive.

CONDITION 82/86 (EASY GOAL - FULL FEEDBACK) -- I just want to

remind you that even though you all have different roles, you should work hard
to achieve your goal of AT LEAST .75. You can keep track of how well you are
doing by comparing this goal the the feedback you receive.

CONDITION 83/87 jQIFFICULT GOAL - DELAYED FEEDBACK) -- I just

want to remind you that even though you all have different roles, you should
work hard to achieve your goal of AT LEAST 1.50. You can keep track of how
well you are doing by comparing this goal to the periodic feedback you receive.

CONDITION 84/88 (QIFFICULT GOAL - FULL FEEDBACK) - I just want

to remind you that even though you all have different roles, you should work
hard to achieve your goal of AT LEAST 1.50. You can keep track of how well
you are doing by comparing this goal to the periodic feedback you receive.

Pick up the general overview booklets. Ask them to turn over their individualized
instruction sheets. Give them the task knowledge test MAKE SURE YOU GIVE
THE RIGHT TEST TO THE RIGHT STATION. Give them a couple of minutes to
complete it. MAKE SURE YOU WRITE IN THE TEAM NUMBER ON THIS
TEST. Then give them the GREEN scantron sheet with the questionnaire. Tell them
to do the first two pages only. Tell them to fill in their PID numbers. Make sure they
use pencils, check to make sure they fill it out completely. Give them about 5 or so
minutes.

When they are finished, pick up the test and scantron. Remind them they are not to
talk to each other. Tell them that there are Q targets total. I ALSO WANT TO
REMIND YOU OF THREE THINGS:
I) try to use the most efficient way of exchanging informatin
2) there may be some communications jamming, and if you experience
this and can’t use the efficient method of communications I taught you,
you still need to find a way to get all of the information you need.

189

21.

22.

23.

24.

25.

26.

Remember you can’t just turn your head and talk to each other. Also
remember that it is critical to get all three pieces for your rule.
3) While you can send text messages to each other during the games,
there are long feedback’s for extended communications.
If any of you have problems with your keyboard, your screen, or mouse, let an
experimenter know because these types of problems are not part of the
experiment.
Wish them good luck, Control F9 to restart the game. Leave the room.

V. POST WERIMENT PROTOCOL

Pick up the individualized instruction sheets and give them back the questionnaire
and the GREEN scantron. Tell them to pick up were they left off. Give them as long
as they need to complete the instrument.

Give them a debriefing form. Thank subjects and answer any questions. Let them
know that they will be told if their team won any money at the end of the semester.

Note if there were station crashes or other noteworthy event in the log at the end of
this binder.

If there are any comments on the training protocol, or a reminder that you want
posted, write it on the next page and it will be added/posted.

The cleanup program will automatically backup files if you type in anything at the
prompt other than an N. Only type in N if there was a huge problem. If you type in
N, you must make sure you follow the instructions on the screen. In addition, you
will need to rename the 1team.log, 2team.log, and 3team.log files XXXlteamlog,
XXX2team.log and XXX3team.log (where XXX is the team number) and m_oy§
them to the c:\teambu subdirectory. If you type N at the prompt and do not do this
procedure, the log files will merge with the next team and this will cause problems.
Copy the log-file onto a disk for back-up. The log file will be in the c:\teambu
subdirectory.

File scantrons.

190

APPENDIX F

191

APPENDIX F

Task Knowledge Test

(Alpha) Team#

 

Please answer the following questions by circling the correct Answer.

1. I am responsible for making judgments on rule Alpha which includes this attribute:
(a) Altitude
(b) Heading Crossing Angle
(c) Radar Cross Section
((1) Range
(e) Electronic Security Measure

2. I am responsible for making judgments on Rule Alpha which includes this attribute, however, I cannot
directly measure this attribute:

(a) Rate Change Altitude

(b) Corridor Status

(c) Radar Cross Section

(d) Heading Crossing Angle

(e) Electronic Security Measure

3. This person is responsible for sending me the attribute I need for my rule but can’t directly measure:
(a) Bravo
(b) Charlie

4. I am responsible for transmitting an attribute to this person:
(a) Bravo
(b) Charlie

5. This is the attribute that I am responsible for transmitting:
(a) Altitude
(b) Speed
(c) Radar Cross Section
((1) Range
(e) Electronic Security Measure

6. If the attributes in rule Alpha are threatening, threatening, and safe, my judgment based on rule Alpha
should be:

(a) Ignore

(b) Monitor

(c) Review

((1) Warn

(e) Ready

(f) Lock-on

(g) Defend

192

(Bravo) Team#

 

Please answer the following Questions by circling the correct answer.

1. I am responsible for making judgments on rule Bravo which includes this attribute:
(a) Number of Targets
(b) nge
(c) Radar Cross Section
(d) Altitude
(e) Electronic Security Measure

2. I am responsible for making judgments on rule Bravo which includes this attribute, however, 1 cannot
directly measure this attribute: ~

(a) Conidor Status

(b) Speed

(c) Radar Cross Section

((1) Number of Targets

(e) Electronic Security Measure

3. This person is responsible for sending me the attribute 1 need for my rule but can’t directly measure:
(a) Charlie
(b) Alpha

4. I am responsible for transmitting an attribute to this person:
(a) Charlie
(b) Alpha

5. This is the attribute that I am responsible for transmitting:
(a) Number of Targets
(b) Rate Change Altitude
(c) Radar Cross Section
((1) Range
(e) Electronic Security Measure

6. If the attributes in mle Bravo are threatening, threatening, and safe, my judgment based on rule Bravo
should be:

(a) Ignore

(b) Monitor

(c) Review

(d) Warn

(e) Ready

(f) Lock-on

(g) Defend

193

(Charlie) Team#

 

Please answer the following guestions bv circling the correct answer.

1. I am responsible for making judgments on rule Charlie which includes this attribute:
(a) Number of Targets
(b) Range
(c) Heading Crossing Angle
(d) Radar Cross Section
(e) Speed

2. I am responsible for making judgments on rule Charlie which includes this attribute, however, I cannot
directly measure this attribute:

(a) Number of Targets

(b) Rate Change Altitude

(C) Range

(d) Heading Crossing Angle

(e) Altitude

3. This person is responsible for sending me the attribute I need for my rule but can’t directly measure:
(a) Alpha
(b) Bravo

4. I am responsible for transmitting an attribute to this person:
(a) Alpha
(b) Bravo

5. This is the attribute that I am responsible for transmitting:
(a) Number of Targets
(b) Corridor Status
(c) Heading Crossing Angle
(d) Range
(e) Altitude

6. If the attributes in rule Charlie are threatening, threatening, and safe, my judgment based on rule Charlie
should be:

(a) Ignore

(b) Monitor

(c) Review

((1) Warn

(e) Ready

(0 Lock-on

(g) Defend

194

APPENDIX G

195

APPENDIX G

Post Experiment Questions

Please read and a_pgwer the following Questions. Please mark only one response to each

guestion on the bubble sheet with a pencil. Please do NOT write on this sheet.

1. How difficult was your assigned goal?

1 2 3 4 5
Very Somewhat Neutral Somewhat Very
Easy Easy Difficult Difficult

2. How challenging was your assigned goal?

1 2 3 4 5
Very Somewhat Neutral Somewhat Very
Unchallenging Unchallenging Challenging Challenging

3. How much feedback did you receive during the task on how well your team was
doing?

1 2 3 4 5
Very A Little Neither a Much Very
Little Little nor Much

a lot

4. How well informed were you during the task about how well your team was doing?

1 2 3 4 5
Very A Little Neither a Much Very
Little Little nor Much

a lot

196

5. How much did communications breakdowns (if any occurred) hurt your team’s
performance?

1 2 3 4 5 6
No No Very Some High Very High
Communications Negative Little Negative Negative Negative
Breakdown Impact Negative Impact Impact Impact

At All Impact

6. How much did communications breakdowns (if any occurred) hurt your team’s
chances of achieving its goal?

1 2 3 4 5 6
No No Very Some High Very High
Communications Negative Little Negative Negative Negative
Breakdown Impact Negative Impact Impact Impact

At All Impact

7. How hard did you work in this task?

1 2 3 4 5
Not Hard Not Very Somewhat Hard Very
at All Hard Hard Hard

8. As a percentage, how much of your effort did you put into this task?

1 2 3 4 5
0-19% 20-39% 40-59% 60-79% 80-100%
9. If your team experienced any communications breakdowns, how hard did you

think about coming up with an alternative way of gathering information?

1 2 3 4 5
Not Hard Not Very Somewhat Hard Very
at All Hard Hard Hard

10. If your team experienced any communications breakdowns, how much of your
thinking went into coming up with an alternative way of gathering information (as a

percentage).

1 2 3 4 5
0-19% 20-39% 40-59% 60-79% 80-100%

197

The following questions refer to communications between ALPHA and BRAVO. Please
use the following scale.

1 2 3 4 5 6 7
0% 1-20% 21-40% 41-60% 61-80% 81-99% 100%

11. Please rate the unreliability of communications between Alpha and Bravo during
trials 4-13 in terms of the percentage of time the link failed.

12. Please rate the unreliability of communications between Alpha and Bravo during
trials 14-23 in terms of the percentage of time the link failed.

13. Please rate the unreliability of communications between Alpha and Bravo during
trials 24-33 in terms of the percentage of time the link failed.

14. Please rate the unreliability of communications between Alpha and Bravo during
trials 34-43 in terms of the percentage of time the link failed.

15. Please rate the unreliability of communications between Alpha and Bravo during
trials 44-53 in terms of the percentage of time the link failed.

 

16. Please rate the unreliability of communications between Alpha and Bravo during
trials 54-63 in terms of the percentage of time the link failed.

17. Please rate the unreliability of communications between Alpha and Bravo during
trials 64-73 in terms of the percentage of time the link failed.

18. Please rate the unreliability of communications between Alpha and Bravo during
trials 74-83 in terms of the percentage of time the link failed.

The following questions refer to communications between BRAVO and CHARLIE.

19. Please rate the unreliability of communications between Bravo and Charlie during
trials 4-13 in terms of the percentage of time the link failed.

20. Please rate the unreliability of commrmications between Bravo and Charlie during
trials 14-23 in terms of the percentage of time the link failed.

21. Please rate the unreliability of communications between Bravo and Charlie during
trials 24-33 in terms of the percentage of time the link failed.

198

l 2 3 4 5 6 7
0% 1-20% 21-40% 41-60% 61-80% 81-99% 100%

22. Please rate the unreliability of communications between Bravo and Charlie during
trials 34-43 in terms of the percentage of time the link failed.

23. Please rate the unreliability of communications between Bravo and Charlie during
trials 44-53 in terms of the percentage of time the link failed.

24. Please rate the unreliability of communications between Bravo and Charlie during
trials 54-63 in terms of the percentage of time the link failed.

25. Please rate the unreliability of communications between Bravo and Charlie during
trials 64-73 in terms of the percentage of time the link failed.

26. Please rate the unreliability of communications between Bravo and Charlie during
trials 74-83 in terms of the percentage of time the link failed.

The following Questions refer to communications between ALPHA and CHARLIE.

27. Please rate the unreliability of communications between Alpha and Charlie during
trials 4-13 in terms of the percentage of time the link failed.

28. Please rate the unreliability of communications between Alpha and Charlie during
trials 14-23 in terms of the percentage of time the link failed.

 

 

29. Please rate the unreliability of communications between Alpha and Charlie during
trials 24-33 in terms of the percentage of time the link failed.

 

30. Please rate the unreliability of communications between Alpha and Charlie during
trials 34-43 in terms of the percentage of time the link failed.

31. Please rate the unreliability of communications between Alpha and Charlie during
trials 44-53 in terms of the percentage of time the link failed.

32. Please rate the unreliability of communications between Alpha and Charlie during
trials 54-63 in terms of the percentage of time the link failed.

33. Please rate the unreliability of communications between Alpha and Charlie during
trials 64-73 in terms of the percentage of time the link failed.

34. Please rate the unreliability of commrmications between Alpha and Charlie during
trials 74-83 in terms of the percentage of time the link failed.

199

LIST OF REFERENCES

200

LIST OF REFERENCES

Abelson (1981). Psychological status of the script concept. American
Psychologist, 36, 715-729.

Ackerman, P. L. (1986). Individual differences in information processing: An
investigation of intellectual abilities and task performance during practice. Intelligence,
_1_(_)_, 101-139.

Ackerman, P. L. (1987). Individual differences in skill learning: AN integration of
psychometric and information processing perspectives. Psychological Bulletin, 102, 3-27.

Allison, G. (1971). Essence of decision: Explaining the Cuban missile crisis.

Boston: Little Brown.

Ammons, R. B. (1956). Effects of knowledge of performance: A survey and
tentative theoretical fourmulation. Joumgl of General Psychology, 54, 279-299.

Anderson, J. R. (1982). Acquisition of cognitive skill. Psychological Review, 89,
369-406.

Anderson, J. R. (1987). Skill acquisition: Compilation of weak-method problem
solutions. Psychological Review, 94, 192-210.

Argote, L., & McGrath, J. E. (1993). Group processes in organizations:
Continuity and change. In C. L. Cooper & I. T. Robertson (Eds), lntemational review of

industrial and organizational psychology (pp. 333-3 89). Chichester, UK: Wiley.

Argyris, C. (1976). Leadership, learning, and changing the status quo.
Organizational Dynamics, 4, 29-43.

Argyris, C., & Schon, D. A. (1978). Organizational learning. Reading, MA:
Addison-Wesley.

201

Ashford, S. J ., & Cummings, L. L. (1983). Feedback as an individual resource:
Personal strategies of creating information. Organizational Behavior and Human
Performance 32 370-398.

 

Atkinson, J. W. (1957). Motivational determinants of risk-taking behavior.
Psychological Review. 64. 359-372.

Bachelder, B. L.. & Denny, M. R. (1977). A theory of intelligence: I. Span and
the complexity of stimulus control. Intelligence, 1, 127-150.

Bachelder, B. L., & Denny, M. R. (1977). A theory of intelligence: 11. The role of
span in a variety of intellectual tasks. Intelligence, 1, 237-256.

Baddeley, A., & Hitch, G. J. (1974). Working memory. In G. Bower (Ed),
Advances in learning and motivatin (V 01. 8, pp. 47-90). New York: Academic Press.

Bandura, A. (1991). Social cognitive theory of self-regulation. Organizational
Behavior and Human Decision Processes 50 248-287.

 

Bandura, A., & Cervone, D. (1986). Differential engagement of self-reactive
influences in cognitive motivation. Organizational Behavior and Human Decision
Processes 38 92-113.

 

Barber, A. E., Hollenbeck, J. R, Tower, S. L., & Phillips, J. M. (1994). The
effects on interview focus on recruitment effectiveness: A field experiment. J__oumal of

Applied Psychology, 79, 886-896.

Barley, S. R. (1986). Technology as an occasion for structuring: Evidence fiom
observations of CT scanners and the social order of radiology departments.

Administrative Science Quarterly, 31, 78-108.

Barnard, C. I. (1938). The functions of the executive. Cambridge, MA: Harvard
University Press.

Barrick, M. R., & Mount, M. K. (1991). The big five personality dimensions and
job performance: A meta-analysis. Personnel Psychology, 44, 1-26.

202

Barry, B., & Stewart, G. L. (1997). Composition, process, and performance in
self-managed groups: The role of personality. Journal of Applied Psychology, 82, 62-78.

Bazerrnan, M. H., Mannix, E. A., & Thompson, L. L. (1988). Groups as mixed-
motive negotiations. Advances in Group Processes, 5, 195-216.

Becker, L. J. (1978). Joint effect of feedback and goal setting on performance: A
field study of residential energy conservation. Journal of Applied Psychology, 63, 428-
433.

Beckhard, R. (1989). Optimizing team-building efforts. In French, W. L., Bell, C.

H., Jr., & Zawacki, R. A. (Eds) Organization development: Theogy. Practice, and
resegch 13rd ed.). pp. 182-189. Homewood, IL: BPI Irwin.

Behling, O., Coady, N., & Hopple, T. G. (1967). Small group adaptation to
unprogrammed change. Organizational Behavior and Human Performance, 2, 73-83.

Berkowitz, L., & Donnerstein, E. (1982). Extemeral validity is more than skin
deep: Some answers to criticisms of laboratory experiments. American Psychologisg, 37,
245-257.

Bettenhausen, K. L. (1991). Five years of group research: What we have learned
and what needs to be addressed. Journal of Management, 17, 345-381.

Beyerlein, M. M., Johnson, D. A., & Beyerlein, S. T. (1995). Advances in
interdisciplipgy studies of work teams (V 01. 2, pp. ix-xiii). Greenwich, CT: JAI Press.

Borgotta, E. F. (1964). The structure of personalin characteristics. Behavioral
Science 12 8-17.

 

Boulding, K. E. (1956). General systems theory: The skeleton of science. General

Systems. Yearbook of the socieg for the advancement of general systems theogy, 1, 11-
17.

203

Brehmer, B., & Hagafors, R. (1986). Use of experts in complex decision making:
A paradigm for the study of staff work. Organizational Behavior and Human Decision
Processes. 38, 181-195.

Brown, J. (195 8). Some tests of the decay theory of immediate memory. British
Journal of Psychology, 10, 12-21.

 

Bryan, W. L., & Harter, N. (1897). Studies in the psysiology and psychology of
the telegraphic language. Psychological Review, 4, 27-53.

Bryan, W. L., & Harter, N. (1899). Studies on the telegraphic language: The
acquisition of a hierarchy of habits. Psychological Review, 6, 345-375.

Byme, J. A. (1993, Sept. 11). Why several heads are better than one. Business
Week, pp. 45-46.

 

Campbell, D. J., & Gingrich, K. F. (1986). The interactive effects of task
complexity and participation on task performance: A field experiment. annizational
Behavior and Human Decision Processes 38 162-180.

 

Campion, M. A., & Lord, R. G., (1982). A control systems conceptualization of
the goal-setting and changing process. Organizational Behavior and Human Decision
Processes 30 265-287.

 

Cantor, N. & Kihlstrom, J. F. (1987). Personalig and social intelligence.
Englewood Cliffs, NJ: Prentice-Hall.

Carver, C. S., & Scheier, M. F. (1981). Attention and self-reguation: A control-
theory approach to human behavior. New York: Springer-Verlag.

Carver, C. S., & Scheier, M. F. (1982). Control Theory: A useful conceptual
framework for personality-social, clinical and health psychology. Psychological Bulle_trp,'
22, 111-135.

Cattell, R. B. (1957). Personaligy and motivation structure and measurement. New
York: World Book.

204

Chakravarthy, B. S. (1982). Adaptation: A promising metaphor for startegic
management. Academy of Management Review, 7, 73 5-744.

Chase, W. G., & Ericsson, K. A. (1981). Skilled memory. In J. R. Anderson (Ed.),
Cognitive Skills and Their Acquisition (pp. 141-189). Hillsdale, NJ: Erlbaum.

Cohen J. (1977). Statistical mwer analysis for the belrpvioral sciences LRev ed.).
New York: Academic Press.

 

Cohen, J ., & Cohen, P. (1983). Applied multiple regmssion/correlation analysis
for the behavioral sciences, 2nd ed. Hillsdale, NJ: Erlbaum.

Cohen, M. D. (1991). Individual learning and organizational routine: Emerging
connections. Organizational Science, 2, 135-139.

Cohen, M. D., & Bacdayan, P. (1996). Organizational routines are stored as
procedural memory: Evidence from a laboratory study. In M. D. Cohen and L. S. Sproull
(Eds), Organizational learning (pp. 403-429), Thousand Oaks, CA: Sage.

Cook, T. D., & Campbell, D. T. (1979). Qpa_si-exmrimentation: Desigp &
analysis issues for field settings. Boston: Houghton Mifflin.

Costa, P. T., & McCrae, R. R. (1992). Revised NEO Personalin Inventog
manual. Odessa, FL: Psychological Assessment Resources.

Costa, P. T., & McCrae, R. R., & Dye, D. A. (1991). Facet scales for
agreeableness and conscientiousness: A revision of the NBC personality inventory.

Personalig and Individm Differences, 12, 887-898.

Cummings, T. G. (1978). Self-regulating work groups: A socio-technical
synthesis. Academy of Mmement Review, 3, 625-634.

‘ Cyert, R. M., & March, J. G. (1992). A behavioral theog of the firm, 2nd ed.
Cambridge, MA: Blackwell.

205

Davis, L. E. (1977). Job design: Future directions. In L. E. Davis & J. C. Taylor
(Eds). Design of jobs (2nd ed. ). Santa Monica, CA: Goodyear.

DeMaris, A. (1995). A tutorial in logistic regression. Journal of Marriage and the
Family, 57, 956-968.

Dewey, J. (1922). Human nature and conduct: An introduction to social
psychology. New York: Holt.

 

Digman, J. M. (1990). Personality structure. Emergence of the five-factor model.
Annual Review of Psychology, 41, 417—440.

Dipboye, R. L., & Flanagan, M. F. (1979). Research settings in industrial and
organizational psychology: Are findings in the field more generalizable than in the
laboratory? American Psychologist, 34, 141-150.

Dodrill, C. B. (1983). Long term reliability of the Wonderlic Personnel Test.
Journal of Consulting and Clinical Psychology, 51, 316-317.

Dumaine, B. (1994). The trouble with teams. Fortune 130 86-93.

 

Earley, P. C., Northcraft, G. B., Lee, C., & Lituchy, T. R. (1990). Impact of
process and outcome feedback on the relation of goal setting to task performance.

Academy of Management Journal, 33, 87-105.

Earley, P. C., & Perry, B. C. (1987). Work plan availability and performance: An
assessment of task strategy priming on subsequent task completion. Organizatinal
Behavior and Human Decision Processes 39 279-302.

 

Earley, P. C., Wojnaroski, P., & Prest, W. (1987). Task planning and energy
expended: Exploration of how goals influence performance. Journal of Applied

Psychology, 72, 107-114.

Eckblad, G. (1981). Scheme theon. New York: Academic Press.

206

Eyring, J. D., Johnson, D. S., & Francis, D. J. (1993). A cross-level units-of-
analysis approach to individual differences in skill acquisition. Journal of Applied

Psychology, 78, 805-814.

F iol, C. M., & Lyles, M. A. (1985). Organizational learning. Academy of
Management Review, 10, 803-813.

F iske, D. W. (1949). Consistency of the factorial structures of personality ratings
from different sources. Journal of Abnormal Social Psychology, 44, 329-344.

Fiske, S. T., & Taylor, S. E. (1991). Social cogm'tion (2nd ed.). New York:
McGraw-Hill.

F itts, P. M. (1964). Perceptual-motor skill learning. In A. W. Melton (Ed.),
Categories of human learning (pp. 243-285). New York: Academic Press.

Fox, J. (1991). Regpession diagpostics. Thoursand Oaks, CA: Sage.

Freeman, J. (1980). The unit problem in organizational research. In W. M. Evan
(Ed.), Frontiers in organization and management (pp. 59-68). New York: Praeger.

Galegher, J ., Kraut, R. E., & Egido, C. (1990). Intellectual teamwork: Social and
technological foundations of coomrative work. Hillsdale, NJ: Erlbaum.

Gersick, C. J. G. (1988). Time and transition in owrk teams: Toward a new model

of group development. Academy of Management Journal, 31, 9—41.

Gersick, C. J. G., & Hackman, J. R. (1990). Habitual routines in task-perforrning
groups. Organizational Behavior and Human Decision Processes, 47, 65-97.

Ginnett, R. C. (1990). The Airline Cockpit Crew. In J .R. Hackman (Ed.), Groups
that work (and those that don't): Creating conditions for effective teamwork. San

Francisco: Jossey-Bass.

207

Gioia, D. A. & Manz, C. C. (1985). Linking cognition and behavior: A script
processing interpretation of vicarious leaming. Academv of Management Review, 10,
527-539.

Gioia, D. A., & Poole, P. P. (1984). Scripts in organizational behavior. Academy
of Management Review, 9, 449—459.

Goldberg, L. R. (1992). The structure of phenotypic personality traits. American
Psychologist, 48, 26—34.

Gough, H. G. (1987). Manual: The California Psychological Inventon. Palo Alto,
CA: Consulting Psychologists Press.

Graen, G. (1976). Role-making processes within complex organizations. In M. D.

Dunnette (ed.) Handbook of industrial and organizational psychology (pp. 1201-1245).
Chicago: Rand McNally.

Guzzo, R. A., & Dickson, M. W. (1996). Teams in organizations: Recent
research on performance and effectiveness. Annual Review of Psychology, 47, 307-338.

Guzzo, R. A., & Shea, G. P. (1992). Group performance and intergroup relations
in organiztions. In M. D. Dunnette and L. M. Hough (Eds), Handbook of industrial and
organizational psychology, 2nd Edition, (V 01. 3, pp. 269-313). Palo Alto, CA: Consulting
Press, Inc.

Hackman, J. R. (1987). The design of work teams. In J. W. Lorsch (Ed.),
Handbook of organizational behavior (pp. 315-342). Englewood Cliffs, NJ: Prentice-Hall.

Hackman, J. R. (1992). Group influences on individuals in organizations. In M. D.
Dunnette and L. M. Hough (Eds), Handbook of industrial and organizational
psychology, 2nd Edition, (V 01. 3, pp. 199-267). Palo Alto, CA: Consulting Press, Inc.

Hackman, J. R., Brousseau, K. R., & Weiss, J. A. (1976). The interaction of task
design and group performance strategies in determining group effectiveness.

Organizational Behavior and Human Performance , 16, 350-365.

208

Hackman, J. R., & Morris, C. G. (1975). Group tasks, group interaction process,
and group performance effectiveness: A review and proposed integration. In L. Berkowitz

(Ed.), Advances in exmrimental social psychology: Vol. 9 (pp. 45-99). New York:

Academic Press.

Hammer, M., & Champy, J. (1993). Reengineering the copporation: A manifesto

for business revolution. New York: Harper Business.

Hardman, M. L., Drew, C. J., & Egan, M. W. (1996). Human Exceptionaligy:
Socieg, school, and family (5th ed). Boston: Allyn & Bacon.

Hartigan, J. A., & Wigdor, A. K. (1989). Fairness in employment testing: Validig

generalization, minorigy issues, and the general aptitude test batter. Washington, D. C.:
National Academy Press.

Hastie, R. (1986). Experimental evidence on group accuracy. In B. Grofrnan & G.
Owen (Eds), Decision research (V 01 2, pp. 129-157). Greenwich, CT: JAI Press.

Hedberg, B. L. T. (1981). How organizations learn and unleam. In P. C. Nystrom
& W. H. Starbuck (Eds), Handbook of organizational desigp (pp. 3-27). London: Oxford

University Press.

Hinsz, V. B. (1990). Cognitive and consensus processes in group recognition

memory performance. Journal of Personalig and Social Psychology, 59, 705-718.

-. ‘Hinsz, V. B., Tindale, R. S., & Vollrath, D. A. (1997). The emerging
conceptualization of groups as information processors. Psychological Bulletin, 121, 43-
64.

Hinsz, v. 13., Vollrath, D. A., Nagao, D. H., & Davis, J. H. (1988). Comparing the

structure of individual and small group perceptions. Intgmational Journal of Small (‘flup
Research 4 159-168.

 

Hirokawa, R. Y. (1990). The role of communication in group decision-making
efficacy. Small Group Research, 21, 190-204.

209

Hoerr, J. (1989, July 10). The pay-off from teamwork: The gains in quality are
substantial. Business Week pp. 56-62.

 

Hogan, R. (1986). Hogan Personalig Inventog. Minneapolis, MN: National
Computer systems.

Hogan, R., & Hogan, J. (1992). Hogan Personality Inventog manual. Tulsa:
Hogan Assessment Systems.

Hollenbeck, J. R. (1989). Control theory and the perception of work
environments: The effects of focus of attention on affective and behavioral reactions to
work. Organizational Behavior and Human Decision Processes. 43. 406-430.

Hollenbeck, J. R., & Brief, A. P. (1987). The effects of individual differences and
goal origin on goal setting and performance. Organizational Behavior and Human
Decision Processes 40 392-414.

 

Hollenbeck, J. R., Ilgen, D. R., Sego, D., Hedlund, J ., Major, D. A., & Phillips, J.
(1995). The multilevel theory of team decision-making: Decision performance in teams
incorporating distributed expertise. Journal of Applied Psychology, 80, 292-316.

Hollenbeck, J. R., Klein, H. J ., O’Leary, A. M., & Wright, P. M. (1989).
Investigation of the construct validity of a self-report measure of goal commitment.

Journal of Applied Psychology, 74, 951-956.

Hollenbeck, J. R., LePine, J. A., & Ilgen, D. R. (1996). Adapting to roles in
decision-making teams. In K. R. Murphy (Ed), Individual differences and behavior in
organizations, pp. 300-333. San Francisco: Jossey-Bass.

Hollenbeck, J. R., & Williams, C. R. (1987). Goal importance, self-focus, and the
goal setting process. Journal of Applied Psychology, 72, 204-211.

Howard, A. (1995). A Framework for work change. In A. Howard (Ed.), _T_h__e
changing nature of work (pp. 3-44), San Francisco, CA: Jossey Bass.

210

Hough, L. M. (1992). The big five personality variables-construct confusion:
Description versus prediction. Human Performance 5 139-155.

 

Hunter, J. E. (1986). Cognitive ability, cognitive aptitudes, job knowledge, and
job performance. Journal of Vocational Behavior 29 340-362.

 

Hunter, J. E., & Hunter, R. F. (1984). Validity and utility of alternate predictors of
performance. Psychological Bulletin, 96, 72-98.

Hutchins, E. (1996). Organizing work by adaptation. In M. D. Cohen and L. S.
Sproull (Eds), Organizational learning (pp. 20-57). Thousand Oaks, CA: Sage.

Hyland, M. E. (1988). Motivational control theory: An integrative framework.
Journal of Personalin and Social Psychology, 55, 642-651.

Ickes, W., & Gonzalez, R. (1994). “Social” cognition and social cognition: From
the subjective to the intersubjective. Small Group Research, 25, 294-315.

Ilgen, D. R., Fisher, C. D., & Taylor, M. S. (1979). Consequences of individual
feedback on behavior in organizations. Journal of Applied Psycholpgy, 64, 349-371.

Ilgen, D. R., & Klein, H. J. (1988). Organizational behavior. In M. Rosenzweig &
L. Porter (Eds), Annual review of psychology: Vol. 40 (pp. 327-351). Palo Alto, CA:
Greenwich, CT: JAI Press.

Ilgen, D. R., LePine, J. A., & Hollenbeck, J. R. (1997). Effective decision making
in multinational teams. In P.C. Earley & M. Erez (Eds), New Persmctives in

lntemational Industrial-Organizational Psychology. San Francisco: Jossey-Bass.

Ilgen, D. R., Major, 0., Hollenbeck, J. R., & Sego, D. (1994). Team research in
the 90’s. In M. Chemers & R. Aymon (Eds), Leadership and organizational effectiveness
(pp. 245-270). New York: Academic Press.

Jackson, D. N. (1974). Personalig Research Form manual. Port Huron, MI:
Research Psychologist Press.

211

Janis, I. L. (1972). Victims of groupthink: A psychological study of foreigh policy
decisions and fiascos. Boston, MA: Houghton Miflin.

Jensen, A. R. (1970). A theory of primary and secondary familial mental
retardation. In N. R. Ellis (Ed.), lntemational Review of Research in Mental Retardation
(V014). New York: Academic Press, 1970.

Jensen, A. R. (1977). An examination of culture bias in the Wonderlic Personnel
Test. Intelligence, 1, 51-64.

Jensen, A. R. (1986). Major contributions g: artifact or reality? Journal of
Vocational Behavior 29 301-331.

 

Kahneman, D. (1973). Attention and effort. Englewood Cliffs, NJ: Prentice-Hall.

Kanfer, R. & Ackerman, P. L. (1989). Motivation and cognitive abilities: An
integrative/aptitude-treatrnent interaction approach to skill acquisition [Monograph].

Journal of Applied Psychology, 74, 657-690.

Kanfer, R., Ackerman, P. L., Murtha, T. C., Dugdale, B., & Nelson, L. (1994).
Goal setting, conditions of practice, and task performance: A resource allocation

perspective. Journal of Applied Psychology, 79, 826-835.

Kanter, R. M. (1983). The change masters. New York: Simon & Schuster.

Kass, R. A., Mitchell, K. J ., Grafton, F. C., & Wing, H. (1983). Factorial validity
of the Armed Services Vocational Aptitude Battery (ASVAB), Forms 8, 9, and 10: 1981

Army applicant sample. Educational and Psychological Measuremeng, 43, 1077-1087.

Katz, D. & Kahn, R. L. (1978). The social psychology of organizations (2nd ed.).
New York: John Wilen & Sons

Keisler, S. & Sproull, L. (1982). Managerial Response to changing environments:
Perspectives on problem sensing from social cognition. Administrative Science

Quarterly, 27, 548-570.

212

Keman, M. C., & Lord, R. G. (1990). Effects of valence, expectancies, and goal-
perforrnance discrepancies in single and multiple goal environments. Journal of Applied

Psychology, 75, 194-203.

Ketchum, L. (1984). How redesigned plants really work. National Productivity
Review 3 246-254.

 

Klein, H. J. (1989). An integrated control theory model of work motivation.
Academy of Management Rgview. 14. 150-172.

Kluger, A. N., & DeNisi, A. (1996). The effects of feedback interventions on
performance: A historical review, a meta-analysis, and a preliminary feedback
intervention theory. Psychological Bulletin, 41, 254-284.

Larson, J. R., Jr., & Christensen, C. (1993). Groups as problem solving units:
Towards a new meaning of social cognition. British Journal of Social Psychology, 32, 5-
30.

Latane, B., Williams, K., & Harkins, S. (1979). Many hands make light the work:
The causes and consequences of social loafing. Journal of Personalin and Social

Psychology, 37, 822-832.

Latham, G. P., & Locke, E. A. (1991). Self-regulation through goal setting.
Organizational Behavior and Human Decision Processes, 50, 212-247.

Latham, G. P., & Saari, L. M. (1982). The importance of union acceptance for
productivity improvement through goal setting. Personnel Psychology, 35, 781-787.

Lemming, R. E. (1922). Tests and norms for vocational guidance at the 15-year-

old performance level. The Psychological Clinic: A Journal of Orthogenics for the
Normal Development of Eveg Child, 14, 193-217.

LePine, J. A., Hollenbeck, J. R., Ilgen, D. R., & Hedlund, J. (in press). The effects
of individual differences on the performance of hierarchical decision making teams:

Much more than g. Journal of Applied Psychology.

213

Levine, J. M., & Moreland, R. L. (1990). Progress in small grup research. Annual
Review of Psychology, 41 , 585-634.

Levine, J. M., Resnick, L. B., & Higgins, E. T., (1993). Social foundations of
cognition. Annual Review of Psychology, 44, 585-612.

Levinthal, D. A. & March, J. G. (1993). The myopia of learning. Strategic
Management Journal, 14, 95-112.

Levitt, B., & March, J. G. (1988). Organizational learning. Annual Review of
Psychology, 14, 319-340.

Lewin, K. (1951). Field theory in social science. New York: Harper & Row.

Locke, E. A., Frederick, E., Buckner, E., & Bobko, P. (1984). Effect of previously
assigned goals on self-set goals and performance. Journal of Applied Psychology, 69,
694-699.

Locke, E. A., & Latham, G. P. (1990). A theog of goal setting and task
grformance. Englewood Cliffs, NJ: Prentice-Hall.

Locke, E. A., Shaw, K. N., Saari, L. M. & Latham, G. P. (1981). Goal setting and
task performance: 1969-1980. Psychological Bulletin, 90, 125-152.

Lord, R. G., & Hanges, P. J. (1987). A control systrns model of organizaitonal
motivation: Theoretical development and applied implications. Behavioral Science 32
161-178.

 

Lord, R. G. & Keman, M. C. (1987). Scripts as determinants of purposeful
behavior in organizations. Academy of Management Review, 12, 265-277.

Lord, R. G., & Maher, K. J. (1991). Cognitive theory in industrial and
organizational psychology. In M. D. Dunnette and L. M. Hough (Eds), Handbook of

Industrial and Organizational Psychology: Vol. 2, (2nd ed., pp.1-62). Palo Alto, CA:
Consulting Press, Inc.

214

Louis, M. R., & Sutton, R. I. (1991). Switching cognitive gears: From habits of
mind to active thinking. Human Relations 44 55- 76.

 

March, J. G., & Simon, H. A. (1992). Organizations, 2nd ed. Cambridge, MA:
Blackwell.

McCall, M. W. Jr., & Bobko, P. (1990). Research methods in the search of
discovery. In M. D. Dunnette and L. M. Hough (Eds), Handbook of Industrial and
Organizational Psychology: Vol. 1, (2nd ed., pp. 381-418). Palo Alto, CA: Consulting
Press, Inc.

McCrae, R. R. (1987). Creativity, divergent thinking, and openness to experience.
Journal of Personalig and Social Psychology, 52, 1258-1265. ,

McCrae, R. R., & Costa, P. T. (1987). Validation fo the five-factor model of
personality across instruments and observers. J oumal of Personalig and Social

Psychology, 52, 81-90.

McGrath, J. E. (1984). Groups: Interaction and pgrformance. Englewood Cliffs,
NJ: Prentice-Hall.

McKelvie, S. J. (1989). The Wonderlic Personnel Test: Reliability and validity in
an academic setting. Psychological Remrts, 65, 161-162.

Meyer, A. (1982). Adapting to environmental jolts. Administrative Science
Quarterly, 27, 515-537.

Miller, D., & Friesen, P. H. (1980). Momentum and revolution in organization
adaptation. Academy of Management Journal, 23, 591-614.

Miller, G. A. (1956). The magical number seven, plus or minus two.
Psychological Review, 63, 81-97.

Miller, J. G. (1955). Toward a general theory for the behavioral sciences.
American Psychologig 10, 513-531.

215

Miller, J. G. (19653). Living systems: Basic concepts. Behavioral Science 10
193-237.

 

Miller, J. G. (1956b). Living systems: Structure and Process. Behlaviorg Science.
fl. 337-379.

Mitchell, T. R., & Silver, W. S. (1990). Individual and group goals when workers
are interdependent: Effects on task strategies and performance. Journal of Applied

Psychology, 75, 185-193.

Mook, D. G. (1983). In defense of external validity. American Psychologisg, 38,
379-3 87.

Moray, N. (1967). Where is capacity limited? A survey and a model. Acta
Psychologicg, 27, 84-92.

Mikulincer, M. (1994). Human learned helplessness: A coping mrspective. New

York: Plenum Press.

Murray, H. A. (1938). Explorations in mrsonaligy. New York: Oxford University

Press.

Navon, D., & Gopher, D. (1979). On the economy of the human-processing
system. Psychological Review, 86, 214-255.

Nelson, R., & Winter, S. (1982). An evolutiopagy theog of economic change.
Cambridge, MA: Belknap.

Newel, A., & Simon, H. A. (1972). Human problem solving. Englewood Cliffs,
N. J .: Prentice-Hall.

Norman, W. T. (1963). Toward an adequate tasxonomy of personality attributes:
Replicated factor structure in peer nomination personality ratings. Journal of Abnormal

Social Psychology, 66, 574-583.

216

Norman, D. A., & Bobrow, D. B. (1975). On data-limited and resource-limited
processes. Cognitive Psychology, 7, 44-64.

O'Brien, G. E., & Owens, A. G. (1969). Effects of organizational structure on
correlations between member abilities and group productivity Journal of Applied

Psychology, 53, 525-530.

O’Leary-Kelly, A. M., Martocchio, J. J., & Frink, D. D. (1994). A review of the
influence of group goals on group performance. Academy of Management Journal, 37,
1285-1301.

Pagell, M. (1997). Unpublished sight visit notes. Michigan State University.

Pashler, H. (1991). Shifting visual attention and selecting motor responses:
Distinct attentional mechanisms. Journal of Exmrimental Psychology: Human Perception
and Performance 17 1023-1040.

 

Pashler, H. (1994). Dual-task interference in simple tasks: Data and theory.
Psychological Bulletin, 116, 220-224.

Paulhaus, D. L., & Martin, C. L. (1988). Functional Flexibility: A new conception
of interpersonal flexibility. Journal of Personalig and Social Psychology, 55, 88-101.

Peabody, D., & Goddberg, L. R. (1989). Some determinants of factor
representations of trait adjectives. Journal of Personalig and Social Psychology.

Peters, T. J., (1988). Thriving on chaos. New York: Knopf.

Peters, T. J. & Waterman, R. J ., Jr. (1982). In search of excellence: Lessons from
America’s best-um companies. New York: Harper and Row.

Peterson, L. R., & Peterson, M. J. (1959). Short-term retention of individual
items. Journal of Exgrimental Psychology, 58, 193-198.

Poole, M. 8., Holmes, M., & Descartis, G. (1991). Conflict management in a
computer-supported meeting environment. Management Science, 37, 926-953. '

217

Poole, M. S. & Roth, J. (1989a). Decision development in small groups IV: Test
of a contingency model. Human Communication Research 15 323-356.

 

Poole, M. S. & Roth, J. (198%). Decision development in small groups V: A
typology of group decision paths. Human Communication Research 15 323-356.

 

Poole, M. S., Seibold, D. R., & McPhee, R. D. (1985). Group decision-making as
a structurational process. Quarterly Journal of Speech, 71, 74-102.

Powers, W. T. (1973). Behavior: The control of mrception. New York: Aldine De
Gruyter.

Press, S. J., & Wilson, S. (1978). Choosing between logestic regression and
discriminant analysis. Journal of me American Statistical Associatipn, 73, 699-705.

Pritchard, R. D., Jones, S. D., Roth, P. L., Stuebing, K. K., Ekeberg, S. E. (1988).
Effects of group feedback, goal setting, and incentives on organizational productivity.

Journal of Applied Psychology, 73, 337-358.

Ree, M. J., & Earles, J. A. (1991). Predicting training success: not much more
than g. Personnel Psychology, 44, 321-332.

Reich, R. B. (1987). Entrepreneurship reconsidered: The team as hero. Harvard
Business Review, 65(3); 77-83.

Roethlisberger, F. J. & Dickson, W. J. (1939). Mmement 111.61 the worker.
Cambridge, MA: Harvard University Press.

Rousseau, D. M. (1985). Issues of level in organizational research: Multi-level
and cross-level perspectives. In L. L. Cummings and B. M. Staf (Eds), Research in
Organizational Behavior: Vol. 7, (pp. 1-37). Greenwich, CT: JAI Press.

Saavedra, R., Barley, P. C., Van Dyne, L. (1993). Journal of Applied Psychology,
]_8_, 61-72.

218

Salas, E., Dickinson, T. L., Converse, S. A, & Tannenbaum, S. I. (1992). Toward
and understanding of team performance and training. In R. W. Swezey & E. Salas (Eds).

Teams: Their training and performance (pp. 3-30). Norwood, NJ: Ablex.

Saville & Holdworth (1990). Occupational Personality Opestionnaire manual.
Esher, Surry: Saville & Holdsworth.

Schneider, W., & Shiffrin, R. M. (1977). Controlled and automatic human
information processing: 1. Detection, search, and attention. Psychological Review, 84, 1-
66.

Shiffiin, R. M. (1976). Capacity limitations in information processing, attention,
and memory. In W. K. Estes (Ed.),_H_gndbookL of learning and cognitive processes:
Memog Processes ( 201 4). Hillsdale, N.J.: Erlbaum Associates.

Shiffrin, R. M., & Schneider, W. (1977). Controlled and automatic human
information processing: 11. Perceptual learning, automic attending, and a general theory.

Psychological Review, 47, 127-190.

Sniezek, J. A. (1992). Groups under uncertainty: An examination of confidence in

group decision making. Organizational Behavior and Human Decision Processes, 52,
124-155.

Sniezek, J. A. & Henry, R. A. (1989). Accuracy and confidence in group
judgment. Organizational Behavior and Human Decision Performance, 43, 1-28.

Starbuck, W. H., & Hedberg, B. L. T. (1977). Saving an organization from a
stagnating environment. In Hans B. Thorelli (ed.), Strategy + structure = mrformance
(pp. 249-258). Bloomington, IN: University Press.

Stasser, G., Taylor, L. A., & Hanna, C. (1989). Information sampling in structured
discussions of three- and six-person groups. Journal of Personalig and Social

Psychology, 57, 67-78.

Stasser, G., & Titus, W. (1985). Polling of unshared information in group decision
making: Biased information sampling during discussion. Journal of Personally and

Social Psychology, 48, 1467-1478.

219

Stasser, G., & Titus, W. (1987). Effects of information load and percentage of
shared information on the dissemination of unshared information during group

discussion. Journal of Personalin and Social Psychology, 53, 81-93.

Staw, B. M., & Boettger, R. D. (1990). Task revision: A neglected form of work
performance. Academy of Management Journal, 33, 534-559.

Steers, R. M. (1975). Task-goal attributes, achievement, and supervisory
performance. Organizational Behavior and Human Performance, 13, 292-302.

Steiner, I. D. (1972). Group process and productivig. New York: Academic

Press.

Steiner, I. D. (1986). Paradigms and groups. In L. Berkowitz (ed.), Advances in
Expgrimental Social Psychology, 19, 251-259.

Streufert, S., & Nogami, G. (1992). Cognitive complexity and team decision

making. In R. W. Swezey & E. Salas (Eds), Teams: Their training and Mormance (pp.
127-151). Norwood, NJ: Ablex.

Sundstrom, E., & Altman, I. (1989). Physical environments and work-group
effectiveness. In L. L. Cummings & B. Staw (Eds), Research in organizational behavior
(V 01 11, pp. 175-209). Greenwich, CT: JAI Press.

Sunstrom, E., DeMeuse, K. P., & Futrell, D. (1990). Work teams: Applications
and effectiveness. American Psychologisg, 45, 120-133.

Taylor, M. S., Fisher, C. D., & Ilgen, D. R. (1984). Individuals’ reactions to
performance feedback in organizations: A control theory perspective. Research in

Personnel and Human Resources Managemeng, 2, 81-124.

Thompson, J. D. (1967). Organizations in actions. New York: McGraw-Hill.

Thurow, L. C. (1983). Dangerous currents. New York: Random House.

220

Tindale, R. S. (1989). Group vs. individual information processing: The effects of
outcome feedback on decision making. Organizational Behavior and Human Decision
Processes. 44, 454-473.

Tjosvold, D., & Tjosvold, M. M. (1995). Cross-functional teamwork: The
channenge of involving professionals. In Beyerlein, M. M., Johnson, D. A., & Beyerlein,

S. T. (Eds), Advances in interdisciplinm studies of work teams (V 01. 2, pp. 1-34).
Greenwich, CT: JAI Press.

Treece, J. B. (1992, Jan 20). Does Chrysler finally have the Jeep that it needs?
Business Week pp. 21-22.

 

Trist, E. L., & Bamforth, K. W. (1951). Some social and psychological
consequences of the longwall method of coal getting. Human Relations 4 3-38.

 

Tuckman, B. W. (1965). Developmental sequences in small groups. Paphological
Bulletin 63 384-399.

 

Tulving, E., & Schacter, D. L. (1990). Priming and human memory systems.
Science 247 301-306.

 

Tziner, A & Eden, D. (1985). Effects of crew composition on crew performance:
Does the whole equal the sum of its parts? Journal of Applied Psychology, 70, 85-93.

US. Department of Education (1994). Sixteenth annual remrt to Congpgss on the

implementation of the Eduggtion for All Hang'capm Children Act. Washington, DC:
US. Government Printing Office.

Van de Ven, A. H., Delbecq, A. L., & Koenig, R. (1976). Detemrinants of
coordination modes within organizations. American Sociological Review, 41, 322-328.

Volh'ath, D. A., Sheppard, B. H., Hinsz, V. B., & Davis, J. H. (1989). Memory
performance by decision-making groups and individuals. Organizational Behavior and
Human Decision Processes 43 289-300.

 

von Bertalanffy, L. (1968). General system theog. New York: Braziller.

221

Von Cranach, M., Ochsenbein, G., & Valach, L. (1986). The group as a self-
active system: Outline of a theory of group action. Euroman Journal of Social

Psychology, 16, 193-229.

Wagner, J. A., III. (1994). Participation’s effects on performance and satisfaction:
A reconsideration of research evidence. Academy of Management Review, 19, 3 12-3 30.

Walsh, J. P., & Ungson, G. R. (1991). Organizational memory. Academy of
Management Review, 16, 57-91.

Walton, R. E. (1985). From control to commitment in the workplace. Harvard
Business Review. 612), 76-84.

Weeless, V. E., & Serpento, S. T. (1982). An analysis of the Wonderlic Personnel
Test (Form A) as a factor in hiring clerical arsonnel in institutions of higher education.

Norgantown, West Virginia: West Virginia University, Office of Institutional Research.

Wegner, D. M. (1987). Transactive memory: A contemporary analysis of the
group mind. In B. Mullen & G. R. Goethals (Eds), Theories of ggoup behavior (pp. 185-
208). New York: Springer-Verlag.

Weick, K. E. (1979). Cognitive processes in organizations. In B. M. Staw & L. L.

Cummings (Eds), Research in Organizatipnal thavior: Vol 1. (pp. 41-74). Greenwich,
CT: JAI Press.

Weick, K. E., & Roberts, K. H. (1993). Collective mind in organizations: Heedful
interrelating on flight decks. Administrative Science Quarterly, 38, 357-381.

Weiner, B. (1980). Human motivation. New York: Holt, Rinehart, & Winston.

Weiner, N. (1948). Cybernetics: Control and communication in the animal and the
machine. Cambridge, MA: MIT Press.

Weingart, L. R. (1992). Impact of group goals, task component complexity, effort,
and planning on group performance. Journal of Applied Psychology, 77, 682-693.

222

Weingart, L. R., & Weldon, E. (1991). Processes that mediate the relationship
between a group goal and group member performance. Human Performance 4 33-54.

 

Weiss, H. M. (1990). Learning theory and industrial and organizational
psychology. In M. D. Dunnette and L. M. Hough (Eds), KsndbookLof industrial and
organizational psychology, 2nd Edition, (V 01. 1, pp. 171-221). Palo Alto, CA: Consulting
Press, Inc.

Weiss, H. M., & Ilgen, D. R. (1985). Routinized behavior in organizations. Ih_e
Journal of Behavioral Economics 24 57-67.

 

Weldon, E., Jehn, K. A., & Pradhan, P. (1991). Processes that mediate the
relationship between a group goal and improved group performance. Journal of

Personalig and Social Psychology, 61, 555-569.

Weldon, E., & Weingart, L. R. (1988). Group goals and group performance.
British Journal of Social Psychology, 32, 307-334.

Welsh, J. R., Jr., Watson, T. w.,& Ree, M. J. (1990). Armed Services Vocational

Aptitude Batteg (ASVAB): Integgative review of validig studies (AF HRL-TR-90-22).
Brooks Air Force Base, TX: US. Air Force Human Resources Laboratory.

 

Whyte, W. F. (1955). Money and motivation. New York: Harper.

Wonderlic & Associates (1983). Wonderlic Personnel Test manual. Northfield IL:
E. F. Wonderlic & Associates.

Wood, R. E. (1986). Task Complexity: Definition of the construct. Organizational
Behavior and Human Decision Processes 37 60-82.

 

Wood, R. E. & Locke, E. A. (1990). Goal setting and strategy effects on complex
tasks. In B. Staw and L. L. Cummings (Eds), Research in Organizational Behavior, Vol
fl (pp. 73-110). Greenwich, CT: JAI Press Inc.

Woodruff, D. (1991, Nov. 4). The racy Viper is already a winner for Chrysler.
Business Week pp. 91-92.

 

223

Woodruff (1993, Dec. 12). Chrysler’s Neon: Is this small car that Detroit couldn’t
bulid? Business Week 116-124.

 

Wright, R. E. (1995). Logistic Regression. In L. G. Grimm & P. R. Yarnold

(Eds), Reading and understanding multivariate statistics (pp. 217-244). American
Psychological Association, Washington DC.

Zaccaro, S. J ., Gilbert, J. A., Thore, K. K., & Mumford, M. D. (1991). Leadership
and social intelligence: Linking social perceptiveness and behavioral flexibility to leader

effectiveness. Leadership Quarterly, 2, 317-342.

Zander, A. W. (1971). Motives and goals in gIoups. New York: Academic Press.

lander, A. W. (1980). The origins and and consequences of group goals. In L.
F estinger (ed.), Retrosmctions on social psychology (pp. 205-235). New York: Oxford
University Press.

Zucker, L. G. (1977). The role of institutionalization in cultural persistence.
American Sociological Review, 726-743.

Zukier, H. & Pepitone, A. (1984). Social roles and strategies in prediction: Some
determinants of the use of base-rate information. Journal of Personaligy and Social

Psychology, 47, 349-360.

224

HICHIGnN STRTE UNIV. LIBRARIES
lllllHillllillll"IllIlllllllllllllllllll”Willi
31293016838892