3..

y.
.

!.v.w:.nv

; .4.

$513" L; “nan

‘ ¥ i
1.

1.3.1.. 1
Juli l I.)

4 Lu
;. 1.. .a L. c...) It;
V Afi. i.“ ‘4

{.:.I\\ .>
:.A .2. i

5...; iémflin

. ”yr :L.
0.0:.” .v, ‘5. . ‘ .

"ENQbEu§%% , . . I

 

lllllll

Ill

I llllllllllllgllm

312930

 

1

This is to certify that the

thesis entitled

COMMUNICATION IN TEAMS: HOW PATTERNS OF
COMMUNICATION INFLUENCE TEAM EFFECTIVENESS

presented by

Lori D. Sheppard

has been accepted towards fulfillment
of the requirements for

M. A. degree in Psychology

 

 

/,

Major professor

Date ”fr/7V

0.7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

 

 

LIBRARY 7

- Mlchigan Stat .
University

 

 

 

 

PLACE IN RETURN BOX
to remove this checkout from your record.
TO AVOID FINE; return on or before date due.

 

MTE DUE DATE DUE DATE DUE

 

*1

MAY25199
im24.

Seminal

AUQMzoW

 

 

WW

 

 

 

 

 

 

 

 

 

 

 

 

1M macs-m4

COMMUNICATION IN TEAMS: HOW PATTERNS OF
COMMUNICATION INFLUENCE TEAM EFFECTIVENESS

By

Lori D. Sheppard

A THESIS

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

MASTER OF ARTS
Department of Psychology

1998

ABSTRACT

COMMUNICATION IN TEAMS: HOW PATTERNS OF
COMMUNICATION INFLUENCE TEAM EFFECTIVENESS

By

Lori D. Sheppard

The frequencies of different communication patterns were examined to determine
the influence of communication on the effectiveness of teams decision making. The
verbal communications of 67 decision-making teams were recorded, transcribed, and
coded using an existing framework, and then sorted into the present framework using four
categories resulting in taskwork, teamwork, off-task social, and off-task other types of
communication. The different types of communications, such as the task related
communications, were hypothesized to be positively related to the accuracy of team
decisions, while the off-task types of communications were hypothesized to be negatively
related. Time was also expected to influence the types of communications to occur, such
that less time to make a decision would result in more on-task communication. in strategy
decisions late in the task. Correlational and regression analyses performed did not
support the hypotheses. The main conclusion of this study is that teams simply did not
engage in enough verbal communications, specifically the off-task communications, to

yield expected results.

ACKNOWLEDGMENTS

I would like to begin by thanking my committee member, Dan Ilgen, John
Hollenbeck, and Ann Marie Ryan. Special thanks to Dan, who served as my ever patient
and continually motivating chair. I couldn’t have done this without your encouragement
and recommendations. I think a special degree should be awarded for being able to
decipher some of your comments on the numerous drafts; but be assured the extra work to
understand what was written was certainly worth while (most of the time!).

I would also like to thank my cohorts. In those periods of time when motivation
and esteem were low, I know that I could always count on my friends for a boost (or some
commiseration). We’re almost there, ladies. Anyone want to go to Asian House or Lou
and Harry’s to celebrate? (Number 18, no rice noodles for Danielle if we go with the
former).

I would also like to thank my family. My dad, Bill Daniels, has served as my
inspiration for higher education. When I was a little girl, I was a bit confused by the title
of doctor, when there was no white coat involved. My mom, Judy, fills the medical side
of the picture. She offers inspiration as she continues her career path in life as a
registered nurse, a kind and patient woman and truly one of my best friends. I know to
some people it’s really scary to hear the words “you’re just like your mother,” but to me
this is not such a bad thing (except when it comes to misplacing things). To Heidi as
well, I hope that you are happy in life and pursue your dreams as I think you are a very
talented individual—you have much more creativity than I and look for beauty and

quality in many things that life has to offer. Thanks for being a great sister, Heidith.

iii

Finally, I would like to thank my husband, Drew. Speaking of following dreams,
you are one who continues to pursue your own dreams. Thank you for all of your support
and constant distractions from the grad student life. You are my rock as well as my sails.
I can only imagine where we’ll end up some day, but as long as we’re together that’s all

that matters. I love you.

iv

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................ vi
LIST OF FIGURES .......................................................................................................... vii
INTRODUCTION .......................................................................................................... 1
Communication .................................................................................................... 1
Time and Decision Making .................................................................................. 2
Purpose ................................................................................................................. 3
Teams ................................................................................................................... 4
Team Effectiveness .............................................................................................. 8
Taskwork and Teamwork ..................................................................................... 11
Summary .............................................................................................................. 21
Time ..................................................................................................................... 22
Hypotheses ........................................................................................................... 25
METHOD ......................................................................................................................... 33
Participants ........................................................................................................... 33
Task ...................................................................................................................... 33
Task Training ....................................................................................................... 35
RESULTS ......................................................................................................................... 4O
Coding .................................................................................................................. 41
Communication Dimensions ................................................................................ 44
Analyses Related to Hypotheses ........................................................................... 49
DISCUSSION ................................................................................................................... 54
Examination of Hypotheses .................................................................................. 54
Examination of Factor Structure .......................................................................... 56
Conclusion ............................................................................................................ 57
APPENDIX: TIDE: Simulation General Overview ........................................................ 62
REFERENCES ................................................................................................................. 71

TABLE 1 :

TABLE 2:

TABLE 3:

TABLE 4:

TABLE 5:

TABLE 6:

TABLE 7:

TABLE 8:

TABLE 9:

LIST OF TABLES

Teamwork Behaviors ...................................................................................... 19
Communication Categories ............................................................................ 25
Description of Voice Coding Schema ............................................................ 42
Rotated Component Matrix ............................................................................ 45
Communication Category Results: Percentage of Rater Agreement .............. 46

Descriptive Statistics and Intercorrelations Among Communication Scales . 48

Frequency of Communication Content by Type of Team .............................. 49
Hypothesis 3 Regression Results .................................................................... 51
Results for Hypotheses 4 and 5: Engagement of Strategy .............................. 53

vi

LIST OF FIGURES

FIGURE 1: Model ........................................................................................................... 26

FIGURE 2: Hypothesis 3: Off-Task Communication and Time Interaction ................... 3O

vii

INTRODUCTION

Few would argue the importance of teams. Teams are found in many places, from
the military to manufacturing plants. Teams provide many services, from flying aircraft
to new destinations to making decisions about public policy. While a common
occurrence in many organizations, there are both positive and negative experiences with
teams.

Many factors lead to effective or ineffective teams. Recently, there is
considerable interest in the social or interactive processes that influence team
effectiveness (e.g. McIntyre & Salas, 1995). Communication is one social process that
occurs in almost all teams. The present study seeks to extend existing research by
focusing on the role of communication in team effectiveness. The process of group
communication is commonly studied in the context of team development (e.g. Gersick,
I988). The effects of communication under circumstances other than development, such
as time limits or deadlines, are not commonly studied. The present study will look at

some of these other circumstances.

Communication Content

The information that team members share when they communicate can be
categorized into two basic, mutually exclusive content domains. These two domains can
be thought of as task and non-task related information. For teams established to perform

some specific task. it is obviously important for team members to share information about

the team task. However, these teams also may discuss non-task related information. The
non-task related information often is important for helping a group of individuals form
into an integrated team (Hackman, 1987). Both types of information, task and non-task,
may influence team effectiveness.

Although teams are used to perform a large number of tasks, they are often used
for decision-making. The effectiveness of decision-making teams is likely to be
influenced by task and non-task related communication. The present study will explore
the effects of communication from both domains (task and non-task) as they affect the

nature and quality of decisions within the decision-making team.

Time and Decision Making

Communication among team members occurs in a finite time frame. Acts of
communication are by their very nature mutually exclusive units. Specifically, at any
point in time, a team member may occupy the role of a communicator. This person, when
communicating, limits the extent to which other communication can occur. Therefore, if
a team member is talking about the task, this person is not at the same time talking about
non-task issues assuming, at this point, that every message falls in only one of the two
dimensions. Since this is a natural constraint, within a given block or subset of time the
proportion of communication devoted to one domain or the other can vary. If one aspect
of communication content occurs a high in proportion of the time, the proportion
represented by the other must be low. Thus, in a very real sense time serves as a

constraint that influences the number of messages that can be communicated within a

finite range of time.

Time as a constraint treats time in the short run. Finite units of time, for example,
five minutes or an hour, can be studied in terms of the number and nature of messages
communicated within the time limits. Yet, time can also be viewed over the long run for
a team. This is where time can be explored in a developmental sense. At a basic level,
teams have a beginning and an end. It is possible to examine the types of communication
that occur during the beginning versus the end of a team’s existence. In the longitudinal
setting, it is likely that team members communicate different types of information at the
aweginning of the team existence than information that is discussed at the end. For
example, at the beginning of a team’s existence, members may be new to the task and
find the need to frequently discuss the task. In contrast, after team members become
familiar with the task, these members may find more time to communicate non-task
related information. The differences in communication content over the entire existence

of a team will also be examined in this study.

Purpose

As stated earlier, the effectiveness of a team is likely to be affected by both task
and non-task communication. The present task is such that communication can occur
verbally or via computer. The amount of time available will moderate the extent to which
each type of communication occurs. This study will look closely at verbal
communications and the moderating effects of time available on the effects of each

communication content domain on team effectiveness.

In this proposal, a model will be presented in which it is suggested that particular
combinations of communication influence team performance. These communications are
examined with respect to time, both from the short term and the long term perspective.

In the remainder of the introduction, I will begin with a brief review of research
on teams relevant to team effectiveness, taskwork and teamwork. Communication is a
major part of the teamwork process. Next, I will lay out a taxonomy of interpersonal
communication. This taxonomy will be studied in three-person teams in a Naval
command and control simulation used to examine the relationship between
communication and team effectiveness. In the simulation, each team member monitors a
specific area for various incoming aircraft, and determines three key characteristics of the
aircraft that may pose a threat to the monitored area. The team members then
communicate information and decide how to deal with the incoming aircraft based on the
perceived level of threat posed by the aircraft. The communications between the team
members will be examined to determine the effects of the different content domains of

communication on the team effectiveness.

Teams

The terms “groups” and “teams” are often used interchangeably in the literature
(Kozlowski, Gully, McHugh, Salas, & Cannon-Bowers, 1996), but there is reason to
separate the two for research purposes. For instance, there are many different kinds of
groups and teams. Groups may be what Argote and McGrath (1993) refer to as “acting,”

groups that perform interdependent activities, or they may be “standing,” groups which

4

are simply aggregates of like individuals performing similar tasks independently (such as
a clerical staff). Definitions of teams typically are more consistent with active rather than
passive descriptions. For example, Morgan, Glickman, Woodard, Blaiwes, & Salas
(1986) define a team as “a distinguishable set of two or more individuals who interact
interdependently and adaptively to achieve specified, shared, and valued objectives”
(p.3). This definition is useful in that it shows that a team is comprised of people, these
people proceed to act interdependently, and the interaction propels the team to the
achievement of some specified goal. The research that follows in this study will look at
teams consistent with Morgan et al.’s definition.

The notion of interdependency among team members is critical to the distinction
between groups and teams, and is often neglected in research (Dyer, 1984). Dyer notes
that the tasks used in laboratory studies can often be performed by one person, and this
does not aid in our understanding of the interdependent nature of teams. In her 1984
review, Dyer lists several characteristics that separate teams, specifically military teams,
from other groups. These characteristics include 1) variations in group size, 2) the
potential for a team history which influences the expectations of team members and 3) the
potential for teams organized by task. In the latter case, members are selected because
they possess specific skills, which leads to the characteristic of assigning members to
roles within a formal team. Often the roles differ in rank (e.g. the teams have a leader).
These three characteristics separate teams from groups in several ways. The hierarchical
nature of certain teams allows for a wider variation in size, because the team can
encompass many different levels within the organization. While the present research will

hold the size of the team constant, it will incorporate role differences both in status and

task. Team histories can vary widely. A team can possess a history that spans many
years, even though some of the members may change. A history could also be a shorter
duration during which a team performs a task multiple times. Task performance is one of
the most salient features that distinguish teams from groups. In the present study, the
aggregates of individuals fit definitions of teams closer than that of groups.

Team tasks often influence how much members must interact (Dyer, 1984;
Hollenbeck, Ilgen, Sego, Hedlund, Major, & Phillips, 1995). The interaction is important
when the team members bring to the team setting different amounts of knowledges, skills,
and abilities to complete the task. Often teams are explicitly structured so that each
member has unique knowledge, skills, or abilities, not shared by all team members. Such
teams are said to possess distributed expertise (Hollenbeck et al., 1995). Distributed
expertise is encountered in the present research setting where teams will work on a
decision-making task.

Although a wide variety of tasks can be done by teams, those that involve some
form of decision-making are extremely common. Typically research has been directed at
one type of decision-making task--one that requires a consensus decision. With this kind
of task, the decision is based upon collective agreement among team members on a task
for which there are no objective standards for judging the correctness of the decision.
Davis (1992a) edited a special issue of Organizational Behavior and Human Decision
Processes devoted entirely to consensus decision research. In this issue, Davis (1992b)
provided a review of forty years of consensus decision research. Consensus decisions
include jury decisions, where a group of individuals are brought together to determine the

guilt or innocence of a peer. Another example of this type of decision is a managerial

6

team deciding upon the best way to carry out a particular project or determine the month’s
meeting agendas.

In contrast to consensus decisions are ones in which a leader or some other person
or subset of persons has the authority to make a decision for the whole team. In teams
that are hierarchical in nature, the decision is usually made by the team leader or
executive, and may or may not represent a consensus opinion. Hierarchies in team
structure often exist. For instance, this structure describes a flight crew which includes a
pilot and copilot, with the pilot clearly in charge of final decisions. Another example of a
hierarchical team is a surgical team with a chief surgeon and support staff. In each case,
the team is formed by members of different status; one member, the leader, has authority
over the other team members. In these teams, the leader is responsible for the overall
decision. It is the pilot’s call on what to do when a problem arises, or it is the surgeon
who decides whether to try a risky procedure in an emergency situation.

In all of the above examples, the tasks have a clear and correct decision option.
Safely landing an aircraft is clearly better than crashing! In some cases, the right
maneuvers or options may not be obvious at the time the decision must be made, but it
becomes clear when it is made. Decisions made by hierarchical teams often are of this
nature. By contrast consensus decisions often have no objectively correct solution. The
process by which the decision is made influences acceptance and often the “correctness”
of the decision. Since the present research is interested in hierarchical teams with
distributed expertise, it uses a task for which there is a correct decision.

While the leader may have the overall say in a decision, other team members may

offer needed support to the leader. As stated earlier, the other team members often

possess unique information or skills which they contribute to the team. These types of
teams have been described as hierarchically structured teams with expertise distributed
across team members (Hollenbeck et al., 1995). The specific knowledges, skills, or
abilities, or distributed expertise brought to a team by individual team members, are

often important in decision making teams, as individuals must share their specific
information to make the best possible decision. The sharing of information between team

members leads to interactive processes and the need for interpersonal communication.

Team Effectiveness

Effectiveness is a broad concept. The focus of team effectiveness has changed
over the years. Effectiveness has been defined in terms of organizational productivity
(e. g. Pritchard, 1992) or team satisfaction (e.g. Hackman & Morris, 1975), to name a few.
There is no commonly accepted defining characteristic of group or team effectiveness
(Goodman, Ravlin, & Schminke, 1987), yet effectiveness is often a concern for research
in teams (Goodman, Ravlin, & Argote, 1986, Goodman et al., 1987).

There are many sources from which to define effectiveness. One distinction that
has been made is the difference between effectiveness and factors leading to
effectiveness. For instance, Bettenhausen (1991) lists several empirical studies that
identify specific factors leading to or influencing effectiveness, such as individual effort.
cohesion, the organizational context, and intergroup relations, to name a few. Others
have created models of team effectiveness, thus defining effectiveness and providing the

factors that influence effectiveness.

Many of the models of team effectiveness follow the classic systems approach, of
inputs leading to processes, and processes leading to outputs (i -) p 9 o). For example,
Guzzo and Shea (1984. cf. Goodman et al., 1987) propose a model with three main
inputs, outcome interdependence, task interdependence and potency, with the process
being task-related interaction, and the output as simply group task effectiveness.
Hackman (1983, cf. Hackman, 1990) presents a more detailed model. Inputs include
organizational context and group design. Processes involve level of effort and amount of
knowledge. The outputs consist of group effectiveness. Another example of the i -) p ->
0 model is that of Gladstein (1984). She offers several group process characteristics, such
as open communication, supportiveness, conflict, discussion of strategy, weighting
individual inputs, and boundary management. These characteristics influence group
effectiveness, which is comprised of performance, and satisfaction (Gladstein, 1984).

While there is agreement on the overall conceptualization and use of models of
team effectiveness based on the i -) p -) o framework, these models are constrained by
the use of specific tasks. In other words, because many models of effectiveness deal with
specific tasks, the inputs are based on that task. The input side of the models face
disagreement primarily because of the wide range of tasks that can be faced by teams.
For example, there is no reason to expect that the skills and abilities needed to play
baseball well as a team overlap much with those needed for a surgical team to perform an
organ transplant. It is difficult to base a generalizable theory of team effectiveness on the
input side of the model due to the fact that the inputs are constrained by the specific task
at hand.

Where as it is difficult to pose a generalizable theory based on inputs, such is not

the case for processes, particularly as it relates to communication. Earlier, two

dimensions of communication were described, taskwork and teamwork. These two
domains capture much of the processes studied in teams. One can look at sociotechnical
theory to substantiate these dimensions. This theory perspective expanded the view of
work as purely mechanical, and contends that a group contains both technical and social
systems (from Guzzo & Shea, 1992). The technical and social systems lay out two main
condition concerns for team processes. One consideration for a team, and therefore the
processes, is the task (the technical). Organizational teams exist to perform some kind of
task. This task condition contends with basic task functions, such as the complexity of
the task, the available resources, and the time needed to complete the task. The other
consideration for the team is the interaction condition (the social). This condition deals
with the team as a whole, and can be conceptualized as the process of being a team. For
example, several authors include processes such as communication and cooperation (e.g.
Campion, Medsker, & Higgs, 1993; McIntyre & Salas, 1995) in their models of team
effectiveness.

While most work on team process identifies two major domains of process
behaviors, those directed at the task and those directed at interpersonal behaviors, there is
a third, less explored area that may also be important to the team. Hackman (1987) notes
that a separate process arises from the other group interaction process--synergy. Synergy
can be thought of as a form of social glue, and like most adhesives this glue can have a
positive effect if it works and a negative effect if it does not. This process results from
interactions occurring between team members, but does not relate directly to the task.
However, even though this is not related directly to the task, the process can have effects

on performance. For example, Hackman notes that in a positive sense this may lead to

10

creative and innovative performance. However, in a negative sense, the same process
could also lead to a decrease in performance. It is important to note that even though this
process occurs from the interactive processes relating to the task, it is not a task-specific
process. This is more of a social behavior that can not be tied closely to the task. In the
classification system introduced at the beginning of this proposal, these behaviors would
be considered off-task. The behaviors would include things like informal discussions
about friends, last night’s baseball game, or what one did over the weekend. These
behaviors can clearly occur in team settings, but can’t be tied directly to the task. As
Hackman implies, these may be positively (or negatively) related to team performance.
Neither the task nor the team behaviors that are frequently described in the literature
capture the off-task social behaviors of this type. As I look more closely at team
processes in decision—making teams, I will explore all three sets of behaviors--taskwork,
teamwork, and off-task social behaviors. At this point it is important to more fully

explain the dominant domains of team processes, taskwork and teamwork.

Taskwork and Teamwork

From the given examples of team effectiveness models, we see that team
characteristics are inputs. These inputs lead to important process interactions that take
place between team members. It has been suggested that research needs to view
effectiveness as a process (Sundstrom, Demeuse, & Futrell, 1990). This view leads us to
focus on the team processes and to a more thorough discussion of taskwork and

teamwork.

ll

Taskwork

There are many definitions for taskwork and teamwork. Taskwork generally
refers to the work on and knowledge of the team task (e.g. Cannon-Bowers, Tannenbaum,
Salas, & Volpe, 1995; Bowers, Salas, Prince, & Brannick, 1992). McIntyre and Salas
(1995) distinguish taskwork and teamwork by regarding the technical aspects of the team
performance as taskwork, and the interactive behaviors among team members as
teamwork. For simplification, this research will consider taskwork as those behaviors
pertaining directly to the team task. This is consistent with others’ views of taskwork

(e.g. McIntyre & Salas, 1995).

Teamwork

The aspect of taskwork is straightforward, in that taskwork involves the actual
performance of the task. Teamwork, in contrast to taskwork, is often more complex.
Definitions of teamwork range from generalizations to very detailed, listed descriptions
of teamwork characteristics. One general, all encompassing definition calls teamwork the
efforts that facilitate interactions to successfully complete taskwork (Rentsch & Hall,
1994). However, this is too general for practical purposes and not much different from
taskwork. This paper will focus on more specific definitions of teamwork, considering
three recent research effort on teams and teamwork. The three research efforts are:
McIntyre and Salas, 1995; Krahl and Dickinson, 1997; and Stevens and Campion, 1994.

McIntyre and Salas (1995) define teamwork as interactive team member behaviors

12

used to achieve desired team goals. They also provide a very detailed explanation of
teamwork complete with essential teamwork behaviors and enabling conditions for
teamwork behaviors. One caveat to their work is the assumption that organizations using
teams would benefit from practical guidance in this endeavor, even if this guidance is not
based on rigorous, empirical research. While this may be practical, the present research
employs the views of McIntyre and Salas in an effort to expand empirical research on
teamwork, using their following guidelines.

Guidelines from McIntyre and Salas (1995) offer several important or essential
behaviors for teamwork. The first of these behaviors is that a team member monitor the
behaviors of his or her teammates in addition to performing his or her own particular job
tasks. This monitoring function ensures that coworkers or teammates are getting the job
done by following correct procedures and not wasting time. The second behavior,
feedback, follows the monitoring behavior. Members let each other know both positive
and negative actions occurring in the team setting, and any member can give and receive
feedback.

Providing effective feedback involves a third essential teamwork behavior,
communication. Communication is effective when it follows a closed-loop pattern,
according to McIntyre and Salas (1995). This means that a sender initiates the message,
the receiver accepts the message and provides feedback, and the sender then double
checks the message to ensure that the intended message was received.

Following communication is the fourth essential team behavior, “back-up
behavior” (McIntyre & Salas, 1995, p. 26). This behavior is especially important to

teamwork because it shows that the team is working as a whole, or more than the sum of

13

the individual team members. When team members help each other, it implies that the
individual member has knowledge of his or her own job in addition to that of a team
member.

The next two team behaviors are what McIntyre and Salas call “enabling
conditions” (p. 27). These behaviors are more attitudinal in nature. The first is the team
viewing itself as a team, in that the success of the team depends on all of the members.
The second behavior is that of sustaining within-team interdependence. This behavior
rests on the degree to which team members believe that individual success is dependent
upon the success of the other team members.

There are two final behaviors essential for teamwork incorporating the notion of
changes occurring over time. The first is that team members exhibit flexible skills as a
function of the circumstances of the team, which can vary, and the second is simply that
teams change over time. While this last point seems obvious, it is worthy of
consideration for two reasons. The first is that, while it is commonly acknowledged that
many things change over time, research tends to take a static view of behavior due to
convenient measurement of a single point in time (although time is becoming more
prevalent in research conditions, e.g. Kozlowski et al., 1996). The second reason to
consider time with respect to teams is that all of the above team behaviors can be
increased or decreased over time, such as monitoring behavior and communication, and
these behaviors, in turn, may affect team performance positively or negatively. Thus, it is
important to assess time or the development of the team when considering the essential
teamwork behaviors.

Krahl and Dickinson (1997) provide a second perspective on teamwork. Rather

l4

than simply presenting a theory of teamwork, these authors actually test a model relating
the relationships and processes underlying teamwork. The major condition necessary for
teamwork, according to their theory, is the presence of task interdependencies, defined
here as the degree to which members must work together.

There are seven major tenants of teamwork in the Krahl and Dickinson (1997)
framework. The first of these is communication. To these authors, communication is a
process variable involving the active exchange of information as well as a linkage to the
other components of teamwork. A second component is coordination, or the merging of
individual actions. Next, we have team orientation and team leadership. Orientation
consists of the interrelationships formed in the teamwork process, and the resulting
attitudes towards others, while leadership is viewed in terms of initiation (leader focus on
the task) and consideration (leader focus on the emotional and social needs of team
members) behaviors of the leader (see Wagner & Hollenbeck, 1995 for a more detailed
explanation). Krahl and Dickinson’s final three teamwork components mirror
components offered by McIntyre and Salas (1995). These are monitoring, feedback and
backup behavior. Similar to that described above, monitoring is the observation of other
team member activities, feedback is the give-seek-receive cycle of information regarding
performance. and backup behavior is that behavior which assists other team members
with their tasks.

Krahl and Dickinson (1997) used a path model to test relationships between task
interdependence and the above teamwork processes. LISREL VIII analyses did not reveal
a significant relationship between task interdependence and communication, but several

other relationships did hold (e.g. monitoring behavior significantly impacts

15

communication). Overall, goodness-of-fit indices indicated that their model fit the data
very well.

A final view of teamwork is presented with Stevens and Campion (1994). These
authors focus on knowledge, skills, and ability (KSAs) requirements for teamwork.
These KSAs are classified in two major categories: Interpersonal and Self-Management.
Since relationships are present in team settings, one section of KSAs has an interpersonal
focus. A second focus on the self-management KSAs is needed due to the fact that many
organizational teams are given some degree of self-management (e. g. Hackman, 1986).
To employ the recommendations of Dyer (1984), the authors focus on the interpersonal
KSAs that are most desirable in the team setting. Stevens and Campion also give
meaningful operationalizations of these KSAs to assist future research.

The interpersonal KSAs proposed by Stevens & Campion (1994) contain three
major categories: conflict resolution, collaborative problem solving, and communication.
Conflict resolution KSAs are needed in situations where members have differences in
opinion, whether these differences refer to individual performance (or lack thereof) or the
overall team goal. The first three specific KSAs relate to this sub-category. They include
the KSAs to recognize desirable and discourage undesirable team conflict, to recognize
the type and source of conflict, and to employ an integrative (win-win) negotiation
strategy (Stevens & Campion, 1994).

Problem solving KSAs are useful in the team setting because teams face greater
problem solving demands than individual-based systems (Stevens & Campion, 1994). In
this subcategory. the collaborative problem solving KSAs are to identify situations

requiring participative problem solving, and to recognize obstacles to collaborative

l6

problem solving.

The final subcategory of the interpersonal KSAs involves communication. These
KSAs include: the KSA to understand communication networks; the KSA to
communicate openly and supportively; the KSA to listen nonevaluatively; the KSA to
maximize consonance between nonverbal and verbal messages; and the KSA to engage in
ritual greetings and small talk. The listing of communication KSAs is the most extensive
of all of the Stevens and Campion (1994) subcategories. Communication appears to
cover many different aspects of team behavior, from the task, in that communication
networks for sending information should be understood, to the non-task, with the ritual
greetings and small talk.

The second major category offered by Stevens and Campion (1994) is self-
management KSAs. The first two KSAs in this category deal with goal setting and
performance management. Specifically, team members need the KSA to establish
specific, challenging, and accepted team goals, as well as the KSA to monitor, evaluate,
and provide feedback on both overall team performance and individual performance. The
final two KSAs are needed to facilitate planning and task coordination. Team members
should possess the KSA to coordinate and synchronize activities and information, and
team members should possess the KSA to establish task and role expectations of
individual team members.

Stevens and Campion (1994) offer a new direction in team and teamwork research
by giving specific KSA requirements for teams. They note that “if the teamwork trend
continues, then the social aspects of work will become more important” (p. 524). Their

KSAs can be used in empirical research to help determine what is useful, or what can

17

help us determine what makes teamwork work.

 

An examination of the major points of the above discussion is shown in Table 1.
Table 1 shows that there is considerable overlap in the teamwork dimensions offered by
the three models reviewed in this proposal. Communication, monitoring. and feedback
are included in all three studies. Back-up behavior and a coordination component occur
in two of the three studies. These dimensions of teamwork will guide the teamwork

communication domain explored in this study.

18

 

 

 

 

 

 

Idiel
least

’

liflnt

llonn

Fted‘

Con

Bat

Ira

Susi

Table 1
Teamwork Behaviors

 

McIntyre & Salas, 1995

Krahl & Dickinson, 1997

Stevens & Campion, 1994

 

Monitoring

Feedback

Communication

Back-up Behavior

Team Recognition

Sustain team interdependence

Monitoring

Feedback

Communication

Back-up Behavior

Team Orientation

Team Leadership

Coordination

KSA to Monitor behavior

KSA to provide Feedback

Communication KSAs

Task Coordination KSAs
Conflict Resolution/Problem

Solving KSAs

 

l9

This study will examine taskwork and teamwork within the context of
communication. Taskwork will involve those communications that pertain directly to the
task, and will incorporate computer mediated as well as verbal task communications. The
inclusion of the computer task communications is explained more fully in the method
section of this study. While this is an exception to the fact that verbal communications
are the primary interest to this study, it is important to capture the task specific
information regardless of how it is passed, and this includes incorporating discrete task
information sent via the computer. This will not affect the off-task communications,
which occur verbally.

Teamwork communications are more complex. given the above review.
Teamwork will include those interpersonal communications that deal with monitoring
and providing feedback to team members, as shown in Table 1. Teamwork
communications will also include those communications that facilitate an important task
function; the function of how to perform the task.

In addition to the interpersonal nature of teamwork is an important task
consideration, which is _h_o_w_ to go about performing the task. Several studies examine
task strategies and strategy development (e.g. Earley, Connoly, & Ekegren, 1989;
Mitchell & Silver, 1990). One component of strategy is the need to coordinate
information (e.g. Mitchell & Silver, 1990). Both Krahl and Dickinson (1997) and
Stevens and Campion (1994) include coordination as specific dimensions of teamwork.
In an unfamiliar task situation, it is unlikely that a person or team holds a specific task

strategy. Therefore, team members must allocate some time for planning or devising a

20

strategy and coordinating member actions to maximize performance on the task (Kernan.
Bruning, & Miller-Guhde, 1994). There are several studies that have found a positive
relationship between strategy and performance in group or team situations (e. g. Hackman,
Brousseau, & Weiss, 1976; Weldon, Jehn, & Pradhan, 1991; Weingart & Weldon, 1991).

These studies support the notion that strategy can be an important part of performance.

Summary

It is the combination of the overlapping teamwork components, communication,
feedback, monitoring, and coordination (see Table 1) that brings together a unified
framework for the present study. Because the present study involves decision making
teams, the most critical teamwork dimension of interest is communication. Monitoring,
feedback, and coordination behaviors must take place via communication in the present
setting, in addition to sharing basic task information expertise. This makes the teamwork
dimension especially interesting, in that it is through teamwork that taskwork also occurs.

Thus far. this review has described teams, with specific consideration for
effectiveness in decision making teams. Effectiveness in an input 9 process 9 output
framework was described, with specific emphasis on the nature and overlap of team
process. Remember that the input side of a model of team effectiveness is strongly
dependent on the task. In a decision making team, team members must share
information. Sharing information is part of the process aspect of a model of
effectiveness. Taskwork and teamwork were described in greater detail as a direction for

generalization in team effectiveness research. Of interest in this research effort is the

21

process of communication, and the different domains of communication that occur in the
decision making team. Broadly, this communication falls into two domains: task and off-
task. More specifically, task communications are further broken down into two
categories: task specific, or Taskwork, and Teamwork. Teamwork consists of on-task
communications related to monitoring, feedback, coordination and strategy information.
The off-task communications are then divided into social information, as off-task social
communications, and anything else, which for completion are categorized as off-task
other communications. These categories are considered to define the domain of
communication messages that may occur in a team setting. Also, remember that
communication behaviors are in many respects related to a zero sum situation, in that if a
team is engaged in communicating one set of messages, they must, by definition not be
addressing the other. The result of this caveat is that these messages are constrained by
time. Therefore, before turning to the hypotheses, one final consideration must be

addressed: time.

:3.
3
(D

 

As noted earlier, teams performing tasks do so over a finite time interval. Often
team members make multiple decisions over time. From the time they are formed, team
members are learning about the task they must do and how to work with each other.
Most team research involving time examines the development of teams over time (e. g.
Gersick, 1988; Tuckman, 1965). However, there is another way to view time. This

perspective views the cyclical nature of team tasks. As Kozlowski et a]. (1996) note.

22

“dramatic, cyclical shifts in task complexity and pace are typical occurrences for teams
performing critical functions” (p. 4). In addition, they note that these shifts place pressure
on individual cognitive resources and limit explicit communication. The present research
explores this cyclical task process. In this process, one can examine times of low task
behavior and times of high task behavior. The cyclical task nature is then combined with
the notion of teamwork behaviors, specifically communication divided into taskwork and
teamwork as the task related communications, and off-task social and off-task other as the
non-task related communications. These communication behaviors can then be examined
over the lifetime of the team.

In addition to time as an overall span needed for a team to perform a task, one can
look at time in a finite sense within separate subtask elements within an entire team task.
More precisely, it was noted that decision making teams make multiple decisions over
time, and therefore each decision must be made over a small segment of time within the
overall time span of the team. Kozlowski et a1. (1996) noted the cyclical nature of tasks
and proposed that these smaller segments, or shifts, can place differing pressures on a
team. One aspect of this pressure can be treated as time pressure, or as an extensive
cognitive demand. Specifically in this research, these pressures are viewed as demand
situations. Demand is operationalized as either a specific decision that requires a great
deal of cognitive effort, or as a time demand in that the team is under time pressure to
make a decision quickly.

This research treats time in two ways. First, time is viewed in the sense that teams
exist over time, and team behaviors can be examined in this overall time span. Within

the communication framework, one can expect different types of communicationutask

23

specific, task strategy, or social—-to occur at different points over the life time of the team.
Second, because the different forms of communication occur in a system with limited
receiving capacity, communication tends to occur at any one time. This means that
within a bounded time frame, the types of communication are dependent in the way that
the items on an ipsative scale are dependent. Time constrains the pattern of
communication behaviors during any given time. During an instance of time, there may
be demands placed on the team in terms of (l) cognitive effort or (2) time demand (in
that there may be little time to make a decision). For example, in a decision making task
the decision may be very straight forward, following an “if / then” type rule such that
there is little cognitive effort placed on a team. On another occasion, the decision may
fall on many complex pieces of information that must be interpreted correctly first before
reaching a decision. This would be an instance of high cognitive demand. In an example
of time demand, it may be that the decision is very straight forward, simply combining
information for a rule as before. However, there could be substantially less time to gather
information, in that even though a decision is relatively easy, the correct decision may not

be reached because there is insufficient time to collect and share information.

24

Hypotheses

To briefly review, the position that has been taken here is that the critical
processes affecting performance in teams whose primary task is that of making decisions
involve the things about which the team members communicate. These communications
can be clustered into those that are focused on the task and those that are not (i.e., off-
task) and represent processes that go on among team members. The task specific
communications were further divided into those directed directly at working on their task,
or Taskwork, and the task monitoring, feedback, coordination and strategy
communications, or Teamwork. The off-task communications were divided into social
communications and, for completeness, all other communications. In sum, the prOposed
mutually exclusive processes are represented by the listing in Table 2.

Table 2
Communication Categories

 

 

 

Category Communication Content
Taskwork On-task, non-social:
Immediate Task information
Teamwork On-task, non-social: Monitoring, Feedback,
Strategy. Coordination
Social Off-task, social
Other Off-task, non-social

 

The model incorporating different dimensions of communication as they relate to

team performance is presented in Figure 1. The hypotheses to be presented below about

25

Emma _. zona—

 

fiBo“ 005.8303

 

wmqq <m. 5:0

 

 

 

 

Zeno”
<98
024652.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

48::
mmmmnaésomm
A0830:
>858ch

 

 

0033:3330:
:_+
Hum—A20} -
:_+
H3322: V
2.... IA .2815 ZmA—wnmv
mgnznmvc V
IN -
034.3.» Y
mega—55$ :w a
i- -
.339

0039:;

 

 

 

26

the effectiveness of quality of decisions made in teams as a function of the nature of their
communications are based on one assumption about time constraints and another about
the development of teams over time. First, as was mentioned above, the four dimensions
of communication are assumed to be primarily mutually exclusive. As a result, when
constrained by time, the four are, by necessity, interdependent. As time devoted to one
form of communication activity increases, that devoted to others must decrease when
time is held constant. The second assumption is that, over the life of the team, the best
mix of the key communications depends upon the point in the team’s development.
Some mixes are better earlier on, and others are better later. The following hypotheses
and the rationales for them address predicted relationships between the key
communication messages and the quality of the teams’ decisions. In all cases, the
accuracy of the team’s decision will be the measure of its quality.

Note that mode is the first input on the left side of the model (see Figure 1).
Mode refers to computer versus voice communications. Mode potentially influences the
volume of communication messages, and possibly the types of messages that occur. For
instance, if a team is restricted to the computer, then it is likely that the amount social
communications will be lower than those teams that can communicate orally. This is
because of the time restriction, in that typing a message usually takes longer than orally
giving the same message. In the sense of the mutually exclusive categories, if one is
typing some sort of task related message, one is not as the same time sending a message
that is social in content. Because the amount of social communications was expected to
be low as a result of restricting messages to the computer, voice communications then

would represent a more conducive environment to exchanging social types of

27

communication. Therefore, voice communications represent the primary interest in this
study, and thus the study was limited to using teams having the capacity to send voice
communications to each other. However, it is important to capture all instances of task
specific communications, and some teams had the opportunity to send task information
via the computer using single keystrokes. Therefore, the only instance of computer
communications used in this study involve the task specific information that is sent via
the computer using single keystrokes. With these considerations in mind, I will now
address the hypotheses included in this study.

Hypothesis 1: The frequency of on-task communications will be
positively correlated with the accuracy of the team decisions. Decision making tasks
of the kind in which these teams will be engaged demand that members gather and
process information prior to making each decision. To fail to gather and share
information with other team members will lead to decisions that are no better than the
base rate for guessing. Thus, Hypothesis 1 simply reflects the obvious that, if teams do
not devote time and effort to working on the task, they will not perform effectively, and
they are likely to perform worse than those that do focus on the task.

Hypothesis 2: The frequency of off-task communications will be negatively
correlated with decision making accuracy.
Hypothesis 2 is a simple extension of the first, in that it takes into account the dependency
among behaviors created by the limits of time. On- and off-task activities are totally and
negatively dependent, and therefore, Hypothesis 2 is the reflection of Hypothesis 1.
Hypothesis 3: The relationship between off-task communications and

decision making accuracy will be moderated by the time available to complete the

28

task.

Although by necessity, within a fixed time period, time that is devoted to off-task
behaviors must be at the expense of on-task behaviors, it does not necessarily mean that
there is remaining sufficient on-task time to perform the task well. Performance is
affected by the absolute amount of time needed to accomplish the task, not the amount
relative to other behaviors. Thus, the amount of time available to accomplish the task
should moderate the negative effect of spending time on off-task behaviors. Under high
time demands, the relationship should be negative; under low time demands it should not.

This interaction is depicted by Figure 2.

29

3.6.8 N. 355283 m” 9.3»? 0033:5830: man .238 38898:.

 

 

 

Em:
r02 :38
Y 8325
m 358 3:8 8 8:688 Hem—n
m
C
A
n
.m
S
.m
e
D
m Em: 2:8 8:83
T. 28mm 3:8 8 8:688 :55
reg Em:

r . .
oE Omrmmw 00585—823

3O

Hypothesis 4: There will be a positive correlation between engagement in
strategy development communications on the early trials and overall team
performance.

(The relationship of early to late trials is based on a comparison between the first quarter
of the experimental trials to the last quarter of the experimental trials). Hypotheses 1
through 3 above are based on the internal logic of the mutually exclusive communications
classification developed for this research. The hypotheses, while important, are simply
extensions of the logic of the classification. The rationale for Hypothesis 4 extends
beyond this classification system. In particular, it is based on findings with goal setting
on complex tasks.

The goal setting literature strongly supports the conclusion that specific, difficult,
attainable goals lead to higher performance than “do your best goals” (Locke & Latham,
1990). Yet, this relationship does not hold on complex tasks (Wood, Mento, & Locke,
1988; Earley, Lituchy & Connoly, 1993). A number of reasons are given for this, but the
most common invokes strategy development (Earley et al., 1993). It is argued that when
people are presented with complex tasks with which they have little experience and also
given specific, difficult goals, the goals focus attention on doing the task and distract
from thinking about ways or strategies for working on the task. Then, the failure to
develop good strategies ends up interfering with performance in the long run.

I have argued earlier that strategic activities in decision making teams share the
same kind of dependence with working directly on the task that has been observed in goal
setting tasks. If the team devotes all of its communications to dealing with immediate

task demands rather than spending some time addressing how they should work together,

31

the team’s performance should also suffer in the long run. Since strategic plans once
developed can be used throughout the lifetime of the team, the effect of failing to devote
time and effort to strategic planning early on in the life of the team should carry on over
time.

Hypothesis 5: There will be a negative correlation between engagement in
strategy development communications on late trials and overall performance.
Hypothesis 5 is a reflection of Hypothesis 4. One primary reason for this hypothesis
stems from the work on building expertise. Over time, as individuals become familiar
with a task, a type of task routine is learned. On many tasks, practice leads to
proceduralization and reduction in resource demands, even though complexity remains
constant (Norman & Bobrow, 1975). Anderson’s ACT* theory (1993) examines just how
individuals become experts based on proceduralization. Practice on a particular task
leads to a compilation of declarative knowledge into subsets of procedural knowledge.
Continued practice leads to automatic, efficient task performance (Anderson, 1993).
Holyoak (1991) links Anderson’s ACT* theory among others to the concept of routine
expertise. Once a task is known, or becomes a routine, familiar problems are solved
quickly. Therefore, teams that know the task should have the basic strategies in place for
solving the team task. Over time, the team procedure should be routine. If a team
engages in strategy development late in the task cycle, this would indicate that the team
does not have the basic knowledge in place and should therefore have lower performance

than teams that establish strategies early in the task.

32

METHOD

Participants

Participants in the 67 teams used for this study included 108 males and 93 females
from a management course in a large Midwestern university. The average age of the
subjects was 21.1 years of age (standard deviation was 2.2 years). Participation in this
experiment was voluntary. All participants received three hours of course credit for their
participation, with the potential for winning cash bonuses for top performing teams.
Individuals reported to the experimental setting and were randomly assigned to three

person teams.

'-I
2.:
(f.
7r

 

The interactive team decision-making task came from TIDE2 (Team Interactive
Decision Exercise for Teams Incorporating Distributed Expertise, Hollenbeck et al.,
1995). This is a software program, designed to simulate a command and control
situation. Teams monitor airspace and make decisions regarding unknown aircraft.
Team members are presented with multiple attribute information, and each member
responds to this information with respect to his or her role within the team.

Team members randomly assigned one of three roles in the team. One role is that
of the leader, or Alpha, while the other two are supporting team members, Bravo and

Charlie. These roles correspond with a military ship or station, an aircraft carrier, a

33

coastal air defense position (CAD), or an Advanced Warning Airborne Command System
(AWACS), respectively.

The task was to monitor airspace for unknown aircraft in order to identify the
nature of the aircraft. The nature of the unknown aircraft refers to the degree of threat
posed by this aircraft. The range of threat is on a seven-point scale, corresponding to the
correct team decision. Non-threatening aircraft are “Ignored,” which is the least
aggressive decision. Increasing in threat, the correct decision is to “Review” or
“Monitor.” The mid-level of threat corresponds to a “Warn.” The upper end of the
threatening continuum is to “Ready,” “Lock-On,” or finally, “Defend.” Defending is the
correct decision when the aircraft is highly threatening on all specific target cue attributes
and is a “weapons away” mode, or the most aggressive decision.

These decisions are based on nine attributes (see Appendix for description of
these attributes). The attribute information is gathered by the team members; each team
member sits at a computer station, and the three stations are networked. Each team
member can measure three of the nine pieces of information. For instance, one team
member measures ran ge, corridor status, and the number of aircraft within the target.

The participants had 54 trials to make decisions regarding the unknown aircraft.
These trials varied in time, from 90 to 180 seconds for the experimental trials. Each
member’s monitor had a clock, counting down the available time for each particular trial.
Members also had audible feedback indicating when time was getting low, starting at 30
seconds and again at 15 seconds. This feedback was a moderately loud beeping noise,
letting members know that time was running out and to make their decisions. Each staff

member makes a judgment decision (ignore to defend) which is sent to the leader, Alpha.

34

Alpha then renders the team decision by combining his or her own decision with the staff

decisions.

Task Training

The team completes training prior to the actual simulation. After random
assignment to a position the experimenter trained the participants to use the
communication equipment. Participants were told that all vocal communications would
be recorded. Each station was equipped with a headset and control box. Alpha, Bravo,
and Charlie stations all corresponded to a colored tab on the control box. To speak to
another team member, a team member was trained to push and hold the corresponding
color to speak to the desired team member. While it was possible to speak to each of the
other team members at one time, participants were instructed to speak to only one
member at a time.

Alpha, Bravo, and Charlie were instructed to follow an established protocol when
sending voice messages by first saying to whom they are speaking, and then to identify
themselves. For instance, if Alpha is talking to Bravo, Alpha must say “Bravo this is
Alpha, do you read me?" or whatever the message. This was done for coding purposes,
so the coders did not have to distinguish between similar voices on the tape (team
members were told this). Each participant had to demonstrate to the experimenter that
they understood the protocol for sending voice messages by sending to and receiving
from messages for each other team member.

Following the communication training, participants were introduced to the task.

35

A general overview handbook was given to each participant (see Appendix). The
experimenter went through each page. First was an introductory explanation to the task,
relating the real world critical incidents (for instance, the shooting of the U. S. S. Stark)
leading to this research. Next, the attributes and judgments were explained, followed by a
description of each position and the unique attributes for which each position is
responsible. Finally, the team outcome was described, based on the accuracy of the
judgment process. Participants were given the opportunity to ask questions. Next,
participants received an individualized instruction sheet for their specific position (Alpha.
Bravo, or Charlie). Again. the unique attributes were described along with a range of
non-threatening to threatening values for each attribute. Alpha, the team leader, was also
given a rule for combining the judgments of all three stations to come up with the best
possible team decision. Participants were given eight minutes to read the instructions,
and were told that they could have all written material available for reference throughout
the simulation. so there was no need to memorize any written rules.

The next step in the training process involved going through the actual TIDE2
simulation. There were 54 trials in which the team made ajudgment on an unknown
aircraft. The first six trials were practice trials during which the experimenter was
available for any questions. Participants learned the computer functions used to measure
attributes, query or ask information from teammates, transfer measured information to
teammates, send written text messages, write text messages to self, and send judgments to
the team leader, Alpha. Alpha was trained to wait to receive judgments from Bravo and
Charlie before rendering the team decision. The first training trial lasted 900 seconds,

with the clock on screen counting down so team members knew exactly how much time

36

remained. This time limit was extremely lenient to give the trainer time to work with the
team members. The computer signaled when time runs low by giving a sharp beeping
sound at thirty seconds and again at fifteen seconds. The subsequent trials varied in time,
so that by the end of the practice trials the amount of time equals that given for the actual
simulation. On the experimental trials, time varied from 90 to 180 seconds.

Following the team (Alpha) judgment, each team member viewed a feedback
screen. The feedback screen showed the given team (Alpha) decision, the correct
decision, and the Bravo and Charlie station decisions. A correct decision was worth two
points, called a hit. One level away from the correct decision was a near miss, worth one
point. Two levels away from the correct decision, a miss, yielded a net of zero. Three
levels away from the correct decision was an incident, and one point was deducted from
the total score. Four or more levels away from the correct decision was a disaster, and
two points were deducted from the team score. The running total score was present, as
well as the number of hits, near misses, misses, incidents, and disasters. There was also a
“team average,” showing if the team is averaging closer to the hit level or the disaster
level. This average was applicable to a measure used for a separate research interest.

Following the six practice trials, the team transitions directly into the actual trials
(7-54). Following the simulation, team members were debriefed and thanked for their
participation. Participants were instructed not to discuss the experiment with any other

students.

37

Time

The data to be reported for this research were collected as part of a larger study
designed to look at the effects of modes of communications on team effectiveness. In
that study, 128 teams each performed 54 trials (of which the first six were practice) of the
decision making task over a three hour time period. The teams were randomly assigned
to one of four communication mode conditions (45 of these teams were no voice
conditions, meaning team members could only use the computers to communicate with
each other). The second type of team was Voice only (N = 22), with only voice
communications used to transmit information; the third type was Strategic (N = 23)
where subjects were trained to verbally ask for target information and then transmit the
information via the computer; and the fourth type of team communication was Free
Choice (N = 22), where subjects were trained to use voice or the computer to seek and
provide information. Since mode was not of interest with respect to the issues addressed
here and because there was no reason to expect that the hypotheses of this study would be
affected by communications mode with one exception, we collapsed over mode. The
exception was the computer only condition. This group was not included because the
hypotheses related to this study examine both task and off-task communications, and in
the computer only teams off-task communications were restricted to typed computer
messages. Because time is needed to type computer messages, and each trial is restricted
by time, we felt that there would not be enough off-task communications to warrant

examination of the computer only teams. Therefore, analyses on the critical variables in

38

this study were examined on the three conditions using the voice communication
equipment. For those hypotheses dealing with time, trial sequence was used as a measure
of time. The first trial after the practice sessions was considered the earliest trial point in
the team’s history (Trial 7) and the last trial (Trial 54) the last point in time for the team.

Time, when indexed by trial, is a continuous variable. Since the hypotheses
dealing with time often were limited to effects expected at one extreme or the other (early
or late), the dimension was split into quartiles and analyses done within the early or late
trials. In this case, the first six trials were practice and treated as such. The remaining 48
trials were sectioned into four quartiles. Trials 7-18 comprised quartile 1, trials 19-30
quartile 2, trials 31-42 quartile 3, and trials 43-54 quartile 4. The first quartile was
considered the early trials and the fourth quartile considered the late trials.

Each of the 54 trials on the task presented the teams with the same task.
However, the length of time to complete the task varied. Within the task, the amount of
time available was expected to affect the ease with which a team could accomplish all of
the subtasks necessary to reach a decision. Thus, the second way that time was measured

in the present study was within task where it was measured in seconds.

39

RESULTS

This study is part of a larger study using the TIDE2 decision—making task. The
purpose of this study was to assess how communication affects the quality of team
decisions. Specifically, the frequency of communication in the categories of taskwork,
teamwork, off-task social, and off-task other were hypothesized to predict effectiveness in
a team decision making task. Each category was considered mutually exclusive, and the
combination of these four categories was expected to capture all communications
occurring between team members. For the purposes of this study, a means of capturing
the frequencies of these types of communications occurring between team members had
to be established.

The frequencies of these communications were based, in large part, on coding
voice transcriptions of each team participating in the decision-making task. Remember
that the entire task consisted of 54 separate decision making trials, and the first six trials
were practice trials. On these practice trials, the first trial was used for training subjects
on the computer aspect of the task, such as how to use the mouse to measure the different
task attributes and to render the decision using pull down menus. Prior to the second
trial, subjects placed headphones on, and their communications were recorded for the
remainder of the trials (2-54). These recordings were then transcribed word for word
using five undergraduate transcribers. Thus, there were hard copies of the recorded
communications available.

An a priori coding system was used for the overall TIDE2 decision-making task

40

(see Table 3 for a description of the coding content). This system was developed for
purposes other than the present study, so to reach the main communication categories
posed in this study there were two main steps. First, the transcribed voice data needed to
be coded and placed into the initial categories shown in Table 3. Second, these categories
needed to be further sorted into the four main categories of interest in this study. To
evaluate the quality of the frequency measures, we will first address the extent to which
the initial coders were able to reliably son the transcribed verbal messages to the initial
coding scheme. Following this, I will then address combining these initial data into the

communication categories relevant to this study.

Coding

Five coders received three initial training sessions. The training began with the
categories into which an utterance of communication would be placed. Coders were also
trained on what constituted an utterance to be considered a unit of communication; a
single sentence could contain more than one utterance of communication. The coders
used electronic scoring sheets to capture the eight main categories in the communication
coding system. The scoring sheets were designed to capture the team, the trial number,
and the team member station along with the utterances of communication in each
category. The data would result in instances representing frequencies for each utterance
of communication within every trial for each member of a team participating in the

decision making task.

41

Table 3
Description of Voice Coding Schema

 

Communication Subcategories:

 

1. Encouragements/ admonitions/ apologies/ other_:_ Anything task related not concerned
with a target or specific coordination issues.

Examples: “We’re not doing very well.” “I’m sorry I screwed up.” “Keep
concentrating.”“My computer is locking up.”

2. Coordination: Anything that somehow deals with coordination or strategy involving
the decision-making task.

Examples: “What can you measure?” “Get your judgments in by the 305. mark.”
3. Target decision information: Communication about the specific trial decision.

Examples: “1 have 2 threats.” “How bad are you looking?” “I want to warn or lock-on.”
“Let’s defend.”

4. Repeat: Anything that requests someone to repeat a statement.
Examples: “What did you say?” “Repeat that last part.”

5. Target Exact attribute information: Communication about exact attribute values
within a specific trial.

Examples: “What is your speed?” “I need radar.” “Speed is 500 mph.” “Radar is 5.”

6. Task Inexact Attribute information: Communication about subjective target values
within a specific trial.

Examples: “15 speed fast?” “What’s radar like?” “ Speed is fast.” “Radar looks bad.”

7. Acknowledgment: Simply adhering to the rules of polite conversation, such as
responding or acknowledging someone.

Examples: “Copy.” “Thank you.”

8. Social: Any communications that are social in nature and have nothing to do with the
task.

99 ‘5

Examples: “I’m hungry. Don’t you think the guy who trained us was really cute?”

 

42

Prior to allowing coders to begin coding the transcriptions of actual experimental
sessions, a subset of transcriptions was used to assess the quality of their ratings. Once
the coders learned the system, each was given a partial transcription from a decision
making team using a subset of 20 trials. These were actual team data transcriptions that
were first coded by the five main experimenters who were able to reach high levels of
agreement when they sorted the communication units into the established categories. The
five coders then also coded the same 20 trials, also reaching agreement with the
experimenters. Then, interrater reliabilities were calculated for each of the eight
categories based on these reliabilities. This index of reliability is based on the alpha
statistic and the assumption that raters may be treated as test items. Based on this
assumption, the following interrater reliabilities were calculated:

Coordination = .87 Decision = .92 Encouragements = .92 Repeats = .86
Acknowledgments = .84 Target Exact = 100 Target Inexact = .98 Social = .85
Based on these reliabilities, it was decided by the experimenters that the raters were
sufficiently trained for the coding task.

Since the coding task was very time consuming, taking about 3 hours to code one
team, each coder was assigned approximately one fifth of the teams. That is, teams were
coded only by one coder. However, to check against possible decay of coding
performance over time, at the end of two weeks each coder was given a subset of trials
(20) from a team that he or she had previously coded. With rare exceptions, the
communication data is far from memorable since subjects are passing very routine
information. There is no reason to expect that a coder would remember his or her ratings

on a trial for any particular team. Therefore, these later identical trials could be used to

43

detect the consistency of raters over time. The test-retest reliabilities were above .85 for
each coder. It was concluded that the coders could reliably sort the communication units
into the categories and that once trained they were able to maintain their performance

over time. The conclusion, that the coders were reliable and consistent, was a necessary

but not sufficient condition for the present study.

Communication Dimensions

Since this study poses relationships based on four main types of communication,
taskwork, teamwork, off-task social and off-task other, the data had to be further
examined for fit into the categories used in this study. More specifically, there was a
need to further identify the quality of the data and whether the data are represented by the
four mutually exclusive domains examined in this study. A first step in this process was
to determine if these eight coded voice communication categories would collapse into the
four main categories of interest. Because some specific task information could be shared
between team members through simple pull-down menus as a part of the task, the
computer task information had to be added to the voice communication data set in order
to provide the total set of communications—communication by voice and communication
via the computer. All of the task information shared via the computer was exact target
attribute information, and was coded as such within the voice data set. In other words, all
computer communications were coded into the category Target Exact Attribute
information (see Table 3). Therefore, the resulting data set to be factor analyzed

consisted of the combination of the voice and computer data into the eight categories

44

described in Table 3.

Results of the factor analysis are shown in Table 4, which is the rotated
component matrix. Principle component analysis with varimax rotation yielded a three
factor solution. The eigenvalues for the three factors were 2.6, 1.4, and 1.2. To examine
the three factors, those variables with loadings above .5 (noted in bold in Table 4) were
used to interpret each factor. Factor one consists of target decision information, repeats,
and acknowledgments. The variables loading on the second factor are coordination and
encouragements, while the third factor consists of target exact and inexact information.

Table 4
Rotated Component Matrix

 

Rotated Component Matrix

Communication

Variable Component
1 2 3
Target Decision .823 .075 -. 176

Target Exact Information -.069 -.387 .753
Target Inexact Information -.023 .296 .543

Coordination .465 .510 .432
Encouragements .363 .784 -.087
Acknowledgments .519 .293 .126
Repeats .839 -.088 .037
Social -.1 13 .073 .097

 

The main conclusion using the factor analytic results is that these communication
component variables do not cleanly fall into the four factor structure presented earlier.
One potential interpretation of these data is that the communication category of taskwork,

or task related information, appears to be multidimensional in nature. Also, the three

45

factors each consist mainly of task related information, while there does not appear to be
a factor representing off-task communication. Since these data are based on the
frequencies of communication data, analyzing the data using expert raters was explored
as a second way to ask whether it was reasonable to consider four categories for the
communications occurring in this study (Cranny & Doherty, 1988).

The eight communication variables used in the factor analysis were analyzed by
expert raters. These ten new raters, all graduate students in psychology, were given
descriptions of taskwork, teamwork, off-task other, and off-task social along with
descriptions of the eight communication categories. Table 5 shows the number of times

each rater listed one of the eight categories into the four primary categories.

Table 5
Communication Category Results: Percentage of Rater Agreement

 

 

 

 

Taskwork Teamwork Off-task Social Off-task Other
Decision (90%) Encouragements (90%) Social ( 100%) Repeat (90%)
Exact Attribute (100%) Coordination ( 100%) Aclnowledgment
Inexact Attribute (100%) (50%)
N=10 raters.

These raters were given a listing of the communication categories (called from
here on subcategories) in the coding schema (See Table 3). Raters then sorted each
subcategory into one of the four main categories, shown in Table 5. There was high
agreement for most categories with the exception of acknowledgments. For this category.
there was 50% agreement that acknowledgments belonged in the category of off-task

other communications. The remaining raters split this category into Social and

46

Teamwork. Based on this information, the researcher went along with the majority and
the acknowledgment subcategory was kept in the Other category.

The decision was made to use four categories for the shared communications in
this study. Several factors influenced this decision. First, although this did not match the
factor analysis, the expert raters were able to reliably sort the eight dimensions into four.
Therefore. the four categories provided a logical grouping of the behaviors. In addition,
past research has suggested that in some cases, expert ratings may be a more reasonable
interpretation of the data than a factor analysis (Cranny & Doherty, 1988). Finally, from
a practical standpoint, the main interest here was not so much the nature of the four
categories, but in whether or not these categories represented reasonable, mutually
exclusive categories. At this point, the data were analyzed using the four categories.
Further consideration of the factor analysis results will take place in the Discussion
section of this thesis.

The first category, Taskwork, combined communication data regarding target
decision information and target attribute information. The second category, teamwork,
resulted from the combination of coordination and the encouragement subcategories of
communication. As a subset of the second category, coordination communications were
examined alone as an index of strategy. The third category, off-task social, was simply
the social communication subcategory. This factor had 100% agreement with the expert
raters. The final category, off-task other, combined the remaining subcategories, repeats
and acknowledgments. This yielded a data set consisting of frequencies of each type of

communication in the four categories.

47

Table 6 gives the descriptive statistics and intercorrelation matrix for the four
main communication content areas. Strategic communications are not included in this
table because strategic communications are comprised solely of coordination
communications, which are a part of the teamwork dimension. The means and standard

deviations are based on the number of utterances per trial for all teams.

Table 6
Descriptive Statistics and Intercorrelations Among Communication Scales

 

M S_D Taskwork Teamwork Social Off-Task Other

Taskwork 9.43 7.2
Teamwork 3.31 3.19 .26**

Social .13 .83 -.02 .l9**
Off-Task 1.32 2.06 .28** .70** .19**
Other

 

N=3216 (67 teams x 48 trials)
” indicates significant correlation (p<.001).

Following consideration of the nature and quality of the communication variables
is the consideration of frequency with which certain communications occur as a result of
the type of team. The method section of this study described that the overall study from
which this study comes included four types of teams. These teams again were Computer
only (not used in the present study; all communication took place via computer text
messages); Voice only, with only voice communications used to transmit information;
Strategic, where subjects were trained to verbally ask for target information and then
transmit the information via the computer; and Free Choice, where subjects were trained

to use voice or the computer to seek and provide information. Table 7 presents the

48

frequency of communication by type of team (excluding computer only teams). The
means and standard deviations of each type of team on the dimensions of taskwork,
teamwork. strategy, social, and other (off-task) communications shows that the types of
communication that occur among team members are quite similar. Therefore, we
collapsed over the three types of teams for all further analyses. The results of this table
are based on the average communication frequencies per trial. The table also shows that
there are more communications related to the actual task, while the other communications

are (I more rare occurrence.

Table 7
Frequency of Communication Content by Type of Team

 

Taskwork Teamwork Strategic Social Other
M SD M SD M SD M SD M SD

 

VoiceOnlyI 15.09 3.94 3.38 1.38 2.47 .78 .07 .12 .85 .98
FreeChoice' 13.09 4.05 3.22 1.66 2.49 1.39 .11 .14 .48 .49

Strategic2 13.11 3.03 3.64 2.27 2.59 1.26 .22 .17 .54 .62

 

1. For Voice Only and Free Choice Teams, N=22
2. For Strategic Teams, N=23

Analyses Related to livpotheses

The primary data for analyses related to the hypotheses were based on'the
frequencies of the types of communications occurring within each trial during the entire
decision making task. which occur during the 48 experimental trials. As stated earlier,

because some hypotheses examine specific target information that can be captured either

49

with voice or computer communications, all forms of task communication must be
included in the analyses. Therefore, for data analyses using “Taskwork”
communications, the addition of those task specific pieces of information sent via the
computer were combined with the verbal taskwork communications.

Hypothesis 1. The first hypothesis predicted that the frequency of on task
communications would be positively correlated with the accuracy of team decisions.
Correlations between the number of task communications on a trial and trial performance
were investigated for several breakdowns of the data across time. Specifically, average
number of task communications per trial were calculated for each team over trials 7
through 54. Average team decision accuracy was also calculated for these trials. This
sequence of 48 trials was broken down into four quartiles of performance, with an
average of task communication and performance calculated for each quartile. Because the
first six trials were considered task training trials, they were excluded from the analyses.
Results of correlation analyses found that the relationship between task communication
and overall performance was not significant (r: .14, ns).

Hypothesis 2. The second hypothesis stated that the frequency of off-task
communications will be negatively correlated with the accuracy of the team decisions.
The relationship between overall off-task communications and overall performance is
basically nonexistent, given a correlation of .01 (ns). In sum, the results do not support
Hypothesis 2. Furthermore, the descriptive data indicate that there was little off-task
communication occurring, given a mean of .71 utterances off-task communication per
trial (standard deviation of .85).

Hypothesis 3. This hypothesis posed that the relationship between off task

50

communication and decision making accuracy would be moderated by the time available
to complete the task. A moderated regression was performed with the frequency of the
off task communications, the time pressure category, and the interaction between the two
variables regressed against decision making accuracy. Table 8 provides the results of the
regression analysis, indicating no relationship between the amount of time available and
the amount of off task communications. Thus, Hypothesis 3 was not supported.

Table 8
Hypothesis 3 Regression Results

 

Model Summary

R R2 AR2

 

Model 1

Off Task communications (OT) .03 .00 .00
Time Pressure Category (TP)

(1=High Time Pressure

2=Low Time Pressure)

Model 2

OT

TP

OT x TP .04 .00 .00

 

Hypotheses 4 and 5. The final two hypotheses relate to the engagement in
strategy communications. Because this is a complex, decision making task, team
members must share multiple pieces of information (task attribute values) in order to
render the best possible decision. Remember from the Method section that hypotheses
dealing with time treat the experimental trials as quartiles. In this consideration Trials 7-
18 comprised quartile 1, trials 19-30 quartile 2, trials 31-42 quartile 3, and trials 43-54
quartile 4. The first quartile was considered the early trials and the fourth quartile

considered the late trials. It was hypothesized that coordinating the sharing of information

51

early in the task would be beneficial to the accuracy of team decisions. In other words,
the earlier team members establish who sends what information to whom, the better team
decisions should be. Specifically, Hypothesis 4 stated that there would be a positive
relationship between the engagement of strategy communication in early trials, or the first
quartile. and decision making accuracy in later trials, quartile 4. Furthermore, teams that
fail to establish the correct strategy early in the team task will probably not make accurate
decisions. Those teams that still engage in strategy communications on later trials
probably do not have a grasp of how to share information efficiently. Thus, Hypothesis 5
stated that there would be a negative relationship between the engagement of strategy
communications on later trials (quartile 4) and overall decision accuracy.

Table 9 shows the descriptive statistics for strategy communication and
performance and the intercorrelation between strategy communications within each
quartile and performance within each quartile. The relationship between strategic
communication within quartile and overall performance is also given in the table. There
was basically no correlation between the engagement of strategy communications in the
first quartile and performance in the last quartile (r=.03, ns). Thus, Hypothesis 4 was not
supported. There was also no relationship between engagement of strategy
communications late in the task and overall performance, (r=-.01, ns). Hypothesis 5 was
not supported. Results indicate a fairly stable pattern of communication across quartiles,
but the frequency of such communications is low in comparison to task specific

communications.

52

Table 9
Results for Hypotheses 4 and 5: Engagement of Strategy

M SD Intercorrelation

Strategy Performance
Q1 Q2 Q3 Q4 All

 

 

Quartile l 2.65 1.41 -.15 .08 .08 .03 -.Ol
Quartile 2 2.21 1.13 .02 .00 .03 .00
Quartile 3 2.10 1.11 .01 .13 .03
Quartile 4 2.16 1.56 .13 -.01
Performance

Quartile l 5.86 .31

Quartile 2 5.96 .25

Quartile 3 5.94 .27

Quartile 4 6.22 .21

N=67.

53

DISCUSSION

The main purpose of this study was to assess how communication affects the
quality of team decisions. In pursuing this question, the frequency of different types of
communications was assessed and sorted into different categories, both task and non-task
related. Disappointingly, no hypotheses were supported. The remainder of this paper

examines the question of why these hypotheses were not supported.

Examination of Hypotheses

Hypothesis 1 stated that the frequency of on task communications would be
positively related to the accuracy of team decisions, while Hypothesis 2 stated the
reflection in that the frequency of off-task communications would be negatively related to
the accuracy of team decisions. These hypotheses were based on the rationale that
because team members must share information regarding the task, those teams that had
more on task communications would have more accurate decisions. Related to this
rationale is the fact that on tasks where time is limited, communications are mutually
exclusive events; if a team is engaged in off-task communications, the team is not at the
same time discussing on task matters. In other words, using time to discuss off-task
things takes away from the time to engage in on task communication. Therefore, the
frequency of off-task communication would be negatively related to the team decision—

making accuracy on any particular task.

54

Overall, Hypothesis 1 was not supported. The relationship between on task
communication and team decisions was not significant when examining the overall
frequency of on-task communications with team performance as the hypothesis stated.
Similarly, Hypothesis 2 was also not supported. An examination of Hypothesis 2 and its
lack of support was most likely due to an overall lack of off-task communication among
team members. Again, a re-examination of Table 5 shows that the mean frequency
amount of social communications was .13 units per trial, and the mean frequency amount
of off-task other communications was 1.32 units per trial, compared with 9.43 and 3.31
mean values for taskwork and teamwork communications, respectively.

Hypothesis 3 stated that the relationship between off-task communication and
decision-making accuracy would be moderated by the time available to complete the task.
In treating the lesser amount of time (90 seconds) as high time demand and the greater
amount of time (150 seconds) available within a trial as low time demand, it was
expected that in the high time demand. teams should focus communication on the task.
Off-task communication would take away from the supposedly more critical job of
communicating about task specific information. However, performing a moderated
regression to test the effects of low and high time demands resulted in no interaction
effects. Again, the lack of support for Hypothesis 3 seems to stem from the overall lack
of off-task communication.

Hypothesis 4 stated that there would be a positive relationship between the
engagement of strategy communications on the early trials and overall team performance.
The rationale for Hypothesis 4 was based on goal setting research in complex tasks (e. g.

Wood et al., 1988; Earley et al., 1993). These studies have found that failure to develop

55

strategies for dealing with complex tasks was detrimental to performance on these tasks
in the long run. However, in the current study there was no relationship between the
frequency of strategic communications in early trials and overall team performance. One
explanation for lack of support for Hypothesis 4 is that the analyses excluded the training
trials. It is possible that teams engaged in some strategic planning (such as who sends
what information to whom) in these early training trials. Therefore, additional analyses
incorporating these trials were performed. However, there was still no support for the
hypothesized relationship between the frequency of strategic communications and
performance using these additional trials in the analyses.

Hypothesis 5 was stated as a reflection of Hypothesis 4, in that there would be a
negative relationship between strategy communications late in the task (the last quartile

of experimental trials) and overall performance. Again, no such relationship was found.

Examination of Factor Structure

One finding that is also important to note in this discussion is the result of the
factor analysis (see Table 4). The results of the factor analysis came up with three
factors, two relating to taskwork (factors one and three) and one relating to teamwork
(factor 2). Thus, the factor analysis yielded categories that did not match clearly with the
four communication categories pose in the present framework.

Future research should focus on a better distinction in the realm of categories of
communication information. Two categories posed in this framework were based on

existing research on taskwork and teamwork (e.g. McIntyre & Salas, 1995), with other

56

information separated into off-task domains. One caveat to this study is that it relied
upon the frequency distribution of different types of communication to sort into the four
main categories. In the present study, the frequency of off-task types of communication
were extremely low (see Table 6). Therefore, to rely upon the frequency distribution to
determine the four categories using factor analysis would render caution in the
interpretation of a factor structure in this setting. However, if the data were such that the
frequencies of more off-task communications occurred, the factor analysis may be more

applicable. Future research should address this issue.

Conclusion

While it was disappointing that the hypothesized relationships in this study were
not supported, there are several possible explanations for these results. The first and most
obvious conclusion is that the amount of off-task (both social and other) verbal
communication was extremely low. The introduction posed that time is indeed a real
constraint (p.2) in that if someone discusses task related information, it cannot be that at
the same time this person discusses non-task related information. It is likely that time
posed an even more serious constraint, thus not producing an environment where much
communication could really occur. Another explanation could be that subjects in this
context do not really feel comfortable talking to each other, given that they may not know
each other well and this is a laboratory setting. The subjects may also have been inhibited
by the headphones or the recording of their voice communications.

A reexamination of Table 6 gives the descriptive statistics for the four main

57

communication categories. The means and standard deviations are based on the average
communication utterances per trial for each of these types of communications for the
teams used in this study. A breakdown of these utterances by type of team is shown in
Table 7. While the off-task social and off-task other communications appear low in
number and may have been an early indication that hypotheses relating to off-task
communications would lack in support, this type of research is not published (now maybe
we know why) and therefore there was no a priori basis to believe that hypotheses would
not be supported. Taskwork and teamwork communications occur more frequently, and
one might expect that these larger numbers would be sufficient to glean effects but alas
they were not. Again, there was no published evidence to lead this research effort to
expect certain numbers to produce hypothesized effects.

Related to the explanation above is one related to task demands. Upon closer
examination of potential avenues for communication, there is a very small need for teams
to communicate. Even though the overall study was structured to capture differences
between mode of communication, using the four different types of teams, the voice
communication teams (with the exception of voice-only teams) were trained on how to
perform everything via the computer. The voice-only teams, where members were not
trained on how to send task-specific infomiation through the computer, had to perform at
minimum 9 communication exchanges per trial. This is determined when one notes that
each of the three members needs to get specific information. IfI was Alpha, for example.
and I didn’t know which team member had my information, I would ask both Charlie and
Bravo. Presumably the others would follow a similar path. Then, only one member

needs to respond to the request, because if Charlie had my information she or he would

58

tell me; there is no need for Bravo to respond. As the game continues, the need for the
initial query of who has my information should be learned, because the information team
members have does not change. When this conclusion is reached, only three
communication exchanges are needed; simply having each member send the task specific
information to the person who needs it. This did not happen often, and of course there
were instances of repeating information, but at minimum, very little voice communication
was necessary.

Another possibility for lack of support is the exclusion of the training trials in the
analyses. Because teams have the opportunity to discuss the task in the training trials and
this task involves the rote exchange of similar information, teams might have established
task-related strategies early in the task, meaning within the training trials. However, as
reported above additional analyses incorporating data from the training trials yielded no
support for the hypothesized relationships. Therefore, the exclusion of these initial trials
was not a sufficient explanation for the observed effects.

Another potential avenue for post-hoe analyses was to examine those teams that
communicated more frequently. Examination of the average frequencies of team
communications per trial indicated a mean level of 13.5 (standard deviation of 7)
utterances per trial. Therefore, I chose an arbitrary cut-off of 20 utterances per trial to
select teams for additional analyses. This resulted in post-hoe analyses for 17 teams
(down from the original 67). Unfortunately, these analyses failed to support the
hypotheses.

From these disappointing results then comes the question: Is it reasonable to

conclude that communication does not lead to effective team decisions? My conclusion

59

is no, it is not reasonable to assume that just because these relationships did not hold in
this particular study that they do not exist overall. This study tried to examine
communication as a process leading to effective decisions in one setting using one task.
The models following the input-process-outcome framework presented in the
introduction focused on specific inputs leading to outputs for singular tasks. The studies
used to define teamwork in this study each labeled communication as a subset of
teamwork, while this study used communication as a means of conducting both taskwork
and teamwork, again trying to capture communication as a process. Because this study
relied upon communication as a process, and the means of performing both teamwork and
taskwork behaviors, and subjects were limited in other forms of interaction due to the
computerized nature of the task, there was simply not enough teamwork communication
displayed to adequately test its relationship with task performance. However, I do think
that there is reason to pursue communication and the different content areas of
communication in future research in terms of how it affects the quality of a team’s
effectiveness.

One interesting aspect of this study was the attempt to capture the off-task
communications and the relationship between these communications and the quality of
team decisions. This was an attempt to empirically examine Hackman’s (1987)
theoretical notion of synergy or the social glue that holds teams together. Obviously.
there was not enough off-task communication occurring in this study to capture the effect,
but that does not mean that it doesn’t exist. Possibly a better place to examine the
relationship between the off-task communication and effectiveness would be an

environment where teams are much more long-standing. In a work environment, there is

60

much more “down time” than was available in the experiment used in this study. For
example, there is the opportunity in an office setting for exchanges to occur at the coffee
machine or at lunch. Over a period of months or years, these interactions add up and may
be more of the synergy notion that Hackman theorized.

Another possibility with existing teams is the incorporation of even more complex
tasks than that of the decision making task used in this study. In research and
development teams, for example, there is the potential for more of the teamwork related
behaviors posed by some authors (e.g. Krahl & Dickinson, 1997). When a team is
involved in building a product, there is potential for the product to fail for one reason or
another. It is those interactions involved in the troubleshooting or the “why” aspect of a
project that also may lead to more of the teamwork behaviors. The potential for team
members to work together in the future may also affect the interactions that take place
between team members.

All of these “potentials” may affect interactions and the types of relationships that
occur in team settings that differ from the teams used in the present study. It is for these
reasons that it is not reasonable to conclude at this point that communication between
team members has no effect on the quality of a team product. It is quite possible that

communication leads to effective teams.

61

APPENDIX

TIDE2 Simulation General Overview

62

General Overview

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 U. S. 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 servicemen 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
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 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 our of
the region, and friendly military aircraft from nations not involved in the conflict also
patrol the area.

Overview of Roles: There are three roles in this simulation, referred to as ALPHA,
BRAVO, and CHARLIE. The leader, ALPHA, is the senior weapons director. 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 of 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 possible 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.

63

SEVEN POSSIBLE DECISIONS

1)

2)

3)

4)

5)

6)

7)

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.

Review: this means attention can be shifted away from this target momentarily.
After a short period of time this target should be returned to in order to update its
status.

Monitor: this means that the target should be continuously tracked.

Warn: this means that a message is sent to the target ordering it to turn away.
Warning targets that should be ignored detracts from the importance of legitimate
warnings. Warning targets that intend to attack is also bad, since the warning makes
it easier for an attacker to locate the target base.

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.

Lock-On: This synchronized 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.

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.

64

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

 

RANGE

Distance from base operations. In
general, targets that are closer are
more threatening.

0-600

 

ALTITUDE

Number of feet target is
above the ground. In general,
targets that are low in altitude are
more threatening.

 

RADAR CROSS SECTION

1 00-99,000

 

Estimated size of the target. In
general, smaller targets are more
threatening.

0-12

 

CORRIDOR STATUS

Miles from center of civilian
corridor. In general, targets far
outside the corridor are more
threatening.

0-25

 

ELECTRONIC SECURITY
MEASURE (ESM)

Indicates threat of radar signals. In
general. targets with high ESM
values are more threatening.

0-999

 

# OF AIRCRAFT WITHIN
THE TARGET

Each target may consist of more
than one aircraft, flying close
together. In general, a higher
number of aircraft is more
threatening.

1-20

 

HEADING CROSSING
ANGLE (HCA)

Indicates direction of target. In
general, the higher the HCA, the
more directly the target is headed
toward the base. which is more
threatening.

0-180

 

RATE OF ALTITUDE
C HAN GE
(RATEAALTITUDE)

Number of feet/minute ascending
or descending. In general, the
higher the rate of altitude
change the more threatening.

0- 10,000

 

 

SPEED

 

Miles per hour. In general, the
faster the target the more
threatening.

 

0-800

 

65

 

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. The team leader,
Alpha, is also responsible for combining recommendations of the team members into
a correct overall team decision.

RULE ALPHA RANGE, CORRIDOR STATUS, # AIRCRAFT WITHIN TARGET
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 of low. Thus, a target can be very
threatening for two out of three of these values, and still be considered safe for this
aspect.

RULE BRAVO: ALTITUDE, HEADING CROSSING ANGLE, SPEED
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 of low. Thus, a target can be very
threatening for two out of three of these values, and still be considered safe for this
aspect.

RULE CHARLIE: RADAR CROSS SECTION, ESM, RATEAALTITUDE
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 of low. Thus, a target can be very
threatening for two out of three of these values, and still be considered safe for this
aspect.

HOW ALPHA OCMBINES RECOMMENDATIONS TO DETERMINE TEAM
,IUDGMENTS.

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
recommendations were correctly made, the overall judgment is probably LOCK-ON.

66

OUTCOMES OF DECISIONS

Your decisions regarding each target are to be made based upon the rules discussed on the
previous page. Once your team leader 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:

 

 

 

 

 

 

 

 

 

 

 

OUTCOME DEFINITION EXAMPLE POINTS

The decision You said defend,

(1) HIT was exactly correct decision was +2
correct. defend.
The decision You said defend,

(2) NEAR MISS was off by one correct answer was +1
level. lock-on.
The decision You said defend,

(3) MISS was off by two correct decision was 0
levels. ready.
The decision You said defend,

(4) INCIDENT was off by three correct decision was -1
levels. warn.
The decision You said defend,

(5) DISASTER was off by more correct decision was
than three either monitor, review, -2
levels. or lock on.

 

67

 

INSTRUCTIONS FOR ALPHA (LEADER)

As the leader of the group, ALPHA has the responsibilities of both specific expertise and
the coordination of information and recommendations from 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
judgments regarding a target’s threat level. While there is some overlap of information
within the team, each team member needs to get certain pieces of information from other
team members. The role of the team members (including leader 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 ALPHA leader, 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. SPECIALIZED EXPERTISE: you are uniquely responsible for the
following information:

 

Between safe Between

and moderate and
TARGET CUE SAFE moderate MODERATE threatening THREATENING

 

 

 

 

RANGE 600-400 399-301 300-200 199-101 100-0
CORRIDOR

STATUS 0-3 4-9 10- 15 16-19 20-25
# OF AIRCRAFT 1-2 3-4 5-7 8-1 1 12-20

WITHIN TARGET

 

 

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

IF ANY 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 (1-2) the Alpha assessment for the target is safe.

. Another example is if the target may have threatening corridor status and number
of targets, but if range is safe (600-400) the Alpha assessment for the target is
safe.

68

 

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. While there is some
overlap of information within the team, each team member needs to get certain pieces of
information from other team members. The role of the team members (including leader
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 responsible for the following information:
Between safe Between
TARGET CUE and moderate and

SAFE moderate MODERATE threatening THREATENING

 

 

 

 

ALTITUDE 99,000- 44,999- 35 .000- 9,999- .000- 100
45,000 35,001 10,000 1,001

HEADING

CROSSING ANGLE 0-60 61-89 90-120 121-149 150-180

(HCA)

SPEED 0-100 101-199 200-300 301-399 400-800

 

 

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

IF ANY ONE OF THESE THREE CUES IS SAFE,

THE OVERALL BRAVO ASSESSMENT IS SAFE.

. For example, a target may have threatening hca and altitude, but if the speed is
safe (0-100) the Bravo assessment for the target is safe.

. Another example is if the target may have threatening hca and Speed, but if
altitude is safe (99,000-45,000) the Bravo assessment for the target is safe.

69

 

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. While there is some
overlap of information within the team, each team member needs to get certain pieces of
information from other team members. The role of the team members (including leader
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 uniquely responsible for the following information:

 

 

 

 

 

Between safe Between
and moderate and

TARGET CUE SAFE moderate MODERATE threatening

THREATENING
RADAR CROSS 12-10 9-8 7-5 4-3 2-0
SECTION(RCS)
ELECTRONIC
SECURITY 0-175 176-224 225-400 401-599 600-999
MEASURE(ESM)
RATE CHANGE 0-2000 2001-2999 3000-5000 5001-5999 6000-10.000
ALTITUDE
(RATEAALT)

 

 

 

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

IF ANY ONE OF THESE THREE CUES IS SAFE,

THE OVERALL CHARLIE ASSESSMENT IS SAFE.

. For example, a target may have threatening RCS and ESM, but if RATEAALT
safe (0-2000), the Charlie assessment of the target is safe.

. Another example is if the target may have ESM and RATEAALT, but if the RCS
is safe (12-10), the Charlie assessment of the target is safe.

70

REFERENCES

Anderson, J .R. (1993). Problem-solving and learning. American Psychologist,
48, 35-44.

\(Argote, L., & McGrath, J.E. (1993). Group process in organizations: Continuity
and change. In C. Cooper & I.T. Robertson (Eds), lntemational review of industrial &
organizational psychology, Vol. 8, 333-389). London: John Wiley.

* Bettenhausen, KL (1991). Five years of groups research: What we have learned
and what needs to be addressed. Journal of Management, 17, 345-381.

Bowers, C., Salas, E., Prince, C., & Brannick, M.T. (1992). Games teams play: A
method for investigating team coordination and performance. Behavior Research
Methods, Instruments, and Computers, 24, 503-506.

‘-’- Campion, M.A., Medsker, G.J. & Higgs, C. (1993). Relations between work
group characteristics and efficiency: Implications for designing effective work groups.
Personnel Psychology, 46, 825-850.

~+ Cannon-Bowers, J.A., Tannenbaum, S.I., Salas, E., & Volpe, CE. (1995).
Defining competencies and establishing team training requirements. In R.A. Guzzo and
E. Salas (Eds), Team effectiveness and decision making in organizations (pp. 333-380).
San Francisco: Jossey-Bass.

Cranny. C.J., & Doherty, ME. (1988). Importance ratings in job analysis: Note
on the misinterpretation of factor analyses. Journal of Applied Psychology, 73 (2), 320-
322.

2*. Davis. J. H. (1992a). Some compelling intuitions about group consensus
decisions, theoretical and empirical research, and interpersonal aggregation phenomena:
Selected examples, 1950-1990. Organizational Behavior and Human Decision
Processes, 52(1), 3-38.

Davis, J. H., Ed. (1992b). Organizational Behavior and Human Decision
Processes, 52.

“ti Dyer, IL 1984). Team research and team training: State of the art review. In

F.A. Muckler (Ed). Human factors review (pp. 285-323). Santa Monica, CA: Human
Factors Society.

71

Earley, P.C., Connolly, T., & Ekegren, G. (1989). Goals, strategy development,
and task performance: Some limits on the efficacy of goal setting. Journal of Applied
Psychology, 74 (1), 24-33.

Gersick, CJ. (1988). Time and transition in work teams: Toward a new model of
group development. Academy of Management Journal, 31, 9-41.

Gladstein, D.L. (1984). Groups in context: A model of task group effectiveness.
Administrative Science Quarterly, 29, 499-517.

4 Goodman, P.S., Ravlin, E., & Schminke, M. (1987). Understanding groups in
organizations. Research in Organizational Behavior, 9, 121-173.

_ i 84 Guzzo, R.A. & Shea, GP. (1992). Group performance and intergroup relations
in organizations. In M. D. Dunnette & L.M. Hough (Eds), Handbook of Industrial and
Organizational Psychology, Vol. 3, 2nd ed., Palo Alto, CA: Consulting Psychologists
Press.

~ Hackman, J.R. (Ed). (1990). Leading groups in organizations. San Francisco:
Jossey-Bass. '

Hackman, J .R. (1986). The psychology of self-management in organizations. In
M.S. Pallak & R.O. Perloff (Eds). Psychology and work: Productivity. change, and
employment. Washington, DC: American Psychological Press.

at, Hackman, JR. (1987). The design of work teams. In J.W. Lorsch (Ed.),
Handbook of organizational behavior. Englewood Cliffs, N.J.: Prentice—Hall.

X Hackman, J.R., Brousseau, K.R., & Weiss, J.A. (1975). The interaction of task
design and group performance strategies in determining group effectiveness.
Organizational Behavior and Human Performance, 16, 350-365.

9* Hackman, J .R. & Morris, CG. (1975). Group tasks, group interaction process,
and group performance effectiveness: A review and proposed integration. In L.
Berkowitz (Ed), Advances in experimental social psychology. New York: Academic
Press.

Holyoak, K]. (1991). Symbolic connectionism: Toward third-generation
theories of expertise. In K.A. Ericsson and J.Smith (Eds), Toward a general theory of
expertise pp. 301-336). Cambridge: Cambridge University Press.

4r Kernan. M.C., Bruning, N.S., & Miller-Guhde, L.M. (1994). Individual and
group performance: Effects of task complexity and information. Human Performance, 7.
273-289.

72

*I'Kozlowski, S.W.J., Gully, S.M., McHugh, P.P., Salas, E., & Cannon-Bowers, J.A.
(1996). A dynamic theory of leadership and team effectiveness: Developmental and task

contingent leader roles. In Research in Personnel and Human Resources Management,
Vol. 14, 253-305. JAI Press, Inc.

3,. Krahl, K., & Dickinson, TL. (1997). Teamwork processes, task
interdependence, and team performance. Paper presented at the 13th annual conference
for the Society of Industrial/Organizational Psychology in St. Louis, MO. April, 1997.

Locke, E.A., & Latham, G.P. (Eds) (1990). A theog of goal setting and task
performance. Englewood Cliffs, NJ: Prentice Hall.

McIntyre, R. A. & Salas, E. (1995). Measuring and managing for effective
teams. In R.A. Guzzo and E. Salas (Eds). Team effectiveness and decision making in
organizations. San Francisco: CA: Jossey-Bass Publishers.

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 (2), 185-193.

Morgan, B. B., Jr., Glickman, A.S., Woodard, E.A., Blaiwes, A., & Salas, E.
(1986). Measurement of team behaviors in a Navy environment (NTSC Report, No. 86-
014). Orlando, FL: Naval Training Systems Center.

Pritchard, RD. (1992). Organizational productivity. In M. D. Dunnette & L.M.
Hough (Eds), Handbook of Industrial and Organizational Psychology, Vol. 3, 2nd ed.,
Palo Alto, CA: Consulting Psychologists Press.

4 Rentsch, J.R., & Hall, RI. (1994). Members of great teams think alike: A model
of team effectiveness and schema similarity among team members. Advances in
lt-zterdisciplinarv Studies of Work Teams, 1. 223-261.

Stevens. M.J., & Campion, MA. (1994). The knowledge, skill, and ability
requirements for teamwork: Implications for human resources management. Journal of

Management, 20(2). 503-530.

* Sundstrom, E.. DeMeuse, K.P., & Futrell, D. (1990). Work teams: Applications
and effectiveness. American Psychologist, 45, 120-133.

Tuckman, B.W. (1965). Developmental sequence in small groups.
Psychological Bulletin, 63, 384-399.

Wagner. J.A., & Hollenbeck, JR. (1995). Management of organizational
behavior. Englewood Cliffs, NJ: Prentice-Hall.

73

Weldon, E., Jehn, K.A., & Pradhan, P. (1991). Processes that mediate the
relationship between a group goal and improved group performance. Journal of
Personality and Social Psychology, 61, 555-569.

Weingart, L.R. & Weldon, E. (1991). Processes that mediate the relationship

between a group goal and group member performance. Human Performance, 4 (1), 33-
54.

Wood, R.E., Mento, A.J., & Locke, EA. (1987). Task complexity as a moderator
of goal effects: A meta-analysis. Journal of Applied Psychology,72, 416-425.

74

 

"‘illiltiltiillilii“