4-3-:,.-. -. ‘

 

‘-uaw-u '5. .vy; 7“ "I;

     

.v. d "'
31

..r.

   
  

_ ..¢---

   
  
  
 
   
    

  

--.. -

   
  

 
 

:-
< ,V. .

  
 
 
 
 

  
  
  

M-
,0. . . ..

“-
._. ,_., .« .
-4...
-},":..+'
'zcﬁn‘i

 

,._.
o

   
 

,

 
  

 
  
  

. v.15

 
 

.ﬂ' .

    
  

.

  

Av

p<

   
      

 

     
 
 
  

-. '17—

  
 
 

“, -. .-.
f:
-

o

 
 
  
  

      
 
 
 

.‘a

a...

 
  
   
 

“-
.
a. -
a .4»:
~n
-

W

m-.

    
  
     

.c. -v ‘
,--.

. $9.2- .
: vv-‘I
‘7‘ - . .—o
O. 45" ‘ -
v — I.
_.
1‘."-‘.'1
4 1.7.21

‘1‘

  
   

’ v—
f.

n
.A-=

,-

-9.
W:-
'35."

4..

  

our

 

   

I.
.—

 

THESIS

till/1WlllllillilllllillWilli/ll!

3 1293 01568 1228

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the

dissertation entitled

THE INFLUENCE OF SEX C(ldPOSITION AND TASK
SEX-LINKAGE (N DEEISION MAKII‘G
IN HIERARCHICAL TEADB

presented by

Jennifer Hedlund

has been accepted towards fulﬁllment
of the requirements for

Ph. D. degree in Psycholgy

 

 

 

Date W7

M5 U is an Afﬁrmative Action/Equal Opportunity Institution

PLACE N RETURN BOX to man this chooltout from your tooord.
TO AVOID FINES return on or More data duo.

DATE DUE DATE DUE DATE DUE

LEE tiygwjlﬁ ﬂ
. l‘ |

#3
ﬁr—Wm

MSU II An Afﬂtmdlvo Action/Equal Opportunity Institulon

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

THE INFLUENCE OF SEX COMPOSITION AND TASK SEX-LINKAGE
ON DECISION MAKING IN HIERARCHICAL TEAMS

By

Jennifer Hedhmd

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Psychology

1997

ABSTRACT

THE INFLUENCE OF SEX COMPOSITION AND TASK SEX-LINKAGE
ON DECISION MAKING IN HIERARCHICAL TEAMS

By

Jennifer Hedlund

As the 21st century approaches, organizations will continue to be challenged by an
increasingly diverse workforce and an expanding emphasis on work teams. These trends
ultimately lead to increased group heterogeneity, the implications of which are not well
known. The present study focused on the transition of men and women into traditionally
male and female-dominated ﬁelds, examining the effects of varying sex compositions on
group processes and performance. There is limited attention to group level variables in
the existing sex composition literature, with the available research generating more
confusion than knowledge. Two factors were suggested to contribute to this conﬁrsion:
(1) a failure to distinguish among speciﬁc representations of the sexes at the level of the
team (e.g., balanced vs. male-majority compositions), and (2) the inﬂuence of moderator
variables. The sex-linkage of a task was proposed as relevant variable for understanding
sex composition eﬁ‘ects. The effects of sex composition and task sex-linkage were tested
on hierarchical decision making teams which have traditionally been neglected in the
literature. Eighty- seven teams participated in one of four sex compositions and two sex-‘
linked tasks. Generally, male-dominated teams outperformed female-dominated teams on
three key decision processes. In addition, male-majority teams performed worse than all-

male teams on two of these variables. The inclusion of a single female member appeared

to negatively affect team members in the male-majority condition, particularly when the
task was male-oriented. Despite differences in the decision making processes, there were
no signiﬁcant diﬂ‘erences in team decision accuracy across compositions or tasks. Further
examination suggested that male and female teams used different methods to arrive at

comparable decision outcomes.

ACKNOWLEDGMENTS

This dissertation represents the ﬁnal step in achieving my long-time goal of
obtaining a Ph.D. It is a trying process at times and often will little reward except the
personal satisfaction of knowmg “I did it.” This accomplishment would not have
happened without the support, encouragement, assistance and occasional amusement
provided by those around me.

First, and foremost, I want to thank the most important person in my life, my soon-
to-be husband, Steve. His love and support was key to helping me overcome my doubts
and fears so that I could complete this project. I also want to thank my mom for believing
in me and reminding me that it’s not important what you do as long as your happy. I
treasure the feelings of pride that my brother, Jay, and my grandmothers, Ethel and Pat,
express toward me. And, I value the unconditional love provided by my cherished
companions, Sheba, Loki, and Dobie. Ihave to especially thank Sheba for catching a
minor error in my dissertation before printing.

There are numerous individuals who contributed to this dissertation from helping
to collect data to editing the ﬁnal draﬁ. I want to thank my chair, Dan Ilgen, for his
guidance and facilitation of this process, and my committee: John Hollenbeck, Steve
Kozlowski, Ann Marie Ryan, and Galen Bodenhausen for their helpﬁrl comments. I want

to recognize all those team lab researchers who helped in the data collection: Dan

Weissbein, Spence Tower, Jason Colquitt, Chris Daly, Matt “Cabin Boy” Taylor, and J eﬁ‘
LePine who I especially thank for his enduring support and encouragement.

I want to thank those whose ﬁiendships I value dearly. I could not have survived
graduate school without the support from Dave, Eleanor, Dennis, Stan, Matt, Sandy,
Mike, and all the other graduate students past and present. I want to thank Ilene for
continuously telling me to get it done, and Anne for thinking graduate school was as
difﬁcult as medical school Of course, the past year and a half would have been much less
tolerable without the friendship of Sandy Landis and the opportunity to spend time with
her son, Alex.

Finally, I would like to summarize my experiences by quoting the poet Robert
Frost:

Two roads diverged in a wood, and I--
I took the one less traveled by,

And that has made all the diﬂ‘erence.

TABLE OF CONTENTS

LIST OF TABLES ..................................................................................................... ix
LIST OF FIGURES .................................................................................................... x
INTRODUCTION ...................................................................................................... 1
Literature Overview ........................................................................................... 2
Focus on Individual Level Eﬂ‘ects .............................................................. 3
Unexplained Inconsistencies ...................................................................... 4

Limited Attention to Moderator Variables ................................................. 4

Study Overview ................................................................................................. 5
DIVERSITY .............................................................................................................. 7
Alternative Perspectives of Diversity .................................................................. 7
Social Categorization ................................................................................ 8
Ethnocentricity .......................................................................................... 8

Summary ........................................................................................................... 9

SEX STEREOTYPING ............................................................................................. 11
Inﬂuence on Perceptions and Behaviors ............................................................. 12
Greater Value and Status Associated with Male Characteristics ......................... l3
Behavioral Conﬁrmation of Stereotypes ............................................................. 14
Summary ........................................................................................................... l4

SEX COMPOSITION EFFECTS ............................................................................... 16
Sex Composition Theories ................................................................................. l7
Kanter’s Theory of Tokenisrn .................................................................... 17
Expectation States Theory ........................................................................ 18
Distinctiveness Theory .............................................................................. 19

Summary .................................................................................................. 20

Individual Level Eﬂeas ..................................................................................... 20
Stereotypic Perceptions ............................................................................ 21
Perceptions of coworkers ................................................................. 21

Perceptions of leaders ...................................................................... 24

Inﬂuence of the task ......................................................................... 27

Summary .......................................................................................... 28

Individual Performance ............................................................................. 29

vi

Participation and Comnnmication Content ................................................ 30

Conformity and Inﬂuence ......................................................................... 3 5
Conformity ...................................................................................... 36

Inﬂuence ......................................................................................... 38

Summary ......................................................................................... 41

Factors Moderating Stereotypic Perceptions and Behaviors ...................... 42
Summary .................................................................................................. 43
Group Level Eﬂ‘ects .......................................................................................... 44
Group Performance .................................................................................. 46
Group Processes ...................................................................................... 51
Summary ................................................................................................. 53

Role of the Task ............................................................................................... 54
Conclusion ....................................................................................................... 56
SEX COMPOSITION AND TEAM DECISION MAKING ....................................... 60
Decision Making Teams .................................................................................... 60
Identiﬁcation of Relevant Group Processes ....................................................... 61
Task Eﬂ‘ects ...................................................................................................... 61
Hierarchical Team Decision Making .................................................................. 62
The Multilevel Theory .............................................................................. 63

The TIDE2 Simulation .............................................................................. 66
Hypotheses ....................................................................................................... 67
Participation Equality ............................................................................... 70
Team Informity ........................................................................................ 72

Staﬂ Validity ............................................................................................ 73
Hierarchical Sensitivity ............................................................................. 75
Team Decision Making Accuracy ............................................................. 78
METHOD ................................................................................................................. 82
Design .............................................................................................................. 82
Participants ....................................................................................................... 82
Manipulations ................................................................................................... 84
Sex Composition ...................................................................................... 84

Task Sex-Linkage .................................................................................... 84
Procedure ......................................................................................................... 87
Measures .......................................................................................................... 88
Participation Equality ............................................................................... 88
Team Informity ........................................................................................ 88

Staﬂ Validity ............................................................................................ 89
Hierarchical Sensitivity ............................................................................. 89
Team Decision Making Accuracy ............................................................. 91

Sex Composition ...................................................................................... 91

RESULTS ................................................................................................................. 93

Participation Equality ........................................................................................ 95
Team Informity ................................................................................................. 99
Staﬂ Validity ..................................................................................................... 99
Hierarchical Sensitivity ..................................................................................... 101
Team Decision Accuracy .................................................................................. 101
The Multilevel Theory ...................................................................................... 103
DISCUSSION .......................................................................................................... 106
Participation Equality ....................................................................................... 108
Individual Participation Rates .................................................................. 109
Task Eﬂ‘ect .............................................................................................. 109
Team Informity ................................................................................................ 112
Dominance Eﬂ‘ect .................................................................................... 112
Interaction of Task, Dominance, and Homogeneity .................................. 114
Staﬂ‘ Validity .................................................................................................... 117
Individual Sex Diﬂ‘erences ....................................................................... 118
Task Eﬂ‘ect .............................................................................................. 121
Hierarchical Sensitivity ..................................................................................... 124
Male and Female Leaders ....................................................................... 124
Dyadic Sensitivity ................................................................................... 126
Team Decision Accuracy ................................................................................. 128
Study Limitations ............................................................................................ 131
Task Sex-Linkage ................................................................................... 131
Participation Equality .............................................................................. l3 5
Generalizability ....................................................................................... 137
Recommendations for Future Research ............................................................ 138
Conclusions ..................................................................................................... 141
FOOTNOTES .......................................................................................................... 143
APPENDD( A
HANDBOOK FOR TEAM MEMBERS:
NAVAL COMMAND AND CONTROL SIMULATTON ......................................... 144
APPENDIX B
HANDBOOK FOR TEAM MEMBERS:
NEONATAL NURSING SIMULATION ................................................................. 154
APPENDD( C
TEAM DITERACTIVE TRAINING SCRIPT .......................................................... 164
LIST OF REFERENCES .......................................................................................... 168

Table 1

Table 2

Table 3

Table 4

Table 5

Table 6

LIST OF TABLES

Study Design and Sample Distribution ..................................................... 83
Outcome Matrix for Determining Team Decision Accuracy ..................... 90

Means, Standard Deviations, and Intercorrelations
Among Variables ..................................................................................... 94

Summary of Hierarchical Regression Analyses for all
Dependent Variables ................................................................................ 96

Regression of Team Decision Accuracy on the
Core Decision Variables ......................................................................... 104

Regression of Team Decision Accuracy on the Core Decision
Variables for Male and Female-Dominated Teams .................................. 132

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

LIST OF FIGURES

The multilevel theory of hierarchical team decision making ........................ 65

Predicted eﬂects of sex composition and task sex-linkage on
hierarchical team decision making .............................................................. 69

Hypothesized effects of sex composition and task sex-linkage
on participation equality ............................................................................ 71

Hypothesized eﬂ‘ects of sex composition and task sex-linkage
on team informity ...................................................................................... 74

Hypothesized eﬂ‘ects of sex composition and task sex-linkage
on staﬂ validity .......................................................................................... 76

Hypothesized eﬂ‘ects of sex composition and task sex-linkage
on hierarchical sensitivity ........................................................................... 79

Hypothesized eﬂects of sex composition and task sex-linkage
on team decision accuracy ......................................................................... 81

Interaction of task sex-linkage and homogeneity on

participation equality ................................................................................. 97
Interaction of dominance and homogeneity on participation equality .......... 98

Interaction of dominance, homogeneity, and task sex-linkage

on team informity ..................................................................................... 100
Interaction of dominance and homogeneity on hierarchical sensitivity ....... 102

Interaction of team informity and hierarchical sensitivity on

team decision accuracy ............................................................................. 105
Individual participation rates for males and females by composition .......... 110
Individual informities of males and females by composition and task ......... 116

Figure 15 Frequencies in use of decision alternatives for males and females .............. 120
Figure 16 Frequencies in use of decision alternatives for the male and female tasks... 122
Figure 17 Dyadic sensitivities for males and females by composition ......................... 127

Figure 18 Direction of discrepancy between leader’s weights and appropriate
weights by composition and task ............................................................... 136

INTRODUCTION

Two important issues will continue to face many organizations as they approach
the 21st century. One is how to effectively manage the increasing diversity of the work
force. Another is how to use teams eﬂ‘ectively to perform organizational functions and
obtain organizational goals. There is no doubt that teams have become an important part
of organizations. Many organizations have redesigned jobs to increase task
interdependence (Jackson, 1992), thereby creating teams whose members are
interdependent. At the same time, the increasing diversity of the workforce leads to new
conﬁgurations of employees. These changes create challenges to traditional work
practices.

One of the many types of diversity that will impact organizations is gender
diversity.1 The number of women entering the work force continues to increase steadily
(Johnston, 1991). By the year 2005, women are predicted to compose 63% of the
workforce (Howard, 1995). Furthermore, women are increasingly entering ﬁelds
traditionally ﬁlled by men (Jackson, 1992). A prominent example of this transition is
women entering combat-related ﬁelds in the military. For example, the Navy, which
implemented a gender neutral recruiting policy in 1994, saw an increase in female recruits

from 12.6% to 22.2% between 1994 and 1995 (Hudson, 1995).

 

adc

2

Although women entering male dominated ﬁelds is more common and
controversial, there is a similar trend for males entering traditionally female dominated
ﬁelds (Craig & Sherif; 1986). For example, an increasing number of males are entering
the ﬁeld of nursing (Floge & Merrill, 1986; Ott, 1989). As a result of these changes, work
groups that were previously homogeneous in terms of the sex of their members will
increasingly consist of members of both sexes.

The issues of diversity and teamwork inevitably converge because increasing
diversity at the organizational level will most likely lead to increasing heterogeneity at the
team level (Jackson, 1992). This heterogeneity has many potential implications for team
performance since employees have traditionally worked with others who are similar to
themselves. Traditional members may apply inappropriate stereotypes to non-traditional
members, and even come to resent the increase of non-traditional members (e.g. females
entering military academies). Given the many potential diﬂiculties associated with
diversity, an important task for researchers is to identify the eﬂ‘ects of group heterogeneity
and to recommend how to manage this heterogeneity to insure eﬂ‘ective team
performance.

Literature Overview

Although group heterogeneity is clearly an important issue facing many
organizations and their employees, knowledge regarding its inﬂuence has been limited.
With the continuing trends toward greater diversity and increased teamwork, researchers
should begin to focus more attention on understanding the eﬂ‘ects of heterogeneous work
groups. Several limitations of the existing literature are identiﬁed along with attempts to

address those limitations.

Focus on Individual Level Eﬂ‘ects

The inﬂuence of diversity at the team level is most directly addressed by research
on group composition. Unfortunately, the focus of much of this research is on the eﬂ‘ects
of composition on the behaviors of individuals within groups rather than on group level
phenomena. There is an abundant literature that explores the eﬂ‘ects of sex composition
on male and female behaviors (e.g., Johnson & Schulman, 1989; Lockheed & Hall, 1976;
Strodtbeck & Mann, 1956), and reactions to male and female minority members (e.g.,
Bradley, 1980; Craig & Sherif, 1986). Research addressing the inﬂuence of sex
composition on group processes and performance, however, has been somewhat scarce.
The intent of the present study is to understand the eﬂects of group heterogeneity on
critical team2 level, rather than individual level variables, using the latter to help further
investigate the nature of those eﬂ‘ects.

The importance of focusing on the team level is that organizations are increasingly
designing work around teams (Hackrnan, 1990; Ilgen, 1994). As a result, performance
outcomes are a ﬁmction of the combined, oﬂen interdependent, efforts of employees.
Therefore, it is no longer reasonable to examine the performance of individuals in
isolation. Further, group interaction may have diﬂerent eﬂ‘ects on individual behaviors
than on group behaviors. For example, each member of a team may exhibit the highest
level of task competency, yet the team as a whole does not exhibit the highest level of
performance. The eﬂ‘ects on group level behaviors are expected to have a more direct

inﬂuence on group performance, and are, therefore, of primary interest.

Unexplained Inconsistencies

In addition to the limited attention to group level variables, the literature
addressing the effects of sex composition on group and individual behaviors is full of
inconsistencies. Many researchers argue that individual behaviors within groups follow
general sex stereotypes such that males are more task-oriented and females more socio-
emotional (e.g., Eagly, 1987). However, numerous studies indicate that this depends on
the group’s composition (e. g., Thamelirrg & Andrews, 1992). Unfortunately, it is not
clear exactly which compositions are associated with sex stereotypic or non-Stereotypic
behaviors.

One reason for some of this uncertainty is the failure to distinguish among various
types of nrixed- sex compositions. When researchers do diﬂerenﬁate between male-
dominated and female-dominated mixed-sex groups, signiﬁcant diﬂerences in individual
and group behaviors are often found (e.g., Craig & Sherif; 1986; Mabry, 1989).
Therefore, it is important to consider the representation of males and females within a
group when examining the eﬂ‘ects of group heterogeneity.

Unfortunately, inconsistencies still remain after taking into consideration the
representation of the sexes. Some research suggests, for example, that male-majority
groups make better decisions (Rogelberg & Rumery, 1994), while others ﬁnd that female-
majority groups perform better (Roberston & Kwong, 1994). Such inconsistencies
suggest the inﬂuence of a moderator variable.

Limited Attentign to Modergtor Variables

 

A handful of studies have examined potential moderators such as the experience

and status of group members (Lockheed & Hall, 1976; Wood & Karten, 1986). The most

 

 

111:

Tel;

5
frequently discussed moderator, however, is the task. Task requirements are oﬂen

suggested to favor certain compositions over others (Deaux, 1984, 1985). Many
inconsistencies in the sex composition literature are potentially attributable to the sex-
linkage of the task. That is, many tasks are linked, or associated, more oﬂen with one sex
than the other. As a result, males and females are expected to have different levels of
competency for diﬂerent tasks. For example, we assume that men are more
knowledgeable about auto mechanics, and women more knowledgeable about child care.
Similarly, secretaries are often assumed to be female and corporate executives male.

Unfortunately, the sex-linkage of tasks has oﬂen been ignored in studies examining
sex conrposition eﬂects. Typically, researchers suggest in hindsight that the nature of the
task may have favored male or female members (e. g., Craig & Sherif, 1986). The task’s
sex-linkage may help explain many previous inconsistencies found in the literature.

Study Overvie_\_v

The present study is designed to expand on the existing literature and address the
limitations discussed above. First, since group processes and performance have received
minimal attention in the literature, one purpose of this study is to examine the eﬂects of
sex composition on several group level variables related to the process of team decision
making. Second, this study will focus on speciﬁc conﬁgurations of males and females
which best reﬂect actual work environments (e.g., all-male, single-female), and which will
enable a better understanding of how slight changes in composition aﬂect group processes
and performance. Finally, this study will attempt to address inconsistencies in the existing
literature by examining the sex-linkage of the task as a potential moderator of the

relationship between sex composition and the team decision making process.

6
Several bodies of literature are reviewed to help identify the potential eﬂects of sex

composition on team decision making eﬂectiveness. This review will begin with a general
discussion of diversity which suggests some potential consequences of increased group
heterogeneity. Next, literature on sex stereotyping is discussed to identify the potential
biases that members may bring with them to a group. This is followed by a review of
research on sex composition, addressing both the individual and group level eﬂ‘ects.
Literature is also reviewed regarding the eﬂ‘ect of the sex-linkage of tasks which has often
been suggested to explain the inconsistent ﬁndings of prior research.

Following a review of the literature, a decision making task and an associated
theory are presented which will be used to address the eﬂects of sex composition and the
task’s sex-linkage. Based on the sex composition literature and the decision making
theory presented, several hypotheses are developed regarding the eﬂ‘ects of sex
composition and the sex linkage of the task on team processes and performance. The
method used to test the hypothesized relationships is then discussed, followed by a
presentation of the data analytic results. Finally, a discussion of these results, study

limitations and directions for future research is provided.

 

 

aj

Dc

DIVERSITY

The term diversity has been used to refer to numerous ways in which individuals
differ ﬁ'om one another. Diversity can refer to diﬂ‘erences in demographic characteristics
(e. g., age, sex, race, ethnicity), culture, personality, beliefs, ability, occupation, expertise,
or tenure. The work force is becoming increasingly diverse along a number of dimensions,
one of which is the sex of employees (Johnston, 1991).

Altmtive Perspectives of Diverm

There are basically two organizational perspectives regarding diversity. The
"value-in-diversity" approach, as suggested by Cox (1993), views diversity as an
advantage to groups and organizations because it broadens the skills, abilities, and
perspectives brought to a group. Maznevski (1994) refers to this as "role-related
diversity. " Role-related diversity is desirable because it brings together individuals with
complementary skills and abilities which is oﬂen the goal of designing work around teams.
Researchers have suggested that heterogeneous groups perform better on decision making
and creative problem solving tasks than homogeneous groups due to the greater range of
alternatives and perspectives that heterogeneous groups bring to the task (Hoﬂinan, 1979;
Hoﬁnan, Harburg, & Maier, 1962; Triandes, Hall, & Ewen, 1965). However, diversity is

not always associated with positive outcomes.

8
The alternative perspective of diversity suggests that by bringing together

individuals who are distinct ﬁ'om one another, diversity can create numerous problems
such as obstructing comrmrnications, inhibiting the contributions of certain members, and
decreasing group cohesion (Jackson, 1992). Demographic forms of diversity, in
particular, can be problematic because they highlight diﬂ‘erences between people which
can lead to stereotyping and discrimination. Assuming that increased diversity of the work
force is inevitable, organizations should be concerned with understanding these potential
problems and identifying ways to overcome them. Two processes help explain why
diversity can create problems among work groups, social categorization (i.e.,
stereotyping) and ethnocentricity.

Diﬂ‘erences in attitudes, values, interactions styles, and cognitive ability are often
associated with different demographic characteristics (Cox, 1993; Jackson, 1992). For
example, females are seen as nurturing and males as aggressive. Although these
associations may not always be valid, they are often used to evaluate the behaviors and
performance of those who appear diﬂ‘erent from us. In heterogeneous groups, these
diﬂ‘erences become highly salient, thus increasing the likelihood that members will be
categorized according to their distinguishing characteristics. Many negative eﬂects of this
categorization have been documented including isolation ﬂom the group (Kanter, 1977)
and inaccurate performance evaluations (Ilgen & Youtz, 1986).

Ethnocentri '
Another issue associated with diversity is ethnocentricity. Ethnocentricity refers to

the tendency to view members of one's own group more positively than other social

9
groups (Cox, 1993 ). There are consistent ﬁndings that people are more attracted to and

feel more comfortable and satisﬁed with members who are similar to themselves (Ziller,
1973). According to Cox, there are certain beneﬁts associated with working with similar
others. People feel that the behaviors of similar others are more certain and predictable.
In addition, people feel more satisﬁed that they are promoting the well-being of their own
social group. Finally, working with similar people provides an initial common ground for
establishing relationships with coworkers (Cox, 1993 ). Diversity, therefore, conﬂicts with
general ethnocentric tendencies.
Su_m_mg_ry

Given the potential problems associated with diversity, why is there a need to
pursue such integration? First, there are political, legal and ethical reasons why jobs can
no longer be restricted to those of a certain race and sex. Second, there are potential
beneﬁts in drawing from a wider range of perspectives, skills and abilities, which diversity
can oﬂ'er. For these reasons, employees will increasingly work with individuals who are
diﬂ‘erent from, rather than similar to, themselves. Unfortunately, due to ethnocentricity
and stereotyping, members of diﬂerent racial, gender, and ethnic backgrounds may have a
diﬂicult time working together as a team. However, given the inevitability of these
changes and the potential beneﬁts, it is important to understand the eﬂ‘ects of diversity and
to identify potential barriers to eﬂecﬁve team performance.

Since the present study focuses on gender diversity, the remaining discussion will
be limited to issues that are primarily relevant to understanding the eﬂects of mixed-sex

work environments. A critical issue in understanding the eﬂects of sex composition is sex

10
stereotyping. Sex stereotyping helps to explain why the transition of men and women into

ﬁelds traditionally dominated by the opposite sex can be problematic.

SEX STEREOTYPIN G

Sex stereotypes are widely held beliefs about the appropriate behaviors of men and
women (Broverman, Vogel, Broverman, Clarkson, & Rosenkrantz, 1972). There are
widely held beliefs that women are more "communal" and men more "agentic" (Eagly,
1987). Communal refers to behaviors associated with concern for the welfare of others
(e. g., nurturing, affectionate, expressive). Agentic behaviors include assertiveness,
independence and instrumentality.

These stereotypes are reﬂected in the division of labor between the sexes. Women
are disproportionately represented in occupations that are perceived to require communal
qualities while men occupy positions associated with agentic qualities. For example, males
are more likely to hold positions of leadership because they are perceived as having the
qualities associated with eﬂecﬁve leadership (e.g., assertiveness, irrstrumentality).
Although many women now serve in leadership roles, there are still a disproportionate
number of men in positions of authority (Eagly, 1987). Some have further suggested that
the diﬂerent characteristics attributed to men and women are promoted in order to
rationalize these occupational inequities (Hoﬂinan & Hurst, 1990).

Sex stereotypes inﬂuence our perceptions and behaviors in a variety of settings,
including work. Although the roles of men and women have changed substantially in the

past twenty years or so, there is evidence that traditional sex stereotypes persist (e. g.,

11

12
Heilman, Block, Martel], & Simon, 1989). In addition, the characteristics associated with

men tend to be more highly valued in our society, and the roles they occupy tend to be
associated with higher status (Broverman et al., 1972). As Eagly (1987) suggests, "A
status and power diﬂerence as pervasive as the one that separates women from men is
bound to have implications for gender stereotypes and roles" (p.23 ). The persistence of
sex stereotypes is often attributed to the process of behavioral conﬁrmation. The issues
surrounding sex stereotypes are discussed further below.
Inﬂuence on Perceptions and Behaviors

Although sex segregation may be declining in many ﬁelds, there are still factors
that attempt to maintain the barriers between the sexes. Men and women, for instance, are
often reprimanded for violating sex-role norms (Costrich, Feinstein, Kidder, Marecek, &
Pascale, 1975; Spence, Helmreich, & Stapp, 1975). Costrich et a]. (1975) found that
individuals who behaved counter to sex-role stereotypes were less popular. In addition,
Spence et a1 ( 1975) found that students indicated a preference for a masculine female on
an objective questionnaire, but their preferences reversed on open-ended questions. These
ﬁndings suggest that although individuals may publicly endorse less traditional views, their
underlying beliefs may reﬂect more traditional stereotypes regarding males and females.

Men and women are also attributed with diﬂerent traits and behaviors. Deaux and
Emswiller (1974) found that for traditionally male tasks, successful performance by
females was attributed to luck more than ability, whereas male success was attributed
more to ability than luck Similarly, Pazy (1986) formd that the causes attributed to
successful performance had a substantial impact on the ranking of both male and female

employees. Given the same attributions for performance, males were ranked signiﬁcantly

13
higher than females. Wiley and Eskilson (1983) obtained somewhat diﬂerent results when

considering promotions. Male promotions were less likely attributed to stable causes than
female promotions. This suggests that respondents may have viewed women as requiring
more skill and ability than men in order to be promoted.

Greater Value and Status Associated with Male Characteristics

Sex stereotypes also have an eﬂ‘ect on distributions of power and perceptions of
status. In general, women tend to be ascribed lower status than men (Eagly & Wood,
1982; Wharton & Baron, 1987 ; Wiley & Eskilson, 1982). For example, Eagly and Wood
had subjects read scenarios in which a comnmnicator was attempting to inﬂuence a
recipient, and asked them to rate the predicted inﬂuence and status of each individual.
Males were ascribed a higher status and believed to earn a higher salary than females.
Subjects also perceived behavioral compliance to be greatest in the male communicator-
female recipient condition. Most of these sex diﬂerences disappeared when information
about job titles was provided, with the exception that males were still seen as earning a
higher salary.

Wiley and Eskilson (1982) examined the eﬂecﬁveness of male and female uses of
reward-based or expert power in attempting to inﬂuence a male or female target. Overall
for female targets, there was no diﬂerence in perceived status between male and female
inﬂuencers. For male targets, however, male inﬂuencers were seen as having higher status
than females. Females attempting to inﬂuence a male were seen as the least powerful,
while males attempting to inﬂuence other males were seen as most powerful. Finally,
males were seen as more eﬂ‘ective when using expert inﬂuence, while females were

perceived as more eﬂective using reward-based power.

14
Finally, research has also shown that traits associated with females tend not to be

associated with leadership. Schein (1973), for example, found a signiﬁcant relationship
between the stereotypes attributed to men and those attributed to managers, but there was
no relationship between the attributes of women and managers. These ﬁndings were more
recently replicated by Heilman et a1. (1989) who further found that when subjects were
asked to identify the characteristics of female managers, these were still less associated
with the attributes of successful managers than were characteristics identiﬁed for male
managers.

Behavioral Conﬁrmation of Stereogpes
Finally, sex stereotypes persist because they are reinforced through the process of

behavioral conﬁrmation. That is, people oﬂen comply with the expectations held by
others as well as themselves (Eagly, 1987). These expectations are conveyed through
verbal and nonverbal means, and conﬁrmatory responses are rewarded. For example,
males may receive more opportunities to lead a group based on the assumption that they
have more leadership skill. This results in more males serving in the role of leader, which
is then perceived as indicating that males are better suited for leadership than females.
Eagly further suggests that expectancy-conﬁrming behavior is more likely to occur when
the expectations are widely shared as is the case with sex stereotypes.
_S_ul_n_ma_ry

The literature on sex stereotyping suggests that an individual’s sex can have
extensive effects on our attitudes, beliefs, and behaviors. Further, these stereotypes have
remained fairly consistent over time. Thus, it is likely that sex stereotypes will continue to

inﬂuence the perceptions of men and women, and thus, their interactions with one another.

15
The eﬂ‘ects of sex composition on the perceptions and behaviors of group members are

discussed next.

SEX COMPOSITION EFFECTS

Research interest in the eﬂ‘ects of group sex composition has a long history (e.g.,
Burtt, 1920), but few researchers recognized sex composition as an important variable in
the design of their studies prior to the 19608. In much of the early literature on groups,
women were not even included as subjects (Dion, 1985). From the late 19603 through the
1970s, there was a surge in the number of studies that speciﬁcally addressed the eﬂects of
group sex composition. This interest seemed to coincide with increasing efforts to change
societal perceptions of women. Research interest tapered oﬂ somewhat in the 1980s and
1990s. This might be considered an indication that sex stereotyping had diminished.
However, given fairly recent evidence regarding the potency of sex stereotypes (e.g.,
Heilman et al, 1989), it is more likely that research interest simply discontinued in the
topic. As a review of the literature will show, the literature on sex composition is laden
with inconsistencies which may have been a major deterrent to researchers.

The review of sex composition eﬂects begins with a discussion of several theories
which attempt to explain how the composition of a group aﬂ‘ects the behaviors of its
members. This is followed by a review of the individual level eﬂects of sex composition
which comprises a majority of the sex composition literature. The effects of composition
on the perceptions of male and female members (e.g., ratings of competence) and on

individual behaviors (e.g., levels of participation) are discussed. The group level eﬂects,

16

17
which represent a smaller portion of the sex composition literature, are reviewed next. A

variety of group process and outcomes are addressed in this literature. Finally, the role of
the task as a moderator of sex composition eﬂects on individual and group level variables
is discussed.
Sex Composition Theories

Several theories have been developed to address the eﬂects of sex composition.
These theories focus primarily on the individual level eﬂects of varying compositions.
That is, they attempt to explain how indivithral behaviors and perceptions are inﬂuenced
by different representations of males and females.
K_anter’s Theory of T omLm

Kanter (1977) developed a theory of tokenism which suggests that the numerical
representation of the sexes can inﬂuence the roles of men and women. First, she
distinguishes between “tmiform,” “skewed,” “tilted,” and “balanced” mixed-sex
compositions. “Skewed” compositions include less than 15% membership of one sex,
“tilted” compositions closer to 35% of one sex, “balanced” groups are composed of equal
proportions of males and females, and in “rmiform” groups members are all of the same
sex The theory of tokenism focuses on skewed compositions which represent the
extreme condition, and “one encountered by large numbers of women in groups and
organizations in which numerical distributions have traditionally favored men” (Kanter,
l977,p.966)

Tokenism refers to a very small representation of one sex. The minority member’s
social category is physically obvious. Further, the individual may be new to the setting as

is the case in many occupations traditionally dominated by one sex. Tokenism, therefore,

 

M.

18
has three primary eﬂ’ects. First, the visibility of minority members is heightened by their

representation. Second, diﬂerences between the majority and minority members tend to
be exaggerated, which can result in the isolation of the minority members. Third, the
behaviors of minority members are distorted to ﬁt stereotypes about that member.

Kanter’s theory was developed based on the experiences of female tokens. She
suggests that the concepts can be applied to other tokens, including males in female-
dominated environments. Several studies, to be discussed in more detail later, have not
found the same eﬂ‘ects for male as for female tokens (Alexander & Thoits, 1985; Floge &
Merrill, 1986). In fact, males sometimes experience positive eﬂects as a result of their
minority representation.

Kanter’s theory also fails to explain the stereotypical behavior of majority
members. In a study by Strotbeck and Mann (1956) cited to support the concepts of
tokenism, men were more task-oriented and women more socio-emotional when working
in mock jury decision making groups with a disproportionate representation of men.
Kanter (1977) suggested that “it was women’s scarcity in skewed groups that pushed
them into classical positions and men’s superiority that gave them the edge in task
performance” (p.970). The stereotypical behavior of men under majority representation is
not consistent with the tenets of tokenism.

Emﬁon Stgtes Theory

Expectation states theory suggests that there are general status diﬂerences
associated with males and females in society (Berger, Fisek, Norman, & Zelditch, 1977),
and, as a result, sex ﬁrnctions as a diﬂuse status characteristic (Lockheed & Hall, 1976;

Meeker & Weitzel-O'Neill, 1977). In other words, general status diﬂ’erences in society

19
inﬂuence performance expectations in more speciﬁc settings (Meeker & Weitzel-O'Neill,

1977). Since men tend to hold positions of higher status than women, they are expected
to be more competent in a variety of settings. This implies that, in work groups, men are
oﬂen perceived as having more leadership skill and intelligence than women. This helps
explain why the experiences of male tokens are typically not as negative as those of female
tokens, and thus expands on Kanter’s theory.

Distinctiveness Theory

Finally, distinctiveness theory, as proposed by McGuire, McGuire, and Winton
(1979), provides another explanation for the eﬂ‘ects of group sex composition on
individual perceptions and behaviors. Unlike tokenism and expectation states theories,
distinctiveness theory focuses on self-perceptions. Basically, the theory states that mixed-
sex groups, particularly skewed groups, serve to highlight sex as a distinguishing
characteristic of members (Cota & Dion, 1986; McGuire et a1, 1979; Ruble & Higgins,
1976). This, in turn, increases the awareness of one’s own sex and the likelihood that
diﬂerences between one’s own sex and that of other members are exaggerated.

Several studies provide evidence of such an effect. McGuire et al (1979) found
that both males and females were more likely to use sex as a characteristic in their self-
descriptions the smaller the representation of their own sex was in their reference group
(e.g., family). Similarly, Cota and Dion (1986) found that more individuals mentioned sex
in their self-descriptions when their sex was in the minority than in the majority.
Therefore, minority representation not only increases the likelihood that other members

will take notice, it also makes one more aware of his or her own sex.

20
may

Both distinctiveness and tokenism theories suggest that the numerical
representation of the sexes plays an important role in inﬂuencing perceptions and
behaviors. Although it is fairly well accepted that the numerical representation of males
and females aﬂects individual behaviors, the negative eﬂ‘ects of minority representation
may not be eiqrerienced equally by the sexes. A major drawback of Kanter’s theory is that
it does not account for the diﬂ‘erent experiences often found for male versus female
tokens. Expectations states theory, which suggests that general status diﬂ’erences in
society between males and females transfer to speciﬁc settings, provides an explanation as
to why men oﬂen experience minority representation more positively than females.
Together, these theories suggest that minority representation will increase the likelihood
that a member is viewed stereotypically, but whether that results in a negative or positive
outcome may depend on the sex of that member.

Tests of the adequacy of these theories are found throughout the literature
addressing the eﬂ’ects of sex composition. The eﬂ’ects of composition on individual
behaviors and perceptions, and the effects on group processes and performance are
reviewed below.

mall Level—Effects

Although understanding relationships between sex composition and group level
variables is the goal of this study, most of the research addresses how the sex composition
of a group aﬂ‘ects behaviors and perceptions of individual male and female team members.
Therefore, individual level eﬂ’ects will be reviewed. First, research examining the eﬂects

of different sex compositions on the perceptions of male and female members is reviewed.

2 1
Second, literature addressing the eﬂ‘ects of sex composition on individual behaviors is

discussed. This literature examines male and female diﬂ’erences in performance,
communication, conformity, and inﬂuence. These studies identify individual level effects
that can potentially inﬂuence group processes and outcomes.

W

Substantial evidence exists suggesting that men and women are viewed diﬂ‘erently
within mixed-sex as compared to same-sex groups. Several studies highlight the
stereotypic assumptions made about group members and how these vary across
compositions. Some of these studies focus on male versus female leaders, and provide
good examples of consistent biases in the treatment of males and females.

Perceptions of coworkers. Several studies have examined the ratings of male and
female coworkers (or performers) under diﬂ’erent composition. Toder (1980) asked
participants to evaluate journal articles on the topics of nutrition, education, urban
planning and language that were authored by a male or female. Evaluations were made
either in the context of all-female or mixed-sex groups. Women in the mixed-sex groups
evaluated female authored articles less favorably than male-authored articles, but women
in all-female groups evaluated male and female authors equally favorably. For the city
planning and nutrition articles, female authors were generally rated lower. For the
education article, lower ratings for female authors were given only by subjects who were
assessed as having traditional sex-role attitudes. Finally, no diﬂ‘erences were found for the
language article. This study indicates that the group's composition can inﬂuence individual
evaluations such that in the presence of men, women may rate other women less favorably.

Furthermore, on at least two topics, female authors were perceived less favorably than

22
male authors, suggesting that other factors (e.g., task content) may be considered when

making evaluations.

In a study by Nemeth, Endicott, and Wachtler (1976) subjects were asked to rate
male and female members on several factors following participation in a jury decision
making task. Each group was composed of six participants with no less than two persons
of each sex. Across all groups, males were rated higher on a number of characteristics
including intelligence, inﬂuence, conﬁdence, independence and aggressiveness, and were
perceived as the leader more often than females. Females were only rated as more likable
than males.

The above studies (Nemeth et al., 1976; Toder, 1980) suggest a general bias in
favor of males which may be inﬂuenced by the sex composition of the group.
Unfortunately, these studies ignored the potential inﬂuence of diﬂ’erent proportions of
males and females. Nemeth et a]. combined male-majority, female-majority and balanced-
sex groups in their analyses, potentially masking certain composition effects. The studies
discussed next take into consideration the proportion of males and females in a group, not
just its heterogeneity.

Three studies addressed the relationship between composition and stereotypic
perceptions in ﬁeld settings. First, Izraeli (1983) compared skewed (20% or less women)
and balanced groups (41 to 60% women) of part-time local union oﬁcers in Israel This
study assessed Kanter’s (197 7) theory of tokenism discussed earlier. Izraeli found that
women in balanced groups rated other women as having more skill than men, but those in
skewed groups viewed men as more skilled. Men rated other men more favorably than

women, regardless of composition. Members of skewed groups were more likely than

23
members of balanced groups to feel that women were elected to represent other women,

while men were elected to represent all members. Finally, women in the skewed groups
felt less inﬂuential than women in the balanced groups. These ﬁndings suggest that
skewed compositions can lead women to perceive themselves as less competent and less
deserving of their positions.

A second study addressed the eﬂ’ects of diﬂ’erent proportions of women on the
responses of coworkers. South, Bonjean, Markam, and Corder (1982) examined the
frequency of contact with and social support received from coworkers and male
supervisors of groups that were part of a government inspection agency. They found that
the greater proportion of female members in the group, the less contact females had with
male coworkers, but the more contact they had with female coworkers. The proportion of
females was not related to contact with male supervisors or the amount of social support
received from male coworkers. As the proportion of females in the group increased,
however, the less support received from female coworkers and male supervisors.
Unfortunately these results may simply reﬂect a change in the base rate of females in the
group. That is, increasing the proportion of females by chance alone should increase the
opportunities for contact with female coworkers and decrease the opportrmities for
contact with male coworkers. However, these ﬁndings do raise the point that the number
of females in a group can inﬂuence the responses of coworkers and supervisors toward
female members.

Finally, Floge and Merrill (1986) studied the perceptual responses to male nurses
and female physicians, which both ﬁt Kanter’s (197 7) deﬁnition of “token” members.

Both male nurses and female physicians were more visible than their opposite sex cohorts,

24
and experienced exclusion from the social networks of the majority. Male nurses,

however, were more likely to be assigned leadership positions without the requisite skills.
They were also treated more like coworkers by male physicians and perceived as more
competent by those physicians than were female nurses Further, female physicians were
attributed with stereotypically feminine traits (e.g., compassion, warmth) and were treated
diﬂerenﬂy than male physicians. These ﬁndings, although solely based on observations
and interviews, suggest that tokens are treated diﬂerenﬂy than non-tokens. However, the
nature of the treatment appears to vary such that male tokens are treated more favorably
than female tokens.

Perceptions of leaders. Perceptions of male and female leaders further highlight
the bias in favor of males. Webber (1976) compared the perceptions of members in male-
majority and female-majority groups following participation for a semester in class project
groups. In the male-majority groups, more men than women were seen as the task leader
and the highest contributor to the group. Women were more often viewed as non-leaders,
social leaders, and the lowest contributors than men In the female-majority groups, men
were again more frequently considered the task leader and the highest contributor than
women. However, more women in the female-majority groups perceived the men to be
non-leaders and the lowest contributor than in the male-majority groups. Women
apparently varied in their perceptions of the single male member of their group. Finally,
women in the female-majority groups were more likely to be viewed as the task leader and
highest contributor than in the nude-majority groups.

Yerby (1975) compared reactions to female leaders in all-female, single-female and

balanced-sex groups. These groups were designed to have either initially all positive or all

25
negative attitudes toward female leadership. Balanced- sex groups with positive attitudes

were most satisﬁed with their female leaders, while single-female groups with positive
attitudes and balanced-sex groups with negative attitudes were the least satisﬁed. These
studies clearly indicate that perceptions of male and female leaders are inﬂuarced by group
composition. Although males are generally perceived as leaders more often than females,
male-dominated groups tend to foster more negative attitudes toward female leadership
than female-dominated groups.

Several studies ﬁrrther illuminate the diﬂ‘erential treatment of male and female
leaders. There is no variance in composition within these studies, but looking across them
allows for some conclusion about sex composition eﬂ’ects. These studies also highlight
additional distinctions between perceptions of male and female leaders.

Kushell and Newton (1986) had male and female leaders portray an autocratic or
democratic style in mixed- sex groups working on survival task (NASA Moon Landing
exercise). Female subordinates were more satisﬁed with democratic leaders, while male
subordinates were more satisﬁed with autocratic leaders. However, there were no
differences in how female and male autocratic or democratic leaders were perceived, or
how female and male leaders were perceived in general.

A similar study by Korabik, Baril, and Watson (1993) found no main effect for sex
in subordinate ratings of leader eﬂ‘ectiveness in groups who worked on the Change of
Work Procedure task However, females using a dominating style were rated as less
eﬂective than males, and males using an obliging style were rated less eﬂ’ective than
females. This ﬁnding suggests that individuals may be evaluated less favorably when their

behaviors are incongruent with expectations regarding members of their sex

26
Goktepe and Schneier (1988) examined the emergence of male and female leaders

in mixed- sex groups working on case studies related to personnel and business policy
issues. Slightly more males than females emerged as leaders, but there were no diﬂerences
in the eﬂ‘ectiveness ratings for male and female leaders. Male leaders, however, were
perceived as more masculine and female leaders as more feminine. This study ﬁrrther
suggests that although male and female leaders may be rated as equally eﬂ‘ective, they are
often perceived as performing diﬂ‘erent ﬁmctions in the role as leader.

A study by Greene, Morrison, and Tischler (1981) also suggests that males and
females are perceived as oﬂ‘ering diﬂ‘erent skills as leaders. They compared male and
female co-leaders of mixed- sex groups attending a two-day conference on group
processes. In each condition, a leader of one sex served as an associate leader to the main
leader of the opposite sex. Co-leader arrangements with female leader and male associate
were perceived as more emotionally responsive than male leaders with female associates.
Male co-leaders, regardless of their authority, were seen as signiﬁcantly more active,
instrumental, and insightful than female co-leaders. The characteristics attributed to male
and female leaders reﬂect traditional sex-role stereotypes.

Similar ﬁndings were obtained by Butler and Geis (1990). They observed males
and females serving as co-leaders or solo leaders of mixed-sex groups who worked on a
survival task (NASA Moon Landing exercise). In general, subordinates were more
pleased with male solo and co-leaders than with female leaders. Male and female leaders
were evaluated as equally competent on the task, but male leaders were viewed as having
more general skill, ability and intelligence than female leaders. Female leaders were rated

as more emotional and sensitive, but also more bossy and dominating than male leaders.

27
Subordinates indicated that they wanted less contribution from female leaders than from

male leaders. Subordinate sex had an eﬂ‘ect on the ratings such that females responded
more positively than males to female leaders. These results again suggest a sex bias
favoring males such that they are perceived more favorably on dimensions related to
leadership. It is also clear that individuals make distinctions between male and female
leadership which are consistent with general sex stereotypes.

Buﬂer and Geis (1990) suggested that the male bias found in their study may have
been attributable to the nature of the task In fact, many of the tasks used in the above
studies can be considered male-oriented (e.g., NASA Moon Landing exercise). Therefore,
the rating of males as better leaders may not simply reﬂect general stereotypes, but the
additional consideration of the task

I_nﬂ_uence of the tags. In support of Butler and Geis’ suggestion, Wentworth and
Anderson (1984) found that the emergence of male leaders more often than female leaders
was dependent on the task Balanced-sex groups worked on a problem involving the
distribution of a large sum of money. The speciﬁc problem was designed to be either
masculine (investing a young cousin's inheritance), feminine (paying for a wedding) or
neutral (entertainment spending for a married couple). Males were more likely to emerge
as leaders on the masculine task, but there were no signiﬁcant diﬂ’erences in the emergence
of male and female leaders on the neutral or feminine tasks. Females, however, did
emerge more often as leaders in the feminine than in the masculine condition.

Task knowledge was also assessed and related to leadership emergence
(Wentworth & Anderson, 1984). Males expressed more knowledge on the masculine task

and females more knowledge on the feminine task On the feminine task, there was a

28
signiﬁcant relationship between knowledge and leadership emergence. On the neutral and

masculine tasks, there was no relationship between knowledge and leadership. This
suggests that while females had to demonstrate knowledge to obtain leadership status,
males did not. Therefore, the bias in favor of males may be stronger on tasks which are
perceived as male-oriented.

gem. The above studies suggest that the perceptions of males and females,
whether they are coworkers or leaders, tend to reﬂect sex stereotypes. In general, males
tend to be rated higher than females on intelligence, instrumentality and leadership (e. g.,
Nemeth et al, 1976; Webber, 1976). In terms of leadership, males and females are often
attributed with diﬂerent qualities. Speciﬁcally, male leaders are seen as instrumental and
female leaders as emotional and supportive (Butler & Geis, 1990; Goktepe & Schneier,
1988; Greene et al., 1981). These perceptions, however, can be inﬂuenced by the sex
composition of one’s reference group. For example, females are rated more favorably as
their representation in the group increases (Israeli, 1983; Toder, 1980). Finally, there is
evidence that these perceptions are inﬂuenced by the sex-relatedness of the task
(Wentworth & Anderson, 1984).

These studies clearly indicate the role of sex stereotypes in individual perceptions.
The question now becomes to what extent are these stereotypes also reﬂected in individual
behaviors. Research addressing the eﬂects of sex composition on individual behaviors
tends to focus on variables for which there are well-established sex stereotypes (e.g.,
participation, inﬂuence, conformity). Therefore, the question most frequently addressed is
the extent to which males and females behave stereotypically under diﬂerent group

compositions.

29
Individual Performance

A few studies have addressed male and female diﬂ‘erences in performance across
various group compositions. Two studies compared male and female performance in
mixed-sex groups only and formd no differences between males and females. Burtt’s
(1920) study examined the accuracy of judgments made alone or in mixed-sex mock juries
regarding subjects who lied or told the truth. There were no diﬂerences between males
and females in their individual judgment accuracy as a ﬁmction of the group.

Timmons (1941) studied balanced-sex groups of high school students who
discussed and ranked solutions to the topic of releasing convicts ﬂom prison. Both males
and females improved in their ranking scores following group discussion, with a non-
signiﬁcant tendency for females to improve more than males. These studies suggest little
diﬂ’erence in the performance of males and females, but are limited in their examination of
composition effects.

In a more comprehensive study of sex composition eﬂects, Alexander and Thoits
(1985) found diﬂ‘erences when they compared the achievement of males and females in
male-dominated, female-dominated and balanced university departments. Departments
were also classiﬁed as either skewed or tilted according to Kanter’s (197 7) deﬁnition. In
male-dominated departments, the achievement of men and women did not diﬂ‘er when the
composition was skewed, that is, when women represented less than 20% of the
department, but women tended to perform better than men in tilted departments where
there was 21 to 35% female representation. In female-dominated departments, men and
women's achievements did not diﬂer when the composition was tilted (21 to 35% male),

but men tended to outperform women when the representation was skewed (less than

30
20% male). In balanced departments, there were no differences between the performance

of men and women. These ﬁndings suggest that the performance of females may suffer
ﬁom small representation. Males, however, do not experience similar performance
deﬁcits rmder minority representation, but instead may actually perform better under such
conditions.

Although the research on individual performance is limited, some conclusions can
be drawn. Both Alexander and Thoits (1985) and Timmons’ (1941) ﬁndings suggest that
the performance of males and females do not diﬂ‘er when the sexes are equally
represented. Further, these ﬁndings were made over 40 years apart, suggesting some
consistency in this eﬂ’ect over time. The contribution of Alexander and Thoits is to show
the complexity of sex composition eﬂ’ects. The numerical representation of the sexes is
clearly important in understanding the eﬂ‘ects of composition. This is further suggested by
research on participation and comrmmication content.

Partic_rp' ation and Communication Content

The focus on sex stereotypic behavior is probably most apparent in research on
individual participation rates and communication content. Much of this literature begins
with the assumption that males and females generally behave stereotypically (i. e., males
are more task-oriented and females more social), and examines the extent to which the
composition inﬂuences this behavior. Therefore, the question typically addressed is
whether mixed-sex compositions induce more or less stereotypic behavior. As will
become clear through the following review, there is no deﬁnitive answer to this question.

Piliavin and Martin (1978) hypothesized that individuals would express less sex-

stereotypic behavior in same-sex than in mixed- sex groups. They observed and coded

3 1
group discussions of social problems over three trials. During the second trial, one

member in each group was encouraged to increase his or her participation by using a green
light to indicate when they should speak. Prior to the intervention, same- sex groups were
found to interact more stereotypically than mixed-sex groups, contrary to their prediction.
Members of all-female groups engaged in more socio-emotional activity while those of all-
male groups were more task-oriented. In mixed- sex groups, both males and females
exhibited less stereotypic behavior.

Following the intervention, mixed-sex groups in which a female was targeted
exhibited no signiﬁcant sex diﬂ‘erences in the amormt of task or socio-emotional activity.
However, when a male was targeted, there was an increase in stereotypic behavior. Males
became signiﬁcantly more task-related and females signiﬁcantly more socio-emotional In
these male target groups, there appeared to be some conﬂict over male dominance which
may have contributed to the more stereotypic behavior exhibited by both males and
females. Finally, in same-sex groups, both males and females experienced an increase in
task activities and a decrease in socio-emotional behaviors following the intervention
(Piliavin & Martin, 1978).

Unlike Piliavin and Martin’s (1978) results, Johnson and Schuhnan (1989) found
individuals to initially behave more stereotypically in mixed- sex groups. Groups of
varying sex compositions were observed while they discussed a scenario about a middle-
aged businessman who is asked for information regarding his membership in the
Comrmmist party at age twenty. The authors compared the top and bottom activity levels
(as rated by group members) across each composition. The top level of task activity for

females decreased as their representation in the group decreased. The bottom activity

32
level for females, however, was not aﬂected by composition. The top task activity level

for males only decreased from balanced to single-male groups. The bottom level of task
activity for males increased as their representation in the group decreased. In balanced-
sex groups, the top male task activity was signiﬁcantly higher than the top female task
activity. For same- sex groups, there were no diﬂerences between top male and female
task activity. Further, in all mixed-sex groups, the bottom level of male task activity was
higher than the bottom level of female task activity.

In terms of socio-emotional activity, the top level of both male and female activity
decreased as their representation in the group decreased. The bottom level of both male
and female socio-emotional activity, however, increased as their representation in the
group decreased. This study suggests that as the representation of one's sex in the group
decreases, task activity levels, or at least ratings of task activity, conform more to sex
stereotypes such that males become more active and females less active. However, males
and females were affected similarly in their levels of socio-emotional activity by the
group’s composition.

Spangler, Gordon, and Pipkin (1978) obtained similar results in a ﬁeld study
comparing law students from two schools, one which had 20% women (skewed) and the
other 33% women (tilted). In the skewed program, women were less likely to speak in
class and more likely to consider quitting school than men. This was not true of the tilted
program. In the skewed program, women also had more diﬁculty interacting with their
professors than men, while no diﬂerences existed between men and women in the tilted
program There were additional diﬂ’erences between the two schools which may have

contributed to the results. For example, one school was private, the other public.

33
However, these ﬁndings are consistent with other results suggesting that minority

representation has a negative aﬂect on females.

Aries (1976) examined diﬂerences in participation, dominance patterns, and
commrmication content as a ﬁmction of the group’s composition. All-male, all-female,
and mixed- sex groups were studied as they met over ﬁve sessions with the task of
becoming acquainted. In mixed- sex groups, males both initiated and received more
interaction than females. In all-male groups, positions of dominance and submissiveness
were constant over time, but in all-female groups dominance was less stable. The amount
of interaction addressed to the group as a whole versus individuals was also viewed as an
indicator of dominance. Members of all-male groups addressed signiﬁcantly more
interactions to the group than members of all-female groups. In mixed-sex groups,
however, males addressed less interactions to the group than they did in all-male groups.

In terms of content, Aries (1976) found that all-male group discussions reﬂected
themes of competition and status, while all-female groups discussed issues of intimacy and
interpersonal relations. In mixed-sex groups, males reduced the number of discussions
relevant to competition and status, and instead were more personally oriented. The
content of female discussions, however, only changed slightly in mixed-sex groups.
Finally, in mixed- sex groups there was greater use of qualiﬁcations and more
defensiveness by both males and females than in same-sex groups, suggesting less
conﬁdence on the part of both sexes in the mixed-sex condition. These ﬁndings suggest
that some stereotypic behaviors increase in mixed-sex environments (e.g., higher male

participation rates), while others decrease (e.g., male dominance and competitiveness). In

34
other words, there may not be a simple relationship between composition and the presence

of sex stereotypic behaviors.

The behaviors exhibited toward other members may also reﬂect sex stereotypes.
Smith-Levin and Brody (1989) studied sex diﬂ‘erences in interruptions during group
discussions. Males were more likely to interrupt a female speaker than a male speaker,
while females attempted to interrupt males and females at an equal rate. Composition had
no eﬂect on the number of interruptions attempted, but did inﬂuence the content of these
interruption. Males used more supportive interruptions in all-male groups, but this
declined as the proportion of women in the group increased. There were no differences in
the number of negative interruptions made by males and females across groups, except
that females were more likely to yield to negative interruptions from males. In mixed-sex
groups, positive interruptions were successful only half the time, but in same-sex groups
they were more than twice as likely to succeed Finally, females had less success than
males when trying to interrupt a male using neutral comments. These ﬁndings suggest
that the behaviors of males and responses of females may reduce the opportunities for
women to participate. Therefore, observed diﬂerences in behaviors may be the result of
the stereotypes held by other members rather than actual sex differences.

Evidence of non- stereotypic behavior by women ﬁrrther supports the argument
that observed diﬂerences reﬂect the inﬂuence of stereotypes rather than actual sex
diﬂerences. Mabry (1989) observed groups composed of 60% females and 40% males
who worked on a human relations task over ﬁve sessions. Women participated more than
men and engaged in more socio-emotional and task-related interactions. Overall, women

appeared to favor same-sex interactions and tended to dominate the group's interactions.

35
The latter is contradictory to many previous ﬁndings and is likely attributable to the

greater proportion of females in the group. These results suggest that females and males
may not always behave stereotypically.

Similar results were obtained by Thune, Manderschied, and Silbergeld (1980) who
observed small therapy groups of either married couples or teachers. In the couples'
group, which had an equal number of males and females, males interacted more ﬂequently
than females and exhibited more instrumental behaviors while females were more socio-
emotional. In the teachers' group, which was composed of a female majority, females
interacted more ﬁ'equently than males and were more instrumental, while males were more
socio-emotional.

Based on the above studies, it appears that males and females do not behave
stereotypically across all situations. Some studies suggest that males and females behave
more stereotypically in same-sex groups (Aries, 1976; Piliavin & Martin, 1978), while
others indicate more stereotypic behavior in mixed-sex groups (Johnson & Schulman,
1989). Two studies suggest that when the group is female-dominated, both males and
females engage in non-stereotypical behavior (Mabry, 1989; Thtme et al., 1980).
Therefore, the extent to which a group is dominated by males or females, rather than
simply its heterogeneity, may determine how stereotypically its members behave.
Research on conformity and inﬂuence provide further evidence that the numerical
representation of the sexes can have a signiﬁcant eﬂ’ect on individual behavior.
Confom' and Inﬂpence

Two frequently studied behaviors in the literature on sex composition are

conformity and inﬂuence. This research highlights both the importance of the dominant

36

sex of the group as well as the task’s role in moderating the effects of various
compositions.

Confomg' . Several studies have compared the conformity of males and females
under diﬂ‘erent compositions and on a variety of tasks. In each study, the opinion of the
majority is manipulated. First, Reitan and Shaw ( 1964) studied the conforming behavior
of individuals in all-male, all-female or balanced-sex groups. The task involved either
choosing a line with the same length as a standard line or the ﬁgure with the greatest area.
The experimenter manipulated the responses of the other group members as they were
seen by participants. Overall, females conformed more than males to the responses of
other members. However, all subjects, regardless of sex, conformed more in balanced-sex
groups. Subjects also eiqiressed more concern about disagreements and were less
conﬁdent of their accuracy in balanced-sex than in same- sex groups.

A similar study by Tuddenham, MacBride, and Zahn (195 8) observed subjects who
made judgments about either visual, information, or opinion problems. Again, the
experimenter controlled the judgments seen by each participant. Overall, females
conformed more than males, with larger sex diﬂ‘erences observed in mixed than same-sex
groups. On the information and opinion tasks, females conformed more in male-majority
groups, while males conformed more in female-majority groups. On the visual task, males
conformed more than females in male-majority groups. These two studies show that
although females have a general tendency to conform more than males, this ﬁnding is
dependent on the group's composition. Further, as shown by Tuddenham et al (1958),
diﬂerences in conformity may also vary as a frmction of the task. Two studies further

examine the task’s inﬂuence on conformity.

37
Feldman- Summers, Montano, Kasprzyk, and Wagner (1980) asked subjects to

respond to discussions by either male majority or female majority groups on topics
predetermined to be either male or female-related. The majority position was designed to
be opposite of that held by most subjects. On female-related issues, subjects were
inﬂuenced more by the female-majority, and on male-related issues they were more
inﬂuenced by the male-majority. In addition, male subjects conformed more than females
on the female-related issue, and female subjects had a tendency to conform more than
males on the male-related issues.

Sistrunk and McDavid (1971) had subjects indicate their level of agreement or
disagreement with items chosen to be masculine, feminine, or neutral Indicated next to
each item was a "majority response" which was used to assess conformity. Males
conformed more on feminine items and females conformed more on masculine items. On
average, there was no diﬂerence in conformity for neutral items. When the sex of the
"majority response" was inchrded, the only diﬂererrce formd was that females conformed
more to female than male sources on masculine items. This ﬁnding may simply indicate
that on rmfamiliar topics, females respond in a similar way to other females.

The above studies suggest that both males and females rely more on the opinions
of others when their sex is inconsistent with the topic under consideration. However, it is
not clear whose Opinion they are more likely to accept. Feldman— Summers et al (1980)
found that individuals conform to the majority whose sex is consistent with the task, while
Sistnmk and McDavid (1971) found that individuals respond consistently with their own

sex. This complexity is further emphasized by research on individual inﬂuence. Given the

38
ﬁndings regarding conformity, it is also expected that sex diﬂerences in inﬂuence will

depend on the composition and task.

Inﬂuence. Similarly to research on conformity, several studies have assessed the
amount of inﬂuence males and females have under diﬂ’erent sex compositions. Three
studies examined the inﬂuence of male and female confederates who argued a position
opposite the rest of the group (Bradley, 1980, 1981; Thameling & Andrews, 1992).
Another study examined the eﬂ‘ect of confederates who attempted to inﬂuence group
members responses on an association task (Ridgeway, 1982). Finally, Craig and Sherif
(1986) manipulated the information given to certain members to study their inﬂuence on a
group ranking task. Each study is discussed in detail below.

First, Bradley (1980) studied the inﬂuence and reactions to male and female
confederate opinion deviates who expressed either high or low competence regarding the
task. All groups were composed of four males and either a male or female confederate.
Females who demonstrated low competence were less inﬂuential than high competence
females and low competence males. Low competence males however, were less
inﬂuential than males who demonstrated high competence. Statements directed toward
females with low demonstrated competence were more dominant and hostile than those
directed toward high competence females or low competence males. However, there were
no signiﬁcant diﬂ‘erences in the behaviors directed toward high competence males and
females. Finally, male opinion deviates were liked more than females opinion deviates
regardless of competence. These results likely reﬂect the fact that groups were composed

of mostly male members.

39
A second study by Bradley (1981) on the inﬂuence of opinion deviates used

groups composed of two males and two females phrs a male or female confederate. In
addition, confederates provided either support for their arguments, qualifying phrases
indicative of rmcertainty, or both. Female opinion deviates who did not use supporting
arguments were less inﬂuential than females who used support and their male counterparts
who did not support their arguments. In addition, females who did not use support were
rated as less intelligent and less knowledgeable than females who used support and males
who did not use support. Females who used qualifying phrases were also seen as less
intelligent and knowledgeable than females who did not use qualiﬁers and males who did.
Finally, male opinion deviates were liked better than female opinion deviates regardless of
the use of support or qualiﬁers.

A similar study by Thameling and Andrews (1992) examined the inﬂuence of male
and female opinion deviates who used qualiﬁers or evidence. As in Bradley's (1981)
study, the use of qualiﬁers was viewed as undermining perceptions of conﬁdence, while
evidence was expected to increase one's perceived credibility and inﬂuence. Confederate
opinion deviates had little inﬂuence regardless of their sex or the behaviors used.
However, statements directed toward female opinion deviates who spoke without
evidence or qualiﬁers were more emotional than those directed toward comparable males.
In addition, members were more cooperative with male deviates who spoke with or
without evidence and without qualiﬁers, than females who acted in a similar manner.

Thameling and Andrews' ﬁndings, along with those of Bradley (1980; 1981),
suggest a potential bias in the way group members respond to male and female opinion

deviates. It appears that members respond more favorably to male deviates regardless of

 

It

f0.

40
the way they present their arguments. Although Thameling and Andrews did not ﬁnd a

diﬂerence in inﬂuence between male and female deviates, both of Bradley's studies did.
Examining the content of the discussion topic provides a potential explanation. In the ﬁrst
study, groups discussed the culpability of alleged plagiarism in a journalism class (Bradley,
1980), while in the second study, the discussion topic involved the appropriateness of
moderate drinking of alcoholic beverages (Bradley, 1981). The topic in Thameling and
Andrews' study, however, was the issue of sexual assault on campus about which females
may have been seen as more knowledgeable. This may explain why males were not more
inﬂuential in the latter study.

In a slightly diﬂerent approach, Ridgeway (1982) manipulated the motivations
portrayed by a confederate who tried to inﬂuence the decisions of members on a word
association task Confederates presented themselves as either self- or group-oriented. It
was proposed that females who presented themselves as group-oriented would have more
success inﬂuencing members than those who were self-oriented. In male groups, female
confederates who were group-oriented achieved much higher inﬂuence than self-oriented
females, although females did not achieve as much inﬂuence as male confederates. In
female groups, male confederates had more inﬂuence than group-oriented females, but
self-oriented females had the highest inﬂuence. Finally, there were no diﬂ‘erences in the
ratings of male and female confederates on their competence or inﬂuence. These ﬁndings
suggest that although males tend to have more inﬂuence, females can alter their behaviors
to gain more inﬂuence, particularly in predominantly male groups.

Finally, in Craig and Sherifs (1986) study, groups ranked items needed for survival

following a plane crash in the desert. In each group, one member, either a male or female,

41
was given a critical piece of information (i.e., an item of high rank) to be used to

determine their level of inﬂuence. Males had more inﬂuence than females only in single-
male groups. Females were more inﬂuential in balanced-sex than in single-male groups,
and males were slightly more inﬂuential in single-male than single-female groups. In
addition, the target item was ranked signiﬁcantly higher in all-male than in the balanced-
sex groups, and slightly higher than in the single-female groups. Overall, men were rated
higher than women on leadership, inﬂuence and aggressiveness.

Craig and Sherif ( 1986) suggest that one reason males may have had less inﬂuence
in the single-female groups is that they were distracted by competition for the female's
attention, commonly referred to as the "rooster eﬂect." The reason men did better in the
single-male condition could be attributed to the nature of the task. The task was rated by
participants as male-oriented, although both males and females were equally interested in
the task. A single male in a group of females may have more inﬂuence when the task is
seen as male-oriented, especially if he is also perceived as more knowledgeable about the
task.

My. Research on conformity and inﬂuence stress the importance not only of
composition, but also of the task, in aﬂecﬁng individual behaviors. Although females are
generally found to conform more and have less inﬂuence than males, this diﬂerence
appears to be limited to certain conditions. For example, males have been found to
conform more than females in female-majority groups and on female-related tasks
(Feldman-Summers et al, 1980; Tuddenham et al, 1958). Craig and Sherifs (1986)
study exempliﬁes the complexity of the relationship between sex composition and

individual behaviors. In their study, males were more inﬂuential in the single-male

 

0!

42
condition (Craig & Sherif, 1986), while other research suggests that males are more likely

to conform in the presence of a female majority (Tuddenham et al, 195 8). As stated
earlier, such inconsistencies are likely attributable to moderator variables. In addition to
the task, other potential moderators have been explored in the literature.

Factors Moderating Stereotypic Perceptions and Behaviors

Several studies have examined conditions which may aﬂect the extent to which sex
stereotypic behaviors are observed in groups. Nemeth et al (197 6) studied groups with
either balanced or skewed sex compositions who performed a jury decision making task in
which they either read or heard criminal cases. In groups that read the case, there was a
tendency for males to send and receive more commrmications than females. These
diﬂerences were observed primarily in statements reﬂecting agreement, suggestions and
opinions. When groups heard the cases, there were no signiﬁcant diﬂ’erences in the
frequency or content of communications between males and females. Nemeth et al
suggest that when information is readily available, as in the written condition, males may
be more dominant, but when information must be stored in memory, as in the audio
condition, everyone's contribution is considered important so that males and females
participate more equally. Therefore, differences in the task demands can inﬂuence the
amount of stereotypic behavior observed in mixed-sex groups.

Wood and Karten (1986) examined the extent to which behaviors are inﬂuenced by
the speciﬁcation of members’ status on an aptitude test. They studied the interactions of
mixed- sex groups who discussed what an individual should do about her roommate's
suspected drug use. The status of male and female members was either speciﬁed as high

or low, or unspeciﬁed. When status was unspeciﬁed, males spoke more than females,

43

engaged in more task-related behavior, and were perceived as more competent than
females. Females, however, engaged in more socio-emotional activity. When status was
speciﬁed, no diﬂerences were observed between male and female interactions. Instead,
the behaviors of high and low status members reﬂected the differences between males and
females formd in the unspeciﬁed status condition. This suggests that unless otherwise
speciﬁed, females are assumed to have lower status than males. In this study, this
assumption was shown to have a substantial inﬂuence of the behaviors of males and
females.

Experience is another variable that may override sex stereotypic assumptions.
Lockheed and Hall (1976) studied the behaviors of males and females in balanced and
same-sex groups. In balanced- sex groups, males were more likely to emerge as leaders,
initiate more task-related acts and have more inﬂuence than females. Among same- sex
groups, no diﬂ‘erences were found between the behaviors of males and females. Following
the ﬁrst part of the study, subjects from same-sex group were reassigned to mixed-sex
groups. Females who gained experience in all-female groups initiated higher task-related
acts when placed in mixed-sex groups compared to females whose experience was initially
in mixed- sex groups. These ﬁndings suggest that mixed-sex environments may inhibit the
behaviors of women unless they are previously given an opportunity to gain experience in
a same- sex environment.

Lam—mag

There are inconsistent ﬁndings regarding the eﬂect of composition on individual

behaviors. Some studies indicate that males and females act less stereotypically in mixed-

sex groups (Aries, 1976; Piliavin & Martin, 1978), while others suggest that stereotypical

44
behavior is more likely to be observed in mixed-sex environments (Lockheed & Hall,

1976; Wood & Karten, 1986). Although no deﬁnitive conclusions can be drawn about the
eﬂ’ects of sex composition, it is clear from this literature that the behaviors of men and
women do not necessarily reﬂect stereotypes about the sexes. There appear to be factors
that moderate the extent to which these stereotypic behaviors are observed Some studies
suggest factors (e.g., status, experience) that reduce the reliance on sex stereotypes
(Lockheed & Hall, 1976; Wood & Karten, 1987). Other studies suggest variables (e.g.,
the task) that actually may alter sex stereotypic behaviors (Nemeth et al, 1976;
Wentworth & Anderson, 1984). Overall, this literature suggests that complex processes
are involved in the relationship between composition and individual behaviors. The same
is true of group level eﬂ‘ects, which are discussed next.
Group Level Eﬂ‘ects

A limited number of studies have examined the eﬂ’ects of different sex
compositions on group performance. These studies include comparisons between all-male
and all-female groups as well as those between same-sex and rrrixed-sex groups. Studies
involving only same- sex groups constitute a large portion of this research and are
primarily interested in differences between males and females, rather than between
homogeneous and heterogeneous groups (Wood, 1987). Since the present study is
interested in the eﬂ’ects of group heterogeneity, the focus of the remaining review will be
research comparing mixed and same-sex groups.

Wood’s (1987) meta-analytic review of sex composition eﬂ‘ects inchrded 13
studies that compared mixed and same-sex groups. Her review was based upon

laboratory studies in which an objective measure of performance (e.g., accuracy,

45
creativity, time to completion, amount of motor activity) was obtained. In addition, she

coded for task type based on Steiner‘s taxonomy (Steiner, 1972), and whether
performance was dependent primarily on task or social activities. Unfortunately, not all of
the studies included in her review examined group level outcomes. Studies by Burtt
(1920), Swanson and Tjosvold (1979) and Timmons (1941) examined male-female
differences in mixed-sex groups and thus were addressed earlier. In addition, three studies
involved dyads (Kerr & Sullaway, 1983; Rosenthal, 197 8; Swanson & Tjosvold, 1979)
which potentially diﬂ‘er from larger groups. However, these studies are reviewed along
with those involving groups of three or more individuals, but with awareness of their
potential differences.

Wood's (1987) meta-analytic results suggested that overall, there is no signiﬁcant
diﬂerence between mixed- sex and same-sex groups in terms of performance. However,
due to the small number of studies, a more accurate conclusion from this review is that
there is substantial variance in ﬁndings across studies. Although characteristics of the task
were coded for by Wood, it was not possible to test for moderators due to the small
number of studies.

One potential explanation for the inconsistency across studies in Wood's review is
that there is variance among diﬂerent types of mixed- sex groups. For example, a group
with a single male may perform very diﬂerenﬂy from a group with equal numbers of males
and females. As discussed previously, distinctions can be made between " skewed,"
"balanced" and even "tilted" nrixed-sex compositions (Kanter, 1977; Martin & Shanahan,

1983). Skewed compositions may have very diﬂerent inﬂuences on groups than tilted or

46
balanced compositions. Further, as shown by the literature on individual level eﬂ‘ects,

male dominated groups may diﬂ‘er from female dominated groups.

Another possible explanation for the variance across studies reviewed by Wood
(1987) is that characteristics of the task other than the conﬁguration of members as
suggested by Steiner (1972) may interact with the speciﬁc composition to have an impact
on performance. As already seen in the individual literature, the task can have a signiﬁcant
inﬂuence on the results. Individual studies, including those not addressed in Wood's meta-
analysis, are reviewed below. Inconsistencies in ﬁndings across studies are identiﬁed
along with factors that may help to explain those inconsistencies.
Group Performepee

Research on group performance involves a number of diﬂ‘erent outcome variables
and has produced a variety of results. Three studies address the productivity of same-sex
versus mixed-sex groups. Kent and McGrath (1969) examined the eﬂ‘ect of composition
across diﬂerent measures of performance. Groups with members of the same-sex or a
male or female majority were studied on production, discussion and problem solving tasks.
For each task, several performance dimensions were measured (e. g., creativity, accuracy,
productivity). Sex composition was signiﬁcantly related to three of eight group product
dimensions. First, all-female and female majority groups generated more action oriented
products than male-dominated groups. Second, same- sex groups were more original than
mixed- sex groups. Third, all-female groups were more optimistic than all-male or mixed-
sex groups. In addition, the task type interacted with the sex composition for ﬁve

performance dimensions, but the nature of these interactions was not reported.

47
Hoﬂinan et al (1962) examined the extent to which the group's composition

aﬂ’ected the development of integrative, or creative, solutions in groups that role played
the Change of Work Procedure problem No diﬂ’erences were found between same and
mixed- sex groups, but all-female groups had a smaller proportion of integrative solutions
than either all-male or mixed-sex groups.

In a study by Kerr and Sullaway (1983) the amount of air pumped into a chamber
was used as an indicator of task motivation. Mixed-sex dyads pumped more air into the
chamber than same- sex dyads. However, same-sex dyads felt more conﬁdent than mixed-
sex dyads of winning a prize for pumping the most air. There was no signiﬁcant diﬂerence
between same- and mixed-sex dyads in members' interest in working with the same partner
again. These ﬁndings suggest that mixed-sex groups may have diﬂerent motivational
features than same-sex groups, thus increasing their productivity.

Based on the above three studies, it is difﬁcult to arrive at any conchrsion
regarding the eﬂ‘ect of composition on productivity. The ﬁndings range from female-
dominated groups having the highest productivity to having the lowest productivity.
Unfortrmately, not enough information is provided about the task used in Kent and
McGrath’s (1969) study to compare to Hoffman et al’s (1962) ﬁndings. The speciﬁc
tasks used by Kent and McGrath may have been more female-oriented than the Change of
Work Procedure task used by Hoﬂinan et al, thus explaining the variable success of all-
female groups in these two studies. Research addressing other performance measures
have similar limitations.

Several studies examined group diﬂ‘erences in the speed of task completion.

Eskilson and Wiley (1976) found no signiﬁcant diﬂ’erence in the speed of puzzle

 

110

ho]

48
completion across groups composed of all-male, all-female, single-male or single-female

members. However, there was a tendency for single-male groups with a female leader to
be faster than any of the other group compositions.

Rosenthal (197 8) studied mixed and same-sex dyads, and found diﬂerences in the
time taken to solve either a math or verbal problem On the math task, all-male dyads
solved problems faster than mixed-sex dyads who, in turn, were faster than all-female
dyads. On the verbal task, mixed-sex dyads were faster than either all-male or all-female
dyads. These ﬁndings suggest that diﬂerences in completion times across groups may
vary as a ﬁrnction of the task Assuming that sex diﬂerences are associated with math and
verbal tasks, this study provides support for the inﬂuence of task sex-orientation.

South (1927) examined both the speed and accuracy of all-male, all-female and
balanced groups on four different tasks. Only minimal information was provided about
these tasks, and no signiﬁcance tests were reported. Therefore, only general conclusions
are presented here. On two tasks, same-sex groups were faster while on the other two
mixed-sex groups were faster. All-female groups were more accurate on a task which
involved interpreting emotions portrayed in photographs and a task which involved
judging English compositions. On these two tasks, balanced-sex groups and all-male
groups were equally accurate. All-male groups were better on a task involving multiple
choice problems which the authors considered more abstract. Balanced-sex groups were
the least accurate on this task On the fourth task, called the Bridge Problem, there were
no diﬂerences in the quality of problem solving across groups. This study further shows
that the eﬂ’ect of the composition on speed may depend on the task, and that the same

holds true for accuracy.

49
Clement and Schiereck (1973) studied the accuracy of same-sex and balanced-sex

groups who worked on a visual signal detection task The seating arrangements of group
members was also manipulated such that male and female members were either alternated
around a table, or were seated next to someone of the same sex Balanced-sex groups in
which same- sex members sat adjacent to one another had signiﬁcantly lower detection
accuracy than same-sex groups Balanced-sex groups with adjacent seating, however, did
not diﬂer signiﬁcantly from groups with alternative seating. In addition, there were no
diﬂerences in accuracy between all-female and all-male groups. These results suggest one
potential explanation for differences in accuracy between mixed and same-sex groups.
That is, coalitions may have formed among same-sex pairs in the adjacent seating
condition which may have hindered effective group performance.

Finally, three studies were identiﬁed that examined the quality of decision making
across diﬂerent compositions. First, although sex composition was not the focus on their
study, Hoﬂinan and Maier (1961) formd that groups containing one or more females
produced higher quality solutions than all-male groups across two tasks. One task
involved developing a method for permitting ﬁve men to cross a heavily-mined road (the
Mined Road problem) and the other was the Change in Work Procedure problem
discussed previously. The focus of the study was the comparison of groups that were
homogeneous or heterogeneous in terms of personality. In the Mined Road problem
diﬂ‘erences were only found for all-male groups, with groups heterogeneous in personality
performing better. The opposite was true for the Change of Work Procedure problem,

where diﬂerences occurred only among youps with one or more females. These ﬁndings

50
suggest that relationships among other variables may be inﬂuenced by the group’s

composition and the task.

A more recent study by Rogelberg and Rumery (1994) compared same, balanced,
and skewed-sex groups. F our-person groups worked on a survival task (the Desert Plane
Crash exercise). Groups were compared on the quality of their decisions (assessed using
expert rankings), the time spent on the task, and group cohesion. Single-female groups
had the highest quality decisions. Balanced-sex and all-male groups had the next highest
performance, followed by single-male and all-female groups. No signiﬁcant diﬂerences
were obtained for time spent on task or interpersonal cohesion across these compositions.

The authors suggested that the task was male oriented because male individuals
performed signiﬁcantly better than females. However, the all-male teams did not have the
highest performance, which the authors attributed to some type of "process loss. " The
male-dominated groups with a single female were able to outperform their all-male
cohorts, but it is unclear what processes may have contributed to this result. Interestingly,
this ﬁnding is inconsistent with an earlier study which suggested that males in single-
female groups tend to be less eﬂ‘ective because they are distracted by competition for the
attention of the female (e.g., Craig & Sherif; 1986).

Finally, a ﬁeld study was identiﬁed that examined the eﬂ'ect of composition on the
quality of group decisions. Robertson and Kwong (1994) studied the eﬂ’ect of group
heterogeneity on the decision making eﬂ’ectiveness of school leadership councils. They
found a marginally signiﬁcant eﬂ‘ect for sex composition, indicating that as the number of

men in the group increased, decision malo'ng eﬂ’ectiveness as measured by self-reports,

5 1
decreased This study ﬁrrther highlights the inconsistencies in ﬁndings regarding the eﬂ‘ect

of composition on performance.

Overall, the literature on group performance suggests that same-sex groups are
sometimes more creative, productive, accurate, and faster than mixed-sex groups (e.g.,
Kent & McGrath, 1969; South, 1927), and others times they are not (e. g., Kerr &
Sullaway, 1983 ). Some studies showed that this relationship depended on the dominant
sex of the group (e.g., Hoﬂinan et al, 1962) or the nature of the task (e.g., Hoﬂinan &
Maier, 1961; South, 1927). Studies addressing decision quality exemplify this complexity,
with some ﬁnding that male-majority groups made better decisions (Rogelberg & Rumery,
1994), and others ﬁnding that female-dominated groups performed better (Robertson &
Kwong, 1994).

Group Proce§_se_s

Research focusing on group processes exhibits similar inconsistencies. In terms of
participation, Patterson and Schaeﬂ‘er (197 7 ) found no diﬂ‘erences between same and
balanced-sex groups in the duration of their interactions or rates of participation.
However, diﬂerences were formd in their interaction distance. All-male groups interacted
most distantly, all-female groups most closely, and balanced-sex groups at an intermediate
distance.

Skvoretz (1988) similarly failed to ﬁnd diﬂerences in participation equality across
groups. He developed mathematical models to test the eﬂects of sex composition on the
equality of participation. He found that in no groups was participation distributed equally,
but that these distributions were not explained well by diﬂerences in composition.

However, there was a positive correlation between sex and participation indicating that

52
men participated more than women and suggests that the model may not have adequately

captured the eﬂects of sex composition on participation. Another study by Johnson and
Schulman (1989) also failed to ﬁnd differences in participation equality for both task and
socio-emotional activity. These ﬁndings seem inconsistent with the diﬂ’erences found at
the individual level between the participation rates of males and females.

Two studies examined differences between groups in terms of the speciﬁc content
of the communications. Piliavin and Martin ( 1978) studied the interaction content of
groups who discussed various social problems. All-female groups had the highest
percentage of socio-emotional acts and all-male groups the highest percentage of task-
related acts, as deﬁned by Bales' (1970) Interaction Process Analysis. The percentages of
task and socio-emotional acts in balanced-sex groups fell in between those of all-male and
all-female groups.

Mabry (1985) compared the interactions of all-male, all-female, male-majority and
female-majority groups on two types of tasks. The ﬁrst task, considered low structured,
involved discussing a human relations case. The second task, considered high structured,
involved rank ordering behaviors according to the level of interpersonal competence they
were perceived to represent. Interactions were again coded according to Bales' ( 1970)
Interaction Process Analysis. A signiﬁcant interaction of task and composition was found
for the category " gives suggestions." All-male groups working on the structured task
gave signiﬁcantly more suggestions than all other groups, with the exception of male-
majority groups. Male-majority groups working on the unstructured task gave the lowest
number of suggestions. Mabry suggested that this ﬁnding might be attributable to the

reference to a woman in the unstructured task which may have led groups to rely more on

53
the single female member, thus reducing the number of suggestions given by males in the

male-majority groups.

In addition to the interaction of task and composition, there were main eﬂects for
composition (Mabry, 1985). All-female groups had signiﬁcantly fewer disagreements than
all other compositions. All-male groups had fewer disagreements than male-majority
groups, but still more than all-female groups. Finally, no signiﬁcant diﬂerences were
found across groups in the amormt of task-related interactions. This study shows that
diﬂ’erences may be found when communications are categorized by content area, and that
these diﬂerences may vary as a frmction of the task.

S_um_ma_ry

The above ﬁndings highlight the complexity of the relationship between sex
composition and group processes and performance. Subtle distinctions in group
composition (e.g., all-male versus single-female) seem to produce very diﬂerent outcomes.
In addition, diﬂ‘erences have been found within and across studies for diﬂerent tasks (e. g,
Hoﬂinan & Maier, 1961; Kent & McGrath, 1969; Rosenthal, 1978; South, 1927). The
exact nature of the tasks eﬂ‘ect, however, is unclear. The sex-orientation of the task has
been discussed by some (e. g., Hoﬁnan & Maier, 1961; Rogelberg & Rumery, 1994) as
having an inﬂuence on these ﬁndings. Speciﬁcally, some tasks may be seen as more
appropriate or associated more with one sex than the other. As a result, group members
may respond in accordance with stereotypes about the competencies of men and women

on diﬂ’erent tasks. Literature addressing the inﬂuence of the task is reviewed next.

 

011

54
Role of the Trik

Many tasks are associated primarily with members of one sex or the other, which
leads to expectations about the appropriate sex to perform a task (Ruble, Cohen, & Ruble,
1984). Men may be assumed to be more competent and put forth more eﬂ‘ort on tasks
traditionally performed by men (e.g., auto repair), while women are expected to be more
competent and more motivated on tasks traditionally performed by women (e.g., child
care). However, an increasing number of men and women are entering ﬁelds traditionally
associated with the opposite sex. This creates an incongruency between stereotypes about
the individual and stereotypes about the task As a result, ”non-traditional" members may
be ignored by or isolated from the majority since they are assumed to have less task-
related competency. In addition, certain members may have higher status in the group
simply because their sex has been traditionally associated with the task (Ridgeway, 1982).

There is a fair amormt of evidence that the type of task interacts with sex
composition to inﬂuence group performance (Martin & Shanahan, 1983). As suggested
by Deaux (1984), “Many observed sex diﬂ’erences are not durable main eﬂ’ects, but rather
are inﬂuenced by task characteristics, resulting in frequent interactions between sex of
subject and sex-linkage of task" (p.107). Unfortrmately, with the exception of the
literature on individual conformity, the sex-linkage of the task has generally been
neglected by researchers. Further, most of the research on the eﬂ‘ects of sex composition
have been studied using masculine, or traditionally male, tasks (Eagly & Carli, 1981).
Therefore, the majority of ﬁndings may not generalize beyond male-oriented contexts.

Several studies suggest that the sex-linkage of the task can have signiﬁcant eﬂ‘ects

on individual perceptions and behaviors. Dames-Farrell, L'Heureux-Barrett, and Conway

55
(1991) found that task performance was evaluated more accurately when the task

_ behaviors were congruent with the sex-type of the job. In other words, behaviors on
male-typed tasks were evaluated more accurately in the context of male-typed
occupations, and female-typed tasks evaluated more accurately in female-typed
occupations.

Mabry (1985) also suggested that the reference to a women in a task description
may aﬂect behaviors in the group. She found that men were less dominant in nrixed-sex
groups only on the task that referenced a women. As a result of this reference, males may
have conceded to the expectation that female members were more knowledgeable about
the task.

Ward (1991) directly tested the notion of the sex-appropriateness of tasks.
Subjects were asked to make judgments about the occupational suitability of male and
female applicants. They found that sex-congruent applicants were rated signiﬁcantly more
suitable than sex-incongruent applicants.

The expectations of one's own behavior may also depend on the task. Karabenick,
Sweeney, and Penrose (1983) found that women expected their performance to be higher
on feminine tasks while men expected higher performance on masculine tasks. Similarly,
Baucom and Banker-Brown (1984) found that women who scored low on masculinity
were more likely to give up and waste time on male-stereotyped tasks.

As the work force becomes increasingly diverse, not only are groups becoming
increasingly heterogeneous in terms of sex, they are doing so in a way that conﬂicts with
traditional stereotypes about the roles of men and women. It is clear, however, that the

sex-linkage of the task has an aﬂ’ect on both the expectations and behaviors of men and

 

Ill

56
women. Throughout the review of sex composition, the task was frequently cited as a

potential factor explaining inconsistencies within and across studies. Given that many
relationships in this literature are rmstable, additional variables need to be considered that
may inﬂuence the results. There is reasonable evidence that the sex-linkage of the task
plays such a role. Therefore, it should prove valuable in understanding the eﬂects of the
group's composition on team processes and performance.
Conclusion

There are strong and pervasive stereotypes regarding the appropriate
characteristics of males and females. These stereotypes aﬂ‘ect performance attributions
(Deaux & Emswiller, 1974; Wiley & Eskilson, 1983), the perceived status and inﬂuence of
men and women (Eagly & Wood, 1982; Wiley & Eskilson, 1982; Wharton & Baron,
1987), and help maintain sex segregation in the workplace through the discouragement of
sex-inappropriate behavior (Costrich et al, 1975).

Research on the perceptions of males and females has generally found that males
are more often seen as leaders (Greene et al, 1981; Nemeth et al, 1976; Webber, 1976),
and are evaluated more favorably than females on a number of dimensions, including
intelligence and inﬂuence (Butler & Geis, 1990; Toder, 1980). These eﬂ‘ects, however,
are not consistent across all situations. Similarly, individual behaviors sometimes reﬂect
sex stereotypes and other times do not.

These behaviors and perceptions are often dependent on the sex composition of
one’s reference group. Composition can inﬂuence the salience of sex as a distinguishing
characteristic of members in the group (Kanter, 1977; McGuire etal, 1979). Certain

representations may increase stereotypic perceptions of males and females. There is

 

ta;

57
substantial evidence that the sex composition of a group aﬂ’ects individual as well as group

level behaviors. The nature of this inﬂuence, however, is not always clear.

In terms of individual behaviors, some studies have found males and females to act
more stereotypically in same-sex groups (Aries, 1976; Mabry, 1989; Piliavin & Martin,
197 8), while others have shown that individuals behave more stereotypically in mixed-sex
groups (Johnson & Schulman, 1989; Smith-Lovin & Brody, 1989). For example, several
studies found a general bias in favor of males such that males were more inﬂuential
regardless of composition (Bradley, 1980, 1981; Thameling & Andrews, 1992).
However, in another study males were only more inﬂuential than females when they were
the minority member in the group (Craig & Sherif; 1986).

At the group level, there is also much variance in ﬁndings across studies that
compare the performance of mixed and same- sex groups. Both mixed-sex and same-sex
groups have been found to be more creative, accurate, productive, and faster (e.g.,
Clement & Schiereck, 1973; Hoﬂrnan et al, 1962). Diﬂ’erences are also found among
various types of mixed-sex compositions (e. g., male-majority, female-majority). However,
no consistent trend has emerged in the types of compositions that perform best (Kent &
McGrath, 1969; Homnan & Maier, 1961; Rogelberg & Rumery, 1994; Robertson &
Kwong, 1994). Similar inconsistencies are found when examining group processes such
as comnnmication content (Mabry, 1985; Piliavin & Martin, 197 8) and participation
equality (Johnson & Schulman, 1989; Skvoretz, 1988).

Attempts to explain these inconsistencies have identiﬁed several potential
moderators of the effects of sex composition. These include the sex-orientation of the

task (W entworth & Anderson, 1984), the speciﬁcation of status (Wood & Karten, 1986),

58
experience (Lockheed & Hall, 1976), and the proportion of males and females in the

group (Alexander & Thoits, 1985; Spangler et al, 1978). For example, men and women
may only act stereotypically when the task favors men (W entworth & Anderson, 1984),
when the status of members is unknown (Wood & Karten, 1986), or when the
representation of females in the group is relatively small (Alexander & Thoits, 1985).

Of these variables, the task seems to be most promising in explaining the majority
of inconsistencies found in the literature. Expectations about one's own competency and
the competency of others may depend on the congruency between the sex-linkage of the
task and the sex of the individual (Karabemck et al, 1983; Ward, 1991). Research on
individual conformity, in particular, exempliﬁes the inﬂuence of the sex-relatedness of the
task (Feldman- Summers et al, 1980; Sistrunk & McDavid, 1971). In general, individuals
conform more to majority inﬂuence, especially when the sex of the majority is consistent
with the sex-linkage of the task

To summarize, three main conclusions can be drawn ﬁ'om the literature on sex
composition. First, individuals often think and behave according to sex stereotypes. For
example, men are often perceived to be more task-oriented and women more socio-
emotional (e. g., Butler & Geis, 1990; Heilman et al, 1989). Second, these behaviors and
perceptions are inﬂuenced by the sex composition of groups. At the individual level,
diﬂ’erences in participation rates and conformity between males and females have been
found to vary depending on the group’s composition (e.g., Feldman-Summers et al, 1980;
Mabry, 1989; Piliavin & Martin, 1978; Reitan & Shaw, 1964). At the group level,
composition aﬂects the content of communications and performance (e. g., Mabry, 1985;

Robertson & Kwong, 1994). Finally, there are numerous inconsistencies within and

59
across studies. Rather than attempt to resolve these inconsistencies, it appears that many

researchers have moved on to other topics. Therefore, the question remains as to the
source of these discrepancies. This study attempts to address this question by examining a
potential moderator of sex composition eﬂects. There is evidence that the sex-linkage of
the task can inﬂuence individual perceptions and behaviors (e.g., Karabenick et al., 1983),

and potentially moderate the eﬂects of sex composition.

SEX COMPOSITION AND TEAM DECISION MAKING

The review of prior research suggests three issues that should be considered when
studying the effects of sex composition on teams. First, given the wide range of groups
that can be studied, there is a need to set boundary conditions on the type of groups to be
addressed. Second, it is important to identify what variables should be studied given what
is known and not known from the existing literature. Finally, there is a need to focus on
an aspect of the task which oﬂ‘ers the most promise of gaining a better understanding of
sex composition eﬂects.

Decision Making Teams

Research on group decision making has a long and diverse history, examining
issues like group size, group polarization and ﬂee discussion (Davis, 1992). The focus of
much of this literature is on consensus decision making in which members typically share
common information and are expected to reach agreement about a course of action. In
many teams, members do not share a common knowledge base, but instead are assigned to
various areas of expertise. In such groups, agreement is often not feasible or even
desirable. Therefore, many groups have a leader who is responsrhle for making the
decision on behalf of the group. Since many work groups do not ﬁt the model of
consensus decision making, there is a need to explore alternative group structures.

Therefore, this study will focus on teams in which knowledge is distributed and one

60

6 1
member is primarily responsible for making the group’s decision. Such groups will be

referred to as hierarchical decision making teams.
Identiﬁcation of Relevant Grogp Processes

In addition to the limited focus on certain types of decision making, there has also
been insufﬁcient attention to group processes in the team decision making literature. As
shown by the previous review of sex composition eﬂ‘ects, individual behaviors have been
more widely considered than group behaviors. The more frequently studied group
behaviors are participation rates and comnnmication content. The importance of these
behaviors to eﬂ‘ective group decision making is rarely addressed Researchers have shown
that certain group processes, such as information sharing and coordination, are
signiﬁcantly related to performance (e.g., Stasser, Taylor, & Hanna, 1989). Therefore,
any inﬂuence that the group’s composition has on these variables may ultimately affect
group performance. For example, mixed-sex compositions may inhibit eﬂective
communication among group members which may reduce the amount of information that
is shared. Groups will be less likely to arrive at a proper decision if information is not
shared.

There are numerous variables which may be inﬂuenced by a group’s composition,
but it is important to focus attention on those which aﬂ‘ect performance and are relevant to
the teams being studied. This study will examine the eﬂ‘ects of group sex composition on
several group processes associated with eﬂ’ective decision making in hierarchical teams.

Task Eﬂ‘ects
Prior research has found that the task appears to inﬂuence the eﬂects of sex

composition on individual and group level behaviors. The tasks used, with some

62
exception (e.g., individual conformity), seem to be chosen with little consideration of their

potential inﬂuence. Of the studies involving multiple tasks (e.g., Hoﬁnan & Maier, 1961;
Toder, 1980), there is little attention to the characteristics of these tasks that might have
contributed to the ﬁndings. Although it is important to examine task characteristics such
as complexity, abstractness, and structure, the sex-linkage of the task is a characteristic
that is uniquely important in this literature. Many researchers have already suggested that
the male or female orientation of a task may explain inconsistent ﬁndings within and across
studies. However, only research on conformity has speciﬁcally tested the eﬂects of task
sex-linkage. Others have used tasks which may reﬂect this variable (e.g., math versus
verbal tasks), but these tasks may diﬂ’er on other characteristics such as structure or
complexity, thus confounding the results. In order to test the effect of the task’s sex-
linkage, one need only change the content from male to female-oriented (e.g., Butler &
Geis, 1990). Examining the sex-linkage of the task will hopefully provide an eiquanation
for the inconsistent results found in the literature.
Hierarchical Team Deeinﬁm MaLn’ng

One of the advantages of focusing on hierarchical team decision making is that
both a theory and a task have been developed to address these types of teams.
Hierarchical decision making teams have certain characteristics that distinguish them from
consensus decision making teams. First, they are composed of a leader and staﬂ members
who together are responsible for making team decisions. The staﬂ members are
responsible for evaluating information and making recommendations based on that
information. The leader is primarily responsible for incorporating staﬂ member

recommendations into a team decision.

63
Second, knowledge regarding the relevant information is distributed within the

team such that members have different expertise. Expertise reﬂects the portion of
available information that a member is primarily responsible for evaluating. All
information available to the team, however, is relevant to the decision, and, therefore, each
member‘s contribution is important for the team to perform eﬂecﬁvely. For example, one
member's information may be critical to the decision making if that information leads to a
diﬂerent conclusion than the rest of the team The failure of one member to contribute his
or her knowledge to the leader may result in a poor team decision.

Third, these teams are characterized by interdependence. That is, members rely on
one another for obtaining some of the information relevant to their expertise. Access to
the pool of information available to the team varies such that members can directly obtain
some, but not all, of the information. Members are responsible for gathering and sharing
information with one another prior to reaching a decision. Therefore, interaction among
team members is needed to make eﬂ‘ective team decisions.

The Multilevel Theogg

A theory was developed which attempts to predict and explain the eﬂ‘ectiveness of
decision making in teams with the above characteristics (Hollenbeck, Ilgen, Sego,
Hedhmd, Major, & Phillips, 1995). This theory requires that a criterion exists for
assessing the accuracy of each decision, and that this criterion reﬂects some integration of
the information on which the decision was made. The multilevel theory then identiﬁes
factors that are expected to directly inﬂuence decision accuracy, referred to as the core
constructs of the multilevel theory (see Figure 1). These core constructs are based on the

deﬁning characteristics of hierarchical decision making teams. First, it is necessary that

64
staﬂ‘ members obtain all of the information relevant to their expertise. Therefore, the ﬁrst

core construct, team infom' , reﬂects the extent to which the team as a whole is
informed on all the relevant information. Second, it is important that staﬂ’ members
interpret the information accurately according to their expertise and supply the leader with
useful recommendations. The variable which captures this effectiveness is staﬂvah ' , or
the extent to which members’ recommendations, on average, predict the correct team
decision. The third core construct addresses the hierarchical nature of the teams. Leaders
are responsible for making the ﬁnal decision for the team, and, therefore, can aﬂect the
accuracy of decisions in the way they incorporate staﬂ members’ recommendations. For
example, a leader may perceive a staff member as less competent than the other members,
and give less weight to that member’s recommendation. Depending on the validity of that
member’s recommendations, the leader’s weight may or may not have been appropriate.
The third core construct, hierarchical senm' ' ' , assesses how appropriately the leader
weighs staﬂ members’ recommendations in arriving at the team decision.

Each of these team-level constructs is an aggregate of a variable at a lower level of
analysis. These lower level constructs often help to explain phenomena at the team level.
Decision infoM’ is the amount of relevant information the team as a whole has obtained
for a particular decision object. Individual valim is the predictive validity of any one
staﬂ member’s recommendations. Finally, dyadic seam ' ' is the appropriateness of the

weight given by the leader to any one staﬂ‘ member’s recommendations.

.wﬁxwﬁ 5302. :83 30.32303? E85 $5338 2E. g

 

 

 

 

 

 

 

 

66
The multilevel theory proposes that all other variables (e.g., task characteristics,

cognitive ability) aﬂ’ect team decision accuracy through the three core constructs. In other
words, the core constructs are viewed as mediators. As shown in Figure 1, these other
variables are categorized into six sets of constructs based on a ﬁamework developed by
McGrath (1976). The theory further suggests that certain categories are likely to
inﬂuence some core constructs more than others. Research testing the multilevel theory
has repeatedly found all three core constructs to be signiﬁcantly related to decision
accuracy, and in combination to account for between 27% and 43% of the variance in
team decision accuracy (Hedlund, 1993; Hollenbeck et al, 1995). The role of the core
constructs as mediators of the relationship between variables commonly addressed in the
group decision making literature (e.g., cohesion, familiarity) and team decision accuracy
has also been supported. For instance, a study on computer-mediated comrmmication
found that the eﬂ‘ect of the communication medium on team decision accuracy was
completely mediated by the core constructs (Hedhrnd, 1993 ). That is, the comrmmication
medium only had an indirect eﬂ’ect on team decision accuracy through its inﬂuence on the
core constructs. The isolation of the three core constructs improves knowledge of the
process by which teams arrive at their decisions and allows for the identiﬁcation of
diﬂ‘erent and possibly conﬂicting ways in which they do so.
The TIDE2 Sinml_eﬁpr_r

The Team Interactive Decision Exercise for Teams Incorporating Distributed
Expertise (TIDEZ) task was developed to study hierarchical decision making teams
(Hollenbeck, Sego, Ilgen, Major, Hedlund, & Phillips, in press). TIDE2 is a computer

simulation designed to study decision making contexts involving rmrltiple cues which are

67
often distributed among members of a team Members commrmicate through nWorked

computers which virtually record every keystroke, message, and decision made by every
member. A team can consist of as many as four individuals, and up to nine decision
making criteria (i. e., cues) can be assigned to the team Decisions can be compared to a
“true score” which is based on the method employed for combining the cues into a
decision.

The TIDE2 program allows for ﬂexibility in terms of the content area about which
decisions are made. For example, a selection task could be sinnrlated in which members
are asked to consider the experience, education, and ability test scores of applicants.
Diﬂ‘erent members could be assigned the responsibility for evaluating each piece of
information. Further, these cues could be weighed diﬂ‘erently in making the decision such
that experience received more weight than test scores. The TIDE2 task not only provides
an ideal tool for studying the types of teams of interest, it is also a favorable mechanism
for manipulating the sex-linkage of the task

Hypotheses

The following hypotheses address the effects of sex composition and task sex-
lirrkage on team decision making. The compositions of interest include all-male, male-
majority, female-majority, and all-female membership. The sex-linkage of the task reﬂects
the issue about which a decision is made, and is either male or female-linked. The female
version involves a nursing task in which decision are made about infants. The male
version is a military task in which decisions are made about aircraft. The teams are all

hierarchically structured with expertise distributed among staff members.

68
The multilevel theory of hierarchical team decision making is used as a framework

for predicting the eﬂ’ects of sex composition and the task sex-linkage on team decision
making (see Figure 2). In addition to the three core constructs identiﬁed by the multilevel
theory, participation equality is also viewed as a critical group process. For the task used
in this study, each member has an equally important role in the team Decisions are based
on all of the information available to a team, and therefore, eﬂ‘ective decision making
depends on the contributions of all team members. Equal participation can help to insure
that all members obtain the information they need to make good recommendations, and
that members have an equal inﬂuence on the ﬁnal team decision. Participation equality is
also expected to be highly inﬂuenced by the sex composition of a team, and thus is a
relevant variable to this study.

As shown in Figure 2, sex composition is expected to aﬂ'ect team decision
accuracy through its eﬂ’ects on the three core constructs and participation equality. The
sex-linkage of the task is viewed as a moderator of the relationship between sex
composition and the decision making processes. The core constructs and participation
equality are expected to directly inﬂuence team decision accuracy.

Predictions of the nnrltilevel theory are used along with the literature on the effects
of sex composition to develop speciﬁc hypotheses regarding the eﬂ'ects of sex
composition and the task sex-linkage on the team decision making process. Since there is
a greater abundance of knowledge regarding individual level eﬂects of composition, these
ﬁndings are used to form hypotheses regarding group level eﬂects. The hypotheses,
however, only address relationships among group level variables as this is the level to

which conclusions about sex composition eﬂ‘ects will be drawn. The variables identiﬁed

69

.waﬂaa nommmoov Ewen 30393.53 no owaxqaéom #3 can 836388 x3. we macaw... 38:55 «Muelzmm

 

 

 

5839‘s wee—e2
eoEooQ Each

 

 

 

esemsm assess:
bG=m> bﬁm
eases as
eonmaﬁubenﬁ mo bans—om

 

 

 

 

 

:oEmomEoU xom

 

 

 

ewasﬁaem
ssh

 

 

 

70
by the multilevel theory (i. e., team informity, staﬂ’ validity, hierarchical sensitivity) have

not been directly addressed in the sex composition literature. Therefore, attempts are
made to draw on knowledge of related constructs in developing the hypotheses.
Particip ation Eguelﬂ'

At the group level, researchers have failed to ﬁnd differences in participation
equality as a fimction of sex composition. However, individual participation rates do vary
as a frmction of the composition, suggesting that it is reasonable to expect diﬂ'erences at
the group level In mixed-sex groups, males are often found to participate at a higher rate
than females (Aries, 1976; Johnson & Schulman, 1989). There is also evidence that
females participate more than males when the group has a female majority (Mabry, 1989;
Thune et al, 1980). Therefore, it is reasonable to expect that both male and female-
majority teams will experience unequal participation.

Differences in participation equality between male and female-majority
teams will likely emerge as a function of the sex-linkage of the task. That is, on the male
task, male-majority members will be more likely to dominate the single female member
than they would on the female task. Similarly, a female-majority is expected to dominate
the group’s interaction more on the female than the male task. As illustrated in Figure 3,
it is hypothesized that:

Hypothesis la: In general, same-sex (homogeneous) teams will participate more
equally than mixed- sex (heterogeneous) teams.

Hypothesis 1b: Within mixed-sex teams, male-majority teams will participate less
equally than female-majority teams on the male task, and female-majority teams will

participate less equally than male-majority teams on the female task.

71

$5260 333693 no owes—dzéom me v5 598988 xommo Scuba conic—ﬂoat»: .m nag

x93. 038$ 673,—. 032

menu:

5
.i’
:13
E

 

 

i
i

wounding—.032

seeseoaaee

     

. /
Beeeeeeeaasom/ /

samba “0986mm
hams unwound

 

 

 

72
Team Infamy

The most closely related variable to team informity in the sex composition
literature is the amount of task-related activity engaged in by a team At the group level,
research ﬁndings are inconclusive regarding between group diﬂ‘erences in task-related
activity. Some studies found between-group diﬂ‘erences in task-related activity (Piliavin &
Martin, 197 8), while others found these differences to depend on task characteristics
(Mabry, 1985). At the individual level, males have a tendency to engage in more task-
related activity in same-sex and balanced-sex groups (Johnson & Schulman, 1989; Thune
et al, 1980), while females are more task-oriented in female-majority groups (Mabry,
1989; Thune et al, 1980). Together, the group and individual ﬁndings suggest that all-
male groups should have the highest level of task-related activity followed by mixed-sex
and all-female teams.

The amount of task-related activity alone, however, does not insure that
information is being shared. High team informity also requires that every member receives
the information that he or she needs. Therefore, participation equality will also likely
inﬂuence team informity. Same-sex teams are expected to participate more equally than
mixed- sex teams, and thus should distribute information more effectively. Combined with
the ﬁndings regarding task-related activity, this suggests that, overall, all-male teams
should have the highest team informities. The extent to which differences emerge among
all-female and mixed-sex teams will likely depend on the task.

The sex-linkage of the task has been shown to aﬂ‘ect perceptions of task-related
competency (Karabenick et al, 1983; Ward, 1991). For example, female members are

generally perceived as less knowledgeable about male-oriented tasks. These perceptions

73
may inﬂuence the distribution of information to certain members, as well as attempts made

by members to obtain information. This failure to provide all members with the
information they need will ultimately reduce a team’s overall informity.

It is expected that the task will inﬂuence the distribution of information to
single male and single female members. Speciﬁcally, on the male task, male-majority
members will fail to share information with the single female member, thus reducing their
overall team informity. Similarly, in female-majority teams, single males are not expected
to obtain all relevant information when working on the female task. Teams are not
expected to be negatively aﬂ’ected when the sex-linkage of the task is consistent with the
sex of the single male or female. Therefore, on the female task, the team informity of
male-majority teams should not diﬂ’er signiﬁcantly from all-male teams. The sex-linkage
of the task is also expected to aﬂ‘ect all-female teams such that they are not expected to
have signiﬁcantly lower informity than all-male teams when working on the female task.
As illustrated in Figure 4, it is hypothesized that:

Hypothesis 2: On the male task, all-male teams will have higher team informity
than all-female and female-majority teams, with male-majority teams having the lowest
team informity. On the female task, all-male, all-female, and male-majority teams will
have higher team informity than female-majority teams.

Staﬂ‘ Validity

Staﬂ’validity reﬂects the ability of members to make useful recommendations
based on the information they have obtained The literature on individual level
performance within groups is most informative regarding the eﬂects of sex composition on

staﬂ‘ validity. From the limited research on individual performance, there appear to be

74

DEF—cud— anu do 036:5-me me Ea notion—doe xom we mwoobo 39:85.69»: .v earn

013. 235m

 

Bessemeaaea

Beeaeeaaz

 

mos-om

[immoral meal

xmﬂ. 2&2

 

Eugene—«2

Baaséaaﬁa

   

KIM-10m weer

 

75
complex relationships between sex composition and sex diﬂ’erences in performance. In

general, males seems to perform better than females under minority representation, while
females fare better than males under majority representation (Alexander & Thoits, 1985).
These ﬁndings suggest that the performance of both males and females should be
enhanced in female-majority groups, which should transfer to higher staff validity at the
group level

Team informity is also relevant to staﬂ validity since being well informed increases
the chance of making good recommendations. Therefore, the expected eﬂ’ects on team
informity are considered, along with ﬁndings regarding male and female performance, to
make predictions regarding staﬂ‘ validity. On the male task, the eﬂ’ects of sex composition
on staff validity are expected to be similar to those for team informity, with the exception
that female-majority teams should also beneﬁt ﬁ'om the enhanced performance of male and
female members. The advantage of female-majority teams is also expected on the female
task, which should reduce the diﬂerence between male and female-majority teams that was
predicted for team informity. Thus, as shown in Figure 5, it is hypothesized that:

Hypothesis 3: On the male task, all-male and female-majority teams will

have higher staﬂvahdiﬁes than all-female and male-majority teams. On the

female task, teams will not diﬂ‘er in terms of staﬂ‘ validity.
Hierarchical Sensrt_rv_rty° ' '

Hierarchical sensitivity is a variable that is particular to the types of teams of
interest in this study. Eﬂective team performance is dependent on the leader’s ability to
appropriately weigh staﬂ‘ members’ recommendations according to their validities. Those

weights, however, may be inﬂuenced by other factors, like a member’s sex, which may not

76

ﬁnial/ﬂea no owmgkom x33 can 8385580 x8 we Scuba woﬂmoﬁogm gm

 

 

 

Vina. 0389a x3e 032

g magnum mobﬁm widow
3Q.—
m S
A m
w W.
,M. ,M.

Bondage—«Sum
wBaEEeQbEE . Bananas—«Sum
emf

 

 

 

77
be a reliable indicator of ability. Although there is no research that is directly relevant to

understanding the eﬂ‘ect of sex composition on hierarchical sensitivity, the literature on
conformity and inﬂuence provide some direction regarding this eﬂ‘ect.

Research on conformity suggests that both males and females conform more in
mixed than same-sex groups (Reitan & Shaw, 1964; Tuddenham et al, 1958). The more
interesting ﬁnding from this literature is that conformity depends on the sex-relatedness of
the task Speciﬁcally, males conform more on male-related tasks while females conform
more on female-related ones (F eldman- Summers et al, 1980; Sistrunk & McDavid, 1971).
In regards to inﬂuence, males tend to have more inﬂuence than females, but this appears to
be limited to male-oriented tasks (Bradley, 1980, 1981; Craig & Sherif, 1986). These
ﬁndings suggest that both the sex-linkage of the task and the sex of the individual may be
considered when evaluating the validity of a member’s recommendation. For example, if a
task requires knowledge of how to sew a button hole, a single female’s opinion will likely
receive more weight than that of any male member.

Research on conformity and inﬂuence suggests that the weights leaders assign to
diﬂerent members will likely be inﬂuenced by the sex of the member and the sex-linkage of
the task At the group level, this implies that leaders of rnixed- sex teams will weigh
members’ recommendations less appropriately than leaders of same- sex teams since their
evaluations are more likely to be clouded by irrelevant information (i e., sex). Leaders of
male-majority teams are expected to underweigh the recommendations of single female
members on the male task, and to overweigh their recommendations on the female task,
thus reducing their overall hierarchical sensitivity. It is further expected that the act of

underweighing recommendations will be greater than the act of overweighing

78

recommendations, and thus, have a more negative eﬂ'ect on sensitivity. Leaders of
female-majority groups are similarly expected to weigh the recommendations of single
male members inappropriately, thus lowering their hierarchical sensitivity. As shown in
Figure 6, it is hypothesized that:

Hypothesis 4a: In general, same-sex teams will have better hierarchical sensitivity
than mixed-sex teams.

Hypothesis 4b: Within mixed-sex teams, male-majority teams will have poorer
hierarchical sensitivity than female-majority teams on the male task. Female-majority
teams will have poorer hierarchical sensitivity than male-majority teams on the female
task.

ImDecisimr Making Accurag

Research examining the eﬂ‘ect of sex composition on group decision making does
not provide any clear direction regarding the eﬂ‘ect of composition on team decision
accuracy. In some cases female-dominated groups make better decisions (Robertson &
Kwong, 1994), while in other cases male-majority groups perform better (Rogelberg &
Rumery, 1994). Similar inconsistencies exist for other outcome variables such as
creativity and productivity. Although the sex composition literature provides little
direction, the predictions of the multilevel theory of hierarchical team decision making
(Hollenbeck et al, 1995) can be used to generate hypotheses regarding the eﬂ‘ects of sex
composition and task sex-linkage on team decision accuracy.

The multilevel theory proposes that the core constructs will mediate the eﬂects of
sex composition and the task sex-linkage on team decision accuracy. Therefore, the

previous four hypotheses can be used to speculate on the nature of these effects. Same-

79

53238 30.39883 no owme—kom xmﬁ can :oEmoQEeo x8 mo Boobs voﬁmoﬁegm .c pawn

 

  
 

e62. 235m
9883 g
Eugene—nae..—
. / 33
/
/ / m.
Bananas: W.
W
emf

 

 

    

ass. as:
g man-em
Begonia:
33
W.
m
/ .m...
“Begonia—«Soul m.
. l
S
w
W.
.3:

 

80
sex teams were generally predicted to participate more equally, have higher team

informity, staﬂ‘ validity, and hierarchical sensitivity than mixed- sex teams which should
result in higher decision accuracy. The exception was all-female teams who were
hypothesized to have lower team informity and staﬁ validity when working on the male
task. It was also proposed that male-majority teams would be lower than female-majority
teams on all four decision processes when working on the male task. Female-majority
teams were expected to be lower than male-majority teams on these same decision
processes when working on the female task, with the exception of staff validity.
Therefore, male-majority teams should have lower decision accuracy than female-majority
teams on the male task, while the opposite is expected on the female task. Based on these
predictions, illustrated in Figure 7, it is hypothesized that:

Hypothesis 5: On the male task, all-male teams will have the highest decision
accuracy, followed by all-female and female-majority teams. Male-majority teams will
have the lowest decision accuracy. On the female task, all-male and all-female teams will
have the highest team decision accuracy, followed by male-majority teams. Female-

majority teams will have the lowest decision accuracy.

81

Scab—eon "86¢on 83“ no owaxnzkom «mg 28 summon—Eco xom me 300%? 3585..ng .5 oclzmmm

 

  

Mmﬂ. 225m

moss: 85m
Baaacnaaaom 33

/
/
l
/ m
/

m

Beganaaz w

m

m

m

 

 

Em

as... 232

meet:

 

Eugenia:

      

I
I
I
Baaaﬁaaaum

bemobv uorsrooq (noel

 

E

E

METHOD

26an
This study involves a 2 (group homogeneity) x 2 (male/female dominance) x 2
(sex-linkage of task) between-subjects design as shown in Table 1. Participants were
assigned to teams to create one of four sex compositions (i. e., all-male, all-female, male-

majority, female-majority), and randomly assigned to one of two sex-linked tasks.

Parngants

A total of 348 participants were recruited ﬁ'om undergraduate management and
psychology classes to form 87 four-person teams. They received course credit for their
participation. Information was obtained regarding demographic characteristics prior to
participation (e. g., age, sex, race, major, status in school) which was used to control, as
much as possible, for other types of heterogeneity among team members. The sample
characteristics were: sex: 47% males, 53% females; race: 81% Caucasian, 23% Aﬁican
American, 5% Asian and 6% other; status in school: 5% sophomore, 55% junior, 35%
seniors, and 5% other; major: 58% business, 15% social science, 7% human ecology, 5%
communications, 14% other; age: 66% ages 19 to 21, 22% ages 22 to 25, 7% age 26 and
over, 5% undisclosed. Attempts were made to balance the number of cases in each of the
8 cells of the design (see Table l), but due to characteristics of the student population and

high absenteeism rates, some cells have fewer cases.

82

83

Table 1
Study Design and Smle Distribution

 

 

 

 

Homtmeneity
Dominance Homgeneous Heterogeneous
Male task
Male All-Male Male-Majority
(9) (11)
Female All-Female Female-Majority
(11) (12)
Female task
Male All-Male Male-Majority
(9) (11)
Female All-Female Female-Majority

(12) (12)

 

Mm’ ulations

Sex Composition

Teams were composed of members who were all-male, all-female, one male and
three females, or one female and three males, fulﬁlling each cell of the homogeneity x
dominance design. No balanced-sex teams were included because the intent of the study
was to assess the effects of lmequal representations of males and females which best
reﬂect the reality of traditionally male and female-dominated ﬁelds. In all conditions, the
leader was a member of the majority sex since leadership by a minority member raises
additional issues which cannot be adequately addressed within the realm of this study.
Task Sex-Linkage

The task used in this study was the TIDE2 sinmlation (Hollenbeck et al, in press)
discussed earlier. The basic structure of this task involves four-person teams who are
responsible for evaluating and making decisions based on nine pieces of information. This
information can reﬂect any topic of interest (e. g., deciding where to go to college).
Therefore, almost any decision making situation can be sinnllated using the TIDE2 task.
For this study, task sex-linkage was manipulated by creating two situations which reﬂect
traditionally male and female-dominated ﬁelds. The male-linked task involved a military
command-and-control exercise while the female-linked task involved neonatal nursing
care. The task is ﬁrst described in detail based on the male version. The female version,
which only involved changes to the task‘s content, is then described.

The male version involved a naval command and control exercise in which team
members were responsible for monitoring aircraﬁ targets. These aircraft ranged in their

level of threat from ﬁiendly to hostile. The team's objective was to evaluate the aireraﬁ's

85

level of threat on a number of dimensions and to decide how to respond to the aircraft.
There were nine pieces of information, or cues, to be used in determining the threat level
of an aircraft. In addition, there were three rules for combining these nine cues. These
rules were used to make one of seven decisions, ranging from ignoring the aircraft to
defending against it (see Appendix A).

Each member was assigned to a speciﬁc role in the team These were the
commanding oﬂicer of the Coastal Air Defense (CAD), a reconnaissance aircraft
(AWAC S), a Cruiser, and a Carrier who served as the team’s leader. The three
subordinate members were each responsible for two of the three rules, one of which was
their primary expertise. For example, the AWACS was responsible for the motion rule and
the location rule, which was also the responsibility of the CAD. This provided some
overlap in expertise which allowed the team to potentially ignore certain members and still
perform reasonably well Since each rule involved three cues, each member was
responsible for obtaining six pieces of information and making an evaluation based on that
information Only members who were responsible for a rule knew how to interpret the
three cues associated with that rule. However, each member initially had access to only
some of those cues; the other cues had to be obtained ﬁom another team member, thus
creating interdependence within the team.

Once each member evaluated his or her information, a recommendation was sent
to the leader. The leader's speciﬁc role was to combine these recommendations, which
were in the form of decision alternatives, into a ﬁnal decision for the team The leader
also had access to some of the cues, as well as knowledge of which members had access

to what cues, and therefore, could assist in the dissemination of information. The leader,

86

however, did not know how to interpret any of the nine cues, and, therefore, relied on the
recommendations of his or her staﬂ‘. Once a decision was made by the leader, the team
received feedback regarding the accuracy of its decision.

In terms of the mechanics of the task, teams responded to a series of aircraft that
appeared sequentially on their computer screens. Each member had a menu ﬁom which to
select one of several ﬁmctions. These included "measure," "query," "transmit," "text" and
"receive." Measure enabled members to obtain information available to them, query
allowed them to ask for information, transmit was used to directly send information, text
was an option to send typewritten messages, and receive allowed members to read any
communications sent to them. There was also a menu for sending in recommendations, or
a decision in the case of the leader.

For the female version of the simulation, the mechanics and requirements of the
task were held constant. That is, members were responsible for evaluating sets of
information based on rules to which they were assigned, and they used a menu to measure,
query, transmit and receive messages. The decision task involved a nursing team working
in a neonatal ward. Instead of aircraft, teams were responsible for monitoring the severity
of health problems of newborns. As with the male version, nine pieces of information and
three rules were used to make decisions regarding the treatment of the infant. The
information was adapted to reﬂect cues and decision alternatives appropriate to neonatal
care (e. g., respiration, heart rate). Each member was assigned a speciﬁc role. The three
staﬂ‘ nurses each specialized in a different area of medicine (e.g., metabolic vs.

respiratory), and the leader was referred to as the nursing supervisor. Instructions for the

nursing task are displayed in Appendix B.

87

Procedure

Participants signed up for a particular time to report to the experiment. Upon
arrival, they were asked to sign in and read and sign a consent form. They were then
assigned to a team and brought into a room with four computers. Each person was
handed a set of instructions to read that described the general purpose of the task, and the
role of the team (see Appendices A and B). Speciﬁc role instructions were also provided
that included the information each member needed to perform the task (see Appendices A
and B). At this point, participants were asked to introduce themselves by stating their
name and role in the team They were asked to enter the names of the other team
members next to their assigned role on a form at each station. These forms helped to
identify each participant while members worked on the computer. Participants were then
given 10 minutes to read over the materials. At the end of the 10 minutes, the research
assistant answered any questions.

Next, participants were trained on how to use the computers to perform the task.
This training involved a 10 minute trial in which a research assistant, using a script,
pointed out the different components of the computer screen and talked participants
through the use of all the ﬁmctions in their menu bars (see Appendix C). Following the
training trial, participants were given two more trials on which to practice. At this point,
the research assistant answered any further questions and made sure everyone understood
how to use the computer. Throughout the training and the simulation, participants were
asked not to talk to one another so that all their communications would be captured by the

computer. To ensure they did not talk, teams were monitored by an intercom

88
At the end of the third practice trial, the simulation was paused and participants

were instructed to take 5 more minutes to read over their instructions. After 5 minutes,
the general task instructions were taken ﬁom participants and they were asked to turn
their speciﬁc role instruction sheets face down. At this time, they were given a
questionnaire that tapped their task knowledge, reactions to the previous training, and
performance expectations for use in research unrelated to this study. Once these measures
were collected, participants were again allowed to refer to their role instruction sheets.
The simulation was started and teams were left to work on their own. Once the team has
completed all trials in the simulation, they were asked to complete a questionnaire as part
of an unrelated study, and then were debriefed and dismissed.

Measures
Parti ' ation E '

Participation equality was measured as the standard deviation of total
communications (queries, transmits, messages) sent by members within the team A log
transformation was performed to normalize the distribution of these standard deviations
before using them in the analyses.

Team Infor_m_rty'

Team informity involved measuring for each trial the proportion of information
obtained by all staﬂ‘ members relative to the information needed by those members to
perform their speciﬁc roles. For example, if CAD needed range and corridor status,
AWACS direction and angle, and Cruiser [FF and radar (see Appendix A), but each only
obtained one of the two pieces of information, then as a team they only obtained 50% of

the necessary information and thus received a score of .50 on decision informity. These

89

decision informity scores were then averaged across all trials to obtain a measure of team
informity.
Staff Vali '

Staﬂ‘ validity was operationalized as the average correlation between staﬂ‘
members’ recommendations and the correct team decision (true score). Each member
performed 33 trials, responding to either 33 aircraft or 33 infants. For each trial, every
member made a recommendation regarding the decision object. Failure to reach a
decision in time was considered a “no call” decision which was equated with the least
aggressive decision in the respective task (“ignore” or “release”). The decisions were

‘61”

scored on an ordinal scale ﬁom for the least aggressive to “‘7” for the most aggressive
action (see Table 2). Each decision object had a predetermined correct value based on the
values assigned to informational cues and the rules for combining those cues. The
recommendations of each staff member were correlated with the true scores over all trials
to produce individual validity scores. At the end of the session, three such validities (one
for each staff member) were calculated, and then these individual validities were averaged
to obtain a measure of team staﬂ‘ validity.
Hierarchical Sensitrm '

Hierarchical sensitivity was determined by comparing two regression equations.
One involved regressing the correct decision on the staﬂ‘ members’ recommendations, the
other regressing the leader’s decision on those recommendations. The regression weights

were then compared across these two equations to assess the appropriateness of the

weights assigned to each member by the leader. Hierarchical sensitivity was then

90

 

v

N

382—0 \ vacuum 6
38:52 \ nee—es .o
382:. \ $.qu .m
333:— \ Eu? .v
BEBE \ BEBE .m
”@338 \ Buses .N

883% \ 0.8an A

 

n

v

m

 

838w Sacco

neuavnoaaooou 533nm

 

 

 

3503‘ now—memon— 83,—. E809 8m x534 080330

N 2%...

91
measured by computing the average of the difference in beta weights across staff

members. A smaller value reﬂects better hierarchical sensitivity.
learn Decig’on Making Accuracy

Team decision accuracy was assessed by comparing the team's decision to the
correct decision for each trial The correct decision, or true score, was determined using
an equation which represented the set of rules on which decisions were based. These rules
describe weighting schemes for combining the cues into an overall evaluation. For
example, in the military task, the “movement rule” stated that an aircraft with high speed,
descending at a sharp angle and coming at a straight direction toward the carrier ship
should be considered highly threatening. The equation was applied to the values assigned
to the cues (e. g., speed of 700 mph) to determine the true score for each aircraft or infant.
The output of the equation was scaled from 1 to 7 to represent the seven decision
alternatives (see Table 2). Difference scores were computed between the team’s decision
and correct decision for each trial. The diﬂ‘erences were then averaged across all trials to
obtain an overall measure of team decision accuracy.
Sex Composition

Sex composition is treated as two variables in the analyses, one that represents the
homogeneity of the team and the other which represents whether the team is male or
female-dominated. This allows both differences between same and mixed-sex teams, as
well as differences between teams with male and female majorities to be addressed. Each
of these variables was contrast coded (.5, -.5) for the analysis based on Cohen and Cohen

(1983). That is, homogeneity was coded -.5 for same-sex teams and +5 for mixed-sex

92
teams. Teams were coded -.5 if they were dominated by male members and +5 if they

were female-dominated.

RESULTS

The means, standard deviations and correlations for all of the variables are
presented in Table 3. Some basic relationships among the variables are readily observed
from the correlation matrix Of the two sex composition variables, only dominance is
signiﬁcantly related to any of the decision making processes or performance. These
correlations suggest that, in general, male-dominated teams performed better on team
informity, staff validity and hierarchical sensitivity than female-dominated teams. The sex-
linkage of the task is signiﬁcantly correlated with participation equality and staff validity.
The three core variables from the multilevel theory are each signiﬁcantly related to team
decision accuracy, and also exhibit signiﬁcant relationships with one another.
Participation equality, however, is not related to any of these variables or team decision
accuracy.

Hierarchical moderated regression analyses were used because the goal was to
understand the relationship between sex composition and the decision making variables
rather than simply to identify mean differences between groups. Regression was also
useful for testing both the individual hypotheses and the overall theory. For each
dependent variable, the main effects for task sex-linkage, composition homogeneity, and
dominance were entered ﬁrst, followed by the two-way and three-way interactions. The

variables of homogeneity and dominance were contrast coded as described above. Task

93

94

Ac. V RSI. .mo. VRa... .2. V R...

.988 5:2. a 3.5352 053:; m3“ no 02.3 .532 < _.

 

 

-- :58. .38.- ..is..- S. 8. 8. 3.- a. as $838“ 838‘ SE .w
-- ta.- ......m~.- 8.- *8. S. 8. we. 25 5358... 3020355 5
-- 1.3m. 8.- :3.- 8.- EN.- 2. are gag-«em e
-- 2. gram.- 2.- 8. a. who gcoegﬁ .m
-- 8.- 8. time. me. 3m 3:38 geaeogm e
-- 3... 8. on :3 8&385 .m
-- S.- on 8... gﬁwoacm .N
-- on o2 ”mg-x8 as... A
a b e n v m N _ a. E ossa>

 

 

 

Sn? 8:325 neg gosadbooeooﬁ 28 332663 335% :32
m 03:.

95
sex-linkage was dummy coded with 1 representing the male-linked task and 2 for the

female task. The results of the hierarchical regression analyses are summarized in Table 4
and described in detail below.
Part'grp' ation Eggﬁy’

It was hypothesized that same-sex teams would participate more equally than
mixed- sex teams, and that male-majority teams working on the male task and female-
majority teams working on the female task would participate the least equally. The
analyses revealed a signiﬁcant main effect for task (A R2 = .11, p < .01) with the direction
of this effect indicating that teams participated more equally on the female than the male
task. There were also signiﬁcant interaction effects for task and homogeneity (A R2 = .04,
p < .05), and for dominance and homogeneity (A R2 = .06, p < .01). An interpretation of
the interaction of task and homogeneity revealed that same-sex teams participated more
equally than mixed-sex teams on the male task, but on the female task mixed-sex teams
participated more equally (see Figure 8). There was more variance in participation
equality across tasks for mixed than same-sex teams. A plot of the interaction of
dominance and homogeneity showed that although all-female teams participated more
equally than all-male teams, there was more equal participation in male-majority than
female-majority teams (see Figure 9). Comparing these ﬁndings to the hypothesized
relationships (see Figure 3), there were differences between same- sex as well as mixed- sex
teams in that all-female teams participated more equally than all-male teams. Further, the
task only moderated the inﬂuence of the team’s homogeneity not its dominance by male or

female members.

96

.~O.VQ..:.:.. .mo.V&..:.. .o~.V& ...

 

 

co. me. co. .32.. oo. .82.. Sumo. 2.8mm. No. 3.5mm. O x I x H b
ﬁo. mc. .35. 3.50. co. Sac. mo. «5&3 . 3.3.09. .........mm. D x m o
8. 8. 8. .8. 8. :8. 8. 2...... 8. 3.2. a x .. m
co. No. cc. inc. co. 339. cc. .132 . 3&5. 3.3.2 . I x .H v
8. 8. .3. .2. :3. :8. .2... .. .32. 8. .32. E. 85.388 m
.o. 8. 8. S. 8. .3. S. 8. 8. .1... E. 3.0838,... N
8. 8. 8. 8. .3. .3. 8. 8. ...:. .... .. E amiss may .
.6933“ 3.35:8 gm...» £88m... 5:560
.8302. Each. 30203.85 Maw :83. 3.8.53.3.—
Nm 4 w. .38. ”m 4 at .35 N... < “m .33 ”x 4 Hz Boa «x < E .35 05.9., .8308? new

 

 

 

 

hm"? 833...; Eocﬁmoﬂ a3...“ momlhﬁe maﬁa-owed $038305 we MESH—3m
.. 2.3

97

5:830 acumﬁomtwn no muoaowcaoa can owe—amkom xmﬁmc 530885 .w oar.

 

wagon was:
3:85
.d
m
.m
x8... 032 w.
u
m.
l m.
‘ ‘ ‘ .III III .M.
| ‘ ‘
x3... 3.35m
35

 

98

5:360 33363.3 no bmoqowcac: 28 3833.2. ac .8583:— d 84.3...

 

 

wobuum woﬁom
.38:

d

m”

3388.038»... .m
m.

U

H

b

I .ml.

Bagoabzz I l m.
.35

 

99
Team Infom'

Next, it was hypothesized that all-nnle teams would have higher team informity
than all other compositions, except for all-female and male-majority teams working on the
female task. Male-majority teams were expected to have the lowest informity on the male
task and female-majority teams the lowest informity on the female task A signiﬁcant main
effect for dominance (A R2 = .11, p < .01) and a signiﬁcant three-way interaction of
dominance, homogeneity, and task sex-linkage (A R2 = .05, p < .05) was found. Overall,
team dominated by males (i e., all-male and male-majority teams) had higher informity
than those dominated by females (i e., all-female and female-majority teams). This effect,
however, was dependent on the task and the team’s homogeneity (see Figure 10). On the
female task, male-dominated teams had higher team informity than female-dominated
teams, regardless of homogeneity. On the male task, however, all-male teams had higher
team informity than all-female teams, but the inclusion of a single female member in the
team resulted in a signiﬁcant decrease in informity such that male-majority teams were less
informed than all-male teams. At the same time, the inclusion of a single male in female-
dominated teams led to an increase in team informity in comparison to all-female teams.
These results are consistent with the hypothesized relationships (see Figure 4), with the
exception that all-female teams had lower team informity than predicted for the female
task.

Staff Vali '

It was proposed that all-male and female-majority teams would have higher staff

validity than all other teams, except for all-female teams working on the female task.

There was a marginally signiﬁcant eﬂ‘ect for the task (A R2 = .04, p < .10), and a

100

5.883 83. no away—nakom M93 38 568329.: .oogﬂaeae 3.8883 .2 owlsmmm

 

 

 

 

v.3. case". .3 252
meoaom waom wBBom meadow
I I
I I 33 85.85.03,. 33
Engages»... I
I I \ \_
I.
\ \
l l
w 3398.03»... w
W W
.M. «A...
Baggage;
am... am...

 

 

101
si ' cant effect for dominance (A R2 = .05, p < .05). The direction of these effects

indicated that teams generally had higher staﬂ‘validities when working on the male than
the female task. Further, male-dominated teams had higher staff validities than female-
dominated teams. None of the predicted interactions were signiﬁcant.
Hierarchical Scum
The hypothesis regarding hierarchical sensitivity suggested that, overall, same-sex
teams would have better hierarchical sensitivity than mixed-sex teams. Male-majority
teams were predicted to have the poorest sensitivity on the male task and female-majority
teams the poorest sensitivity on the female task There were marginally signiﬁcant eﬁ‘ects
for dominance (A R2 = .04, p < .10), and for the interaction of homogeneity and
dominance (A R2 = .04, p < .10). Male-dominated teams had better hierarchical sensitivity
than female-dominated teams. The dominance eﬂ‘ect, however, was greater in
homogeneous than heterogeneous teams. That is, all-male teams had better hierarchical
sensitivity than all-female teams, but there was little difference in the hierarchical
sensitivity of male-majority and female-majority teams (see Figure 11). Contrary to the
hypothesized eﬂ‘ects (see Figure 6), there were diﬂ‘erences between all-male and all-female
teams rather than diﬂ‘erences between male and female-majority teams, and there was no
apparent advantage for same-sex teams in terms of hierarchical sensitivity.
learn Decis_ion Accuracy
Based on Hypotheses 1 through 4, it was proposed that same-sex teams would
make more accurate decisions than mixed- sex teams. Male-majority teams were expected

to have the lowest team decision accuracy on the male task, and female-majority teams the

102

53238 Roma—230.3 no Emu—“omens: was cog—idea no 832283 .: earn

wobﬁm woﬁom

 

Baasﬁaaaom

.IIIIII-IIIIIIIIIIIIIIIIJ

Eugene—«2

 

>93

Mumsms mwmmm

@m

103
lowest on the female task. Neither of the composition factors or the task sex-linkage had

a signiﬁcant effect on team decision accuracy (see Table 4).
The Multlevel Theory

Hypothesis 5, which predicted composition and task effects on team decision
accuracy, was based on the assumption that the four decision making processes were
related to performance. Each core variable was entered into the regression equation
followed by their two-way interactions.3 The interactions were included based on previous
research that found some interactions among the core variables (e. g., staﬂ‘ validity x
hierarchical sensitivity) to account for additional variance in decision accuracy. Result of
the regression analysis showed that each of the decision making variables was signiﬁcantly
related to team decision accuracy and together with their interactions accounted for 52%
of the variance in decision accuracy (see Table 5). A signiﬁcant interaction eﬂ‘ect was
found for team informity and hierarchical sensitivity and was interpreted to indicate that
teams were most accurate when they had both high team informity and hierarchical
sensitivity (see Figure 12). Interestingly, these were also the two core variables most
affected by sex composition and the task The mediational role of the core constructs was
not examined since neither sex composition or task sex-linkage had a signiﬁcant eﬂ‘ect on

team decision accuracy.

104

Table 5
Regression of Team Decision Accuracy on the Core Decision Variables (N=87)

 

 

 

 

Step Independent Variable Totale A R2 p
1 Team Informity (TI) .21" .21M -.39
2 Staﬁ‘Validity (SV) .43** .22" -.54
3 Hierarchical Sensitivity (HS) .47M .04* -.68
4 TI x SV .48“ .01 -.33
5 TIst .51" .03* .52
6 SV x HS .52" .01 .49

 

*p<.05. **p<.01.

105

$8.303 .8662. :88 no 33333 30.30333 was bane: Emoﬂo 8:882: .2 oar...—

mumbi mambi

 

33

Koamoov norstoeq meal

 

£883 ash is

 

ﬁi

DISCUSSION

The purpose of this study was to examine the effects of changing conﬁgurations of
males and females on group processes and performance. Same-sex teams and teams with
a single male or single female were compared on several variables identiﬁed as relevant to
decision making eﬁ‘ectiveness in teams characterized as hierarchical with distributed
expertise. Numerous inconsistencies in the literature suggest that sex composition effects
are complex and may depend on the inﬂuence of other variables. The sex-linkage of the
task was examined as a potential moderator of the effects of sex composition based upon
several studies that found different composition eﬂ‘ects across tasks. In addition, changes
in composition are arguably more visible in ﬁelds traditionally dominated by one sex

The sex composition of a team affected all four decision making variables. The
dominance factor had a consistent and signiﬁcant eﬂ‘ect on team informity, staﬁ‘ validity,
and hierarchical sensitivity indicating that male-dominated teams ('L e., all-male and male-
majority teams) were generally higher than female-dominated teams (ie., all-female and
female-majority teams) on all three variables.

The dominant sex of the team also interacted with team homogeneity to affect
participation equality and hierarchical sensitivity. For participation equality, this
interaction indicated that in same-sex teams, all-female teams participated more equally

than all-male teams, but in mixed-sex teams the opposite occurred. That is, male-majority

106

107
teams participated more equally than female-majority teams The interaction of

dominance and homogeneity produced a different effect on hierarchical sensitivity in that
the male advantage occurred primarily in same-sex teams. All-male teams had better
hierarchical sensitivity than all-female teams, but there was little difference in the
hierarchical sensitivities of male-majority and female-majority teams. This effect seemed
to be attributable to the lower hierarchical sensitivity of male-majority compared to all-
male teams.

The task sex-linkage moderated the effects of composition for the variables of
participation equality and team informity. For participation equality, mixed-sex teams
participated more equally than same-sex teams when working on the female task, but less
equally when working on the male task. For team informity, there was a signiﬁcant three-
way interaction among the task, dominance, and homogeneity factors. On the female task,
teams dominated by males had higher team informities than those dominated by females
regardless of the team’s homogeneity. On the male task, however, male-majority teams
had substantially lower team informities compared with all-male teams, and female-
majority teams had slightly higher team informities compared with all-female teams.

The task also had unexpected direct effects on participation equality and staff
validity. The nature of these effects were such that team members generally participated
more equally on the female than the male task, but they generally had higher staﬂ‘validities
on the male than the female task.

Although the task and composition aﬁ‘ected the core decision processes, there
were no signiﬁcant effects on team decision making accuracy. Several possible reasons

for this will be explored. First, a more detailed discussion of the composition and task

108
effects on each of the four decision making variables is provided, including attempts to

identify some underlying causes of the observed effects.
Parti ' ation E u '

For participation equality, there was no general dominance eﬂ‘ect as there was with
the other three decision making variables. Instead, there were differences between male
and female-dominated teams which varied depending on the team’s homogeneity. All-
female teams participated more equally than all-male teams, but there was more equal
participation among male-majority than female-majority teams.

The observed difference between all-male and all-female teams was not predicted
although it was not inconsistent with previous ﬁndings in the literature. Some researchers
have found that male teams tend to exhibit more stable patterns of dominance than female
teams (e. g., Aries, 1976). That is, certain members emerge as dominant and maintain that
position throughout the team’s existence. Female team members are more" likely to share
positions of dominance. The sharing should result in a more equal distribution of
participation over the lifespan of a team. Based on existing knowledge, it is reasonable to
ﬁnd that all-female teams participated more equally than all-male teams.

The more puzzling ﬁnding is that this difference did not transfer to mixed- sex
teams. Male-majority teams were observed to participate more equally than female-
majority teams. The change in composition from all-female to female-majority reduced
participation equality, but the change from an all-male to a male-majority composition led
to more equal participation. This raises the question of how the inclusion of a single
member of the opposite sex affected the distribution of participation among team

members, and why did this effect diﬁ‘er for predominantly male and female teams.

109
Individual Particip ation Rates

The effect of sex corrrposition on participation equality was ﬁrrther explored by
examining the individual participation rates of team members within different
compositions. Speciﬁcally, the participation rates of males and females were compared
across compositions (see Figure 13). There was a signiﬁcant sex diﬂ‘erence in
participation rates for same-sex teams with males participating at a higher rate than
females. This sex difference did not consistently emerge across compositions. In mixed-
sex teams, there was a signiﬁcant interaction effect for sex and dominance which indicated
that males participated more than females in female-majority teams but not in male-
majority teams. One possible explanation for the higher participation of males in female-
majority teams is that the single male may have assumed a dominant position and the
majority females more submissive positions in the team. The differences in male and
female participation rates for across compositions appears to explain the differences in
participation equality observed at the team level Male-majority teams experienced more
equal participation than female-majority teams because males and females differed less in
their participation rates.

Task Eﬂ‘ect

In addition to the composition effects on participation equality, there were
signiﬁcant effects involving the sex-linkage of the task. There was a signiﬁcant main
effect for the task indicating that teams participated more equally on the female than the
male task. This effect was larger in mixed-sex than same-sex teams as reﬂected by the
signiﬁcant interaction of homogeneity and task. Mixed- sex teams participated more

equally on the female task but less equally on the male task than same-sex teams.

110

 

mac—mag E 332 D

 

88-83 “Mao—£2 xomoﬁmm

 

02

Inc:

[rom—

[-02

[-03

[:02

 

 

[:02

 

 

 

o:

 

 

 

0M2

detaches E meadow was 838 Sm maﬁa nomﬁamomﬁuq 1322?: g

uorssos rad 310V

1 l 1
One possible explanation for this ﬁnding is that the discrepancy between male and

female participation rates within mixed-sex teams was greater on the male than the female
task. The mean participation rates for males and female working on the male task are 177
and 166 acts per session respectively, and for the female task 139 and 134 respectively.
Although there appears to be a larger diﬂerence in participation rates between males and
females on the male task, this eﬂ‘ect was not signiﬁcant. However, there was a signiﬁcant
task eﬂ‘ect indicating that team members consistently participated at a higher rate on the
male than the female task, and a signiﬁcant effect for member sex which revealed that
males generally participated more than females.

Since the male-female diﬂerenﬁal in participation rates did not appear to readily
explain the observed interaction eﬂect of the task and homogeneity, it was necessary to
focus on understanding the inﬂuence of the task. The observed task eﬂ‘ect on participation
equality may be attributable to the lower participation rates of staﬂ‘ members when
working on the female as compared to the male task. The question is why did the task
have such an eﬂ‘ect on individual members’ behaviors. One suggestion is that the
diﬂerence in task orientation often attributed to males and females may transfer to tasks
which are diﬂ‘erentially associated with the sexes. prarticipants behave in a manner
consistent with the sex-linkage of the task, both males and females may appear more task-
oriented on male than female tasks. Since the majority of participation in this study can be
categorized as task-oriented, the higher participation rates on the male task may reﬂect a
higher task-orientation associated with the male than the female task.

Another possible explanation is that the male task motivated members to

participate more because it was more interesting than the female task. Whatever the

l 12
reason is for the lower participation levels of individuals working on the female task, it

likely reduced the amount of variance in staﬂ‘ members’ participation, which resulted in the
more equal participation observed at the team level. Such an eﬂ‘ect may be greater in
mixed-sex teams because they have more potential to vary in their distn’bution of
participation, which would provide a possible explanation for the observed interaction of
the task and homogeneity.

Team Infom'

Two signiﬁcant eﬂ‘ects were observed for team informity: a main eﬂ’ect for
dominance and a three-way interaction of dominance, homogeneity, and the task.
Potential explanations for the dominance eﬂect are discussed, followed by an examination
of the interaction eﬂ‘ect.

Dominance Eﬂ‘ect

Teams dominated by male members generally had higher team informity than those
dominated by female members. Possible explanations for this eﬂ‘ect may exist at diﬂ‘erent
levels of analysis. At the team level, certain compositions may foster a more productive
environment and increase the amormt of information obtained by members. There may
also be individual diﬂ‘erences between males and females which leads teams dominated by
male members to outperform those dominated by females.

In terms of team level explanations, researchers have formd diﬂ‘erent compositions
to vary in their task-orientation (Piliavin & Martin, 197 8) and competitiveness (Aries,

197 6). Aries formd, for example, that all-male teams exhibited a sense of competitiveness
which was not observed in other compositions. In the present study, teams dominated by

male members may have been more competitive than those dominated by females,

l 13
resulting in greater eﬂ‘ort to obtain task-relevant information. A competitive attitude was

in fact observed in all-male teams in the form of verbal outbursts in reaction to successﬁrl
outcomes (i. e., “hits”), and not so successful outcomes (i. e., “disasters”). This was in
spite of the fact that members were instructed to communicate solely through typewritten
messages. Unfortunately, no comparable expression to these verbal outbursts appeared in
the typewritten messages, and, therefore, this observation could not be tested.

Although team level explanations seem reasonable, there remains the possibility
that the dominance eﬂ‘ect reﬂects an individual sex difference. If this is the case,
diﬂerences should exist in the amount of information acquired by males and females
regardless of the composition. An examination of the informity scores of individual team
members4 revealed a signiﬁcant sex diﬂ‘erence such that males were more informed than
females (M = .68 and .48 respectively, p > .01). The question becomes why did males
have higher informity scores than females?

It is widely believed that males are more task-oriented than females (Eagly, 1987).
One possible explanation is that males focused more of their time on task-relevant activity
and became more informed than females. There is research, however, that contradicts the
higher task-orientation of males and suggests that the diﬂerence in task-orientation
between males and females may depend on other factors (Johnson & Schulman, 1989;
Mabry, 1989). Both the composition and task are factors that have been suggested in the
literature. The inﬂuences of both variables are discussed in detail below, but one aspect of
the task deserves mention here. The general task, regardless of its sex-linkage, was a
computer sinnrlation. It can be argued that males are generally more interested and

familiar with computer games. Therefore, males may have been more adept at using the

1 14
computer to obtain information or may have simply enjoyed working on the task more

than females. This could have contributed to their higher informities.

An individual sex diﬂerence may ermlain the dominance effect on team informity,
but members could not attain high informity by acting independently. Some of the
information had to be received from other team members. This suggests that a
combination of individual and team level explanations may best characterize the results
obtained here. For example, males may initially be more interested in the task than
females, and by working with other males they become more motivated and involved in
the task.

Interaction of Task. Domi_n_ancmaM Homogenm'

The higher team informity of male-dominated teams did not hold across all
conditions. For the male task, male-majority teams had lower informity scores than all-
male and female-majority teams. In other words, the inclusion of a single female in a
predominantly male team seemed to suppress the general male advantage indicated by the
dominance eﬂ‘ect.

Two aspects of this effect deserve explanation. First, why did male-majority teams
experience a substantial decrement in team informity compared to all-male teams?
Second, why did this only occur on the male version of the task? One potential
explanation is that the single females in male-majority teams did not receive relevant
information from other members and lowered the team’s overall informity. The extent to
which this occurred could be tested by examining the informity scores of individual team

members within the diﬂerent composition and task conditions.

l 15
Figure 14 shows the individual informities of male versus female members by task

and composition. This data reveals that on the male task, males rather than females were
negatively aﬂected by the male-majority composition. Further, males working in male-
majority teams were less informed than the those working in all-male teams. This eﬂect
did not occur on the female task Instead, the individual informities of males were higher
in the male-majority than in the all-male condition. Majority males were not negatively
aﬂected by working with a single female on the female task as they were on the male task.
In addition, the single female members were more informed than females in any other
condition (see Figure 14).

One possible explanation for the diﬂ‘erences found for male-majority teams across
the two tasks is that a single female working on a male task stands out as incongruent and
draws more attention than a single female working on a female task. This is consistent
with the distraction theory, referred to by Craig and Sherif(1986) as the “rooster eﬂ‘ect,”
in which males shift their attention ﬂom the task to the female. In this study, the males
may have directed their attention to the female member rather than obtaining important
task-related information.

An alternative explanation is that the presence of a single female aﬂ‘ected the
motivation of team members. On the female task, the single female may be viewed as an
asset, raising the performance expectations and motivations of the team But on the male
task she may be viewed as a liability, lowering members’ expectations and motivations.

It is interesting to note that members of female-dominated teams also experienced
diﬂ‘erent levels of informity as a function of the composition and task In female-majority

teams, both male and female members had low individual informities when working on the

116

.me as 82838 3 noise ea menace 82833 33:5 3

SEE. oﬁaom Saab 232

 

 

 

season a as: U season n as: D

 

 

 

 

 

Wagon Hate—m2 xomroaam mag...— mgéiz x3685

  
  

nnnnnnnnn
aaaaaaaaaaa
oooooooooooo

 

...........
uuuuuuuuuuuu
.naa

 

 

 

 

\\\\\
O
V)

 

 

 

pouuogur moored

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

potmogur moored

1 17
female task On the male task, however, females had comparable informities to other

compositions, but the single male members were more informed than in all-male or male-
majority conditions (see Figure 14).

These ﬁndings suggest that male and female teams are similarly aﬂected by
changes in composition when it comes to informity. When the sex of the minority member
is inconsistent with the task, staﬂ‘ members become less informed. But when that
member’s sex is consistent with the task, staﬂ‘ members appear to respond more positively.
Further research is needed to identify whether these changes in behavior are attributable to
attentional or motivational factors. These behavioral changes do help explain the
interaction effect whereby male-majority teams experienced lower team informity than all-
male teams on the male but not the female task.

Staﬂ‘ Valim

For staﬂ‘ validity, which measures how well team members’ reconnnendations
predict the correct team decision, there were no diﬂerences between same and mixed- sex
compositions. Composition did affect staﬂ‘ validity in that teams dominated by male
members had higher staﬂvahdiﬁes than those dominated by females. As with team
informity, this dominance eﬂ‘ect may be attributable to individual as well as team level
sources. Although obtaining information was somewhat dependent on the actions of other
team members, evaluating that information was a fairly independent activity given the
task’s structure. It is reasonable to emect that the difference found between male and
female-dominated teams may reﬂect an underlying sex diﬂ‘erence. In fact, the mean
individual validities of males and females regardless of composition or task were .52 and

.45 respectively (p < .01).

1 18
Individuﬁal Sex Diﬂ‘erengg

A reasonable explanation for the higher validities of males is that they were better
able to process the information in the form it was presented The task can be
characterized as a multiple cue probability task in which members are required to interpret
numerical values according to decision rules Numerous researchers have proposed the
existence of sex diﬂerences in speciﬁc cognitive abilities. For example, Hyde, F ennema,
and Lamon ( 1990) suggest that males are better at problem solving than females.
However, it is impossible to conclude that the diﬂ‘erences observed in this task were due
to a sex diﬂ‘erence in cognitive ability without further research. A study of individuals
working independently on the same decision making task found no signiﬁcant diﬂerence
between males and females (M = .67 and .68 respectively), thus questioning the accuracy
of this explanation.

Another alternative explanation for the diﬂ‘erence observed between the validities
of males and females. Validity involved not only evaluating one’s information, but also
choosing a recommendation from seven decision alternatives. These decision alternatives
ranged on a continuum from least to most aggressive action. The least aggressive decision
reﬂected either ignoring the aircraﬂ or releasing the newborn. The most aggressive action
was to defend against the aircraft or to operate on the infant. A fairly well-established
diﬂerence between males and females has been found for aggressiveness. That is, males
are generally more aggressive and assertive than females (Maccoby & Jacklin, 1974). This
diﬂerence may have emerged in the use of decision alternatives such that males made more

aggressive decisions more oﬂen than females.

119

Unlike differences in ability, diﬂ‘erences in the use of decision alternatives can be
tested with the present data. The frequency with which males and females used each of
the seven decision alternatives were compared (see Figure 15). Females were more likely
than males to use alternatives 3 and 4 which reﬂect moderate courses of action (e. g.,
monitoring the infant’s health or aircraﬂ’s threat). The more aggressive alternatives (5
through 7) were more frequently used by males.

These diﬂ‘erences could have aﬂ‘ected validities in two primary ways. First, the
recommendations of females may have been less predictive of the correct decision simply
because they were less willing to use the more aggressive decision alternatives. Second,
the use of the moderate decision alternative can also be considered less risky since it
reduces one’s degree of inaccuracy (see Table 2). For example, in choosing alternative 4,
the most alternatives one can be ﬂom the correct decision is three. Ifone selects
alternative 6, it is possible to be ﬁve alternatives from the correct response. By choosing
less risky decision alternatives females may have reduced their inaccuracy, but at the same
time constricted the variance in their decisions and lowered their validities.

The data on use of decision alternatives provide evidence that males and females
diﬂ‘ered in a way that could have affected their validities. They do not, however, rule out
the possibility of team level inﬂuences. Diﬂ‘erences in riskiness or aggressiveness may be
fostered by team environments, particular those dominated by one sex or the other. Bauer
and Turner (1974) found that all-male teams were riskier in their decision making than all-
female teams. Therefore, characteristics of the team may have contributed to the

diﬂ‘erence observed between males and females in their use of decision alternatives.

.8380.“ was 838 new 838:8“? nommmoovmo om: a 36:2.on .2 baa

 

moaﬁom E 832 D _

 

333:3“? season

120

b o m

 

 

 

v

 

 

 

m

 

 

 

 

 

 

H

 

 

 

 

 

 

C

Hm
mm
mm

d aﬁeroAv

sotouonbor

121
Task Eﬂ‘ect

In addition to the dominance eﬂ’ect on staff validity, there was a direct task eﬂect
indicating that teams had lower staﬂvahdiﬁes on the female than the male task. This
ﬁnding suggests an unexpected task diﬂerence that aﬂected the recommendations made by
staﬂ members. As with the dominance effect, one potential explanation is that teams
working on the female task diﬂ‘ered from those working on the male task in their use of
the seven decision alternatives. Staﬂ members may have chosen less risky decision
alternatives when working on the female as opposed to the male task simply due to the
nature of the decision object. As discussed above, certain decision alternatives (e. g.,
operate, ignore) can be considered more risky because they increase the probability of
obtaining a disastrous outcome. Team members may have been less willing to risk disaster
when dealing with an infant than an aircraft. Similarly, some alternatives can be viewed as
more aggressive (e.g., operate, defend) than others. A military task may simply elicit more
aggressive responses than a nursing task. Diﬂ‘erences in the use of decision alternatives
between the male and female task could contribute to the observed task eﬂ‘ect on staﬂ‘
validity.

To assess whether such differences occurred, the frequency of using each of the
seven decision alternative were compared across tasks. As shown in Figure 16, members
appeared to make more conservative recommendations about twice as oﬂen as risky ones,
but there was no signiﬁcant diﬂerence between the two tasks in the use of risky or
aggressive decision alternatives. Therefore, task diﬂ‘erences in staff validity do not appear

to be attributable to diﬂ‘erences in the use of decision alternatives by staﬂ‘ members.

122

£me 0180“ was 038 05 8m 82358? 2.303% 3: E 838:”.on .2 “SEE

 

v68 225m E x23 23% ﬂ:

 

833:6“? commmooﬂ
n c w v m N

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

sorouanbor :1 aﬁeroAV

123

Since diﬂ‘erences did not occur in the use of decision alternatives, the only other
feasible explanation is that members were better able to make valid recommendations on
the male than the female task There is reason to suspect that the information on the male
task was easier to evaluate and combine into a decision recommendation than information
on the female task Information in the male task may have been more intuitive than in the
female task5 For example, in the male task, members were asked to combine speed, angle
and direction into a judgment, and in the female task they were asked to combine heart
rate, blood pressure and temperature. While the former closely resembles the way
information is interpreted by actual military teams, the latter does not accurately depict the
way nursing staﬂ‘s process information.

In addition, the way information is combined may have impeded members working
on the female task. In the military task, an aircraft is to be considered non-threatening if
any one characteristic in a combination rule (e. g., speed, angle or direction) is non-
threatening. The same rule was applied in the nursing task However, an infant’s health
may be considered in danger if any one of the characteristics poses a threat. A nurse
typically does not ignore an infant who has high temperature and blood pressure even with
a normal heart rate. Therefore, the judgments staﬂ‘ members made may have reﬂected
their own interpretations rather than the rules they were instructed to use.

This raises the question of whether there were diﬂ‘erences in familiarity with the
two tasks. Many of the computer and video games that students play involve some type
of conﬂict, including military combat. One would be hard pressed, however, to ﬁnd a
nursing game in an arcade or on a home computer. The general familiarity with military

games could explain the higher validity of members who worked on the male task in this

124
study. Their prior experiences may have led them to more easily interpret the military-

related data than the information about the health of newborns.

There are a number of feasible explanations for the dominance and task eﬂ‘ects on
staﬂ validity. Post hoc analyses suggested that the difference between male and female-
dominated teams is likely attributable to diﬂ‘erences in the decision alternative used by
males and females. Unfortunately, the data does not similarly account for the diﬂerence
between the male and female tasks. Prior experiences and expectations may be the most
logical explanation for the task’s eﬂ‘ect on staff validity.

Hierarchical Sensitrv_rty' '

The task had an eﬂ‘ect on all three variables previously discussed, but the effects on
hierarchical sensitivity were limited to the composition variables. There was a direct effect
for dominance which revealed that teams dominated by male members generally had
higher hierarchical sensitivities than those dominated by females. This eﬂ'ect was
dependent on the team’s homogeneity such that diﬂerences between male and female-
dominated teams were greater in same than mixed-sex compositions. In other words, all-
male and all-female teams diﬂered more in terms of hierarchical sensitivity than male-
majority and female-majority teams.

Maleand Female Lerﬁers

As with team informity and staﬂ validity, understanding the dominance eﬂ‘ect may
call for individual level explanations. Hierarchical sensitivity primarily captures the
activities of the team’s leader, that is, how eﬂecﬁvely he or she uses staﬂ members’

recommendations. It is possible that the dominance eﬂ’ect reﬂects a diﬂerence between

125

male and female leaders. This again raises the issue that a general sex diﬂ‘erence may have
inﬂuenced team level results.

There are any number of factors that could explain the diﬂ‘erences in hierarchical
sensitivity. Male and female leaders may diﬂer in their ability to attend to task-relevant
information or to identify which staﬂ‘ members are more reliable than others. They may
also differ in their decision making strategies. For example, female leaders may average
staﬂ‘ members’ recommendations while male leaders use a weighting scheme. While there
is no data that directly measures the use of task-relevant or irrelevant information, it is
possible to assess, to some extent, the use of different decision strategies.

If female leaders did use more of an averaging approach than male leaders there
should be less variability in the weights they assign to staff members’ recommendations.
An examination of this data reveals no signiﬁcant diﬂ‘erence in the standard deviation of
weights assigned by male and female leaders to their staﬂ members (M = .15 and .16
respectively). Both male and female leaders appear to vary in the weights assigned to staﬂ
members. The apparent diﬂerence between male and female leaders is not readily
explained. The interaction of dominance and homogeneity discussed next may provide
some insight.

The diﬂerence in hierarchical sensitivity between male and female-dominated teams
was less pronounced in mixed-sex teams. Male-majority teams had lower hierarchical
sensitivity than all-male teams, which reduced the diﬂ‘erence between male and female-
majority teams. A better understanding of this effect requires determining how the
sensitivities of male leaders were affected by the inclusion of a single female member in the

team

126
Madic Sengtrvity' ' '

One way to address the question of why male-majority teams had lower
hierarchical sensitivity than all-male teams was to examine the appropriateness of the
weights assigned by a leader to individual staﬂ members, which are referred to as dyadic
sensitivities. The dyadic sensitivities for male and female staff members were compared
across diﬂ‘erent compositions (see Figure 17). It was expected that leaders of male-
majority teams would underweigh the recommendations of the single female, thus
lowering their dyadic sensitivity. As shown in Figure 17, leaders of male-majority teams
actually weighed the male staﬂ‘ members less appropriately than the single female
members. More importantly, leaders weighed male staﬂ‘ members less appropriately in
male-majority than in all-male teams.

Although there was little difference in the hierarchical sensitivities of all-female and
female-majority teams, an exarrrination of the dyadic sensitivities reveals an interesting
eﬂ‘ect. Leaders weighed female staﬂ‘ members more appropriately in female-majority than
in all-female teams. However, they weighed the single male members far less
appropriately than the female members. These effects apparently canceled out one another
so that little difference emerged in the hierarchical sensitivity of female-majority and all-
female teams.

The dyadic sensitivity data suggests that both male and female leaders were
inﬂuenced by the sex of their staﬂ‘ members. The composition appears to have had a more
substantial and negative inﬂuence on the sensitivities of male leaders. As proposed earlier
with regard to team informity, a single female member may serve as a distraction to the

rest of the team A team’s composition may inﬂuence where its members focus their

127

doEmoaSoo 3 mac—«Bow was momma 8m 8:15:38 235 .: oﬂwmm

 

moEEom E 332 B

$8.83 Fad—«2 xombﬁwm

 

 

[rmud

 

.390

D\.
Val

 

Inwnd

 

rtmwd

[lewd

 

 

30.
V0

 

 

wwd

 

 

Nd >904

hmntwas

3E

128
attention and their ability to perform eﬂectively. For example, all-male teams may be

more task-focused than all-female or male-majority teams which assists the ability of
leaders to eﬂecﬁvely weigh staﬂ members’ recommendations. The difference between the
hierarchical sensitivity of male and female-dominated teams may be attributable to both
individual differences and team level inﬂuences. Identifying more speciﬁc explanations is a
task for future research.
Team Decision Accuracy

The eﬂ‘ects of the task and composition on the decision making variables only
address one aspect of a larger theory (see Figure 2). Speciﬁcally, these variables were
proposed to mediate the effects of the composition and task on the ultimate outcome of
the team’s performance, decision accuracy. Several conditions must be met in order to
test for mediation. First, the task and composition nnrst aﬂ‘ect the core decision making
variables which the previous discussion conﬁrmed Second, those variables must relate
signiﬁcantly to decision accuracy. These relationships were also conﬁrmed by the data,
with the exception of participation equality. That is, team informity, staﬂ‘ validity and
hierarchical sensitivity each contributed to a signiﬁcant portion of the variance in team
decision accuracy (see Table 5). Third, there must be a task or composition eﬂ‘ect to
mediate. This is where the theory is not supported Desp'ne support for the ﬁrst two
conditions, there were no signiﬁcant task or composition effects on decision accuracy.
There was nothing to be mediated by the core decision making variables.

The lack of a task or composition effect on team decision accuracy is puzzling
given the consistent dominance effect on the three core decision making variables.

Although male-dominated teams appeared to have an advantage over female-dominated

129

teams in the decision making process, their ﬁnal outcomes did not diﬂer. There are
several possible explanations for this discrepancy.

One possibility is that there was not enough statistical power to assess the
signiﬁcance of task or composition eﬂ‘ects on decision accuracy. The percent variance
accounted for by any of these factors or their interactions was no greater than 3%.
Therefore, regardless of signiﬁcance, there was no substantial relationship to be mediated
by the core variables.

The next question is why did the composition eﬂects on the core decision variables
fail to transfer to an eﬂect on decision accuracy? The most likely rationale is that female
dominated teams somehow compensated for their generally lower scores on team
informity, staﬂ‘ validity and hierarchical sensitivity. It is possible that these variables failed
to capture some process or processes that female-dominated teams were better at than
male-dominated teams. For example, leaders of female teams may acted to oﬂ‘set the poor
performance of their staff members.

Furthermore, other variables may inﬂuence decision accuracy without relating to
the core variables. Previous research on hierarchical decision making teams has found
some variables to have a direct eﬂect on team decision accuracy alter accounting for the
core decision making variables (Hollenbeck et al., 1995). One variable that had a direct
effect on decision accuracy was group cohesion. Thus, female-dominated teams may have
positively inﬂuenced decision accuracy by being more cohesive than male-dominated
teams. Other factors that female-dominated teams have been found to exhibit more than
male-dominated teams include optimism and agreement (Kent & McGrath, 1969; Mabry,

1985). In addition, female-dominated teams in the present study participated more equally

130
than male-dominated teams. Participation equality, however, was not a signiﬁcant

predictor of decision accuracy. Although it is reasonable to expect that female-dominated
teams may have outperformed male-dominated teams on variables like cohesion and
agreement, it is unlikely that such variables would have a large enough inﬂuence on
decision accuracy to offset the signiﬁcant diﬂ‘erences on the core decision making
variables.

An alternative explanation for the comparable decision accuracy of male and
female-dominated teams is that leaders of female teams somehow atoned for lower scores
on the core variables in the process of making the ﬁnal team decision. The results for
hierarchical sensitivity indicated that female leaders weighed their staﬂ‘ members’
recommendation less appropriately than male leaders. But they may have compensated for
this by incorporating their own evaluation of the data into making the decision. The
extent to which leaders weighed information in addition to staﬂ’ members’
recommendations was examined. This involved entering the nine attributes associated
with the decision object into a regression equation following the three staﬂ‘ members’
recommendations. There was no indication ﬁ'om this analysis that male or female leaders
were inﬂuenced by information other than staﬂ members’ recommendations.

Since it appears unlikely that female-dominated teams could have easily
compensated for lower team informity, staff validity and hierarchical sensitivity, it is
possible that these teams used a different style of decision making which is not represented
by the nnrltilevel theory. In order to test the adequacy of the theory in explaining female
team decision making, the relationship of the core variables to decision accuracy was

tested separately for male and female teams.

131

As shown in Table 6, the core variables accounted for more variance in decision
accuracy for female-dominated teams than for male-dominated teams (total R2 = .61 and
.53 respectively). However, a diﬂ‘erent pattern of relationships was found. In male-
domirrated teams, the best predictor of decision accuracy was team informity, while in
female-dominated teams it was staff validity. For male teams, the key to decision accuracy
was obtaining all the relevant information, but for female teams it was making good
recommendations to the leader. In addition to this diﬂ‘erence, hierarchical sensitivity was
only a signiﬁcant predictor in female-dominated teams, as was the interaction of team
informity and hierarchical sensitivity.

Although these results do not answer the question of why male and female-
dominated teams did not diﬂer in decision accuracy, they do suggest that male and female
teams arrive at their decisions through diﬂ‘erent processes. Further theory and research is
needed to determine the potential diﬂ‘erences between these processes.

Study Limitations

Although a number of interesting observations arose from this study, there remain
many unanswered questions due, in part, to the limited contribution of the task sex-linkage
and participation equality to the understanding of sex composition eﬂ‘ects. Some potential
problems with these variables are discussed next.

Task Sex-Linkage

With the exception of team informity, the sex-linkage of the task did not moderate
the eﬂ‘ects of composition as predicted. Diﬂ‘erences were expected to primarily occur
between male-majority and female-majority teams rather than between all-male and all-

female, with the nature of these diﬂ‘erences depending on the task sex-linkage. Instead,

132

Table 6
Regression of Team Decisgrn Accgacy on the Core Decision Variables for Male and
Fernjale-Domlated Tea_m_s

 

Step Independent Variable Total RT A R2 [i

 

Male-dominated teams (r_r = 40)

 

1 Team Informity (TI) .30" .30" -.32

2 Staﬂ’ Validity (SV) .46“ .16" -.34
Hierarchical Sensitivity (HS) .49“ .03 -. 15
TIx SV .51" .02 -.79
TIxHS .51“ .00 -.35
SV x HS .53" .02 .65

Female-dominated teams (_r_r = 47)

Team Informity (TI) .15“ .15" -.61
Staﬂ‘Validity (SV) .44** .29" -.7s
Hierarchical Sensitivity (HS) .51** .07* - 1.29
TIx SV .52" .01 -.35
TI x HS .59" .07* .88
SV x HS .61“ .02 .98

 

*p< .05. **p< .01.

133
teams dominated by males had a general advantage over those dominated by females. One

explanation for this eﬂ‘ect is that the general structure of both tasks favored the decision
making style of male teams. In other words, the female task may not have been as female-
oriented as it was intended.

The nursing task was designed to be equivalent in structure and corrrplexity to the
military task in order to isolate the eﬂ‘ects of stereotypes and avoid possible confounds of
additional task diﬂerences. Unfortunately, the eﬂ‘orts that were made to ensure
equivalency may have undermined the sex-linkage manipulation. The nursing task may
have been de-feminized because it did not involve the characteristics typically associated
with infant care (e. g., compassion, sensitivity). Even though the ﬁrture practice of
medicine may involve experts from remote locations who make decisions regarding
patients over computer terminals, the form of medical care most people are familiar with
involves hands-on treatment.

To avoid potential confounds, the tasks were piloted and modiﬁed to make sure
that they were equal in diﬁculty. Although pilot testing did not reveal any signiﬁcant
differences, this was conducted using individuals. At the team level, the task was adapted
to the hierarchical structure. Information was distributed according to staﬂ‘ members’
expertise and the ﬁnal decision was made by the leader. As suggested earlier, different
processes may have been involved in female and male team decision making. When the
task was adapted to the team level, the eﬂon to develop a female-oriented task may have
been compromised by a structure that was viewed as male-oriented. The trade-oﬂ‘ in
manipulating task sex-linkage is between creating equivalent, but perhaps superﬁcial tasks,

or developing realistic tasks which may diﬂer in number of additional ways. The latter

134
creates potential confounds which can make it diﬂicult to identify the exact cause of any

observed eﬂ‘ects.

In addition to generally failing to produce the intended eﬂ‘ects, the task had some
unexpected effects that suggest other problems with the sex-linkage manipulation. Teams
participated more equally on the female task and had lower staﬂ'validities than on the male
task. The difference in participation equality was attributed to the lower participation
rates of members when working on the female task It is not clear, however, why team
members participated less on the female than the male task The eﬂ‘ect on staﬂ validity
appeared to indicate the tasks diﬂ‘ered in difﬁculty. As suggested earlier, the nursing task
may have been counter-intuitive and somewhat superﬁcial to participants, making the task
more diﬂicult than the rrrilitary task Additionally, the military task may have been more
interesting or familiar to participants given the wide availability of war-type games. This
provides a reasonable explanation for both the lower participation and lower validities of
participants who worked on the female as opposed to the male task.

Although there are potential limitations with the sex-linkage manipulation, there is
evidence that the two tasks were viewed stereotypically. First, the task’s inﬂuence on the
individual informity of team members is consistent with the sex-linkage manipulation.
Members of mixed-sex teams responded diﬂ‘erently depending on the task. Their
informities were higher when the single member’s sex was consistent with the sex-linkage
of the task and lower when it was inconsistent.

Second, the data on dyadic sensitivity provides support for the sex-linkage
manipulation. Typically, the absohrte diﬂerence between the weights assigned by the

leader and the appropriate weight based on the correct decision is used to assess dyadic

135
sensitivity. However, one can also look at the direction of the weights to identify whether

leaders underweigh or overweigh their staﬂ‘s’ recommendations (see Figure 18). These
weights reveal that females were overweighed on the female task and underweighed on
the male task Similarly, males were underweighed on the female task, yet weighed fairly
appropriately on the male task. These ﬁndings suggest that, at least from the leader’s
perspective, the task did inﬂuence members’ perceptions. Females appeared to be viewed
as more competent on the female task and males more competent on the male task. The
sex-linkage manipulation may have had some of its intended eﬂ‘ects on the perceptions of
team members but not to the extent that was initially expected.

Particip ation Eguﬂg’

The other variable that contributed little to the understanding of sex composition
effects was the equality of members’ participation. Participation equality was included
along with the core decision making variables with the expectation that it would add to the
prediction of decision accuracy. It was argued that since each member’s contribution was
considered important to eﬂ‘ective team performance, a measure of participation equality
would capture an additional team process. Although teams did vary in the equality of their
participation, it was not a signiﬁcant predictor of team decision making accuracy. An
additional justiﬁcation for including participation equality as a key variable was that it
would promote team informity and hierarchical sensitivity. Unfortunately, participation
equality also failed to relate to these variables. For the decision making task used in this
study, participation equality does not appear to be a relevant variable. This does not rule
out the possibility that it may be important to other types of decision making tasks such as

consensus. For tasks in which participation equality is a critical determinant of eﬂ‘ective

136

.xmﬁ can 836388 3 0.3303 “Bum—mega was 8.363 .8032 80250; guanobmmvme 83025 an ommﬁ

nah eased nae 2e:

 

 

Essen a as: E messed a as: D

 

 

32.80.; .3203: 8.8.0an Ezéom @2032 x8085

  
    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

137
team performance, the composition effects may be different than those observed in the

present study.
Generalizab' '

In addition to limitations with the task and participation equality variables, there
are some potential limitations with regards to generalizing the ﬁndings here to other
situations. First, the participants in this study were undergraduate students whose
perceptions and motivations likely diﬂ‘er from employees in actual military or nursing
environments. For example, tmdergraduates may attmd more to cues indicative of a
social rather than working relationship. This would support the distraction theory
proposed to explain the lower informity and hierarchical sensitivity of male-majority
relative to all-male teams. That is, in the presence of a single female, males focus their
attention on her rather than the task

Second, motivational diﬂ‘erences are likely to exist between tmdergraduates and
actual employees even when efforts are made to provide incentives for good performance.
Participants in this study were randomly assigned to the male or female task, yet
employees typically choose their occupation. Therefore, individuals in military and
nursing ﬁelds, particularly non-traditional members, are likely to be more interested and
motivated to perform well.

Finally, the participants in this study may have less well-deﬁned stereotypes about
particular occupations than those who work in such environments. Although certain
majors may be characterized by disproportionate numbers of men or women, the
undergraduates in this sample were fairly integrated in terms of sex. However, military

units and nursing wards have been traditionally dominated by one sex which could result in

138
persistent stereotypes being developed about the type of persons suited for such

occupations. For example, many members of traditionally all-male military academies
maintain strong beliefs about the qualiﬁcations of females. This suggests that the sex-
linkage manipulation may not elicit the same stereotypes in undergraduates as it would in
actual employees.

The second generalizability issue involves the focus on hierarchical team decision
making. Although many sex composition studies use decision making tasks and address
groups with leaders, they typically do not examine the same variables or focus on the same
level of analysis as the present study. Most of the previous research addresses consensus
decision making, largely ignoring alternatives such a hierarchical decision making.
Different factors are likely to inﬂuence the quality of decisions made through consensus as
opposed to those made hierarchically. For example, in hierarchical decision making teams,
it is not important that members agree in their judgments, but in consensus decision
making, this is a critical determinant of effective team performance. This means that the
effects of sex composition on variables in one case may not generalize to another.
Therefore, conclusions regarding the eﬂ‘ects of sex composition and the task sex-linkage in
this study should not be generalized beyond hierarchical decision making contexts.

Recommendations for Future Reseaih

As the previous discussion indicates, a number of questions remain regarding the
effects of sex composition on team decision making. These questions can be categorized
into three main areas. First, research is needed to better understand the underlying causes
of the observed composition effects on the core decision making variables. Second, there

is a need to explain why there was no eﬁ‘ect on team decision accuracy. Third, the

139
potential inﬂuence of task characteristics on sex composition eﬂ‘ects should continue to be

explored. More speciﬁc issues to be address in each of these areas are discussed below.

For each of the decision making variables, potential explanations for sex
composition effects were identiﬁed at both the individual and team level. These inchided
individual sex differences (e. g., task-orientation) and diﬂ‘erences as a ﬁmction of the
composition (e. g., competitiveness). Data were available to address some of these issues
such as whether males were more aggressive in their decision making than females. The
majority of feasible explanations, however, could not be readily tested. Therefore,
numerous research questions can be generated from this study.

A basic question to be addressed is to what extent the observed effects are
attributable to individual and team level factors. It was suggested, for example, that the
higher informity of male-dominated teams may have been attributable to the higher task-
orientation of individual males. This raises the question of whether the differences
observed between compositions are simply reﬂections of individual differences between
males and females. Another ﬁnding was that male leaders appeared to be more negatively
aﬂ‘ected by mixed-sex compositions than females leaders. Future research should explore
how individuals might be diﬂerenﬁally aﬂ‘ected by the team’s composition. Finally, the
data seemed to suggest that male-majority teams were less motivated or task-focused than
all-male teams. Do different compositions or tasks lead to diﬂ‘erent levels of motivation
or interest- on the part of team members?

The second direction for future research involves the question of why sex
composition did not affect team decision accuracy. It was suggested that female-

dominated teams compensated for their lower performance on the core decision making

140

variables by outperforming male-dominated teams on some other factor related to decision
accuracy. It seems unlikely that a variable or combination of variables could offset the
amormt of variance in decision accuracy accounted for by the core variables. However,
post hoc analyses suggested that female and male-dominated teams arrived at their
decisions differently. A more in depth investigation of the process by which male and
female teams make decisions is needed. This might include examining the way males and
females process information, and whether female teams apply a different structure, like
consensus, to the decision making task.

The third area in which ﬁirther research is needed involves the inﬂuence of the
task. Although there were potential limitations with the sex-linkage manipulation, there
was evidence that team members responded differently to the two tasks. The informities
of staﬂ‘ members appeared to be inﬂuenced by the task in a way consistent with its sex-
linkage. The task eﬂ‘ects on participation rates and individual validities, however,
suggested additional differences between the tasks that may have overshadowed the sex-
linkage eﬂ‘ects. One direction for ﬁiture research is to develop better manipulations of the
sex-linkage construct. New manipulations should attempt to deal with problems of
artiﬁciality and potential differences in familiarity or interest with the task.

In addition to sex-linkage, this study suggests that the structure of a task may
inﬂuence sex composition eﬂ‘ects. Researchers have found that females prefer more
participative styles of leadership and males more autocratic styles (e. g., Jago & Vroom,
1982). The hierarchical decision making structure in this study may have been perceived as
a male-oriented style of decision making. Therefore, the general task structure may have

interfered with the sex-linkage manipulation. Since decision making styles may be

141
differentially associated with the sexes, the structure of the task might provide another

direction for studying the effects of sex composition. Research is needed that compares
how different compositions perform using hierarchical versus consensus decision making
structures.

These are some of the issues to be addressed by ﬁrture research. There are more
general questions which also need to be considered. For instance, how do teams
composed of equal numbers of males and female compare to those studied here? What
happens with larger team sizes? Are there differences between leaders who are members
of the majority sex versus the minority? And, are differences likely to emerge between
teams when working on a sex-neutral task Answers to these questions will hopeﬁrlly
provide more insight into the potential effects of changing workforce demographics.

Conclusions

The main question addressed by this study was how teams composed of different
representations of males and females perform on tasks traditionally associated with one
sex or the other. In attempting to answer this question, a number of interesting
observations were made regarding the effects of sex compositions on team decision
making.

In general, teams dominated by male members outperformed those dominated by
females on all three core variables related to team decision accuracy. However, teams
with a male majority were often less effective than all-male teams. Staff members were
weighed less appropriately in male-majority than in all-male teams, and they were less
informed in male-majority teams when working on the male task. The behaviors of team

members were clearly affected by the composition and task.

142

Despite the diﬂ‘erences found for the core variables, there were no composition
effects on team decision accuracy. Male and female-dominated teams appear to arrive at
the same outcome through diﬂ‘erent processes. This implies that different task structures
may be appropriate for different compositions or alternatively, that teams should be
composed to best ﬁt the structure of the task.

These results raise some important issues to be considered by researchers and
practitioners. When conducting team studies, researchers should be aware of the potential
effects diﬂ‘erent sex compositions may have on the results. Depending on the research
question, it may be desirable to control for composition in the study design or data
analysis. In terms of designing work teams, practitioners should be aware that certain
compositions may have negative effects on the behaviors of team members, and that
certain compositions may be better suited for a particular task than others.

As the workforce becomes increasingly diverse, and in particular, men and women
continue to enter ﬁelds traditionally dominated by the opposite sex, additional efforts will
be needed to develop a better understanding of the effects of changing compositions on
team performance and the behaviors of team members. These efforts will hopefully ensure

the successful transition of men and women into non-traditional roles.

FOOTNOTES

1The terms gender and sex reﬂect the common uses in the literature (e. g., gender
diversity, sex composition). There is no intent here to distinguish between psychological
and physiological characteristics. Teams were composed according to the self-reported
sex of participants as either male or female.

2Throughout this manuscript, the terms group and team are used interchangeably
to refer to a variety of collectives with a common objective. The term team, however, is
primarily used in reference to the present study to be consistent with existing research and
theory on hierarchical decision making.

3Since participation equality did not correlate signiﬁcantly with team decision
accuracy, it was not included in the regression with the other core decision making
variables.

4Individual informity measures the percent of trials on which a staﬂ‘ member
obtained all of the requisite information. For example, if a staff member only acquires two
of the six pieces of information on every trial, his or her individual informity is 33%.

SPretests were conducted using individuals to ensure that both tasks were
comparable in terms of diﬁcnlty. The tasks were considered equivalent once there were

no signiﬁcant differences in individual performance across the two tasks.

143

APPENDICES

APPENDIX A

APPENDIX A

HANDBOOK FOR TEAM MEMBERS
NAVAL COMMAND AND CONTROL SIMULATION

Experiment #14

Team Eﬂ‘ectiveness Research Laboratory
Michigan State University
East Lansing, MI
Spring 1995

144

145

INTRODUCTION
Theyearisl996andyouareapartofaU.S.navalearriergroup'scormnandandcmtrolteamstationed
intheMiddleEast. Aregimalcmﬂiabetwemnmnaﬁaumﬂnsarmhasrecaﬂybrokmormand
mummismpmmmmcmmmmmmmmmm.
Ashistoryindieates,thisisahighlysensitivetask Forexarrple,inl987,failurebyaoormnandand
connoltaamtoquiddyandaccmatelyidentifyaplaneasdrrwemng,allowedanlraqijettoacddentally
ﬁretonxocetmissilesintoﬂreFriyteU.S.S.Stark,ldlling37 Ameriean servicemananderipplingthe
vessel. Oneyearlater,acormnandandcanroltamerrorresultedinﬂieU.S.S.CruiserVincames
acddanaﬂyslmoﬁngdownmhmumpassmyrplmeldﬂmgwonmocandvihms. Anyrepeatof
mistakesofﬂnsldnduaﬂpmbablyleadmawiﬂrdrawalofAmaieanforcesﬁmnﬂiearea. Sudra
withdrawal would have disastrous economic and politieal ramiﬁeations that would spread well beyond
thisregion.

THETASKFORCE
Yournavaleamergrupismarmyofslnpsmlmesmdodiasuppomngmntswnhdnpmposeof
protectingapproximatelyl96,0005quaremilesofooean. Inordertocontrolsudralargearea,radar
smveiﬂmrceisneoessarysoﬂntﬂreearﬁergmpisrnsmpnsedbyﬂiemny. Theearriergroupis
composedoffourmitswhoprovidemostoftheradarcoverage. Thesemitsarelinkedtogaherbyan
dectrorncdatanetworksoﬂmtdreyeansrpplybitsandpiecesofcriﬁealinfonmticnconoemingpossible
aremyplanestoead'rother. Essarﬁaﬂy,ﬂresefommntsmustcormnmneateandcoordinatewhatthey
memﬂidrnxhﬁmnlmdam,mmmﬂntmmemoumnmamovaaﬂasmmofmeaimaﬁ

TEAM MISSION - Moritaggg’ Air Space
TheteamofvvludryouareapartwiﬂplaytherolesofCormnandingOmcasofthefourunitswhich
cormosetheearriergroup'soormnandandcormolteam. Yourmissionistomonitortheairspace
smaMgﬂreearﬁergroupmmldngsmeﬂratnatalslﬁpsamndaﬂadred. Inperforningtlisrole,
younnntmdcccrtﬁndntyoudomtdlowlouofﬁfemdﬁngﬁomnddmorhmmﬁond
attacksonslipaindletaskforcc. Atthesameﬁnmitis also ofparamomrtimortancethatyoudo
mtinadverturﬂyshootdownﬁimrﬂynﬁtaryn’nnﬂoranydvilianircnﬁ. Manypassenger
ﬂightsmoveinmidornofﬂieregimandﬁimdlynnhtaryairaaﬂﬁomnaﬁmsnotmvdvedindre
conﬂiaalsopatrolthearea.

W
Therearefourrolesinﬂnssnnulaﬁmmeforeadrmemberofafourpersonteam. Theleaderisthe
CorrmrandingOﬁcer(CO)oftheAircraﬂCanier. Theotherteammanbersarethecormnanding
oﬁcers ofthe AWACS air recomaissanoe plane, the Cnriser, andthe CAD (Coastal Air Defense unit).
Theteam'smskismdeddevdmrespmsemeeamamdiomdmalmwwammcmmngaimaﬂ
targets. TheAWACS,Cnnser,andCADwiﬂmakereoonnnardationstoﬂreCanier,whowiﬂdiar
maketheﬁnaldecisionfortheteam.Teammanbersbasedra'rdedsimsmdatatheycollectby
measuring characteristics oftargets that enter the earrier group's airspace using sophistieated radar and
oﬂrerelearancdevicesﬂiatmanbersofdieteamareexpertsmhnerprding Aircraftthatarebeing
mdcedammﬂedmrgas.1haeamwvmpmsiblemdcsmnnkefmeadrmmrgalhese
rspmsmamgadedmtamofﬂrdragyssivmwsmﬂmaeisawmneampmsebrmdrarcmﬂ.
Eadrofﬂieseisdeseribedbelow,movingﬁ’omleasttomostaggressive.

146

SEVEN POSSIBLE DECISIONS

l) IGNORE: Tlﬁsmemsﬂratmﬁndreraumﬁmslwuldbedevaedmﬂretargamdmsteadfoms
shouldbedirectedmodrerpossibletargetsintheam. Neverigroreatargetthatnnghtpossiblyattadt.
Thiswmrldmostassmedlyleadtolossoflives.

2) REVIEW: TlﬁsmemsauarﬁmembeslnﬂedawayﬁmnﬂnstarganmranmﬂyAﬁerashon
periodofﬁmeﬂﬁstargetshorddberdumedtomordertotpdateitsstatus. Alargenumberoftargas
eanbeinreviewstams,however,ﬂ1isnrnnberisﬁnite.

3) MONITOR: Tlﬁsmemuﬂratdietargashmﬂdbecmﬁnummlyuadced.1hesystamﬂmdoﬂns

mammapabledmmgmmammmmmmmmmmm
overallpatroleapacity.

4) WARN: Tlﬁsmwrsthatamessageissenttothetargetorderingittotwnaway. Warningtargets

ﬂratshouldbeignreddaradsﬁomﬂreumormoflegiﬁmatewammgs. Wantingtargetsthatintend
toattaekisalsobad,sincethewamingmakesiteasierforanattadcertoloeatetheship.

5) READY: Tlﬁsmaansmgetnnoadefmsiveposmemdmsadefmsiveweapmsmammnaﬁc. A
shipinarmdiedpositimisrarelyvulnerabletoattadc Thisstanceshouldnotbetakmtonon-
mreatanngmrgassmwwmpmssamammnaﬁceanﬁmnnﬂakaﬂyanmocanmrgasﬂmﬂymo
closetotheearriergmup. Aslnpmdnsposiﬁathowever,eamotreadilytakeoﬂ‘msiveacﬁmtoward
othertargds.

6) LOCK-ON: This syndnauzesmdarandattadtweapmssoﬂiatﬂreweapmsﬁxﬂianselvesmﬂre
target. AslﬁpatIpdc-Onposiﬁmeantakeoﬂ‘arsiveadimatamomansnotice. Theeapacitytotradt
odramrgasissemdywrmnredmmaeislndr-Qrmasir'nglemrmhowem. Thus,this should
bereservedfortargetsﬂratarealmostcertaintobeﬂrreaternng

7) DEFEND: This is ”weapmsaway'andmeanstoattackthetargawithrrnssiles ordqrthcharges.
Adefenddedsioneamotbeabortedonoeinitiatedandthusmustmlybeusedwhenenemyattackis

147

CHARACTERISTICS OF AIRBORNE TARGETS
Airbometargetseanbemsuredcnnineattributes. Thesearelistedbdowalcngwiththerangesof

 

 

 

 

 

 

 

 

 

possible values for these attributes:
(1) Speed Faster targets are morethreatening. l00to 800mph.
(2) Altitude Lower targets are more tlnreatening. 35,000 to 5,000 ft.
(3) Size Smaller targets are more threatening 65 to 10 mtr.
(4) Angle Rapidly descending targets are more threatening +15 dgs (rapid ascent) to -15
dgs (rapid descent).
(5) IFF IFF stands for "Identiﬁcation Friend of Foe” and identiﬁes whether an aircraft is
civiliarn, para-military or rrnilitary. .2 MHz (civilian) to 1.8 MHz (military).
(6) Direction Targets coming straight towards carrier are more tlnreatening. +30 dgs (passing
far to the left or right) to 00 dgs (coming straight in).
(7) Corridor A corridor is a 20 mile wide “safe lane” open to commercial air traﬁc, and is
Status expressed in terms of miles away from the center of the corridor.
0 nn’ (in the middle ofit) to 30 mi (way onrt ofit).
(8) Radar Targets with weapons radar are more threatening
Type Class 1 (weather radar only) to Class 9 (weapons radar).
(9) Range Aircraft that are closer to the earrier are more threating.

 

 

 

200mitolmi.

 

 

148

DETERMINING THE LEVEL OF THREAT FOR A DECISION RULE

In general, the degree to whicln an incoming target is threatening depends on its standing on these
nine attributes. These nine attributes combine irnto three simple rules wlnich are used to determine
the danger associated with any target. The commanding ofﬁcers of the CAD, AWACS, and
Cruiser are each respornsible for combining these attributes irnto the rules. Eadn oﬁcer is
respornsible for their area of expertise plus an additional rule representing the expertise of another
menber. This ensures that all information will be sumciently represented in the team. Eadn
ofﬁcer must make an evaluation of the aircraft based on these rules. The Carrier is responsible for
combining the recommendations from the CAD, AWACS, and Cruiser into an overall team
decision.

Location Rule (CAD & AWACS):
ALTITUDE, CORRIDOR STATUS, and RANGE go together to determine the locatiorn of the

aircraft. Aircraﬂ are threatening only if they are low (low value on altitude), outside commercial
traﬁc lanes (lnigh value on corridor status), and close (low value on range) to the carrier. If any
one of these three values are non-threatening, then the aircraﬂ is to be considered norne-threatening
in terms of the location rule.

Movement Rule (AWACS & CRUISER):
SPEED, ANGLE, AND DIRECTION go together to detemnine the movement of the aircrait.

Aircraft are threatening only if they are going fast (high value on speed), descending (low value on
angle), and coming straight in to the Carrier (low value on direction). If any one of these three
values are non-threatening, then the aircraft is to be considered non-threatening in terms of the
movenent rule.

W

SIZE, IFF, and RADAR TYPE go together to detemnine the eategory of aircratt. Aircraft are
threatening only if they are small (low value on size), military (lnigh value on IFF) and carrying
weapons radar (high value on radar). If any one of these three values are non-threatening, then
the aircraft is to be considered non-threatening in terms of the category nnle.

 

COMBINING RECOMMENDATIONS TO DETERMINE TEAM DECISIONS

The three rules combine to determine the overall threat represented by the aircraft. For example, if
the CAD recommends a DEFEND, and the AWACS recommends a DEFEND, and the Cruiser
recommends a MONITOR - the Carrier should probably clnoose LOCK-ON or DEFEND since
two teammates recommend DEFEND and the other teammate is recommending an intermediate
course of actiorn.

Another example would be if the CAD recommends an IGNORE, and the AWACS recommends a
MONITOR, and the Cnriser recommends an IGNORE - the Carrier should probably choose
IGNORE or REVIEW since two teammates are recommending an IGNORE and the other
teammate is recommending an intermediate course of action.

Another example would be if the CAD recommends a DEFEND, and the AWACS recommends an
WARN, and the Cruiser recommends an IGNORE - since the three recommendations are spread
out, the Carrier should probably pick an irntermediate judgment such as MONITOR, WARN or
READY.

149

OUTCOMES OF DECISIONS
Yourdecisiorns regardingeadntargetaretobemadebasedupontheinfonnation ornthedimensions listed
above. According to rules described irn this section, there are ﬁve possible evaluative outcomes
associated with the accuracy or your decisions (scoring is done automatieally by the computer). The ﬁve

 

 

 

 

 

 

possible outcomes indude:
OUTCOME DEFINITION EXAMPLE SCORE
(1) HIT The decisiorn was exactly correct You said defend, correct +2
answer was defend
(2) NEAR MISS The decision was oﬂ‘ by one level You said defend, correct +1
answer was lock-on
(3) MISS The decisiorn was oﬂ‘ by two You said defend, correct 0
levels answer was ready
(4) INCIDENT The decision was oﬂ' by three You said defend, correct -1
levels answer was wam
(5) DISASTER The decision was cﬂ‘ by more You said defend, correct -2

 

 

 

than three levels

 

answer was eitlner monitor,
review, or ignore

 

 

 

150

CARRIER - INSTRUCTIONS FOR AIRCRAFT TARGETS

As the Carrier, you are the team’s leader and the person who must make the team's ﬁnal
decision. Each of your teammates (CAD, AWACS, and Cruiser) is responsible for sending you a
recommendation based on their assigned rules. Each rule requires three pieces of information. Each of
your teammates is able to directly measure one or two of the attributes that make up a rule; the remaining
information is measured by another member of the team. You are able to measure some of these attributes
(Angle, Corridor Status and Radar) in order to provide team members with information that they may
be missing

The following tables illustrate: l) the decision rules each team member is responsible for; 2) the
attributes which make up each decision rule; 3) the attributes the member can measure directly, and 4) the
attributes they need to make an informed recommendation. Since this information will not be posted at
your station, you will need to try to memorize what information each position has and what each needs.
This is important because your role is to help your teammates get the information they need to make
accurate recommendations. Remember, you are going to base your decision on your teammates
recommendations - you eannot be accurate unless they are.

CAD is r 'ble for the LO ATION and MOTION RULES

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LOCATION RULE: CAD IS ABLE TO CAD NEEDS:
MEASURE:

ALTITUDE ALTITUDE

RANGE RANGE

CORRIDOR STATUS CORRIDOR STATUS

MOTION RULE:

SPEED SPEED

DIRECTION DIRECTION

ANGLE ANGLE

AWACS is ramble for the MOTION and CATEGORY RULES

MOTION RULE: AWACS IS ABLE TO AWACS NEEDS:
MEASURE:

SPEED SPEED

DIRECTION DIRECTION

ANGLE ANGLE

CATEGORY RULE:

SIZE SIZE

IFF IFF

RADAR RADAR

 

CRUISER is rmnsible for the CATEGORY and LOCATION RULES

 

 

 

 

 

 

 

 

 

 

 

 

CATEGORY RULE: CRUISER IS ABLE TO CRUISER NEEDS:
MEASURE:

SIZE SIZE

IFF IFF

RADAR RADAR

LOCATION RULE:

ALTITUDE ALTITUDE

RANGE RANGE

CORRIDOR STATUS CORRIDOR STATUS

 

 

15]

CAD — INSTRUCTIONS FOR AIRCRAFT TARGETS

YOUR SPECIFIC ROLE

As the CAD, your role is to assess the aircraft targets on several attributes related to two decision rules
(LOCATION and MOTION) and then make a recommendation to the Carrier regarding the stance that
the team should take toward the target. You can directly measure one or two ofthe attributes that make
up each rule. Information on all three of the attributes in each rule is necessary in order to make an
accurate assessment of the target’s level of threat. Other members of the team have access to the attributes
you are missing Your unique knowledge concerns how to interpret and combine the attributes to assess
the LOCATION and MOTION of the target. If any attribute in a rule is non-threatening, the target is non-
thremeningonthatrule even ifthe othertwoattributesare inthe highthreateningrange.

Therangeofvalues anddegreeofthreatassociatedwitheachattributeare shownbelow. Youwill needto
try to memorize the end points of the value ranges on each scale in order to identify when the aircraft goes
from non-threatening to somewhat threatening and to very-threatening This information will not be
postedatyourworkstation-—youneedtohavethesevaluescommittedtomemory.

The bolded values are critical values, because they determine when the aircraft moves from one status
(non-threatening to somewhat threatening; or somewhat threatening to very threatening). Remember -
if an aircraft is non-threatening on any one value that makes up a nnle it is non-threatening overall
on that rule, regardless of the values for the other two attributes.

A'TTRIBUTES FOR THE LOCATION RULE:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Altitude 35k 25k -----15k 5k
(thou. ft) No threat Some threat
High threat
Range 20f.L 100 50 1
(mi) Some threat High threat
No threat
Corridor Status 0 10 20 30
(mi) No threat Some threat High Threat
A'I'TRIBUTES FOR THE MOTION RULE:
Speed 10LL 300 550 800
(mph) No threm Some threat
High threat
Direction 30 20 10 0
(€185) Some threat High threat
No threat
Angle +15 +6 -6 -15

 

 

 

(dgs) No threat Some tlnreat High Threat

152

AWACS - INSTRUCTIONS FOR AIRCRAFT TARGETS

YOUR SPECIFIC ROLE

As the AWACS, your role is to assess the aircraft targets on several attributes related to two decision rules
(MOTION and CATEGORY) and then make a recommendation to the Carrier regarding the stance that
the team should take toward the target. You can directly measure one or two of the attributes that make
upeachrule. Informationonallthreeoftheattributes ineach rule is necessary in orderto makean
accurate assessment of the target’s level of threat. Other members of the team have access to the attributes
you are missing Your unique knowledge concerns how to interpret and combine the attributes to assess
the MOTION and CATEGORY of the target If any attribute in a rule is non-threatening, the target is
non-threatening on that rule even if the other two attributes are in the high threatening range.

The range ofvalues anddegreeofthreatassociatedwith each attributeare shownbelow. You will needto
try to memorize the end points of the value ranges on each scale in order to identify when the aircraft goes
from non-threatening to somewhat threatening and to very-threatening This information will not be
posted at your work station - you need to have these values committed to memory.

The bolded values are critical values, because they determine when the aircraft moves from one status
(non-threatening to somewhat threatening; or somewhat threatening to very thremening). Remember -
if an aircraft is non-threatening on any one value that makes up a nrle it is non-threatening overall
on that rule, regardless of the values for the other two attributes.

A'T'TRIBUTES FOR THE MOTION RULE:

 

 

 

 

 

 

 

 

 

Smd 100 300 550 800
(mph) No threat Some threat

High threat
Direction 30 20 10 9
((185) Some threat High threat

No threat

Angle +15 +6 -6 -15
(dgs) No threat Some threat High Threat

ATTRIBUTES FOR THE CATEGORY RULE:

 

 

 

 

 

 

 

 

 

Size 65 40 20 10
(mtrs) No threat Some tlnreat

High threat
IFF .2 .7 1.2 1.8
(MHz) Some threat High threat

No threat

Radar 1 3 7 9
(class) No threat Some threat High Threat

153
CRUISER - INSTRUCTIONS FOR AIRCRAFT TARGETS

YOUR SPECIFIC ROLE

As the Cruiser, your role is to assess the aircraft targets on several attributes related to two decision rules
(CATEGORY and LOCATION) and then make a recommendation to the Carrier regarding the stance
that the team should take toward the target. You can directly measure one or two of the attributes that
make up each rule. Information on all three ofthe attributes in each rule is necessary in order to make an
accurate assessment of the target’s level of threat. Other members of the team have access to the attributes
you are missing Your unique knowledge concerns how to interpret and combine the attributes to assess
the CATEGORY and LOCATION of the target. If any attribute in a rnnle is non-threatening the target is
non-threatening on that rule even if the other two attributes are in the high threatening range.

The range ofvalues anddegreeofthrw associatedwith each attributeare shownbelow. You will need to
try to memorize the end points of the value ranges on each scale in order to identify when the aircraft goes
from non-threatening to somewhat threatening and to very-threatening This information will not be
posted at your work station - you need to have these values committed to memory.

The bolded values are critical values, because they determine when the aircraft moves from one status
(non-threatening to somewhat threatening; or somewhat tlnreatening to very threatening). Remember -
if an aircraft is non-threatening on any one value that makes up a rule it is non-threatening overall
on that rule, regardless of the values for the other two attributes.

ATTRIBUTES FOR THE CATEGORY RULE:

 

 

 

 

 

 

 

 

 

Size 65 40 20 10

(mtrs) No threat Some threat
High threat
[FF .2 .7 1.2 1.8
(MHz) Some threat High threat
No threat

Radar 1 3 7 9
(class) No threat Some threat High Threat

ATTRIBUTES FOR THE LOCATION RULE:

 

 

 

 

 

Altitude 35k mmrsk 5k
(thou. ft) No threat Some threat
High threat
Range 200 100 50 1
(mi) Some threat High threat
No threat
Corridor Status 0 10 20 30

 

 

 

(mi) No threat Some threat High Threat

APPENDIX B

APPENDIX B

HANDBOOK FOR TEAM MEMBERS
NEONATAL NURSING SIMULATION

Experiment #14

Team Effectiveness Research Laboratory
Michigan State University
East Lansing, MI
Spring 1995

154

 

155

INTRODUCTION
Theyearis1996andyouareapanofamenatalnmsingteameaﬂedmadtymﬂnesouﬂneastenUS.
whichlnsrecentlyexpeiencedarnepidenicofprenatm'ebirths. Researdnesaredilimrtlytryingto
isolatethepotertialenvirornnentaleausesofthisproblen. Intherneartime,UnivesityHospitalis
cenﬁenedvnﬁnmabmmaﬂylngnrateoflow-birﬂrwegitaBmmfarnswlndn usuallyexpeiencea
varietyoflifethreateninghealﬂnproblens. Caringforhigh—risknevbomsisanextrenelysensitiveissue.
Asof1995,ﬂneU.S.nnfartnwnalityrateetceedsﬂnatofmmyodnenaﬁms. Forbothmoralarnd
economicreasons,itishrpoﬂartﬂnatﬂneenretqriMcbehmdledeﬂ’ecﬁvdymmdweﬂnemfmn
mortalityrate.

THETASKFORCE
Yomnmsmgtammvdvewvedspedahesdegwimwnesnppmeaﬂwlwsepmpmeismmre
forhighriskinfants. Themosttedmologieallyadvancedmeasmingandmanitoringeqnnpmenis
availabletoyoutoassessthehealthoftbesenewboms. Infomnationaborteadninfant’shealﬂnisente’ed
intoacomputerdatabasetowhidnyourteamhasaccess. Yourteamiscomposedofthreespedalists
andasupervisor. Togeher,ﬂwfmnofyoumustcemnmneuemndmordimteﬂneun‘mmanmobtained
fromthedatabasetomakeanoveallassessmertofeadninfant.

TEAMMISSION-Mogm ' Egkﬂ m
Theteamofwlndnyouareapartwillplaytheroleofthefom'menbesofanecrnatalnnn'singteam.

Yourtaskistonnentorhigln-nsknnfamstlnateaetlneimensiveearewardofUnivesityHospital,arnd
makededsionsregardingtheirtreatment. Inperforningtlisrole,yonmnstmﬂrecert-°nthatyondo
notallowloss ofliferesritingﬁomtheprernatmereleaseofinfantsﬁ'omﬂnehospital. Atthesame
mitisdsonnportantthatyondonotoperateoninfamswhodonotreqﬁnsurgerysinceﬂisis
anextremelyriskyproeedm'e. Marnyinfantswillmoveinarndoutofthewardmdeyom'eare—you
wanttoensureﬂnatallnewbomsreceivedneappropnateearedneyreqnnre.

OVERVIEWQFROLES
Therearefourrolesinthis simulationoneforeadnmenberofafourpersonteam. Theleadeisthe

NursingSupervisor(NS)oftheNeonatalWard. TheodneteammenbesindudetheHeartSpecialist
(HS), theLung Specialist (LS), arndthe Metabolic Specialist (MS). Theteam's task is to decidewhat
respmseﬂnenmsmgteamshonddnunketowardmfamsinﬂnenematalward. TheHearLLungand
MetabolicSpecialistswillmakerecemnendaﬁenstotlneNmsingSupevisor,whowiﬂﬂnenmakeﬂne
ﬁnaldecisionfortheteam.Teammenbesbasetheirdedsionsendatatheycolleebymeasuring
dumdeisﬁeofnevbomsMetedmneeWmﬁlheemsuremeoMnedﬁensmﬁsﬁeﬂed
eqnnpmemﬂnenmsesareexpetsininnwreingﬂnisinfonnatim. Therearesevenpossiblednoicesto
makeforeadninfanttobeevaluated.Theseresponsesaregradedirntermsoftheirseveityarndtlneeis
oneconoeresponseforeadninfant.Eadnoftheseisdescribedbelow,movingfromleasttomostsevere.

156

SEVENPOSS LEDE NS

1) RELEASE: Thismeansallowtlnenewbomtobereleasedfromthehospital. Attentiornshouldbe
directedtowardotheinfantsinﬂneward. Neverdeaseaninfanttluntmightrequiresurgey. A
newbomdnatisrdeasedprenannelyemnbeplacednnahfeﬂneatenngsimaﬁm.

2) DOWNGRADE: ’lhisrneansmovethenewbomoutofintensiveeare. However, cominuetodneck
ontheinfantandnpdateitsstatus.

3) MONTTOR: Thismeansyouwiﬂcontinuetomonitoraﬂsignsforhealththrsats. Thisisarn
appropriaterespmsewhentheeissomethreﬂtothenewbom’shealth,buttlneinfantisnotyetin
seriousdanger. Madmmngnmnemnmfmnsdinnmshemeteam’soveaﬂeapadtymwaluaemfmns.

4) INCUBATE: Tlﬁsmemnstoplaceﬂnemfartmamdatedenvirennentoensuremedenal
factors(e.g,tenpeature)threatenthernewbom’shealtln. Theearelimitednumbersofirmbatorsinthe
neornatalward,sothededsientoirnenbatemustberesevedforthoseudnoneedsudneare.

5) THERAPY: This mnstryusingvarious physieal manipulationstoreducehigh risk symptoms.
The'apyisnctparticularlyusefulforedrenehealtlnproblens,andeanleadtofnnthedeteioratimin
dnenewbom’shealth. Howeve,itisbetetoavoidmedieatimandsmgeyifthehealdnpmblensean
beconoaedthroughdnerapy.

6) MEDICATE: Tlﬁsmvolveshmsﬁngﬂnenewbmnwiﬂnmedieaﬁmsmﬂnehcpeofcoﬂeemgﬂne
problenwiﬂnoutsurge'y. Medieatingdneinfartdoesnaensmethatsmgeywillbeavoided;many
dnngmhavehamnfulsidedfedsdnatmaynewssitateancpeaﬁm. Tlneeforernnedieationshould
enlybeusediftheearealnnostcetainﬂnreatstotheirnfant’shealth.

7) OPERATE: Thismeansperfonrnngsnngeytocorrecttoinfant’.shealthproblens Surgeyis
endrenelyriskyandshouldonlybeusedwhentheinfarnt’sccnditionisdeariylife—thnwening

157

CHARACTERISTICSOFINFANTS
Infantscanbemeasnuedenninehealthattributes. Thesearelistedbelowalongwiththerangesof

 

 

 

 

 

 

 

 

 

possible values for these attributes:
(1) Breathing Faster respiratiorns are morethreatening. 20to 80rpm.
(2) Heart Faster heart rates are more threatening. 40 to 130 bpm.
Rate
(3) Glucose Lower glucose levels are more threatening. 100 to 10 mg%.
(4) pH Balance Lower pH balances are more threatening. 7.8 to 7.0 pH.
(5) Calcium Lower calcium levels are more tlnreatening. 10 to 4 mydl.
(6) Mean Air Higher MAP is more threatening. 5 to 15 cmHZO.
Pressure
(MAP)
(7) Temperature Higher tenperatures are more threatening 97 to 103 dgs.
(8) Magnesium Lower magnesium levels are more threatening. 2.5 to 0 mEq.
(9) Blood Higher blood pressure is more threatening. 20 to 95 mmHg.

 

 

Pressure

 

 

 

158

DETERMINING THE LEVEL OF THREAT FOR A DECISION RULE

In general, the degree to whicln an infarnt’s health is threatening depends on its standing on these
nine attributes. These nine attributes combine irnto three simple rules whicln are used to detemnine
the danger associated with any newborn. The nurse specialists of the heart, lnmgs, arnd metabolism
are eacln respornsible for combining these attributes into these rules. Each nurse is responsible for
their area of expertise plus an additional rule representing the expertise of another menber. This
ensures that all irnfonnation will be sufﬁciently represented in the team. Eadn nurse must make an
evaluation of the newbom’s cornditiorn based on these rules. The nursing Supervisor is responsible
for combining the recommendations from the nursing specialists irnto an overall team decision.

Cardiovascular Rule (E & MS):
HEART RATE, BLOOD PRES SURE, and TEMPERATURE go together to determine the

cardiovascular condition of the infant. Infants are nnnder high threat if their heart rate is fast,
blood pressure is high and temperature is high. If any one of these three values falls irn the norn-
threatening range, tlnen the infant is to be considered non-threatening in terms of the eardiovascular
rule.

Res irato Rule S&HS:
BREATHING, MAP, and pH BALANCE go together to determine the repiratory conditiorn of the

infarnt. Infants are nmder lnigh threat only if their respirations are fast, their MAP is high, and
tlneir pH balance is low. If any one of these three values falls in the non-threatening range, then
the infarnt is to be cornsidered norn-tlnreatening in terms oftlne respiratory rule.

Homeostasis Rule (MS & LS):
GLUCOSE, CALCIUM, AND MAGNESIUM go together to determine the homoestatic conditiorn

of the infant. Infants are under high threat only if their glucose is low, calcium level is low, and
magnesiunn level is low. If any one of these three values falls in the non-threatening range, then
the infarnt is to be cornsidered non-threatening in terms of the homeostasis rule.

COMBINING RECOMMENDATIONS TO DETERMINE TEAM DECISIONS

The three rules combine to deermine the overall threat of the infant’s health. For example, if the
HS recommends to OPERATE, and the LS recommends to OPERATE, and the MS recommends
to MONITOR - the Supervisor should probably dnoose MEDICATE or OPERATE since two
teammates recommend OPERATE and the other teammate is reconnnending an irnternnediate
course of action.

Another example would be if the HS recommends a RELEASE, and the LS recommends a
MONITOR, and the MS recommends a RELEASE -- the Supervisor should probably choose
RELEASE or DOWNGRADE since two teammates are recommending a RELEASE and the other
teammate is recommending an intermediate course of action.

Another example would be if the HS recommends to OPERATE, and the LS recommends a
HNICUBATE, and the MS recommends a RELEASE - since the three recommendations are spread
onrt, the Supervisor should probably pick an irntennediate judgment such as MONITOR,
INCUBATE or THERAPY.

159

OUTCOMES OF DECISIONS
Yourdee'sions regardingeadninfant areto bemadebased upontheinfonnation onﬂnedimensions listed
above. Accordirng to rules described irn this sedorn, there are ﬁve possible evaluative outcomes
associated with the accuracy or your decisions (scoring is done automatically bytlne compute). The ﬁve

 

 

possible outcomes indude:
OUTCOME DEFINITION EXAMPLE SCORE
(l) HIT The decisiorn was exactly correct You said operate, correct +2

answer was operate

 

 

 

 

 

 

 

than three levels

 

answer was either monitor,

downgrade, or release

 

(2) NEAR MISS The decisiorn was oﬂ~ by one level You said operate, correct +1
answer was medicate
(3) MISS The decisiorn was oﬂ‘ by two You said operate, correct 0
levels answer was therapy
(4) INCIDENT The decisiorn was oﬂ‘ by three You said operate, correct -1
levels answer was incubate
(5) DISASTER The decisiorn was oﬂ' by more You said operate, correct -2

 

 

160

NURSE SUPERZQOR - INSTRUCTIONS FOR NEONATAL TASK
As the Nurse Supervisor, you are the team’s leader and the person who must make the team's ﬁnal

decision. Each of your teammates (HS, LS and MS) is responsible for sending you a recommendation
based on their assigned rules. Each rule requires three pieces of information Each ofyour teammates is
wle to directly measure one or two of the attributes that make up a rule; the remaining information is
measured by another member of the team. You are able to measure some of these attributes
(Temperature, pH Balance and Magnesium) in order to provide team members with information that

theymaybemissrng

The following tables illustrate: 1) the decision roles each team member is responsible for; 2) the attributes
which make up each decision rule; 3) the attributes the member can measure directly; and 4) the attributes
they need to make an informed recommendation. Since this information will not be posted at your station,
you will need to try to memorize what information each position has and what each needs. This is
important because your role is to help your teammates get the information they need to make accurate
recommendations. Remember, you are going to base your decision on your teammates recommendations -
-youcannotbeaccurateunlesstheyare.

HS is rmible for the CARDIOVASCULAR and RESPIRATORY RULES

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CARDIOVASCULAR HS IS ABLE TO HS NEEDS:

RULE: MEASURE:

HEART RATE HEART RATE

BLOOD PRESSURE BLOOD PRESSURE

TEMPERATURE TEMPERATURE

RESPIRATORY

RULE:

BREATHING BREATHING

MAP MAP

pH BALANCE pH BALANCE
LS is remnsible for the RESPIRATORY and HOMEOSTASIS RULES

RESPIRATORY LS IS ABLE TO MEASURE: LS NEEDS:

RULE:

BREATHING BREATHING

MAP MAP

pH BALANCE pH BALANCE

HOMEOSTASIS

RULE:

GLUCOSE GLUCOSE

CALCIUM CALCIUM

MAGNESIUM MAGNESIUM
MS is nsible for the HOMEOSTASIS and CARDIOVASCULAR RULES

HOMEOSTASIS MS IS ABLE TO MS NEEDS:

RULE: MEASURE:

GLUCOSE GLUCOSE

CALCIUM CALCIUM

MAGNESIUM MAGNESIUM

CARDIOVASCULAR

RULE:

HEART RATE HEART RATE

BLOOD PRESSURE BLOOD PRESSURE

TEMPERATURE TEMPERATURE

 

 

 

 

16 l
HEART SPECIALIST (Ii—St INSTRUCTIONS FOR NEONATAL TASK

YOUR SPECIFIC ROLE

As the HS, your role is to assess the infant’s health on several attributes related to two decision rules
(CARDIOVASCULAR and RESPIRATORY) and then make a recommendation to the Supervisor (N 8)
regarding the stance that the team should take toward the infant. You can directly meaure one or two of
the attributes that make up each rule. Information on all three ofthe attributes in each rule is necessary in
order to make an accurate assessment of the infant’s level of health threat. Other members of the team
have access to the attributes you are missing Your unique knowledge concerns how to interpret and
combine the attrrbutes to assess the CARDIOVASCULAR and RESPIRATORY state of the infant. If any
attribute in a rule is non-threatening, the infant’s health is non-threatening on that nnle even if the other
two attributes are in the high threatening range.

The range of values and degree of threat associated with each attribute are shown below. You will need to
try to memorize the end points ofthe value ranges on each scale in order to identify when the infant’s
health goes ﬁom non-threatening to somewhat threatening and to very-threatening This information will
notbepostedatyourworkstation—youneedtohavethesevalues cornmittedtomemory.

The bolded values are critical values, because they detemnine when the infant’s health moves from one
status (non-threatening to somewhat threatening; or somewhat threatening to very threatening).
Remember — if an infant’s health is non-threatening on any one value that makes up a rule it is non-
threatening overall on that rule, regardless of the values for the other two attributes.

ATTRIBUTES FOR THE CARDIOVASCULAR RULE:

 

 

 

 

 

 

 

 

 

Heart Rate 40 70 100 130
(913111) No tlnreat Some threat

High threat
Blood Pressure 20 45 70 95
(mmHg) Some tlnreat High threat

No threat

Temperature 97 99 101 103
(dgs) No threat Some threat High Threat

ATTRIBUTES FOR THE RESPIRATORY RULE:

 

 

 

 

Breathing 20 40 50 80
(rpm) No threat Some threat
High threat
MAP 5 8 12 15
(cmHZO) Some threat High threat
No threw
pH Balance 7.8 7.5 7.3 7.0

 

 

 

(pH) No threat Some threat High Threat

162

LUNG SPECIALIST - INSTRU I N R NEONATAL T K

YOUR SPECIFIC ROLE

As the LS, your role is to assess the infant’s health on several attributes related to two decision rules
(RESPIRATORY AND HOMEOSTASIS) and then make a recommendation to the Supervisor (NS)
regarding the stance that the team should take toward the infant. You can directly measure one or two of
the attributes that make up each nnle. Information on all three of the attributes in each rule is necessary in
order to make an accurate assessment of the infant’s level of health threat. Other members of the team
have access to the attributes you are missing Your unique knowledge concerns how to interpret and
combine the attributes to assess the RESPIRATORY and HOMEOSTATIC state of the infant. If any
attribute in a rule is non-threatening, the infant’s health is non-threatening on that rule even if the other
two attributes are in the high threatening range.

The rangeofvalues anddegreeofthreatassociatedwith eachattributeareshownbelow. You will needto
try to memorize the end points of the value ranges on each scale in order to identify when the infant’s
health goes ﬁ'om non-threatening to somewhat threatening and to very-threatening This information will
notbepostedatyourworkstation-youneedtohavethesevalues cornmittedtomemory.

The bolded values are critical values, because they determine when the infant’s health moves from one
status (non-threatening to somewhat threading; or somewhat threatening to very threatening).
Remember — if an infant’s health is non-threatening on any one value that makes up a nnle it is non-
threatening overall on that rule, regardless of the values for the other two attributes.

ATTRIBUTES FOR THE RESPIRATORY RULE:

 

 

 

 

 

 

 

 

Breathing 2t} 40 60 80
(rpm) No threat Some threat

High threat
MAP 5 8 12 15
(cmHZO) Some threat High threat

No threat

pH Balance 7.8 7.5 7.3 7.0
(pH) No threat Some threat High Threat

ATTRIBUTES FOR THE HOMEOSTASIS RULE:

 

 

 

Glucose 100 70 40 10
(mg%) No threat Some threat
High threat
Calcium 10 8 6 4
(mg/d1) Some threat High threat
No threat
Magnesium 2.5 1.5 .5 0

 

 

 

(mEq) No threat Some threat High Threat

163

METABOLIC SPECIALIST (m) — INSTRUCTIONS FOR NEONATAL TASK

YOUR SPECIFIC ROLE

As the MS, your role is to assess the infant’s health on several attributes related to two decision rules
(HOMEOSTASIS and CARDIOVASCULAR) and then make a recommendation to the Supervisor (NS)
regarding the stance that the team should take toward the infant. You can directly measure one or two of
the attributes that make up each rnnle. Infornnation on all three of the attributes in each rnnle is necessary in
order to make an accurate assessment of the infant’s level of health threat. Other members of the team
have access to the attributes you are missing Your unique knowledge concerns how to interpret and
combine the attributes to assess the HOMEOSTATIC and CARDIOVASCULAR state of the infant. If any
attribute in a rule is non-threatening, the infant’s health is non-threatening on that rule even if the other
two attributes are in the high threatening range.

The range of values and degree of threat associated with each attribute are shown below. You will need to
try to memorize the end points of the value ranges on each scale in order to identify when the infant’s
health goes from non-threatening to somewhat threatening and to very-threatening This information will
notbepostedatyourworkstationnyou needtohavethesevaluescommittedtomemory.

The bolded values are critical values, because they deterrrnine when the infant’s health moves from one
status (non-threatening to somewhat threatening; or somewhat threatening to very threatening).
Remember — if an infant’s health is non-threatening on any one value that makes up a rule it is non-
threatening overall on that rule, regardless of the values for the other two attributes.

ATTRIBUTES FOR THE HOMEOSTASIS RULE:

 

 

 

 

 

 

 

 

 

 

 

 

 

Glucose 100 70 40 10
(mg%) No threat Some threat
High threat
Calcium 10 8 6 4
(mg/d1) Some threat High threat
No threat
Magnesium 2.5 1.5 .5 0
(mEq) No threat Some threat High Threat
ATTRIBUTES FOR THE CARDIOVASCULAR RULE:
Heart Rate 40 70 100 130
(bpm) No threat Some threat
High threat
Blood Pressure 20 45 70 95
(mmHg) Some threat High threat
No threat
Temperature 97 99 101 103
(dgs) No threat Some threat High Threat

APPENDIX C

APPENDIX C

TEAM INTERACTIVE TRAINING SCRIPT

(CNC/NEO Task)

Tell subjects: "The interactive training involves primarily the ﬁrst game of the simulation
which is 10 minutes long. You will then have two additional trials to practice on. Please
follow along closely and do not get ahead of me so that we can cover all the material."
Make sure everyone has their keyboard and mouse nearby. Then bring up the ﬁrst trial by
pressing the spacebar on the Carrier. Make sure everyone is on the bhne icon screen and the
correct task

1

2

Poinrt out the icons, game If, time clock, and menu bar.

First, explain that they will use their mouse to access the menu bar. "To measure
attributes, drag the mouse and click on the Measure menu Notice the 3 attributes you
canmeasureareinthismenu. Tomeasureanattribnrtesjustpointtothedesired
attribute and click on it. A gray box will appear on the lower left corner ofyour
screen showingthe attributevahne. Thiswilldisappearin3 seconds— don‘t worry, you
can review the information later. Nowmeaarre a second atuib ."

"After you have measured two attributes, open the measure menu again and click on
the Measure Summary option. This is a surmnary box that displays the attributes you
havemeasured. Itwillstayopenforonlyafewseconds Youcanalsoaccessthis
snunmaryboxbyhittingtheFZkeyonyour keyboard."

"Next, we will cover the Query option You will notice that you can only measure a
few attributes. Other people inn the team have the additional information you need
Therefore, you may want to ask for this information, which you can do using Query.
Click on the Query option, but do not proceed any further yet! In the Query menu
youwillsee allthe attributes When you click on one ofthese attributes, youwill get a
box asking you who you want to Query, inn other words, who you want to ask for the
information just selected. You need to make a choice quickly by clicking on a station.
Forthispradcetriallwilltellyouwho to send a anery.

Carrier should Query the CAD I NurseSpvr should Query the HeartSpec
CAD should Query the AWACS I HeartSpec should Query LungSpec
AWACS should Query the Cruiser I LungSpec should Query MetabSpec
Cruiser should Query the Carrier I MetabSpec should Query NurseSpvr

Now select an attribute and send a query to the above mentioned stations by clicking
on the appropriate station in the gray box "

164

 

165

"The (by abbreviation slnould show up in the lower left corner of each screen To get
the query you will use the Receive menu. Click on the Receive menu and click on the
Q to read the Query. Make sue you pay attention to who sent the query and what
theyaskedfor— oncethebox disappearsit cannotberecalled.

"Now you will send somethinng to the person who queried you. To do this you will use
the Transmit option. Open the 'ltansmit menu Notice that the only attribtes slnown
in this menu are the attributes you have measured You cannot send your teammates
attributesyou cannotmeasureorhaven'tmeasuredyet. Whathant you to do is
select an attribute and send it to the same person who queried you. Ifyou don‘t have
the information requested, for now send something you do have. If you don‘t
remember who queried you, here's who you transmit to:

Carrier to Cruiser I NurseSpvr to MetabSpec
CAD to Carrier I HeartSpec to NurseSpvr
AWACS to CAD I LungSpec to HeartSpec
Cruiser to AWACS I MetabSpec to LungSpec

"Notice the Tim abbreviation inn the lower left part ofthe screen There can be three
types of messages displayed here. QRY indicates a Query or that someone is asking
youformformaﬁomTRdeicatesaTransmitorthatsomeoneissendingyou
information, and MSG indicates a Text Message which we'll go over in a few minutes

ThekeyistouseReceive anytime you see a messageinthelowerleﬁ comer ofthe
screen. Now go to the Receive menu and get the Transnnit sent to you. After you
receivedthetransnit,youcanpresstheF2keyagairnandseethatthistransrnitted
irnformation is now in you summary box"

"While no talking is allowed during the simulation, you can send typewritten messages
to one another using the Text option nmder the Transmit menu Before you select this
option, I want to tell you who to send a text message to:
Carrier / Nnn'seSpvr send a text message to the CAD / HeartSpec
CAD I HeartSpec send a text message to the AWACS l LungSpec
AWACS / LungSpec send a text message to the Cruiser l MetabSpec
Cruiser I MetabSpec send a text message to the Carrier I NurseSpvr

Now select text, then the station you want to send a message to. When you get to the
box with the long black strip, you can type inn a short message (e.g., Hello) and then
send the message by pressing Enter.

"At this poinnt everyone should have an Msg abbreviation in the lower left comer of
your screen Again, you receive this message by seledng the Receive menu When
you receive the message you have the option of Log or Don't Log. Log is the deﬁnult
andmeansthatthemessagewillbe savedto a summaryﬁle. To seethis summaryﬁle,
click on Measure and the choose LOG File. This is a summary box of all the

10

ll

12

13

14

166

messagesyousave. YoucanalsoaccesstheIDGFilebyhittingtheFSkeyonyour
keyboard. Themessagesyou savewillremaininthesnunmaryboxforthe entire
sirmnlation unless you delete them To delete simply highlight the desired message
usingthearrowkeysonyourkeyboardandhittheDeletekey.

"You can also send a text message to yourself You may want to do so to reminnd
yourself of something for later inn the sinnrlation. To do this simply select Transmit
then Text and to the ﬂu right select Log File. Then just type a message to yourself and
hit enter. You do not have to receive this message - it will automatically go to your
snnmmarybox. To seethisrnessage alongwiththose othershave sent you,justhitthe
F3 key.

"So ﬁnr you‘ve learned how to Measure an attribute, how to request information ﬁom
someone else using Query, how to send specmc irnformation using Transmit, how to
send and log Text Messages, and how to Receive queries, transmits and messages.
The last thirng to cover is how to make Judgments.

"With 30 seconds left, the clock starts beepinng irndicating a judgnnernt should be sent by
outlyirng stations (CAD, AWAC, Cruiser / HeartSpec, angSpec, MetabSpec). _Th_is
judgment must reach the Carrie / NurseSpvr with enoggh time left to make a
team decision. Everyone except the Carrier I NurseSpvr should open the Judgment
menu." (Explainn to the Carrier I NurseSpvr that if the he or she has the menu bar
highlighted or a box open, he/she will not receive incoming judgments) For these
pradce trials I will tell you what judgrnernt to make.

"To make a judgment, you will click on the decision you believe to be correct and then
hit OK to Verify. Ifyou accidently select the wrong decision, you can click on Cancel
and resubmit yourjudgment Once the judgment is sent, you cannot change it even if
thereistimeleﬂ onthe clock. Youcan submityourjudgrnernt at anytime durirngthe
trial,butifyoufailto sendin ajudgmentbythetimethe clocknmsout yourdecision
will be registered as a No Call which equals an Ignore / Release. Right now everyone
except the Cannier I Nnn'seSpvr should choose an Ignore l Release.

"Now its time for the Carrier / NurseSpvr to make its Judgment. Notice that when an
icontumsreditmeansthat stationhasmade a decision andthe decisionisindicatedin
the middle of the icon. These reconnnendations from the CAD I HeartSpec, AWACS I
ergSpec and Cruiser I MetabSpec are what the Carrier I NurseSpvr based the team
decision on, so make sure your judgments get to the Carrier / NurseSpvr in plenty of
time before he/ she has to make the ﬁnal decision The Carrier / NurseSpvr should now
make an Ignore / Release decision

Once the Carrier I NurseSpvr has made the decision the next screen will be the
feedback screen. There are different types of information on this screen irnchrding: the
team decision (Ignore l Release), the correct decision (Ignore l Release), the
outcome which was a Hit meanirng you made the right decision The individual
station judgments along the leﬁ side of the screen are the recommendations you made

15

16

17

167

to the Carrier I NurseSpvr and should all be Ignores I Releases The performance
history on the right tells you the number of trials and outcomes. The performance
score at the bottom is based on 2 points for a "Hit", 1 for "Near Miss", 0 for "Miss", -
1 for "Incidert", and -2 for "Disaster". Depending on the outcomes, the backgronmd
color will vary, for instance green means Hit, and black means Disaster.

At this time I want to make three ve'y important points.

(1) With 30 seconds it is very important that non-Carrier / non-Supervisor stations
make theirjudgments quickly so as to leave time for the Carrier I NurseSpvr to make
the team decision

(2) When the time is nmning down and a judgment was already made, the non-Carrier I
non- Supervisor stations need to have their menu bar clear and be out of any boxes inn
order to proceed to feedback and the next trial. To make snn'e everythinng's clear you
can sinnply hit the ESC key. Ifyou notice negative time on the clock or are still inn
feedback when others have proceeded to the next game, you need to contact a
researcher immediately.

(3) When you are in a box, the clock will appear ﬂow, but it is actually conmting
down. Be aware!

(4) There is no talking during the experinnent. All communications must take place
through the computer.

"The next trial will automatically appear on your screens Go ahead and practice on
thenexttwotrials Noticethatthereislesstimeforthistrial Aﬁerthepracticetnials,
the amormt of time you have to make decision will vary so you will need to pay
attention to the clock " (Let them make their own decisions) "I will answer any
questionschnringthistime, but aﬁerthatyouwillnotbe allowed to talk You willbe
monitored through an irntercom and eveything you type is recorded by the computer."
(lining these two trials, make are eveyone understands how to use the computer and
that there are no misunderstandings)

At the erd of trial #3 hit Ctrl—F9 on the Carrier's / NurseSpvr’s machine and answer
any ﬁnal questions Refer back to the protocol, item #7.

LIST OF REFERENCES

Alexander, V.D., & Thoits, PA. (1985). Token achievement: An examination of
proportional representation and performance outcomes. Soworceg 64. 332-340.

Aries, E. (1976). Interaction patterns and themes of male, female, and mixed
groups. Small Group Behavior, 7, 7-18.

Bales, RF. (1970). Person ' and innt ersonal behavior. New York: Holt,
Reinhart & Winnston.

Bames-Farrell, J.L., L'Heureux-Barrett, T.J., & Conway, J.M. (1991). Impact of
gender-related job features on the accurate evaluation of performance innformation.

Organizational Behavior and Human Decision Processes, 48, 23-35.

Baucom, D.H., & Danker-Brown, P. (1984). Sex role identity and sex
stereotyped tasks in the development of learned helplessness in women. Journal of

Personm and Social chhology, 46, 422-430.

Bauer, R.H., & Turner, J.H. (1974). Betting behavior inn sexually homogeneous
and heterogeneous groups. chhological Reports, 34, 251-258.

Berger, 1, Fisek, H., Norman, R., & Zelditch, M (1977). Status characteristics, a
social interaction: An mectation—states approach. New York: Elsevier.

Bradley, PH. (1980). Sex, competence and opinion deviation: Ann expectation
states approach. Communication Monograph; 47, 101-110.

Bradley, PH. (1981). The folk-linguistics of women's speech: An empirical
examinnation. Comnnmication Monographs, 48, 73-90.

Broverman, I.K., Vogel, S.R., Broverman, D.M., Clarkson, F.E., & Rosenkrantz,
RS. (1972). Sex-role stereotypes: A current appraisal Journal of Social Issues 28 59-
78.

 

Burtt, HE. (1920). Sex differences inn the effect of discussion. Journal of
Experimental chhology, 3, 390-395.

168

169

Butler, D., & Geis, EL. (1990). Nonverbal aﬂ‘ect responses to male and female
leaders: Implications for leadership evaluations. Journal of Personm and Social

Pyohology, 58, 48-59.

Clement. D.E., & Schiereck, 1’]. Jr. (1973). Sex composition and group
performance in a visual detection task. Memog and Cogmp' 'on 1, 251-255.

Cohen, J., & Cohen, P. (1983). lied ° 1e re ession/correlation an sis
for the behavioral sciences. Hillsdale, NJ: Lawrence Erlbaum Associates.

Costrich, N., Feinsteinn, J., Kidder, L., Marecek, J., Pascale, L. (1975). When
stereotypes hurt: Three studies of penalties for sex-role reversals. Journal of
Experimental Social chhology, 11, 520-530.

Cota, A.A., & Dion, KL. (1986). Salience of gender and sex composition of ad
hoc groups: An experimental test of distinctiveness theory. Journal of Personmy' and

Social Psychology, 50, 770-776.

Cox, T. (1993). Cultural diverm' in organizations: Theogg research and
practice. San Francisco: Berrett-Koehler Publishers.

Craig, J.M., & Sherif, CW. (1986). The effectiveness of men and women in
problenn-solving groups as a ﬁnnction of group gender composition. Sex Role; 14. 453-
466.

Davis, 111 (1992). Some compelling intuitions about group consensus decisions,
theoretical and empirical research, and innterpersonal aggregation phenomena: Selected

examples, 1950-1990. Organizational Behavior and Human Decision Processes, 52, 3-3 8.

Deaux, K (1984). From inndividual differences to social categories: Analysis of a
decade's research on gender. American Psycholognst’ , 39, 105-116.

Deaux, K. (1985). Sex and gender. Annual Review of chhology, 36, 49-81.

Deaux, K., & Emswiller, T. (1974). Explanations of successful performance on
sex-linked tasks: What is skill for the male is luck for the female. Jounna_l_of Perm

and Social chhology, 29, 80-85.

Dion, KL. (1985). Sex, gender, and groups: Selected issues. In V.E. O'Leary,
ILK. Unger, and BS. Wallston (Eds. ), Womg, gender, and social pgchology. Hillsdale,
NJ: Lawrence Erlbaum

Eagly, AH. (1978). Sex diﬂ‘erence inn inﬂuenceability. chhologn'cal Bulletin 85,

 

86-116.

170

Eagly, AH. (1987). Sex diﬂ‘erences in social behavior: A soci_al-role
integpretation. Hillsdale, NJ: Erlbaum

Eagly, AH, & Carli, LL. (1981). Sex of researcher and sex-typed
communications as determinants of sex differences in inﬂuenceability. A meta-analysis of
social innﬂuence studies. Psychological Bullﬂ,’ 90, 1-20.

Eagly, A.H., & Wood, W. (1982). Inferred sex diﬂ'erences inn status as a
determinant of gender stereotypes about social inﬂuence. Joumsl of Personal_rty° and

Social Psychology, 43, 915-928.

Eskilson, A, & Wiley, M. G. (1976). Sex composition and leadership in small
groups. Sociomggg, 39, 183-194.

Feldman—Summers, S., Montano, D.E., Kasprzyk, D., & Wagner, B. (1980).
Inﬂuence attempts when competing views are gender-related: Sex as credibility.

Pachology of Women Mi. 311-320.

Floge, L., & Merrill, D. (1986). Tokenism reconsidered: Male nurses and female
physicians in a hospital setting. Socisl Forces, 64, 925-947.

 

Goktepe, J.R, & Schneier, CE. (1988). Sex and gender effects in evaluating
emergent leaders inn small groups. Sex Roles 19, 29-36.

Greene, L.R, Morrison, RL., & Tischler, N.G. (1981). Gender and authority:
Effects on perceptions of snnall group co-leaders. Snnall Group Behavior 12 401-413.

 

Hacknnan, JR (1990). Gron_1ps that work (and those that don’t). San Francisco:

Jossey-Bass.

Hedlund, J. (1993). Computer-mediated versus face-to-fsce communication in
hierarchical team decision making. Unpublished master’s thesis, Michigan State

University, East Lansing, Michigan.

Heilman, M.E., Block, C.J., Martel], RF., Simon, MC. (1989). Has anything
changed? Current characterizations of men, women, and managers. Journal of Applied
chhology, 74, 935-942.

Hoffman, C.R, & Hurst, N. (1990). Gender stereotypes: Perception or
rationalization? Jomil of Personm and Social chhology, 58, 197-208.

Hoﬂinan, LR (1979). Applying experimental research on group problem solving
to organizations. Journal of Applied Behavioral Science. 15. 375-391.

171

Homnan, L.R, & Maier, N.RF. (1961). Quality and acceptance of problem
sohntions by members of homogeneous and heterogeneous groups. I ouma1 of Abnormal

and Social Pachology, 62, 401-407.

Homnan, L.R, Harburg, E., & Maier, N.RF. (1962). Differences and
disagreements as factors inn creative problem solvinng. Jouraal of Abnormal and Social

Psgchology, 64, 206-214.

Hollenbeck, J.R, Sego, D.J., Ilgen, D.R, Major, M.A, Hedlund, J., & Phillips. J.
(in press). Team decision making accuracy under diﬁcult conditions: Construct
validation of potential manipulations and measures using the TIDE2 simulation. In M. T.
Brannick, E. Salas, & C. Prince (Eds), Team performance assessment and measurement:
Theog, methods and applications. Hillsdale, NJ: Erlbaum

 

Hollenbeck, J.R, Ilgen, D.R, Sego, D.J., Hedhmd, 1, Major, D.A, & Phillips, J.
(1995). Multilevel theory of team decision making: Decision performance in teams
incorporating distributed expertise. Journal of Applied Psychology, 80, 292-316.

Howard, A (1995). A framework for change. In A Howard (Ed), The changm' g
nature of work. San Francisco: Jossey-Bass.

Hudson, N. (1995, February 27). Pentagon: Goals met, but quality is suffering.
Air Force Timss.

Hyde, J.S., Fennema, E., & Lamon, S. (1990). Gender differences inn mathematics
performance: A meta-analysis. Pszchological Bulletina107. 139-155.

Ilgen, DR (1994). Jobs and roles: Accepting and copinng with the changing
structure of organizations. In M. G. Rumsey, & C.B. Harris (Eds), Selection and
classiﬁcation. Hillsdale, NJ: Lawrence Erlbaum Associates.

Ilgen, D.R, & Youtz, MA. (1986). Factors affecting the evaluation and
development of minorities in organizations. Research i_nn PersonneL and Human Resource
Management, 4, 307-337.

Izraeli, D.N. (1983). Sex effects or structural effects? An empirical test of
Kanten‘s theory of proportions. Social Forces, 61. 153- 165.

Jackson, SE. (1992). Team inn organizational settings: Issues inn managing an
increasingly diverse work force. In S. Worchel, W. Wood, & J .A Simpson (Eds), Group
process and producm. Newbury Park: Sage Publications.

Jago, AG., & Broom, V.H. (1982). Sex differences in inncidence and evaluation
of participative leader behavior. Journal of Applied Pachology, 67, 776-783.

172

Johnson, RA, & Schulman, G.I. (1989). Gender-role composition and role
entrapment in decision-making groups. Gendeaﬂd Society, 3. 355-372.

Johnston, W. (1991). Global work force 2000: The new world labor market.
Harvard Businness Revisw. 69. 115-127.

Kanter, RM. (1977). Some effects of proportions on group life: Skewed ratios
and responses to token women. American Jourmnl of Sociologx 82, 965-990.

Karabenick, S.A, Sweeney, C., & Penrose, G. (1983). Preferences for skill
versus chance-detemnined activities: The inﬂuence of gender and task sex-typing. Journal
of Research an Personam 17, 125-142.

Kent, RN., & McGrath, J.E. (1969). Task and group characteristics as factors
inﬂuencing group performance. Journal of Experimental Social Psgchology, 5, 429-440.

Kerr, N.L., & Sullaway, ME. (1983). Group sex composition and member task
motivation. Sex Roles 9, 403-417.

Korabik, K, Baril, G.L., & Watson, C. (1993). Managers' conﬂict management
style and leadership effectiveness: The moderating eﬂ‘ects of gender. Sex Roles, 29. 405-
420.

Kushell, E., & Newton, R (1986). Gender, leadership style, and subordinnate
satisfaction: An experiment. Sex Roles. 14. 203-209.

 

Lockheed, M.E., & Hall, KP. (1976). Conceptualizing sex as a status
characteristic: Applications to leadership training strategies. Journal of Social Issues 32
1 1 1-124.

 

Mabry, EA (1985). The effects of gender composition and task structure on
small group innteraction. Small Group Behavior. 1@ 7 5-96.

Mabry, EA (1989). Some theoretical implications of female and male interaction
in unstructured small groups. Small Group Behavior. 20. 536-550.

Maccoby, E.E., & Jacklin, C.N. (1974). The pgchology of sex diﬂ‘egences.
Stanford, CA: Stanford University Press.

Martin, P.Y., & Shanahan, KA (1983). Transcending the effects of sex
composition in small groups. In B.G. Reed and CD. Garvinn (Eds), Groapwork with
women/GroupwoLvLith m_en: An overview of gender issues in social gaoupwork practice.
New York: Hayworth.

Mamevski, ML. (1994). Understanding our differences: Performance in
decision-making groups with diverse members. Human Relations 47 531-552.

 

173

McGrath, J .E. (1976). Stress and behavior in organizations. In M.D. Dunnette
(Ed. ), Handbook of industrial and organizational pgchology. Chicago: Rand McNally.

McGuire, W.J., McGuire, C.V., & Winton, W. (1979). Effects of household sex
composition on the salience of one's gender inn the spontaneous self-concept. Journal of

Experimental Social Psychology, 15, 77-90.

Meeker, B.F., & Weitzel-O'Neill, RA (1977). Sex roles and innterpersonal
behavior inn task-oriented groups. American Sociological Reviewa42. 91-105.

Nemeth, C., Endicott, J., & Wachtler, J. (1976). From the ‘50s to the '70s:
Women in jury dehberations Sociomm 39, 293-304.

Ott, EM. (1989). Effects of the male-female ratio at work: Policewomen and
male nurses. Pachology of Women Qua—ﬁery, 13, 41-57.

Patterson, M.L., & Schaeﬂ‘er, RE. (1977). Effects of size and sex composition
on interaction distance, participation, and satisfaction in small groups. Small Group
Behavior 8 433-442.

 

Pazy, A (1986). The persistence of pro-male bias despite identical information
regarding causes of success. Organizational Behavior and Human Decﬁon Processes 38,
366-377.

Piliavin, J.A, & Martin, RR (1978). The effects of the sex composition of
groups on style of social interaction. Sex Roles 4, 281-296.

Reitan, HT, & Shaw, ME. (1964). Group membership, sex composition of the
group, and conformity behavior. Journal of Social Psychology, 64, 45-51.

Ridgeway, CL. (1982). Status inn groups: The importance of motivation.
American Sociological Review. 47. 76-88.

Robertson, P.J., & Kwong, SS. (1994). Decision making in school-based
leadership councils: The impact of council membership diversity. The UQan Review. 26.
41-54.

Rogelberg, S.G., & Rumery, SM. (1994). Gender diversity, team decision
quality, time-on-task and innterpersonal cohesion. Paper presented at the 9th Annual
Conference of the Society for Industrial and Organizational Psychology, Nashville,
Tennessee.

Rosenthal, SF. (1978). The relationship of attraction and sex composition to
performance and nonperformance experimental outcomes in dyads. Sex Roles 4 887-
898.

 

174

Ruble, D.N., & Higgins, ET. (1976). Effects of group sex composition on self-
presentation and sex-typing. Journal of Social Issues 32, 125-132.

Ruble, T.L., Cohen, R, & Ruble, D.N. (1984). Sex stereotypes: Occupational
barriers for women. American Behavioral Scientist 27 339-356.

 

Schein, V.E. (1973). The relationship between sex role stereotypes and requisite
management characteristics. Journal of Applied chhology, 57, 95-100.

Sistnmk, F ., & McDavid, J.W. (1971). Sex variables inn conforming behavior.
Journal of Personﬂﬂ' and Social Psychology, 17, 200-207.

Skvoretz, J. (1988). Models of participation in status-differentiated groups.
Social Pachology Qaarterbg, 51, 43-57.

Smith-Lovin, L., & Brody, C. (1989). Interruptions in group discussions: The
effects of gender and group connposition. American Sociological Review, 54, 424-435.

 

South, B.G. ( 1927). Some psychological aspects of committee work. Journal of
Applied chhology, 11, 437-464.

South, S.J., Bonjean, C., Markham, W., & Corder, J. (1982). Social structure and
inntergroup interaction. American Sociological Review. 41 587-599.

Spangler, E., Gordon, M.A., & Pipkin, RM. (1978). Token women: An
empirical test of Kanter’s hypothesis. American Joml of Sociology, 84, 160-170.

Spence, J.T., Helmreich, R, & Stapp, J. (1975). Likability, sex-role congruence
of interest, and competence: It all depends on how you ask. Journal of Applied Social
chhology, 5, 93-109.

Stasser, G., Taylor, LA, & Hanna, C. (1989). Information sampling inn
structured and unstructured discussions of three- and six-person groups. Journal of
PersoaM’ and Social Psgchology, 57, 67-78.

Steiner, ID. (1972). Group process and produMy' ' . New York, NY: Academic
Press.

Strodtbeck, F .L. & Mann, RD. (1956). Sex role differentiation in jury
dehberation. Sociomm 19, 3-11.

Swanson, M.A, & Tjosvold, D. (1979). The eﬂ‘ects of unequal competence and
sex on achievement and self-presentation. Sex Roles 5 27 9-285.

 

175

Thameling, C.L., & Andrews, RH (1992). Majority responses to opinion
deviates: A comnumicative analysis. Small Group Research, 23, 475-502.

Thune, E.S., Manderscheid, RW., & Silbergeld, S. (1980). Status or sex roles as
determinants of interaction patterns in small mixed-groups. Journal of Soci_al Psychology,
l_12_, 51-65.

Timmons, W.M. ( 1941). Sex diﬂ‘erences in discussion. Speech Monoggaphs, 8,
68-7 5.

Toder, N.L. (1980). The eﬂ‘ects of the sexual composition of a group on
discrimination against women and sex-role attitudes. Psychology of Women Quaraerbz, 5,
292-3 10.

Triandis, HC., Hall, E.R, & Ewen, RB. (1965). Member heterogeneity and
dyadic creativity. Human Relations 18, 33-35.

Tuddenham, RD., MacBride, P., & Zahn, V. (1958). The innﬂuence of sex
composition of the group upon yielding to a distorted norm Journal of Psychology, 46,
243-25 1.

Ward, NJ. (1991). Occupational suitability bias for fnnll-time and part-time
employment in sex-typed jobs. Sex Roles. 25. 81-89.

Webber, R (1976). Perceptions and behaviors in mixed sex work teams.
Industrial Relations 15 121-129.

 

Wentworth, D.K, & Anderson, LR (1984). Emergent leadership as a function
of sex and task type. Sex Roles 11, 513-524.

Wharton, AS., & Baron, IN. (1987). So happy together? The impact of gender
segregation on men at work American Sociological Review, 52, 574-5 87.

Wiley, M., & Eskilson, A (1982). Copinng in the corporation: Sex role
constrainnts. Journal of Applied Social Psichology, 12, 1-11.

Wiley, M., & Eskilson, A (1983). Scaling the corporate ladder: Sex differences

inn expectations for performance, power and mobility. Social Psychology Qaarterbg 46,
3 5 1-3 59.

Wood, W. (1987). Meta-analytic review of sex differences in group performance.
chhologjcal Bulletm' , 102, 53-71.

Wood, W., & Karten, SJ. (1986). Sex differences in interaction style as a
product of perceived sex diﬂ‘erences in competence. Journal of Personalﬂ' and Social

Psichology, 50, 341-347.

176

Wood, W., Polak, D., & Aiken, C. (1985). Sex differences in group task
performance. Journal of Personam and 80% Psychology, 48, 63-71.

 

Yerby, J. (197 5). Attitude, task, and sex composition as variables affecting
female leadership in small problem-solving groups Speech Mommas 42, 160- 168.

Ziller, RC. (1973). Homogeneity and heterogeneity of group membership. In
C.H. McClintock (Ed ), Experimental social psichology. New York: Holt, Rinehart &
Winston.