 

 

 

‘Ir

 

 

v”,
t) :9.
195:6;
. {2'14

1

~—4

I
. an;

\1
‘1
.‘1‘

a“
on-

""353"

‘41
1,.
2."
Lg»,

34w...
‘Ah’l
. .1.
. 4.9»:
1 . Pu . .
uxﬁwmwww...
“.0"- .DD. .

rim...” . . i .

I'lfalo A
.. 1!:4
00

.Ill“:
1!!

.9 5.15.1
lb
;V|r. up .
t I!!!

i4»...

Mm

f-I:
, C

.. . {1,3 H.
i? .4}. | .uanrdﬁ... .
n tit-rntcr!‘ I.‘
guWEhnln 1A" (H... ‘ . ‘ . “H
. . . MCI“- ul . {turd-WWW!” ‘vﬂWIPf- A v0

L #4,: -- ....c.. Wu... .1 2

ll .

Trims

( g ,)
'\

lllllllllll llllllll'llllll'll Illllllllll

31293 01712 9267

This is to certify that the

dissertation entitled

THE EFFECT OF TASK TYPE ON THE RELATIONSHIP
BETWEEN EFFICACY BELIEFS AND PERFORMANCE

presented by

Sandra Elaine Moritz

has been accepted towards fulﬁllment
of the requirements for

Ph.D. Kinesiology

degree in

 

 

Major professor 2

MM \3; 1795

M511 is an Afﬁrmative Action/Equal Opportunity Institution 0- 12771

 

 

 

' [LIBRARY

[ Michigan‘sme
University

 

 

PLACE IN RETURN BOX

 

 

 

 

to remove this checkout from your record.
TO AVOID FINES return on or before date due.
DATE DUE , DATE DUE DATE DUE
A23 1 3 1:9
JlngliE’AiﬂﬂQ 7 T‘ i T

 

 

 

 

 

 

 

 

 

 

THE EFFECT OF TASK TYPE ON THE RELATIONSHIP BETWEEN
EFFICACY BELIEFS AND PERFORMANCE
By

Sandra Elaine Moritz

A DISSERTATION
Submitted to
Michigan State University

in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Kinesiology

1998

ABSTRACT
THE EFFECT OF TASK TYPE ON THE RELATIONSHIP BETWEEN
EFFICACY BELIEFS AND PERFORMANCE
By

Sandra Elaine Moritz

In sport, efficacy beliefs are regarded as one of the more important factors affecting an
individual’s motivation and performance. The purpose Of this dissertation was to examine
the relationships among self-efﬁcacy, collective efﬁcacy, and individual and team
performance in a less interdependent (additive) versus a more interdependent bowling
task. A second purpose of the dissertation was to determine if the relationships between
eﬁicacy and performance differed according to the level of analysis (i.e., individual vs
team). The participants comprised 250 students who were enrolled in bowling classes.
They were randomly assigned to two person teams. As expected, task type moderated the
relationship between collective efﬁcacy and performance, whereby collective efﬁcacy
was a signiﬁcant predictor of performance in the interdependent condition but not for the
additive condition. This ﬁnding, however, was dependent on the efﬁcacy and
performance measures used, and was only evident for the ﬁrst time period. The results
were the same for both levels of analysis. Task type did not moderate the relationship

between self-efﬁcacy and performance at either level of analysis.

To my family:

Mom, Dad, Leslie, Brother Dean, Brother-In-Law Dean, Krista, Stephanie and Darren

“Only one thing is ever guaranteed,
that is you will deﬁnitely not achieve the goal if you don't take the shot. ”

Wayne Gretzky

ACKNOWLEDGMENTS
I owe my efﬁcacy to a number of people...

To Deb Feltz. As my program director, and the chair of my dissertation committee, you
brought out the best in me. Thank you for your inﬂuence, and your example. I see you in
myself and in my ways. I appreciate the positive characteristics you passed on to me.

To Dan Ilgen, Steve Kozlowski, and Steve Raudenbush. When I put my dissertation
committee together, I *shot for the stars.* I am pleased that you all agreed to help me
through this process, and this dissertation beneﬁted from your thoughtful comments and
suggestions. I never thought about “Levels of Analysis” issues until I took Dan and Steve
K.’s class. And look what happened! Thank you all so much.

To Eva. We made it through yet another degree! We are ﬁnally graduating -- Grade 21 in
Gym! Over the past 5 years, our relationship has evolved personally, and academically. I
truly love you like a sister.

To Victoria. I am so happy that we met! Its not often that you met a person and you know
that you will be friends with that person forever. I feel that way with you. Keep singing
our favorite Billy Joel tune.

To Shannon. I don’t think that there is anything that you don’t know about me. One of
the best qualities of our friendship is the unconditional support that we gave, and will
keep giving to each other.

To Martin. Based on the way our relationship started in Denver, I never expected that we
would be where we are today. Your wit, charm, and other “things” and “stuff” are
incomparable. You’ve proven that “regression to the mean” is just a statistical term.

To My Grad School Friends -- Kelly, Karin, Roop, Barry and Geffrey -- thanks for the
GrEaT times. I will miss you guys.

To Verna and Jan. Thanks so much for all your help over the years. Both of you made
alot of things easy for me.

There are so many other people who deserve recognition: Bob Malina, Craig Hall, John
Haubenstricker, Paul Vogel, Ralph Levine, and Kathleen Martin, and Carl Watson have
all inﬂuenced me in ways that can only make me a better person, teacher and researcher.

And To My Family. I’ve always done my own thing, and you have always been there for
me. It was hard for me to move to the “other side” of the world; I thought that I would
miss out on so much. I know now that wherever I go, I am never too far away from any
of you. I thank you all for talking to me when I was down, and celebrating with me when
I was happy. And for loving me in the special ways that only you do.

iv

TABLE OF CONTENTS

 

LIST OF TABLES .................................................................................................. viii
LIST OF FIGURES ................................................................................................ ix
1. INTRODUCTION ............................................................................................... 1
Statement Of the Problem ............................................................................ 1
Nature of the Problem ................................................................................. 1
Purpose and Hypotheses .............................................................................. 12
11. REVIEW OF THE LITERATURE ..................................................................... 14
Self-Efﬁcacy ................................................................................................ 15
Sources of self-efﬁcacy ................................................................... 15
Self-efﬁcacy and performance ........................................................ l6
Collective Efﬁcacy ...................................................................................... 18
Deﬁnitions of collective efﬁcacy .................................................... 19
Differentiation of collective efﬁcacy from other constructs ........... 20
Differences between collective efﬁcacy and selféefticacy .............. 21
Sources of Collective Efﬁcacy .................................................................... 23
Collective Efﬁcacy and Performance .......................................................... 24
Collective Efﬁcacy Research ....................................................................... 26
Collective efﬁcacy research - Outside of sport ............................... 30
Collective efﬁcacy research - Sport studies ...................................... 34
Measurement of Collective Efﬁcacy ............................................................ 42
Levels Issues ................................................................................................. 52
Task Type as a Moderator of the Efﬁcacy - Performance Relationship ...... 52
III. METHOD .......................................................................................................... 60
Overview ...................................................................................................... 6O
Participants and Task ................................................................................... 62
Measures ...................................................................................................... 64
Performance measures .................................................................... 64
Self-efﬁcacy questionnaires ............................................................ 65
Collective Efﬁcacy questionnaires ................................................. 65
Team Concept and incentive measures ........................ ‘ ................. . 66
Procedure .................................................................................................... 67

IV. RESULTS ......................................................................................................... 69

Descriptive Statistics for Measures ............................................................. 69

Moderator Analyses .................................................................................... 77

Time 1: Individual level ................................................................... 77

Time 1: Team level ......................................................................... 80

Time 2: Individual level ................................................................... 84

Time 2: Team level .......................................................................... 84

Past Performance ............................................................................. 89

V. DISCUSSION ..................................................................................................... 91
APPENDICES ............................................................................... 100
A. Pilot Study ........................................................................................ 101

Table A1: Pilot Study: Individual Level: Means, standard
deviations, and ranges for self-efﬁcacy, collective efﬁcacy

and individual performance for the additive and

interdependent groups ........................................................... 102

Table A2: Pilot Study: Individual Level: Correlations
between self-efficacy, collective efficacy, and performance

for the additive and interdependent groups .......................... 103
B. Bowling Instructions ........................................................................ 104
C. Performance Measures / Tournament Scoresheet ............................. 106
D. Self-efﬁcacy Questionnaires ............................................................. 108
E. Collective Efﬁcacy Questionnaires ................................................... 111
F. Team Concept and Incentive Measures ........................................... 116
G. University Approval for Use of Human Subjects ............................ 118
H. Consent Form ................................................................................... 120
I. Summary of Reduced Sample Analysis for Time 1 ................... 122
II.

Table 11: Reduced Sample: Individual Level: Means,

Standard Deviations, and Ranges for Self-Efﬁcacy,

Collective Efﬁcacy and Individual Performance

for the Additive and Interdependent Groups ....................... 123

Table 12: Reduced Sample: Individual Level: Correlations
Between Self-Efﬁcacy, Collective Efﬁcacy, and
Performance for the Additive and Interdependent Groups .. 124

Table 13: Team Level: Means, Standard Deviations,

and Ranges for Self-Efﬁcacy, Collective Efﬁcacy and

Team Performance for the Additive Group and

Interdependent Groups ........................................................ 125

vi

Table 14: Reduced Sample: Individual Level: Moderation
Analysis between Efﬁcacy and Performance (Links A
and B) .................................................................................... 126

Table 15: Reduced Sample: Team level: Correlations

among Aggregated Self-Efficacy, Collective Efﬁcacy

and Team Performance for the Additive and Interdependent
Groups ................................................................................... 127

Table 16: Reduced Sample: Team Level: Moderation
Analyses Between Efﬁcacy and Performance
(Links C and D) .................................................................... 128

LIST OF REFERENCES ......................................................................................... 129

vii

LIST OF TABLES

TABLE
1. Summary of Collective Efficacy Research in Chronological Order ........... 27
2. Individual Level: Means, Standard Deviations, and Ranges for

10.

ll.

12.

13.

Self-Efficacy, Collective Efﬁcacy and Individual Performance
for the Additive and Interdependent Groups .............................................. . 71

Individual Level: Correlations Between Self-Efﬁcacy, Collective
Efﬁcacy, and Performance for the Additive Group ...................................... 72

Individual Level: Correlations Between Self-Efﬁcacy, Collective
Efﬁcacy, and Performance for the Interdependent Group ............................ 73

Team Level: Means, Standard Deviations, and Ranges for Self-Efﬁcacy,
Collective Efﬁcacy and Team Performance for the Additive Group ........... 74

Team Level: Means, Standard Deviations, and Ranges for Self-Efﬁcacy,
Collective Efﬁcacy and Team Performance for the Interdependent Group...75

Means and Standard Deviations for Team Concept and Incentive Measures
for Task Type for Time 1 and Time 2 ........................................................... 76

Time 1: Individual Level: Moderation Analysis between Efﬁcacy
and Performance (Links A and B) ................................................................. 79

Time 1: Team level: Correlations among Aggregated Self-Efﬁcacy,
Collective Efﬁcacy and Team Performance for the Additive and
Interdependent Groups .................................................................................. 82

Time 1: Team Level: Moderation Analyses Between Efﬁcacy
and Performance (Links C and D) ....................................... g .......................... 83

Time 2: Individual Level: Moderation Analysis Between Efﬁcacy
and Performance (Links A and B) ................................................................ 86

Time 2: Team level: Correlations among Self-Efficacy, Collective
Efﬁcacy and Team Performance for the Additive and Interdependent
Groups ........................................................................................................... 87

Time 2: Team Level: Moderation Analysis Between Efficacy and
Performance (Links C and D) ...................................................................... 88

viii

LIST OF FIGURES
FIGURE

1. Overview: The Relationship Among Efficacy and Performance ............... 19

CHAPTER I

INTRODUCTION

S_tat_ement of the Problem

To date, no research has investigated the relationships among aggregated self-
efﬁcacy, collective efﬁcacy and performance with independent-type sports (e.g., bowling,
golf, gymnastics) or compared interdependent and independent-type activities in terms of
these relationships (Little & Madigan, 1997). Therefore, the purpose of this dissertation
was to examine the relationships among aggregated self-efﬁcacy, collective efﬁcacy, and
team performance in a less interdependent (additive) versus a more interdependent task.
Based upon theory and past research, it was hypothesized that task type would moderate
the relationship between efﬁcacy beliefs and performance.

The second purpose of this study was to determine if the relationships among
efﬁcacy and performance differed according to the level of analysis. Researchers have
suggested that self-efficacy, collective efﬁcacy and performance may comprise a multi-
level model (e.g., Lindsley, Brass, & Thomas, 1995, Prussia & Kinicki, 1996; Silver &
Bufanio, 1996). At the individual level of analysis, the relationships among self-efﬁcacy,
(disaggregated) collective efﬁcacy and individual performance were examined. At the
team level of analysis, the relationships among aggregated self-efﬁcacy, collective
efﬁcacy, and team performance were considered
Nature of the Problem

Self-efﬁcacy theory has provided the impetus for a number of research studies

across psychological domains due primarily to its signiﬁcance in affecting people’s

motivation and performance (Bandura, 1986, 1997; Feltz, 1988, 1992; Gist & Mitchell,
1992; Maddux, 1995). Self-efﬁcacy is deﬁned as one's belief in one's ability to
successfully perform a speciﬁc behavior or set of behaviors which are required to obtain
a certain outcome (Bandura, 1977, 1986). Thus, self-efﬁcacy is a belief in one’s
capabilities to produce given levels of performance rather than a global trait that accounts
for overall performance optimism (Bandura, 1997).

The relationship between self-efﬁcacy and performance has been studied almost
exclusively at the individual level despite the fact that in many sports, individuals
perform as members of a team rather than as individual competitors (F eltz, 1992).
Researchers have suggested that the self-efﬁcacy and performance relationship can be
extended from the individual level to the team (or group) level where collective efﬁcacy
is recognised as an important determinant of team performance (Bandura, 1986, 1997).

There has been little development of collective efﬁcacy, and much less is known
about the relationship between collective efﬁcacy and performance. Only recently has
Bandura elaborated upon the collective efﬁcacy construct, and more fully deﬁned it as “a
group’s shared belief in its conjoint capabilities to organize and execute the courses of
action required to produce given levels of attainments” (Bandura, 1997, p.477). The most
obvious difference between self-efﬁcacy and collective efﬁcacy is that the former refers
to people’s judgments of individual capabilities and effort; whereas the latter refers to
people’s judgments of team capabilities (F eltz & Chase, 1998). That is, self-efﬁcacy is an
individual level phenomena, while collective efﬁcacy is conceptualized as a team level

attribute.

Although Bandura (1997) asserted that collective efﬁcacy is rooted in self-
efﬁcacy, research has shown that the relationship is modest (cf Watson & Chemers,
1998). That self-efﬁcacy and collective efﬁcacy are related, but are also distinct
judgments, has implications for researchers interested in these constructs. Most of the
self-efficacy research has focused primarily on an individual level, and as such, it can
provide only a partial guide to the antecedents and consequences of collective efﬁcacy.
Collective efﬁcacy is a team based construct, and therefore, must be considered on the
team level. The issue of what differences might occur at the team level has not been
resolved (Porter, 1992).

The conceptualization of efﬁcacy as a team based construct raises a number of
methodological and measurement issues. In the past, researchers generally accepted that
collective efﬁcacy referred to people’s perceptions about the collective as a whole, but it
was not clear whether these beliefs were shared, and therefore represented a team level
rather than an individual level phenomenon (Zaccaro, Blair, Peterson, & Zazanis, 1995).
Recently, however, Bandura (1997) has clearly stated that collective efﬁcacy refers to
shared beliefs and is a team level attribute. Thus, researchers should measure this
construct in accordance with this conceptualization.

There are a number of ways to assess collective efﬁcacy at the team level. For
example, individuals can make ratings of the team’s collective efﬁcacy, or the team can
make collective efﬁcacy judgments together. When the former method is used,
researchers typically aggregate individual level perceptions to represent the higher level
construct. However, in order to justify the aggregation, it is ﬁrst necessary to examine the

degree of consensus (or agreement) at the individual level (Feltz & Lirgg, in press; Gully,

Devine, & Whitney, 1995; James, Demaree, & Wolf, 1984; Kozlowski & Hattrup, 1992).
Only when there exists an acceptable level of consensus can individual data be
aggregated to form a team level construct. Failure to consider consensus when
aggregating data on one level to represent a higher level of analysis may result in an
aggregation bias (James, 1982).

The second method that can be used to assess collective efﬁcacy requires that the
team members make the judgment together. This approach eliminates the calculation of
statistical indicators of agreement thereby avoiding aggregation issues. It is believed by
some (e.g., Guzzo et al., 1993) to be an acceptable way of measuring a team’s collective
belief. However, this method also has serious drawbacks. For example, according to
Bandura (1997), team members are rarely of one mind in their appraisals of matters. He
pointed out that forming a consensual judgment of a team’s efﬁcacy by team discussion
is subject to social persuasion and conformity pressures. For example, it may be possible
that a more prestigious team member could inﬂuence the judgment in such a way that it
does not accurately represent the views of many of its members. Thus, team members
may publicly consent to a response without truly believing it (Guzzo et al., 1993). Forced
consensus may also be highly misleading, especially with a large collective.

In accordance with other researchers (e. g., Prussia & Kinicki, 1996), the teams in
this dissertation completed the collective efﬁcacy measures together, rather than
individually. The teams in this dissertation were dyads, and therefore social persuasion
issues and conformity pressures were likely very minimal. Using this method also
eliminated the need for a consensus analysis for the collective efﬁcacy measures.

However, it created the need to demonstrate that the dyads functioned as a team. More

speciﬁcally, measures were constructed to assess each team members’ perceptions that
they felt they were a team, that they thought they worked together, and that they thought
they were coordinating their efforts. Furthermore, in order to investigate collective
efﬁcacy at the individual level of analysis, the collective efﬁcacy scores needed to be
“disaggregated.” To disaggregate team level collective efficacy scores, the individuals
comprising a team are assigned the same score.

Levels of analysis issues have been ignored in sport psychology research. Most
researchers have focused either on the individual athlete or the sport team, seldom on
both. Although relationships between individual level variables may differ ﬁom team
level processes, it is likely that some individual relationships can be generalized to the
team level. It is impOrtant to note that hypothesized relationships at the team level of
analysis are not intended to replace individual level ﬁndings (George, 1990). More
speciﬁcally, George stated that team level relations are not necessarily more important
than those found at the individual level of analysis, but rather represent a different way of
looking at the same constructs.

Research in sport has demonstrated the existence of a relationship between self-
efﬁcacy and individual performance, and between collective efficacy and team
performance. A recent meta-analysis that investigated the relationship between self-
efﬁcacy and performance in sport settings found that the correlation between these
constructs was positive and moderate in size (r = .37) (Moritz, Mack, & Feltz, 1996). In
addition, although considerably less research has been conducted on collective efﬁcacy
beliefs as they relate to team performance in sport, a positive relationship generally has

been found in both ﬁeld and laboratory settings (F eltz & Lirgg, in press; Hodges &

Carton, 1992; Lichacz & Partington, 1996; Paskevich, 1995; Spink, 1990; Watson &
Chemers, 1998). Research from other psychological domains has also demonstrated that
a relationship exists between collective efﬁcacy and performance (e.g., Kerr, 1989;
Parker, 1994; Prussia & Kinicki, 1996; Riggs & Knight, 1994; Silver & Bufanio, 1996).

A number of researchers, in domains other than Sport, have suggested that self-
efﬁcacy, collective efﬁcacy and performance may comprise a multi-level model (e.g.,
Lindsley et al., 1995, Prussia & Kinicki, 1996; Silver & Bufanio, 1996). A multi-level
conception of eﬂicacy does not simply identify self-efﬁcacy as an individual level
phenomenon and collective efﬁcacy as a group level attribute (Moritz & Watson, 1998).
Rather, a multi-level model postulates the relationship between context, efﬁcacy and
behavior within and across levels.

To date, only two, unpublished studies have simultaneously explored self-
efﬁcacy, collective efﬁcacy and performance. In their multi-level analysis of collective
efﬁcacy, Watson and Chemers (1998) investigated the antecedents and consequences of
collective efﬁcacy in college basketball teams. Twenty-eight teams participated in the
study (336 individuals). Collective efﬁcacy and self-efﬁcacy were assessed twice during
the season. In order to demonstrate that collective efﬁcacy was a shared belief, the degree
of consensus was examined for each team. Data were subsequently analyzed using
Hierarchical Linear Modeling (HLM) techniques. l-ILM allows for the testing of data
with a hierarchically nested structure (i.e., individuals nested within teams). Only the
results related to efﬁcacy and performance are described here.

First, this study Supported the notion of collective efﬁcacy as a shared belief. A

high level of within team consensus on collective efﬁcacy was found. Second, although

considerable stability in collective efﬁcacy judgments were noted over two time periods,
the degree to which collective efﬁcacy judgments were shared and their antecedents
changed over time. More speciﬁcally at the beginning of the season, collective efﬁcacy
was less shared (i.e., within-team variance was larger compared to between-team
variance) than later in the season.

With respect to the relationship between self-efﬁcacy and collective efﬁcacy,
Watson and Chemers demonstrated that it depended on the average self-efﬁcacy of the
team. For teams with high average self-efﬁcacy, the relationship between players’ self-
efﬁcacy and perceptions of collective efﬁcacy was positive. For teams with low average
self-efﬁcacy the relationship between players’ self-efﬁcacy and collective efﬁcacy was
negative.

With respect to performance, Watson and Chemers found that very early
collective efﬁcacy judgments were persistent over time, and impacted overall team
performance. Previous performance was unrelated to collective efﬁcacy at the beginning
of a season. However, collective efﬁcacy judgments at the beginning of the season
predicted later collective efﬁcacy beliefs and overall team performance at season’s end.
On an individual level, stronger individual performances (as measured by the average
number of points a player scored per game) were predicted by high self-efﬁcacy but not
by collective efﬁcacy.

In the set>ond study, the relationship between self-efﬁcacy, collective efﬁcacy and
performance was examined by Porter (1992). The participants in this study were 35 work
teams (150 individuals) from a manufacturing company. Performance was the total

number of parts assembled by the team each day. In order to aggregate individual self-

efﬁcacy and collective efﬁcacy ratings to the team level, consensus indexes were
estimated using the measures suggested by James et a1. (1984). Porter found that neither
aggregated self-efﬁcacy or aggregated collective efﬁcacy predicted team performance in
a hierarchical multiple regression analysis, although the constructs were correlated with
each other.

Other researchers have suggested that the relative predictiveness of self-efﬁcacy
and collective efﬁcacy may vary depending on the degree of interdependence in the
production of team attainments (cf Bandura, 1997). That is, collective efﬁcacy may be
especially relevant to team performance when team attainments require high
interdependent effort because members’ beliefs about the team encompass the
coordinative and interactive dynamics that operate within a team. Furthermore, self-
efﬁcacy may be adequate when team attainments represent largely the summed
contributions of individual members. Thus, collective efﬁcacy scores may be more
relevant in sports requiring high levels of interdependence than in sports with little or no
task interdependence (e.g., additive) (Feltz & Lirgg, in press; George & Feltz, 1995). For
sports with little interdependence (e.g., additive tasks), aggregated self-efﬁcacy should be
more predictive of team performance.

The model that guided this dissertation is illustrated in Figure l. The ﬁgure is
separated according to the level of analysis (i.e., individual and team). The bottom
portion illustrates the individual level relationships, and the top portion shows the team
level relationships. For both levels of analysis, task type is shown as a moderating

variable. Each link in the model is discussed below.

 

 

 

 

 

 

 

 

    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

gage... .. ................ .mmﬁo }
332.22. , m 832835 emo_em._-.._mw

" u

_ _

. _

_m>m._ 3:222: " u
.

n ....... " oa>k "
.............. “I.-- - ..-i--..-.....i”...:..: xmm... limit-..
_o>o._ Eco... " . u

u u

_ _

_ _

< - a

8.385th A _ >825 >omo_=m.:mm.
Emma. a _ 9:02.00 8886?.

 

ooﬁgoﬁom use moonm maeE< aEmcouﬂom 2:. ”33.550

Ea a

The relationship between self-efﬁcacy and individual performance at the
individual level of analysis is represented by Link A. There is considerable theoretical
(e.g., Bandura, 1986, 1997) and empirical (cf Moritz et al., 1996) support for this
relationship. It is unlikely that the type of task a person is performing (in terms of its
interdependency) would affect the self-efﬁcacy - individual performance relationship.
More speciﬁcally, an individual’s efﬁcacy beliefs regarding his/her performance should
be the same regardless of whether another person’s performance scores were being added
to his/her scores (i.e., an additive task), or if one was dependent on someone else’s
performance (i.e., an interdependent task). Similarly, the relationship between
disaggregated collective efﬁcacy and individual performance (Link B) is not expected to
be affected by task type. There is no theoretical justiﬁcation to predict a moderator effect
for task type on the relationship between disaggregated collective efﬁcacy and individual
performance.

At the team level of analysis, task type was expected to affect the relationship
between aggregated self-efﬁcacy and team performance (Link C), and the relationship
between collective efﬁcacy and team performance (Link D). Support for the differential
predictiveness of aggregated self-efﬁcacy and collective efﬁcacy beliefs based on task
interdependence in a sport setting is provided by Feltz and Lirgg (in press). They
examined the relationships among aggregated self-efﬁcacy, aggregated collective
efﬁcacy and team performance across an entire season of intercollegiate ice hockey.
Results of this study revealed that aggregated team and aggregated self-efﬁcacy were
positively correlated (; =.60), and that team performance was correlated with aggregated

collective efﬁcacy (1' =30) but uncorrelated with aggregated self-efﬁcacy (r =.03). Their

10

analysis revealed that aggregated collective efﬁcacy was the strongest predictor of team
performance, while aggregated self—efﬁcacy was signiﬁcant, but in the negative direction.

The F eltz and Lirgg (in press) study was not able to test whether these
relationships would hold for teams when performance was less interdependent. Testing
the moderating effect of task-type on the relationship between efﬁcacy beliefs and
performance poses some challenging issues. A number of factors such as team size,
motor skill classiﬁcation, and level of difﬁculty need to be controlled.

Deciding on which task to use in this dissertation was very challenging. I wanted
to manipulate interdependence, but also control for as many factors as I could (i.e., team
size, motor skill classiﬁcation, level of difﬁculty). For example, there is no degree of
interdependence in singles’ tennis, and there is a higher degree of interdependence in
doubles’ tennis. However, the shift from singles’ to doubles’ is confounded with group
size (1 player to 2 players). In some sports, there are inherent variations of
interdependence with respect to the skills needed to perform the task. For example, in ice
hockey, penalty shots have no interdependence, while most other aspects of the game are
higher in interdependence. Similarly, serving in volleyball is a skill that can be performed
independently, while all other offensive and defensive strategies demand coaction among
teammates. The problem with these sports is that as the degree of interdependence
needed to complete the skills changes, so to does the motor skill classiﬁcation. More
speciﬁcally, penalty shots, and serves are both closed skills, while the other aspects of the
game can be characterized as open skills. One may also argue that the different skills vary
in task difﬁculty. Scoring on a penalty shot is relatively easier that scoring when both

teams are playing at even strength.

11

I decided to use two bowling tasks for this study. Bowling was an ideal setting for
research of this kind. When played in its traditional format, it can be considered as an
additive task, whereby each individual team member contributes equally to team
performance. Under these conditions, interdependence among members is low, and team
performance is determined by summing the scores of individual team members. Bowling
can also be manipulated to be a conjunctive task (i.e., all team members must perform
some speciﬁc action which contributes to the product; Steiner, 1972). This modiﬁcation
results in a higher degree of interdependence among team members compared to the
additive task. In the interdependent condition, the two bowlers alternated between
bowling ﬁrst and second for each frame, like a “Scotch Bowling” activity (see Pangrazi
& Darst, 1997).

Firmse and Hymtheses

This dissertation extends the research on self-efﬁcacy, collective efficacy and
performance by examining task type as a potential moderator (see Figure l). A moderator
is a variable that affects the strength and direction of a relationship between an
independent variable (e.g., self-efﬁcacy, collective efﬁcacy), and a dependent variable
(e.g., performance) (Baron & Kenny, 1986). Moderation implies that a relationship
between two variables changes as a function of the moderating variable.

Task type was hypothesized to moderate the relationships between aggregated
self-efﬁcacy and team performance (Link C), and collective efﬁcacy and team
performance (Link D). Based on previous research and efﬁcacy theory, at the team level
of analysis, I hypothesized that collective efﬁcacy would be a more important predictor

of team performance compared to aggregated self-efﬁcacy for the interdependent task. I

12

did not expect to ﬁnd a relationship between collective efﬁcacy and team performance
for the additive task (Link D). With respect to Link C, I expected to ﬁnd that aggregated
self-efﬁcacy would be a stronger predictor of team performance in the additive task, but I
did not expect to ﬁnd a relationship between aggregated self-efﬁcacy and team
performance for the interdependent task. On an individual level of analysis, the
relationships between self-efﬁcacy and individual performance (Link A), and
disaggregated collective efﬁcacy and individual performance (Link B) were hypothesized
not to be moderated by task type.

A second purpose of this dissertation was to consider the efﬁcacy - performance
relationships within a multi-level model. Even though multi-level models may specify
relationships at one level of analysis that are suggested by empirical results or theory at
another level of analysis (Rousseau, 1985), it is not necessary for the individual and team
level relationships to be exactly similar. That is, it may be possible that a certain variable,
like task type, will affect the relationship at the team level, but will not affect the
relationship at the individual level. Consistent with this line of thought, I hypothesized
that task type would only be a moderator at the team level of analysis, but not at the
individual level of analysis. The relationships were tested at two time periods to

determine whether they would replicate. Therefore, “time” was not incorporated into

Figure l.

13

CHAPTER 11
REVIEW OF THE LITERATURE

Self-efﬁcacy theory has provided the impetus for a number of research studies
across psychological domains due primarily to its signiﬁcance in affecting people’s
motivation and performance (Bandura, 1986, 1997; Feltz, 1988, 1992; Gist & Mitchell,
1992; Maddux, 1995). To date, the relationship between self-efﬁcacy and performance
has occurred almost exclusively at the individual level despite the fact that in many
settings, such as business, military and sport, individuals perform as members of a team
rather than as individual competitors (F eltz, 1992). Researchers have suggested that the
self-efﬁcacy and performance relationship can be extended from the individual level to
the team level, and collective efﬁcacy has been recognised as an important determinant of
team performance (Bandura, 1986, 1997).

This chapter reviews relevant literature pertaining to self-efﬁcacy, collective
efﬁcacy and performance within sport settings. Although it was contextually relevant to
introduce efﬁcacy research from other domains, this inclusion was deemed as
complimentary to the examination of efﬁcacy and performance in sport, and therefore, in
no way, does this review attempt to cover all of the work on efﬁcacy theory. Rather, a
general discussion of the self-efﬁcacy construct, and its relationship with performance is
presented. Due to the limited quantity of research on collective efﬁcacy, this construct is
considered more fully. The collective efﬁcacy section draws parallels with self-efﬁcacy,
and also highlights the unique nature of collective efﬁcacy as a team construct. Next,

levels of analysis issues that affect team research are presented. I conclude with a multi-

l4

level model between efﬁcacy and performance. In this section, task type is introduced as
a moderator of the efﬁcacy - performance relationship.
SEEM

The construct of self-efﬁcacy was developed within the framework of social
cognitive theory (Bandura, 1977, 1986). Self-efﬁcacy is deﬁned as one's beliefs in one's
ability to successfully perform a speciﬁc behavior or set of behaviors which are required
to obtain a certain outcome (Bandura, 1977, 1986). Thus, self-efﬁcacy is a belief in one’s
capabilities to produce given levels of performance rather than a global trait that accounts
for overall performance optimism. Self-efﬁcacy is not concerned with the skills an
individual has, but rather with the judgments of what one can do with whatever skills
he/she possesses (Bandura, 1986). That is, self-efﬁcacy judgments are about what one
can do, not what one has.
Sources of Self-Efﬁcacy

Judgments of efﬁcacy, whether accurate or faulty, are based on six principal
sources of information (Bandura, 1977, 1982, 1986; Maddux, 1995). These sources are:
performance experiences, vicarious experiences, imaginal experiences, verbal or social
persuasion, physiological states, and emotional states. The six categories of efﬁcacy
information are not mutually exclusive in terms of the information they provide, although
some are more inﬂuential than others. Generally, as consecutive outcomes accumulate,
personal experience becomes the primary inﬂuence on efﬁcacy judgments, and possible
counteracting inﬂuences such as verbal persuasion and vicarious experience lose their
potency (Lindsley et al., 1995). Thus, the most important and most powerful source of

efﬁcacy is derived ﬁ'om previous performances. Generally, success at a task, behavior or

15

skill strengthens self-efﬁcacy, whereas perceptions of failure diminish self-efﬁcacy
(Bandura, 1977, 1982, 1986; Feltz, 1992; Maddux, 1995).
Self-Efﬁcacy and Performance

Bandura (1986) proposed a reciprocal relationship between self-efﬁcacy and
performance, According to Bandura (1977), “mastery expectations inﬂuence performance
and are, in turn, altered by the cumulative effects of one’s efforts” (p.194). Researchers
have supported this notion (Feltz, 1994). In sport performance, self-efﬁcacy has been
shown to predict initial performance; however, as one gains experience on the task,
performance also becomes a strong predictor of both ﬁlture performance and self-efﬁcacy
(Feltz, 1982; 1988; F eltz & Mugno, 1983; McAuley, 1985). For example, Feltz (1982)
found that as participants progressed over trials in learning a back dive, performance
became a stronger inﬂuence on self-efﬁcacy than self-efﬁcacy became on performance,
and Feltz (1982), and McAuley (1985) both found that performance correlates stronger
with post-performance self-efﬁcacy scores compared with pre-performance self-efﬁcacy
scores. In their meta-analyses, Moritz et a1. (1996) and Multon, Brown and Lent (1991),
also found that correlations for post-performance selfuefﬁcacy and performance measures
were signiﬁcantly larger that correlations between pre-performance self-efﬁcacy
measures and performance.

The importance of self-efﬁcacy in sport has been repeatedly demonstrated. For
example, in correlational studies, self-efﬁcacy has been shown to be an important
discriminating factor between “qualiﬁers” and “non-qualiﬁers” in Olympic gymnastics
(Mahoney & Avener, 1977), as well as between successful and unsuccessful Big 10

wrestlers (Gould, Weiss, & Weinberg, 1981). Furthermore, in recent meta-analyses, self-

16

efﬁcacy was shown to have a positive and moderate correlation with performance across
sport and exercise settings (Moritz et al., 1996). Finally, experimental and path analytic
studies have demonstrated self-efﬁcacy to be a major determinant of athletic performance
(e.g., Feltz, 1982, 1988; Feltz & Mugno, 1983; George, 1994; McAuley, 1985).

Self-efﬁcacy beliefs are a major determinant of behavior only when people have
sufﬁcient incentives to act on their self-perception of efﬁcacy and when they possess the
requisite subskills (Bandura, 1986). Discrepancies between efﬁcacy beliefs and
performance are most likely to occur when tasks or situations are ambiguous or when one
has little information on which to base efﬁcacy judgments, such as when one is learning a
skill (Bandura, 1977). Bandura stated that self-efficacy beliefs will exceed actual
performance when there is little incentive to perform the activity, or when physical and
social constraints are imposed on performance. An individual may have the necessary
skill and high self-efﬁcacy, but no incentive to perform. This ﬁnding does not mean that
self-efﬁcacy has no inﬂuence whatsoever in the absence of skills. On the contrary,
people’s beliefs in their ability to learn can inﬂuence skill development.

Most of the efﬁcacy research, to date, has assumed the presence of proper
incentives rather than assessing and controlling for this factor (F eltz, 1992). However,
there are methods available for the assessment of incentives (cf Feltz & Chase, 1998). For
example, one method that has been used for the assessment of incentives is a measure of
one’s perceived importance to perform well on the task (George, F eltz & Chase, 1992).
These researchers found that only when they used participants who perceived the task to

be at least moderately important did they ﬁnd a self-efﬁcacy - performance relationship.

17

Collective Efﬁcacy

To date, almost all of the efﬁcacy research has been focused on the individual
level. Bandura (1986, 1997) suggested that the efﬁcacy - performance relationship can be
extended from the individual level to the Collective level. There has been little
development of collective efﬁcacy, either as a theoretical construct or in terms of its
measurement (Paskevich, 1995; Zaccaro et al., 1995), and much less is known regarding
the relationship between collective efﬁcacy and performance although Bandura (1986,
1997) has suggested it is an important determinant of team performance. This lack of
attention and understanding is attributable to problems in initial conceptions of collective
efﬁcacy, its treatment as a mere extension of self-efﬁcacy theory to larger aggregations,
and the difﬁculty of conducting research on teams or teams (George & Feltz, 1995;
Zaccaro et al., 1995). Thus, more research is needed on efﬁcacy beliefs as they relate to
team performance in sport (Bandura, 1977, 1986; Feltz, 1992; Spink, 1990).

To better understand the concept of a team having a Collective efﬁcacy for various
skills, it is important to establish clear conceptual and operational deﬁnitions of teams. In
this dissertation, no distinction is made between the terms “group” and “team,” and the
terms are used interchangeably (cf Guzzo & Shea, 1992).A number of deﬁnitions for the
terms group or team exist, and these deﬁnitions are often accompanied by many different
ideas about the minimum set of criteria that constitute a team. Perceptions of “teamness,”
coordination, and working together have been identiﬁed as team characteristics (Zaccaro
et a1. 1995). A team is deﬁned in this dissertation as “a social unit consisting of a number
of individuals who stand in role and status relationships to one another, stabilized in some

degree of time, and who possess a set of values or norms of their own regulating their

18

behavior, at least in matters of consequence to the group” (Sherif & Sherif, 1969; p. 131).
Sport teams, by virtue of the rules associated with playing a team sport, the distinctive
player roles, and a common motivational goal of performance, display these
characteristics (Paskevich, 1995).

Deﬁnitions of Collective Efﬁcacy

Since the concept of collective efﬁcacy was ﬁrst introduced by Bandura in 1986,
there have been many ways in which the construct has been deﬁned. Thus, it is necessary
to determine what collective efﬁcacy is, and how it differs from other constructs. It was
until recently that Bandura elaborated upon the collective efﬁcacy construct, and more
fully deﬁned it as “a group’s shared belief in its conjoint capabilities to organize and
execute the courses of action required to produce given levels of attainment” (Bandura,
1997, p. 477). Like self-efﬁcacy, Bandura considers collective efﬁcacy to be task-
speciﬁc.

Similar operationalizations of collective efﬁcacy have been proposed by other
researchers. For example, Lindsley et a1. (1995) deﬁned collective efﬁcacy as the group’s
collective belief that it can successfully perform a speciﬁc task. Paskevich (1995) stated
that collective efﬁcacy is the perception of the team regarding a team skill, and that it is
reﬂected in the team’s belief in their ability as a team to produce desired results.

The work of Zaccaro et a1. (1995) reﬂects a relatively unique conceptualization of
collective efficacy. These authors believe that a deﬁnition of collective efficacy must
acknowledge the notion of collective coordination, and the integration of individual
contributions to the collective effort. They deﬁned collective efﬁcacy as “a sense of

collective competence shared among members when allocating, coordinating, and

19

integrating their resources as a successful, concerted response to speciﬁc situational
demands” (p.309). To Zaccaro and his colleagues, collective efﬁcacy refers to a team
member’s beliefs not only about how well each and every other team member can
marshal individual resources to accomplish the team task, but also how well team
members can coordinate and combine their resources.

In some deﬁnitions, collective efficacy has been deﬁned as a team member’s
judgment of team capabilities (see Feltz & Chase, 1998; George & Feltz, 1995; Weldon
& Weingart, 1993). For example, Bandura (1986) has also referred to collective efﬁcacy
as judgments that people make about a group’s level of competency. This differentiation
between team and individual beliefs has implications for collective efﬁcacy measurement
and subsequent research.

Differentiation of collective efﬁcacy from other constructs. Before discussing the
differences and similarities between collective efﬁcacy and self-efﬁcacy, it is important
to differentiate collective efﬁcacy from other team constructs. First, the concept of
collective efﬁcacy is not new to the organizational behavior ﬁeld (Shamir, 1990).
Collective efficacy has been a part of the old concept of group “morale” which was found
to enhance the performance of military units, small teams and organizations. The notion
of morale, however, was vaguely conceived and too broad. The concept of collective
efﬁcacy is analytically narrower and cleaner, and more suitable for inclusion in cognitive
models of motivation.

There are other constructs that warrant differentiation from collective efﬁcacy.
For example, Zander and Medow’s (1963) conceptualization of “group aspiration levels”

refer to a score, or a performance level that a team expects to achieve. It differs from

20

collective efﬁcacy in that the former are exact statements of performance goals, rather
than cognitive beliefs about a team’s capability. Another team construct is “collective
esteem,” which is operationalized as the extent to which individuals generally evaluate
their social group positively (Crocker & Luhtanen, 1990). Collective esteem refers to the
value of a team rather than a team’s expected effectiveness.

A construct very similar to collective efficacy is group potency (Guzzo, Yost,
Campbell, Shea, 1993; Guzzo & Shea, 1992). Group potency is deﬁned as a generalized
collective belief in a group that it can be effective across multiple tasks encountered by
groups in complex environments. Group potency differs from collective efficacy in that it
is a more global construct which encompasses all types of tasks in a number of situations.
Guzzo and Shea (1992) conceptualized group potency as an attribute of the group, and
advocated for group measures of the construct. According to these authors, this group-
focus served to differentiate potency from collective efficacy. The differences between
group potency and collective efﬁcacy are minimal, and seem to be more of an issue for
measurement. Measures of group potency are much broader in scOpe compared to the
task-speciﬁc focus of collective efficacy.

Differences between collective efﬁcacy and self-efﬁcacy. It is also worthwhile to
differentiate collective efficacy from self-efﬁcacy. The most obvious difference between
these constructs is that self-efficacy refers to people’s judgments of individual
capabilities and effort, whereas collective efﬁcacy refers to people’s judgments of team
capabilities. According to Bandura (1997), self-efﬁcacy and collective efﬁcacy differ in
the unit of agency. Self-efﬁcacy is an individual level phenomena, while collective

efﬁcacy exists as a team level attribute.

21

Bandura (1986, 1997) stated that collective efﬁcacy is rooted in self-efficacy.
More speciﬁcally, he stated that one’s knowledge of personal efﬁcacy is not unrelated to
perceived collective efﬁcacy. A team that has a strong sense of collective efﬁcacy can
enhance the perceived task-speciﬁc efficacy of its members, but a team with a weak
sense of collective efﬁcacy may not totally undermine the perceived self-efﬁcacy of its
more resilient members. Team losses, for instance, may decrease a team’s efﬁcacy levels,
but should not affect the individual efﬁcacies of its more resilient members. As well,
members of a team who have weak beliefs in their own individual capabilities are not
easily transformed into a strong collective force (Bandura, 1982, 1986).

Research has shown, however, that the relationship between collective efficacy
and self-efficacy is modest (Feltz & Lirgg, in press; cf Watson & Chemers, 1998). That
self-efficacy and collective efﬁcacy are related, but are also distinct judgments, has
implications for researchers interested in these constructs. Most of the self-efﬁcacy
research has focused primarily on an individual level, and as such, it can only provide a
partial guide to the antecedents and consequents of collective efﬁcacy. Collective
efﬁcacy is a team based construct, and therefore must be considered on a team level. The
issue of what differences might occur at the team level has not been resolved (Porter,
1992). Bandura (1997) noted that sociocognitive determinants operate in much the same
way at the collective level as they do at the individual level, and that both forms of
efﬁcacy serve similar functions and operate through similar processes. Thus, self-efﬁcacy
and collective efﬁcacy should share similarities in terms of their antecedents and

consequences (cf Feltz & Lirgg, in press; Zaccaro etal., 1995).

22

_So_urc_£s of Collective Efﬁcacv

Collective efﬁcacy is thought to be inﬂuenced by diverse sources of information
and determined in part by events and experiences similar to those that determine self-
efﬁcacy (Bandura, 1997; George & Feltz, 1995; Feltz & Chase, 1998; Feltz & Lirgg, in
press; Zaccaro et al., 1995). In addition, it is also likely that qualities of the collective as a
whole add a number of other variables to those associated with self-efﬁcacy (Zaccaro et
al., 1995). Expectations of collective efﬁcacy have been hypothesized to be derived from
the following ﬁve sources of information: performance accomplishments, vicarious
experiences, verbal persuasion, physiological states, and inﬂuences within the team
(Zaccaro etal., 1995).

As with self-efﬁcacy, mastery experiences of the team are predicted to be the
most powerful source of efficacy information (F eltz, 1994; Feltz & Chase, 1998; Feltz &
Lirgg, in press; George & Feltz, 1995; Zaccaro et al., 1995). Teams that have an
outstanding record of performance undoubtedly cultivate strong percepts of efﬁcacy
among their members. Likewise, a serious. performance failure could decrease the
collective efﬁcacy of its membership which, in turn, could inﬂuence subsequent failures
(Feltz & Lirgg, in press; see also Lindsley et al., 1995). In a recent study, Feltz and Lirgg
(in press) found that past team performance affected aggregated collective efficacy
beliefs to a larger degree than aggregated self-efﬁcacy beliefs. More speciﬁcally, they
found that although wins and losses affected aggregated collective efﬁcacy more than
aggregated self-efﬁcacy, losses affected aggregated collective efﬁcacy more in

comparison to aggregated self-efﬁcacy.

23

The pattern of prior performances has also been hypothesized to inﬂuence the
strength of collective efﬁcacy (Zaccaro et al., 1995). According to Zaccaro et al., team
members need to share a signiﬁcant number of performance experiences in order to
develop a coherent and consistent sense of collective efficacy.

Although research on the sources of collective efficacy is just starting to emerge,
one recent study suggests that the sources of collective and self—efficacy may differ
slightly (Chase, Feltz, Tully, & Lirgg, 1997). Using data from 43 women collegiate
basketball players prior to 12 games, these researchers investigated the frequency of
occurrence of ﬁve different efficacy sources: past performance, physiological/emotional
states, social comparison, outside sources, and persuasion. The results revealed that past
performance, and physiological/emotional states were both ranked as the most important
sources of self- and collective efﬁcacy. The only difference was that self-efﬁcacy was
inﬂuenced by sources outside of sport more than collective efﬁcacy, and that collective
efﬁcacy was inﬂuenced more by social comparison than self-efﬁcacy. More speciﬁcally,
individual players perceived events such as school demands as more of an inﬂuence on
their personal efﬁcacy than their team’s efﬁcacy. In addition, social comparisons (e.g.,
“We’re a stronger team”) were found to be more important than sources outside of sport
for collective efﬁcacy information. For both self- and collective efﬁcacy, persuasion was
ranked as the least used source for efﬁcacy information.

Collective Efﬁcacy and Performance

Like self-efﬁcacy, collective efficacy is hypothesized to inﬂuence what people

choose to do as a team, how much effort they put into their team endeavours, and their

staying power when collective efforts fail to produce quick results or encounter forcible

24

Opposition (Bandura, 1982, 1986, 1997). Zaccaro et a1. (1995) hypothesized that
highly efﬁcacious teams should be more likely to persist in the face of collective
difﬁculties and obstacles, and be willing to set more diﬂicult goals and be more
committed to these goals, as well as accept more difﬁcult challenges for the team. In
addition, high collective efﬁcacy should also facilitate a team’s responses to
environmental stress, promote persistence and perseverance in the face of signiﬁcant
demands, and may be linked to greater readiness for risk taking.

The bottom line with respect to collective efﬁcacy is that highly efﬁcacious teams
should perform better and persist longer than teams having lower collective efﬁcacy
(Feltz & Lirgg, in press; George & Feltz, 1995 ; Spink, 1990; Zaccaro et al., 1995).
George and Feltz (1995) stated that the relationship between collective efﬁcacy and team
performance is likely to be reciprocal. That is, similar to the relationship between self-
efficacy and performance, successful team performance should positively inﬂuence
percepts of collective efﬁcacy, which in turn are likely to lead to behaviors and actions
that enhance the ability of the team to succeed in the future.

According to Bandura (1997), the stronger the beliefs people hold about their
collective capabilities, the more they achieve. He stated that this is true regardless of
whether the team’s sense of efﬁcacy develops naturally or is created experimentally, and
that the relationship holds across domains (e.g., education, organizational, athletic).
Although little research has been conducted on collective efﬁcacy beliefs as they relate to
team performance in sport, in the studies that have been conducted, a positive

relationship has been found between collective efficacy and performance in both the ﬁeld

25

and laboratory (F eltz & Lirgg, in press; Hodges & Carton, 1992; Lichacz & Partington,
1996; Paskevich, 1995; Spink, 1990; Watson & Chemers, 1998).
Collective Efﬁcacy Research

Bandura (1982, 1986) called for a broad and comprehensive research effort with
respect to collective efﬁcacy. He stated that the greatest progress will be made in
elucidating the development, decline, and restoration of collective efﬁcacy and how it
affects team functioning; that is, if measures of perceived collective efﬁcacy are tied to
explicit indices of performance. To date, however, there are only a few studies that have
examined the collective efﬁcacy construct. The purpose of this section is describe this
research. Table 1 provides a brief summary of the collective efﬁcacy research. In this
table, only the relationships between self-efﬁcacy and collective efficacy, and collective

efﬁcacy and performance are noted.

26

Table 1. Summarv of Collective Efﬁcacy Research in Chronological Order.

 

 

 

 

 

 

 

 

Authors/ Setting Deﬁnition of Collective Type of Results
Date Efﬁcacy Study
Kerr, 1989 Social Rating of which group Exper. Under certain experimental
Psyc. was more likely to conditions, for which personal and
receive an investment collective efficacy do not decline,
payoff. people still presume that they do (no
values reported).

Spink , 1990 Sport Operationalized as a Corr. Individuals who were high in
construct which reﬂects collective efficacy were on teams that
the fact that teams oﬁen ﬁnished higher in the tournament than
have collective individuals who were in low
expectations for success collective efficacy.

Hodges & Sport Cited Bandura, 1986. Exper. Experimental triads high in collective

Carron, 1992 efﬁcacy improved their performance

on the muscular endurance task
following a failure experience,
whereas triads low in collective
efﬁcacy experienced a performance
decrement.

Teams with high efﬁcacy persisted
longer than low efﬁcacy teams
Both high and low efﬁcacy teams
decreased their collective efﬁcacy
following failure.

Porter, 1992 [/0 Extension of self- Corr. Neither aggregated self-efficacy or

(Unpublished efﬁcacy deﬁnition to aggregated collective efﬁcacy

Dissertation) team level predicted performance in regression

analysis.

[’5 between aggregated self-efﬁcacy
and performance ranged between —.05
- -.55

[’5 between aggregated collective
efﬁcacy ranged between -.59 - .17

Parker, 1994 Educ. Judgments that people Corr. [’5 between selfffﬁcacy and
make about a team’s collective efﬁcacy ranged ﬁom .60 -
level of competency. .93

[’5 between teacher’s collective
efﬁcacy and student performance
ranged between .29 - .61.

Riggs & [/0 The beliefs that Corr./ r between self-efﬁcacy and collective

Knight, 1994 individuals hold Path efficacy = .18
concerning the ability Analysis

 

 

of their team to
successfully perform
it’s work tasks

 

 

Performance not assessed.

 

27

 

 

 

 

 

 

 

 

Paskevich, Sport Deﬁned as a sense of Corr. Collective efﬁcacy mediated the
1995 collective competence relationship between cohesion and
shared among team performance, at early season, but not
members when later in the season.
allocating, Evidence for the independent effects
coordinating, and of both collective efﬁcacy and
integrating their cohesion on performance.
resources as a (No correlations listed)
successful, concerted
response to situational
demands.
Lichacz & Sport The judgment by team Exper. Correlation between the individual
Partington, members of the team’s perceptions of collective efﬁcacy and
1996 capabilities to organize team performance was .27.
and execute courses of
action required to attain High efﬁcacy conditions
designated types of demonstrated larger means for
performance. collective performance than the low
collective efﬁcacy teams
Prussia & [/0 Individual perceptions Exper. Correlations between team
Kinicki, 1996 regarding a team’s effectiveness and collective efﬁcacy
ability to perform in a were .60, and .74.
particular situation
Silver & 1/0 The collective belief in Exper. Correlations between efﬁcacy and
Bufanio, 1996 a team that is can be performance were .62, and .37.
successful.
Past performance correlated positively
with subsequent judgements of
collective efﬁcacy Q’s = .80 and .67)
Little & l/O Members’ judgments of Repeated Collective efﬁcacy was related to
Madigan, 1997 team capabilities, or an measures team performance
assessment of the ANOVA
team’s collective ability
to perform a job at hand
Watson & Sport Cited Bandura (1997) HLM Relationship between self-efﬁcacy

Chemers, 1998

 

 

 

 

and collective efﬁcacy depended on
the average self-efﬁcacy of the team
(for teams with high average self-
efﬁcacy, relationship was positive for
teams with low average self-efficacy,
relationship was negative).

Previous performance was unrelated
to collective efﬁcacy at the beginning
of the season. '

Collective efﬁcacy at beginning of
season predicted later collective
efﬁcacy beliefs and overall team
performance at season’s end.

 

28

 

 

 

Feltz & Lirgg,
in press

 

Sport

 

Degree of conﬁdence
that player’s had in
their team’s ability to
outperform their
opponents on important
game competencies.

 

Corr.

 

Team and player efﬁcacy were
positively correlated (r=.60),

Team performance was correlated
with collective efﬁcacy (r=.30) but
uncorrelated with player efﬁcacy
(r=.03).

 

29

 

Collective efﬁcacy research - Outside of sport. Those who have investigated
collective efﬁcacy outside of sport have found that collective efﬁcacy is related to group
size (Kerr, 1989); that teacher’s collective efﬁcacy is related to student achievement
(Parker, 1994); that success/failure inﬂuence collective efﬁcacy (Riggs & Knight, 1994);
and that collective efﬁcacy is related to group effectiveness (Prussia & Kinicki, 1996).
With respect to the relationship between collective efﬁcacy and performance, generally a
positive relationship (Little & Madigan, 1997; Silver & Bufanio, 1996), or equivocal
results have been reported (Porter, 1992).

For example, Riggs and Knight (1994) used path analysis to describe the
relationships between perceived group success-failure, individual and collective levels of
job-speciﬁc efﬁcacy and outcome expectancy, satisfaction, and organizational
commitment (one’s acceptance and willingness to work for organizational goals) with a
sample Of 59 work groups. Collective efﬁcacy was deﬁned as the beliefs that individuals
hold concerning the ability of their group to successﬁilly perform it’s work tasks, and as
such was an individually held belief about the group in which each participant belonged.

The results Of their path analysis demonstrated that group success-failure played a
direct, dominant role in the determination of self- (r =3 7) and collective efﬁcacy (r =
.56). Other results indicated that self-efﬁcacy failed to be linked with satisfaction (; = -
.09) or organizational commitment (.13 = -. 12), and that the links between collective
efﬁcacy and satisfaction, and collective efﬁcacy and organizational commitment were the
weakest in their model (r’s = .48, .64 respectively). With respect to the relationship

between self-efﬁcacy and collective efﬁcacy, a correlation of .18 was reported.

30

Prussia and Kinicki (1996) investigated how perceived colleCtive efﬁcacy
operates in concert with other sociocognitive determinants, namely group goals and
affective evaluative reactions, in determining group effectiveness. Their model predicted
that group affective evaluations, group goals, and collective efﬁcacy would mediate the
inﬂuences of performance feedback and vicarious experience on group effectiveness.
Collective efﬁcacy was deﬁned as individual perceptions regarding a group’s ability to
perform in a particular situation, and group effectiveness was evaluated as the process
behaviors and the output quantity of novel brainstorming tasks.

Four-person, concocted groups (r_r = 81) were lead to believe that their group
performed either above or below the normative productivity standard. In addition, half of
the groups viewed a brainstorming information video, while the other half did not (i.e., a
2 X 2 design). The results of this study found that feedback, vicarious experience, and
group effectiveness at Time 1 all affected collective efﬁcacy (the correlation between
group effectiveness and collective efﬁcacy was .60). With respect to feedback, as
expected, a positive relationship was found between collective efﬁcacy and positive
feedback, and a negative relationship was found between negative feedback and
collective efﬁcacy. In addition, collective efﬁcacy affected group goals (r = .80) and
group effectiveness at Time 2 (r = .74).

Upon further analysis, a number of mediating relationships were evident. For
example, collective efﬁcacy was found to mediate the relationship between group
eﬂ‘ectiveness and feedback, and also partially mediated the relationship between
vicarious experience and group effectiveness. Lastly, collective efﬁcacy was also found

to mediate the relationship between feedback and group goals.

31

Recently, Silver and Bufanio (1996) investigated the impact of collective efﬁcacy
and group goals on group performance. Collective efﬁcacy was deﬁned as the collective
belief in a group that it can be successful, and was assessed by having the groups make
judgments about their capabilities to complete a construction task (i.e., building model
trucks from Lego construction pieces). Twenty-ﬁve groups composed of undergraduate
students participated in this study. The groups completed the collective efﬁcacy and
group goal level measures prior to completing the task. The groups completed the task 3
times; one performance measure was used as a pre-trial measure. .

The results revealed that collective efﬁcacy was related positively to group
performance, on both experimental trials. The correlations between efficacy and
performance were .62, and .37, respectively. In addition, hierarchical regression analyses
showed that efﬁcacy accounted for a signiﬁcant proportion of the variance in group goals
beyond that explained by previous task performance. The total variance explained using
these two predictors was 60% and 77% for trials one and two, respectively. Past
performance was also correlated positively with subsequent judgments of collective
efﬁcacy for both trials (r’s = .80 and .67, respectively). Finally, to determine the
proportion of variance in task performance accounted for by past performance, collective
efﬁcacy, and group goals, multiple regression analyses were conducted on past
performance. Although the total variance accounted for by these 3 variables was 52% and
44% (for Trials 1 and 2), collective efﬁcacy and group goals added unique predicted
variance beyond that accounted for by past performance, but only for Trial 1.

A comment on the Silver & Bufanio study by Kaplan (1997) suggested that Silver

and Bufanio may have wrongly interpreted their results. More speciﬁcally, Kaplan noted

32

Silver and Bufanio hypothesized that the correlation of efﬁcacy and goals with
subsequent performance would become progressively stronger over time. Initial
performance, efﬁcacy, and goal measures did predict the performance scores for the ﬁrst
trial. However, by the next trial, the additional predictive value of efficacy and goals was
gone. That is, the variance in performance explained by past performance increased,
whereas the variance explained by efficacy and goals decreased. Rather than growing
stronger, as the hypothesized model implied, the correlation of efﬁcacy and goals with
subsequent performance grew weaker, and the second efficacy and goal measures
accounted for no additional variance in a multiple regression analyses beyond that
contributed by the second performance score alone. Kaplan concluded that the results of
the Silver and Bufanio study provide stronger evidence against the inﬂuence of collective
efﬁcacy and goals on performance than in favor of such an inﬂuence.

Porter (1992) examined the relationship between self-efficacy, collective efﬁcacy
and performance using 35 work groups (150 individuals) from a manufacturing company.
Performance was the total number of parts assembled by the group each day. Porter’s
study was conducted on the individual and group levels. In order to justify the
aggregation of individual level perceptions to the group level, she ﬁrst calculated the
degree of consensus for each group using the method recommended by James et a1.
(1984). Porter found that neither aggregated self-efﬁcacy or aggregated collective
efﬁcacy predicted performance in a hierarchical multiple regression analysis. However,
correlations between aggregated self-efﬁcacy and group performance ranged between -
.05 and -.55, whereas correlations between aggregated collective efficacy and group

performance ranged between -.58 to .17.

33

Collective efﬁcacxregarch - Sporgtudie; More research has been conducted
on collective efﬁcacy in the sport domain and all have found collective efﬁcacy to be an
important factor in team performance. Spink (1990) investigated the relationship between
group cohesion and collective efﬁcacy in elite and recreational volleyball teams. In this
study, collective efﬁcacy was operationalized as a construct which reﬂects the fact that
teams often have collective expectations for success. It was assessed with the following
two questions: “What placing do you expect to attain in Supervolley (the tournament)?”
and “How conﬁdent are you that your team will attain this placing?” The participants
who expected their team to ﬁnish ﬁrst formed the “high” collective efﬁcacy team,
whereas the participants who expected their team not to place in the top three formed the
“low” efﬁcacy team. Ninety-two volleyball players participated in this study.

Results for the elite sample demonstrated that the high and low collective efﬁcacy
teams could be differentiated by cohesion beliefs. Speciﬁcally, Attractions to Group -
Task (ATG-T) and Group Integration - Social (GI-S) were greater for team members who
were high in collective efﬁcacy. For the recreational volleyball players, however, none of
the cohesion variables discriminated between the high collective efﬁcacy and low
collective efﬁcacy teams.

Spink also reported a relationship between collective efﬁcacy and team
performance. Speciﬁcally, in a post hoc examination of the ﬁnal tournament placings, he
found that “high collective efﬁcacy teams ﬁnished higher in the tournament than low
collective efﬁcacy teams” (p.308). He suggested that the cohesiveness might be an
important determinant of collective efﬁcacy, whereby the greater the perception of group

cohesion, the greater the collective efficacy.

34

The study conducted by Hodges and Carron (1992) investigated the effects of
collective efﬁcacy on performance using a laboratory task. Experimental triads competed
against a confederate team on a muscular endurance task (i.e., holding a medicine ball
aloft). The 51 triads were divided into high and low collective efﬁcacy teams. Collective
efﬁcacy was manipulated by telling the high efﬁcacy teams that their performance was
superior to a confederate team. Low collective efficacy teams were told that their
performance were substantially inferior to the confederate teams. In all cases, the
confederate team won the competition. Collective efﬁcacy was assessed by asking the
teams to give a team response to the following questions: “what do you think your team’s
chances are of winning?” and “how conﬁdent are you of your prediction?” An overall
efﬁcacy measure was computed by multiplying the two scores together. Results
demonstrated that the manipulation was successful in producing high and low efﬁcacy
teams.

Hodges and Carron (1992) found support for the hypothesis that experimental
triads high in collective efﬁcacy improved their performance on the muscular endurance
task following a failure experience, whereas triads low in collective efﬁcacy experienced
a performance decrement. In addition, the teams with high efﬁcacy persisted longer than
low efﬁcacy teams. Furthermore, both teams (high and low efﬁcacy teams) decreased
their collective efﬁcacy following failure.

A series of three studies examining a number of collective efﬁcacy issues were
conducted recently by Paskevich (1995). In his ﬁrst study, Paskevich developed a multi-
dimensional measure lof collective efﬁcacy. The collective efﬁcacy measure contained

separate subscales for task (offensive, defensive, transition), communication, motivation,

35

conﬁdence in the face of obstacles, conﬁdence in skills in overcoming obstacles
associated with teammates, and general collective efﬁcacy. Reliabilities for each
subscale were greater than .85 (in all 3 studies), and the correlations between the
subscales ranged from .41 to .82.

This collective efﬁcacy measure was subsequently used to examine the
relationship between cohesion and collective efﬁcacy. Paskevich hypothesized that
because collective efﬁcacy measures focus on speciﬁc skills and abilities that the team
performs, collective efﬁcacy would predict athletes who were high or low in their
perceptions of the task-related measures of cohesion. The sample used in this study
comprised 70 volleyball players from 7 teams.

Results from discriminant frmction analyses revealed that two of the group
cohesion constructs (GI-T, and ATG-T) were able to differentiate between the high and
low collective efﬁcacy teams. When the constructs were reversed, four of the collective
efﬁcacy subscales (communication, motivation, overcoming Obstacles in association with
teammates, and general collective efﬁcacy) were able to differentiate the teams high and
low in group integration — task (GI-T).

A correlational analysis revealed that GI-T and ATG-T demonstrated a number of
signiﬁcant relationships with the collective efﬁcacy subscales. In addition, GI-S was also
correlated with the communication and the motivation subscales of the collective efﬁcacy
measure. ATG-S was not signiﬁcantly correlated to any of the collective efﬁcacy
subscales. Thus, individuals high in collective efﬁcacy had greater levels of task-related

cohesion than those individuals with lower levels of collective efﬁcacy.

36

In his second study, Paskevich examined collective efﬁcacy data from both team
and individual perspectives. The purpose of this study was to clarify the degree of sharing
reﬂected in collective efﬁcacy beliefs and cohesion beliefs, and to determine whether
collective efﬁcacy and cohesion were best represented as an individual, a team
phenomena, or some combination of both. Two hundred and eighty-seven individuals
from 25 collegiate level volleyball teams participated in this study. Through the use of
Kenny and LaVoie’s (1985) LEVEL program, Paskevich used intraclass correlations to
test for non-independence of team member responses.

Paskevich found that all measures of collective efﬁcacy and cohesion appeared to
be more of a function of team processes and interaction than individual level processes.
He stated that because the majority of the larger correlations and percentage of variance
accounted for were found at the adjusted team level, this fact should provide elementary
evidence, as suggested by Steiner (1972), Shaw (1981) and Zaccaro et a1. (1995), that the
team is something more than the simple collection of individuals. He recommended
assessing collective efﬁcacy as team level construct.

In his ﬁnal study, Paskevich examined the direction of the relationships between
collective efﬁcacy, cohesion and performance-outcome. The participants for this study
were the same as those who participated in study two (i.e., 25 teams). Winning
percentage was used as the performance-outcome variables.

Paskevich found that perceptions of collective efﬁcacy and cohesion increased
over the course of a season, with later season efﬁcacy values being larger than early
season values. This is consistent with Feltz and Lirgg (in press). He also found initial

evidence that the task subscale of the collective efﬁcacy measure mediated the

37

relationship between cohesion (GI-T) and performance, at early season, but not later in
the season. However, there was also evidence for the independent effects of both
collective efﬁcacy and cohesion on performance.

Another study was conducted by Lichacz and Partington (1996). These authors
investigated the relationship between collective efﬁcacy and performance for “true” and
“ad hoc” teams. The participants in this study comprised 8 teams. The four true teams
were made up of individuals who belong to either the same rowing team (2 teams), or the
same basketball team (2 teams). Thus, these teams had a history of playing together. The
ad hoc teams were made up of individuals who either had some previous team sport
experience (2 teams) or who did not (2 teams). The ad hoc teams did not know each other
before the study. The experimental task was a rope-pulling task. All participants pulled
on a rope, individually, and then collectively with their other team members. The
collective pulls and the sum of the respective individual pulls were used to derive a
collective performance efﬁciency index (i.e., collective pull divided by sum of individual
pulls).

Perceptions of collective efﬁcacy were manipulated through performance
feedback during the experimental session. The participants were given differential
feedback after each of their individual and collective pulls. Participants in the low
efﬁcacy condition were told that their pulls were 10% below the standards set by hi gh-
level athletes, or non-athletes (depending upon the composition of the team), while those
in the high-level efﬁcacy condition were told that their pulls were 10% above the
standards set by high-level athletes or non-athletes. In addition, perceptions of collective

efﬁcacy were assessed by the following question: “Do you think that your team will

38

achieve a better pull than the normative pull?” Ratings were made from 0 (no chance of
obtaining the best score) to 10 (will obtain the best score). The participants in the high
collective efﬁcacy teams all displayed mean collective efﬁcacy scores. greater than the
participants in the low collective efﬁcacy teams. Thus, they concluded that the
experimental manipulation was successful. The correlation between the individual
perceptions of collective efficacy and team performance was .27. In addition, the teams
from the high efﬁcacy conditions demonstrated larger means for performance than the
low collective efﬁcacy teams.

More recently, in their multi-level analysis of collective efﬁcacy, Watson and
Chemers (1998) investigated the antecedents and consequences of collective efficacy in
college basketball teams. Twenty-eight teams participated in the study (336 individuals).
Collective efﬁcacy and self-efﬁcacy were assessed twice during the season. In order to
demonstrate that collective efﬁcacy was a shared belief, the procedures outlined by James
et al. (1984) were used. Data were analyzed using Hierarchical Linear Modeling (HLM)
techniques. HLM allows for the testing of data with a hierarchically nested structure (i.e.,
individuals nested within teams). Only the results related to efficacy and performance are
described here.

First, this study supported the notion of collective efﬁcacy as a shared belief. A
high level of within team agreement on collective efﬁcacy was documented. Second,
although considerable stability in collective efﬁcacy judgments were noted over the two
time periods, the degree to which collective efﬁcacy judgments were shared and their
antecedents changed over time. More speciﬁcally at the beginning of the season,

collective efficacy was more personalized than later in the season. Watson and Chemers

39

also showed that the relationship between self-efﬁcacy and collective efficacy depended
on the average self-efﬁcacy of the team. For teams with high average self—efﬁcacy, the
relationship between players’ self-efﬁcacy and perceptions of collective efﬁcacy was
positive. For teams with low average self-efficacy the relationship between players’ self-
efﬁcacy and collective efﬁcacy was negative.

With respect to performance, Watson and Chemers found that previous
performance was unrelated to collective efﬁcacy at the beginning of a season,. However,
collective efﬁcacy judgments at the beginning of the season predicted later collective
efﬁcacy beliefs and overall team performance at season’s end. On an individual level,
stronger individual performances (as measured by the average number of points a player
scored per game) were predicted by high self-efﬁcacy but not by collective efﬁcacy.

Feltz and Lirgg (in press) examined the pattern of collective and self-efficacy
across a season of competition, and examined the relationships among self-efﬁcacy,
collective efﬁcacy and team performance. This study was conducted on the team level,
and therefore the collective efﬁcacy and self-efﬁcacy constructs comprised aggregated
individual perceptions. The participants in this study were 159 ice hockey players
representing 6 teams. The task-speciﬁc efﬁcacy measure assessed the degree of
conﬁdence that player’s had in their team’s ability to outperform their opponents on
important game competencies. To determine if athletes held shared beliefs within their
teams regarding collective efﬁcacy perceptions, these authors conducted a consensual
analysis in accordance with the procedures outlined by James et al. (1984).

To compare how collective and self-efficacy beliefs varied across the season,

polynomial trend analyses were used. A signiﬁcant cubic trend was found. The following

40

relationship was evident for the entire sample: all teams increased in collective efﬁcacy
across the ﬁrst half of the season, then collective efﬁcacy decreased at the midpoint of
season, and then, at the end of the season, collective efﬁcacy increased to a level higher
than initial values. The same trend was evident for self-efﬁcacy, except for a more subtle
decline in the second half of the season.

Although all teams demonstrated a high degree of consensus in their collective
efﬁcacy beliefs, because of the small number of teams, team by game was used as the
unit of analysis in regression analyses. Bivariate correlations revealed that collective and
self-efﬁcacy were positively correlated (r=.60), and that team performance was correlated
with collective efﬁcacy (r=.30) but uncorrelated with self-efﬁcacy (r=.03). The regression
analysis revealed that collective efﬁcacy was the strongest predictor of performance,
while self-efﬁcacy was signiﬁcant, but in the negative direction.

A secondary purpose of the Feltz and Lirgg study was to investigate if winning or
losing affected efﬁcacy judgments. They found that wins and losses affected collective
eﬁicacy more than self-efﬁcacy, but losses affected collective efﬁcacy more in
comparison to self-efficacy. That is, winners increased collective efﬁcacy beliefs
following a win, and losers signiﬁcantly decreased their perceptions of team’s abilities
following a loss. Results for self-efﬁcacy supported the same trend, but were not
signiﬁcant. Thus, they concluded that past team performance affects collective efﬁcacy
beliefs to a greater extent than self-efficacy beliefs. This study supported the association
between collective efficacy and team performance, and the fact that it was conducted

over a full season gives credence that this association was not spurious.

41

As one can see, the research using the collective efﬁcacy construct has been
varied, and diverse. Collective efﬁcacy has attracted researchers from social, educational,
industrial/organizational and sport psychology. One problem that plagues this research,
however, is the inconsistencies in the conceptualizations and measurement of collective
efﬁcacy. In some studies, researchers have treated collective efﬁcacy as a team level
phenomena (e.g., F eltz & Lirgg, in press), while others conceptualized collective efﬁcacy
as individual level perception (e.g., Prussia & Kinicki, 1996; Riggs & Knight, 1994). In
one case, collective efﬁcacy was conceptualized as having both individual and team level
effects (i.e., Paskevich, 1995). Adding to the confusion, is that Bandura (1997) and others
(e.g., Gist, 1987; Guzzo et al., 1993; Lindsley et al., 1995) have identiﬁed a number of
methods that can be used to assess collective efﬁcacy. Thus, the next section of this
chapter addresses the myriad of ways collective efﬁcacy can be assessed.

Measurement of Collective Efﬁcacy

Bandura (1977, 1986) recommended that self-efﬁcacy be measured along the
dimensions of magnitude (level), strength and generality. The same format is applicable
to collective efﬁcacy measures. Also similar to self-efﬁcacy measures, measures of
perceived collective efﬁcacy should be closely tied to explicit indices of team
performance (George & Feltz, 1995). That is, researchers need to conduct a contextual
analysis of the competency areas within sport teams prior to developing their
questionnaire items.

Bandura (1986, 1997) called for a “broad and comprehensive research effort” (p.
452) with respect to the construct of collective efficacy. He stated that advancement in

collectiveefﬁcacy research requires the “development of suitable tools for measuring

42

perceptions of collective efficacy to achieve varying levels of results” (p. 452). To date,
collective efﬁcacy measures have been developed in a number of ways, and there is alot
of confusion regarding the appropriate way to assess the construct.

The ﬁrst method that has been suggested as a means to assess collective efﬁcacy
involves aggregating the perceptions of individual self-efﬁcacy beliefs. That is,
individual self-efﬁcacy beliefs are averaged to form a team measure which represents
collective efﬁcacy. However, Bandura (1986; 1997) and others (Feltz, 1994; Feltz &
Chase, 1998; F eltz & Lirgg, in press; Lindsley et al, 1995; Paskevich, 1995; Zaccaro et
al, 1995) have pointed out that collective efﬁcacy may be insufﬁciently represented by
the sum of the perceived personal efﬁcaeies of participants.

One rationale for why collective efﬁcacy is more than the sum of individual
efﬁcacy expectations is based upon Steiner’s (1972) work with group performance.
Steiner found that group performance was more than the sum of the individual’s abilities,
and showed clearly that group performance is typically overestimated by the simple
summation of member abilities because coordination and motivational processes are not
taken into account (i.e., actual group productivity equals potential productivity minus any
social process losses). The existence of a group structure and the associated coordinated
and interactive processes that occur relative to matters of consequence to the group imply
that researchers are unrealistic in their estimates of group function when they simply sum
its member abilities (cf Steiner, 1972). Similarly, Paskevich (1995) and Zaccaro et a1.
(1995) argued that merely summing judgments of abilities of other group members on a

task misses the group’s interaction, coordination and integration factors.

43

Bandura (1997) stated that team ﬁrnctioning is a product of the interactive and
coordinative dynamics of it’s members and these interactive dynamics create an emergent
property that is more than the sum of the individual attributes. He also argued that the
sum of the team members’ individual efﬁcacy expectations would not be an accurate
representation of the collective efficacy construct under most circumstances. In addition,
Bandura stated that there are certain conditions under which aggregated personal and
wholistic judgements of collective efficacy are likely to diverge. A weak link in an
activity that has to be performed interdependently can spell team failure even though the
remaining individuals are highly efﬁcacious. Similarly, a collection of supremely
efﬁcacious members may perform poorly as a unit if they do not work well together. In
both of these cases, the aggregate of personal efﬁcacies would overpredict the level of
team performance. Furthermore, when a key function for team success is performed by a
highly efﬁcacious individual, members will have a higher opinion of their team’s
capability than what they can do personally. Where success rests in the hands of a few
extraordinary efﬁcacious members in critical positions, aggregated personal efﬁcacies
would underpredict team attainments.

A second method to assess collective efﬁcacy involves aggregating individuals’
own perceptions of collective efﬁcacy (Lindsley et al., 1995). To assess collective
efﬁcacy in this manner, team members would be asked to respond to this type of
statement: “rate our conﬁdence that your team can beat the opposing team.” This
approach of aggregating individual beliefs about the collective to assess collective
efﬁcacy is not without criticism, however. The major drawback with this method is that

the wording of the items reﬂects individuals’ perceptions of their team’s capability rather

44

than a team’s efﬁcacy per se (Earley, 1993). That is, respondents are asked their
individual perception of the collective efﬁcacy of the team (Lindsley et al., 1995).
Lindsley et al. (1995) stated that this is a useful strategy if it is assumed that individuals
may not be knowledgeable of the team’s collective perception, or it is assumed that the
“collective min ” is a differentiated system in which no individual (part) has access to the
collective whole.

Bandura (1997), however, stated that the methods of aggregating individual
perceptions of self-efficacy (Method 1) and aggregating individual perceptions of the
team (Method 2) are not as distinct as they might appear. He stated that when people are
judging perceived efﬁcacy in a team endeavour they are not plumbing an abstract team
mind in which the members are detached from each other, and that an assessment focus at
the individual level is steeped in processes operating within the team. Furthermore, a
focus at the team level does not remove any thought about the individuals that contribute
to the collective effort. Thus, he stated that these two indices of collective efﬁcacy would
be at least moderately correlated. Some support for this proposition was provided by
Feltz and Lirgg (in press). In their study, they assessed efficacy by aggregating individual
self-efﬁcacy beliefs, and aggregating individual beliefs about the collective. They found
that these two measures were positively correlated (r = .60).

Another assessment method is to use individuals as informants to estimate the
team’s collective efﬁcacy. This procedure is similar to the one just described, however, it
involves assessing each team member’s appraisal of their team’s ability that it can
perform speciﬁc tasks successfully (Lindsley et al., 1995). For example, an item may

read “rate your team’s conﬁdence that your team can beat the opposing team.” Thus,

45

when items are phrased in this way, the focus is on the team’s belief, rather than
individual level beliefs. However, it is still an individual’s perception or rating. Similar to
the second method, individual beliefs regarding the team’s ability would be aggregated to
produce one index of collective efﬁcacy.

According to Lindsley et al. (1995), although the empirical difference between
this method and the second method may be minor when respondents share beliefs about
the team’s capabilities, the methods for assessing reliability differ. That is, focusing on
the team’s belief, rather than individual beliefs, avoids many of the debates and pitfalls of
multi-level analysis. The rationale for using individuals as informants to estimate a
team’s collective efﬁcacy is based on the notion that there are certain cognitions that
team members have which are quite different and distinguishable from the beliefs they
experience as individuals in isolation, or in other contexts outside the team (Lindsley et
al., 1995). These cognitions are collective, team-based beliefs arising from an
individual’s ability to cognitively consider social entities larger than himself or herself.
Thus, they are emergent properties of the team rather than the individual, and cannot be
reduced to their constituent parts.

This method is partially consistent with what. Zaccaro et al. (1995) and Paskevich
(1995) recommend for assessing collective efficacy. That is, these authors concur with
Lindsley et al. (1995) that collective efﬁcacy measures should focus on the collective as a
whole. However, they also recommend that collective efficacy measures assess the
various aspects that may inﬂuence how well a team perceives that it can work together.
Because collective efﬁcacy beliefs can be inﬂuenced by a variety of factors including

differences in task competencies, physical qualities, social skills and a team’s

46

coordination and integration capabilities, they believe that collective efﬁcacy should be
conceptualized as multi-dimensional in nature, and assessed with multi-dimensional
measures.

In his study, Paskevich (1995) developed a multi-dimensional measure of
collective efﬁcacy for volleyball. The scales were constructed to be sport and context
speciﬁc. The questionnaire items pertained to the various general components of practises
and games that concern offensive and defensive team capabilities/strategies. The items
also assessed the team’s conﬁdence in it’s skills and abilities to make the team’s
communication, coordination, persistence, and barrier reduction strategies effective.
Thus, Paskevich’s measure was composed of eight collective efﬁcacy subscales.

Although Zaccaro et al. (1995) and Paskevich (1995) believe that collective
efﬁcacy must be assessed in this multi-dimensional format in order to capture the variety
of forces that may impact collective efﬁcacy beliefs, Bandura (1997) may disagree.
According to Bandura, the wholistic judgment of a team member’s appraisal of their
team’s capability encompasses the coordinative and interactive inﬂuences operating
within a team. Thus, the multi-dimensional measures of collective efﬁcacy are not
needed.

One important ﬁnal point, however, that is especially evident in Paskevieh’s
(1995) study, is that in both cases (i.e., if one asks respondents to rate his/her perceptions
of the team, or rate his/her teams’ perceptions), individuals are still responsible for
providing the rating. More speciﬁcally, although the items on each of Paskevich’s
collective efficacy subscales were phrased as “rate your team’s conﬁdence...,“ in his

description of each subscale he stated that the items “assessed the athlete’s perceptions of

47

the team’s conﬁdence...” It can be argued that although team features, events and
processes provide the situational stimuli that are a determinant of collective efﬁcacy
beliefs, it is the individuals who attempt to extract the psychological meaning and
interpret these factors and processes, and therefore the perception, interpretation and
meaning of collective efﬁcacy beliefs should occur at the individual level (cf Kozlowski
& Hattrup, 1992). Thus, despite the fact that items may be worded as “rate your team’s
conﬁdence...” rather than “rate your conﬁdence in the team...” the appropriate level of
analysis would be the individual until consensus was demonstrated. The issue of
consensus or agreement is paramount in collective efﬁcacy research and will be discussed
in more detail later.

In any event, some researchers (e.g., Prussia & Kinicki, 1996), have argued that
the use of aggregation procedures represent surrogates of team level measures, and are
not team level measures per se. These authors would advocate for collective efﬁcacy
measures that require team members to provide a single response as a team, thereby
theoretically treating the team as an entity. Thus, another method to assess collective
efﬁcacy is to have the team members make the collective efficacy judgment together.
This approach eliminates the calculation of statistical indicators of agreement and avoids
the aggregation issue. It is believed by some (e.g., Guzzo et al., 1993) to be an acceptable
way of measuring a team’s collective belief. However, it also has drawbacks. For
example, according to Bandura (1997), team members are rarely of one mind in their
appraisals of matters. He pointed out that forming a consensual judgment of a team’s
efﬁcacy by team discussion is subject to social persuasion and conformity pressures. For

example, it is possible that a more prestigious team member could inﬂuence the judgment

48

in such a way that it does not accurately represent the views of many of its members.
Thus, team members may publicly consent to a response without truly believing it
(Guzzo et al., 1993). Forced consensus may be highly misleading, especially with a large
collective.

It is Obvious that the conceptualization of collective efﬁcacy raises a number of
methodological and interpretive issues. Before Bandura’s new chapter, it was generally
accepted that collective efficacy referred to people’s perceptions about the collective as a
whole, but it was not clear whether these beliefs were shared, and therefore represented a
team level rather than an individual level phenomenon (Zaccaro et al., 1995). Bandura
(1997) stated clearly that collective efﬁcacy does refer to shared beliefs, and is a team
level attribute.

The speciﬁcation of collective efficacy as a team level attribute requires that team
members be sufﬁciently similar with respect to their collective efﬁcacy beliefs so that
they may be characterised as a whole (Klein, Dansereau, & Hall, 1994). Accordingly,
homogeneity of beliefs is considered to be a prerequisite for asserting that the construct in
fact applies to that team (Dansereau, Alutto, & Yammarino, 1984). Thus, the theoretical
conceptualization of collective efﬁcacy as a shared belief (Bandura, 1997; Zaccaro et al.,
1995), implies that team members’ collective efﬁcacy beliefs should be sufﬁciently
homogeneous.

According to Klein et al. (1994), asserting that members of a team are
homogenous with respect to a theoretical construct, like collective efﬁcacy, means that a
researcher is predicting that team members’ collective efﬁcacy beliefs are identical.

However, Bandura (1997) stated that commonality of efficacy belief among team

49

members does not mean that every member is of exactly the same view on every aspect
of team functioning. He stated that although complete uniformity would be rare, the
differences in efﬁcacy beliefs between teams should be greater than the variation within
teams (Bandura, 1997). Thus, in proposing relationships, the focus should be on the
variation between teams, and that hypotheses should conform to the prediction that
differences between teams on one construct of the theory are related to differences
between teams on other constructs of the theory.

By conceptualizing collective efﬁcacy as shared beliefs, researchers need to
assess the degree of agreement or consensus of team member’s perceptions in order to
validate the collective efﬁcacy construct. In his recent writings, Bandura (1997) stated
that collective efﬁcacy is best characterised by a representative value for the beliefs of
it’s members and the degree of variability or consensus around that central belief. That is,
the differences in efﬁcacy beliefs between teams should be greater than the variation

Now that collective efficacy has been conceptualized to be a team level
phenomenon, importance must be placed on how individual perceptions of collective
competence are shared and aggregated in a conceptually appropriate fashion (Shamir,
1990). Feltz and Chase (1998) stated that although aggregated individual data may be an
appropriate way to assess collective efﬁcacy, it is ﬁrst necessary to examine the degree of
consensus at the individual level. Gully et al. (1995) recommend that the individual data
be aggregated to form a team level construct only when there exists an acceptable degree
of consensus. There are a number of methods that can be used to assess the degree of

consensus or homogeneity of collective efﬁcacy judgments. These statistical tests fall

50

into two rough categories: (a) those that assess the extent of agreement within a single
team, (b) those that assess the extent of agreement or consensus by contrasting within-
and between-team variance. For example, one method that can be used in consensus
analysis involves using an index of within team interrater agreement (James et al., 1984).
Other methods are based on analysis of variance models, and related measures of
association like eta-squared, and different forms of the intraclass correlation (Kenny &
LaVoie, 1985). Although they employ different criteria, these indicators are designed to
answer a common question: does the variability within teams differ from what would be
expected from chance?

Just how much inter-member agreement is enough? If one adopts the James et a1.
(1984) approach where team members are viewed as multiple judges rating the same
stimulus, then guidelines from classical test theory can be applied to inform the judgment
of whether an observed level of inter-member agreement is sufﬁcient. Nunnally (1978)
stated the sufﬁciency of a measure depends on the use to which the measure is put.
Generally, coefﬁcients of .50 to .80 can be quite sufﬁcient for research purposes, but
coefﬁcients in excess of .90 may be required for certain applied purposes.

One question, then, is what would a researcher do if there was no consensus in
efﬁcacy beliefs among team members? Bandura (1997) stated that when individuals
differ as widely in efﬁcacy beliefs within teams as they do between teams, there is no
distinguishable shared efﬁcacy attribute for the team as a whole. This is in agreement
with Guzzo and Shea (1993) who stated that low agreement may indicate that the team is
not a consequential entity to its members, or that the unity of the team is doubtful. Thus, a

researcher might properly chose to exclude from further analysis those teams having

51

insufﬁcient inter-member agreement. Zaccaro et al. (1995) stated that collective efficacy
may be comprised of both individual and team level components. A team level
phenomenon would be said to exist if there is little variability between team members’
perceptions of their collective competence, and this homogeneity cannot be attributed to
external factors. If however, there is considerable heterogeneity or variability in
individual perceptions of the team, then it is possible that individual level processes are
being expressed.

Level; Issues

The phrase “level of analysis” was coined to refer to the unit in which data are
assigned for hypothesis testing and statistical analysis (Rousseau, 1985). Thus, the level
of analysis describes the treatment of data during statistical procedures (Klein etal.,
1994). “Level of measurement,” is used to describe the actual source of the data (i.e., the
unit to which data are directly attached). For example, self-report measures generally
result in individual level data, whereas the number of team members in a team is a team
level attribute (Klein et al., 1994; Rousseau, 1985). If the level of measurement is the
individual, but individual scores are aggregated by using team means in data analysis,
then the level of statistical analysis is the team.

A critical ﬁrst step in addressing levels issues is the speciﬁcation of the level of
one’s theory (Klein et al., 1994; Rousseau, 1985). The level of theory shapes the
fundamental nature of the constructs of the theory by describing the target (e. g.,
individual, team, organization) that a theorist or researcher aims to depict and explain,
and specifying the level to which generalizations are made. More simply stated, when

specifying the level of the theory, a researcher needs to ask him/herself the following

52

questions: (a) to what level do you want to make generalizations, and (b) what are the
constructs of interest (i.e., the individual or team)? The answers to these questions drive
the wording of questionnaire items and the form of data collection and analysis. The
failure to specify the level(s) to which a theory applies leads to imprecision within the
theory, and confusion during data collection and analysis to test the theory (Klein et al.,
1994; Rousseau, 1985).

In addition to the level of one’s theory, the research question is equally important
to consider. More speciﬁcally, although Bandura (1997) conceptualized collective
efﬁcacy as a team level construct, at times it may be best to collective efﬁcacy as an
individual belief. For example, Feltz (1994) conceptualized social loaﬁng as a possible
“dark side” of collective efﬁcacy. In order to investigate the relationship between
collective efﬁcacy, and social loaﬁng, it would be necessary to measure both
performance, and collective efﬁcacy at the individual level. That is, one would measure
individual perceptions of collective efﬁcacy (i.e., conﬁdence in one’s team), and correlate
this with individual performance; in this case, the level of analysis would be the
individual (F eltz & Chase, 1998).

Levels issues create particular problems when the level of theory (i.e., level of the
research question), level of measurement, and/or the level of statistical analysis are not
the same (Klein et al, 1994). A Merl of the wronglevel (i.e., misspeciﬁcation) occurs
when a mismatch between the focal level and the level of analysis occurs. In this case,
relationships are attributed to a level other than the actual behavioral or responsive unit.
That is, the use of individual-level data to say something about the team may be an

unjustiﬁed shift in level. For example, by operationalizing and measuring group cohesion

53

at the individual level and generalizing ﬁndings to the team level, a researcher is
committing a fallacy of the wrong level.

Equally signiﬁcant is the problem of spuriousness or agggegation bias.
Aggregation bias refers to a class of errors in which an apparent relationship is an artifact
of the data combination method (James, 1982). For example, if individual level data are
summarized as team means without ensuring the homogeneity of responses at the
individual level, then aggregation bias becomes a potentially severe problem (Gully et al.,
1995). Indik (1968) argued that variables at the same level of analysis should be more
highly related than variables at different levels of analysis, and Ostroff (1992)
demonstrated that relationships at higher levels of analysis are often stronger than those
from individual level analyses due to statistical artefacts, biased estimates, elimination of
error variances, or actual differences between individual and higher level constructs.
Thus, there are a myriad of methodological reasons that suggest that relationships at the
team level will be stronger than relationships at the individual level, especially when
aggregation is performed without consideration for consensus (cf Gully etal., 1995).

In some cases, a theory may operate on a number of levels (Klein et al, 1994;
Rousseau, 1985). The generation of these multi-level models have tremendous
advantages in that they are uniquely powerful and parsimonious. Multi-level models
postulate relationships among variables which apply at two or more levels (Rousseau,
1985). Thus, they describe relations at one level that are generalizable to other levels.
One beneﬁt of this approach is that multi-level models may generate testable hypotheses
at one level of analysis that are suggested by empirical results or theory at another level

of analysis. In addition, multi-level thinking forces a theorist and/or researcher to

54

examine individuals, teams and organizations as parts of a whole, each affecting and
being affected by the other, rather than as separate conceptual categories (Lindsley et al.,
1995)

Recently, it has been suggested that self-efﬁcacy, collective efﬁcacy and
individual and team performance may comprise a multi-level model (cf Feltz & Lirgg, in
press; Lindsley et al., 1995, Prussia & Kinicki, 1996; Silver & Bufanio, 1996). A number
of individual studies have demonstrated the existence of a relationship between self-
efﬁcacy and individual performance, and initial research on collective efﬁcacy has
demonstrated that a similar relationship exists between collective efﬁcacy and team
performance. A multi-level conception of efﬁcacy, however, does not simply identify
self-efﬁcacy as an individual level phenomenon and collective efficacy as a group level
attribute (Moritz & Watson, 1998). Instead, a multi-level model of efﬁcacy postulates the
relationship between context, efﬁcacy and behavior within and across levels. Further, it is
not necessary for the individual level and team level relationships to be exactly the same.
It is possible that a certain variable will not affect the relationships among constructs at
one level, but yet, the same variable may act as an intervening variable at another level.
Thus, there may be potential discontinuities when one shiﬁs from the individual to the
team level with respect to the efﬁcacy - performance relationship. One of the variables
that may affect the efﬁcacy - performance relationship is task type.

Task Tm as a Moderator of the Efﬁcacy - Perfmance Relat_i9_ns_hip

The type of task has been identiﬁed by a number of researchers as a potential

moderator of the efﬁcacy - performance relationship (cf Bandura, 1997; Feltz, 1988,

1994; F eltz & Chase, 1998; Feltz & Lirgg, in press; George & Feltz, 1995; Little &

55

Madigan, 1997; Paskevich, 1995; Porter, 1992; Zaccaro et al., 1995). According to
Steiner (1972), the optimal process of converting a group’s resources into group products
depends heavily on the nature of the task a group is performing. Steiner distinguished
between unitary and divisible tasks. Unitary tasks are those that cannot be divided into
subtasks, and are characterized by a single group output. For example, rope-pulling,
which is often used in tug-of-wars, is a unitary task. Divisible tasks can be broken down
into sub-tasks that can be assigned to different group members. For example, playing a
football game is a divisible task.

Several types of unitary tasks can be distinguished, each of which has
implications for how individuals ought to combine their resources in a single group
output. Additive tasks involve summing individual inputs to yield a group product.
Completing an additive task requires every member of a group to perform the same
function. For example, team bowling may be considered as an additive task -- each
member of the team bowls his or her own game, and the ﬁnal team score is derived by
summing the individual scores. The group product for compensatory tasks is the average
of individual performances. For example, taking the average of each bowling team
member’s scores may be considered as compensatory. Disj unctive tasks require that only
one member do the task for the group to succeed (i.e., the potential productivity of a
group is determined entirely by the resources of its most competent member). Often,
these tasks typically require a single speciﬁc answer to an either/or, yes/no type problem.
For example, refereeing a tennis match can be considered to be disjunctive. Although
there are a number of ofﬁcials, a “call” (e.g., in or out) is typically made by only one

member of the team of ofﬁcials. Conjunctive tasks require that each group member must

56

individually succeed if the group is to succeed (i.e., the potential productivity of the
group is determined by its least productive member). That is, conjunctive tasks are
completed only when all of the group members perform some speciﬁc action. Many team
sports, like hockey and basketball can be considered as conjunctive. Finally, discretionary
tasks are those tasks where the group can decide how individual inputs will relate to the
group outcome. That is, the group members are free to choose the method by which they
will combine their individual inputs.

Bandura (1997) noted that the manner in which the individual contributions
combine to affect team performance is dependent upon the degree of task
interdependence among the individual tasks. He suggested that the predictiveness of
collective efﬁcacy may depend upon the degree of interdependence required by the team.
Teams are interdependent in the sense that an event which affects one member is likely to
affect all members. Interdependence can be conceptualized as varying along a continuum
from low to high.

Shaw (1971) deﬁned group tasks in terms of the degree to which integrated action
of group members is required to complete the task. At a minimum, individuals often
require resources and support from their surrounding social environment, even if there is
little or no coaction (i.e., members work in the same location, but do not interact with one
another) or interaction with other individuals. That is, under conditions of minimal task
interdependence, individual tasks are totally independent, and each one is unaffected by
the others (i.e., members of the group do not have to rely on each other to perform their
job though they have shared goals and provide mutual social support). Group

performance is simply the sum of individual performances. Bandura (1997) stated that

57

under these conditions, collective efﬁcacy should approximate the sum of individual self-
efﬁcacies. Thus, under low interdependence, an aggregate of individual efﬁcacies should
have predictive value.

Zaccaro et al. (1995) stated that even on such tasks that require minimal
interdependence, the independence of action does not mean that individual team members
are not inﬂuenced in some manner by the activities of other team members. He stated that
members may be affected by the speed and intensity of their peer’s actions, and may alter
the strength and nature of their own responses in accordance with earlier team activities.
Thus, collective efﬁcacy beliefs may still be contingent on other member’s responses.

At a higher level of interdependence, individuals may coact to produce an
aggregated product. While there may be little interaction, characteristics of the aggregate
(e.g., team size, homogeneity) may still inﬂuence its member’s beliefs about personal
mastery and hence the nature of their performance. At the most complex level of
interdependence, individuals often behave in complete concert with each other. Such
behavior typically has the characteristics of complete integration, coordination, and
synchronization. Individual actions are fully dependent upon the actions of others to
produce a collective outcome. In this case, team performance is not the sum of individual
performance, and in fact, individual performances often cannot be distinguished from one
another. Thus, only the collective or aggregate product is identiﬁable.

Bandura (1990) stated that the predictiveness of individual and collective efﬁcacy
may depend on the degree of integrated action of team members required to achieve
success. Perceived collective efﬁcacy beliefs may not show predictive superiority over

aggregated players own perceived abilities in sports with little or no task interdependence

58

because there are fewer co-ordination efforts to assess (cf Feltz & Lirgg, in press). Other
researchers have also suggested that aggregated perceived collective efﬁcacy scores may
be more relevant in sports requiring high levels of interdependence (e.g., conjunctive
tasks) than in sports with little or no task interdependence (George & Feltz, 1995).

Initial support for the differential predictiveness of collective efﬁcacy beliefs
based on task interdependence is provided Feltz and Lirgg (in press). These authors stated
that on highly interactive tasks, team member’s beliefs in their team’s ability to perform
successfully will be a better predictor of collective efﬁcacy than aggregated self-
efﬁcacies because member’s beliefs about the collective encompass the coordinative and
interactive dynamics that operate within a team Their results supported this proposition.
To date, no collective efﬁcacy research has been conducted using tasks that are low in

interdependence.

59

CHAPTER III
METHOD
Overview

The purpose of this dissertation was to examine the relationships among
aggregated self-efﬁcacy, collective efﬁcacy, and team performance in a less
interdependent (additive) versus a more interdependent condition in bowling. Based
upon theory and past research, it was hypothesized that task type would moderate the
relationships between aggregated self-efﬁcacy and team performance (Link C) and
between collective efﬁcacy and team performance (Link D). Task type was not
hypothesized to moderate the relationships between self-efﬁcacy and individual
performance (Link A) and between disaggregated collective efﬁcacy and individual
performance (Link B).

The second purpose of this study was to determine if the relationshipsamong
efﬁcacy and performance differed according to the level of analysis. Researchers have
suggested that self-efﬁcacy, collective efﬁcacy and performance may comprise a multi-
level model (e.g., Lindsley, Brass, & Thomas, 1995, Prussia & Kinicki, 1996; Silver &
Bufanio, 1996). At the individual level of analysis, the relationships among self-efﬁcacy,
(disaggregated) collective efﬁcacy and individual performance were examined. At the
team level of analysis, the relationships among aggregated self-efﬁcacy, collective
efﬁcacy, and team performance were considered (see Figure 1). It was hypothesized that
task type would be a moderator of performance for the team level relationships (Links C
and D), but not for the individual level relationships (Links A and B). These

relationships were tested at two time periods to see if they would replicate.

60

First, a pilot study was conducted to determine if the bowling manipulation was
effective, to determine the best measures of self- and collective efﬁcacy, to determine the
appropriateness of the aggregation of individual responses to a team level construct (cf
Little & Madigan, 1997), and to make other procedural reﬁnements. The participants in
the pilot study were 42 students (Males: a = 26; Females: a = 16) who were enrolled in
one of two introductory bowling courses. Two-member bowling teams were constructed
based on ability to ensure adequate heterogeneity of efﬁcacy beliefs and performance
scores between teams. All data were collected during the “tournament period” of the
course with one class bowling in the additive format and the other bowling in the
interdependent format.

Three self-efﬁcacy (SE-Pins, SE-Spares, SE—Strikes) and four collective efficacy
(CE-Pins, CE-Spares, CE-Strikes, CE-Points) measures were used in the pilot study. The
reliability for each measure, as assessed via Cronbach’s Alpha was acceptable; all
Alpha’s were over .90. The means, standard deviations, and ranges for all efﬁcacy and
performance measures for each task type are contained in Appendix A, Table A1. The
intercorrelations among these variables can be found in Appendix A, Table A2.
Performance measures included number of pins, number of strikes, number of spares, and
total points.

Consensus was tested using two methods: Pearson correlations were computed for
each team for each measure (Shrout, 1995), and a multiple item index of within group
agreement or consensus (rm); James et al., 1984) was calculated. Using the criteria of
m z .80, and g 2 .80 (cf George, 1990) excluded a number of teams that could be used

in analysis (a = 10 for Pins, a = 12 for Strikes, a = 11 for Spares). Therefore a different

61

method of assessing collective efﬁcacy was proposed for the larger study whereby
teammates completed the measures together rather than individually. This procedure
eliminated the need for consensus for the collective efﬁcacy measures. Assessing
collective efﬁcacy in this manner, however, created the need to demonstrate that the
partners functioned as a team. More speciﬁcally, did the partners feel that they were a
team? Did they perceive that they worked together? Did they perceive that they
coordinated their efforts? Measures were constructed to assess these perceptions.

Although consensus analyses were not needed for collective efﬁcacy, they were
needed before self-efﬁcacy beliefs could be aggregated to the team level. In the larger
study, the criteria for consensus for aggregation purposes was lowered to .70 (see Porter,
1992). Results from individual and team level analyses showed that there were enough
trends in the data to indicate that the type of bowling would moderate the efﬁcacy-
performance relationship. Thus, the larger study was undertaken with the above-
mentioned modiﬁcations.

Migrants and Task .

Two hundred and ninety students from introductory and advanced bowling
activity courses participated in this study, however only data ﬁ'om 250 individuals (179
males, and 70 females; one person did not indicate his/her gender) were useable. The
other 40 participants were excluded because they (or their partners) did not complete the
ﬁrst portion of this study. One hundred and twenty ﬁve two-person teams were used for
the ﬁrst part of this study. The participants ranged in age from 18 to 39 years (M= 20.15,
_S_I; = 2.33). The majority of the sample indicated that they were recreational bowlers

(66%). The average number Of years bowled was 5.59 (SD = 4.26), and the self-reported

62

mean for bowling averages was 117.47 (§_D = 20.67). The same participants were used in
the second part of this study (Time 2). However, only 160 participants (80 teams) had
complete data (117 males, and 43 females). Even with the missing data, the Time 2
sample was comparable to the ﬁrst, in that they ranged in age ﬂom 18 to 39 years (M:
20.15, SD = 2.65). The majority of the sample indicated that they were recreational
bowlers (68%), and the average number of years bowled was 5.74 (SD = 4.14). The mean
for self-reported bowling averages was 116.93 (S2 = 21.37).

In the additive condition, each team member contributed equally to team
performance. That is, each participant bowled a 10 ﬂame game, and team performance
was the sum of the two individual scores (the performance score ranged ﬂom 0 to 600
points). In the interdependent condition, the two bowlers alternated between bowling ﬁrst
and second for each ﬂame, like a “Scotch Bowling” activity (see Pangrazi & Darst,
1997). In this task, however, if Bowler A bowled the ﬁrst ball, unless he/she bowled a
strike, Bowler B was responsible for bowling the second ball in that frame. In the next
ﬂame, Bowler B bowled the ﬁrst ball, (regardless of whether he/she bowled the second
ball in the preceding ﬂame). Unless Bowler B bowled a strike, Bowler A was responsible
for delivering the second ball in this ﬂame etc. The instructions which were given to the
bowlers can be found in Appendix B. I imposed these rules so that I could control how
many balls were delivered by each team member; I wanted to ensure that each bowler
started the same number of ﬂames (i.e., 10 each) in this task. One game was designed to
be 20 ﬂames long so that the ﬁnal performance score would be comparable to the

additive condition (i.e., range between 0 to 600 points).

63

With respect to the degree of interdependence in the two bowling tasks, the
second task was considered to be more interdependent compared to the ﬁrst task because
it required more coordination between teammates (cf Zaccaro et al., 1995). More
speciﬁcally, in this condition, the bowlers were dependent on their teammates for all
“marks” (i.e., spares, multiple strikes) and for total points. Both of the tasks, however,
produced the same performance measures (i.e., number of pins dropped on ﬁrst ball,
number of strikes and number of spares), and these performance measures were available
for both the individual and team levels. Total points, however, was an exception.
Although team scores were available for both teams, individual performance scores were
not available for the additive team.

Measures

Performance Mea_sures. The performance measures consisted of the scoresheets
used in the tournament (see Appendix C). At the individual level, the number of strikes
and spares bowled by each individual per game were recorded. In addition, the average
number of pins dropped on each ﬁrst ball was computed. At the team level, the number of
spares and strikes bowled by each team was used as a performance measure. For
example, if Bowler A bowled 3 strikes and Bowler B bowled 2 strikes, the team “strike”
performance measure was 5. The average number of pins knocked down on the ﬁrst balls
was also computed as a team score. In this case, if Bowler A averaged 5 pins, and Bowler
B averaged 7 pins, the team “pin” performance measure was 6. “Total points per game”
were collected but not used because it was impossible to construct a corresponding self-
efﬁcacy measure with the same range of points (i.e., 0 to 600), even though collective

efﬁcacy measures were available. For the additive team, total score was the sum of the

64

individual members’ scores; for the interdependent team, the ﬁnal score for each 20-
ﬂame game was used.

Self-efﬁcacy question_r_rafires. Three self-efﬁcacy measures were used in this study
(see Appendix D). All of the measures were constructed in accordance with Bandura’s
(1977, 1986) recommendations. That is, the measures were task-speciﬁc, and were
hierarchically arranged to represent increasing levels of complexity. Participants
indicated the “level’ of their self-efﬁcacy beliefs, as well as the “strength” of their beliefs
using a 0 to 10 probability scale where 0 = “I cannot do this at all” and 10 = “Very
certain I can do this.” All of the self-efﬁcacy measures started with the phrase “Rate your
conﬁdence that you can...” The individual items pertained to: (a) bowl a strike in one of
your ﬂames (up to all 10 ﬂames); (b) bowl a spare in one of your ﬂames (up to all 10
ﬂames), and (c) drop at least one pin with your ﬁrst ball consistently (up to 10 pins). The
reliability for each measure, as assessed via Cronbach’s Alpha was acceptable; the
coefﬁcients ranged ﬂom .89 to .96.

Collective efﬁcacy questionnaires. The four collective efﬁcacy measures that
were used in this study were constructed in the same way as the self—efﬁcacy measures
(see Appendix E). However, this time the participants were asked to collectively “Rate
your conﬁdence that your team” can (a) bowl a strike in one of your ﬂames (up to all 20
ﬂames); (b) bowl a spare in one of your ﬂames (up to all 20 ﬂames), and, (c) drop at
least one pin with your ﬁrst ball consistently (up to 10 pins). The reliability for each
measure, as assessed via Cronbach’s Alpha was acceptable; the coefﬁcients ranged ﬂom
.89 to .96. In addition, a collective efﬁcacy measure was constructed to assess each

teams’ conﬁdence that their team could average a certain number of points in this game.

65

This measure ranged ﬂom 25 to 600 points; the baseline was 25 points, and the scale
increased in 25 point increments. The collective efficacy measures were anchored with 0
= “We are certain that we cannot do this at all” and 10 j: “We are very certain that we can
do this.” The reliability for this measure, as assessed via Cronbach’s Alpha was .89. In
accordance with other researchers (e.g., Prussia & Kinicki, 1996), and based on the
results of the pilot study, all of the collective efficacy measures were completed by the
team.

Team concth and incentive measures. Efficacy beliefs are a major determinant of
behavior only when people have sufﬁcient incentives to act on their self-perception of
efﬁcacy and when they possess the requisite subskills (Bandura, 1986). In order to
determine if the participants in this study had sufficient incentives to perform their best,
and to ensure that the study participants felt that they were part of their respective “team,”
a questionnaire was developed to assess a number of “team-relat ” characteristics. There
were ﬁve items on this questionnaire, and all were constructed using the same anchors as
the self-efﬁcacy and collective emcacy measures (see Appendix F). Question I asked
participants to indicate the degree to which they felt that they and their partner were a
“team.” Question 2 asked participants to indicate the degree to which they felt that they
and their partner were “working together” in this tournament. The third question asked
participants to indicate the degree to which they felt that they and their partner were
“coordinating their efforts” in this tournament. Question 4 asked each individual to
indicate how much effort they felt that they were putting into this tournament. The last

question asked each individual to indicate how important it was for them to bowl well

66

during this tournament. All of these questions were completed individually by each
participant.
Procedure

Approval to conduct this study was granted by the Institutional Review Board
(see Appendix G). The study participants were recruited from bowling classes offered
through the Department of Kinesiology. On average, there are 50 students in each class,
and there are 8 sections Offered per semester. The length of the classes is 50 min and
there are typically 30 classes per course.

All individuals registered in the bowling classes were asked to volunteer for the
study. They were told that the purpose of the study was to determine how self-confidence
and team conﬁdence are related to team performance in a bowling tournament. The
individuals who agreed to participate in the study were required to complete a consent
form (see Appendix H).

All consenting participants were then randomly assigned to two member bowling
teams (within their respective course sections). The experimental manipulation occurred
during the latter portion of the semester. Teams ﬂom certain sections competed in the
additive task (i.e., all team members bowled one game, and team performance was the
sum of each member’s score). In the other sections, the teams competed in the modiﬁed
interdependent format.

All bowling teams were given “practice” sessions. These sessions were designed
to give the teams experience in completing the collective efﬁcacy questionnaires, as well
as to become familiar with his/her bowling partner and the bowling task. Depending on

the course section, three or four practice sessions were given. After the practice sessions,

67

the tournament started. All of the measures were completed by the participants before
they bowled. The tournament lasted for 2 days thereby resulting in 2 sets of data (T irne 1
and Time 2). Prizes were awarded to the winning teams in each section based on the team

performance data.

68

CHAPTER IV
RESULTS

The results are presented in 2 sections. First, descriptive statistics are reported, In
this section, means, standard deviations, and ranges are presented for all measures (self-
efﬁcacy, collective efﬁcacy, performance, and team concept and incentive measures).
The results are presented separately for each task type (additive and interdependent), for
each level of analysis (individual and team) and for each time period (Time 1 and Time
2).

The second section presents the results of the moderator analyses. This section is
divided into Time 1 and Time 2. Within each time period, the results of the moderator
analyses are presented for the individual and the team level. Because two of the three
team level constructs (i.e., self-efﬁcacy and performance) are composed of aggregated
individual level responses, the individual level analyses are presented ﬁrst. All of the
statistical analyses were conducted with SPSS Release 7.0. The consensus analyses were
done using Microsoft Excel Version 5.0c.

Descriptive Statistics for Mca_sures

There were three self-efﬁcacy (SE-Pins, SE-Spares, SE-Strikes) and four
collective efﬁcacy measures (CE-Pins, CE-Spares, CE-Strikes, CE-Points) used in this
study. For each measure, two scores were computed (strength and level). The correlations
between these measures ranged ﬂom .56 to .89, and all were signiﬁcant (p < .01).
Because of the high correlation between the measures, only the strength scores were used
in subsequent analyses. The individual level means, standard deviations, and ranges for

the self-efﬁcacy, collective efﬁcacy and performance measures can be found in Table 2.

69

The individual level correlations between these variables are presented in Tables 3 and 4.
Descriptive results for the team level data are presented in Tables 5 and 6. Table 7 reports
the results of the incentive analyses for both time periods.

With respect to the incentive analyses, a one-way ANOVA using task type as the
independent variable showed that the additive and interdependent teams did not differ
statistically on any of the ﬁve items (2 >.05). Except for the item pertaining to the
importance of bowling well in this tournament, all of the means listed in Table 7 are
above 8.00, which correspond to the “we deﬁnitely...” anchor in the questionnaire. The
means for the importance of bowling well in this tournament are lower than the other
items, but are still above 7.0. Thus, there was an acceptable level of interest in the

bowling task.

70

.eoﬁmocmwmmme a: 8% domommm 3:80:00 .ooeagotoa 32:32: I A: ”.3850 2:80:00 u m0 ”doaommodﬁm u Mm deem 03:23

 

 

 

.8.“ #0 n m .3on 8333285 .89 em .I. m "N 08E. ”macaw 986cm 8.: cm H m £5on “coeaoaoﬁﬂﬁ .8.“ o: n : 08C. ”mam
odd: .. 9:. 3.3. doin— odﬂ . ode 3.3 2.5— 550:”:
06 . od w: 3.: c6 - cd :3 3.: 3. .. ed on: an: ad . 3. d: 32 nee—Emu“:
o.” . ed ow: Ed 06 . od we: wed o.» .. ed 3.: wad 3. . o.o K: and 3.89m":
ed - We 8.: and ed - Nd on: 86 md - ed 3....— Sd ad - «an :N: and 83x:
v.» . .d 8.: 26 3 - 3” S.— mm.m v.5 . md 3.: «ed vd - a: on: an aSoméU
md - 3 av: 86 ed - 3 3.: and ad - :d 3.— 36 a.» - ed mm: and gammo
3. .. d: mm: 26 3. - 3 mm: 91. 3. - 5.: am: one 3. . v: we: 86 moimﬁo
3: - ed cm: Gd dd . 3. mm: was ad .. dd 8.: a.» 3: - v6 3.: 3d 8:35
ed . nd 3.: $6 3: - ed mm: 36 3:- m: a»: 26 3: .. m: 2..— mmé moaqmém
md . _.N 3.: 34. cd - ed av: mod o.» - od an: Ed N.» . 9o 3.. 33V membmmm
ed . Nd am: Sun 3: - Nd :m: «dd 5d - dd d: cod dd - in 53 and «Sammm
owes: mm. 3 owﬁm mm. .5. owes: arm. .2 omen: ma 2
N 08:. : 08C. N 08:. : 08E.
835328.: 2563‘ 8:302

 

 

 

8:80 Eotdegoﬁowﬁ Ea 3:63.

05 e8 Budgeted 3636:: 28 085m 9:80:00 oeoﬁmdﬁm com me as: use 338309 Bﬁéﬂm 882 :33 1.52sz

 

N 038.

71

.ooﬁgotom 36565" a $0880 2:80:00" mo £085»

 

 

 

I.:om u Mm Avon Iv N 08C. 88m 2s Ecowaﬁ 2: 26am 80520.58 A3" .3 H 08:. 89a 08 Enema—v oﬁ 3209 80332.80 .802
II :3. 830.186. tom. 2. :3”. INN. .wN. :Om. uNN.$EomIE
3:6. I 3. 35¢. ..:.&m. we. INN. *eN. 3. «L m. *oN. madam I an
:50. 8. II 1.3.. 2. 3: RN. :hn. 3. znm. 3. movmbm I n:
:3. :5”. 8.. 3. II :mm. co. 3mm. «5N. ON. «42‘. :3”. «Sm I n:
:Nm. *cN. N~. *Nm. II :vm. the. 22K. :3. 23%. tax. $50& I m0
ihN. a. mN. :. .33.. .II 85v. :3”. :60. :QN. . :3... 85 I m0
Ion. IwN. :. :3”. :3. :3. II :0”. :3. :8. 83m. magnum I m0
:2. «N. 2. Scan: :5. :nm. 85w: II :3. :3. turn. .3me I m0
Ewan. .nN. MN. «5N. *«mv. uvwv. «Ln. *«Nm. II :bn. .«No. 85 I mm
Nv. *nN. IoN. :wv. «Inn. mN. :ow. 2.an :ow. II :Nw. madam I Mm
:wm. a. a. :Nv. 1.0%. 2. :No. :mw. . :60. 8.5%. II moﬁbm I Mm
358. a ”85 . a maﬁa . a ”E. a 358.8 "5.26 «Isaac 355-8 8: . mm 83mm . mm aim .mm

 

 

 

 

223
m 23

72

005800.000.“ 322305 n a: $00080 2:000:00 u m0 S00080$0m
H mm .60 u 3 m 025. 88% 0.8 3.8920 06 9600 0020200000 Act u 3 ~ 08:. 88m 0.8 Ecomﬁv 05 323 mucus—00.80 "mam

 

.- 8. :8. 2. z. 8. 2. S. .2. .2. $an . a

2. .. 005m. 0:. co. 2. 2. 0o. 2. :3. mow—tum . A:
:cm. :mn. u- :vm. :wN. 00mm. 03. :NN. 00m“. “zom. mEm . A:
Nﬁ. co. 2. 3 .39.. 00am. :on. 005m. 36¢. 00;. ﬁﬁom . m0
om. :. :oN. 000m. .. :vm. «0N? inn. 00am. «03.. 85 . m0
2. 8. 2. :mo. :0“. .. :3. :3. :3. :2. $an . 00
2. no. 2. .230. :3. :NB. .. 00mm. 00mm. :3. movmbm .. m0
:. 3. 2. 009‘. :No. :3. 85¢. a- 01%. 01%. 0:5 . mm
cm. 2. mo. :NV. 003:. :06. 00cm. the. u. .550. magnum .. mm
:. vo. N_. :hn. «3m. 2.0m. :mo. 00mm. 00:. .. moﬁbm - mm

 

sham é “3:00 é «5%.: 058-00 25-00 0:30-00 $55.00 85-00 80am .00 £30 .00
:05 E020 02003 05 08 0808080?” 0:0 0005m 0380:00 omoﬁmaﬁmdogﬁmm macaw—0:00 ”_0>0.H #030365
v 2an

 

 

73

Table 5

Team Level: Means Standjard Deviations, and Ranges for Self-Efﬁcacy, Collective

 

Efﬁcacy and Team Perfoerce for the Additive Group.

 

 

Measure Time 1 Time 2

M S_12 _g_Ran e n M £12 _g_Ran 6 Q
SE-Pins 8.08 1.01 3.9 - 9.4 36 8.00 1.03 5.5 - 9.3 32
SE-Strikes 4.86 1.40 1.7 - 8.1 38 5.20 1.45 2.6 - 8.0 30
SE-Spares 6.41 1.30 3.9 - 9.7 36 6.13 1.59 2.8 - 8.9 27
CE—Pins 8.20 1.22 5.1 - 10.0 36 8.21 1.10 5.3 - 9.8 32
CE-Strikes 4.74 1.59 1.4 - 7.6 38 4.64 1.63 1.7 - 7.3 30
CE-Spares 5.87 1.49 2.6 - 8.8 36 5.52 1.49 3.1 - 7.9 27
CE-Points 5.31 1.3 1.8 - 9.4 40 5.43 1.14 3.3 - 7.4 32
TP-Pins 6.95 0.83 4.9 - 8.7 36 7.05 0.92 4.5 - 8.5 32
TP-Spares 5.67 2.32 2.0 - 12.0 36 5.56 3.29 1.0 - 15.0 27
TP-Strikes 2.92 1.73 0.0 - 9.0 38 3.13 2.24 0.0 - 9.0 30
TP-Points 241.88 39.10 168.0 -358.0 40 248.47 47.24 1600-3410 32

 

Note: The number of teams used for each measure varies. Only those teams that
demonstrated consensus for individual level beliefs were used in the team analyses. SE =
Self-efﬁcacy; CE = Collective efﬁcacy; TP = Team performance. .

74

Table 6

learn Level: Means. Standard Deviations, and Ranges for Self-Efﬁcacy, Collective
Efﬁcacy and Team Performance for the Interdependent Group.

 

 

Measure Time 1 Time 2

M. L _g_Ran e u M 52 _g_Ran e p.
SE-Pins 7.96 1.07 4.4 - 9.7 81 7.83 1.03 5.1 - 9.3 45
SE-Strikes 5.02 1.17 2.8 - 8.3 80 4.90 1.27 2.6 - 7.6 46
SE-Spares 6.18 1 15 2.8 - 8.8 78 5.86 1.33 2.9 - 8.3 46
CE-Pins 7.85 1.27 4.2 - 9.9 81 7.69 1.42 3.6 - 9.3 45
CE-Strikes 4.39 1.31 1 1 - 7.6 80 4.26 1.50 1.9 - 7.5 46
CE-Spares 5.50 1.40 1.7 - 8.5 78 5.10 1.45 1.8 - 8.3 46
CE-Points 5.53 l 1 3.5 - 8.1 85 5.16 1.03 3.1 - 8.4 48
TP-Pins 6.69 0.93 4.5 - 8.8 81 6.55 0.83 4.9 - 8.4 45
TP-Spares 4.90 2.36 0.0 - 12.0 78 5.00 2.59 1.0 - 14.0 46
TP- Strikes 2.85 1.90 0.0 - 10.0 80 2.96 1.73 0.0 - 7.0 46
TP-Points 232.92 40.30 164.0 -37l.0 85 231.63 40.15 156.0 -371.0 48

 

Note: The number of teams used for each measure varies. Only those teams that

demonstrated consensus for individual level beliefs were used in the team analyses. SE =

Self-efﬁcacy; CE = Collective efﬁcacy; TP = Team performance.

75

Table 7

Means and Standard Deviations for Team Concept and Incentive Megsures for Task Tyne

at Time 1 and Time 2

 

 

 

Item Additive Interdependent
Time 1 Time 2 Time 1 Time 2

M S_D M §D M 5.12 M §2

Degree to which you feel that you and 8.39 1.85 8.31 1 84 8.62 1.63 8.69 1.54

your partner are a “team.”

Degree to which you and your partner are 8.30 2.02 8.45 1.67 8.73 1.42 8.55 1.71

“working together”

Degree in which you and your partner are 8.25 2.05 8.23 1.98 8.50 1.63 8.30 1.86

“coordinating your efforts”

Amount of effort you put in this 8.28 1.50 8.23 1.73 8.20 1.60 8.06 1.58

tournament

Importance of bowling well in this 7.58 2.60 7.66 1.99 7.29 2.41 7.28 2.21

tournament

 

Note: there were no statistically signiﬁcant differences between the additive and

interdependent team on any item (2 >05). Time 1: n = 250; Time 2: g = 160.

76

MorAanes

Time 1: Individual level. I hypothesized that task type would not moderate the
relationship between self-efﬁcacy and individual performance (Link A) or the
relationship between disaggregated collective efﬁcacy and individual performance (Link
B). In order to test these hypotheses, the procedures described by Baron and Kenny
(1986) were used. Regression analyses were run for each performance measure (i.e., Pins,
Spares and Strikes). The dependent variable for each analysis was the performance
measure; whereas the predictor variables were SE, CE, Task, SE by Task, and CE by
Task. All 125 teams (250 individuals) were used in this analysis.

In order to test for moderating relationships at the individual level, collective
efﬁcacy scores were “disaggregated.” That is, for each team, the team members that
comprised the team were given the same collective efﬁcacy score. The E test for the
change in _R_2 was used to test the hypothesis that task type was a moderator variable. The
results of these analyses are presented in Table 8.

The results of the individual level analysis show that a signiﬁcant moderator
effect was obtained only when Pins was used as the dependent variable. More
speciﬁcally, the CE by Task interaction (Link B) produced a signiﬁcant increase in the
_R_2. Follow-up bivariate regression analyses demonstrate that, for the additive condition,
CE-Pins was not a signiﬁcant predictor of Pin performance (13%, Adjusted 33 = .02, E
(1,76) = 1.21, 2 =28; [3 = .13; n = 80); whereas it was for the interdependent condition
(28%, Adjusted ﬁ= .08, E (1,168) = 13.92, p =.003; [3 = .28; g = 170). When Spares or

Strikes were used as the dependent variable, there were no moderating effects. The

77

relationship between self-efﬁcacy and perfomrance (Link A) was not moderated by task

type for any of the performance measures (Pins, Spares, or Strikes).

78

$00050 0.600200 n m0 £00050£0m H mm ”942..

 

 

8. S.- 3. 3.- 8. 3. 88. 3. 8. 38.00 a
8. 2. . 2. 8. 8. 8. 8. 3. 8. 38.8 0.
8. 2 .- 8. 8.. K. 2. 88. 3. 8. 3.8 a
3. 8. :. 8. 8. 2. 38. 3. 8. 00 a
8. 2. :. 2. 8. 3.3 3. 3. 8. mm :
3me

8. 8.- 2. 8.- 8. 8.2 m8. 3. 8. 23.00 G
8. 8.- 2. 3.- 8. 22 8°. 8. 8. 023.8 a.
8. 8. a. 82 2. 8.” 3. 8. 3. 0.3- a
8. 2. 2. 2. 8. m8. 88. 8. 8. 00 a
3. 3. 2. 3. 88. 2.2 8. 8. 8. 00 c
020%

3. 8.2 2. 8. 3. 3... 8. 8. 8. 0.3.00 a
:. s.- 2. 8.- 8. 2 .o 88. 8. 8. 2.8-.00 0.
2.. . 8.- 8.2 8.- 8. 32 3°. 8. R. .08. a
8. 2 .- 2. :.. 8. 88 3. 3. R. 00 a
3. 8. 2. 8. 38. :.2 8. 8. 8. mm a
28

re .32 0 a 000 m m m. 8.82.5 .0 m 2832

 

 

 

 

. m 000 .0. 00205 00§§0m00m 0:0 >000mwm 0003000 338300002 ”—0.02 Ellsﬂzuaﬁil; .0. E. E...

w 030-2.

79

Time 1: Team level. This analysis was used to investigate the hypotheses
corresponding to Links C and D in Figure 1. Although it was not necessary to aggregate
the individual level collective efﬁcacy perceptions to represent a higher level construct,
this analysis was needed for self-efﬁcacy. The results of the consensus analysis resulted
in a few teams being excluded from the subsequent regression analyses (for Pins: 3 of
excluded teams = 8; for Strikes: n of excluded teams = 7; for Spares: n of excluded teams
= 11). Correlations among aggregated self-efﬁcacy, collective efficacy and team
performance for Pins, Spares and Strikes, at the team level, can be found in Table 9.
There were 117, 114, 118, and 125 teams used in the Pin, Spare, Strike, and total Points
analyses, respectively.

The team level regression analyses were the same as the individual level analyses;
that is, the team performance measures were used as the dependent variables, and
aggregated self-efﬁcacy and collective efﬁcacy scores were the predictor variables. Task
type was entered as a moderator variable (see Table 10). For this analysis, I expected that
collective efﬁcacy would be a more important predictor of team performance for the
interdependent task (Link D). I did not expect to ﬁnd a relationship between collective
efﬁcacy and team performance for the additive task, but rather that aggregated self-
efﬁcacy would be a stronger predictor of team performance for the additive task (Link
C).

The results of the team level analysis were the same as the individual level
analysis. The only signiﬁcant moderator effects were obtained for Pins. Follow-up
bivariate regression analyses demonstrated that, in the additive condition, CB-Pins was

not a signiﬁcant predictor of team Pin performance (13%, Adjusted 33 = .02, 1: (1,34) =

80

0.58, p =.45; B = .13; g = 36), whereas it was for the interdependent condition (39%,
Adjusted 1:2 = .15, E (1,79) = 13.45, p =.OOO4; B = .38; g = 81). When Spares and Strikes
were used as the dependent variable, there were no moderating effects. Task type did not
moderate the relationship between aggregated self-efﬁcacy and team performance, for
any of the performance measures.

Task type also did not moderate the relationship between CE-Points and
performance at the team level. This analysis was only done at the team level because the
interdependent group did not have individual level performance data.

Therefore, for Time 1, for performance as measured by Pins, task type did
moderate the relationship between collective efﬁcacy and performance (supporting Link
D). This moderating effect was evident at both levels of analysis (i.e., Links B and D).
Thus, the effect of task type on the collective efﬁcacy - performance relationship was
isomorphic (i.e., functionally similar) across levels. Task type did not moderate the
relationship between self-efﬁcacy and individual performance at the individual level
(Link A) or between aggregated self-efﬁcacy and team performance at the team level

(Link C).

81

Table 9

Time 1: Team level: Correlations among Aggregated Self-Efﬁcacy, Collective Efﬁcacy
and Team Performance for the Additive and Interdependent Groups

 

 

CE Performance

Additive Group
Pins (Ll = 36)

SE .81 ** .32

CE .13
Spares (r_r = 36)

SE .77** .35

CE .34
Strikes (Q = 38)

SE .68” .20

CE .19
Points (n = 40)

CE .51 **
Interdependent Group
Pins (g = 81)

SE .71 * * .27*

CE .38"
Spares (n = 78)

SE .65“ .10

CE - .04
Strikes (r_1 = 80)

SE .65" .22

CE .20
Points (1; = 85)

CE .28"

 

Note: SE = Self-efﬁcacy; CE = Collective efﬁcacy.

82

$00050 9:200:00 u mu $00050$0m H mm ”Bdrm

 

 

2.. 8.- 3.0 2.2.. 2.. 8. 3o. 2. 8. 0.3.00 8
8. 2. 8.8 8.2 :. 8.3 8. 2. 8. 00,2- 8
8. 3.. 3.2. 2.2 8. 8.2 2. 2. 8. 00 c
058

8. 8. 8. 8. 8. 83. 88°. 8. 8. 28-700 8
8. 2. 8. :. 8. 2. 83. 8. 8. 23.8 c.
8. 2 .- 2.2 8.- 8. 8. 88. 8. _ 8. 25 8
8. 2. 8. 2. 8. 8. 83. 8. 8. mu 8
8. 2. 8. 2. 8. 3.8 3. 3. 8. mm 2
nova-um

8. 8.- 2.. 8.- 8. 2.2 23. 8. 8. 0.88.00 a
a. 8. 8. 8. 8. 8.2 8. 8. 8. 23.8 a.
8. 2.. 3.8 2.3 2. 8.3 8. 8. 8. 08H 8
3. 2. :.. 8. 8. 8. 88. 3. 8. 00 a
2.. 8. 2.. 2.. 3. 3.2. 3. 3. 8. mm 2
880m

8. 8.3 8. 8. 8. 8.8 3. 2. 2.. 023-20 8
8. 2.3- 8. 8.. 8. 088. _88. 2. 8. 23.8 c.
2.. 8.- 8.2 2..- 8. 82 8. 2. 8. 88 8
8. 8.- 8. 8.- 8. 28 8. :. 8. 00 8
3. 8. 3. 8. 88. 3.2 8. 8. 8. mm 2
8:

.0.-8.“ .26. m a me m a M 308805 «a M 0.8232

 

 

Q 000 O 83.2 00008-8202 000 000mm— 000§0m 00m .82 00380002 _0> 8002. 2 08E.
2 03mg.

83

Time 2: Indiﬁduaﬂevel. Using the same statistical procedures, the results of the
individual level moderator analyses are shown in Table 11. The number of individuals
used for these analyses varied according to the measures used. There were 159
participants for Pins and Spares, and 160 participants for Strikes. Task type did not
moderate any of the efficacy - performance relationships (i.e., Links A and B). In order to
replicate the Time 1 results, task type should have been a signiﬁcant moderator of the
disaggregated collective efﬁcacy - individual performance relationship (Link B) when
Pins was used. Task type by itself did account for a signiﬁcant portion of the variance in
Pin performance. However, when all predictors were entered into the analysis, this Beta
weight for task type was no longer statistically signiﬁcant. Task type did not moderate
the relationship between disaggregated collective efﬁcacy and performance for Spares or
Strikes. It also did not moderate the self-efﬁcacy - individual performance relationship
(Link A) for any of the performance measures (i.e., Pins, Spares, and Strikes).

Time 2: Team level. Similar to Time 1, a consensus analysis was conducted on
the individual level self-efﬁcacy perceptions prior to aggregation. The results of the
consensus analysis resulted in a few teams being excluded from the subsequent
regression analyses (for Pins: 3 of excluded teams = 3; for Strikes: n of excluded teams =
4; for Spares: n of excluded teams = 7). Thus, the number of teams used in each analysis
were 77, 73, 76, and 80 for Pins, Spares, Strikes and Points, respectively. Correlations
among aggregated self-efﬁcacy, collective efﬁcacy and team performance for Pins,
Spares, Strikes and Points for the additive and interdependent groups are found in Table

12.

84

The results of the moderator analysis (see Table 13) show that task type was not a
signiﬁcant moderator in any of the analyses. Thus, the signiﬁcant results for Pins from
Time 1 were not replicated for the relationship between collective efficacy - team
performance (Link D). Similar to the individual level analysis for Time 2, task type by
itself accounted for a signiﬁcant portion of the variance in Pin performance, but when all
predictors were entered into the analysis, the Beta weight for task type was no longer
statistically signiﬁcant.

The failure to replicate the collective efﬁcacy - performance relationships at both
levels of analysis (i.e., the individual and the team) maybe due the decrease in statistical
power that was associated with the decrease in sample size from Time 1 to Time 2. In
order to test this postulation, all of the Time 1 analyses were re-run using only those
participants who were present for both time periods. The n for the individual analyses
was 160, and the n’s for the team analyses varied according to the measures used. The
results of the analyses using the reduced sample were similar to those reported in the
Time 1 section. Links B and D were signiﬁcantly moderated by task type (for Pins only).
Links A and C were not moderated by task type. Summaries of these analyses are

presented in Appendix I.

85

800050 30002200 u m0 ”200050-225 H mm 2%

 

 

8. 8. 2. 8. 8. 88. 888. 8. 2. 23.80 a
2.. 8. - 8. 2. . 8. 32 88. 8. 2. 23.8 c.
8. 8. 2.. 8. 8. 8. 38. 8. 2. 23. 8
8. 8. 2. 3. 8. 8. N8. 8. 2. mo 8
8. 2. 2. 2. 2. 2.8 8. 8. 2. mm 2
nova-am

2.. 8. - 8. 2.- 8. 8. 38. 8. 8. 008.00 a
8. 8. 8. 8. :.. 2. 28. 8. 8. 23.78 0.
8. 2. 22 8. 3.. 8. 38. 8. 8. 25 a
2. 8. 2. 8. 2. 82 8. 8. 8. 00 a
8. 2. z. 2. 38. 8.8 8. 8. 8. mm 2
8.80m

2. 8. 2. 8. 2. 2.8- 8. 2. 8. 008.00 9
8. 22- 2. 8.- 2. 82 8. :. 8. 25.8 a.
8. 8. 82 2 2. 8. 82. 8. 2. 8. 23.8 a
8. 3. - 2. 8.- 8. 82. 8. 8. 8. 00 a
8. 8. 2. 8. 88. 88 8. 8. 8. mm c
a:

22 as. m a 00m 0 a .0. 805205 .0 2 2282

 

 

 

 

omommm 000§0m 2m 20522 00080002 20.62 02200205 um 08:-

: 0309

86

Table 12

Time 2: Team level: Correlations among Aggregated Self-Efﬁcacy, Collective Efﬁcacy
and Team Performance for the Additive and Interdependent Groups

 

 

CE Performance

Additive Group
Pins (g = 32)

SE .56" .34

CE .20
Spares (1_1 = 27)

SE .68" .26

CE .29
Strikes (n = 30)

SE .80" .l 1

CE .19
Points (13 = 32)

CE .42* *
Interdependent Group
Pins (r_1 = 45)

SE .73" .33*

CE .41"
Spares (n = 46)

SE .78" .20

CE .18
Strikes (g = 46)

SE .77" .10

CE .15
Points (9 = 48)

CE .23

 

Note: SE = Self-efﬁcacy; CE = Collective efﬁcacy.

87

Sumo-Eu 2:60:00 u m0 Somaﬁoﬁom H mm ”oluolz

 

 

mm. an: an.» 3.». mm. ca. 8. 2. mm. 2.8-FED Am
on. 2.”. 2.? mod». 2. cm." mo. 3. 5m. 28H AN
2o. 9.. 5.6 «22 oo. cod :- 2. cm. m0 2
250m

on. mm. . 3. 2. - on. m 2. moo. mo. 3. 28-2....mo G
S. E. on. 3. Na. 8. Soc. 8. S. x8hamm 3.
mm. 8. on; we. mm. no. So. mo. :- xmﬂ- Am
2m. 2m. mm. mm. vm. on; No. no. 2. mo R
$- 2. - Nv. 2. - om. mm. 8. 3. 2. mm 2
825m

ow. S.- $. 2. . om. co. So. no. um. awn-Hams G
v». 3.- on. 2. - mm. om. go. no. hm. ﬁat-um ?
me. mm. 36 Em.— an. an. voc- 8. cm. V282. Am
mm. mm. 3. 3.. 2m. 8.— 8. 8. ea. mu 3
8. 2. 2.. 8. 8. 23. 8. 8. 8. mm 2
88am

3. 3. 2m. 2m. 2m. no.2 8. 2m. 3.. 2,8th 3
2m. :.7 mm. mm: 8. mm. moo. om. mv. 28.2...mm Av
cm. 3. - 5.— mo: we. 32¢ no. cm. 3. 28-2. Am
ma. S. 2. 3. mo. 2: 8. m 2. mm. mo Am
2. X... E. om. moo. Rd N.- E. cm. mm 2
SE

Tow 35— m a mmm m a M Eon—82: L— M 0.92:2

 

“O 2.8 U de—n— ooGwEOfom mud-m

 

085m coogom 2m :23 20:80on 233 :2on ”m 08E.

 

2 £an

88

_P_ast performance

The results of the above analyses demonstrated that task type moderated the
relationship between (disaggregated) collective efficacy and individual performance
(Link B), and the relationship between collective efﬁcacyand team performance (Link
D), but only when Pins was used as the dependent variable. These results were found
only for Time 1. Consistent with previous research in the sport psychology literature (i.e.,
Feltz, 1982; McAuley, 1985), it is likely that performance at Time 1 may have affected
the relationship between efﬁcacy beliefs and performance at Time 2. In order to examine
the speciﬁc effects of efﬁcacy beliefs on Time 2 Pin performance and to simultaneously
take into account the effects of past performance (Time 1 Pin performance), I calculated
semipartial correlations (cf Pedhazur, 1982). More speciﬁcally, the semipartial
correlations represented the relationship between efﬁcacy and Time 2 performance aﬁer
Time 1 performance was partialed out from Time 2 performance.

These analyses were conducted with the “Pin” variables (SE-Pins, CE-Pins and
Pin performance) because these measures were the only ones to show statistical
signiﬁcance in the previous analyses. Both the individual and team levels of analysis
were examined. The sample size in these analyses was reduced because only those
participants (and teams) who had complete data for both Time 1 and Time 2, and
demonstrated consensus on the appropriate efﬁcacy measures were used. Thus, for the
additive group, n = 64 at the individual level of analysis, and g = 30 at the team level of
analysis. For the interdependent group, n = 96 for the individual level analyses, and g =

43 at the team level of analysis.

89

At the individual level of analysis, for the interdependent group, the semipartial
correlation for disaggregated CE-Pins and Time 2 Pin performance was .23, which was
higher than the semipartial correlation between self-efﬁcacy and Time 2 Pin Performance
(r = .07). For the additive group, the semipartial correlations for disaggregated CE-Pins
and Time 2 Pin performance, and self-efﬁcacy and Time 2 Pin performance were similar,
r = .15 and 12, respectively.

At the team level of analysis, the results for the interdependent group were the
same as the individual level of analysis. The semipartial correlation for CE-Pins and
Time 2 Pin performance was .29, which was higher than the semipartial correlation
between self-efﬁcacy and Time 2 Pin Performance (r = .26). For the additive group, the
semipartial correlations for aggregated SE-Pins and Time 2 Pin performance was .33,

whereas the semipartial correlation for CE-Pins and Time 2 Pin performance was .26.

9O

CHAPTER V
DISCUSSION

The purpose of this dissertation was to investigate the relationships among self-
efﬁcacy, collective efﬁcacy and performance in an additive task versus an interdependent
task, and to test whether or not the relationships differed according to the level of analysis
(i.e., individual vs team). The results of this dissertation have implications for theory,
application and measurement, thereby contributing to the efﬁcacy literature in several
ways.

This study was the ﬁrst to examine the predictive utility of collective efﬁcacy in
an additive setting, and to compare it to a more interdependent setting. The ﬁndings
suggest that task type is a signiﬁcant moderator of collective efﬁcacy and performance, at
both levels of analysis. However, these results were found only when the teams ﬁrst
performed the bowling task (Time 1), and were found only when Pins was used as the
dependent variable. It did not affect the relationship between self-efﬁcacy and
performance, at either level of analysis.

With respect to the relationships among self-efﬁcacy, collective efﬁcacy and
performance, the results of this study support the strong association among these
constructs, and are consistent with what other researchers found within and outside of
sport (e.g., Feltz & Lirgg, in press). However, I expected that the Time 2 analyses would
replicate the relationships found at Time 1. The failure to replicate may be due to the
decreased statistical power associated with the smaller sample size at Time 2. More
speciﬁcally, when the correlations are examined, the correlations between some of the

constructs in the additive group are higher than those found between the same constructs

91

in the interdependent group, but the additive group correlations are not statistically
signiﬁcant and the interdependent group correlations are statistically signiﬁcant.
Furthermore, the regression analyses that were used to investigate the moderating effects
require power. The Time 2 analyses showed the same trends, but the overall results were
not signiﬁcant (see also Appendix I results). Thus, there may not have been adequate
power to detect the moderating effects at Time 2.

Interestingly, the results of this study seem to be dependent on the efﬁcacy and
performance measures. That is, statistically signiﬁcant results were found only when
“Pins” was used as the performance measure, and when “Pins” was used as the focus of
the efﬁcacy measures. Three other efﬁcacy measures were constructed and these
measures corresponded with three performance measures (i.e., Spares, Strikes, and
Points). Although all of the measures were constructed in accordance with Bandura’s
(1986) recommendations, and were concordant (Moritz et al., 1996), the relationships
among the constructs were not evident. The relationships between efﬁcacy and
performance for Pins were consistent throughout the entire analyses. I am not sure why
the other measures failed to show the desired relationships. Bowlers may have been more
accurate in their efﬁcacy judgments regarding the number of pins they could drop than in
the number of strikes or spares they could bowl. Given that the bowlers were beginners,
they did not have much consistency in bowling spares or strikes. In addition, Pins was a
more variable performance measure, than were Spares or Strikes.

In order to circumvent levels of analysis issues with respect to the collective
efﬁcacy measures team members, in this study, made the judgments together. This .

procedure was adequate given that the teams were dyads. However, Bandura (1997) has

92

cautioned researchers against using this procedure when team size is an issue. Team
members are rarely of one mind in their appraisals of matters, and therefore forming a
consensual judgment of a team’s efﬁcacy by team discussion may be subject to social
persuasion and conformity pressures. For example, it is possible that a more prestigious
team member could inﬂuence the judgment in such a way that it does not accurately
represent the views of many of its members. Thus, team members may publicly consent
to a response without truly believing it (Guzzo et al., 1993). The bottom line is that a
forced consensus can be highly misleading, and may be unmanageable with large
collectives.

In this dissertation, consensus analyses were still needed in order to justify the
aggregation of the individual level self-efﬁcacy perceptions to the team level. The
consensus analyses resulted in a few teams being excluded from the team analyses.
However, this number was minimal, and may actually be expected given the
experimental design of the study. More speciﬁcally, the participants in this study were
randomly assigned to two-member bowling teams. They bowled together for
approximately 8 h before competing in the “tournament” phase of this study. In natural
sport settings, sport teams are never randomly determined, and they practice together for
extended periods of time, much longer than 8 hours! That so many teams agreed on their
efﬁcacy beliefs may lend credence to the existence of a sport “culture.”

The results of this study also supported the levels of analysis hypothesis. When it
was a signiﬁcant moderator, the effect of task type was surprisingly apparent at both the
individual and team level of analysis. Thus, this study provides empirical evidence which

supports a multi-level conceptualization of these constructs.

93

Interestingly, the relationship between collective efﬁcacy and performance was
isomorphic across levels of analysis. Isomorphism implies that constructs mean the same
thing across levels, and it exists when the same functional relationship, or functional
process, can be used to represent constructs at more than one level. In this case, task type
moderated the relationship between disaggregated collective efﬁcacy and individual
performance and the relationship between collective efﬁcacy and team performance. I did
not expect task type to moderate the relationship between disaggregated collective
efﬁcacy and individual performance. More speciﬁcally, I thought that one’s belief in
his/her team would not impact one’s individual performance.

Levels of task interdependence range on a continuum from low to high (Zaccaro
et al., 1995). At a minimum, individuals often require resources and support from their
surrounding environment, even if there is little or no coaction with other individuals. A
higher level of interdependence is found in settings where individuals coact to produce an
aggregated product. While there still may be little interaction, characteristics of the
aggregate may still inﬂuence beliefs and performance. At the most complex level of
interdependence, individuals behave in complete concert with each other. Individual
actions are fully dependent upon the actions of others to produce a collective outcome
and such actions cannot be distinguished from one another; only a collective product is
identiﬁable.

The tasks used in this study were designed to represented varying degrees of
interdependence. The additive task had low interdependence among team members -- it
could be completed by each team member working independently, and at his or her own

speed (Zaccaro et al., 1995). The interdependent task, although constrained by

94

considerations such as task difﬁculty, team size, and type of motor skill, was designed to
foster a greater sense of interdependence by requiring team members to coordinate their
efforts, and rely on their partner during the toumament. The degree of coordination in this
task was considerably less than in other sports (i.e., hockey, basketball), although it was
designed to be greater than the additive condition.

It is noteworthy to mention that all of the participants in this study perceived that
they had to coordinate their efforts and work together with their partner in order to be
successful in tournament, not just the interdependent team. Zaccaro et al. (1995) stated
that low degrees of interdependence do not mean that individual team members are not
inﬂuenced. in some manner by the activities of the other team members. Thus, it is not
surprising that the additive team did not differ in their perceptions of coordinated effort
compared to the interdependent team. However, task type still differentially affected the
relationship between collective efﬁcacy and performance. Thus, the results of this study
seem to suggest that it is more than just the perceptions of coordination and “working
together” that account for the moderating inﬂuence of task type on the collective efﬁcacy
- performance relationship. More research is needed to determine what, exactly,
accounted for the difference. A within-subjects design where participants rate the degree
of interdependence for different tasks seems to be the logical extension of this research.

Interestingly, the relationship between aggregated self-efﬁcacy and team
performance was not moderated by task type. I expected aggregated self-efﬁcacy to
predict performance for the additive group, but not for the interdependent group.
Aggregated self-efﬁcacy may be an adequate predictor of team performance when team

attainments represent largely the summed contributions of individual members (F eltz &

95

Lirgg, in press; George & Feltz, 1995). Even though the majority of the correlations
between aggregated self-efﬁcacy and team performance were higher than the correlations
between collective efﬁcacy and team performance for the additive group, task type was
not a signiﬁcant moderator.

The ﬁndings from this study suggest that collective efﬁcacy is a stronger predictor
of team performance for interdependent tasks compared to additive tasks. The
implications of these ﬁndings are most pronounced when one considers applied settings.
That is, when designing interventions to develop, maintain or enhance efﬁcacy,
psychologists should consider the setting in which participants are working. The results
of this study indicate that in interdependent tasks, interventions should be directed
towards the entire team. For example, team goals may be used in interdependent sports.
Ultimately, efﬁcacy inﬂuences performance, whereby those who are highly efﬁcacious
will exert more effort, out-perform, and out-persist their low efﬁcacy counterparts
(Bandura, 1986, 1997).

The implications of these ﬁndings may be extended outside of the sport domain
as well. A common strategy in teaching and coaching is the collaborative or cooperative
approach (Pangrazi & Darst, 1997). This approach is characterized by placing students in
teams in order to enable the students to work together toward common goals. The student
work is arranged so there is an interdependence in the achievement of team goals. The
emphasis is placed on team outcomes rather than on individual outcomes, and the success
of the approach rests on the assumption that the students perceive that they cannot
complete the task alone. This description of student work teams. is analogous to the

interdependence found in some sports settings. Similar to sport, if teachers and coaches

96

want to develop student efﬁcacy in interdependent settings, they may be advised to focus
on collective efficacy enhancement strategies.

Concentrating on developing, maintaining, and enhancing collective efﬁcacy does
not imply that there will be no effects on self-efﬁcacy beliefs. Researchers (e.g., Feltz &
Lirgg, in press; Watson & Chemers, 1998) have shown that self-efﬁcacy and collective
efﬁcacy are related. The correlations between these constructs in this dissertation provide
additional support that self-efﬁcacy and collective efficacy are positively related, but also
that they are distinct constructs. Future researchers may consider examining the effects of
collective efﬁcacy interventions on self-efﬁcacy beliefs and vice versa-

In sport, this is one of only two studies to sirnultaneously explore the relationships
between self-efﬁcacy, collective efﬁcacy and performance on both the individual and
team levels (see also Watson & Chemers, 1998). The generation of multi-level models
have tremendous advantages in that they are uniquely powerful and parsimonious. Multi-
level models postulate relationships among variables which apply at two or more levels
(Rousseau, 1985), and they describe relations at one level that are generalizable to other
levels. One beneﬁt of this approach is that multi-level models may generate testable
hypotheses at one level of analysis that are suggested by empirical results or theory at
another level of analysis. Given the vast literature on self-efﬁcacy and performance, the
research possibilities for collective efﬁcacy and performance are boundless. Furthermore,
multi-level thinking forces a theorist and/or researcher to examine individuals, teams and
organizations as parts of a whole, each affecting and being affected by the other, rather
than as separate conceptual categories (Lindsley et al., 1995). Efﬁcacy research could

beneﬁt from such an approach.

97

Future studies could examine the same relationships as in this study but with
existing sports teams that are additive, such as archery, track, or swimming. In some
Sports, both additive and interdependent tasks exist andare performed by the same
individuals. For instance, in baseball and soﬁball, batting or catching a ﬂy ball are
essentially additive tasks, whereas most ﬁelding (e.g., making a double play) is
interdependent. Situations such as these provide a natural setting for the examination of
the effect of task type on efﬁcacy and performance. Furthermore, these settings offer the
researcher the advantages of a within-subjects design.

This dissertation demonstrated that task type moderated the relationship between
collective efﬁcacy and performance. Researchers might also consider other factors as
potential moderators of this relationship. For example, varying levels of task uncertainty
and
task complexity may differentially affect the association between collective efﬁcacy and
performance. The basic premise is that as task uncertainty or complexity increases,
individuals will rely more heavily on team performance when arriving at efﬁcacy
judgments (Lindsley et al., 1995). Thus, one might expect the relationship between
collective efﬁcacy and performance to be stronger under conditions of high task
uncertainty or complexity than under conditions of low task uncertainty or complexity.

Additional constructs which are suited for examination within a multi-level model
of efﬁcacy and performance include team size. In this dissertation, care was taken to
ensure the team size remained constant under the different task types. However, Zaccaro
et al. (1995) stated that members of smaller teams are generally better able to coordinate

their activities than their counterparts in larger teams. He stated that as team size

98

increases, individuals participate less, exhibit greater disagreements and dissension, and
are absent more oﬁen. As these factors occur, the team’s sense of collective efﬁcacy
would be expected to decline, accompanied by decreases in individuals contributions to
the team. Support for this proposition is found in studies of social loaﬁng that have
demonstrated that as team size increases, individual effort and performance declines
(Hill, 1982; Latane, Williams, & Harkins, 1979). With respect to the effects that team
size can have on the efficacy - performance relationship, Lindsley et al. (1995) stated that
as the size of the team increases, it is likely that the relationship between self-efficacy and
collective efﬁcacy will decrease. In addition, they stated that the relationships between
variables at adjacent levels of analysis (i.e., self-efﬁcacy - individual performance;
collective efﬁcacy - team performance) would be stronger than relationships between
those same variables at non-adj acent levels of analysis (i.e., self-efﬁcacy - team
performance).

There is an alternative suggestion regarding the effect of team size on the efﬁcacy
- performance relationship. More speciﬁcally, a large team can mean that more resources
are available to the team. The greater the number of different resources teams can apply
to a task, the stronger would be the probability of success. In such circumstances, then,
team size may be positively associated with members’ perceptions of collective efﬁcacy.
To date, the moderating impact of team size has not been investigated in any of the
efﬁcacy - performance relationships. There are a number of research options for those
who are interested in pursuing how other factors may affect the collective efﬁcacy -

performance relationship.

99

APPENDICES

100

APPENDIX A
Table A1: Pilot Study - Individual Level: Means, standard deviations, and ranges for self-
efﬁcacy, collective efﬁcacy and individual and team performance measures for task

teams.

Table A2: Pilot Study: Correlations between self-efﬁcacy, collective efﬁcacy, and
performance measures.

l0l

Table A1

Pilot Study - Individual Level: Means. Standard Deviations. and Ranges for Self-
Efﬁcacy, Collective Efficacy and Individual Performance for the Additive and
Interdependent Groups.

 

 

Measure Additive (g = 16) Interdependent (_n = 26)
M _S_12 Range M S_D Range
SE-Pins 76.59 14.20 53.0 - 96.0 74.85 15.64 43.0 - 100.0

SE-Strikes 46.56 20.97 14.0 - 91.0 42.21 17.69 15.0 - 76.0
SE-Spares 63.94 19.22 22.0 - 96.0 61.62 19.49 33.0 - 91.0

CE-Pins 75.97 15.19 52.0 - 93.0 76.98 13.53 47.0 - 100.0
CE-Strikes 91.34 44.99 20.0 - 157.0 96.23 37.17 37.0 - 166.0
CE-Spares 111.19 49.07 10.0 - 200.0 120.23 38.65 57.0 - 193.0

CE—Points 120.91 40.59 55.0 - 195.0 127.54 33.74 63.0 - 196.0

1P - Pins 6.33 0.91 4.9 - 7.7 6.72 1.12 4.8 8.5
[P - Spares 1.94 1.44 0.0 - 5.0 2.69 1.49 0.0 - 6.0
IP- Strikes 1.19 1.22 0.0 - 3.0 1.08 1.06 0.0 - 4.0

IP-Points 107.00 22.19 68.0 - 155.0

 

Note: SE = Self-Efﬁcacy, CE = Collective Efﬁcacy, IP = Individual Performance

102

8528028 23232: n 2 282-28 3888 u mo 5828228 u 8 22 u 8 82m 328m
2: 8.2 08 320mg 06 9.0% 88:22-80 .GN u 8 928w. 252203285 2.: 8m 08 Ecowﬁc 06 323 82226200 ”Mad-Z

 

 

1 hm. 83.. Nm. cm. on. mm. mm. on. .. 3. mﬁm . n:
mm. 1 Nm. Vm. 0N. omv. 3.. ON. mm. 5N. maﬁam .- n:
mN. :... 1 mm. 2. *3... "em. on. em. 2. nova-am - A:
«V. awn. .52.. 350a - A:

.00. enn- mo. 3... 1 .23.. 3.3.. ::. hm. :mn- .mv. ﬂﬁom .. m0
ﬁn. ON. 3.. 04. mm. 1 :Nb. :mo. aimo. «9:. :mw. warm .. m0

enm- VN. —o.- 00.. :Nb. hm. 1 :Na. :Vw. :ah. :mo. 8.8mm .. m0
3.. ov. mo. mm. :2.- Nv. :N». 1 the. fun”. 2.. :- moﬁbm - m0
3. no. :- a. :8. «abh- 2. mm. 1 on. :09 man— . Mm
”v. an. so. am. 2.2.. mm. awn. 3.3. 5v. 1 1.2.. madam . Mm
mm. am. av. Nn. :Nm. :wc. mv. :aw. inn. 35$. 1 ova-um - Mm

8: .8an 328m 358 88 8E 28m 225m 85 23m 9225

1 A: 1 a u n: - 5 .m0 .m0 .m0 .m0 - Mm . Mm -Mm

 

 

9.2032382 05 8m cognac-tom was 085m 9:66:00

 

 

082$ng cock-20m 222228.80 233 226368 2 85m SE

103

AWE

Bowling Instructions

104

APPENDIX B:

TOURNAMENT RULES: Additive Team

Teams consist of two players (Player A and Player B)
Player A and Player B will each bowl a full (10 frame) game

Team score will be determined by adding Player A and Player B’s scores together
(out of 600 points)

TOURNAMENT RULES: Interdeﬁpﬁendent Team
0 Teams consist of two players (Player A and Player B)
0 Players A and B will ALTERNATE who bowls ﬁrst in each frame
For example, for Frames 1, 3, 5, 7, and 9: Player A bowls ﬁrst
for Frames 2, 4, 6, 8, and 10: Player B bowls ﬁrst

If Player A i§_ bowling ﬁrst:
If Player A strikes, then only Player A records a score for that frame.
If Player A does not strike, then Player B bowls the second ball in that frame.

If Player B was bowlig ﬁrst:
If Player B strikes, then only Player B records a score for that frame.
If Player B does not strike, then Player A bowls the second ball in that frame.

FRAME 10

In Frame 10: Player B bowls ﬁrst, if he/she strikes, Player A bowls the second ball, and
Player B bowls the third ball.

In Frame 10: Player B bowls ﬁrst, if he/she does NOT strike, then Player A bowls the
second ball. If Player A SPARES, then Player A also bowls the third ball.

In Frame 10: Player B bowls ﬁrst, if he/she does NOT strike, then Player A bowls the
second ball. If Player A does not Spare, then no more balls are bowled.

0 Two games will be played, whereby each team member is Player A in one game, and

Player B in the other game.
0 Team scores will be the sum of the scores in both games (out of 600 pins).

105

APPENDIX C

Performance Measures/Tournament Scoresheet

lO6

Date:

APPENDIX C:

Performance Memes/TournamentScoreﬂeet

 

 

Names of Players:

 

Summary of Game:

Final TEAM Score:

For PLAYER A:

Individual Score:

 

 

 

Number of Strikes:
Number of Spares:

 

 

Average number of pins dropped on ONLY ﬁrst balls:

For PLAYER. B:

Individual Score:

 

Number of Strikes:
Number of Spares:

 

 

Average number of pins dropped on ONLY ﬁrst balls:

107

APPENDIX D

Self-Efﬁcacy Questionnaires

108

oH
OH
2
0H
0H
H:
OH
OH
CH
CH

oxcxoxmmmaooxcx
00000000000000000000
5555555555
\oxoxosoxoxoxooxoxo
WWWWWWWWWW
vvvvvvvvvv-

$5 $5 00
on 58 H 58.50 55> 53 H 55.50 “23058

0H
OH
OH
oH
0H
OH
OH
oH
oH
oH

maxoxoxaoaoma
00000000000000wa
5555555555
\oxoxosooxoxoxosoxo
WWWWWWWWWW
vvvvvvvvvv

£5 £5 0H0
00 58 H 55.50 55> :8 H 55.80 323058

mmmgom MPH mmH<ZZOHHmmDO mUZmQHnHZOO-m.HmHm

MMMMMMMMMM
NNNNNNNNNN

MMMMMMMMMM
NNNNNNNNNN

”Q XHQZmaHonx

v—iv—v—I—Iv—II—‘w—‘u—o'q—q

OOOOOOCOOC

_3305
00 0258 H

OOOOOOOOOO

_3305
00 «0550 H

383 .5930 :0 5 05% 0 H38 :00
855.3 5030 0 0H 80% 0 H300 56
85 .5930 a 5 80% 0 H300 55
883 5930 5 5 80% 0 H300 :00
85.3 .5930 0 0H 0.5% 0 H300 55
353 .5930 n 5 0.5% 0 H300 50
859a 5930 v 5 05% 0 H300 55
853 .5930 m 5 80% 0 H300 50
850.: .5930 a 5 0.5% 0 H300 H50
«253 .5930 H 5 0.5% 0 H300 50

. 0% ~05 0055.500 .50 83H

” OH. mHHHH ”HOHH

 

850a .5930 :0 5 can“ 0 H38 50

353 .59» .30 m 5 8.3... 0 H300 80
35.2.3 .5930 m 5 8:9. 0 H300 56
853 5930 H. 5 8:9. 0 H300 50
850.0 .5930 0 5 3:5 0 H300 56
358.3 .5930 n 5 8:9. 0 _300 50
888 .50» .30 v 5 8:5 0 H300 500
88$ 59» H0 m 5 SHE... 0 H300 56
883 5930 m 5 Saba a H300 55
«0.55 .59» H0 H 5 8:5 0 H300 :00
330300qu 5935M

. mHZ< H. mHH.H.H. m ..H

 

109

0H
OH
OH
oH
oH
oH
oH
0H
0H
0H

O\O\O\O\O\O\O\O\O\O\
oooooooooooooooooooo

55
00 08 H 5850 b0>

5555555555

soxoxoxososoxososow
WWWWVIWWWWW
vvvvvvvvvv

m5 8
50 H 53.50 HangoEom

mmmmmmmmmm

NNNNNNC‘IOINN

~~~~~~~~~~

OOOOOOOOOO

5 a 55
05 .0550 H

5585500 :3 55 .50.» .53 0.50 0H :0 005 50
550.5500 :3 En .59» .53 A0.50 0 055 5 005 000
5585500 :3 «0.5 .590 553 050 w 385 5 005 000
5585500 HHS Eu .59» 553 A050 5 035 .0 005 .50
5585500 HHS Eu .59» 553 050 0 «55 «a 005 56
558558 HHS 00.6 .59» 553 0.50 n 085 «a 005 50
5585500 :3 05c .59» 553 050 0 505 5 005 55
5585500 =3 55 .59» 5H3 0.50 m «80H 5 005 56
5585500 :3 8.6 .59» 553 «50 N «80H 5 005 56

5585500 :3 00.6 .59» 553 50 H 502 5 005 50
.5 05 00555000 50 80.0
. :00. mHHH.H. 500

 

ll0

APPENDIX E

Collective Efﬁcacy Measures

1]]

oH
oH
oH
oH
0H
3
H:
0H
OH.
OH
. oH
oH
OH
2
oH
oH
oH
OH
H:
oH
0H
OH
OH
oH

QQGOOGOQGOO‘O‘GO‘OQaQGQGQO‘O‘
oooooooooooooooocooocooooooooooooooooooooooooooo

55 00 :00 0.5
55 50:00 50> 05 0 3

050m 55 5 5500 80 05803 000
050» 55 5 5500 2% 0w80>0 55
050» 55 5 5500 own 05530 000
050» 55 5 5500 wmn 05526 000
05% 55 5 5500 com 0555 :00
080» 55 5 5500 2.0 0555 000
05.5 55 5 5500 one own—96 000
05% 55 5 5500 «NV 05805 56
05% 55 5 5500 80 5896 000
050» 55 5 5500 n5m 055$ :00
050m 55 5 5500 own 0555 000
050» 55 5 3500 mmn 0555 :00
050» 55 5 5500 com 0555 55
050» 55 5 5500 2a 055% 500
05% 55 5 8500 own 0555 00.0
080» 55 5 5500 nmm 09295 .50
05% 55 5 5500 com 5826 000
050» 55 5 5500 m5H 0w80>0 000
05% 55 5 00500 onH 0w80>0 :00
050m 55 5 5500 2: 0w80>0 55
08% 55 5 8500 SH 5826 000

25% 55 5 $500 «5 05505 55

0.5% 55 5 3500 on 05505 000

05% 55 5 3500 mm 5826 :00

555555555555555555555555
coocxoxooowsosoxoxosoxooxosoxoxoxosoxoxo
WWWWWWWWWWWWWWVIWWWV‘IWWWWW
vvvvvv-vvvvvvvvvvvvvvvvvv
MMMMMMMMMMMMMMMMMMMMMMMM
NNNNNNNNNNNNNNNNNNNNNNNN
HHHHHHHHHHHHHHHHHHHHHHHH
oooooooooooooooooooooooo

:0

 

5 55 05
55 5550 03
05 .50 9s 55 5550 55 55.50
053050... 05 03 80 03 331%

mH>H<mHH OZHSOm MOnH mmH<ZZOH5mmHDO mUZmnHHnHZOU Emu.

”m NHQmenH<

112

o.
o.
o.
o.
o.
o.
. o.
o.
o.
o.
o.
o.
o.
o.
o.
c.
o.
o.
o.
o.

QOO‘GQO‘O‘QO‘O‘O‘O‘O‘O‘O‘QOO‘O‘O‘

55 05 550 03
55 55500 bo> 25 03

oooooooooooooooooooooooooooooooooooooooo

(\(‘l‘l‘l‘l‘t‘l‘l‘l‘l‘l‘l‘l‘hl‘l‘l‘bh

W‘OO‘O‘OO‘O‘O‘OOO‘OW‘OOO‘O‘Obw

WWWWWWWWWWWWWWWWWWWW

VV’V’VV’V’VVVVVVVVVVVVVV

55 05 050 53 555.80

5535505 5.5 53

MMMMMMMMMMMMMMMMMMMM

NNNNNNNNNNNNNNNNNNNN

u—o—Iu—lv—iv—OI—Iv—OI—i—ﬂ—Ov—dv—‘u—II—‘o—O—nu—l—c—c—i

OOOOOOOOOOOOOOOOOCOO

..5
5 55 05
50550 03
55 55:00
055 03

.50 55 0000555

50555 .50» 00 ..5 5 05.55 5 .305 :50
5055.5 .50» 00 0. 5 55.55 5 .305 :5
SE55 :5» .50 5. 5 8.55 5 .305 :50
50855 .50» .50 E 5 8:55 5 .305 :50
50555 50» .50 5. 5 8.55 5 .305 :50
50555 50» .50 n. 5 05.55 5 .305 :50
5555.5 .5030 v. 5 05.55 5 .305 550
50555 .50» .50 m. 5 8.55 5 .305 550
50855 .5030 N. 5 05.55 5 .305 50
50555.5 .5030 2 5 8.55 5 .305 550
50855 .50» .50 o. 5 05.55 5 .305 550
5055.5 .50» .50 0 5 8.55 5 .305 050
5085.5 .50» .50 5 5 8.5.55 5 .305 :50
55555 .50» .50 h 5 8.55 5 .305 550
50555 .50» .50 5 5 8.55 5 .305 550
5085.5 50» .«0 m 5 05.55 5 .305 550
50555 50» 00 v 5 8:55 5 .305 :50
8:55 .50» .50 m 5 55.55 5 .305 550
50555 .50» .50 N 5 8.55 5 .305 550
50855 .50» .50 . 5 8.55 5 .305 :50

  

 

”HZmE<ZMDOH mat. MO".

113

o.
o.
o.
o.
o.
o.
. c.
o.
o.
o.
o.
o.
o.
o.
o.
o.
o.
o.
o.
o.

O‘O‘ONO‘QGONGGO\O\O\O\O\O\O\O\QO\G
OOOOOOOOOOOOOOOOOOQOOOOOOOOOOOOOOOOOOOO

55 05 .50 03
55 55.500 .0? 05 03

[\hhhbbbhhht‘hhhl‘l‘hl‘l‘h

55555555555555555555
WWWWWWWWWWWWWWWWWWV‘IW
vvvvvvvvvvvvvvvvvvvv

55 0.0 050 03 55500
5530505 05 03

MMMMMMMMMMMMMMMMMMMM

NNNNNNNNNNNNNNNNNNNN

v—ﬂv—Qw—v—Iv—ﬂv-ﬂr-ﬂu—ﬂv-Ov-ﬂv—‘I-‘v-iv—Ov—i—ﬂ—dv—I—‘u—o

BOOOOOOOOOOOOOOOOOOOO

_

5 55 05
.0550 03
55 555500

25 03

5055.... .50.. 50 ..5 5 0505 5 .305 550
5055.... .50.. .50 0. 5 0505 5 .305 550
5055.5 .50.. 00 5. 5 0505 5 .305 :50
5055.5 50. 00 5. 5 0505 5 .305 :50
50555 50.. 50 5. 5 0.505 5 .305 550
5055.... .50.. .50 5. 5 0.505 5 .305 :50
5055.5 .50.. .50 v. 5 0.505 5 .305 550
5055.5 50.. .50 m. 5 0505 5 .305 :50
50555 .50.. .50 N. 5 0505 5 .305 :50
5055.5 .50.. .50 .. 5 0.505 5 .305 :50
5055.... .50.. .50 c. 5 0505 5 .305 550
5055.5 50.. ..0 0 5 0505 5 .305 550
50555 50.. 00. 5 5 0505 5 .305 550
50555 .50.. 00 5 5 0.505 5 .305 :50
5055.5 50.. ..0 5 5 050m 5 .305 550
5055.5 .50.. .50 n 5 0.505 5 .305 550
5055.... .50.. 00 v 5 0.505 5 .305 :50
5055,... .50.. 00 m 5 0505 5 .305 550
5055.5 .50.. .50 N 5 0.505 5 .305 050
5055.5 .50.. .50 . 5 0505 5 .305 :50

 

114

o.
o.
o.
o.
o.
o.
o.
o.
o.
o.

O\O\O\O\O\O\O\O\O\C\
OOOOOOOOOOOOOOOOOOOO

55 05 050 03
55 55.500 b0> 0.5 03

(\bbbbhhbbh

5555555555
WWWWWWWWWW
vvvvvvvvvv

55 05 550 03 55:00
5530505 05 03

MMMMMMMMMM

NNNNNNNNNN

F-lv—tv—‘I-Iiv-Ov—lv—iv—‘u—nu—n

CCOOOOOOOO

5
5 55 8
«2.50 03

“$5 Gmﬂtoo

05 03

5.5355500 :55 55.. .50.. 5.3 550 o. ..5 005 :50
550555500 :55 55.. 50. 5.3 550 0 550. 5 00.5 :50
550555500 :55 5.5 .50.. 5.3 550 5 550. 5 005 550
550555000 :55 55.. .50.. 5.3 550 5 550. 5 005 550
550555500 :55 5.5 .50.. 5.3 550 5 550. 5 005 :50
550555500 :55 2.5... .50.. 5.3 550 5 550. 5 005 :50
550555000 :55 5.5.. .50.. 5.3 550 v .550. 5 00.5 .50
550555500 :55 5.5... 50.. 5.3 550 m 550. 5 005 :50
550555500 :55 5.5 .50.. 5.3 550 m 550. 5 005 .50

550555000 :55 5.5.. .50.. 5.3 50 . 550. 5 005 :50

 

”853 .50 was OOGOUCGOO .m2<m.ﬁ .50 3mm

 

92.52.55.520... m5... m0“.

115

APPENDIX F

Team Concept and Incentive Measures

116

o. 0 5 N. 5 5 v 5 N . 5
55.0005. ..._0> 55.0005. 505.0502 55:005. .02
50555.50. 55 5 :03 .305 0. 50.. 0. 5 .. 55.0005. 305 05055 050.0 .5

o. 0 5 5 5 5 v 5 N . o
5.55 555 550 380502 550 02
5055.50. 55 5 55.50 50.. 05 .0050 5055 30... .v
o. 0 5 5 5 5 v 5 N . o
=5
.5 500.50 50
5.8.50 50 5.5.50 50 5555.508
05550000 505550.. 03 05550000 5530505 03 . .00 05 03.

50555.50. 55 5 :50050 .50.. 55555500.. 0.5 .0550 .50.. .005 50.. 50.53 5 00.50.. 05 050.55 050.0 .5

o. 0 5 5 5 5 .5 5 N . c
:5 .5 005050.
005050. 505050. 550.003
0.003 505.50.. 03 0.503 5530505 03 .0: 05 0?

50555.50. 55 5 :..05050. 550.503.. 05 .0550 .50.. .55 50.. 50.53 0. 00.50.. 05 050.55 050.0 .N

o. 0 5 N. 5 5 v 5 N . o
..5 5 550.
550. 5 0.5 005.50 .5 .55 E50. 5 0.5 505.50 .5 .55 5 0.5 03 00...
05 00... .00.. 5.8.5.5.. . 05 00... .00.. 5530505 . .00.. .05 0.0 .

:58... 5 0.5 .0550 .50.. .055 00.. 55 .00.. 50.. 50.53 0. 00.50.. 05 050.55 050.0 ..
55.239292. thm502oo mm. 0...

5.0.33. .0.. 055555050

u... Magma?

117

APPENDIX G

University Approval for Use of Human Subjects

118

APPENDIX G:

MICHIGAN STATE

U I“ l \I E [K S I 1' Y

June 16. 1997

10: Deborah L. Feltz
138 IN Sports Circle

RE: IRE“: 97-327
TITLE: THE RELATIONSHIP BETWEEN SELF-EFFICACY.
COLLECTIVE EFFICACY PERFORMANCE IR SPORT
REVISION REQUESTED: N/A
CATEGORY: 1-C
APPROVAL DATE: 06/16/97

The University Committee on Research Involving Human Subjects'(UCRIﬁS)
review of this project is complete. I an pleased to advise that the
rights and welfare of the human sub ects appear to be adequately
pﬁotected and methods to obtain informed consent are rgpriate.

re g

abgrefore, the UCRIBS approved this project and any v s ons listed
ve.

 

RENEHAL: UCRIHS approval is valid for one calendar year. beginning with
the approval date shown above. Investigators planning to
continue a project be one year must use the green renewal
form (enclosed with original a roval letter or when a
project is renewed) to seek t certification. There is a
maXimnm.of four such expedit renewatg‘gogsible. Investigators

wis to continue a project beyond ime need to submit it
again or complete revzew.

RSVISIdﬁé: UCRIHS must review any changes in procedures involving human
subjects. prior to initiation of change. If this is done at
the time o renewal, please use the een renewal form. To

revise an approved protocol at an 0 er time during the year
send your written request to the Chair, requesting revised
,approval and referenc the project's IRE i and title. Include
in t a description of the change and any revised

ins ruments. consent forms or advertisements that are applicable.

PROBLEMS]

CHANGES: Should either of the follow arise during the course of the
work. investigators must noti y UCRIES promptly: (1) roblems
( aints. e c.) ving

unexpected side effects comp invol
subjects or (2) changes in the research environment or

new
information indicating greater risk to the human sub ects than
existed when the protocol was previously reviewed approved

If we can be of any future help, lease do not hesitate to contact us
at (517)355-2180 or ﬁll (SI?)4 2- 171.

W

 

lkumsﬁnhds
m’ m a.
was

  

UknbnSHIUﬁeuu

zﬁll‘l . 515
Enusehiuﬂﬁpa
QmMMMS

$N§§ﬂﬂl
FM 51 ”432-1171

   

: Sandra 8. Moritz

Imuuynsuunndy
nushnuuuanmy

anu550a55 '
OGUbaaﬁmiuein
,.,.

119

APPENDIX H

Consent Form

120

APPENDIX H
Dear Bowler,

I would like to request your participation in a research project that is being conducted by the Department of
Physical Education and Exercise Science (PEES) at Michigan State University. The research project will help
me fulﬁll part of the requirements for my Doctoral Degree. The purpose of the study is to determine how
self-conﬁdence and team conﬁdence are related to team performance in bowling.

This project will be completed during your regularly scheduled PES Bowling class. If you are willing to
participate, you will be asked to complete 2 questionnaires (one self-conﬁdence and one team conﬁdence
questionnaire). You will complete the self-conﬁdence questionnaire once. This questionnaire contain 30
items, and you will be asked to circle the most appropriate response. The team conﬁdence questionnaire
contains 74 items, and again, you will be asked to circle the most appropriate response. However, for the
team conﬁdence questionnaire, you and your partner will complete the questionnaire together for a total of 3
times. That is, you will be given one questionnaire for your team on 3 occasions. These questionnaires should
take approximately 20 minutes to complete. In addition, you will be asked to participate in a “team” bowling
tournament. This tournament will also be completed during your regularly scheduled bowling classes.

All of the questionnaires, and your bowling performance results will be treated with strict conﬁdence. They
will only be used by the experimenter listed below, and the project supervisor from the PEES Department.
By signing this consent form, you are agreeing that your results may be used for scientiﬁc purposes,
including publication in scientiﬁc and sport speciﬁc journals, as long as your anonymity is maintained. On
your request, the results can be forwarded to you upon project completion.

Participation in this project is completely voluntary, and you may choose not to participate at all. You may
refuse to participate in any of the procedures, and/or refuse to answer any questions. You may also
discontinue your participation in this study at any time without repercussion. There are no risks involved to
you through your participation.

Participation in the study will not entail any “extra” course work for you, nor will it adversely affect your
ﬁnal grade for the course.

Thank you for your cooperation.
I have read and I understand the above, and I agree to participate in the study.

Signature of Participant:
Date:

 

 

Signature of parent or guardian if participant is under 18 years of age.

 

Sandra Moritz, B.Sc., M.A. Deborah Feltz, Ph.D.

Doctoral Candidate Professor and Department Chair
PEES Department PEES Department

Michigan State University Michigan State University

Inquiries: (517)432-1416

121

APPENDIX I
Summary of Reduced Sample Analysis for Time 1

Table 11: Reduced Sample: Individual Level: Means, Standard Deviations, and Ranges for
Self-Efﬁcacy, Collective Efﬁcacy and Individual Performance for the Additive and
Interdependent Groups

Table 12: Reduced Sample: Individual Level: Correlations Between Self-Efﬁcacy,
Collective Efﬁcacy, and Performance for the Additive and Interdependent Groups

Table 13: Team Level: Means, Standard Deviations, and Ranges for Self-Efﬁcacy,
Collective Efﬁcacy and Team Performance for the Additive and Interdependent Groups.

' Table I4: Reduced Sample: Individual Level: Moderation Analysis between Efﬁcacy and
Performance (Links A and B).

Table 15: Reduced Sample: Team level: Correlations among Aggregated Self-Efﬁcacy,
Collective Efﬁcacy and Team Performance for the Additive and Interdependent Groups

Table I6: Reduced Sample: Team Level: Moderation Analyses Between Efﬁcacy and
Performance (Links C and D)

122

APPENDIX I:

Table 11

Reduced Sample: Individugl Level: Means. Standard Deviations, and Ranges for Self-
Efﬁcacv, Collective Efﬁcacy and Individual Performgnce for the Additive and
Interdegandent Groups

 

 

Additive Interdependent
Time 1 Time I
M S12 Range M S_Q Range

SE-Pins 7.87 1.36 3.40 - 9.70 7.82 1.33 3.20 - 10.00
SE—Strikes 4.82 1.78 0.00 - 8.20 4.70 1.38 2.00 - 8.60
SE-Spares 6.37 1.65 1.50 - 10.00 5.89 1.50 2.40 - 8.80
CE-Pins 8.12 1.43 4.40 - 10.00 7.74 1.36 4.20 - 9.60
CE-Strikes 4.62 1.67 1.40 - 7.55 4.02 1.32 1.10 - 6.65
CE-Spares 5.64 1.55 2.55 - 8.15 5.21 1.48 1.70 - 8.35
CE—Points 5.44 1.37 1.83 - 9.38 5.14 0.94 3.46 - 7.83
1P: Pins 6.88 1.15 3.70 - 8.70 6.73 1.12 3.90 - 9.40
1P: Spares 2.86 1.76 0.00 - 7.00 2.51 1.47 0.00 - 6.00
IP: Strikes 1.42 1.10 0.00 - 6.00 1.28 1.21 0.00 - 5.00
1P: Points 121.58 25.17 77.00 - 187.00

 

Note: n = 96 for interdependent group, n = 64 for additive group. SE = Self—efﬁcacy; CE
= Collective efﬁcacy; IP = Individual performance. Collective Efficacy data is
disaggregated.

123

 

 

doggota H5232: u E 300080 03000—30 u m0 $0850.23 H mm Ave n
5 @5on 03:200. 05 0.8 Raowﬂv 05 0.690 30:20:00 ”Ga n 3 95% 803039005 05 03 Ecowﬁu 08. 323 0:330:00 ”082

 

1 “Samoa .. A:
the. . 1 no. 0mm. :NN. 2. mo. :mm. 2. 0mm. f. «08am - a
:mo. vor 1 :wm. 3. Vet mo. . mo. aof No. CC. m0vmbm . &
:mw. 005m. 005m. 1 :Nm. a. 0.0m. «VN. 2. 2. we. 35 - a:
:hm. .33.. N". 0 ~ m. l 2.3.. :mm. :hw. :30. come. :Nm. ﬂﬁom . m0

No. no. no: :. :Nm. 1 :on. :Vm. :oo. «cmv. 8.. ~ M. £5 .. m0
:wm. :3”. 2 . .1. _v. :05. :hm. n. :ne. :3». 2.00. :ov. madam .. m0
com. :3”. 2. VN. .32.. 01%. :0”. 1 «emv. .33. Elk. moﬁbm . m0
ﬂ. 2. we. awm. :wm. :nb. :Nn. :30. 1 :3. :No. 83 - mm
:30. :wm. .oN. «on. :on. VN. game. each. :Nm. 1 :on. moamm - mm
~N. cm. 2. 2. 0.5V. «on. fan. 33m. 00mm. comb. .. movmbm - Mm

 

as£é 83% A: aim A: are“: 058-8 25-00 ”$8.00 80.3me 35 2% mesa .3 ”3:3 .mm

 

NH 28d.

124

00528803 Ema. u AS. $00080 3:00:00 H MD $0850.30m H mm 000305 8000 05 E 000: 0003
$028 _0>0_ 3:23—05 .8.“ 03:00:00 0803:0800 006 3800 00.05 bco .0092, 050008 £000 .8.“ @003 0800“ me 0028:: 05. 0|qu
3 842m - 8.0.2 Vném 34mm on 8.3m . 8&2 86m and: ﬂﬁom . “E.

125

 

 

3 93. - 86 $4 $.m om cod - oo.o a: ca.~ Sim RE.
2. 8.8 . 8.— Vm.N 86 mm 8.2 .. cod wed and 83am AB.
3 n5.” - 34. $6 $6 on on.” - oaé :3 mad SEAS.
3 m3. - Sam 3d 36 mm and - mm; mm; 36 ﬂﬁoméu
3 mm.w - o: o: and mm m; .. mm.~ «2 5d mﬁaqmmo
3. 3.0 - 24 mm; moé on was - 92 $4 2.6 m0xEm-mU
3. cod - omé 3..— NE. on 8.3 .. 36 mm; mm.» mﬁméo
3 no.» - end on; 36 mm on.» - mmé a: ”v.0 00:26-3
3. 3.5 - 2mm 0: $4. on no.» - on; $2 3.». 000mbw-mm
3 ohm - ové v: 3: on ovd - nnd we; no.” 0.5QO
m 0mg 1w .2 m 005m mm .2.
200530005 0>Evu< 0.5802

 

 

 

 

 

03:25. 05 com 00§§ot0m 800,—. 28 Nomowmm 0380—60 Soaoﬁmuﬁm com manna 28 8200300 “0.828% £802 “_0>0q EmuH
2 033.

”a E .92. u a .8. u a .88 ..... 2.8.: m .8. u
80303285 ”2030885 200,—...m0 280$:me 05 8m 82205

.2 2. u a .8. u m .8. u 38.: w. .8. u H. 88.88 8: 885 328...
0 $2 8m 0028800 mo ”965
3.26:5 300050 260280 u mo “200050.920m H mm 9012.

Mm. 20085.83 00558.23 Em 2 0282.9, 0ﬁ m

 

 

8. 8. 2. 8. 8. 88 88. 8. 2. 0.3.00 a
2.. E. - 2. :. - 2.. 8.8 So. 8. :. garmm a.
8. 2. 8. 8. G. 38 88. 8o. 8. gay 2
a. 8. 2 2. 8. 8. 2.8 so. 08. 8. mo 3
8. 2. 2. 8. 8. 8.8 88. 88. 8. mm 2
82 u a 828m

2. 9.. - _m. 8. - 2. 82 S. 2. 8. 0.3.8.00 8
8. 3. - om. 8. - 8. 8.2 8. :. 8. 0.5.8 c.
2. 8. 2: 2:. on. 3:. 8°. 2. 8. 82. a
a. 2. 2. 2. 8. 88 8. 2. a. 00 a
8. 8. 2. on. 8. 8.2 2. S. 8. mm 2
32 u a 83%

8. 8. 2. 8. 8. 8.8 8. 2. 8. 0.3.700 a
8. 2.2- 2. 8..- 8. 83 8. 8. 8. 85.28 a.
8. 2 _. 2.2 cm. on. 88 m8. 3. 2. 0.3.8 a.
2.. 8. - 8. 8. - 8. 28 8°. 8. 2. mo 9
8. 8. 2. a. 8. on.“ 8. 8. 2. mm c
A 82 u a 82

we :52 m a 00m 0 a M 808803 R a 28222

 

 

085m 200253 20 85‘ 2222032 20.64 132305 “2 88m @0030m

 

E 033.

126

Table 15

Reduced Sample: Team level: Correlations among Aggregated Self-Efﬁcacy, Collective

Efficacy and Team Performance for Additive 4nd Interdependent Grows

 

 

CE Performance

Additive Group
Pins (n = 30)

SE .83" .24

CE .04
Spares (n = 28)

SE .77" .41*

CE .42*
Strikes (_n = 30)

SE .74** .11

CE .1 1
Points (r_1 = 32)

CE .51”
Interdependent Group
Pins (r_1 = 46)

SE .74" .21

CE .25
Spares (n = 46)

SE .71 ** .31 *

CE .08
Strikes m = 44)

SE .74" -.05

CE .12
Points (13 = 48)

CE .34*

 

Note: SE = Self-efﬁcacy; CE = Collective efﬁcacy.

127

.222... 2. N2. ".22 22. u 22N. 5| .2 .82.... -mm 2228.223 .2. .2285 22222222.. 222 22.2 .22N. I 2
.3. u m .mw.mI I 2.2.2.. 22 N do." .Nlm 228.222.2223 822228282022 22222 222 8282.222, 222 .20 $3 8.2 22022225822 no 222620 2202202222022:
2822222202222 222222.30 2§ow222w2m 02.22 2022 2022222222 2.222.226.2292 2082.220 2:28:00" m0 $0850.22omu mm wl2oz

 

22. 22. . 2.2.2 222 . 22. 8. 8. 2N 2.2.. , 2222.822
22. 8.- 8.22 22.2 - 22. 8.2 22. 2. 2.2.. 222223
8. 22.. 2N... N22 8. 22.2.2 2 2. 2 2. 22.. mo 22

82-322522
22. N2. 22. 22. 22. 22. 2o. 22. NN. 2222.220 22
2N. 22. - 2.. 22. - 22. 22.. 28. 2.2. 2. 225.2222.
22. 2N. 2.2 22. 2.2.. 2 2. N8. 8. 22. 22222 22
22. 2o. 2N. 2o. 22. 82 8. 8. 22. mo 2N
22. 8. N2. 22. 22. 2.2. 282. 228. No. 222 22

A222. n 9 2822.25

128

 

22. 2.22.2 - 22. 22. - 22. 22.N 8. 22. 22.. 225.2622
22.. 22. 22. 22. 222. 22.. 2o. 2 22. 22222.22 22.
22. 2N. N22 2N2 .N2.. 22. 22. 2. 22. 2222.2 22
2.2. 2N. 2.2.. 22.. 2.2. NN. 822. 2.2. 2.2. mo 2N
2.. 2N. 2.2. 2.2.. 8. 2N2 2 2.2. 2.2. 22. mm 22
22.2. n 5 22.2225

2:. SN 2N. 2.2.. 2. 22. N 2.2. 2 2. 22. 22222.26 22
22. 32 - 22. 2.. - N2. 222. 8. 8. 2N. 2222222222.
22. N2. - 222.2 NN. - 2N. 2N 2 No. 8. 2N. 222.22. 22
22. 22.. - NN. 22. - 22. 22. 28. 22. 2N. mo 2N
8. 22. 2N. 22.. 8. 8.2. 8. 8. 2N. mm 22
222. u a 2222.2

2.52 2252 .22 2 22mm 22 22 2 2582222 .22 22 22222222222

 

 

2 032.2.

REFERENCES

129

REFERENCES

Bandura, A. (1977). Self-efﬁcacy: Toward a unifying theory of behavioral
change. Psychological Bulletin, 84, 191-215.

Bandura, A. (1982). Self-efﬁcacy mechanism in human agency. American
Psychologist, 37, 122-147.

Bandura, A. (1986). Social foundatimf though_t_and action: A social cogitive
My. Englewood Cliffs, NJ: Prentice Hall.

Bandura, A. (1990). Some reﬂections in reﬂections. Psychological Inguig, 1,
101-105.

Bandura, A. (1997). Self-efﬁcacy: The exercise of control. New York: Freeman.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable
distinction in social psychological research: Conceptual, strategic, and statistical
considerations. Journal of Personalig and Social Psychology, 51, 1173-1182.

Chase, M.A., Feltz, D.L., Tully, D.C., & Lirgg, C. D. (1997). Sources of
collective and individual efﬁcacy of collegiate athletes. Manuscript submitted for
publication.

Crocker, J ., & Luhtanen, R. (1990). Collective self-esteem and in-team basis.
Journal of Personality a_n_d Social Psychology, 58, 60-47.

Dansereau, F., Alutto, J .A., Yammarino, F.J. (1984). Theog testing in
organizational behavior: The variant mg. Englewood Cliffs, NJ: Prentice-Hall.

Barley, PC. (1993). East meets West meets mideast: Further explorations of
collectivistic and individualistic work teams. Academy of Management Journal. 9, 319-

348.

130

Feltz, D.L. (1982). Path analysis of the causal elements in Bandura's theory of self-
efﬁcacy and an anxiety-based model of avoidance behavior. Journal of Personalig and
SocﬂPsvchologv, 42. 764-781.

Feltz, D.L. (1988). Self-conﬁdence and sport performance. In KB. Pandolf (Ed.).,
Exercise and Sport Science Reviews (pp. 423-457). New York: Macmillan.

Feltz, D.L. (1992). Understanding motivation in sport: A self-efﬁcacy perspective.
In G.C. Roberts, (Ed.)., MotivLﬁon in sport and exercise. (pp 93 - 105). Champaign, IE:
Human Kinetics.

F eltz, D.L. (1994). Self-conﬁdence and performance. In D. Druckman & R.A.
Bjork (Eds.)., Learning, remembering. believing: enhancing human perforrn_a_n_c_§ (pp.
173-206). Washington, DC: National Academy Press.

Feltz, D.L. & Chase, M.A. (1998). The measurement of self-efﬁcacy and
conﬁdence in sport. In J .L. Duda (Ed.)., Advancements in sport and exercise psychology
measurement (pp. 65 - 80). Morgantown, WV: Fitness Information Technology.

Feltz, D.L., & Lirgg, CD. (in press). Perceived team and player efﬁcacy in
hockey. Journal of Applied Psychology.

Feltz, D.L., & Mugno, DA. (1983). A replication of the path analysis of the causal
elements in Bandura's theory of self-efﬁcacy and the inﬂuence of autonomic perception.
Journal of Smrt Pachology, 5, 263-277.

Forsyth, DR. (1990). Group dynamics (2nd Ed.). Paciﬁc Grove, CA:
Brooks/Cole.

George, J .M. (1990). Personality, affect and behavior in groups. Journal of

Applied Psychology, 75, 107-116.

131

George, TR (1994). Self-conﬁdence and baseball performance: A causal
examination of self-efﬁcacy theory. Journal of Spprt and Exercise Psychology, 16, 381-399.

George, T.R., & F eltz, D.L.(1995). Motivation in sport from a collective efﬁcacy
perspective. International Journal of Smrt Psychology, 25, 98-116.

George, T.R., Feltz, D.L., & Chase, M.A. (1992). The effects of model similarity
on self-efﬁcacy and muscular endurance: A second look. Journal of Sport and Exercise
_P_sychologv. 14. 237-248.

Gist, ME. (1987). Self-efﬁcacy: Implications for organizational behavior and
human resource management. Academy of Mangement Review. 17. 472-485.

Gist, M.E., & Mitchell, TR. (1992). Self-efﬁcacy: A theoretical analysis of its
determinants and malleability. Academy of Management Review. 17. 183-211.

Gould, D., Weiss, M. R., Weinberg, RS. (1981). Psychological characteristics of
successful and nonsuccessful Big-Ten wrestlers. Journal of Sport Pavehology,_3, 69-81.

Gully, S.M., Devine, D.J., & Whitney, DJ. (1995). A meta-analysis of cohesion
and performance: Effects of level of analysis and task interdependence. Small Group

Research 26, 497-520.

 

Guzzo, R.A. & Shea, GP. (1992). Group performance and intergroup relations in
organizations. In M.D. Dunnette, & L.M. Hough (Eds.)., I-Iandbook of industrialand
organizational psychology (pp. 269-313). Palo Alto, CA: Consulting Psychologists Press.

Guzzo, R.A., Yost, P.R., Campbell, R.J., & Shea, GP (1993). Potency in groups:
Articulating a construct. British Journal of Social Psychology, 32, 87-106.

Hill, G.W. (1982). Group versus individual performance: Are N + 1 heads better

than one? Jouraal of Applied Psvcholgv. 91. 517-539.

132

Hodges, L., & Carron, A.V. (1992). Collective efﬁcacy and team performance.
International Joumal of Sport PsychologLﬁ, 48-59.

Indik, B. (1968). The scope of the problem and some suggestions toward a
solution. In B.P. Indik & F.K. Berrien (Eds), People. grogps and organizations. New
York, NY: Teachers College Press.

James, LR. (1982). Aggregation bias in estimates of perceptual agreement.
Journal of Applied Psychology, 67, 219-229. _

James, L.R., Demaree, R.G., & Wolf, G. (1984). Estimating within-group
interrater reliability with and without response bias. Journal of Applied Psychology, 69,
85-98.

Kaplan, LT. (1997). Relationships between group efﬁcacy, goals, and
performance: A comment on Silver and Bufanio (SGR, 1996). Small Group Research, 28,
556-558.

Kenny, D.A., & La Voie, L. (1985). Separating individual and group effects.
Journal of Personalitvgand Social Psychology, 48, 339-348.

, Kerr, N.L. (1989). Illusions of efﬁcacy: The effects of group size on perceived
efﬁcacy in social dilemmas. Journal of Experimental Social Psvcholggygi, 287-313.

Klein, K.J., Dansereau, F., & Hall, RJ. (1994). Levels issues in theory
development, data collection, and analysis. Academy of Management Review. 19. 195-
229.

Kozlowski, S.W.J., & Hattrup, K. (1992). A disagreement about within-group
agreement: Disentangling issues of consistency versus consensus. Journal of Applied

Psychology. 77. 161-167.

l33

Latane, B., Williams, K., & Harkins, S. (1979). Many hands make light the work:
The causes and consequences of social loaﬁng. Jo_um_al of Personﬂty and Social
Psychologv. 37. 822

Lichaz, F.M., & Partington, J .T. (1996). Collective efficacy and true team
performance. International Journ_al of Sport Psycholagv. 27. 146-158.

Lindsley, D.H., Brass, D.J., & Thomas, J.B. (1995). Efﬁcacy-performance spirals:
A multilevel perspective. Academy of Management Review. 20, 645-678.

Little, B.L., & Madigan, RM. (1997). The relationship between collective
efﬁcacy and performance in manufacturing work groups. Small Group Research. 28. 517-
534.

Maddux, J.B. (1995). Self-efﬁcacy theory: An introduction. In J .E. Maddux (Ed.).,
Self-efﬁcacy. adaptation, and adjustment: Theory. research and application (pp. 3-33).
New York: Plenum Press.

Mahoney, M.J., & Avener, M. (1977). Psychology of the elite athlete: An

exploratory study. Cogpitive Therapy and Research, 1, 135-141.

McAuley, E. (1985). Modelling and self-efﬁcacy: A test of Bandura's model.

Journal of Spgrt Psychology, 7, 283-295.

Moritz, S.E., Mack, D.E., & Feltz, D.L. (1996). A meta-analytic investigation of the
self-efﬁcacy and performance relationship in sport and exercise. Journal of Applied Smrt
Psychology, 8, S53.

Moritz, S.E., & Watson, C. B. Levels of analysis issues in group psychology:

Towards multilevel theory of efficacy. Manuscript in preparation.

I34

Multon, K.D., Brown, S.D., & Lent, RW. (1991). Relation of self-efﬁcacy beliefs
to academic outcomes: A meta-analytic investigation. Journal of Counsellm' g Psychology,
3, 30-38.

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

Ostroff, C. (1992). The relationship between satisfaction, attitudes, and
performance: An organizational level analysis. Journal of Applied Psychology. 77. 963-
974.

Pangrazi, R.P., & Darst, P.W. (1997). Dypamic physical education for seconm
school students. Needham Heights, MS: Allyn & Bacon.

Parker, L. E. (1994). Working together: Perceived self- and collective-efﬁcacy at
the workplace. Journal of Applied Social Psychology. 24. 43-59.

Paskevich, D.M. (1995). Conceptual and measurement factors of collective
effrcacv in its relation to cohesion and mrformance outcome. Unpublished doctoral
dissertation. University of Waterloo, Ontario.

Pedhazur, EJ. (1982). Multiple Regr_ession in Behavioral Research (2nd Ed.).

New York, NY: Holt, Rinehart and Winston.

Porter, G. (1992). Collective efﬁcacy, and esteem: Measurement and study of

‘

‘self’ attributes at a gpoup level. Unpublished doctoral dissertation, The Ohio State
University, Ohio.
Prussia, G.B., & Kinicki, AJ. (1996). A motivational investigation of group

effectiveness using social-cognitive theory. Journal of Appliedmchology, 81, 187-198.

135

 

Riggs, M.L., & Knight, RA. (1994). The impact of perceived group success-
failure on motivational beliefs and attitudes: A causal model. Journal of Applied
Psychology, 79, 755-766.

Rousseau, D.M. (1985). Issues'of level in organizational research: Multi-level and
cross level perspectives. Research in Organisational Behavior, 7, l-37.

Shamir, B. (1990). Calculations, values, and identities: The sources of

collectivistic work motivation. Human Relations 43, 313-332.

 

Shaw, ME. (1971). Group dynamics: The psvcholagLof small group behavior.
New York: McGraw-Hill.

Sherif, M., & Sherif, CW. (1969). Social psvcholoav. New York, NY: Harper &
Row, Publishers.

Shrout, PE. (1995). Measuring the degree of consensus in personality judgments.
In P.E. Shrout, & S. Fiske (Eds), Personality research methods and theory. (pp. 79-92).
Hillsdale, NJ: Lawrence Erlbaum.

Silver, W.S., & Bufanio, KM. (1996). The impact of group efficacy and group
goals on group task performance. §_r_n_all Group Research. 27. 347-359.

Spink, KS. (1990). Group cohesion and collective efﬁcacy of volleyball teams.
Journal of Sportﬁ and Exercise Psvchology,_1_2, 301-311.

Steiner, ID. (1972). Group process and productivity. New York: Academic Press.

Watson, C.B., & Chemers, M.M. (1998). The rise of shared perceptions: A
multilevel analysis of collective efﬁcacy. Manuscript submitted for publication.

Weldon, E., & Weingart, LR. (1993). Group goals and group performance.

British Journal of Social Psycholggy. 32. 307-334.

136

Zaccaro, S.J., Blair, V., Peterson, C., & Zazanis, M. (1995). Collective efﬁcacy.
In J .E. Maddux (Ed.)., Self-efﬁcacy. adaptation, and adjustment: Theorygesearch apa
application (pp. 305-328). New York: Plenum Press.

Zander , A. (1971). Motives and goals iagroups. New York, NY: Academic
Press.

Zander, A., & Medow, H. (1963). Individual and group levels of aspiration.

Human Relations. 16. 89-105.

137

 

 

 

"143.5“, i]
xiv . tall. . . 1)!
\Is'.lnll\|..!|pou

ii I
1 . riﬂed lilyuﬂlﬂrl
{1.2) "Ruled. nﬂhwﬂuJa‘uJ
v r. n lﬂnuil .

‘53.

n.)

islem as...
i.