TEAMMATE EFFICACY AND TEAMMATE TRUST:
AN EXAMINATION OF TEAM DYNAMICS IN VOLLEYBALL DEFENSE
By
Lori Dithurbide

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements for the degree of
DOCTOR OF PHILOSOPHY
Kinesiology
2011

ABSTRACT
TEAMMATE EFFICACY AND TEAMMATE TRUST:
AN EXAMINATION OF TEAM DYNAMICS IN VOLLEYBALL DEFENSE
By
Lori Dithurbide
Very little research in the context of sport has examined the constructs of teammate
efficacy and teammate trust. Past research in the counseling and organization psychology areas
suggest that efficacy in other group members (Lent & Lopez, 2002) and trust in teammates
(Dirks, 1999) could have a significant impact on team performance and other group dynamic
constructs. The purpose of this study was to examine the constructs of teammate efficacy, its
relationship to self-efficacy, collective efficacy and team performance, and furthermore to
examine teammate trust as a potential moderator between teammate efficacy and team
performance. Eighteen girl’s club volleyball teams were studied across an entire competitive
season answering questionnaires of efficacy, trust and backing-up behaviors at three time points
during their season. It was found that that the relationship between collective efficacy and
teammate efficacy became stronger over the course of a competitive season. Teammate efficacy
was not found to be a significant predictor of team performance. Teammate efficacy and
teammate trust were highly correlated (i.e., showed multicollinearity) and therefore a moderating
relationship between these variables to predict team performance was not tested. It was also
found that decreased teammate trust significantly predicted increased feelings of having to backup or cover for teammates while accounting for collective efficacy. Findings from this study
contribute to a better understanding of these constructs in the sporting arena and offer a starting
point for future research in this area.

Dedication
To my family and friends who have always shown the greatest efficacy and trust in me.

iii

Acknowledgments
A special thank you to my advisor and dissertation committee chair Dr. Deborah Feltz.
Thank you to my dissertation committee members, Drs. Dan Gould, Dale McKnight, and John
Hollenbeck for providing guidance and direction on this project.

Thank you also to the athletes, coaches, and club directors who volunteered their time and effort
to make this study possible.

Lastly, this project would not have been possible without the help of Kyra Dewyer, Lauren
Dinverno, Jason Giecek, Kaitlynn Osborn, Lisa Griffiths, and Graig Chow.
Your help is very much appreciated!

iv

TABLE OF CONTENTS
List of Tables

vii

Chapter 1
Introduction
Purpose of the Study
Contextual Factors
Hypotheses
Research Questions
Delimitations
Definitions

1
9
9
10
10
10
11

Chapter 2
Review of Literature
Self-efficacy and Performance
Collective Efficacy and Performance
Other Efficacy (Teammate Efficacy)
Trust

12
12
14
18
23

Chapter 3
Method
Participants
Measures
Procedure
Data Analyses

29
29
29
35
37

Chapter 4
Results
Descriptives
Hypothesis Testing
Summary of Hypotheses and Results

40
40
41
46

Chapter 5
Discussion
Strengths and Limitations
Implications
Future Directions
Conclusion

48
53
56
57
58

Appendices
Appendix A
Appendix B
Appendix C
Appendix D
Appendix E

78
80
82
84
86

v

Appendix F

88

References

90

vi

LIST OF TABLES
Table 1

Athlete Demographics

61

Table 2

Means and Standard Variations for Team Variables

62

Table 3

Team Level Bivariate Correlations Between Aggregated
Self-Efficacy, Aggregated Collective Efficacy, Teammate Efficacy
and Teammate Trust

63

Between Level Bivariate Correlations Between Aggregated
Self-Efficacy, Aggregated Collective Efficacy, Teammate Efficacy
and Teammate Trust

64

Correlations at Time 1 Between Collective Efficacy, Teammate
Efficacy, Teammate Trust, and Pre- and Post- Performance Measures

65

Correlations at Time 2 Between Collective Efficacy, Teammate
Efficacy, Teammate Trust, and Pre- and Post- Performance Measures

66

Correlations at Time 3 Between Collective Efficacy, Teammate
Efficacy, Teammate Trust, and Pre- and Post- Performance Measures

67

Correlations at the Between Level Between Collective Efficacy,
Teammate Efficacy, Teammate Trust, and Pre- and Post- Performance
Measures

68

Table 9

Multilevel Model for Mean Points Scored Against

69

Table 10

Explanatory Power at the Within Team and Between Team Levels
for Mean Points Scored Against

70

Table 11

Multilevel Model for Mean Point Differential

71

Table 12

Explanatory Power at the Within Team and Between Team Levels
for Mean Point Differential

72

Table 13

Multilevel Model for Backing-up Behaviors

73

Table 14

Explanatory Power at the Within Team and Between Team Levels
for Backing-up Behaviors

74

Table 15

Multilevel Model for Causal Dimension Scale-Teams: Stability

75

Table 16

Explanatory Power at the Within Team and Between Team Levels
for Stability

76

Table 4

Table 5
Table 6
Table 7
Table 8

vii

Chapter 1
Introduction
The nature of interdependent sport teams, is just that: teammates are dependent upon each
other in order to perform successfully. No matter how talented an athlete may be, or how much
he or she may hypothetically carry the team on his or her shoulders, one athlete cannot win a
team’s championship. Athletes who compete in interdependent sport teams, such as basketball,
hockey, baseball, and volleyball must depend on their teammates to fulfill their own roles within
the team in order for the team to perform well. In order for teams to be successful, each
individual athlete must carry his/her own weight and successfully perform his/her role. This
means that each teammate must, in some ways, trust that other teammates will fulfill their own
tasks. When athletes are confident in their teammates’ abilities and trust that they will perform to
their potential, they can then focus on their own tasks and in turn, the team, as a whole should
perform to its potential.
Before I introduce the construct of teammate efficacy, it is important to present a brief
overview of the construct of self-efficacy to provide a foundation on which teammate efficacy is
based with regard to this study. Efficacy beliefs are the core of Social Cognitive Theory (SCT).
In SCT, human behavior is referred to as a triadic, dynamic, and reciprocal interaction of
personal factors, behavior, and the environment (Bandura, 1986). According to SCT, an
individual's behavior is uniquely determined by each of these three factors (Bandura, 1986).
Individuals possess self-beliefs that allow them to control their thoughts, feelings, and actions,
which in turn control how they behave (Bandura, 1986). Bandura defined self-efficacy as a
person’s judgment of his/her ability to organize and accomplish courses of action deemed
necessary to perform to one’s potential. However, a large number of sports, and other goal-

1

striving tasks for that matter, are not performed by a single person. Collective efficacy is “a
group’s shared belief in their conjoint capabilities to organize and execute the courses of action
required to produce given levels of attainments” (Bandura, 1997, p. 476). The collective efficacy
beliefs a team holds can influence the tasks teams choose to complete, the amount of effort the
team puts forth toward a task, how a team perseveres in the face of adversity and the team’s
ability to bounce back after a set-back (Bandura, 1997). In short, efficacy beliefs do not represent
the level of ability that individuals or teams possess; rather they are what athletes believe they
can do with their abilities in a given circumstance (Bandura, 1997).
Although efficacy belief research in sport and physical activity has focused mainly on
self-efficacy and collective efficacy, the main focus of this dissertation is other-efficacy, or more
specifically, teammate efficacy. Lent & Lopez (2002) have proposed a tripartite view of
relational efficacy beliefs: self-efficacy, other-efficacy, and relation-inferred self-efficacy
(RISE). This proposition was put forth due to Lent and Lopez’s beliefs that the interpersonal
context in the study of self-efficacy was neglected. Lent and Lopez contend that there are many
situations involving close relationships that can serve as a context for building, maintaining,
regaining, and harming one’s sense of efficacy. Even though their focus is on counseling
psychology, one of these situations or contexts can be team sport (Feltz, Short, & Sullivan,
2008). Lent and Lopez define close relationships as “the extent that relationship partners have
mutual impact on one another” (p.258) and “there is ‘high interdependence’ between those in a
close relationship” (p.258). The team sport context, especially sports that are high in
interdependence such as volleyball, provide a good example of a close relationship.
Lent and Lopez (2002) define other-efficacy (i.e., teammate efficacy in this study) beliefs
as each partner’s view of the other’s capabilities and state that these beliefs can be influenced by

2

multiple factors such as perceptions of the other’s performances (e.g., my teammate performed
successfully in the past, therefore I have higher beliefs in their capabilities), beliefs about the
confidence and capabilities of similar others (e.g., I hold low capability beliefs of very similar
athletes to my teammate, therefore my teammate efficacy beliefs towards this said teammate are
low) and third-party views of the other’s capabilities (e.g., coach’s beliefs in the teammate). In
turn, other-efficacy beliefs can potentially influence the type and amount of effort expended in
joint pursuits (e.g., winning a game), reliance on other’s feedback, and persistence intentions.
Although admittedly related, Lent and Lopez (2002) also make a point to differentiate
between collective and other-efficacy. Where collective efficacy is a group (or corporate as
stated by the authors) level agency, other-efficacy is an individual level of agency. Otherefficacy entails an individual’s beliefs about the capability or effectiveness of others in situations
where partners or teammates must work together to produce group outcomes. Because otherefficacy involves the beliefs that partners or teammates hold about each other’s separate
capabilities, it is an individual level construct. Collective efficacy beliefs refer to the shared
perceptions of the group about the conjoint capabilities to produce an outcome. However, Feltz
et al. (2008) argue that Lent and Lopez are measuring other efficacy in the same way as
collective efficacy except that it is in dyad teams rather than teams of three or more (i.e., “Rate
your confidence in your team’s [partner’s] capabilities to perform at a given level”). Because
Lent and Lopez’s main focus was on dyadic relationships, the differences between the two
constructs can be clearer than in larger groups such as sport teams. Specifically, other-efficacy
does not include one’s beliefs in one’s own abilities as a part of the team; the individual is
removed from the perception. Collective efficacy assumes the individual conjointly with the
remaining team members; therefore there is no separation between ‘myself’ and ‘them’. How

3

confident one is in one’s dyad team and how confident one is in one’s partner are clearly
different, but the differences may not be significant with larger teams and when rating one’s
teammates is done at the aggregate level.
Research on other-, or teammate efficacy, has been limited in the sport context, especially
in relation to performance outcomes. Beauchamp and Whinton (2005) examined self- and otherefficacy and their relationship to performance in the dyadic relationship between rider and horse
in equestrian eventing. The authors hypothesized and results supported that other-efficacy would
be able to explain unique variation in riding performance above and beyond that explained by
self-efficacy beliefs alone.
While additional research has examined other-efficacy in sport (see Jackson, Beauchamp,
& Knapp, 2007; Jackson, Knapp, & Beauchamp, 2008), the equestrian study, to my knowledge,
is the only one to examine the link to performance. Consequently, this is an area in the sport
literature that could use more attention. In addition, Lent and Lopez (2002) state that when
successful performances require competencies from an interdependent group, it is advantageous
for the members to possess favorable beliefs about both their own and teammates’ capabilities.
In fact, in these types of situations, they posit that collective efficacy might be conceptually
equivalent to the sum of self-efficacy and other-efficacy ratings. However, in interdependent
teams, the coordination and interaction between each athlete is far more complex than simply the
sum of its parts, therefore Lent and Lopez’s proposition may not hold true in these
circumstances.
Highly interdependent teammates must often depend on each other to fulfill their roles
and successfully perform tasks. Especially in interactive and interdependent sport teams, athletes
must trust their teammates to do their job. Consequently, a lack of trust in each other could

4

potentially lead to changes in cognitive beliefs (i.e., collective efficacy beliefs, worry,
motivation, and satisfaction) and behavioral changes (i.e., backing-up behaviors, and
performance). Where efficacy in one’s teammate represents the beliefs that a teammate can
perform a certain task, or their capabilities, trust in one’s teammate (for the purpose of this study)
represents the beliefs that a teammate will perform a certain task. Even though an athlete has the
capabilities to perform a task, it is not guaranteed that they will when it matters most. Factors
such as anxiety, fatigue, arousal levels, and motivation may influence whether or not an athlete
actually performs up to his/her potential. Although trust seemingly is an important factor in
interdependent groups such as sport teams, very little research has been conducted in this
context.
The concept of trust has been examined on many societal levels but for the purpose of
this dissertation, I will focus on the interpersonal level of trust and more specifically, task-related
trust (i.e., I trust my teammate will perform a task) as opposed to social trust (i.e., I trust my
teammate with my secrets). Because so little research has examined trust in sport teams
(exception: Dirks, 2000), the majority of the theoretical background, framework, and empirical
evidence is derived from the organizational setting. Mayer, Davis, and Schoorman (1995)
defined trust as “a willingness of a party to be vulnerable to the actions of another party based on
the expectation that the other will perform a particular action important to the trustor,
irrespective of the ability to monitor or control that party” (p. 712). Additionally, Rousseau,
Sitkin, Burt, and Camerer (1998) proposed a cross-disciplinary definition of trust as “a
psychological state comprising the intention to accept vulnerability based upon positive
expectations of the intentions or behavior of another” (p. 395). Dirks and Ferrin (2001) further
conceptualize trust as a psychological state, such as a belief or attitude, toward another

5

individual, and not a dispositional construct. In interdependent sport teams, the other individuals
are one’s teammates. An athlete must trust his or her teammate to behave or perform in a way
that works best for the team. For the purpose of this study, teammate trust is conceptualized
along these propositions but is more specifically defined as the belief or expectation that one’s
teammate will effectively perform a particular action that is necessary for one’s own, and the
team’s benefit. It should be clarified that this definition and study does not include perceptions of
benevolence or integrity of the other party (e.g. Mayer, Davis, & Schoorman, 1995), but is solely
about the effective capability of the other person on one’s (or the team’s) behalf. Therefore this
type of trust is actually a trusting belief-competence and was selected because it related closely
to how effective team members are in doing their part on the team. Hereafter, the term “trust”
will refer to the trusting belief-competence type of trust, unless it refers to literature that uses a
broader trust definition.
Although trust has been studied and linked to many constructs in the organizational
setting, much like sport, performance has been a popular outcome variable. Even 35 years ago,
trust was a frequently cited determinant of group performance (Golembiewski & McConkie,
1975). An example in sport is Dirks’ (2000) study of NCAA basketball teams where he
examined the effect of trust in the coach and trust in teammates on team performance. Results
showed a significant relationship between trust in the coach and team performance but no
significant direct relationship was found between trust in teammates and team performance.
Dirks and colleagues (Dirks, 1999; Dirks & Ferrin, 2001) have however argued that even
though the direct relationship between trust and team performance has dominated the literature,
in some circumstances trust may serve better as a moderator between other variables and team
performance. A review of the trust literature comparing both the main (i.e., direct) effects and

6

moderating (i.e., indirect) effects of trust on group outcomes, Dirks and Ferrin (2001) found that
the direct effects of trust on performance outcomes were weak and inconsistent. In exploring the
moderating model of trust, the authors examined if trust provided the conditions under which
higher performance and other outcomes were likely to occur.
According to Dirks and Ferrin’s (2001) argument, trust represents the experiences and
knowledge about an athlete’s teammate. Consequently, trust will affect how an athlete assesses
the future behavior of a teammate and how he or she interprets the actions of a teammate. Their
first proposition is that trust will moderate the relationship between motivational constructs and
outcomes such as performance. This is based on the idea that trust can impact performance but
does not directly cause one to perform or behave in a certain manner. Evidence consistent with
this proposition included a study that found no main effect between trust and performance but
found a moderating effecting of trust between group members’ motivation and group processes
and outcomes (Dirks, 1999).
Research has found a robust relationship between efficacy beliefs (self-, collective, and
recently, other efficacy) and performance. Also, efficacy beliefs are posited to be the primary
determinant of an individual’s level of motivation to accomplish a certain goal (Bandura, 1977,
1997). This is manifested by the choice of activity an individual makes, the effort one puts forth
towards the task and how much one persists in the face of failure. The accordant hypothesis that
might follow is that the level of trust in one’s teammates moderates the relationship between
efficacy beliefs and performance.
For instance, an athlete participating in a sports team may have high levels of beliefs in
his or her teammate’s ability to perform a certain task that is important to the team’s performance
(i.e., efficacy beliefs). However, if this athlete does not trust that his or her teammate will

7

perform this task, for example, in an important game situation when the athlete and/or team is
vulnerable, the team’s performance may not reach its potential, or the distrusting athlete may
compensate by performing differently (e.g., playing “out of position”). If athletes perceive a need
to back-up other teammates, they may put themselves in a position that may compromise their
own and their team’s performance. However, if the team is highly coordinated, backing-up
correctly could also lead to improved performance. In addition, the level of trust an athlete has
for his or her teammates may have an affect on how one attributes success or failure of the team.
Therefore, if there is no trust between teammates, an athlete may not exert as much effort
towards the team’s goal because he or she cannot depend on his or her teammates to do the same.
On the other hand, if efficacy beliefs are high and there is trust between teammates, athletes may
put forth more effort towards the task and worry less about their teammates, allowing them to
focus on their roles and tasks towards the group’s goal. This may consequently lead to an
increase in performance.
A second proposition made by Dirks and Ferrin (2001) states that trust will moderate the
relationship between a teammate’s action and the truster’s (athlete) response. In other words,
trust can affect how one interprets another person’s actions. Consequently, if an athlete has high
levels of trust towards a teammate, he or she is more likely to interpret the teammate’s behavior
favorably than if there were low levels of trust. Consequently, if an athlete holds high levels of
trust towards a teammate, perhaps efficacy beliefs towards this teammate may remain stable over
time, as the perceptions of even questionable performances stay favorable. Furthermore, if trust
levels are lower, perceptions of the teammate’s behaviors should be negative, even if the efficacy
beliefs towards this teammate are high. These perceptions of behavior from teammates could

8

lead to changes in affect, conflict within the group, and potentially, changes in behavior in other
teammates, such as having to back-up these teammates or cover for them on the court or field.
In sum, the sporting arena offers the ideal context to examine the highly interdependent
and relational constructs of teammate efficacy and teammate trust; neither of which have been
studied in the team sport domain to date. Research in the counseling and organizational settings
suggest that teammate efficacy and teammate trust may have significant effects not only on team
performance but also on other team processes. It is for these reasons, and based on this
foundation that I chose to further the research in the sport domain.
Purpose of the Study
The purpose of this dissertation is to examine the constructs of teammate efficacy, its
relationship to self-efficacy, collective efficacy and team performance, and furthermore to
examine teammate trust as a potential moderator between teammate efficacy and team
performance.
Contextual Factors
The context chosen for the current study was sport. Team sports provide an ideal context
in which to examine teammate efficacy and teammate trust, because sport teams epitomize the
intact group. They perform a meaningful task where each member shares a common goal and
member interdependence is very high.
Female club volleyball was selected for this study for both conceptual and practical
reasons. First, volleyball is a highly interactive and interdependent team sport where efficacy
beliefs and trust can play an integral part in the performance of the team. Secondly, volleyball,
especially club volleyball in the Mid-West region, is played mostly if not entirely by females,

9

therefore essentially controlling for gender differences. Finally, club volleyball was chosen
because of the timing of their season, which begins shortly after the end of the holiday season.
Hypotheses
1. Collective efficacy and teammate efficacy will be positively related to each other but
remain two distinct variables (show no multicollinearity).
2. Teammate efficacy is a significant predictor of team performance while controlling for
collective efficacy, trust in the coach, and past performance.
3. Teammate trust will moderate the relationship between teammate efficacy and team
performance where increased levels of teammate trust will strengthen the relationship
between teammate efficacy and team performance, and decreased levels of teammate
trust will weaken it.
4. Teammate trust will negatively predict feelings of having to back-up or cover for
teammates while controlling for collective efficacy.
Research Questions
1. Are there differences in the relationships between efficacy beliefs (self-, collective, and
teammate) and trust over time?
2. Do teammate trust and/or trust in the coach have an effect on team attributions while
controlling for subjective performance?
Delimitations
Because of the specificity of the sample, the findings will be limited to female club-level
interdependent team sports, and thus may not completely generalize to other populations. For
example, the results may not be applicable to collegiate level or male volleyball teams. In
addition, the results may not be applicable to sports with a lower level of team interdependence,

10

such as swimming or gymnastics. Lastly, because some of the measures will be situation-specific
to volleyball at the club level, they are not appropriate for other sports.
Definitions
Collective efficacy: an individual’s belief in his/her team’s capabilities to organize and execute
the courses of action necessary to produce certain levels of attainment.
Other efficacy: an individual’s belief in another’s capabilities to organize and execute the courses
of action necessary to produce certain levels of attainment. For this dissertation, see Teammate
efficacy.
Self-efficacy: an individual’s belief of his/her capabilities to organize and execute the courses of
action necessary to produce certain levels of attainment.
Team performance: a team’s win percentage as well as the average points scored against a team
where a higher win percentage and lower average points scored against reflect greater team
performance.
Teammate efficacy: an individual’s belief in his or her teammates’ capabilities to organize and
execute the courses of action necessary to produce certain levels of attainment.
Teammate trust: the belief or expectation that one’s teammate will effectively perform a
particular action that is necessary for one’s own, and the team’s benefit.

11

Chapter 2
Review of Literature
The purpose of this chapter is to provide a review of the literature that is relevant to the
variables and hypotheses of this study. The chapter begins by providing a foundation for the
hypotheses, which represents an overview of the literature supporting the efficacy beliefperformance relationship. Next, support is presented for the inclusion and importance of other’s,
or teammate efficacy belief in the research field of efficacy belief theory. Lastly, this is followed
by a summary of the literature on trust including general theoretical background, research on
trust in groups or organizational teams (where more of the related research has been conducted),
as well as how this research links to the sport domain.
The efficacy judgments made by individuals with regards to their own or of their teams’
abilities are the result of a complex process of self-, or team, appraisal and persuasion that is
dependent on a cognitive process of varying sources of efficacy information (Bandura, 1990).
These varying sources include, but are not limited to, past performance accomplishments or
mastery experiences, vicarious experiences, verbal persuasion, and physiological states
(Bandura, 1977, 1997). Of all sources linked to efficacy beliefs, past performance has been
deemed the most influential (Bandura, 1997). Additionally, in the sport psychology literature,
performance also has been the dependent variable of greatest interest and research regarding the
predictive strength of self- and collective efficacy beliefs (Feltz et al., 2008).
Self-efficacy and Performance
The study of self-efficacy and performance in sport began in the late 1970s with two
distinct lines of research: the Feltz series of studies and the Weinberg series of studies (Feltz et
al., 2008). Feltz and her colleagues focused more on the relationships between self-efficacy and

12

performance across time (Feltz, 1982, 1988; Feltz, Landers, & Reader, 1979; Feltz & Mungo,
1983); whereas Weinberg and his colleagues focused more on experimental methods using
competitive tasks (Weinberg, 1985; Weinberg, Gould, & Jackson, 1979; Weinberg, Gould,
Yukelson, & Jackson, 1981; Weinberg, Yukelson, & Jackson, 1980). Since this early period,
efficacy-belief research in sport has demonstrated wide generalizability due to the diverse
methodologies used, measurements used, and with a variety of sport and motor tasks. Overall, a
consistent moderate relationship has been found between self-efficacy and sport performance,
which has been supported by a meta-analytic review (Moritz, Feltz, Fahrbach, & Mack, 2000).
Results from this meta-analysis also indicated that when a non-significant or negative
relationship was found between self-efficacy and performance it was likely due to measurement
or methodological issues such as non-concordant efficacy belief and performance measures or
too much time between the assessment of each variable. The relationship between self-efficacy
and performance has also shown to be a reciprocal one. Not only does self-efficacy positively
affect performance accomplishments, but performance also has shown to increase one’s efficacy
beliefs in a temporarily recursive fashion (Feltz, 1982; McAuley, 1985).
It should be noted however that arguments have been made against the positive
relationship between efficacy and performance. Vancouver and his colleagues argued for a
negative relationship between efficacy and performance (Vancouver, Thompson, Tischner, &
Putka, 2002) where better performance leads to increased levels of self-efficacy, which in turn
leads to complacency and overconfidence for the following performance. This results in poorer
performance whereas lower levels of self-efficacy leads to more practice and more effort to
improve in future performances. Vancouver and his colleagues found that self-efficacy was
negatively related to performance within-person over time (Vancouver, Thompson, & Williams,

13

2001; Vancouver, et al., 2002; Vancouver & Kendall, 2006) but positively related to
performance at the between-person level, using computer games and academic performance as
the outcome.
Further research in the sporting and physical activity context did not support Vancouver’s
(Vancouver, Thompson, & Williams, 2001; Vancouver, et al., 2002; Vancouver & Kendall,
2006) propositions. Gilson (Gilson, Chow, & Feltz, in press) found that when using a less
cognitive task (i.e., squat lifting performance), self-efficacy was positively related to
performance at both the within-person and between-person levels when controlling for the
participants’ past performance. In addition, research both in individual and team sport has
examined the relationship between collective efficacy and performance across time and withinperson/within-team and significant positive relationships have been consistently found (Feltz,
Chow, & Hepler, 2008; Myers, Feltz, & Short, 2004; Myers, Payment, Feltz, 2004).
Collective Efficacy and Performance.
The collective efficacy and performance relationship has been studied similarly to the
self-efficacy and performance relationship. There have been experimental and field study
designs, and results have shown similar significant positive findings. Collective efficacy and
performance research in sport began in the early 1990s when Spink (1990) examined the
relationship between both variables in volleyball teams. Using a sample of elite volleyball teams,
Spink found that teams higher in collective efficacy (determined prior to the commencement of a
tournament) finished in significantly higher rankings at the end of the tournament than teams low
in collective efficacy.
In contrast to Spink’s field study, Hodges and Carron (1992) investigated group
performance in female and male triads with different levels of collective efficacy in a muscular

14

endurance task in a laboratory setting. Groups were determined by randomly assigning each
participant to a high or low efficacy group by providing bogus feedback after a pre-test session.
Using this experimental design, the researchers were able to manipulate and control for each
group’s level of collective efficacy. To test the impact of the level of collective efficacy, the
competition was manipulated so that each group would constantly lose to a confederate group.
Results of the study showed that the high efficacy groups improved their performance following
failure whereas low efficacy groups showed a decrease in performance after failure.
Extending this research, Lichacz and Partington (1996) examined performance
differences between true (rowing crews and members of the same basketball team) and ad hoc
groups (comprised of members with no playing history). This study used a rope-pulling task,
which was completed by 25 male undergraduate students. Collective efficacy was manipulated
through bogus feedback. Results showed that the group’s performance was significantly affected
by both factors, although more so by the group’s history than the manipulation through bogus
feedback. Groups with higher levels of collective efficacy and those who were members of true
groups performed better. No other study had examined collective efficacy by incorporating true
groups and performance history in a controlled laboratory setting at the time of the study.
Collective efficacy and team performance research has not been limited to laboratory
studies. Research has been extended into the field, adding to Spink’s original study, with Feltz
and Lirgg (1998). They examined team efficacy with competitive ice hockey teams throughout
an entire season. The authors followed approximately 160 male ice hockey players throughout an
entire competitive hockey season, measuring self-efficacy, collective efficacy, and performance
prior to every game. Collective efficacy was measured by using a hockey specific measure
regarding specific tasks and actions that must be accomplished during a match for a team to be

15

successful in their attempt to win. Performance was measured using the teams’ win and loss
records. Results of the study supported the researchers’ hypotheses that aggregated team efficacy
scores were a stronger predictor of team performance than aggregated self-efficacy scores. It also
supported the notion of a link between collective efficacy and team performance found in
laboratory-based studies. The authors also found that following a failure (i.e., team loss), team
member’s collective efficacy decreased, but their individual efficacy belief did not.
Myers, Payment, and Feltz (2004) continued with this research in ice hockey by
examining the reciprocal relationship between collective efficacy and team performance in
women’s ice hockey. Collective efficacy beliefs and performance assessments were completed
on weekends where teams played each other twice (i.e., both Friday and Saturday). The results
indicated that the level of collective efficacy moderately and positively influenced the team’s
performance for the same day performance (Saturday) when controlling for the previous day’s
(Friday) performance. Results also indicated that previous day’s performance has a small and
positive influence on the next day’s collective efficacy scores. The latter finding supports and
extends the finding of Feltz and Lirgg (1998) that there is a positive influence of previous
performance on subsequent collective efficacy across teams. This factor had been missing and
limiting in similar past research. Notable limitations to the Myers, Payment and Feltz study
include a small number of games and teams therefore restricting the ability to generalize the
findings.
Myers, Feltz, and Short (2004) extended the aforementioned study through their study of
collective efficacy and performance in football teams. Offensive football players (N = 197) from
10 different university teams completed self- and collective efficacy measures within 24 hours
prior to each Saturday afternoon game. Performance measures were obtained by analyzing

16

statistics provided by conference headquarters. Results suggested that aggregated collective
efficacy prior to performance positively influenced subsequent offensive performance, and that
previous offensive performance negatively influenced subsequent aggregated collective efficacy
within teams and across games. Self-efficacy did not show a similar pattern. Within weeks and
across teams, aggregated collective efficacy prior to performance also was a positive predictor of
subsequent offensive performance, and previous offensive performance was a positive predictor
of subsequent aggregated collective efficacy. Consistent with Feltz and Lirgg (1998), aggregated
collective efficacy appeared to positively influence offensive performance whereas aggregated
self-efficacy did not within teams and across games. This line of research demonstrates the
consistent and robust relationship between collective efficacy and sport performance that has
been found in both the laboratory and field setting.
Extending the work on the collective efficacy-performance relationship, research has
examined the role of group cohesion and team-referent causal attributions. Heuzé, Raimbault,
and Fontayne (2006) examined the mediating effects in the relationships between group
cohesion, collective efficacy, and individual performance in professional basketball teams. Male
basketball players (N = 154) completed the French version of the Group Environment
Questionnaire, and a 27-item basketball specific collective efficacy measure, while performance
was assessed by individual statistics. The authors found that collective efficacy mediated the
relationship between the Group Integration-Task (GI-T) dimension of group cohesion and
performance and GI-T mediated the relationship between performance and collective efficacy. (It
should be noted that all analyses in this study were conducted at the individual level, which
poses a limitation to its findings. Both cohesion and collective efficacy are defined as grouplevel constructs and should be examined at the group-level to account for the interdependence of

17

the data). Thus, this finding may indicate that the team’s sense of cohesion towards its task does
not directly predict individual basketball performance, but is cognitively processed as an efficacy
judgment about the team’s collectively capabilities, which in turn, predicts performance.
Other Efficacy (Teammate Efficacy).
Lent & Lopez (2002) have proposed a tripartite view of relational efficacy beliefs: selfefficacy, other-efficacy, and relation-inferred self-efficacy (RISE). This proposition was put
forth due to Lent and Lopez’s beliefs that the interpersonal context in the study of self-efficacy
was neglected. Lent and Lopez contend that there are many situations involving close
relationships that can serve as a context for building, maintaining, regaining, and harming one’s
sense of efficacy. One of these situations or contexts can be team sport. Teammates often spend
many hours training and competing together. Lent and Lopez define close relationships as “the
extent that relationship partners have mutual impact on one another” (p.258) and “there is ‘high
interdependence’ between those in a close relationship” (p.258). The team sport context,
especially sports that are high in interdependence such as volleyball, provide a good example of
a close relationship. Because self-efficacy theory is rooted in social cognitive theory, it is
assumed that the social environment has great influence on our efficacy perceptions, may they be
towards the self, the team, or our teammates. The close relationships that individuals hold with
each other are then consequently a part of the social environment.
Lent and Lopez (2002) define other-efficacy beliefs as each partner’s view of the other’s
capabilities and state that these beliefs are influenced by multiple factors such as perceptions of
the other’s performances, beliefs about the efficacy of similar others and third-party views of the
other’s capabilities (e.g., coach’s beliefs in the teammate). In turn, other-efficacy beliefs can

18

potentially influence the type and amount of effort expended in joint pursuits (e.g., winning a
game), reliance on other’s feedback, and persistence intentions.
Although admittedly related, Lent and Lopez (2002) also make a point to differentiate
between collective and other-efficacy. Where collective efficacy is a group (or corporate as
stated by the authors) level agency, other-efficacy is an individual level of agency. Otherefficacy entail individual’s beliefs about the efficacy of others in situations where partners or
teammates must work together to produce group outcomes. Because other-efficacy involves the
beliefs that partners or teammates hold about each other’s separate capabilities, it is an
individual level construct. That is, other-efficacy (i.e., teammate efficacy) is one’s beliefs about
their teammate’s capabilities. Collective efficacy refers to the shared perceptions of the group
about the conjoint capabilities to produce an outcome, and although it also is measured at the
individual level, it is aggregated up to the group level to represent the collective construct
(Bandura, 1997). Because Lent and Lopez’s main focus was on dyadic relationships, the
differences between the two constructs can be clearer than in larger groups such as sport teams.
However, it should be noted that other-efficacy does not include one’s beliefs in one’s own
abilities; the individual is removed from the perception. Collective efficacy includes the
individual conjointly with the remaining team members; therefore there is no separation between
‘myself’ and ‘them’.
Lent and Lopez (2002) also raise an interesting point with regards to the congruency of
self- and other-efficacy beliefs. They state that when successful performances require
competencies from an interdependent group, it is advantageous for the members to possess
favorable beliefs about both their own and teammates’ capabilities. In fact, in these types of
situations, they posit that collective efficacy might be conceptually equivalent to the sum of self-

19

efficacy and other-efficacy ratings. However, in interdependent teams, the coordination and
interaction between each athlete is far more complex than simply the sum of its parts, therefore
Lent and Lopez’s proposition may not hold true in these circumstances.
Feltz and colleagues (2008) note, however, a difference between Lent and Lopez’s (2002)
conceptualization and operationalization of the concept of other-efficacy. Their previous use of
an other-efficacy scale had a stem that was ability-focused assessment (i.e., rate your partners’
ability), whereas they conceptualize the concept more in line with a confidence-focused
assessment (i.e., rate your partner’s confidence in his/her ability). This raises a similar argument
put forth by Myers and Feltz (2007) with regards to the measurement of collective efficacy that
researchers should aggregate individual responses to collective efficacy items. Item stems should
not ask an individual to assess the team’s confidence in its capabilities because individuals are
not as capable of assessing the group’s beliefs about its abilities as they are in assessing their
own beliefs about the group’s abilities. That is, the group, an inanimate social system cannot
have beliefs, whereas the individuals who make up such a group can. Because this study goes
beyond the dyadic relationship and will be examining teams of more than two members, I chose
to measure other-efficacy, or teammate efficacy, using the ability-focused approach.
The study of other- or teammate efficacy in the sporting context has been limited to only
a handful of recent publications. The first one by Beauchamp and Whinton (2005) examined
self- and other-efficacy in the dyadic relationship between rider and horse in equestrian eventing.
Riders (N = 187) competing in 1-day eventing (combination of dressage, show-jumping, and
cross-country disciplines) competitions participated in the study. Task specific self- and other
(horse) efficacy beliefs were assessed prior to each stage of the competition. The overall purpose
of the study was to examine the relationships between self-efficacy, other-efficacy, and

20

performance. The authors hypothesized that other-efficacy would be able to explain unique
variation in riding performance above and beyond that explained by self-efficacy beliefs alone.
With regards to the dressage portion of the competition, Beauchamp and Whinton found that
other-efficacy was able to explain unique variation in dressage performance beyond that
explained by self-efficacy. The two other events showed no significant relationship between selfefficacy and performance, however the authors noted that this was likely due to the lack of
variation in the performance scores. However, one could suggest that in this situation, selfefficacy may mediate the relationship between other-efficacy (belief in the horse’s capabilities)
and performance.
Although not examining Lent and Lopez’s (2002) tripartite view of efficacy in relation to
performance, researchers Jackson, Knapp, and Beauchamp (Jackson, Beauchamp, & Knapp,
2007; Jackson, Knapp, & Beauchamp, 2008) have furthered the research on other-efficacy in the
sporting context. The first of the two studies examined the interrelationships among the three
forms of relational efficacy within junior tennis dyads. Furthermore, Jackson and colleagues
(2007) examined the relationships between all three forms of efficacy beliefs and the athletes’
perceptions of their commitment to and satisfaction with the current partnerships. In addition,
Jackson and colleagues also examined actor and partner effects. That is, actor effects are when
both predictor and outcome variables occur within an individual (e.g., self-efficacy beliefs
influence individual performance) and partner effects are when the predictor variable affects the
outcome in the partner (e.g. other-efficacy beliefs that an individual holds about their partner
affects how the partner performs). In sum, the three forms of efficacy identified in the tripartite
view of efficacy held by one individual may influence the motives and behaviors toward a
partner, and may also impact the motives and behaviors of that partner. Results showed that self-

21

efficacy and other-efficacy predicted athlete commitment and satisfaction, respectively. Results
also indicated actor effects for both other-efficacy and RISE in relation to self-efficacy so that
believing in one’s partner’s capabilities to perform effectively leads to an elevated sense of one’s
own self-efficacy. With regards to partner effects, only a relationship between self-efficacy and
commitment was found. That is, the results suggest that elevated levels of self-efficacy
experiences by one athlete may increase levels of commitment to the relationship by his or her
partner. Therefore, relational efficacy beliefs may have implications not only for the individual
holding the belief, but other team members as well.
Using a qualitative approach, Jackson and his colleagues (2008) furthered the sport
literature of the tripartite view of efficacy belief by examining the antecedents and consequences
of self-efficacy, other-efficacy and RISE within six international-level athlete dyads. Results
from the interviews revealed four higher-order categories across the three efficacy constructs
containing themes regarding: oneself, one’s partner, the dyad, or external factors. Not
surprisingly, perceptions regarding oneself were most frequently cited as sources of self-efficacy.
These included past individual performance achievements, experience, and pre-competition
preparation. Verbal persuasion from their partner as well as the two intrarelationship cognitions
of other-efficacy and RISE were also found to be sources of self-efficacy. Sources of otherefficacy included experience with partner, partner and dyad past performance accomplishments,
comparison with past partners and other athletes as well as both psychological and physiological
perceptions of the partner (e.g., calmness, strength, etc). Finally, sources of RISE included ones’
own self-efficacy beliefs, one’s partner’s verbal and non-verbal behavior, as well as one’s own
motivation, past mastery accomplishments and physiological factors.

22

With regards to consequences of each efficacy beliefs, there was a distinction between
those that related to intrapersonal and interpersonal concepts. Self-efficacy outcomes included
improved affect, individual performance, greater effort and motivation, as well as greater ability
to concentrate on task-related aspects of their performance. Consequences of other-efficacy
included more open and positive verbal behavior toward the partner, greater responsiveness to
the partner, and enhanced satisfaction in the relationship. However, not all consequences for
other-efficacy were positive. Some athletes thought that elevated levels of other-efficacy might
be related to potential negative affective responses, and even possible breakdowns in
relationships. Lastly, consequences of RISE included enhanced self-efficacy beliefs, relationship
persistence intentions, greater relationship satisfaction and elevated motivation in relational
contexts.
Trust
As mentioned in Chapter 1, the concept of trust has been examined in many societal
levels but for the purpose of this dissertation, I will focus on the interpersonal level of trusting
belief-competence type of trust. Because so little research has examined trust in sport teams, the
majority of the theoretical background, framework, and empirical evidence will derive from the
organizational setting.
Although trust has been studied and linked to many constructs in the organizational
setting, much like sport, performance has been a popular outcome variable. Even 35 years ago,
trust was a frequently cited determinant of group performance (Golembiewski and McConkie,
1975). However, Dirks and colleagues (Dirks, 1999; Dirks & Ferrin, 2001) have argued that even
though the direct relationship between trust and team performance has dominated the literature,
in some circumstances trust may serve better as a moderator between other variables and team

23

performance. Dirks and Ferrin (2001) conducted a review of the trust literature comparing both
the main (i.e., direct) effects and moderating (i.e., indirect) effects of trust on group outcomes.
The review for the direct effects model included 43 studies where 29 of these studies had
behavioral and performance outcomes. Dirks and Ferrin argued that most main effect models
contend that trust about another individual affects how one behaves in interactions with the
trustee. Mayer et al. (1995), supporting this contention, provided a model explaining that a
person’s trust towards another leads to a willingness to risk, which in turn leads to risk taking.
That is, higher levels of trust in a partner increase the likelihood that one will take a risk with the
said partner and/or increases the amount of risk that is assumed. Risk-taking then leads to
positive outcomes such as increased performance. Dirks and Ferrin (2001) found that the direct
effects of trust on performance outcomes were weak and inconsistent, and given these findings
they suggest that more research is needed to better understand the effects of trust on behavioral
and performance outcomes.
In exploring the moderating model of trust, the authors examined if trust provided the
conditions under which higher performance and other outcomes were likely to occur. According
to Dirks and Ferrin’s (2001) argument, trust represents the experiences and knowledge about an
athlete’s teammate. Consequently, trust will affect how an athlete assesses the future behavior of
a teammate and how he or she interprets the actions of a teammate. Their first of two
propositions is that trust will moderate the relationship between motivational constructs and
outcomes such as performance. This is based on the idea that trust can impact performance but
does not directly cause one to perform or behave in a certain manner. Evidence consistent with
this proposition included a study that found no main effect between trust and performance but

24

found a moderating effect of trust between group members’ motivation and group processes and
outcomes (Dirks, 1999).
A second proposition made by Dirks and Ferrin (2001) states that trust will moderate the
relationship between a teammate’s action and the truster’s (athlete) response. In other words,
trust can affect how one interprets another person’s actions. Consequently, if an athlete has high
levels of trust towards a teammate, he or she is more likely to interpret the teammate’s behavior
favorably than if there were low levels of trust.
The evidence that was presented in support for both propositions stemmed from an array
of situations and settings, including both field and experimental studies. The consistency in
results across these settings show potential in both propositions to provide an integrative
understanding of how trust operates thus future research is needed.
A more recent study by Shen and Chen (2007) examined team trust, leadership, and team
performance in both service and manufacturing industries. The authors define team trust as “a
dynamic concept which describes faith in or positive expectations for the thoughts, words and
actions of others and the willingness to rely on others and take risks, and signifies the existence
of team-to-individual and individual-to-team interdependence” (Shen & Chen, 2007, p.644).
Questionnaires measuring each of the variables were sent to companies in both these industries
and results indicated that team trust had a positive relationship to team performance (measured
by task performance and cooperation satisfaction). However the significant relationship only
occurred between relational trust (most similar to interpersonal trust which is the focus of this
dissertation) and team performance in service based industries. Team trust was also found to be
significantly related to team leadership style. That is a concerned type of leadership was

25

significantly related to relational trust in both industries. Lastly, an equation model showed that
team trust was an intervening variable between leadership and team performance.
Also examining trust, leadership and team performance, Dirks (2000) used collegiate
basketball teams as his study sample. The purpose of this study was first to examine the
assumption that a team’s trust in its leader has a significant effect on the team’s performance,
and second, was to examine the more complex mediating properties of trust in the leader on post
and future team performance. Using survey and archival data from Division I and Division III
NCAA basketball teams, results indicated that trust in the leader was in fact significantly related
to team performance. Results also indicated that trust in teammates was not significantly related
to team performance when controlling for other confounding variables (i.e., prior performance,
trust in leader, and team talent). This is in line with research previously mentioned in this
dissertation that trust in teammates or partners does not have a direct effect on performance.
Trust in teammates may play more of a moderating role with other variables and their effect on
team performance. Support was also found for the mediating role of trust in the leader between
past and future performances. The author states that one of the reasons that the inertia in team
performance can be maintained is because past performance affects the trust a team has in its
leader, which then affects team performance.
An interesting extension to his line of trust research, Dirks and his colleague Skarlicki
(2009) most recently examined the relationship between being perceived as trustworthy by others
and its effect on individual performance. To examine this relationship, the authors used social
exchange theory (Blau, 1964) as a framework, which states that individuals take part in exchange
relationships because they expect that they can gain benefits over time. Using both a field and
laboratory study, they found that being seen as trustworthy by coworkers was positively related

26

to the trustee’s performance. In the field study, the interaction between a coworker’s perceptions
of an individual’s capability and integrity predicted his or her performance. That is, perceptions
of capability were more positively related to performance when the individual was viewed as
high in integrity. The laboratory study further explored these results via social exchange theory.
Results from this study confirmed the results from the first (i.e., field study). Furthermore, the
results from the second study suggested that trust mediates the effect of the capability and
integrity on an individual’s willingness to provide resources to a coworker. This demonstrates
that trust is an underlying mechanism in this relationship.
As I have mentioned previously in this chapter and Chapter one, very little research has
been conducted in the sport domain examining the construct of trust. Other than Dirks’ (2000)
aforementioned study in NCAA basketball, no other study has really examined teammate trust as
a group dynamic construct. Past studies have compared trusting behaviors between starters and
non-starters in youth basketball behaviors (McGowan & McGowan, 1991) without really
describing what “trusting” means; other studies have examined the development of general
interpersonal trust through athletic participation (Clark & Gronbech, 1987) where trust was
measured using the “trust fall” often used in team or trust building exercises. Trust in this
instance is not specifically direct towards your teammates in the performance domain. Another
way researchers have examined variants of trust in the sport domain is through the relationship
between participants and organizations. For example, Lynch, Jonson, and Dibbeen (2007)
examined trust in adventure racing. More specifically, their study examined trust in event
organizers using a criminal case in adventure racing where the death of a participant was
determined the responsibility of the event organizers as the main example.

27

As a whole, teammate trust in the sport domain has attracted very little research attention,
and yet, in other comparable domains (e.g. organizational settings), has been extensively studied.
Moreover, in such settings, trust has been found to be an important factor in team performance.
This dissertation aims to further the trust literature in the sport and athletic context.

28

Chapter 3
Method
Participants
Girls club volleyball teams from Michigan, northern Indiana, and the Chicago area were
contacted for recruitment in the study (approximately 30 clubs ranging from 1 to 30 teams each),
18 of who agreed to participate in the study. Two of the 18 teams completed only two time
points of the study, but due to the nature of the analysis all 18 teams were used in the study and
subsequent analyses.
The participants in this study were athletes (N = 180) and their head coaches (N = 18)
from the recruited teams. Teams competed in age groups ranging from U14 to U18 (U14 = 3,
U15 = 1; U16 = 5, U17 = 5, U18 = 4). An ANOVA was run to determine if there were
significant differences between each age group in the study’s variables and no significant
differences were found. Athletes ranged in age from 12 to 18 years (M = 15.27, SD = 1.38), were
members of their respective teams for 0 to 5 years (M = .93, SD = .72), and had played
competitive volleyball starting from the age of 7 to 16 years (M = 11.75, SD = 1.62).
Athletes competed in either a Winter season (January – March), Spring season (April –
June) or with teams that stayed intact for the entirety of both seasons. An ANOVA was run to
verify if the season in which the team competed produced any significant differences in the
study’s variables. No significant differences were found.
Measures
Demographics. A brief demographic questionnaire was administered at Time 1 of the
study and included basic demographic questions such as age, height, usual position played on the
court, starting status, volleyball playing experience (measured by number of years playing the

29

sport), the length of time the participants have been playing with their team and teammates, and
if they play volleyball year round (see Appendix A).
Self-efficacy. A non-hierarchical teammate efficacy scale was constructed specifically
for this study. As per Bandura’s (2006) recommendations, experts in volleyball (collegiate-level
coaches) were consulted and were instructed to produce a list of items of volleyball defensive
sub-skills required by each teammate for the team to be successful. These items were then
piloted with current volleyball players to ensure their relevance and the basic understanding of
the questionnaire. Specifically for the self-efficacy scale, the instructions told participants to rate
their confidence in their own abilities to effectively perform the defensive tasks listed below.
This adheres to Bandura’s (2006) guideline that efficacy scales should be written in terms of can
do to properly assess efficacy beliefs. A total of 5 items were generated (see Appendix B) and
addressed various defensive volleyball skills (e.g., Pass a tough serve; Defend against a hard
driven attack). A Likert-scale ranging from 0% (not at all confident) to 100% (extremely
confident) was used as Bandura (2006) argues that the 11-point scale (i.e., 0 to 10) is more
sensitive and reliable than scales using fewer categories. Even though arguments have been made
towards a narrower response scale (i.e., Myers & Feltz, 2007), the 11-point scale was chosen
because of the longitudinal nature of the data. This scale may be more sensitive to changes that
occur over the course of the season. The options were changed from percentages to a 0-10 scale
at the point of data entry. Percentages were used, as they may be more tangible to young athletes.
The internal consistency reliability for the self-efficacy scale was .85 at Time 1, .87 at Time 2,
and .92 at Time 3.
Collective efficacy. This scale included the same items and response scale as the selfefficacy scale, but instructed participants to rate their confidence in their team’s ability as a

30

whole to effectively perform the defensive tasks (Appendix B). The internal consistency
reliability for the collective efficacy scale was .89 at Time 1, .88 at Time 2, and .94 at Time 3.
Teammate efficacy. This scale also included the same items as the self- and collective
efficacy measures. However, the teammate efficacy scale instructed participants to rate their
confidence in each one of their teammates’ abilities to effectively perform the defensive tasks
listed below (see Appendix C). Because teammate efficacy is an individual-level construct about
specific teammates, each participant responded to the scale for each one of their teammates. The
stem used for the scale was “I believe that my teammate CAN…”. The specific teammate in
question was listed at the top of the scale. Because items were specific to each teammate, and
entered and computed through the use of team matrices, reliability coefficients were not
computed.
Teammate trust. The teammate trust measure mirrored the teammate efficacy measure.
All items used in the teammate efficacy measure were also included in the teammate trust
measure (see Appendix C). However the scale instructed the participants to rate how much they
trusted that their teammates would effectively perform the defensive task during competition.
The stem used for this scale was “I TRUST that my teammate [name] WILL…” where 0%
indicated they did not trust at all and 100% indicated that they completely trusted their
teammates in performing the skills listed in the items. Options were also converted to a 0-10point scale at data entry.
Participants were also asked to answer each of these questions with regards to each
individual teammate because trust is a very individualized belief about another person. An
athlete may trust each teammate differently, especially depending on the position they play. For
example, the skills listed in the measure are skills that all volleyball players have some

31

experience with, however some athletes, depending on their position, may have more experience
performing the skill. This could lead to an increase in ability in the skill and also allow
teammates to form more realistic judgments with regards to the skill. Again, because items were
specific to each teammate, and entered and computed through the use of team matrices,
reliability coefficients were not computed.
Backing-up behavior. In order to begin the exploration of backing-up behaviors in the
sport context, a question was constructed to ask participants their agreement with regard to how
often the participant feels like she has to back-up or cover more of the court for the teammate in
question. To keep consistency with the other scales, an 11-point Likert scale was also used where
0% indicates complete disagreement and 100% indicates complete agreement (Appendix C).
Again, these options were converted to a 0-10-point scale at the time of data entry.
Trust in coach. An adaptation to Dirks (2000) Trust in the Leader measure was used to
assess trust in the coach. The scale used in this study was reduced to four items to reflect more of
the performance and competence factors of trust in a coach, and therefore did not include the
more personal items from the original scale (see Appendix D). The items were rated on a 1 to 10point Likert scale where 0 represented strong disagreement with the statement and 10
represented strong agreement with the statement. A confirmatory factor analysis supported the
2

one-factor structure at each of the three time points (Time 1: ! = 3.76, CFI = .99, RMSEA = .08;
2

2

Time 2: ! = 1.53, CFI = 1.00, RMSEA = .00; Time 3: ! = 9.32, CFI = .98, RMSEA = .17). It
should be noted that even though the RMSEA was not ideal for Time 3, the chi-square and CFI
still indicated an acceptable model fit. Internal consistency reliabilities were .69 for Time 1, .87
at Time 2, and .87 at Time 3.

32

Team attributions. Greenlees, Lane, Thelwell, Holder, & Hobson’s (2005) Causal
Dimension Scale for Teams (CDS-T) was used for the assessment of team-referent causal
attributions in this study (see Appendix E). The CDS-T comprises of 16 items, divided into four
subscales, addressing the participant’s perceived main cause for the team’s performance. The
four subscales include Locus of Causality (items 1, 6, 9, and 13), Stability (items 3, 7, 11, and
15), Team Controllability (items 2, 4, 10, and 14), and External Controllability (items 5, 8, 12,
and 16). Each item was rated on a 9-point Likert scale (ranging from 1 to 9) where higher scores
are anchored by statements reflecting more internal (e.g. “Caused by an aspect of your team”),
more stable (e.g. “Permanent”), more team controllable (e.g. “Your team can control”), and more
externally controllable (e.g. “Controllable by people outside your team”) attributions.
Consequently, lower scores are anchored by statements reflecting less internal (e.g. “Caused by
an aspect of the situation”), less stable (e.g. “Temporary”), less team controllable (e.g. “Your
team cannot control), and less externally controllable (e.g. “Over which no-one has control”)
attributions. For the purpose of this study, and the age of the study’s participants, small
modifications in the wording of some items were changed in order to assure comprehension of
the questions. For example, the word “varies” was replaced with “changes”, and “regulate” was
replaced with “ adjust.” The internal consistency reliabilities for Locus of Causality were .71 at
Time 1, .79 at Time 2, and .81 at Time 3. For Team Controllability, the internal consistency
reliabilities were .77 at Time 1, .92 at Time 2, and .90 at Time 3. The internal consistency
reliabilities for Stability were .83 at Time 1, and .82 at both Time 2 and 3. Lastly, the internal
consistency reliabilities for External Controllability were .72 at Time 1, .75 at Time 2 and .78 at
Time 3.

33

Performance. Performance was measured both subjectively and objectively for this
study. The subjective performance measure was included in Greenless and colleagues’ (2005)
CDS-T where athletes were asked to indicate how they felt their team had performed so far this
season. Subjective performance was measured using a 5-point Likert scale ranging from 0 (not
well at all) to 4 (very well). Objective performance was measured using 4 indices. First as the
team’s match winning percentage calculated by dividing the number of match wins by the
number of matches played. Second, because each match could include 2-3 games, the game
winning percentage was also calculated as a performance index. Third, mean point differential
was also calculated by subtracting the number of points scored against from the number of points
scored for and calculating the mean across each game. Lastly, because this study focuses solely
on the defensive aspects of volleyball, the total points scored against each team were obtained.
Wins and losses are absolute measures of team performance and therefore may not be the best
indicator of team performance. A team may have won a game but still had 20 points scored
against them versus another game where they only had 5 points scored against them. This
example demonstrates that even though the team won both games, their defensive performance
in each may have been quite different. The win/loss records and scores for each team were
obtained from an online scoring website and from the coach when online scores were not
available.
Bivariate correlations for all performance indices, at all time points were computed to
ensure each index was distinct from each other. Even though some performance indices were
highly correlated at one time point (e.g., mean points scores against and mean point differential
at Time 0, r = -.972), the high correlations were not consistent across time points (e.g., mean

34

points scores against and mean point differential at Time 3, r = -.619). Consequently, all five
performance indices were kept and used in subsequent analyses.
Coach rankings. In an attempt to verify the accuracy of the teammate efficacy and
teammate trust measures, coaches were asked to complete rankings for each of the defensive
skills used in the efficacy and trust measures for each of their athletes. Coaches were asked to list
in order, from most effective to least effective, each player for each defensive skill. Lastly,
coaches were asked to also indicate three athletes from the team who would be characterized as
the “go to” girls whom the coach could count on in a very crucial game situation (Appendix F).
Bivariate correlations were computed between coach rankings and teammate efficacy at
all time points. Results indicated a significant negative relationship for all correlations (Time 1: r
= -.572, p< .0001; Time 2: r = -.630, p<.0001; Time 3: r = -.548, p<.0001) indicating that lower
rankings (represented by higher numbers) were significantly related to lower ratings of teammate
efficacy (represented by lower numbers).
Procedure
Permission to conduct this study was obtained from the Institutional Review Board for
Human Subject Research. Following approval, volleyball club directors were contacted via email
inviting the teams from their club to participate in the study. A brief description of the study and
procedures were provided in the initial email. Once directors agreed to have their clubs
participate, coaches were invited to have their team participate. With the invitation from the club
director, the researcher also attended club coach meetings and other club functions to invite
coaches from each club to participate. Again, a brief description of the study and procedures
were provided. Once coaches agreed to participate, parental informed consent forms were sent to

35

the athletes’ parents. Because the teammate measures required team members’ names, coaches
were asked to provide a team roster.
Data collection involved three time points where Time 1 data were collected around the
beginning of the competitive season, Time 2 data were collected around the mid-point of the
season, and Time 3 data were collected around the end of the season. Due to the briefness of
some teams’ seasons, it was ensured that at least one tournament was played between each time
point. In addition, no matter how long a team’s season was, the interval of time between each
data collection point remained consistent within each team. Three time points were included to
account for the fact that at the beginning of the season, teammates may not be familiar with each
other’s abilities thus requiring each athlete to make assumptions with regard to each others’ and
the team’s ability. Furthermore, trust beliefs of teammate may be conservative or inflated at the
beginning of the season; therefore, it is important to have measuring points at different times of
the season (McKnight, Cummings, & Chervany, 1998). Collecting data at three time points also
allowed for the examination of the potential development of teammate efficacy, and teammate
trust within each team. Because these constructs are dynamic yet often becoming more stable
towards the end of the season once the athletes become familiar with each other (Jung & Sosik,
2003), it is most beneficial to examine the data at multiple time points. Questionnaires were
administered at practice and not immediately before or after a competition so that competitionspecific responses were avoided.
Once parental consent was obtained for all athletes, Time 1 data collection was scheduled
with the coach at a regular practice time. It should be noted that only two athletes (from the same
team) did not have parental consent forms and therefore did not participate in the study. At Time
1, athletes completed informed assent forms and were administered demographics, self-,

36

collective, and teammate efficacy measures as well as teammate trust, backing-up behavior, trust
in the coach questionnaires, and a causal attribution scale for teams. Time 2 and Time 3 included
all measures except the demographics. If a team member did not attend the practice when the
questionnaires were being administered, a questionnaire packet, instructions, and a stamped and
addressed envelope were left with the coach. The coach was instructed to provide the
information to the team member, have them complete the measure on their own, insert and seal it
in the envelope and return it to the researcher by postal mail. Participants were guaranteed
confidentiality of their responses and were instructed to complete the questionnaires individually
without conversing with teammates. Coaches also signed consent forms and completed coach
rankings at each time point.
Team performance scores were gathered through online resources or by email from the
coach. Objective performance scores were gathered at up to 4 time points. If teams had already
played prior to Time 1 data collection, this performance was considered at Time 0. Each
subsequent performance scores were numbered the same as the data collection time occurring
prior to the performance. For example, a team who played matches before the first data
collection time and also after the last data collection time, had a data collection pattern as
follows: Time 0 performance – Time 1 data – Time 1 performance – Time 2 data – Time 2
performance – Time 3 data – Time 3 performance. Once the data were collected, data files for
each team were constructed, including team matrices for teammate efficacy, teammate trust, and
backing-up behavior.
Data Analyses
Prior to any analysis, team matrices were constructed for the teammate efficacy,
teammate trust, and backing-up variables. Athletes were removed from the matrices if they

37

showed response bias where they simply circled the same number for each question for each
teammate. Matrix means were then calculated for each and entered into the team level data file.
Means for each variable were computed for each athlete to represent each athlete’s evaluation of
each teammate.
Descriptive statistics (i.e., mean, standard deviations) were calculated for all variables.
Demographic variable descriptive statistics were calculated at the individual level, while all other
variable descriptive statistics were calculated at both the individual and team level. Bivariate
correlations were calculated between teammate efficacy and teammate trust at both the
individual and team level at each of the three time points. Bivariate correlations were also
calculated between self-efficacy, collective efficacy, teammate efficacy, and teammate trust at
both the team level (at all three time points) and between team level (testing Hypothesis 1).
Bivariate correlations were then calculated between all five performance measures (i.e.,
subjective performance, match winning percentage, game winning percentage, mean points
scored against, and mean point differential) at the team level for each time point. Lastly,
bivariate correlations were calculated between mean coach rankings for each of the five
defensive skills and teammate efficacy mean scores for each athlete, at each of the three time
points. One-way ANOVAs by team age group and by season (i.e., Winter season, Spring season,
combined season) were then calculated for each variable at each time point. Internal consistency
reliabilities were computed for scales that had multiple items and confirmatory factor analyses
were computed for the Trust in Coach scale as well as for the Causal Dimension Scale for Teams
(CDS-T) (Greenless et al., 2005).
Due to the nested and longitudinal nature of the data, multilevel modeling was used
where Level 1 represented differences within teams over time and Level 2 represented

38

differences between teams. Multilevel statistical techniques such as hierarchical linear modeling
(HLM) account for the interdependence of the data and allow researchers to simultaneously
examine relationships at each level and across levels, while determining the amount of variation
at each level (Raudenbush & Bryk, 2002). The first step of model building involved imposing an
unconditional model where no individual, subgroup, or team level predictors were entered into
the model. The purpose of fitting the unconditional model was to determine the amount of
variation in the dependent variable to ensure HLM is an appropriate form of analysis. HLM was
used to test Hypothesis 2 where separate analyses were run for each performance measure.
Teammate efficacy, collective efficacy, trust in coach, and past performance (when available)
were entered as predictors at both Level 1 (within teams) and Level 2 (between teams). Because
of the preliminary results, Hypothesis 3 could not be tested (see Results chapter). HLM was also
used to test Hypothesis 4 where backing-up behavior was entered as the dependent variable and
teammate trust and collective efficacy were entered as predictors at both Level 1 and Level 2.
Bivariate correlations were calculated to answer Research Question 1 where self-efficacy,
collective efficacy, teammate efficacy, and teammate trust were examined at each of the three
time points. Lastly, HLM was also used to answer Research Question 2 where each of the four
dimension of causal attributions in teams were entered as the dependent variable in separate
analyses and time, teammate trust, trust in the coach, and subjective performance as predictors at
both Level 1 and Level 2.

39

Chapter 4
Results
Descriptives
Descriptive statistics for the athletes are presented in Table 1. Means and standard
deviations for team level variables are presented in Table 2. In terms of performance, teams rated
themselves as having performed “so-so” to “well” over the course of the season with scores
1

ranging from 1.50 to 3.57 . The objective performance scores ranged from 0% to 100% for both
the match and game winning percentage; 13.00 points to 26.31 points for mean points scored
against; and -23.00 points to 10.83 points for mean point differential. Teams tended to have
moderately high aggregated self-efficacy beliefs in their defensive volleyball skills with scores
ranging from 6.08 to 9.04 as well as moderate to high aggregated collective efficacy beliefs in
the team’s defensive volleyball skills with scores ranging from 5.73 to 9.20. Teammate efficacy
scores and teammate trust scores showed similar patterns with scores ranging from 5.51 to 9.13
and 5.27 to 9.42 respectively. Teams showed high levels of trust in their coach with scores
ranging from 5.91 to 9.89 and teams did not seem to feel as though they needed to back up their
teammates often with scores ranging from .83 to 4.61. With regard to team attribution
dimensions, teams tended to attribute their performances to more internal, more team
controllable, less stable, and less externally controllable factors where scores ranged from 5.28 to
8.36 for Locus of Causality, 6.11 to 8.56 for Team Controllability, 2.25 to 6.25 for Stability, and
2.19 to 5.44 for External Controllability.
Tables 3 and 4 show the bivariate correlations between aggregated self-efficacy,
aggregated collective efficacy, teammate efficacy, and teammate trust at the team (within) level
1

All score ranges include scores across all time points within the mentioned variable.
40

and between level, respectively. Pearson Product Moment correlations showed at the between
team level significant relationships between teammate efficacy and teammate trust (r = .955),
teammate efficacy and collective efficacy (r = .72), collective efficacy and self-efficacy (r = .56),
and collective efficacy and teammate trust (r = .71). At the within team level (i.e., across time),
Pearson Product Moment correlations revealed interesting results. The relationship between
teammate efficacy and collective efficacy became stronger as the season progressed (Time 1: r =
.40; Time 2: r = .63; Time 3: r = .90). Not surprisingly since the relationship between teammate
trust and teammate efficacy was so strong, the same pattern emerged between teammate trust and
collective efficacy (Time 1: r = .39; Time 2: r = .69; Time 3: r = .89). Lastly, the same pattern
also emerged in the relationship between self-efficacy and collective efficacy (Time 1: r = .29;
Time 2: r = .66; Time 3: r = .76).
Table 5, Table 6, and Table 7 show the correlations at each Time 1, Time 2, and Time 3,
respectively, between collective efficacy, teammate efficacy, teammate trust and performance
indices before and after each time point. Of note, there are no significant relationships between
collective efficacy and performance, nor are there any significant relationships between
teammate efficacy (or teammate trust) and performance. Table 8 shows the same correlations at
the between team level. Results show the same non-significance.
Hypothesis Testing
The correlational results provided partial support for Hypothesis 1 that collective efficacy
and teammate efficacy would be positively related to each other but remain two distinct variables
(show no multicollinearity). At the between team level, teammate efficacy and collective
efficacy had a moderately strong positive relationship. However, at the within team level at Time

41

1 the relationship was non-significant, at Time 2 the relationship was moderately positive, and at
Time 3 the relationship was so strong and positive that it would suggest multicollinearity.
Multiple linear regression and HLM were used to test Hypothesis 2: Teammate efficacy
is a significant predictor of team performance while controlling for collective efficacy, trust in
the coach, and past performance. Because subjective performance was assessed at the same time
points as the predictor variables and asked participants to evaluate the team’s performance so far
during their competitive season, teams only evaluated their performance at three time points. In
addition, because subjective performance is an outcome variable, only performance evaluations
at Time 2 data collection (Time 1 performance) and at Time 3 data collection (Time 2
performance) could be inserted into a regression model. Consequently, multiple linear regression
was used to test Hypothesis 2 for subjective performance as HLM is not appropriate for
longitudinal data with less than three time points. In the first regression model, subjective
performance Time 1 (the team’s performance between the first and second data collection
sessions) was entered as the outcome variable. Subjective performance Time 0 (the team’s
performance prior to the first data collection session), aggregated collective efficacy at Time 1,
teammate efficacy at Time 1, and trust in the coach at Time 1 were entered as predictors. The
model was not significant. For the second regression model, subjective performance Time 2 (the
team’s performance between the second and third data collection sessions) was entered as the
outcome variable. Subjective performance Time 1, aggregated collective efficacy at Time 2,
teammate efficacy at Time 2, and trust in the coach at Time 2 were entered as predictors. Once
again, the model was not significant.
Similar to the multilevel analysis procedures described in the data analyses section of
Chapter 3, an unconditional model was run with the remaining four performance measures. The

42

unconditional models (i.e. variance components) for both the match winning percentage and
game winning percentage were non-significant indicating that HLM was not an appropriate form
of analysis, therefore both forms of performance measures were removed from any further
analysis.
The unconditional model for the mean points scored against indicated that 31.6% of the
2

variance was due to between-team differences and was significant (! = 38.20, p < .01). Mean
points scored against was entered as the outcome variable and collective efficacy, teammate
efficacy, trust in the coach, and past performance (as a control variable) were entered and grand
mean centered at both Level 1 and Level 2. Time was also entered in this model as a Level 1
predictor. Individual level data (i.e., age) were not entered in the models because all other
variables were analyzed at the team level (within teams over time). Table 9 presents the results
of this model where the only significant predictor of means points scored against was past
performance (! = .71, p < .01). Table 10 presents the percentage of variance accounted for at the
within and between levels. For mean points scored against, the percentage of variance explained
was 9% at the within level and 1.4% at the between level. Adjusting for the within-team (i.e., 1 –
ICC) and between-team (i.e., ICC) variance, the percentage of total variance accounted for was
6.6% in mean points scored against (within level = 6.2%, between level = -.4%). It should be
noted that because the variance of both within- and between-level predictors was smaller than
just the within-level predictors alone at the within level, Snijders & Bosker’s (1999) adjusted
formula was used to calculate the percentage of variance accounted for using a mean number of
team members (9 athletes).
The unconditional model for the mean point differential indicated that 24.5% of the
2

variance was due to between-team differences and was significant (! = 32.65, p < .05). Mean

43

point differential was entered as the outcome variable and collective efficacy, teammate efficacy,
trust in the coach, and past performance were entered and grand mean centered at both Level 1
and Level 2. Time was once again entered as a Level 1 predictor. Table 11 presents the results of
this model where no significant predictors were found. Thus, Hypothesis 2 was not supported.
Table 12 presents the percentage of variance accounted for at the within and between levels. For
mean points scored against, the percentage of variance explained was 3.5% at the within level
and 1.2% at the between level. Adjusting for the within-team (i.e., 1 – ICC) and between-team
(i.e., ICC) variance, the percentage of total variance accounted for was 2.9% in mean point
differential (within level = 2.6%, between level = .3%). It should again be noted that because the
variance of both within- and between-level predictors was smaller than just the within-level
predictors alone at the within level, Snijders & Bosker’s (1999) adjusted formula was again used
to calculate the percentage of variance accounted for using a mean number of 9 team members.
Hypothesis 3 stated that teammate trust would moderate the relationship between
teammate efficacy and team performance. Due to the very strong positive relationship between
teammate trust and teammate efficacy (r = .95, p< .0001), this hypothesis could not be
statistically tested as the correlation coefficient indicates multicollinearity.
HLM was again used to test Hypothesis 4, which stated that teammate trust would
negatively predict feelings of having to back-up or cover for teammates while accounting for
collective efficacy. The unconditional model for backing-up behaviors indicated that 69.9% of
2

the variance was due to between-team differences and was significant (! = 134.17, p < .001).
Backing-up behavior was entered as the outcome variable and teammate trust and collective
efficacy were entered as predictors at both Level 1 and Level 2, and Time was entered only at
Level 1 as a predictor. Table 13 presents the results of this model where teammate trust was a

44

significant predictor at the within team level (! = -.93, p < .01). No other predictors were found
to be significant therefore providing support for Hypothesis 4. Table 14 presents the percentage
of variance accounted for at the within and between levels. For backing-up behaviors, the
percentage of variance explained was 50% at the within level and 3.8% at the between level.
Adjusting for the within-team (i.e., 1 – ICC) and between-team (i.e., ICC) variance, the
percentage of total variance accounted for was 17.7% in backing-up behaviors (within level =
15%, between level = 2.7%).
With regards to the research questions posed in this study, some interesting results were
found. The first question asked if there were differences in the relationships between efficacy
(self-, collective, and teammate) and trust over time. Table 3 presents the bivariate correlations
of each of these variables at all three time points. As mentioned earlier in this chapter, a number
of these relationships strengthened over the competitive season. Of note, and as presented
previously, the relationship between teammate efficacy and collective efficacy became stronger
as the season progressed (Time 1: r = .40; Time 2: r = .63; Time 3: r = .90) as did the
relationship between teammate trust and collective efficacy (Time 1: r = .39; Time 2: r = .69;
Time 3: r = .89).
Lastly, to answer Research Question 2 (Do teammate trust and/or trust in the coach have
an effect on team attributions while accounting for subjective performance?), HLM was used by
running a 2-level model for each of the four team causal dimensions. The unconditional models
2

for each of the dimensions were significant: Locus of Causality (! = 27.52, p = .05), Team
2

2

2

Controllability (! = 27.35, p = .05), Stability (! = 64.12, p < .001), External Controllability (!
= 48.50, p < .001). The unconditional models also indicated that 18.1% of the variance was due
to between-team differences for Locus of Causality, 17.7% for Team Controllability, 49.8% for
45

Stability, and 39.2% for External Controllability. Each team causal dimension was entered in its
own model as the outcome variable. Teammate trust, trust in the coach, and subjective
performance were added as predictors at both Level 1 and Level 2 for each of the four models.
Time was also added as a Level 1 predictor for each of the four models. The team controllability
and external controllability models were not significant. The model for Locus of Causality was
also non-significant however, trust in the coach approached significance at Level 1 (! = .35, p =
.087). The model for Stability was significant. Table 15 presents the results of this model where
trust in the coach was a significant predictor at Level 2 (! = -.86, p < .05), indicating that higher
levels of trust in the coach predicted less stable team attributions. No other predictors were found
to be significant. Table 16 presents the percentage of variance accounted for at the within and
between levels. For stability, the percentage of variance explained was 29.5% at the within level
and 20.5% at the between level. Adjusting for the within-team (i.e., 1 – ICC) and between-team
(i.e., ICC) variance, the percentage of total variance accounted for was 25% in stability (within
level = 14.8%, between level = 10.2%).
Summary of Hypotheses and Results
Hypothesis 1 addressed the relationship between collective and teammate efficacy.
Partial support was provided for this hypothesis as the between level correlation was moderately
strong and significant, however over time, the strength of the relationship increased from nonsignificant to multicollinear. Hypothesis 2 examined the predictive strength of teammate efficacy
on team performance while controlling for collective efficacy, trust in the coach and past
performance. Hypothesis 2 was not supported. Hypothesis 3 posited that teammate trust would
moderate the relationship between teammate efficacy and team performance. This hypothesis
could not be tested, and is therefore not supported, because of the multicollinearity between

46

teammate efficacy and teammate trust, as well as the non-significant prediction of teammate
efficacy of team performance. Hypothesis 4 posited that teammate trust would negatively predict
feelings of having to back-up teammates while controlling for collective efficacy. Support for
Hypothesis 4 was found. Research Question 1 was examined and it does seem that there are
differences in the relationships between efficacy and trust over the course of a competitive
season. Lastly, teammate trust and trust in the coach did not seem to have much effect on team
attributions, except for the causal dimension of stability.

47

Chapter 5
Discussion
Even though confidence and trust in one’s teammates is fairly common in anecdotal
counts of sport teams, it has rarely been studied in the sport context. To address this gap in the
research, this dissertation examined the constructs of teammate efficacy, its relationship to selfefficacy, collective efficacy and team performance. Furthermore this study aimed to examine
teammate trust as a potential moderator between teammate efficacy and team performance.
Findings from this study contribute to a better understanding of these constructs in the sporting
arena and offer a starting point for future research in this area. This chapter discusses the
findings of this dissertation, discusses strengths and limitations of the study, identifies
implications of these results, and presents future research directions.
Results showed that the relationship between collective efficacy and teammate efficacy
became stronger over the course of a competitive season. Teammate efficacy and collective
efficacy were not found to be significant predictors of team performance. Teammate efficacy and
teammate trust were highly correlated (i.e., showed multicollinearity) and therefore a moderating
relationship between these variables to predict team performance was not tested. Results also
showed that decreased teammate trust significantly predicted increased feelings of having to
back-up or cover for teammates while accounting for collective efficacy. Lastly, no variables of
trust were found to significantly predict team attributions.
Teammate trust, along with teammate efficacy were the focal points of this study and
following Dirks and Ferrin’s (2001) propositions, it was hypothesized that there would be a
moderating relationship between the two. Consequently, it would be assumed that these variables
would be distinct from one another. However, the results did not demonstrate this. Teammate

48

efficacy and teammate trust were so highly correlated that there was no statistical distinction
between the two. This was possibly due to the age group of the participants. Because of the mean
age of the participants being around 15 years, conceptually distinguishing between teammate
efficacy and teammate trust may be difficult. Efforts were made through written and oral
instructions and definitions to explain the difference between the two variables however having
the participants respond to similar questions for each variable may have confused them. It should
be noted, however, that even though the two were very highly correlated, and the magnitudes of
both were similar, teammate trust was slightly less than teammate efficacy at all three time points
(see Table 1). Past research on efficacy beliefs in sport has mostly yielded high levels of efficacy
beliefs with athletes rarely using the lower half of the response scale (Feltz & Chase, 1998).
Myers and Feltz (2007) have argued that the psychometric functioning of collective efficacy
scales could be improved if the rating categories were reduced because of the sparse responses at
the lower end of the scales. Perhaps teammate trust may be a more salient construct with more
wide spread response variability than teammate efficacy and should be considered in future
research. On the other hand, teammate efficacy and teammate trust may in fact be conceptually
very similar. Because the trust measure focused solely on trusting belief-competence and the
teammate efficacy and teammate measures were so similar to each other, it is likely that the
participants had difficulty discriminating between the two. Consequently, as measured in this
study, they measured the same construct. Perhaps if the teammate trust also included the
benevolence and integrity features of trust (Mayer, Davis, & Schoorman, 1995), was measured
on a different scale, and also was not measured directly following teammate efficacy, a greater
level of discriminant validity would occur. Future research should also continue to examine the
psychometric properties of each, and examine the conceptual similarities and differences.

49

Dirks and Ferrin (2001) also stated that evidence supporting a direct relationship between
trust and performance was weak and inconsistent. The results from this study seem to support
this claim. No significant correlations were found between teammate trust and either of the five
performance indices both at the within and between level.
It was somewhat surprising that teammate efficacy was not a significant predictor of any
of the team performance measures. Even more surprising that collective efficacy was not a
significant predictor of performance either because past research has shown a positive and robust
relationship between the two (e.g., Myers, Feltz, & Short, 2004; Myers, Payment, & Feltz, 2004).
One could assume that this was due to the small sample size, however studies with similar
sample size have shown significant positive relationships (Feltz & Lirgg, 1998). It could also be
due to the somewhat stable yet subtle increase in both teammate and collective efficacy across
the duration of the season and the fairly unstable performance indices. Teams could be playing
more difficult opponents towards the end of the season resulting in a decrease in performance
according to the performance indices. However, no significant relationships or predictions were
found with subjective performance evaluations, which also showed a slight increase over the
course of the season.
Most studies examining the relationship between efficacy beliefs and team performance
have been completed using collegiate or older samples. Perhaps adolescents competing at the
club level cannot perceive their team’s or teammate’s abilities as accurately as adults can.
Children only begin to distinguish between their own effort and ability in their late childhood
years (Nicolls, 1990), therefore the mean age of about 15 years could be indicative of a lack of
accurate perceptions. In addition, at the club level, coaches are encouraged to focus on

50

development and effort, therefore efficacy beliefs may again be slightly inflated as the results
may suggest (see Table 2).
Another interesting finding was the change in strength of relationship between collective
and teammate efficacy (see Table 3). At the beginning of the season, there was no significant
relationship between the two variables, however, as the season progressed, the relationship
became quite strong and significant. This may suggest a form of de-individuation on the part of
the athletes as the season moves along. For example, when the team comes together at try-outs or
within the first week or so of practice, there is more of a sense of ‘me and them’ (i.e. teammates),
therefore collective efficacy and teammate efficacy are two separate entities. However as the
season progresses, the sense of “us” grows where the athletes feel more and more as part of the
team. It was mentioned in Chapter 1 that Feltz and colleagues (2008) argue that Lent and Lopez
(2002) measure other efficacy in the same way as collective efficacy except that it is in dyads
rather than teams of three or more members. In addition, because Lent and Lopez’s main focus
was on dyadic relationships, the differences between the two constructs can be clearer than in
larger groups such as sport teams. This proposition is partially supported as the relationship
between teammate efficacy and collective efficacy is very high at the end of the competitive
season. A second, and realistic reason for these results is that after completing the measures once
or twice already, athletes did not take as much care in their responses compared to prior data
collection times. Athletes could have been rushing through the items and simply answered each
with more consistency than in prior data collection times.
As mentioned in Chapter 1, Dirks and Ferrin’s (2001) second proposition was that trust
will moderate the relationship between a teammate’s action and the truster’s response. Therefore
trust can have an effect on how athletes interprets their teammate’s actions. If an athlete has a

51

decreased feeling of trust in a teammate, perceptions of this teammate’s behaviors should also be
low. Consequently, these decreased perceptions could lead to changes in behavior in the truster
such as having to back-up or cover for the teammate in question. This proposition led to
Hypothesis 4, which was supported. While accounting for collective efficacy, teammate trust in a
particular teammate significantly predicted feelings of having to back-up this said teammate;
athletes who had decreased levels of trust in their teammate, felt they had to back-up or cover for
this teammate.
Very little research has been conducted examining backing-up behaviors, however a
handful of studies in the organizational context should be noted. In their first study, Porter,
Hollenbeck, Ilgen, Ellis, West and Moon (2003) defined backing up behaviors as “the
discretionary provision of resources and task-related effort to another member of one’s team that
is intended to help that team member obtain the goals as defined by his or her role when it is
apparent that the team member is failing to reach those goals” (p. 391-392). This initial study and
a follow-up by Porter (2005) found that when there was a mismatch in workloads and team
capacities, backing up behaviors were likely to have positive effects on team performance. On
the other hand, a more recent study by Barnes, Hollenbeck, Wagner, DeRue, Nahrgang, and
Schwind (2008) contended and found that backing-up behavior leads to both initial and
subsequent costs to taskwork (i.e., a team’s interaction with equipment and the task at hand).
More specifically Barnes and colleagues (2008) found that the relationship between backing up
behavior and the backup provider’s neglected work was moderated by workload distribution
where the relationship was stronger when the workload was evenly distributed between team
members.

52

These findings could be extended to the sporting context where backing up a teammate
may result in both positive and negative consequences in team performance depending on the
workload distribution. For example, if a volleyball player is being targeted at every opponent’s
serve, the workload is not evenly distributed therefore backing up behaviors from a teammate
who is not being targeted may have positive effects on team performance. However, if the
workload is evenly distributed and coordination is not at its optimal level, backing up behaviors
may indeed have negative consequences on team performance.
Furthermore, Barnes and colleagues (2008) also suggest that backing up behaviors can
lead to more long term negative effects such as decreased motivation, dependence, and social
loafing from the teammate who is receiving the backing up. They state that individuals who
receive high levels of backup from teammates will interpret their own abilities as inadequate and
may lead to decrease motivation in future tasks. In addition, when individuals are consistently
being backed up they will no longer feel personally accountable for their incomplete tasks. Once
more, these findings could be extended to the sporting context and future research should
examine these relationships using different sport teams.
Strengths and Limitations
This study examined efficacy and trust beliefs in an adolescent sample across an entire
competitive season. The advantages or strengths of a longitudinal design include the ability to
examine relationships over time, such as an entire competitive season. Because group dynamics
such as efficacy beliefs and trust are in fact dynamic but becoming more stable towards the end
of the season (Jung & Sosik, 2003), using a longitudinal design allows researchers to identify
any changes in variables over the course of a season. Additionally, as mentioned in a previous
chapter, trust beliefs of teammate may be conservative or inflated at the beginning of the season

53

therefore it is important to have measuring points at different times of the season (McKnight, et
al., 1998).
Another strength associated with this study pertains to the use of multilevel modeling
techniques (e.g. HLM). Due to the nested and longitudinal nature of the data, HLM was used to
simultaneously analyze the data and account for the interdependence of the data at each level and
across levels, while determining the amount of variation at each level (Raudenbush & Bryk,
2002).
Additionally, this study measured multiple performance indices, using both objective and
subjective measures. Often, studies will simply use a team’s or athlete’s winning percentage as
an indicator of performance, or a subjective measure which is subject to response bias.
Moreover, in this study winning percentages were not deemed appropriate to use with multilevel
modeling. Because this study examined specifically defensive skills in volleyball, more
defensive performance indices were also used. It is important to have congruence between
efficacy and performance measures in order to ensure a study’s validity (Feltz et al., 2008).
Future research should consider implying multiple performance measures, both objective and
subjective and not simply a team or athlete’s winning percentage.
Lastly, no other study in the sport context has examined the construct of trust. Even
though anecdotally trust is often mentioned as an important factor leading to peak team
performance, it has rarely been examined scientifically. In addition, very little research has been
conducted examining the construct of teammate efficacy (i.e. other-efficacy, Lent & Lopez,
2002) outside of the counseling psychology domain. This investigation offers a foundation on
which future studies can be based upon to further understand these variables in the sport context.

54

Although there are several strengths associated with this study, there are also some
limitations that should be mentioned. The first limitation in this study is the sample size.
Although the number of teams that participated in this 3-time point longitudinal study (N = 18) is
somewhat typical for sport psychology research, it was not favorable for statistical purposes.
Even though multiple volleyball clubs were contacted and invited to participate (approximately
30 clubs ranging from 1 to 30 teams each), only a small portion of these agreed to participate
(i.e., director, coach, parents, and athlete agreement). When a team has a relatively short
competitive season (e.g., 3 months), giving up potential practice time to complete surveys was
not favorable to many directors and coaches. Large sample sizes are difficult to acquire in group
field research because one team may include several members; therefore a sample size of 50
teams may actually include over 500 participants. The small sample size reduces the statistical
power in a study and therefore more hypotheses may have been significant had there been a
larger sample size. There are no known well-developed techniques or statistical software to
determine statistical power or effect sizes for hierarchical linear modeling or multilevel
modeling. Consequently, any calculations of these would only be speculation and unreliable.
Future studies with larger sample sizes may be better able to test the hypotheses proposed in this
dissertation and have the statistical power required to find significant differences.
A second limitation in this study could be the variability in team goals. Even though all
teams competed at the club level, it was apparent to the researcher that some teams had set
higher goals than others (i.e., attending national championships versus playing well and having
fun). Even within teams, coaches, athletes, and teammates could have had different goals for
themselves and the team. No measure was taken in this study to account for this variability and

55

future research should include team goals and/or expectations as a control variable to account for
potential confounders.
Another limitation is the proximity of each of the time points for a number of teams.
Because some teams participated in relatively short seasons, the amount of time between each
data collection session was only 2 weeks while other teams participating in longer seasons had
over a month between time points (time intervals were consistent within each team). There were
no significant differences in the variables between teams competing in different seasons.
However, two weeks could be too short of a time period to show much change or development in
efficacy beliefs and trust. Future studies could examine teams that compete in longer seasons and
potentially measure each variable at more than three time points during the competitive season.
Implications
This study has shown that levels of trust in teammates significantly predict feelings of
having to back-up or cover for teammates. That is, if an athlete has lower levels of trust in a
teammate, she feels that she must cover or back-up this teammate. For example, if a volleyball
player does not trust a teammate to effectively play a serve from her opponent, she may move
out of her own position to cover more space from her teammate. As mentioned previously, the
consequence of this change in position could be positive or negative depending on the workload
of the backup provider and the communication and coordination skills of the team (Barnes et al.,
2008). Coaches should address these issues with their team and make it clear with their athletes
when providing backup for teammates is most beneficial. Because there may be long-term
consequences with providing backup (e.g. dependency and lack of accountability), coaches and
athletes should continuously remind their team what every person’s role is and the importance of
each teammate fulfilling the role to the best of her ability.

56

Future Directions
Because this study is the first to examine teammate efficacy, teammate trust and backing
up behaviors in sport, there are several directions that could be followed and investigated in
future studies. First, because the sample used in the current study was quite specific, future
research should examine similar research questions in different samples. As previously
mentioned, the mean age of the current participants was just above 15 years of age, which could
have contributed to an inability to differentiate between efficacy and trust. Future research
should examine any differences between these constructs in more mature samples.
Qualitative methods such as interviews and focus groups with athletes and coaches may
also give more insight on perceptions of each of these constructs. Interviews conducted with
athletes and coaches could serve as a better means to explore the differences between efficacy
and trust or if these differences even exist. These methods can provide an answer to the “why”
questions in this area of research as well as dig deeper into the meaning of trust, its development
and consequences in sport teams. In addition, due to the difficulties in measurement, qualitative
methods may also be helpful in examining the effect of backing up behaviors on team
performance.
Past research mentioned in this dissertation (Dirks, 2000) found that trust in the leader or
coach plays a significant part in team performance. Examining this aspect of trust in sport team
dynamics is also an area of future research. The data from this study showed that trust in the
coach decreased over the course of the season. Future studies could examine why this may
happen and if this phenomena is true in different age groups where coaches are not just volunteer
parents, but have multi-million dollar contracts.

57

Another line of research stemming from this initial study could be the examination of
trust and performance in terms of different team roles. For example, is it more important (in
terms of performance) to trust your goaltender than a fourth line forward in ice hockey? Some
team roles have a greater effect on the team’s outcome therefore trusting the teammates in these
specific roles may have a greater effect on collective efficacy, team cohesion, and team
performance.
Until now, backing up behaviors have been studied in an experimental setting. Although
complex, future research could begin to investigate how one could measure backing up behaviors
in a field setting. Observational techniques and evaluation of the coach and teammates could
provide insight into how, and when it is mostly likely that an athlete provides backup for his or
her teammate. Once appropriate assessment and measurement techniques are validated, more
complex research designs and hypotheses can be examined and tested.
Lastly, additional group dynamic variables can be compared with teammate trust in
future research thus leading to the development of a model of trust in team sport. Antecedents
such as team interdependence, team goal orientation (e.g. Porter, 2005), communication,
cohesion, coordination, playing experience, and past performance could be tested as sources of
trust in teams. Consequences or outputs of teammate trust could include performance,
enjoyment, cohesion, commitment, and coordination. There is plenty of anecdotal evidence from
athletes that demonstrates the perceived importance of trust in team sport, however it has not
received empirical attention. The opportunities for future research in this area are subsequently
abundant.
Conclusion

58

This study examined the constructs of teammate efficacy, its relationship to self-efficacy,
collective efficacy and team performance. Measures of efficacy belief, trust, and performance
were gathered from girls club volleyball teams at three time points across a competitive season.
Results showed that teammate efficacy and teammate trust were too statistically similar to
differentiate. Results also showed that teammate efficacy did not predict team performance.
However, decreased levels of teammate trust led to increased feelings of having to provide
backup for teammates. Findings from this study contribute to a better understanding of these
constructs in the sporting arena and offer a starting point for future research in this area.

59

TABLES

60

Table 1
Athlete Demographics
Variable

Mean

SD

Age

15.29

1.42

Height

66.34 inches

2.94

Starting status*

2.97

1.00

Age when began playing competitively

11.74

1.62

Number of seasons with team

.93

.72

*Athletes were asked to choose one option: Never = 0, Not a whole lot = 1, About half the time
= 2, A lot of the time = 3, Always = 4

61

Table 2
Means and Standard Variations for Team Variables (N = 18)
Variable

Time 0
M

Subjective performance

Time 1
SD

M

Time 2
SD

M

Time 3
SD

2.50

.50

2.72

.51

2.84

.56

Match winning %

37.48

39.32

44.91

28.22

41.99

Game winning %

32.73

32.16

45.74

23.98

Mean pts scored against

21.62

2.82

20.45

2.78

Mean point differential

-.06

5.41

-.31

Aggregated self-efficacy

-

-

7.22

.51

Aggregated collective

-

-

6.85

Teammate efficacy

-

-

Teammate trust

-

Trust in coach

M

SD

-

-

30.91

48.71

27.46

37.27

27.23

50.86

23.28

21.86

2.59

21.82

5.01

4.70 -3.03

7.08

.47

5.16

7.56

.61

8.02

.63

.54

7.41

.57

7.82

.67

7.02

.52

7.18

.52

7.49

.71

-

6.72

.50

7.03

.59

7.34

.81

-

-

8.97

.46

8.54

1.05

8.55

1.20

Backing-up

-

-

2.99

.78

2.96

.81

2.60

.94

CDS – T: Locus of

-

-

6.96

.58

6.85

.73

7.21

.58

-

-

7.53

.49

7.18

.67

7.50

.57

CDS – Team: Stability

-

-

3.79

1.05

4.01

.86

4.32

.82

CDS – Team: External

-

-

3.86

.59

4.10

.68

3.95

.71

efficacy

Causality
CDS – T: Team
Controllable

Controllable

62

Table 3
Team Level Bivariate Correlations Between Aggregated Self-Efficacy (SE), Aggregated Collective Efficacy (CE), Teammate Efficacy
(TE) and Teammate Trust (TT).
Variables

1

2

3

4

5

6

7

8

9

10

11

1. Time 1 SE
2. Time 2 SE

.78**

3. Time 3 SE

.56*

.68*

4. Time 1 CE

.29

.27

.44

5. Time 2 CE

.38

.66**

.59*

.58*

6. Time 3 CE

.11

.32

.76**

.49

.68**

7. Time 1 TE

.35

.68**

.75**

.40

.64**

.63**

8. Time 2 TE

.13

.51*

.69**

.37

.63**

.74**

.83**

9. Time 3 TE

-.02

.28

.70**

.47

.55*

.90**

.72**

.84**

10. Time 1 TT

.32

.57*

.76**

.39

.62**

.64**

.93**

.77**

.67**

11. Time 2 TT

-.19

.39

.63**

.27

.69**

.80**

.73**

.93**

.84**

.75**

12. Time 3 TT

-.85

.22

.69**

.39

.52*

.89**

.64**

.78**

.97**

.65**

Note. ** p< .01; * p< .05

63

.84**

Table 4
Between Level Bivariate Correlations Between Aggregated Self-Efficacy (SE), Aggregated
Collective Efficacy (CE), Teammate Efficacy (TE) and Teammate Trust (TT).
Variables

1

2

3

1. Aggregated Self-Efficacy
2. Aggregated Collective Efficacy

.56*

3. Teammate Efficacy

.52*

.72**

4. Teammate Trust

.44

.71**

Note. ** p< .01; * p< .05

64

.95*

4

Table 5
Correlations at Time 1 Between Collective Efficacy (CE), Teammate Efficacy (TE), Teammate Trust (TT), and Pre- and PostPerformance Measures.
Variables

1

2

3

4

5

6

7

8

9

10

11

12

1. CE
2. TE

.40

3. TT

.39

.93!

4. Pre Subjective Performance

.43

.18

.24

5. Pre Match Winning %

.27

.21

.17

.43

6. Pre Game Winning %

.11

-.06

-.15

.20

.70!

7. Pre Mean Pts Scored

-.32

-.28

-.32

-.48

-.83!

-.51

8. Pre Mean Pt Differential

.32

.42

.46

.41

.82!

.49

-.97!

9. Post Subjective Performance

.00

.24

.22

.00

.52

.37

-.47

.67*

10. Post Match Winning %

-.04

.07

.00

.18

.76!

.48

-.62*

.59*

.40

11. Post Game Winning %

.04

.12

.00

.16

.78!

.56*

-.54

.54

.43

.96!

12. Post Mean Pts Scored

-.06

-.47

-.45

-.19

-.49

-.23

.58*

-.68*

-.61!

-.46

-.51*

.17

.31

.32

.38

.61*

.36

-.80!

.82!

.56*

.67!

.64!

Against

Against
13. Post Mean Pt Differential
Note. ! p< .01; * p< .05

65

-.80!

Table 6
Correlations at Time 2 Between Collective Efficacy (CE), Teammate Efficacy (TE), Teammate Trust (TT), and Pre- and PostPerformance Measures.
Variables

1

2

3

4

5

6

7

8

9

10

11

12

1. CE
2. TE

.63!

3. TT

.69!

.93!

4. Pre Subjective Performance

.43

.25

.45

5. Pre Match Winning %

.10

.04

.12

.40

6. Pre Game Winning %

.19

.16

.21

.43

.96!

7. Pre Mean Pts Scored

-.41

-.32

-.43

-.61!

-.46

-.51*

8. Pre Mean Pt Differential

.42

.08

.24

.56*

.67!

.64!

-.80!

9. Post Subjective Performance

.28

.26

.37

.77!

.07

.16

-.25

.14

10. Post Match Winning %

.30

.00

.04

.23

-.07

-.04

-.35

.31

.45

11. Post Game Winning %

.23

.17

.14

.06

.20

.34

-.09

.03

.42

.61!

12. Post Mean Pts Scored

-.50*

-.32

-.42

-.53*

-.24

-.23

.69*

-.65!

-.53*

-.59*

-.22

.44

.08

.25

.46

.27

.29

-.51*

.54*

.31

.31

.14

Against

Against
13. Post Mean Pt Differential
Note. ! p< .01; * p< .05

66

-.74!

Table 7
Correlations at Time 3 Between Collective Efficacy (CE), Teammate Efficacy (TE), Teammate Trust (TT), and Pre- and PostPerformance Measures.
Variables

1

2

3

4

5

6

7

8

9

10

11

1. CE
2. TE

.90!

3. TT

.89!

.97!

4. Pre Subjective Performance

.35

.42

.40

5. Pre Match Winning %

.19

.14

.12

.45

6. Pre Game Winning %

.18

.23

.18

.42

.61!

7. Pre Mean Pts Scored Against

-.45

-.35

-.35

-.53*

-.59*

-.22

8. Pre Mean Pt Differential

.35

.09

.11

.31

.31

.14

-.74!

9. Post Match Winning %

.26

.01

.09

-.15

.31

-.05

-.44

.39

10. Post Game Winning %

.10

-.09

.00

-.04

.43

.00

-.46

.27

.93!

11. Post Mean Pts Scored Against

.00

.06

-.07

-.28

-.39

-.15

.50

-.37

-.65!

-.69!

12. Post Mean Pt Differential

.11

.00

.06

.06

.60*

.11

-.48

.09

.80!

.93!

Note. ! p< .01; * p< .05

67

-.61*

Table 8
Correlations at the Between Level Between Collective Efficacy (CE), Teammate Efficacy (TE), Teammate Trust (TT), and Pre- and
Post- Performance Measures.
Variables

1

2

3

4

5

6

7

8

9

10

1. CE
2. TE

.72!

3. TT

.71!

.95!

4. Pre Match Winning %

.10

.01

.00

5. Pre Game Winning %

.08

.00

-.51

.82!

6. Pre Mean Pts Scored Against

-.24

-.28

-.31

-.76!

-.40

7. Pre Mean Pt Differential

.36

.13

.23

.73!

.49

-.77!

8. Post Match Winning %

.29

.07

.12

.80!

.62!

-.67*

.81!

9. Post Game Winning %

.26

.09

.09

.79!

.85!

-.47*

.62!

.86!

10. Post Mean Pts Scored Against

-.16

-.20

-.30

-.72!

-.40

.86!

-.85!

-.80!

-.59!

11. Post Mean Pt Differential

.23

.06

.14

.59*

.43

-.62!

.92!

.77!

.61!

Note. ! p< .01; * p< .05

68

-.78!

Table 9
Multilevel Model for Mean Points Scored Against
Fixed Effect

Coefficient

SE

t

Mean points scored against

21.19

2.62

8.09**

Collective efficacy

-4.50

2.84

-1.58

Teammate efficacy

4.25

3.62

1.17

.87

1.60

.54

Time

-.03

1.30

-.03

Collective efficacy

3.49

2.27

1.54

Teammate efficacy

-3.71

2.94

-1.26

Trust in coach

-.29

1.22

-.24

Past Performance

.71

.23

Trust in coach

(mean points scored against)
Note. ** p< .01; * p< .05

69

3.07**

Table 10
Explanatory Power at the Within Team and Between Team Levels for Mean Points Scored
Against
Variance Estimates
Mean Points

No

Within-Level

Scored

Within- and

Predictors Predictors Only

Percentage of

Between-

Variance

Total

Level

Against

Percentage of

Accounted for Variance

Predictors
Within

8.81

11.57

Between

4.07

11.39

.13

.34

Total

Accounted for
9%
1.4%

6.2%
.4%
6.6%

70

Table 11
Multilevel Model for Mean Point Differential
Fixed Effect

Coefficient

SE

t

Mean point differential

-3.92

4.01

-0.978

Collective efficacy

5.45

5.14

1.06

Teammate efficacy

2.76

6.23

.44

Trust in coach

-2.62

2.73

-.96

1.46

1.96

.75

Collective efficacy

.62

3.90

.16

Teammate efficacy

-5.41

5.23

-1.03

Trust in coach

.87

2.04

.43

Past Performance

.10

.17

.63

Time

(mean point differential)
Note. ** p< .01; * p< .05

71

Table 12
Explanatory Power at the Within Team and Between Team Levels for Mean Point Differential
Variance Estimates
Mean Point

No

Within-Level

Differential

Within- and

Predictors Predictors Only

Percentage of

Percentage of

Between-

Variance

Total

Level

Accounted for Variance

Predictors
Within
Between

Accounted for

25.21

32.58

32.19

3.5%

2.6%

8.20

.04

.06

1.2%

.3%

Total

2.9%

72

Table 13
Multilevel Model for Backing-up Behaviors
Fixed Effect

Coefficient

SE

t

Backing-up Behaviors

2.86

.25

11.41**

Collective efficacy

.48

.52

.92

Teammate trust

.14

.50

.30

Time

-.00

.09

-.02

Collective efficacy

.20

.17

1.14

Teammate trust

-.93

.20

-4.57**

Note. ** p< .01; * p< .05

73

Table 14
Explanatory Power at the Within Team and Between Team Levels for Backing-up Behaviors
Variance Estimates
Backing-up

No

Within-Level

Behaviors

Within- and

Predictors Predictors Only

Percentage of

Percentage of

Between-

Variance

Total

Level

Accounted for Variance

Predictors

Accounted for

Within

.22

.11

.11

50%

15%

Between

.50

.52

.50

3.8%

2.7%

Total

17.7%

74

Table 15
Multilevel Model for Causal Dimension Scale-Teams: Stability
Fixed Effect

Coefficient

SE

t

Stability

3.74

.35

10.55**

Teammate trust

-.59

.51

-1.16

Trust in coach

-.86

.31

-2.81**

Subjective performance

.95

.58

1.65

Time

.17

.16

1.04

Teammate trust

.47

.39

1.20

Trust in the coach

.32

.21

1.50

Subjective performance

-.10

.33

-.31

Note. ** p< .01; * p< .05

75

Table 16
Explanatory Power at the Within Team and Between Team Levels for Stability
Variance Estimates
Stability

No

Within-Level

Percentage of

Percentage of

Between-

Variance

Total Variance

Level

Predictors Predictors Only

Within- and

Accounted for

Accounted for

Predictors
Within

.44

.41

.31

29.5%

14.8%

Between

.44

.39

.31

20.5%

10.2%

Total

25%

76

APPENDICES

77

Appendix A
Demographics

78

AGE:
HEIGHT:
WHAT POSITION DO YOU USUALLY PLAY?:
HOW OFTEN DO YOU START A GAME (circle one):
Never

Not a whole lot

About half the time

A lot of the time

HOW OLD WERE YOU WHEN YOU STARTED PLAYING COMPETITIVE
VOLLEYBALL?:
FOR HOW MANY SEASONS HAVE YOU PLAYED WITH YOUR TEAM?:
DO YOU PLAY COMPETITIVE VOLLEYBALL YEAR-ROUND?(circle one):
Yes

No

79

Always

Appendix B
Self-efficacy and Collective Efficacy Measures

80

Rate your confidence in YOUR OWN ability to effectively perform the following defensive tasks: Circle ONE for each question
Not at All Confident

Extremely Confident

1. Pass a tough serve 0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

2. Pick up a tip

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

3. Defend against
0%
a hard driven attack

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

4. Dive/roll when
0%
necessary to play a ball

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

5. Back up my
0%
10%
20%
30%
teammates if they are out of position or miss a shot

40%

50%

60%

70%

80%

90%

100%

Rate your confidence in YOUR TEAM’S ability as a whole to effectively perform the following defensive tasks:
Circle ONE for each question
Not at All Confident

Extremely Confident

1. Pass a tough serve 0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

2. Pick up a tip

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

3. Defend against a 0%
hard driven attack

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

4. Dive/roll when
0%
necessary to play a ball

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

5. Back each
0%
10%
20%
30%
other up if they are out of position or miss a shot

40%

50%

60%

70%

80%

90%

100%

81

Appendix C
Teammate Efficacy, Teammate Trust, and Backing-up Measures

82

Rate your confidence in EACH of your TEAMMATES’ ability to perform these defensive tasks:
I believe that my teammate
NAME
CAN correctly…
1. Pass a tough serve
2. Pick up a tip
3. Defend against a hard
driven attack
4. Dive/roll when necessary
to play a ball
5. Back me up if I am out of
position or miss a shot

Not at All Confident
0%
10%
20%
0%
10%
20%
0%
10%
20%

30%
30%
30%

40%
40%
40%

50%
50%
50%

60%
60%
60%

70%
70%
70%

Extremely Confident
80%
90%
100%
80%
90%
100%
80%
90%
100%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Rate how much you TRUST that EACH OF YOUR TEAMMATES will CORRECTLY and EFFECTIVELY perform the
following defensive tasks during a very crucial game situation:
I TRUST that my teammate
NAME
WILL…
1. Pass a tough serve
2. Pick up a tip
3. Defend against a hard
driven attack
4. Dive/roll when necessary
to play a ball
5. Back me up if I am out of
position or miss a shot

Do Not Trust at All
0%
10%
20%
0%
10%
20%
0%
10%
20%

30%
30%
30%

40%
40%
40%

50%
50%
50%

60%
60%
60%

70%
70%
70%

80%
80%
80%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Rate how much you agree with the following statement:
Completely Disagree
I often feel like I have to
back-up or cover more of the
court for this teammate
0%
10%
20%

Completely Trust
90%
100%
90%
100%
90%
100%

Completely Agree
30%

83

40%

50%

60%

70%

80%

90%

100%

Appendix D
Trust in Coach Measure

84

Rate how much you TRUST YOUR HEAD COACH:
Circle ONE NUMBER for each question
1. My coach approaches his/her job with expertise and dedication.
Strongly Disagree
0
2.

1

2

Strongly Agree
3

4

5

6

7

8

9

10

Given my coach's past performance, I see no reason to doubt his/her competence.
Strongly Disagree
0

3.

1

2

Strongly Agree
3

4

5

6

7

8

9

10

I can rely on the coach not to make my job (as a player) more difficult by poor coaching.
Strongly Disagree
0

4.

1

2

Strongly Agree
3

4

5

6

7

8

9

10

I respect my coach’s ability to make the most of our team’s talents.
Strongly Disagree
0

1

2

Strongly Agree
3

4

5

6

7

8

9

85

10

Appendix E
Causal Dimension Scale – Teams (CDS-T)

86

Attribution Scale
Please answer all questions as honestly as possible. All answers will be treated with the strictest of confidence.
1. How well do you feel your team has performed so far this season (please circle one)?
not well at all

not well

so-so

well

very well

2. Please write what you feel to be the most likely CAUSE of your team’s performance so far this season.
________________________________________________________________________
Now please complete the following questions: Think about the cause you wrote above.
The items below concern your opinions of this cause or causes of your team’s performance.
Please circle one number for each of the following questions.
Is the cause above something:
1. Caused by something in your team
2. Your team can do something about
3. Permanent (won’t change)
4. Your team can control
5. Controllable by people outside your team
6. Inside the team
7. That doesn’t change over time
8. Under the power of people outside the team
9. Due to your team
10. Over which your team has power
11. Unchangeable
12. People outside the team can adjust
13. About your team
14. Controllable by your team
15. Stays the same across games
16. Determined by people outside the team

9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9

8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8

7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7

6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6

5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5

4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4

87

3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3

2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2

1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1

Caused by something outside the team
Your team can do nothing about
Temporary (will change)
Your team cannot control
That no one can control
Outside the team
That changes over time
Not under the power of people outside the team
Due to stuff outside the team
Over which your team has no power
Changeable
People outside the team cannot adjust
About the situation
Not controllable by your team
Changes from game to game
Not determined by people outside the team

Appendix F
Coach Rankings

88

For each of the following defensive skills, rate all your athletes from MOST EFFECTIVE (1.) to LEAST EFFECTIVE (15.)
Pass a tough serve
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.

_
_
_
_
_
_
_
_
_
_
_
_

Dive/roll when necessary to play a ball
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.

_
_
_
_
_
_
_
_
_
_
_
_
_
_
_

Pick up a tip
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.

_
_
_
_
_
_
_
_
_
_
_
_

Back up a teammates if they are out of
position or miss a shot
1.
_
2.
_
3.
_
4.
_
5.
_
6.
_
7.
_
8.
_
9.
_
10.
_
11.
_
12.
_
13.
_
14.
_
15.
_

89

Defend against a hard driven attack
1.
_
2.
_
3.
_
4.
_
5.
_
6.
_
7.
_
8.
_
9.
_
10.
_
11.
_
12.
_

Name up to three of your athletes
you would consider your “go-to”
girls who you could count on if
your team was in a very crucial
game situation (1. = first “go-to”
girl; 2. = 2nd “go-to” girl, etc).
1.
2.
3.

_
_
_

REFERENCES

90

References
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change.
Psychological Review, 84, 191-215.
Bandura, A. (1986). Social Foundations of thought and action: A social cognitive theory.
Englewood Cliffs, NJ: Prentice Hall.
Bandura, A. (1990). Perceived self-efficacy in the exercise of personal agency. Journal of
Applied Sport Psychology, 2, 128-163.
Bandura, A. (1997). Self-efficacy: The exercise of control. New York: Freeman.
Bandura, A. (2006). Guide for creating self-efficacy scales. In F. Pajares & T. Urdan (eds.), Selfefficacy beliefs of adolescents (pp. 307-337). Greenwich, CT: Information Age
Publishing.
Barnes, C.M., Hollenbeck, J.R., Wagner, D.T., DeRue, D.S., Nahrgang, J.D., & Scwind, K.M.,
(2008). Harmful help: The costs of backing-up behavior in teams. Journal of Applied
Psychology, 93, 529-539. doi: 10.1037/0021-9010.93.3.529
Beauchamp, M.R., & Whinton, L.C. (2005). Self-efficacy and other-efficacy in dyadic
performance: Riding as one in equestrian eventing. Journal of Sport and Exercise
Psychology, 27, 245-252.
Blau, P. (1964). Exchange and power in social life. New York, NY: Wiley.
Clark, M., & Gronbech, C.E. (1987). The effect of age, sex, and participation in age group
athletics on the development of trust in children. International Journal of Sport
Psychology, 18, 181-187.
Dirks, K.T. (1999). The effect of interpersonal trust on work group performance. Journal of
Applied Psychology, 84, 445-455.
Dirks, K.T. (2000). Trust in leadership and team performance: Evidence from NCAA basketball.
Journal of Applied Psychology, 85, 1004-1012.
Dirks, K.T., & Ferrin, D.L. (2001). The role of trust in organizational settings. Organization
Science, 12, 450-467.
Dirks, K.T., & Ferrin, D.L. (2002). Trust in Leadership: Meta-analytic findings and
implications for research and practice. Journal of Applied Psychology, 87, 611-628.
Dirks, K.T., & Skarlicki, D.P. (2009). The relationship between being perceived as trustworthy
by coworkers and individual performers. Journal of Management, 35, 136-157. doi:
10.1177/0149206308321545

91

Feltz, D.L, (1982). A path analysis of the causal elements in Bandura’s theory of self-efficacy
and an anxiety-based model of avoidance behavior. Journal of Personality and Social
Psychology, 42 764-781.
Feltz, D.L. (1988). Gender differences in the causal elements of self-efficacy on a highavoidance motor task. Journal of Sport and Exercise Psychology, 10, 151-166.
Feltz, D.L., Chow, G.M., & Hepler, T.J. (2008). Path analysis of self-efficacy and diving
performance revisited. Journal of Sport & Exercise Psychology, 30, 401-411.
Feltz, D.L., Landers, D.M., & Reader, U. (1979). Enhancing self-efficacy in high-avoidance
motor tasks: A comparison of modeling techniques. Journal of Sport Psychology, 1, 112122.
Feltz, D.L., & Lirgg, C.D. (1998). Perceived team and player efficacy in hockey. Journal of
Applied Psychology, 83(4), 557-564.
Feltz, D.L., & Mugno, D.A. (1983). A replication of the path analysis of the causal elements in
Bandura’s theory of self-efficacy and the influence of automatic perception. Journal of
Sport Psychology, 5, 263-277.
Feltz, D.L., Short, S.E., & Sullivan, P.J. (2008). Self-efficacy in sport: Research and strategies
for working with athletes, teams, and coaches. Champaign, IL: Human Kinetics.
Gilson, T.A., Chow, G.M, & Feltz, D.L. (in press). Self-efficacy and athletic squat performance:
Positive or negative influences at the within and between levels of analysis? Journal of
Applied Social Psychology.
Golembiewski, R., & McConkie, M. (1975). The centrality of interpersonal trust in group
process. In C. Cooper (Ed.), Theories of group process (pp. 131-185). New York: Wiley.
Greenlees, I.A., Lane, A., Thelwell, R., Holder, T., & Hobson, G. (2005). Team-referent
attributions among sport performers. Research Quarterly for Exercise and Sport, 76, 477487.
Heuzé, J.P., Raimbault, N. & Fontayne, P. (2006). Relationships between cohesion, collective
efficacy and performance in professional basketball teams: An examination of mediating
effects. Journal of Sports Sciences, 24, 59-68.
Hogdes, L., & Carron, A. V. (1992). Collective efficacy and group performance. International
Journal of Sport Psychology, 23, 48-59.
Jackson, B., Beauchamp, M.R., & Knapp, P. (2007). Relational efficacy beliefs in athlete dyads:
An investigation using actor-partner interdependence models. Journal of Sport and
Exercise Psychology, 29, 170-189.

92

Jackson, B., Knapp, P., & Beauchamp, M.R. (2008). Origins and consequences of tripartite
efficacy beliefs within elite athlete dyads. Journal of Sport and Exercise Psychology, 30,
512-540.
Jung, D.I., & Sosik, J.J. (2003). Group potency and collective efficacy: Examining their
predictive validity, level of analysis, and effects of performance feedback on future group
performance. Group and Organization Management, 28, 366-391. doi:
10.1177/1059601102250821
Lent, R.W., & Lopez, F.G. (2002). Cognitive ties that bind: A tripartite view of efficacy
beliefs in growth-promoting relationships. Journal of Social and Clinical Psychology, 21,
279-291.
Lichacz, F. M., & Partington, J. T. (1996). Collective efficacy and true group performance.
International Journal of Sport Psychology, 27, 146-158.
Lynch, P., Jonson, P., & Dibbeen, M. (2007). Exploring relationships of trust in ‘adventure’
recreation. Leisure Studies, 26, 47-64. doi: 10.1080/02614360500284767
Mayer, R.C., Davis, J.H., & Schoorman, F.D. (1995). An integrative model of organizational
trust. Academy of Management Review, 20, 709-734.
McAuley, E. (1985). Modeling and self-efficacy: A test of Bandura’s model. Journal of Sport
Psychology, 7, 283-295.
McGowan, S.J., & McGowan, R.W. (1991). Trust and adolescent sports: Starters vs
nonstarters. Perceptual and Motor Skills, 73, 714. doi: 10.2466/PMS.73.7.714-714
McKnight, D.H., Cummings, L.L., & Chervany, N.L. (1998). Initial trust formation in new
organizational relationships. The Academy of Management Review, 23, 473-490. doi:
10.2307/259290
Moritz, S.E., Feltz, D.L., Fahrbach, K.R., & Mack, D.E. (2000). The relation of self-efficacy
measures to sport performance: A meta-analytic review. Research Quarterly for Exercise
and Sport, 71, 280-294.
Myers, N.D., & Feltz, D.L. (2007). From self-efficacy to collective efficacy in sport: Transitional
issues. In G. Tenenbaum & R.C. Eklund (Eds.), Handbook of sport psychology (3rd ed.)
(pp. 799-819). New York, NY: Wiley.
Myers, N.D., Feltz, D.L., & Short, S.E. (2004). Collective efficacy and team performance: A
longitudinal study of collegiate football teams. Group Dynamics: Theory, Research, and
Practice, 8, 126-138.
Myers, N.D., Payment, C.A., & Feltz, D.L. (2004). Reciprocal relationships between collective

93

efficacy and team performance in women’s ice hockey. Group Dynamics: Theory,
Research, and Practice, 8, 182-195.
Nicolls, J.G. (1990). What is ability and why are we mindful of it? A developmental perspective.
In R.J. Sternberg & J. Kolligian (Eds.), Competence considered. New Haven, CT: Yale
University Press.
Porter, C.O.L.H. (2005). Goal orientation: Effects on backing up behavior, performance,
efficacy, and commitment in teams. Journal of Applied Psychology, 90, 811-818. doi:
10.1037/0021-9010.90.4.811
Porter, C.O.L.H., Hollenbeck, J.R., Ilgen, D.R., Ellis, A.P.J., West, B.J., & Moon, H. (2003).
Backing up behaviors in teams: The role of personality and legitimacy of need. Journal
of Applied Psychology, 88, 391-403. doi: 10.1037/0021-9010.88.3.391
Raudenbush, S.W., & Bryk, A.S. (2002). Hierarchical linear models: Applications and data
analysis methods. Newbury Park, CA: Sage.
Rousseau, D.M., Sitkin, S.B., Burt, R.S., & Camerer, C. (1998). Not so different after all: A
cross-discipline view of trust. Academy of Management Review, 23, 393-404.
Shen, M.J., & Chen, M.C. (2007). The relationship of leadership, team trust and team
performance: A comparison of the service and manufacturing industries. Social
Behavior and Personality, 35, 643-658.
Snijders, T., & Bosker, R. (1999). Multilevel analysis: An introduction to basic and advanced
multilevel modeling. Newbury Park, CA: Sage.
Spink, K.S. (1990). Group cohesion and collective efficacy of volleyball teams. Journal of Sport
and Exercise Psychology, 12, 301-311.
Vancouver, J.B., & Kendall, L.N. (2006). When self-efficacy negatively relates to motivation
and performance in a learning context. Journal of Applied Psychology, 91, 1146-1153.
doi: 10.1037/0021-9010.91.5.1146
Vancouver, J.B., Thompson, C.M., Tischner, E.C., & Putka, D.J. (2002). Two studies examining
the negative effect of self-efficacy on performance. Journal of Applied Psychology, 87,
506-516. doi: 10.1037/0021-9010.87.3.506
Vancouver, J.B., Thompson, C.M., & Williams, A.A. (2001). The changing signs in the
relationships among self-efficacy, personal goals, and performance. Journal of Applied
Psychology, 86, 605-620. doi: 10.1037/0021-9010.86.4.605
Weinberg, R. (1985). Relationship between self-efficacy and cognitive strategies in enhancing
endurance performance. International Journal of Sport Psychology, 17, 280-292.

94

Weinberg, R., Gould, D., & Jackson, A. (1979). Expectations and performance: An empirical test
of Bandura’s self-efficacy theory. Journal of Sport Psychology, 1, 320-331.
Weinberg, R.S., Gould, D., Yukelson, D., & Jackson, A. (1981). The effects of pre-existing and
manipulated self-efficacy on a competitive muscular endurance task. Journal of Sport
Psychology, 3, 345-354.
Weinberg, R.S., Yukelson, D., & Jackson, A. (1980). Effect of public and private efficacy
expectations on competitive performance. Journal of Sport Psychology, 2, 340-349.

95