A confirmatory factor analysis of the coach behavior scale for sport.
Sullivan, Philip J. ; Whitaker-Campbell, Tammy ; Bloom, Gordon A. 等
The coaching process in sport is in part defined by the outcomes
which the athlete participants accrue. Although the most obvious of
these outcomes may be team or individual performance, it is also
recognized that coaching may affect a variety of psychological
characteristics which may extend beyond sport performance. Particularly
within youth sport, research has focused on how coaches influence the
behaviors and attitudes of participants (Fraser-Thomas & Cote, 2009;
Weiss & Fretwell, 2005). These outcomes, including athletes'
self-perception, self-confidence, self-esteem, commitment, and enjoyment
and adherence to sport (Newin, Bloom, & Loughead, 2008; Smith &
Smoll, 2002; Weiss, Barber, Sisley, & Ebbeck, 1991) can be either
positively or negatively affected by youth sport coaching. Cote and
Gilbert's (2009) definition of coaching effectiveness highlighted
the use of the coaching practice to optimize these positive outcomes,
fostering the personal growth and development of athletes. Because such
outcomes of effective youth sport coaching are obviously significant, it
is important that valid and reliable measures of youth sport coaching
exist so as to support this important applied and basic research.
Various coaching models have been developed to conceptually
describe the interaction between coaches' behaviors and
athletes' performance and personal development (e.g.,
Multidimensional Model of Leadership (MML): Chelladurai, 1978; Coaching
Model (CM): Cote, Salmela, Trudel, Baria, & Russell, 1995;
Medational Model of Leadership: Smoll & Smith, 1989). Subsequently,
assessment tools for these conceptual models have been generated and
validated, although there are various results for the support of the
psychometric properties of these scales.
Chelladurai's MML is one of the most widely used models of
effective leadership in sport, and one of the primary aspects of
leadership studied has been coaching. The MML was created to provide a
framework to identify effective leadership behaviors specific for sport
settings. Central to the model are three states of coaches'
behaviors: required, actual, and preferred by athletes. These behaviors
are influenced by three antecedents: situational characteristics, coach
characteristics, and athlete characteristics. The model's outcome
is that athletes' satisfaction or performance is related to the
congruence between the coaches' behaviors and antecedent
components. The Leadership Scale for Sports (LSS) was built to test the
relationship between the constructs in the MML and its application to
the predictions of leadership effectiveness in sport (Chelladurai,
2007). It assesses five dimensions of coaching leadership: training and
instruction, democratic behavior, autocratic behavior, social support,
and positive feedback (Chelladurai & Saleh, 1980). Since its
development, research has supported the content and construct validity
of the LSS (Chelladurai & Carron, 1981; Chelladurai & Riemer,
1998; Dwyer & Fischer, 1988; Price & Weiss, 2000; Zhang, Jensen,
& Mann, 1997).
Although the MML and the LLS have been used to research coaching,
they were not exclusively designed for this context. Smoll and
Smith's (1989) meditational model of leadership was developed
specifically to support research in sport coaching. The model proposes
athletes' perceptions and recollection of their coaches'
behaviors mediates their attitudes towards the coach and their
experience in sport. This model focuses not only on situational factors
and observable behaviors, but also on the cognitive processes and
individual difference which affect the relationship between antecedent,
leader behaviors, and outcomes. The Coaching Behavior Assessment System
(CBAS) is an observation evaluation tool designed by Smith, Smoll, and
Hunt (1977) to measure coaches' leadership behaviors. Twelve
dimensions of coaching behavior compose two main classes--reactive and
spontaneous behaviors (Smith et al., 1977). Reactive behaviors are
immediate responses to preceding events usually in response to an
athlete's or teams behavior and performance (e.g., responses to
desirable performances, mistakes, and misbehaviors). Spontaneous
behaviors include game-related (i.e., general technical instruction,
general encouragement, and organization) and game-irrelevant behaviors
(i.e., general communication). In a review of Smith and Smoll's
work, Chelladurai and Riemer (1998) concluded that this 12 dimension
scheme captures most of the meaningful coaching behaviors. Results
gathered from studies using the CBAS guided the development of the
mediational model of leadership (Smoll & Smith, 1989).
The CM is another widely recognized model that has contributed to
the coaching literature. Like Smoll and Smith (1989) model, the CM
focuses exclusively on sport. Furthermore, because one of its focal
points was athlete development, it is particularly applicable to youth
sport. Created by Cote and colleagues (1995), the CM identifies the
variables that influence coaches' behaviors and helps to understand
how coaches' behavior influences athlete development. Central to
the CM are the elements of training, competition, and organization.
These elements are monitored and adjusted by the coach's mental
model of the athletes' potential, which is determined by the
coach's assessment of three peripheral components: coach's
personal characteristics, athlete's personal characteristics, and
contextual factors. The peripheral components are the coach's
representation of the ultimate goal of athlete development.
The Coaching Behavior Scale for Sport (CBS-S) is an assessment
developed to measure the elements of the CM (Cote, Yardley, Hay,
Sedgwick, & Baker, 1999). The items for the CBS-S were generated in
a series of qualitative studies with coaches and athletes (Bloom,
Durand-Bush, & Salmela, 1997; Cote, 1998; Sedgwick, Cote, &
Dowd, 1997). These items were examined for face validity by a panel of
experts including both scholars and coaches. A preliminary version of
the CBS-S comprising these items was administered to two different
samples (105 rowers and 205 participants in a wide variety of sports).
Two separate exploratory factor analyses (EFA) with independent samples
revealed a seven factor model (i.e., physical training and conditioning,
technical skills, mental preparation, competition strategies, goal
setting, personal rapport, and negative personal rapport) that fit the
CM (Cote et al., 1999). Training behavior was reflected in physical
training and conditioning, technical skills, and mental preparation.
Competitive behaviors were measured by competition strategies.
Organizational behavior was reflected in goal setting, and personal
rapport. The scale also showed some acceptable psychometric properties.
The Cronbach's alphas for all scales exceeded .85, showing
excellent internal consistency, and a three week test-retest trial
showed reliability of .68 or above for all positive scales, although the
r for negative rapport was somewhat low, at .45 (Cote et al., 1999).
Research using the CBS-S has found that the construct of
athlete-centered coaching behavior was related to several significant
consequences (Baker, Cote, & Hawes, 2000; Baker, Yardley, &
Cote, 2003). More specifically, all seven of the CBS-S scores were
significant predictors of athletes' satisfaction with their coaches
(Baker et al., 2003). Furthermore, type of sport moderated six of these
relationships (all except physical training), where coaching behavior
was a more significant predictor of satisfaction for team sport than
individual sport athletes. The outcome of athlete satisfaction was
significantly related to CBS-S behaviors of negative rapport and
competition strategies (Baker et al., 2000).
The current seven-factor structure of the CBS-S is the result of
exploratory factor analyses (EFA) (Cote et al., 1999). Although the EFA
produced results consistent with the theoretical framework of the CM and
has shown acceptable levels of internal consistency, it has not been
confirmed using CFA procedures. Whereas an EFA examines the correlations
among variables to suggest a viable structure of latent variables or
themes within the data, a CFA tests the fit of a particular model to the
data (Tabachnick & Fidell, 2007). The CFA will indicate which
variables load into the factors of the particular model, and which of
these correlate. Thus, an EFA is appropriate when no model is known or
suggested and a CFA is appropriate when previous empirical or conceptual
work suggests a model (Kline, 2005). The primary purpose of the present
study was to extend the psychometric support for the CBS-S by confirming
the seven-factor model through CFA.
A secondary purpose for this analysis was to examine potential
invariance by gender with respect to the factor structure of the CBS-S.
Previous research has suggested that males and females may have
different perceptions and expectations of coaches (Amorose & Horn,
2000; Beam, Serwatka, & Wilson, 2004). For example, male collegiate
athletes preferred more social support than female athletes, whereas
their female counterparts preferred greater training and instruction
(Beam et al., 2004). Within youth sport, both adolescent athletes and
their parents have gender preferences for coaches. More specifically,
male athletes and their parents preferred male coaches, whereas gender
of the coach was not important to female athletes and their parents
(Martin, Dale, & Jackson, 2001). Based on the potential of gender
differences on athlete perceptions of coaching behaviors, an analysis of
invariant factor structure was also conducted to see if the factor
structure of the CBS-S varied by male or female youth sport athletes.
Method
Participants
Three hundred and forty five youth athletes (193 male, 152 female)
who participated in either basketball (n = 124), ice hockey (n = 114) or
soccer (n = 99) completed the questionnaire (8 participants did not
indicate sport). The athletes ranged in age from 13 to 18 years old,
with a mean age of 15.17 years (SD = 1.28), and had averaged 7.96 years
(SD = 2.81) playing their sport.
Measures
The CBS-S (Cote et al., 1999) is a 47-item scale that measures
seven interrelated aspects of coaching behavior from the athlete's
point of view. All items are answered on a 7-point scale frequency based
scale from 1 (never) to 7 (always). Physical training and conditioning
(e.g., provides me with structured training sessions), and competition
strategies (e.g., keeps me focused in competitions) were each measured
by seven items. Technical skills (e.g., give me reinforcement about
correct technique), and negative rapport (e.g., yells at me when angry)
both had eight items. Goal setting (e.g., helps me set short term goals)
and positive rapport (e.g., is a good listener) were both measured by
six items each, and mental preparation (e.g., provides advice on how to
stay focused) had five items. As previously mentioned, research has
supported the reliability, face, and construct validity of the CBS-S
(Cote et al., 1999). With the current sample, all factors appeared to be
internal consistent, with Cronbach's alphas ranging from .83 to
.94.
Procedures
Upon receiving institutional review clearance, researchers
contacted youth sport league administrators via their publically
available contact information. The league administrators were asked to
forward a message to the coaches containing an explanation of the study
and the principle investigator's contact information. Once the
interested coaches replied to the message, the investigator attended a
team meeting or practice to explain the nature of the study to their
athletes. Interested athletes were given a questionnaire package that
included a letter of invitation, a consent form for their parents to
sign, and the CBS-S. Athletes were instructed to take the package home,
and, if their parents consented, to complete the survey. The surveys
were collected at the following practice. Each athlete who was solicited
consented to participate in the study and returned a completed survey.
Results
The data were examined for assumptions of multivariate data
analysis. All variables were normally distributed, and there were no
instances of multicollinearity in the data set. Statistics for Variance
Inflation Factor (VIF) and tolerance were within the criteria of less
than 10 and greater than 0.2, respectively, for all variables.
Furthermore no correlations among the variables exceeded .90. Table 1
gives the descriptive statistics of the CBS-S items. Several of the
items were characterized by high levels of kurtosis and negative
skewness. Specifically, these were the items that comprised the factor
of Negative Rapport; most of the athletes responded with a 1 (never) on
the 7 point scale frequency. However, these items were retained because
previous work has supported their face and construct validity (Baker et
al., 2000; Baker et al., 2003; Cote et al., 1999). Furthermore,
indicators of negative coaching behavior (e.g., my coach intimidates me
physically) were considered to be important even if not frequently
reported. To accommodate for this abnormal distribution, the subsequent
CFA was conducted with a heterogeneous kurtosis RLS method to
accommodate for this abnormality in the data (Byrne, 2008). All
structural equation model (SEM) analyses were conducted using EQS 6.1.
Accepted criteria for excellent goodness of fit for a CFA include
values above 0.95 on the CFI and other goodness of fit indexes (Hu &
Bentler, 1999), and 0.05 or less for the RMSEA (Browne & Cudeck,
1993). The results of the CFA showed an excellent fit of the model to
the data (NNFI = .984, CFI = .985; RMSEA = .044). Each of the factors
loaded significantly on its requisite items.
The only other proposed model for this data in the literature is
that the seven factors of the CBS-S comprise the three conceptual
factors of the CM--namely competitive, organizational, and training
coaching elements. Specifically, organizational behaviors comprise goal
setting, positive rapport and negative rapport; training behaviors
consists of mental preparation, technical skills, and physical training
and conditioning; and competitive strategies is the solitary factor for
competitive behaviors (Cote et al., 1999). For comparative purposes,
this model was tested. The results for this analysis revealed indices
approaching an excellent fit of the model to the data (NNFI = .945, CFI
= .948; RMSEA = .080). In comparison to the above results for the
seven-factor model, these indices were not as strong. The difference
between models was statistically significant change in chi-square
([DELTA][[chi square].sub.(10)] = 1602.73, p <.05). These results
suggest that the model of seven separate, but interrelated, factors of
coaching behavior may be the most appropriate model for this data. Table
2 gives the standardized factor loadings and errors for this solution.
Analyses of invariance by gender for the seven-factor model
followed the suggestions of Byrne (2008), which entailed testing the fit
of the model to the data of both samples, then gradually constraining
the model so that certain aspects could be freely estimated yet not vary
between groups. More specifically, first the factor loadings were
constrained and then the error variances were constrained. Each of these
models generated goodness of fit indicators that were interpreted to see
if the model fitted adequately to both groups. Table 3 summarizes the
fit indices for the two samples for each step of the analysis of
invariance. The model sustained indicators of excellent fit to both
samples throughout these procedures, showing that the factor structure
of the CBS-S did not vary between males and females.
Discussion
The primary purpose of this study was to examine the factor
structure of the CBS-S with a sample of youth sport athletes, and a
secondary purpose was to examine the invariance of this factor structure
between genders. The results showed an excellent fit of the model to the
data with this sample, which does not vary between males and females.
However, these analyses were based on a method of SEM that acknowledged
the abnormality within the data due to the fact that the scale includes
some items that, by design, will typically have skewed results (i.e.,
indicators of negative rapport with one's coach). The current
results support the factor structure of the seven-factor measurement
model upon which the CBS-S is based. These seven factors appear to
comprise a valid measurement model for athletes' perspective of
coaching behavior in youth sport, for both male and female athletes.
Factor analyses examine the correlations among items in a
measurement to indicate how the presence of more parsimonious themes
within a multi-item measure. With respect to the CBS-S, its items were
generated from a series of qualitative interviews with athletes,
coaches, and coach educators (Cote et al., 1999). Although the resultant
47 items were all important from the perspective of those who generated
them, it was highly unlikely that the CBS-S measured 47 distinct
concepts. The results of the current analyses, which are consistent with
the exploratory analysis of Cote et al., support that a more
parsimonious factor structure exists to represent the range of coaching
behaviors assessed by these original qualitative studies.
Both the EFAs conducted by Cote et al. (1999) and the current CFA
support the presence of seven key factors in the athlete-centered
measurement of coaching behavior in youth sport--physical training and
conditioning, technical skills, mental preparation, competition
strategies, goal setting, personal rapport, and negative personal
rapport. The contribution of the current analysis was confirming these
interrelated factors with a distinct sample from the EFA study. The
Cronbach's alphas in the current study also supported the
reliability of these factors, which was again consistent with Cote et
al.'s earlier research. Furthermore, the current analysis revealed
that the seven-factor model suggested by Cote et al. was equally valid
and reliable for young male and female athletes.
Results from the current study supported that the measurement model
of the CBS-S was a good fit for both male and female youth sport
participants. We cannot conclude that perceptions of coaching behavior
were similar for males and females since the scale includes behaviors
favored by both males (social support) and females (training and
instruction). It was beyond the purpose of this study to analyze gender
differences in the perceptions and expectations of coaches. Therefore,
it would be appropriate to measure these issues within youth sport.
With respect to psychometric properties, the CBS-S compares
favorably with other measures of coaching behavior, particularly with
respect to factor structure. A review of the literature revealed that
the CBAS has never been subjected to a CFA. This is understandable
considering that the instrument was designed as an observational measure
and it would not be reasonable to assume that a study based on coach
observation would include a sample large enough for a CFA. The factor
structure of the LSS (and its revised version, the RLSS) has been
supported by CFA (Sullivan & Kent, 2003; Zhang et al., 1997), but
not to any examination of invariance. Our analysis supports the CBS-S as
an equally valid and reliable measure to these other scales, and a
useful alternative for researchers, particularly for research focused on
youth sport, and athlete-centered assessments of coaching behavior.
The assessment of coaches' behaviors is essential for the
profession's development and to increase coach effectiveness.
Current definitions of coach effectiveness highlight that good coaching
goes beyond teaching sports skills and includes the person's
holistic development (Cote & Gilbert, 2009; Vallee & Bloom,
2005). The current study provides additional validity to an assessment
tool which can help disseminate good coaching habits and promote youth
athletes' development outcomes that includes improved
self-perception, self-confidence, self-esteem, commitment, and enjoyment
and adherence to sport.
Finally, it must be acknowledged that there are several limitations
in the current study. This sample comprised three sports, all of which
are interactive team sports. Furthermore, the teams and athletes were
all from similar urban areas, which may have affected the organizational
culture of the sport club, or the likelihood that coaches had been
certified. Although the current study does support the psychometric
properties of the CBS-S, future research with a different and more
heterogeneous sample of youth sport athletes could help to address these
limitations.
References
Amorose, A. J., & Flom, T. S. (2000). Intrinsic motivation:
Relationships with collegiate athletes' gender, scholarship status,
and perceptions of their coaches' behavior. Journal of Sport &
Exercise Psychology, 22, 63-84.
Baker, J., Cote, J., & Hawes, R. (2000). The relationship
between coaching behaviours and sport anxiety in athletes. Journal of
Science and Medicine in Sport, 3, 110-119.
Baker, J., Yardley, J., & Cote, J. (2003). Coach behaviors and
athlete satisfaction in team and individual sports. International
Journal of Sport Psychology, 43, 226-239.
Beam, J. W., Serwatka, T. S., & Wilson, W. J. (2004). Preferred
leadership of NCAA division I and II intercollegiate student-athletes.
Journal of Sport Behavior, 27, 3-17.
Bloom, G. A., Durand-Bush, N., & Salmela, J. H. (1997). Pre-and
postcompetition routines of expert coaches of team sports. The Sport
Psychologist, 11, 127-141.
Browne, T. W., & Cudeck, R. (1993). Alternative ways of
assessing model fit. In K. A. Bollen & J. S. Long (Eds.), Testing
structural equation models (pp. 136-162). Newbury Park, CA: Sage.
Byrne, B. M. (2008). Structural equation modeling with EQS. New
York: Routledge.
Chelladurai, P. (1978). A contingency model of leadership in
athletics. Unpublished doctoral dissertation, Department of Management
Sciences, University of Waterloo, Canada.
Chelladurai, P. (2007). Leadership in sports. In G. Tenenbaum &
R. C. Eklund (Eds.), Handbook of sport psychology (3rd ed., pp.
113-135). Hoboken, NJ: John Wiley.
Chelladurai, P., & Carron, A. V. (1981). Applicability to youth
sport of the Leadership Scale for Sports. Perceptual & Motor Skills,
53, 361-362.
Chelladurai, P., & Riemer, H. A. (1998). Measurements of
leadership. In J. L. Duda (Ed.), Advances in sport and exercise
psychology (pp. 227-253). Morgantown, WV: Fitness Information
Technology.
Chelladurai, P., & Saleh, S. (1980). Dimensions of leader
behavior in sports: Development of a leadership scale. Journal of Sport
Psychology, 2, 34-45.
Cote, J. (1998). Coaching research and intervention: An
introduction to the special issue. Avante, 4, 1-15.
Cote, J., & Gilbert, W. (2009). An integrative definition of
coaching effectiveness and expertise. International Journal of Sports
Science & Coaching, 4, 307-323.
Cote, J., Salmela, J., Trudel, P., Baria, A., & Russell, S.
(1995). The coaching model: A grounded assessment of expert gymnastic
coaches' knowledge. Journal of Sport & Exercise Psychology, 17,
1-17.
Cote, J., Yardley, J., Hay, J., Sedgwick, W., & Baker, J.
(1999). An exploratory examination of the coaching behavior scale for
sport. Avante, 5, 82-92.
Dwyer, J. M., & Fischer, D. G. (1988). Psychometric properties
of the coach's version of leadership scale for sports. Perceptual
& Motor Skills, 67, 795-798.
Fraser-Thomas, J., & Cote, J. (2009). Understanding
adolescents' positive and negative developmental experiences in
sport. The Sport Psychologist, 23, 3-23.
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit
indexes in covariance structure analysis: Conventional criteria versus
new alternatives. Structural Equation Modeling: A Multidisciplinary
Journal, 6, 1-55.
Kline, R. B. (2005). Principles and practice of structural equation
modeling (2nd ed.). New York: Guilford.
Martin, S. B., Dale, G. A., & Jackson, A. W. (2001). Youth
coaching preferences of adolescent athletes and their parents. Journal
of Sport Behavior, 24, 197-212.
Newin, J., Bloom, G. A., & Loughead, T. M. (2008). Youth ice
hockey coaches' perceptions of a team-building intervention
program. The Sport Psychologist, 22, 54-72.
Price, M. S., & Weiss, M. R. (2000). Relationships among coach
burnout, coach behaviors, and athletes' psychological responses.
The Sport Psychologist, 14, 391-409.
Sedgwick, W. A., Cote, J., & Dowd, J. (1997). Confidence
building strategies used by Canadian high level rowing coaches. Avante,
3, 80-92.
Smoll, F. L. & Smith, R. E. (1989). Leadership behaviors in
sport: A theoretical model and research paradigm. Journal of Applied
Social Psychology, 19, 1522-1551.
Smith, R. E., & Smoll, F. L. (2002). Way to go coach! A
scientifically proven approach to youth-sports coaching effectiveness
(2nd ed.). Palo Alto, CA: Warde.
Smith, R. E., Smoll, F. L., & Hunt, E. (1977). A system for the
behavioral assessment of athletic coaches. Research Quarterly, 48,
401-407.
Sullivan, P. J., & Kent, A. (2003). Coaching efficacy as a
predictor of leadership style in intercollegiate athletics. Journal of
Applied Sport Psychology, 15, 1-11.
Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate
statistics (5th Ed.). Boston: Pearson.
Vallee, C. N., & Bloom, G. A. (2005). Building a successful
university program: Key and common elements of expert coaches. Journal
of Applied Sport Psychology, 17, 179-196.
Weiss, M. R., Barber, H., Sisley, B. L., & Ebbeck, V. (1991).
Developing competence and confidence in novice female coaches: 11.
Perceptions of ability and affective experiences following a season-long
coaching internship. Journal of Sport & Exercise Psychology, 13,
336-363.
Weiss, M. R., & Fretwell, S. D. (2005). The
parent-coach/child-athlete relationship in youth sport: Cordial,
contentious, or conundrum? Research Quarterly for Exercise & Sport,
76, 286-305.
Zhang, J., Jensen, B., & Mann, B. (1997). Modification and
revision of the Leadership Scale for Sports. Journal of Sport Behavior,
20, 105-122.
Philip J. Sullivan
Tammy Whitaker-Campbell
Brock University
Gordon A. Bloom
William R. Falcao
McGill University
Address correspondence to: Philip Sullivan, Department of
Kinesiology, Brock University, St. Catharines, ON, L2S 3A1. Email:
phil.sullivan@brocku.ca
Table 1
Descriptive statistic of the CBS-S variables.
CBS-S M (SD) Skewness Kurtosis CBS-S
Items Items
1 5.28 (1.52) -0.90 0.36 25
2 5.13 (1.59) -0.70 -0.11 26
3 4.73 (1.76) -0.45 -0.72 27
4 4.82 (1.66) -0.42 -0.65 28
5 5.77 (1.37) -1.11 0.64 29
6 5.72 (1.42) -1.20 0.93 30
7 4.33 (1.97) -0.32 -1.00 31
8 5.97 (1.13) -1.01 0.42 32
9 6.02 (1.15) -1.23 1.00 33
10 5.80 (1.22) -0.95 0.48 34
11 5.98 (1.17) -1.22 1.05 35
12 5.80 (1.34) -1.23 1.08 36
13 5.85 (1.30) -1.28 1.44 37
14 5.86 (1.23) -0.93 0.10 38
15 5.58 (1.30) -0.84 0.33 39
16 4.82 (1.60) -0.50 -0.44 40
17 5.09 (1.70) -0.65 -0.51 41
18 5.28 (1.52) -0.94 0.28 42
19 5.28 (1.51) -0.80 0.19 43
20 5.52 (1.50) -1.03 0.70 44
21 5.07 (1.56) -0.63 -0.10 45
22 4.81 (1.68) -0.47 -0.56 46
23 4.73 (1.72) -0.39 -0.71 47
24 4.25 (1.91) -0.09 -1.15
CBS-S M (SD) Skewness Kurtosis
Items
1 4.76 (1.73) -0.49 -0.63
2 4.77 (1.78) -0.49 -0.74
3 5.79 (1.26) -0.98 0.37
4 5.76 (1.35) -1.16 0.88
5 5.74 (1.32) -1.29 1.47
6 5.50 (1.44) -1.08 0.80
7 5.36 (1.53) -0.90 0.09
8 5.85 (1.31) -1.32 1.56
9 5.74 (1.45) -1.18 0.70
10 5.71 (1.41) -1.20 1.09
11 5.70 (1.42) -1.14 0.99
12 5.28 (1.74) -0.84 -0.32
13 5.08 (1.82) -0.70 -0.57
14 5.28 (1.72) -0.91 -0.04
15 5.38 (1.79) -1.10 0.25
16 1.97 (1.49) 1.75 2.40
17 2.26 (1.57) 1.34 0.98
18 1.81 (1.33) 2.09 4.11
19 2.19 (1.61) 1.38 0.94
20 1.39 (1.02) 3.69 15.13
21 1.32 (0.88) 3.62 14.53
22 1.42 (1.02) 3.45 13.15
23 1.63 (1.28) 2.47 5.79
24
Note. CBS-S variables are measured on a scale of 1-7 with higher
numbers indicating greater frequency.
Table 2
Standardized factor loadings and error terms.
CBS-S Factor Factor Error CBS-S
Items loading Items
1 Technical skills .766 .643 27
2 .737 .676 28
3 .767 .641 29
4 .689 .724 30
5 .527 .850 31
6 .644 .765 32
7 .520 .854 33
8 Physical training .731 .683 34
9 and conditioning .748 .664 35
10 .753 .659 36
11 .781 .625 37
12 .647 .762 38
13 .660 .751 39
14 .694 .720 40
15 .645 .764 41
16 Mental preparation .760 .649 42
17 .757 .653 43
18 .770 .638 44
19 .784 .621 45
20 .762 .647 46
21 Goal setting .735 .678 47
22 .766 .642
23 .795 .607
24 .742 .670
25 .763 .646
26 .741 .671
CBS-S Factor Factor Error
Items Loading
1 Competitive .784 .621
2 strategies .736 .677
3 .761 .649
4 .678 .735
5 .671 .742
6 .682 .731
7 .588 .809
8 Positive Rapport .767 .641
9 .736 .677
10 .717 .697
11 .709 .706
12 .709 .705
13 .604 .797
14 Negative Rapport .505 .863
15 .593 .806
16 .630 .777
17 .598 .801
18 .578 .816
19 .654 .756
20 .635 .772
21 .626 .780
22
23
24
25
26
Note. Heterogeneous Kurtosis solution. All factor loadings significant
at p < .05.
Table 3
Summary of analysis of invariance
Model NNFI CFI RMSEA
Full sample (N = 345) .984 .985 .044
Female sample (n = 152) .988 .988 .033
Male sample (n = 193) .988 .988 .043
Multi-group model .940 .943 .083
Factor loadings constrained .943 .944 .081
Error variances constrained .943 .944 .080
