Commitment in Sport: The Role of Coaching Style and Autonomous vs. Controlled Motivation

Abstract

This study examined the relationships amongst coaching style (autonomy-supportive; controlling), athlete motivation (autonomous; controlled), sport commitment (enthusiastic; constrained), and intention to continue in sport. Participants were 154 competitive football/soccer players (M = 22.21 years). Partial least squares structural equation modelling (PLS-SEM) revealed positive associations among adaptive motivational (autonomy-supportive coaching style, autonomous motivation) and commitment (enthusiastic) variables. Similarly, positive links were found among maladaptive motivational (controlling coaching style, controlled motivation) and commitment (constrained) constructs. Additionally, results indicated differential relationships between enthusiastic and constrained dimensions of commitment and intention to continue. Findings are discussed in terms of their theoretical and practical implications.

Full text

RUNNING HEAD: COMMITMENT IN SPORT
Commitment in Sport: The Role of Coaching Style and Autonomous versus Controlled
Motivation
Liam O’Neil & Ken Hodge*
University of Otago, New Zealand
*Corresponding Author:
Ken Hodge, Ph.D.
School of Physical Education, Sport, & Exercise Sciences
University of Otago
46 Union Street, Dunedin, New Zealand
Email: ken.hodge@otago.ac.nz

COMMITMENT IN SPORT 1
Abstract
1
This study examined the relationships amongst coaching style (autonomy-supportive;
2
controlling), athlete motivation (autonomous; controlled), sport commitment (enthusiastic;
3
constrained), and intention to continue in sport. Participants were 154 competitive
4
football/soccer players (M = 22.21 years). Partial least squares structural equation modelling
5
(PLS-SEM) revealed positive associations among adaptive motivational (autonomy-supportive
6
coaching style, autonomous motivation) and commitment (enthusiastic) variables. Similarly,
7
positive links were found among maladaptive motivational (controlling coaching style,
8
controlled motivation) and commitment (constrained) constructs. Additionally, results
9
indicated differential relationships between enthusiastic and constrained dimensions of
10
commitment and intention to continue. Findings are discussed in terms of their theoretical and
11
practical implications.
12
Lay Summary: This study investigated the relationships among motivational factors (coach
13
interpersonal style; athlete motivational orientation), commitment perceptions toward sport,
14
and intentions to continue in sport (n = 154; M = 22.21 years). Overall, findings highlighted the
15
important combined role of both commitment and motivation in understanding and explaining
16
athletes' sport continuation.
17
18
19
20
21
22
23

COMMITMENT IN SPORT 2
Introduction
24
Sport commitment refers to the psychological state underlying athletes’ persistence in
25
sport over time (Scanlan, Chow, Sousa, Scanlan, & Knifsend, 2016). This psychological state
26
reflects a motivational force which binds athletes to their chosen sport, and consequently, is
27
said to facilitate patterns of continued participation. Over the past decade, scholars have
28
provided considerable evidence supporting the role of commitment as an important
29
psychological construct underpinning athletes’ sport persistence (e.g., Casper, Gray, & Stellino,
30
2007; Scanlan et al., 2016; Weiss & Weiss, 2006). The Sport Commitment Model (SCM;
31
Scanlan et al., 2016) represents the major theoretical framework guiding much of this sport
32
commitment research. Developed to examine commitment in sport and to explain why athletes
33
persist in this setting, the SCM distinguishes between two dimensions of sport commitment
34
(i.e., enthusiastic; constrained). Enthusiastic commitment represents athletes’ “desire and
35
resolve to persist in a sport over time” (Scanlan et al., 2016, p. 235). Constrained commitment,
36
by contrast, represents athletes’ “perceptions of obligation to persist in a sport over time”
37
(Scanlan et al., 2016, p. 235).
38
It could be argued that these two commitment dimensions share conceptual parallels
39
with the two broad types of motivation conceptualised in Self-Determination Theory (SDT;
40
Ryan & Deci, 2017): autonomous motivation (experiencing a full sense of volition, agency,
41
choice, and self-endorsement when engaging in an activity or behaviour) and controlled
42
motivation (feeling either externally or internally pressured or compelled to engage in an
43
activity or behaviour). Enthusiastic commitment, in a similar manner to autonomous
44
motivation, reflects engagement in sport due to feelings of agency, choice, and interest in the
45
activity itself (i.e., “want to”). While constrained commitment, in a similar fashion to controlled
46
motivation, represents sport involvement resulting from feelings of external control and
47
obligation on behalf of the participant (i.e., “have to”). Although scarce, research has provided
48

COMMITMENT IN SPORT 3
initial empirical support for the theoretical connection between SCM commitment and SDT
49
motivation (Garcia-Mas et al., 2010; Weiss & Weiss, 2003, 2006).
50
While certain features of multidimensional SCM commitment and SDT motivation
51
share conceptual similarities, they ultimately represent distinct psychological variables. From
52
a theoretical viewpoint, motivation is considered a global variable which energises and directs
53
behaviour (Ryan & Deci, 2017). As a result, autonomous and controlled motivation are viewed
54
as distal psychological variables influencing motivated behaviour at the contextual level. This
55
might explain why despite reporting significant associations between different forms of
56
motivation (i.e., autonomous vs. controlled) and sport persistence, empirical research has found
57
the amount of variance explained to be modest/moderate. On the other hand, sport commitment
58
is regarded as a psychological state which binds athletes to a specific commitment target (e.g.,
59
club, team, sporting activity). Therefore, enthusiastic and constrained commitment are
60
considered proximal psychological variables which have a direct, immediate influence on
61
motivated behaviour (Boiché & Sarrazin, 2009). Jackson, Gucciardi, Hodge, and Dimmock
62
(2017) summarised this distinction by asserting that an athlete may exhibit motivation “for”
63
one’s sport and commitment “to” one’s sport.
64
Within the SDT framework, several authors have contended that coaches’ interpersonal
65
style (“coaching style”) plays a particularly influential role in shaping athlete motivation and
66
subsequent behaviour. SDT research has primarily focused on the degree to which coaches
67
employ an autonomy-supportive or controlling coaching style – although coaches can adopt
68
both styles to varying degrees (e.g., Fenton, Duda, Quested, & Barrett, 2014; Haerens et al.,
69
2018; Ntoumanis, Barkoukis, Gucciardi, & Chan, 2017). An autonomy-supportive coaching
70
style characterises coaches who acknowledge their athletes’ thoughts and feelings, encourage
71
choice, self-initiation, regulation of one’s own behaviour, and minimise the use of pressure and
72
demands to control others (Gagné, Ryan, & Bargmann, 2003). In contrast, controlling coaches
73

COMMITMENT IN SPORT 4
may behave in a coercive, pressuring, and authoritarian manner to impose a specific and
74
preconceived way of thinking and behaving upon their athletes (Bartholomew et al., 2010).
75
SDT literature provides substantial evidence supporting the positive relationship between an
76
autonomy-supportive style and autonomous motivation (e.g., Fenton et al., 2014; Haerens et
77
al., 2018; Hodge & Lonsdale, 2011), and positive association between a controlling style and
78
controlled motivation (Fenton et al., 2014; Haerens et al., 2018; Pelletier et al., 2001).
79
Based on theoretical predictions from both the SCM and SDT, the following hypotheses
80
were tested: i) an autonomy-supportive coaching style would be positively associated with
81
autonomous motivation, which, in turn, would be positively associated with enthusiastic
82
commitment; ii) controlling coaching style would be positively associated with controlled
83
motivation, which, in turn, would be positively associated with constrained commitment; iii)
84
both enthusiastic and constrained commitment would be positively associated with intention to
85
continue in sport; iv) enthusiastic and constrained commitment would act as mediating
86
variables between autonomous and controlled motivation and intention to continue in sport.
87
These hypotheses were tested via the specification of two theoretically opposing path models:
88
a hypothesised model reflecting the aforementioned postulates (coaching style → motivation
89
→ commitment → intention), and an alternative model in which commitment precedes
90
motivation (coaching style → commitment → motivation → intention).
91
Methods
92
Participants
93
Participants were 154 football/soccer players (96 males, 52 females, six individuals did
94
not report gender), with an average age of 22.21 years (SD = 5.51; range = 17-51), who had
95
participated in football for an average of 15.6 years (range = 4-46 yrs, SD = 5.49). Athletes
96
were predominantly of New Zealand European (Caucasian) descent (n = 123, 79.87% of the
97

COMMITMENT IN SPORT 5
sample). Athletes had been competing under the guidance of their current coach for an average
98
1.8 seasons (range = 1-18 seasons, SD = 2.63).
99
Measures
100
Autonomy-supportive coaching style. An adapted version of the 14-item Health Care
101
Climate Questionnaire (HCCQ; Williams, Cox, Kouides, & Deci, 1999) was used to assess
102
athletes’ perceptions of autonomy-supportive strategies exhibited by their coach. Participants
103
responded using a seven-point Likert scale (1 = Strongly disagree; 7 = Strongly agree). Several
104
studies have supported the psychometric properties of adapted versions of the HCCQ in the
105
sport context (e.g., Hodge & Lonsdale, 2011; Ntoumanis et al., 2017).
106
Controlling coaching style. The 15-item Controlling Coach Behaviors Scale (CCBS;
107
Bartholomew et al., 2010) was employed to assess athletes’ perceptions of controlling
108
behaviours and strategies conveyed by their coach. Participants provided answers using a
109
seven-point Likert scale (1 = Strongly disagree; 7 = Strongly agree). The CCBS contains four
110
subscales measuring: controlling use of rewards, negative conditional regard, intimidation, and
111
excessive personal control. Past research provides robust evidence supporting the psychometric
112
properties of the CCBS and the calculation of an overall score (e.g., Bartholomew et al., 2010;
113
Ntoumanis et al., 2017).
114
Autonomous and controlled motivation. The 24-item Behavioral Regulation in Sport
115
Questionnaire-6 (BRSQ-6; Lonsdale, Hodge, & Rose, 2008) was used to assess autonomous
116
and controlled motivation at the contextual level. The BRSQ-6 includes subscales measuring
117
amotivation (AM), external regulation (EX), introjected regulation (IJ), identified regulation
118
(ID), integrated regulation (IG), and intrinsic motivation (IM). Participant responses were
119
recorded using a seven-point Likert scale (1 = Not at all true; 7 = Very true). Research has
120
demonstrated support for the psychometric properties of the BRSQ (e.g., Hodge & Lonsdale,
121

COMMITMENT IN SPORT 6
2011; Lonsdale et al., 2008). Scores for autonomous motivation were calculated using the
122
following formula: 2 × IM + IG + ID; while controlled motivation was calculated using: 2 × IJ
123
+ 2 × EX (Lonsdale et al., 2008).
124
Enthusiastic and constrained commitment. The Sport Commitment Questionnaire-2
125
(SCQ-2; Scanlan et al., 2016) was employed to measure athletes’ commitment to their current
126
sport participation. The 11-item SCQ-2 contains two subscales measuring the two dimensions
127
of sport commitment proposed in the SCM (Scanlan et al., 2013, 2016): enthusiastic
128
commitment and constrained commitment. Participants responded to each item using a five-
129
point Likert scale (1 = Strongly disagree; 5 = Strongly agree). Preliminary research has
130
documented support for the psychometric properties of the SCQ-2 (Scanlan et al., 2016).
131
Intention to continue. Two items measuring intention to continue in sport were adapted
132
from Gucciardi and Jackson (2015; e.g., “I intend to continue to participate in football next
133
season”, “Will you continue to participate in football next season?”). Responses were given by
134
participants using a seven-point Likert scale (1 = Extremely unlikely / Definitely plan not to; 7
135
= Extremely likely / Definitely plan to). Gucciardi and Jackson (2015) found these measurement
136
items to possess adequate psychometric properties.
137
Procedure
138
Teams were contacted and informed of the purpose of the study. With permission, the
139
first author was provided an opportunity to meet with athletes immediately before an agreed
140
upon training session. After informing athletes of the voluntary, anonymous nature of
141
participation and obtaining informed consent, the questionnaire was administrated to willing
142
participants in the absence of coaches. Data collection occurred from early to mid-season.
143
Ethical approval for this study was granted by the researchers’ university ethics committee.
144
Data Analysis
145

COMMITMENT IN SPORT 7
A hypothesised (coaching style → motivation → commitment → intention) and
146
alternative (coaching style → commitment → motivation → intention) path model was tested
147
using partial least squares structural equation modelling (PLS-SEM) in SmartPLS 3 (Ringle,
148
Wende, & Becker, 2015). The PLS-SEM approach follows a two-stage process: measurement
149
and structural model evaluation. The measurement model evaluation involved testing normality
150
of data distribution (univariate skewness and kurtosis), internal consistency reliability
151
(composite reliability and Cronbach’s alpha), convergent validity (average variance extracted),
152
and discriminant validity (indicator cross-loadings and heterotrait-monotrait ratio). The
153
structural model evaluation involved estimating multicollinearity (variance inflation factor),
154
path coefficients, variance, effect sizes (Cohen’s f 2 and q2), predictive relevance (Stone-
155
Geisser’s Q2), and absolute model fit (SRMR). Mediation hypotheses were tested by specifying
156
a combined effects model incorporating total, direct, and indirect effects. The significance of
157
these effects was estimated by constructing 95% bias-corrected and accelerated confidence
158
intervals based on a bootstrapping procedure with 5,000 resamples. To compare hypothesised
159
and alternative model predictions, we tested whether types of motivation mediated the
160
relationship between commitment dimensions and intention to continue. For more information
161
on statistical procedures used in this study, see Supplement 1.
162
Results
163
Descriptive statistics, internal reliability estimates, and bivariate correlations are
164
presented in Table 1. Preliminary analyses supported the reliability and validity of the
165
measurement model (see Supplement 2). In the hypothesised model (see Figure 1), expected
166
positive paths were significantly different from zero, while negative paths were not significant,
167
apart from the path linking autonomous motivation to constrained commitment (path estimate
168
= -.46, p < .001). Generally, the hypothesised model reported stronger relationships between
169
path model constructs compared to the alternative model (see Figure 1 & 2). Most notably,
170

COMMITMENT IN SPORT 8
enthusiastic commitment (path estimate = .53, p < .001) shared a stronger relationship with
171
intention to continue than autonomous motivation (path estimate = .34, p < .001). Although
172
both path models demonstrated acceptable/adequate fit in the PLS-SEM context (Hair, Hult,
173
Ringle, & Sarstedt, 2017), the hypothesised model (SRMR = .09) fit the data slightly better than
174
the alternative model (SRMR = .10). A central criterion for each model’s assessment was the
175
coefficient of determination (R2). Across models, we found a greater proportion of variance
176
explained in the outcome variable (i.e., intention to continue) from the hypothesised model (R2
177
= .38) as opposed to the alternative model (R2 = .20). Overall, these results informed our
178
decision to accept the hypothesised model over the alternative model.
179
As shown in Table 2, the path between autonomous motivation and intention to continue
180
displayed the only significant indirect effect of the two hypothesised model mediational
181
processes. An examination of the specific indirect pathways associated with this indirect effect
182
revealed that both the path from autonomous motivation → enthusiastic commitment (path
183
estimate = .61, p < .001, 54% of indirect effect) and from enthusiastic commitment → intention
184
to continue (path estimate = .53, p < .001, 46% of indirect effect) accounted for large
185
proportions of this mediated effect. These results provided substantial evidence supporting the
186
presence of complete mediation. The mediation results of the alternative model further
187
supported this inference with the indirect effect from enthusiastic commitment to intention to
188
continue via autonomous motivation being non-significant (see Table 2).
189
Discussion
190
Our findings provided empirical support that commitment and motivation are
191
conceptually distinguishable variables. This was demonstrated by three separate statistical
192
measures used in model evaluation. First, low VIF values revealed minimum multicollinearity
193
issues among path model constructs (see Supplement 2). Second, moderate-strong bivariate
194

COMMITMENT IN SPORT 9
correlations (.62 & .47; see Table 1) were found between both autonomous motivation and
195
enthusiastic commitment and controlled motivation and constrained commitment – these
196
correlations did not meet the critical threshold (r > .80) to indicate the presence of significant
197
conceptual overlap between constructs. Finally, low HTMT ratios were observed among types
198
of commitment and motivation when assessing construct-level discriminant validity (see
199
Supplement 2). Altogether these results substantiated our contention that autonomous and
200
controlled motivation are conceptually distinct from enthusiastic and constrained commitment.
201
Autonomy-Supportive Coaching, Autonomous Motivation, & Enthusiastic Commitment
202
Results revealed that an autonomy-supportive coaching style had a weak, positive
203
association with autonomous motivation (path estimate = .22), which, in turn, had a strong,
204
positive relationship with enthusiastic commitment (path estimate = .61). Taken collectively,
205
these results corroborate those of Weiss and Weiss (2003, 2006) who found athletes
206
demonstrating attraction-based (i.e., enthusiastic) commitment reported significantly higher
207
perceptions of coach support and intrinsic motivation – the core behavioural regulation of
208
autonomous motivation – than athletes displaying entrapment-based (i.e., constrained)
209
commitment. These findings provided full support for Hypothesis 1.
210
The strength of the relationship between autonomous motivation and enthusiastic
211
commitment may be explained with regard to the psychological process of internalisation which
212
is central to SDT. According to Ryan and Deci (2017), internalisation refers to the process
213
whereby individuals actively assimilate values, beliefs, or behavioural regulations from
214
external sources and transform them into their own. At higher levels of internalisation (i.e.,
215
autonomous motivation; enthusiastic commitment), behaviours are experienced as being
216
volitionally endorsed and self-regulated; whereas at lower levels of internalisation (i.e.,
217
controlled motivation; constrained commitment), behaviour is perceived as being primarily
218

COMMITMENT IN SPORT 10
externally controlled and regulated (Ryan & Deci, 2017). Organisational psychology research
219
has demonstrated support for this contention with scholars suggesting that an individual’s
220
innate internalisation process might act as the mechanism linking commitment and motivation
221
(Gagné, Chemolli, Forest, & Koestner, 2008; Meyer, Becker, & Vandenberghe, 2004). For
222
instance, Gagné et al. (2008, p. 237) postulated that “more internalised forms of commitment
223
may develop through internalisation of motivation”. In the sport context, motivation is assumed
224
to influence an athlete’s type of sport commitment through the extent to which the values and
225
behavioural regulations of an athlete’s sporting experience are internalised. It seems that
226
athletes exhibiting autonomous motivation (i.e., high internalisation) might experience a
227
willingness to engage in sport which potentially leads to perceptions of “wanting to” (i.e.,
228
enthusiastic commitment) continue sport participation.
229
Controlling Coaching, Controlled Motivation, & Constrained Commitment
230
Hypothesis 2 was fully supported. Results indicated coaches controlling style had a
231
moderate, positive relationship with athletes’ controlled motivation (path estimate = .42). This
232
finding was consistent with previous SDT research illustrating a positive association between
233
coaches controlling style and controlled motivation (Fenton et al., 2014; Haerens et al., 2018;
234
Pelletier et al., 2001). Results further revealed controlled motivation to have a moderate,
235
positive association with constrained commitment (path estimate = .38). This corroborated
236
Weiss and Weiss’ (2003, 2006) findings that athletes exhibiting entrapment-based (i.e.,
237
constrained) commitment reported higher levels of extrinsic motivation and amotivation than
238
athletes exhibiting attraction-based (i.e., enthusiastic) commitment. This finding further
239
substantiated the argument for internalisation in explaining the commitment-motivation
240
relationship. From an SDT perspective, controlling coaching behaviours and the resulting
241
maladaptive internalisation process means athletes perceived their behaviours as being
242
regulated by external factors (i.e., controlled motivation; constrained commitment). Extending
243

COMMITMENT IN SPORT 11
the psychological process of internalisation to explain the development of constrained
244
commitment, it was postulated that when exhibiting controlled motivation (i.e., low
245
internalisation), athletes might sense an obligation to participate in sport and thereby experience
246
perceptions of “having to” maintain their involvement in sport (i.e., constrained commitment).
247
Commitment Dimensions and Intention to Continue in Sport
248
The hypothesis linking commitment dimensions with athletes’ intention to continue in
249
sport was partially supported. Results showed enthusiastic commitment to have a strong,
250
positive association with intention to continue in sport (path estimate = .53). This finding
251
provided support for a positive association between enthusiastic commitment and self-reported
252
behavioural persistence and corroborated initial empirical support for the positive association
253
between an athlete’s initial intention and subsequent behavioural persistence in sport (path
254
estimate = .50; Gucciardi & Jackson, 2015). Across path models, enthusiastic commitment
255
(path estimate = .53) shared a stronger relationship with intention to continue than autonomous
256
motivation (path estimate = .34). This offered additional support for enthusiastic commitment’s
257
position as a proximal variable in relation to sport persistence (Boiché & Sarrazin, 2009).
258
Contrary to SCM premises, constrained commitment was found to have a weak,
259
negative association with intention to continue in sport (path estimate = -.16). Despite this
260
association only approaching significance (p = .08), our finding was consistent with Weiss and
261
Weiss’ (2006) longitudinal study which found that a greater percentage of athletes who reported
262
“having to” commit to sport (i.e., constrained commitment) one year earlier, had discontinued
263
their participation compared to athletes who reported “wanting to” commit to sport (i.e.,
264
enthusiastic commitment). In addition, this finding supported research in the exercise domain
265
which has found only the “want to” dimension of commitment to be associated with positive
266
exercise behaviour (e.g., Wilson et al., 2004). As noted earlier, it is reasonable to expect
267

COMMITMENT IN SPORT 12
differential outcomes to result from the two commitment dimensions (Wilson et al., 2004). It
268
could be argued that enthusiastic commitment might lead to longer-term intentions to continue,
269
whereas constrained commitment may be linked with shorter-term intentions to continue.
270
Alternatively, athletes who experience constrained commitment might feel stronger intentions
271
toward dropout and conversely weaker intentions to continue (Guillet et al., 2002).
272
Enthusiastic and Constrained Commitment: Psychological Mediators?
273
This was the first study to examine the hypothesis that sport commitment may act as a
274
psychological mechanism mediating the relationship between motivation and intention to
275
continue in sport. Overall, this fourth hypothesis was partially supported. We hypothesised that
276
enthusiastic commitment would mediate the relationship between autonomous motivation and
277
intention to continue in sport. Findings provided considerable support for this part of
278
Hypothesis 4 (see Table 2), empirically substantiating the role of enthusiastic commitment as
279
completely mediating the relationship between autonomous motivation and intention to
280
continue. This mediation process was further supported by findings that autonomous motivation
281
did not mediate the relationship between enthusiastic commitment and intention to continue.
282
An important implication of this finding is that autonomous motivation may not necessarily
283
lead to sport persistence under circumstances when an athlete fails to express a firm desire to
284
remain in their given sport (i.e., enthusiastic commitment). Constrained commitment was also
285
examined as a potential mediating variable between controlled motivation and intention to
286
continue. This relationship was rejected with the indirect effect from controlled motivation to
287
intention to continue in sport via constrained commitment being non-significant (p > .05).
288
Strengths, Limitations, & Future Research Directions
289
This study extended previous research examining sport persistence by assessing both
290
commitment dimensions and integrating them with autonomous and controlled motivation.
291

COMMITMENT IN SPORT 13
Nevertheless, this study was not without limitations. First, this study used cross-sectional, self-
292
report data, preventing causal relationships being inferred from these findings. Second, this
293
study was limited by focusing on athletes’ intentions rather than actual behaviour. Third, the
294
study sample was recruited primarily from a young adult population. Finally, this study was
295
limited insofar as it only examined one team sport (i.e., football/soccer).
296
A primary applied implication of the current study was the apparent utility of SDT
297
concepts in helping promote an optimal commitment state (i.e., enthusiastic commitment) for
298
athletes’ sport persistence. Based on our findings, coaches should consider adopting autonomy-
299
supportive strategies to foster the development of autonomous motivation and, in turn,
300
enthusiastic commitment. These adaptive motivational conditions are expected to facilitate
301
athletes’ long-term engagement in sport. There is also reason to believe that the provision of
302
overall need-support (i.e., autonomy-support, competence-support, relatedness-support) might
303
further enhance athletes’ enthusiastic commitment by tapping into additional sources of this
304
commitment dimension (i.e., desire to excel, sport enjoyment, valuable opportunities). For
305
example, it is reasonable to expect that competence-supportive coaching behaviours and
306
strategies would strengthen athletes’ desire to excel in sport. In our ongoing effort to encourage
307
athletes’ persistence in sport, it is also important to ensure we do not selectively target the
308
development of autonomous motivation or enthusiastic commitment by itself. Each construct
309
seems to play a crucial role in underpinning this participation behaviour and thus warrants equal
310
attention by practitioners. Future research employing longitudinal designs is necessary when
311
examining these constructs in order to directly examine the behavioural consequence of sport
312
persistence over time. This would be particularly useful in determining the temporal
313
relationship between commitment dimensions (i.e., enthusiastic vs. constrained) and
314
behavioural persistence in sport. As noted earlier, future research might hypothesise constrained
315
commitment to predict short-term sport persistence, and enthusiastic commitment to predict
316

COMMITMENT IN SPORT 14
long-term sport persistence. Future research should also consider experimentally examining the
317
efficacy of SDT-based interventions aimed at promoting sport commitment.
318
This study provided novel insights into the links between commitment, motivation, and
319
athletes’ intention to continue in sport. In describing the correlates of this continuance intention,
320
motivation was viewed as a distal psychological variable forming the underlying basis of
321
athletes’ intention to continue; commitment was considered a proximal psychological variable
322
exerting a direct, immediate influence on intentions to continue. Our findings demonstrated
323
support for this contention, highlighting the role of commitment as a potential psychological
324
mechanism mediating the relationship between motivation and intention to continue in sport.
325
References
326
Bartholomew, K. J., Ntoumanis, N., & Thøgersen-Ntoumani, C. (2010). The controlling
327
interpersonal style in a coaching context: Development and initial validation of a
328
psychometric scale. Journal of Sport & Exercise Psychology, 32, 193–216.
329
Boiché, J. C. S., & Sarrazin, P. G. (2009). Proximal and distal factors associated with dropout
330
versus maintained participation in organized sport. Journal of Sports Science &
331
Medicine, 8, 9-16.
332
Casper, J. M., Gray, D. P., & Stellino, M. B. (2007). A Sport Commitment Model perspective
333
on adult tennis players’ participation frequency and purchase intention. Sport
334
Management Review, 10, 253-278. doi: 10.1016/S1441-3523(07)70014-1
335
Fenton, S. A. M., Duda, J. L., Quested, E., & Barrett, T. (2014). Coach autonomy support
336
predicts autonomous motivation and daily moderate-to-vigorous physical activity and
337
sedentary time in youth sport participants. Psychology of Sport & Exercise, 15, 453-
338
463. doi: 10.1016/j.psychsport.2014.04.005
339

COMMITMENT IN SPORT 15
Gagné, M., Chemolli, E., Forest, J., & Koestner, R. (2008). A temporal analysis of the relation
340
between organisational commitment and work motivation. Psychologica Belgica, 48,
341
219-241. doi: 10.5334/pb-48-2-3-219
342
Gagné, M., Ryan, R. M., & Bargmann, K. (2003). Autonomy support and need satisfaction in
343
the motivation and well-being of gymnasts. Journal of Applied Sport Psychology, 15,
344
372-390. doi: 10.1080/10413200390238031
345
Garcia-Mas, A., Palou, P., Gili, M., Ponseti, X., Borras, P. A., Vidal, J., Cruz, J., Torregrosa,
346
M., Villamarín, F., & Sousa, C. (2010). Commitment, enjoyment and motivation in
347
young soccer competitive players. The Spanish Journal of Psychology, 13, 609-616.
348
Gucciardi, D. F., & Jackson, B. (2015). Understanding sport continuation: An integration of the
349
theories of planned behaviour and basic psychological needs. Journal of Science &
350
Medicine in Sport, 18, 31-36. doi: 10.1016/j.jsams.2013.11.011
351
Guillet, E., Sarrazin, P., Carpenter, P. J., Trouilloud, D., & Cury, F. (2002). Predicting
352
persistence or withdrawal in female handballers with social exchange theory.
353
International Journal of Psychology, 37, 92-104. doi: 10.1080/00207590143000243
354
Haerens, L., Vansteenkiste, M., De Meester, A., Delrue, J., Tallir, I., Vande Broek, G., Goris,
355
W., Aelterman, N. (2018). Different combinations of perceived autonomy support and
356
control: Identifying the most optimal motivating style. Physical Education & Sport
357
Pedagogy, 23, 16-36. doi: 10.1080/17408989.2017.1346070
358
Hair, J. F., Hult, G. T. M., Ringle, C. M., & Sarstedt, M. (2017). A primer on partial least
359
squares structural equation modeling (PLS-SEM) (2nd ed.). Thousand Oaks, CA: Sage.
360
Hodge, K., & Lonsdale, C. (2011). Prosocial and antisocial behavior in sport: The role of
361
coaching style, autonomous vs. controlled motivation, and moral disengagement.
362
Journal of Sport & Exercise Psychology, 33, 527–547. doi: 10.1123/jsep.33.4.527
363

COMMITMENT IN SPORT 16
Jackson, B., Gucciardi, D. F., Hodge, K., & Dimmock, J. A. (2017). Commitment in sport and
364
exercise: Implications for individual, group, and organizational functioning. In C. R. D.
365
Wagstaff (Ed.), The organizational psychology of sport: Key issues and practical
366
applications (pp. 11-32). Abingdon, UK: Routledge.
367
Lonsdale, C., Hodge, K., & Rose, E. A. (2008). The development of the Behavioral Regulation
368
in Sport Questionnaire (BRSQ): Instrument development and initial validity evidence.
369
Journal of Sport & Exercise Psychology, 30, 323–355. doi: 10.1123/jsep.30.3.323
370
Meyer, J. P., Becker, T. E., & Vandenberghe, C. (2004). Employee commitment and
371
motivation: A conceptual analysis and integrative model. Journal of Applied
372
Psychology, 89, 991-1007. doi: 10.1037/0021-9010.89.6.991
373
Ntoumanis, N., Barkoukis, V., Gucciardi, D. F., & Chan, D. K. C. (2017). Linking coach
374
interpersonal style with athlete doping intentions and doping use: A prospective study.
375
Journal of Sport & Exercise Psychology, 39, 188-198. doi: 10.1123/jsep.2016-0243
376
Pelletier, L. G., Fortier, M. S., Vallerand, R. J., & Brière, N. M. (2001). Associations among
377
perceived autonomy support, forms of self-regulation, and persistence: A prospective
378
study. Motivation & Emotion, 25, 279–306. doi: 10.1023/a:1014805132406
379
Ringle, C. M., Wende, S., & Becker, J. -M. (2015). SmartPLS 3 [Computer software]. Retrieved
380
from http://smartpls.com
381
Ryan, R. M., & Deci, E. L. (2017). Self-Determination Theory: Basic psychological needs in
382
motivation, development, and wellness. New York, NY: Guilford Press.
383
Scanlan, T. K., Chow, G. M., Sousa, C., Scanlan, L. A., & Knifsend, C. A. (2016). The
384
development of the Sport Commitment Questionnaire-2 (English version). Psychology
385
of Sport & Exercise, 22, 233-246. doi: 10.1016/j.psychsport.2015.08.002
386
Scanlan, T. K., Russell, D. G., Scanlan, L. A., Klunchoo, T. J., & Chow, G. M. (2013). Project
387
on Elite Athlete Commitment (PEAK): IV. Identification of new candidate commitment
388

COMMITMENT IN SPORT 17
sources in the Sport Commitment Model. Journal of Sport & Exercise Psychology, 35,
389
525-535. doi: 10.1123/jsep.35.5.525
390
Weiss, W. M., & Weiss, M. R. (2003). Attraction- and entrapment-based commitment among
391
competitive female gymnasts. Journal of Sport & Exercise Psychology, 25, 229-247.
392
Weiss, W. M., & Weiss, M. R. (2006). A longitudinal analysis of commitment among
393
competitive female gymnasts. Psychology of Sport & Exercise, 7, 309-323.
394
Williams, G. C., Cox, E. M., Kouides, R., & Deci, E. L. (1999). Presenting the facts about
395
smoking to adolescents: Effects of an autonomy-supportive style. Archives of Pediatrics
396
& Adolescent Medicine, 153, 959-964. doi: 10.1001/archpedi.153.9.959
397
Wilson, P. M., Rodgers, W. M., Carpenter, P. J., Hall, C., Hardy, J., & Fraser, S. N. (2004).
398
The relationship between commitment and exercise behavior. Psychology of Sport &
399
Exercise, 5, 405-421. doi: 10.1016/S1469-0292(03)00035-9
400

COMMITMENT IN SPORT 18
Table 1. Descriptive Statistics, Internal Reliability Estimates, and Bivariate Correlations Among Coaching Style, Motivation, Commitment, and
Intention to Continue.
Variable
M
SD
1
2
3
4
5
6
7
1 Autonomy-Supportive Coaching Style
5.40
1.04
(.96)
2 Controlling Coaching Style
2.26
.93
-.36**
(.91)
3 Autonomous Motivation
23.69
4.44
.25**
-.14
(.84)
4 Controlled Motivation
9.84
6.62
-.13
.41***
-.18*
(.92)
5 Enthusiastic Commitment
4.28
.67
.11
-.13
.62***
-.16
(.92)
6 Constrained Commitment
1.79
.75
-.15
.25**
-.53***
.47***
-.39**
(.86)
7 Intention to Continue
6.34
1.26
.12
-.09
.42***
-.02
.60***
-.32**
(.99)
Note: Internal reliability estimates (composite reliabilities) provided on the diagonal in parentheses.
* p < .05, ** p < .01, *** p < .001.

COMMITMENT IN SPORT 19
Table 2. Examining Mediation: Total, Direct, and Indirect Effects.
Relationship
Total
Effect
Direct
Effect
Indirect
Effect
Hypothesised Model
Autonomous Motivation → Intention to Continue
(via Enthusiastic Commitment)
.43***
.03
.40***
Controlled Motivation → Intention to Continue
(via Constrained Commitment)
.07
.15
-.08
Alternative Model
Enthusiastic Commitment → Intention to Continue
(via Autonomous Motivation)
.55***
.53***
.02
Constrained Commitment → Intention to Continue
(via Controlled Motivation)
-.10
-.16
.06
Note: *** p < .001.

COMMITMENT IN SPORT 20
Figure 1. Hypothesised model of coaching style, motivation, commitment, and intention to continue.
Note: * p < .05, *** p < .001
.38***
.53***
R2 = .38
-.16
.61***
.42***
R2 = .39
-.06
.22*
.03
-.05
-.46***
R2 = .42
R2 = .17
R2 = .06
-.39***

COMMITMENT IN SPORT 21
Figure 2. Alternative model of coaching style, motivation, commitment, and intention to continue.
Note: * p < .05, *** p < .001
R2 = .48
.48***
.34***
R2 = .20
.15
.49***
.22*
-.11
.06
-.07
-.34***
.02
R2 = .23
R2 = .07
R2 = .02
-.39***