You are on page 1of 14

Journal of Sports Sciences

ISSN: (Print) (Online) Journal homepage: www.tandfonline.com/journals/rjsp20

The role of quality relationships and


communication strategies for the fulfilment of
secure and insecure athletes’ basic psychological
needs

Louise Davis, Sophia Jowett, Daniel Sörman & Rebecka Ekelund

To cite this article: Louise Davis, Sophia Jowett, Daniel Sörman & Rebecka Ekelund (2022) The
role of quality relationships and communication strategies for the fulfilment of secure and
insecure athletes’ basic psychological needs, Journal of Sports Sciences, 40:21, 2424-2436, DOI:
10.1080/02640414.2022.2162240

To link to this article: https://doi.org/10.1080/02640414.2022.2162240

© 2023 The Author(s). Published by Informa


UK Limited, trading as Taylor & Francis
Group.

Published online: 08 Jan 2023.

Submit your article to this journal

Article views: 3982

View related articles

View Crossmark data

Citing articles: 1 View citing articles

Full Terms & Conditions of access and use can be found at


https://www.tandfonline.com/action/journalInformation?journalCode=rjsp20
JOURNAL OF SPORTS SCIENCES
2022, VOL. 40, NO. 21, 2424–2436
https://doi.org/10.1080/02640414.2022.2162240

RESEARCH ARTICLE

The role of quality relationships and communication strategies for the fulfilment of
secure and insecure athletes’ basic psychological needs
Louise Davisa, Sophia Jowettb, Daniel Sörmanc and Rebecka Ekelunda
a
Department of Psychology, Umeå University and Umeå School of Sports Science, Umeå, Sweden; bSchool of Sport, Exercise and Health Sciences,
Loughborough University, Loughborough, UK; cDepartment of Health, Education and Technology, Luleå University of Technology, Luleå, Sweden

ABSTRACT ARTICLE HISTORY


The correlates of coach-athlete relationship quality have been the focus of research for over a decade; Received 17 September 2021
however, little is known about the mediating and moderating mechanisms underlying these associations. Revised 9 December 2022
The present study conducted a moderated mediation analysis to examine (a) the mediating role of Accepted 19 December 2022
communication strategies (via COMPASS) on the association between the quality of the coach-athlete KEYWORDS
relationship and athlete psychological needs satisfaction and (b) whether individual differences in Attachment; relationship
athletes’ attachment style (secure, anxious, avoidant) moderates the mediational relationship. 350 quality; needs satisfaction;
Swedish athletes representing a range of sports and competition levels completed a multi-section coaching
questionnaire. Mediation and moderation analysis partially found that coach-athlete relationship quality
and athletes basic psychological needs were associated via the COMPASS strategies of support, motiva­
tion, assurance and openness. It was also found that athletes secure attachment with their coach
significantly moderated the mediated effects of motivation and support. These findings highlight the
practical utility of motivation, support, openness and assurance strategies in enhancing the quality of the
coach-athlete relationship. Moreover, these findings demonstrate that the attachment orientation of
athletes towards their coaches play a significant role in determining what communication strategies to
use to enhance both the relationship quality and an athlete’s competence, autonomy and relatedness.

The relationship formed with a coach is central to an athlete’s employed to develop and maintain high quality coach-athlete
sporting experience (Davis et al., 2019b). It is not uncommon relationships.
for athletes to acknowledge the significance of their coach The quality of the coach–athlete relationship is defined and
following major performance success in competitive events, operationalised as a social environment, whereby athletes’
during post competition interviews, via social media (e.g., and coaches’ feelings, thoughts, and behaviours are mutually
Twitter), or in the narratives found in their autobiographies. and causally inter-connected (Jowett, 2007). These feelings,
This was recently demonstrated by Lauren Steadman upon thoughts, and behaviours are reflected in Jowett’s (2007),
winning Triathlon gold in the Tokyo 2020 Paralympics, “words 3 + 1Cs framework, shaped by the interpersonal dimensions
cannot describe how special it was to have my coach by my of closeness (emotional connection manifested in feelings of
side yesterday . . . he believed in me, encouraged me, pushed trust, like, respect, appreciation), commitment (thoughts to
me, guided me, and above all he’s always been there for me” maintain a close tie over time), and complementarity (beha­
(Steadman, 2021). Although Steadman struggled to articulate vioural interactions that are responsive, receptive, comforta­
the unique connection with her coach, it is clear that the quality ble, and friendly). The 3Cs signify the extent to which the
of the coach-athlete relationship is a key determinant to health coach and the athlete perceive the relationship to be of high
and performance at the highest levels of sport competition. quality. Co-orientation represents the +1C in the model and
Extant research outlines that good quality relationships are provides an insight into the degree to which the coach and
associated with positive intrapersonal (e.g., wellbeing; Davis & the athlete perceive the relationship quality to be (dis) similar
Jowett, 2014), interpersonal (e.g., performance; Hampson & (Lorimer & Jowett, 2012). Co-orientation acknowledges the
Jowett, 2014), and developmental outcomes (e.g., Vella et al., relationship from two different perspectives: (a) the direct
2013). From the perspective of athletes, harmonious and stable perspective reflects coaches’ and athletes’ own perceptions
coach-athlete relationships are important as they provide regarding the 3Cs (e.g., “I trust my coach/athlete”); and (b), the
security during turbulence in sport and/or life (Li et al., 2020; meta perspective reflects coaches’ and athletes’ perceptions
Wachsmuth et al., 2018). Although coach-athlete relationships of how the other perceives the 3Cs (e.g., “My coach/athlete
have garnered increasing research attention, studies have lar­ trusts me”).
gely focused on characteristics that constitute high quality In an attempt to accelerate and map a pathway for focused
relationships and its correlates (e.g., Jowett & Shanmugam, research surrounding the quality of the coach-athlete relation­
2016). The present study aims to advance knowledge by focus­ ship, Jowett and Poczwardowski (2007) put forward an inte­
ing on the strategies, processes, and methods that can be grated research model (IRM) illustrating how the quality of the

CONTACT Louise Davis louise.davis@umu.se Department of Psychology, Umeå University, Umeå, Sweden
© 2023 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://creativecommons.org/licenses/by-nc-nd/4.0/),
which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed, or built upon in any way.
JOURNAL OF SPORTS SCIENCES 2425

coach-athlete relationship (3Cs) represents various social phe­ coach-athlete relationships; conversely, absence of these stra­
nomena (i.e., communication, personality, motivation) that can tegies can undermine relationship quality (as defined by
either determine its quality or be determined by the quality of the 3Cs).
the relationship. Accordingly, the model highlights three inter­ Although research examining coach-athlete relationship
related layers outlining coach-athlete relationship quality. The quality is extensive, limited research has focused on the strate­
top layer identifies antecedent variables with the potential to gies coaches’ and athletes’ use to maintain relationship quality
determine the quality of the coach-athlete relationship (second and subsequent outcomes. Early research conducted by Rhind
layer). These antecedent variables include individual differ­ and Jowett (2011, 2012) noted that higher levels of closeness
ences (e.g., personality, age, gender), the social-cultural context within the coach-athlete relationship were associated with
(e.g., culturally defined norms, values), as well as relational COMPASS strategies that promote open channels of commu­
characteristics (e.g., duration, type). The middle/second layer nication and engagement with wider social networks including
of the model describes the quality of the coach-athlete relation­ peers, support staff, and/or parents. Further, higher levels of
ship itself (e.g., 3Cs), whilst the bottom layer includes outcomes commitment were associated with greater use of motivational,
or consequences of relationship quality. These outcomes support, and assurance strategies; whilst complementarity was
include intrapersonal (e.g., well-being, motivation, perfor­ associated with greater engagement in the use of conflict
mance), interpersonal (e.g., relationship satisfaction, interper­ management and preventative strategies. More recently,
sonal conflict), and group (e.g., team cohesion, collective Davis et al. (2019a) observed that COMPASS strategies can act
efficacy) factors. The three layers are interrelated, whereby as mechanisms that transfer the effects of relationship quality
antecedent variables (top layer) such as athletes’ gender, per­ onto dimensions of sport satisfaction. Specifically, support and
sonality, and experience may affect the quality of the relation­ motivational strategies transferred the effects of relationship
ship (middle layer) and in return the quality of the coach- quality onto athletes’ satisfaction with personal treatment,
athlete relationship affects athlete outcomes (third layer). training and instruction, and performance; these strategies
To date, both antecedent and outcomes variables of the were most instrumental when the quality of the coach-athlete
coach-athlete relationship quality defined by the 3 + 1Cs have relationship was perceived to be poor.
been widely examined. For example, the quality of the coach- Preliminary research recommends further scrutiny of
athlete relationship has been found to correlate with antece­ COMPASS strategies to evaluate their role as mechanisms in
dents of personality (Yang et al., 2015), individual differences in influencing relationship quality and associated outcomes
attachment (Davis et al., 2013), gender and age (Davis et al., (Davis et al., 2019a). Therefore, the focus of the present study
2019b; Jowett & Nezlek, 2012), as well as outcomes including was on COMPASS as a medium that transfers the effects of the
cognitive performance (Davis et al., 2018), flourishing and thriv­ coach-athlete relationship quality onto athletes’ basic psycho­
ing (Gosai et al., 2021), motivation (Felton & Jowett, 2013b), logical needs satisfaction (i.e., autonomy, competence, and
collective efficacy (Hampson & Jowett, 2014), psychological relatedness). Previous studies have identified associations
needs satisfaction (Felton & Jowett, 2013a), interpersonal con­ between quality relationships (i.e., middle/second layer of
flict (Wachsmuth et al., 2018), and athlete burnout (Davis et al., IRM) and athletes’ psychological needs satisfaction (i.e., third
2018, 2019b). layer of IRM; Felton & Jowett, 2013b; Riley & Smith, 2011), this
The integrated research model highlights that the main study explores whether COMPASS (i.e., auxiliary layer of com­
middle layer of relationship quality is further sandwiched munication of the IRM) facilitates these associations.
between two auxiliary layers of interpersonal communication Furthermore, individual difference characteristics such as per­
(see, Jowett & Poczwardowski, 2007). Communication (i.e., ver­ sonality (Jackson et al., 2011) and attachment styles (Davis et al.,
bal or nonverbal, intended or unintended, effective or ineffec­ 2013) are prevalent antecedents of coach-athlete relationship
tive), is the process by which the relationship is formed and quality and athletes’ psychological need satisfaction, therefore,
maintained (Sullivan et al., 2014). Subsequently, an effective this study seeks to explore whether attachment orientations
high-quality relationship requires the maintenance of effective (i.e., first layer of IRM) moderates the proposed mediation.
high-quality communication. Rhind and Jowett (2010) propose Attachment styles are considered personal tendencies
that coaches and athletes employ seven key communication related to athletes’ and coaches’ ways of relating, interacting,
strategies to maintain an optimal relationship. These seven and communicating (Davis & Jowett, 2014). There are three
strategies presented within the COMPASS model include con­ distinct attachment styles: secure, anxious-ambivalent, and
flict management (e.g., efforts to identify, discussing, resolve, avoidant (Ainsworth et al., 1978). A secure attachment charac­
and monitor potential areas of conflict and disagreement), terises individuals who display confidence in the availability of
openness (e.g., efforts to engage in open lines of communica­ significant others (e.g., parents, peers, coach) and feel comfor­
tion), motivation (e.g., providing reasons for the other partner table with relational closeness. In contrast, an anxious ambiva­
to stay in the relationship), preventative (e.g., efforts to discuss lent attachment style characterises individuals who display
expectations and what to do if these are not met), assurance a lack of confidence in the availability of significant others
(e.g., efforts to demonstrate commitment to the relationship), (i.e., fear of rejection), yet a strong need for relational closeness.
support (e.g., efforts to help each other through difficult and Finally, an avoidant attachment characterises individuals who
challenging times), and social networks (e.g., create opportu­ also display a lack of confidence in the availability of significant
nities to develop strong bonds with significant others that others and avoid any form of relational closeness. Extensive
prevent the coach and athlete operating in a “bubble”). Use research indicates attachment styles offer a means of explain­
of these strategies support the maintenance of high-quality ing relationship behaviours across the lifespan with diverse
2426 L. DAVIS ET AL.

significant relationship partners (e.g., romantics partners, tea­ allows for further understanding of the mediating effect by
chers, close friends, sports coaches) that provide protection and determining under what conditions or variables the mediation
support when needed (Bergin & Bergin, 2009; Collins & Read, may be significant (Preacher et al., 2007). The proposed mod­
1990; Davis & Jowett, 2014; Mayseless, 2010). erated mediational model is illustrated in Figure 1 and the
Research focusing on attachment styles between coaches formulated hypotheses are as follow:
and athletes has revolved around investigating the links
between coach-athlete attachment and the quality of the H1. The association between the quality of the coach-athlete
coach-athlete relationship (Davis et al., 2013a), athletes’ affec­ relationship (direct and meta perspectives) and athletes’ per­
tive well-being (Davis & Jowett, 2014; Felton & Jowett, 2013b), ceived psychological needs (perceived autonomy, competence
sport satisfaction (Davis & Jowett, 2010), eating psychopathol­ and relatedness) will be mediated by each of the seven
ogy (Shanmugam et al., 2014), and athletic thriving (Davis et al., COMPASS communication strategies.
2021). In particular, athletes who reported a secure attachment
towards their sports coach evaluated the quality of their coach- H2. The association between the quality of the coach-athlete
athlete relationship (3Cs) more positively, whilst avoidant ath­ relationship and athlete psychological needs via COMPASS
letes tended to evaluate their relationship more negatively strategies will be positively moderated by athletes’ secure
(Davis et al., 2013). In relation to psychological needs satisfac­ attachment towards their coach and inversely moderated by
tion, athletes reporting high levels of avoidant attachment can athletes’ insecure attachment (anxious and avoidance).
also experience significant reductions in psychological need
satisfaction (Felton & Jowett, 2017). Although, athletes with
an avoidant attachment style indicate they can function opti­
mally if they perceive that their basic psychological needs were Methods
satisfied (Felton & Jowett, 2013b). Taken collectively, research
suggests athletes’ attachment styles to their coach may offer Participants
insight into the associations between relationship quality, com­ Participants were 350 Swedish athletes; 159 (45.4%) males, 190
munication, and psychological needs satisfaction. (54.3%) females, and 1 (0.3%) categorized as other. The mean
age of the sample was 19.18 years (SD = 3.98 years, range = 15–
45 years) and athletes ‘were recruited from a range of sports
The Present Study including both individual sports (46.9%; e.g., swimming, cross-
As such, the present study is guided by the integrated research country skiing, golf, athletics) team sports (52.3%;(e.g., ice-
model (Jowett & Poczwardowski, 2007) with the aim to under­ hockey, handball, football, floorball, volleyball) and unspecified
take: (a) an analysis of the mediating role of communication (0.8%). Participants competed in their sport at various levels
strategies (COMPASS) on the association between coach- including regional (22.6%), national (49.1%) and international
athlete relationship quality and athlete psychological needs level (26.3%), and 2% had reported “other”. Participants
satisfaction (i.e., competence, autonomy, relatedness); and (b) reported they had been involved in their sport for a mean
an examination of whether athletes’ attachment style moder­ length of 11.04 years (SD = 5.46) and had an average relation­
ates the mediational influence of communication strategies on ship length with their principal sports coach of 2.59 years
the association between coach-athlete relationship quality and (SD = 2.65), with a mean of 8.57 hours spent with their coach
athlete psychological needs satisfaction. Moderated mediation in training each week (SD = 6.32).

Figure 1. The hypothetical model.


JOURNAL OF SPORTS SCIENCES 2427

Measures consistency (Rhind & Jowett, 2010). For the present study, the
items were translated into Swedish using a parallel back trans­
Coach-athlete relationship quality
lation process. Cronbach’s alpha was measured for each strat­
The Coach-Athlete Relationship Questionnaire (CART-Q; Jowett,
egy; Conflict management (α = .81), Openness (α = .80)
2009; Jowett & Ntoumanis, 2004) was used to measure athletes
motivational, (α = .92), preventative (α = .84), assurance
direct and meta-perceptions of the quality of the coach-athlete
(α = .78), support (α = .89), and social networks (α = .79).
relationship. The 11-item direct perspective and the corre­
Similarly, to the CART-Q, the factorial validity of the items was
sponding 11-item meta-perspective of the CART
examined. Fit indices from the CFA of the Swedish version of
Q represented closeness (4 items, e.g., I trust my coach; 4
the 28-item, 7 factor model of the CARM-Q was adequate with
items, e.g., My coach trusts me), commitment (3 items, e.g.,
regard to all fit indices (χ2/df = 2.70, p < .001; CFI = .91; TLI = .90,
I am committed to my coach; 3 items, e.g., My coach is com­
RMSEA = .07, 90% CI = .06, .08). All items loaded significantly
mitted to me), and complementarity (4 items, e.g., When I am
(for all p < .001) on its latent variables (i.e., conflict manage­
coached by my coach, I am responsive to his/her efforts; 4 items
ment, openness, motivational, preventative, assurance, sup­
e.g., My coach is responsive to my efforts when he/she coaches
port, and social networks).
me). Responses were rated on a 7-point scale (1 = Strongly
Disagree, 7 = Strongly Agree). Previous studies have displayed
Coach-athlete attachment
sound psychometric properties of the CART-Q (see, Jowett &
The Coach–Athlete Attachment Scale (Davis & Jowett, 2013)
Ntoumanis, 2004). As in previous research (e.g., Davis & Jowett,
was used to measure athletes’ attachment to their sports coach
2013; Davis et al., 2019b), a composite approach was imple­
across three orthogonal styles. The CASS contains 19 items,
mented for direct and meta relationship quality. For the present
designed to measure athletes’ secure (5 items, e.g., I know
study, the items were translated into Swedish using a parallel
I can rely on my coach”) anxious (7 items, e.g., I worry that
back translation process. Cronbach’s alpha was measured for
I won’t fulfil my coaches’ expectations) and avoidant (7 items,
the two perspectives: direct perspective (α = .94) and meta
e.g., I do not turn to my coach for reassurance) attachment
perspective (α = .95) respectively. Additionally, the factorial
styles. Responses were rated on a 7-point scale (1 = strongly
validity of the items of the CART-Q were examined, since the
disagree, 7 = strongly agree) and relevant to how they feel
instrument has not been statistically scrutinized in the Swedish
towards their principal sports coach. Davis and Jowett (2013,
language. To test the validity of the items with this sample,
2014) have provided sound psychometric properties of validity
a second order CFA model was performed, where the 3Cs were
and reliability. For the present study, the items were translated
subsumed under a general category namely coach-athlete rela­
into Swedish using a parallel back translation process.
tionship quality. Analysis showed good fit for the direct per­
Cronbach’s alpha was measured for each attachment style;
spective: χ2/df = 3.52, p < .001; CFI = .97; TLI = 96, RMSEA = .08,
Secure (α = .89), anxiety (α = .89) and avoidant, (α = .88). The
90% CI = .07, .10. All items loaded significantly on its latent
factorial validity of the items was examined. Fit indices from the
construct (p < .001) following the addition of correlations
CFA of the 19 item, 3-factor model were adequate (χ2/df = 2.67,
between error terms within factors; closeness (“I like my
p < .001; CFI = .94; TLI = .93, RMSEA = .07, 90% CI = .06, .08) after
coach” and “I appreciate my coach’s sacrifices in order to
two correlations between error terms within factors were
improve performance”) and complementarity (When I am coa­
added; anxious (“I often worry that my coach does not value
ched by my coach: “I am responsive to his/her efforts”, “I am
me as much as I value him/her” and “ I worry a fair amount that
ready to do my best”, and “I adopt a friendly stance”). For the
my coach will leave me to train elsewhere”) and secure (“ I know
meta perspective, analysis also showed good model fit for CFI
that my coach is loyal to me” and “I find it easy to interact with
(.95), and adequate with regard to TLI (.93). However, fit indices
my coach”). All items loaded significantly (for all p < .001) on its
for RMSEA (.11, CI = .10, .13) and χ2/df = 5.51, p = .000 was still
latent variables (i.e., secure, anxious, and avoidant).
poor.
Basic psychological need satisfaction
Coach-athlete relationship maintenance strategies The Basic Need Satisfaction in Sport Scale (BNSSS; Ng et al.,
The Coach-Athlete Relationship Maintenance Questionnaire 2011) was used to measure athletes’ basic psychological needs
(CARM-Q; Rhind & Jowett, 2012) was used to assess the strate­ satisfaction. Specifically, 10 items measured athlete’s autonomy
gies athletes use to communicate with their coach. The 28 item satisfaction (e.g., in my sport, I get opportunities to make
questionnaire measures seven COMPASS communication stra­ choices), five items measured competence satisfaction (e.g.,
tegies including: conflict management (5 items, e.g., I am I am skilled at my sport) and five items measured relatedness
understanding during disagreements), openness (4 items, e.g., satisfaction (e.g., In my sport, I feel close to other people).
I am open about my feelings), motivational (5 items, e.g., I show Responses were rated on a 7-point scale (1 = Not true at all,
that I am motivated to work hard with my coach), preventative 7 = very true). Ng et al. (2011) provided support for the valida­
(4 items, e.g., I tell my coach what I expect from him/her), tion of the scale both for internal consistency, and for the factor
assurance (3 items, e.g., I show my coach that s/he can count structure. For the present study, the items were translated into
on me), support (3 items, e.g., I give my coach support when Swedish using a parallel back translation process. Cronbach’s
things are not going well), and social networks (4 items, e.g., alpha was measured for competence (α = .86), autonomy
I like to spend time with our mutual friends). Responses were (α = .83) and relatedness (α = .85). The factorial validity of the
rated on a 7-point scale (1 = Strongly Disagree, 7 = Strongly 20 item, 3-factor CFA model was adequate with regard to all fit
Agree). The CARM-Q has high construct validity and internal indices (χ2/df = 3.25, p < .001; CFI = .91; TLI = .90, RMSEA = .08,
2428 L. DAVIS ET AL.

90% CI = .07, .09) following addition of correlations between statistics including alpha reliability coefficients, means, standard
error terms within the factor of autonomy (between items “In deviation, skewness, and kurtosis were calculated for each vari­
my sport, I have a say in how things are done” “ In my sport, able included in this study. The suggested threshold for Skewness
I can take part in the decision making process” and “ In my is 2, and for kurtosis is 7 (Finney & Distefano, 2006). Structural
sport, I get opportunities to make decisions” and between the Equation Modelling (SEM) was employed in IBM AMOS 26 using
items “ I feel I participate in my sport willingly” and “I choose to maximum likelihood estimation to examine the proposed media­
participate in my sport according to my own free will”) and tion and moderated mediation models of relationship quality
relatedness (between items “ In my sport, I feel close to other (direct and meta) on athletes perceived basic needs satisfaction
people” and “ I have close relationships with people in my (competence, autonomy, relatedness). Because the mediation
sport”). All items loaded significantly (p < .001) on their respec­ model includes multiple mediators, specific indirect effects were
tive latent factors (i.e., autonomy, competence, and calculated by user-defined estimates and evaluated by multiply­
relatedness). ing the two effects associated with this pathway. Furthermore,
the multiple moderated mediation effects were analysed by
using user-defined estimates. To evaluate the conditional indirect
Procedure effects, we followed the procedure suggested by Hayes (2013).
Values of the moderator variable representing “low”, “moderate”
A cross sectional, questionnaire-based design was employed
and “high” were picked using one standard deviation below the
using a purposeful and convenient sampling technique.
mean, the mean, and a standard deviation above the mean.
Approval to conduct this study was granted by the
To examine the potential moderated mediation (interaction)
Swedish ethical review board. Sport-organizations and
effects, factors (CART-Q, CARM-Q, and CAAS) were mean centred
sports clubs were contacted via phone and/or email and
(z score-transformed). In mediation models, we regressed the
informed about the nature of the study; permission to
basic psychological needs (BNSSS) variables (competence, auton­
approach athletes was sought out. Upon consent, one of
omy and relatedness) on the CART-Q (3Cs) to investigate relation­
two methods for data collection were adopted. First, a date
ship quality as a significant direct predictor. Each CARM-Q
and time for the research team to visit the sports clubs
mediator variable (conflict management, openness, motivational,
closest to the first author were arranged. Upon meeting
preventative, assurance, support, social networks) were regressed
the participants at the start of a training session, the study’s
on CART-Q to confirm CART-Q as a significant predictor of the
aims were explained, confidentiality and anonymity assured
CARM-Q mediators, and we as well regressed BNSSS on CARM-Q
and written consent obtained. A multi-section questionnaire
to fully investigate the mediating effects of the model. In moder­
was then distributed in paper and pencil format to all
ated mediation models, attachment styles (secure, anxious, avoi­
participants over the age of 15. In line with the Swedish
dant) were included to investigate if the factors moderate the
Act (2003; 460) and the Swedish Ethical Review Authority
mediating effects (moderated mediation model). This process
(Etikprövnings myndigheten) parent consent was not
was repeated for a moderated mediation model of direct relation­
required as young people of 15 years and older give their
ship quality and for meta relationship quality, resulting in model 1
own consent in Sweden. Participants were asked to com­
and model 2.
plete the questionnaire independently from their coach and
Analyses with IBM Amos SPSS-26 (Arbuckle, 2016) allowed for
peers; members of the research team were on hand to
verification of the measurement and structural models using
supervise any queries. This process took approximately
goodness of fit indices. Using recommendations made by Hu
20 minutes. For those athletes’ who could not be contacted
and Bentler (1999) and Marsh (2007), models were evaluated by
face to face, and in an attempt to reach out to participants
use of chi-square statistic divided by the degrees of freedom (χ2/
across Sweden, a second method of data collection that
df), Comparative Fit Index (CFI), Tucker Lewis index (TLI), and
involved a web-based survey (i.e., SurveyMonkey) was
Root Mean Squared Error of Approximation (RMSEA).
employed. Sport-clubs and organizations distributed the
Conventional cut-off criteria were used to estimate adequate fit
web-based survey to their athletes. The web-based survey
(CFI > 0.90 and TLI > 0.90, RMSEA < 0.08; Kline, 2010) and very
explained the purpose, voluntary nature, anonymity and
good fit (CFI > 0.95 and TLI > 0.95, and RMSEA < 0.06; Hooper
confidentiality of the study, as well as provided instructions
et al. (2008); Hu & Bentler, 1999). For χ2/df, the suggested upper
on how to complete the questionnaire online. Upon con­
thresholds ranged from 2.0 (Tabachnick & Fidell, 2007) to 5.0
sent and after providing a personal code, the multi-section
(Wheaton et al., 1977). To test direct, indirect, and total effects of
questionnaire became available. Following completion, the
the mediation and moderated mediation models, 95% bias cor­
participants’ data were electronically sent to a secure data­
rected bootstrapped confidence intervals (BC CI) were calculated
base for analysis.
based on 1000 resamples, which allowed to interpret how accu­
rate the sample statistic reflected the population parameters
Data Analysis (Preacher & Kelley, 2011). Statistical power of final models (mod­
erated mediation) for the test of close fit were calculated based
Analyses were conducted using SPSS 26 (IBM, 2019) and IBM on RMSEA using significance criterion (α), sample size (n), and
Amos 26 (Arbuckle, 2016). SPSS 26 was used to screen for the model degrees of freedom (df) following the recommendations
proportion of missing data, univariate and multivariate outliers by MacCallum et al. (1996). An illustration of the theoretical
and to compute the variables in the study. In addition, descriptive framework used as a base for analyses is presented in Figure 1.
JOURNAL OF SPORTS SCIENCES 2429

Results moderately associated with motivational (r = .56, p < .01),


support (r = .51, p < .01), and assurance (r = .55, p < .01)
Descriptive statistics and bivariate correlations
strategies.
Descriptive statistics including calculations of alpha reliability
coefficients, means, standard deviations, skewness, and kurto­
Mediated, and moderated mediated effects of the
sis along with Pearson’s correlations of all variables included in
associations between direct relationship quality and basic
this study are presented in Table 1. The means and standard
psychological needs (model 1)
deviations revealed that on average, athletes tended to display
moderate -to -high perceived relationship quality from both Mediation
a direct and meta perspective as well as reporting the use of The first model included direct paths from direct relationship
maintenance strategies as moderate to high. A similar pattern quality on basic need satisfaction (in terms of competence,
was found for basic psychological needs with athletes demon­ autonomy and relatedness) as well as indirect relationships
strating moderate to high levels of autonomy, competence and mediated by relational maintenance strategies including con­
relatedness. Finally, the athletes also tended to display high flict management, openness, motivation, preventative, assur­
levels of attachment security, moderate levels ofattachment ance, support, and social networks. The model provided
avoidance and low levels of attachment anxiety. adequate model fit to the data with regard to χ2/df (2.34), CFI
Bivariate correlations were computed to assess the asso­ (.94), TLI (.92), and RMSEA (.06, CI = .05, .07). Direct relationship
ciations between the study variables. Statistically significant quality was significantly related to all three outcomes of basic
correlations were found between relationship quality (direct psychological needs satisfaction including competence, auton­
and meta), relationship maintenance strategies and athletes omy and relatedness. Direct relationship quality was also sig­
basic psychological needs of autonomy, competence and nificantly related to all seven mediators including conflict
relatedness. Furthermore, athlete’s attachment styles of management, openness, motivational, preventative, assurance,
secure, anxiety and avoidance were significantly correlated support and social networks. Some of the mediators were also
with all study variables and the directions of the correla­ directly related to the outcome variables. Support was nega­
tions were as expected. Perhaps of particular interest for tively related to competence and autonomy and motivational
this study, correlational analyses showed that athletes’ was related to competence, autonomy, and relatedness.
direct-perspective of relationship quality was most strongly Openness was related to competence, and finally assurance
associated with a secure attachment style (r = .71, p < .01), was related to autonomy.
as well as motivational (r = 49, p < .01), support (r = .49, Next, specific indirect effects of direct relationship quality on
p < .01), and assurance (r = .48, p < .01) strategies, and the the three outcomes of basic psychological needs satisfaction
satisfaction of the basic psychological need of autonomy through each of the seven relational maintenance strategies
(r = .53, p < .01). Moreover, the meta-perspective of rela­ were examined. The analysis showed that motivational strate­
tionship quality was most strongly correlated with a secure gies positively mediated specific indirect effects on all three
attachment (r = .74, p < .01) as well as with openness outcomes including competence, autonomy and relatedness.
strategies (r = .58, p < .01) and the satisfaction of the Further, the analysis showed that openness mediated indirect
basic psychological need of autonomy (r = .60, p < .01). effects on competence and finally assurance positively
The meta-perspective of relationship quality was also mediated a specific indirect effect on autonomy. All indirect

Table 1. Descriptive statistics including alpha reliability coefficients (α), means (Ma), standard deviation (SD), skewness, kurtosis, and Pearson correlation matrix for
variables included in this study.
Factor α Ma SD Skewnessb Kurtosisc 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Coach-athlete relationship quality
1. Cart-Q Direct .94 5.67 1.16 −1.50 2.37 -
2. Cart-Q Meta .95 5.51 1.20 −1.10 1.05 .84** -
Coach-Athlete Relationship Maintenance Strategies
3. Conflict Management .81 5.60 1.33 −0.73 0.36 .45** .47** -
4. Openness .80 4.85 1.36 −0.44 −0.42 .44** .58** .35** -
5. Motivational .92 6.07 1.05 −1.34 1.90 .49** .56** .51** .54** -
6. Preventative .84 3.75 1.54 −0.18 −0.58 .35** .44** .19** .64** .38** -
7. Assurance .78 4.44 1.48 −0.09 −0.61 .48** .55** .30** .60** .47** .70** -
8. Support .89 4.42 1.67 −0.21 −0.65 .49** .51** .26** .56** .35** .57** .74** -
9. Social Networks .79 3.65 1.51 0.49 −0.33 .35** .43** .17** .48** .24** .48** .54** .55** -
Coach-athlete attachment
10. Secure .89 5.33 1.41 −0.90 0.41 .71** .74** .36** .48** .40** .36** .44** .47** .37** -
11. Anxiety .89 2.02 1.19 1.48 1.80 .37** .45** .26** .25** .35** .18** 0.20** −.13* .16** .39** -
12. Avoidant .88 3.49 1.48 0.28 −0.67 .55** .57** .23** .48** .39** .43** −.51** .47** .40** .50** .32** -
Basic Psychological Need Satisfaction
13. Autonomy .83 5.47 0.85 −1.16 2.05 .53** .60** .37** .47** .65** .36** .46** .31** .28** .45** .39** .43** -
14. Competence .86 5.94 0.93 −1.31 2.62 .35** .43** .29** .41** .60** .28** .30** .20** .22** .26** .29** .29** .65** -
15. Relatedness .85 6.28 0.89 −1.81 4.37 .38** .40** .31** .32** .49** .20** .23** .17** .18** .28** .33** .19** .50** .57**
Note: **p < .01 *p < .05.
2430 L. DAVIS ET AL.

Table 2. Estimates of significant direct and indirect effects of Cart-Q (direct) on basic psychological needs with bias-corrected bootstrapped confidence
intervals and p-values.
Mediation model β B Boot S.E. 95% BCLL 95% BCUL P
Direct
Cart-Q (D) – Competence .18 .14 .06 .03 .26 .034
Cart-Q (D) – Autonomy .32 .28 .08 .16 .47 .005
Cart-Q (D) – Relatedness .32 .23 .06 .09 .33 .015
Cart-Q (D) – Conflict Management . 45 .45 .06 .32 .55 .019
Cart-Q (D) – Openness .44 .44 .06 .32 .55 .012
Cart-Q (D) – Motivational .49 .49 .07 .36 .63 .011
Cart-Q (D) – Preventative .35 .35 .05 .23 .44 .019
Cart-Q (D) – Assurance .48 .48 .05 .37 .59 .012
Cart-Q (D) – Support .49 .49 .06 .38 .59 .011
Cart-Q (D) – Social Network .35 .35 .05 .24 .44 .009
Support – Competence −.17 −.13 .05 −.25 −.0 .011
Support – Autonomy −.16 −.14 .06 −.28 −.03 .009
Motivation – Competence .51 .38 .07 .23 .52 .008
Motivation – Autonomy .57 .51 .07 .38 .69 .004
Motivation – Relatedness .37 .26 .07 .12 .39 .015
Openness – Competence .12 .09 .05 .02 .21 .022
Assurance – Autonomy .16 .14 .06 .02 .24 .039
Indirect
Cart-Q (D) – Support – Competence −.08 −.06 .02 −.12 −.02 .011
Cart-Q (D) – Support – Autonomy −.08 −.07 .03 −.13 −.02 .009
Cart-Q (D) – Motivation – Competence .25 .19 .05 .10 .30 .009
Cart-Q (D) – Motivation – Autonomy .28 .25 .05 .16 .39 .007
Cart-Q (D) – Motivation – Relatedness .18 .13 .04 .06 .22 .006
Cart-Q (D) – Openness – Competence .05 .04 .02 .01 .09 .023
Cart-Q (D) – Assurance – Autonomy .08 .07 .03 .01 .12 .042
β = Standardized beta, B = Unstandardized beta, S.E. = Standard Error, BCLL = Bias-corrected bootstrap confidence interval lower limit, BCUL = Bias-
corrected bootstrap confidence interval upper limit.

effects of direct relationship quality on the three basic needs Similarly, the indirect effect of a support-focused strategy on
satisfaction scales through conflict management, preventative autonomy was somewhat stronger among individuals with
and social networks were insignificant. A breakdown of all a higher secure attachment compared to a low secure attach­
significant findings can be found in Table 2. ment. Additionally, it was also found that a secure attachment
style moderates the indirect effect of motivational focused
Moderated mediation strategies between direct relationship quality and competence,
Following assessment of the mediation model, we turned to autonomy, and relatedness. Evaluation of conditional indirect
tests of moderated mediation to examine whether athlete’s effects showed that under the condition of high secure attach­
attachment styles towards their coach moderated the ment, the indirect effect of motivational strategy was some­
mediated effect of relationship maintenance strategies what stronger on competence, autonomy, and relatedness
between the quality of the coach-athlete relationship and ath­ than it was under low levels of a secure attachment on compe­
letes perceived basic need satisfaction. For this part of the tence, autonomy, and relatedness. Attachment styles did not
analysis, only the relationship maintenance strategies that indi­ moderate any indirect effects of assurance or openness on
cated significant indirect effects were investigated. Therefore, basic psychological need satisfaction. The final model, includ­
secure, anxious and avoidant attachment styles were investi­ ing both mediation effects and moderated mediation effects,
gated as moderators of the indirect effects of motivation, sup­ explained 49% (R2) of the variance in athletes perceived com­
port, openness and assurance strategies on the relationship petence, 67% in autonomy, and 45% in relatedness. All signifi­
between direct relationship quality and athletes perceived cant moderated mediation effects are presented in Table 3.
competence, autonomy and relatedness. The moderated med­
iation model showed adequate model fit; χ2/df (2.09), CFI (.94),
TLI (.91), and RMSEA (.06, CI = .05, .06). The power of the Mediated, and moderated mediated effects of the
moderated mediation model, based on RMSEA (if .08), associations between meta relationship quality and basic
α = 0.05, n = 350, and df = 386, shows that the power is psychological needs (model 2)
essentially 1.0, which is larger than needed to assure power of
at least 0.80 for rejecting the hypothesis of close fit in the Mediation
population. The second structural model was similar to the first model but
Results showed that a secure attachment style to the coach included the meta relationship quality perspective instead of
moderates the indirect effect of a support-focused strategy direct relationship quality. The model provided adequate
between direct relationship quality and perceived competence model fit to the data with regard to χ2/df (2.39), CFI (.94), TLI
and autonomy. Results showed the overall (negative) indirect (.91), and RMSEA (.06, CI = .06, .07). Meta relationship quality
effect of a support-focused strategy for perceived competence was significantly related to all three outcomes of basic psycho­
was stronger, as indicated by beta weight, under high levels of logical needs satisfaction including competence, autonomy
secure attachment than for those with low secure attachment. and relatedness. Meta relationship quality was also significantly
JOURNAL OF SPORTS SCIENCES 2431

Table 3. Estimates of significant moderated indirect effects of Cart-Q-direct on Basic psychological needs with bias-corrected bootstrapped confidence intervals and
p-values.
Moderated mediation model β B Boot S.E. 95% BCLL 95% BCUL P
Cart-Q (D) – Support – Competence (secure) −.03 −.02 .01 −.05 −.01 .025
+ 1 SD (High) − .08 .03 −.16 −.03
− 1 SD (Low) − .04 .02 −.08 −.01
Cart-Q (D) – Support – Autonomy (secure) −.04 −.02 .01 −.05 −.01 .032
+ 1 SD (High) −.08 .04 −.16 −.03
− 1 SD (Low) −.04 .02 −.09 −.01
Cart-Q (D) – Motivation – Competence (secure) .09 .05 .02 .02 .09 .028
+ 1 SD (High) .19 .08 .09 .29
− 1 SD (Low) .09 .05 .02 .17
Cart-Q (D) – Motivation – Autonomy (secure) .11 .07 .03 .03 .13 .023
+ 1 SD (High) .27 .08 .17 .41
− 1 SD (Low) .14 .06 .04 .26
Cart-Q (D) – Motivation – Relatedness (secure) .07 .03 .02 .01 .06 .032
+ 1 SD (High) .13 .05 .06 .22
− 1 SD (Low) .06 .03 .02 .14
a
Moderator variable in brackets, β = Standardized beta, B = Unstandardized beta, S.E. = Standard Error, BCLL = Bias-corrected bootstrap confidence interval lower
limit, BCUL = Bias-corrected bootstrap confidence interval upper limit, 1 SD = one standard deviation above or below the mean of the moderator variable.

related to all seven mediators including conflict management, preventative, assurance and social networks were insignificant.
openness, motivational, preventative, assurance, support and All significant findings are presented in Table 4.
social networks. The motivational strategy was significantly
related to competence, autonomy, as well as relatedness,
Moderated Mediation
whereas support was directly and negatively related to compe­
Following assessment of the mediation model, we turned to
tence, and autonomy.
tests of moderated mediation to examine whether athletes
Next, specific indirect effects of meta relationship quality on
attachment styles towards their coach moderated the
the three outcomes of basic needs satisfaction through each of
the seven relational maintenance strategies were examined. mediated effect of relationship maintenance strategies
The analysis showed that motivational strategies positively between meta quality of the coach-athlete relationship and
mediated specific indirect effect on all three outcomes includ­ athletes perceived basic need satisfaction. For this part of the
ing competence, autonomy and relatedness. Further, the ana­ analysis, only the relationship maintenance strategies that indi­
lysis showed that support negatively mediated specific indirect cated significant indirect effects were investigated. Therefore,
effects on perceived competence and autonomy. All indirect secure, anxious and avoidant attachment styles were tested as
effects of meta relationship quality on the three basic needs moderators of the indirect effects of motivation, and support
satisfaction scales through conflict management, openness, strategies on the relationship between meta relationship

Table 4. Estimates of significant direct and indirect effects of Cart-Q (meta) on basic psychological needs with bias-corrected bootstrapped confidence
intervals and p-values.
Mediation model β B Boot S.E. 95% BCLL 95% BCUL P
Direct
Cart-Q (M) – Competence .25 .18 .05 .08 .28 .010
Cart-Q (M) – Autonomy .39 .35 .06 .22 .54 .004
Cart-Q (M) – Relatedness .31 .22 .09 .08 .32 .020
Cart-Q (M) – Conflict Management .47 .47 .06 .34 .57 .023
Cart-Q (M) – Openess .58 .58 .04 .45 .65 .036
Cart-Q (M) – Motivation .56 .56 .06 .42 .72 .009
Cart-Q (M) – Preventative .44 .44 .05 .33 .52 .023
Cart-Q (M) – Assurance .55 .55 .04 .46 .65 .015
Cart-Q (M) – Support .51 .51 .05 .40 .61 .012
Cart-Q (M) – Social Network .43 .43 .05 .34 .65 .015
Motivation – Competence .49 .37 .07 .22 .51 .009
Motivation – Autonomy .54 .49 .07 .35 .66 .006
Motivation – Relatedness .37 .26 .07 .13 .39 .014
Support – Competence −.16 −.12 .04 −.23 −.04 .008
Support – Autonomy −.12 −.11 .05 −.20 −.02 .022
Indirect
Cart-Q (M) – Support – Competence −.08 −.06 .02 −.12 −.02 .006
Cart-Q (M) – Support – Autonomy −.06 −.06 .03 −.11 −.01 .022
Cart-Q (M) – Motivation – Competence .27 .21 .05 .11 .33 .007
Cart-Q (M) – Motivation – Autonomy .32 .28 .06 .19 .41 .006
Cart-Q (M) – Motivation – Relatedness .21 .15 .05 .07 .25 .007
β = Standardized beta, B = Unstandardized beta, S.E. = Standard Error, BCLL = Bias-corrected bootstrap confidence interval lower limit, BCUL = Bias-
corrected bootstrap confidence interval upper limit.
2432 L. DAVIS ET AL.

Table 5. Estimates of significant moderated indirect effects of Cart-Q-meta on Basic psychological needs with bias-corrected bootstrapped confidence intervals and
p-values.
Moderated mediation model β B Boot S.E. 95% BCLL 95% BCUL P
CartQ (M) – Support – Autonomy (secure) −.03 −.02 .01 −.05 −.00 .049
+ 1 SD (High) −.07 .03 −.13 −.02
− 1 SD (Low) −.03 .02 −.08 −.01
CartQ (M) – Motivation – Competence (secure) .11 .06 .02 .03 .10 .016
+ 1 SD (High) .23 .06 .14 .36
− 1 SD (Low) .17 .05 .06 .22
CartQ (M) – Motivation – Autonomy (secure) .12 .08 .03 .04 .12 .008
+ 1 SD (High) .32 .07 .22 .48
− 1 SD (Low) .17 .06 .07 .28
CartQ (M) – Motivation – Relatedness (secure) .08 .04 .02 .02 .08 .005
+ 1 SD (High) .16 .05 .07 .24
− 1 SD (Low) .09 .04 .03 .16
a
Moderator variable in brackets, β = Standardized beta, B = Unstandardized beta, S.E. = Standard Error, BCLL = Bias-corrected bootstrap confidence interval lower
limit, BCUL = Bias-corrected bootstrap confidence interval upper limit, 1 SD = one standard deviation above or below the mean of the moderator variable

quality and athletes perceived competence, autonomy and Mediated effects of COMPASS strategies on the
relatedness. The moderated mediation model showed ade­ associations between relationship quality and basic
quate model fit; χ2/df (2.09), CFI (.94), TLI (.91), and RMSEA psychological needs
(.06, CI = .05, .06).Furthermore, similar to the moderated media­ The study’s first hypothesis stated that each of the seven com­
tion model including direct relationship quality, the model with munication strategies identified as conflict management, open­
meta relationship quality showed high power (1.0) based on ness, motivational, preventative, assurance, support, and social
RMSEA. networks (COMPASS) would mediate the relationship between
The results showed that a secure attachment style to the the quality of the coach-athlete relationship as defined by the
coach moderates the indirect effect of a support-focused strat­ 3Cs and athletes’ psychological needs satisfaction. Taking into
egy between meta relationship quality and autonomy. consideration both direct and meta perspectives of relationship
Evaluation of the interaction effects suggests that (negative) quality (model 1 and 2), the direct effects illustrated that both
indirect effects of a support-focused strategy on autonomy are perspectives were significantly associated with satisfaction of
slightly stronger under high levels of secure attachment than all three basic psychological needs and support findings of
under the condition of low secure attachment. Attachment previous studies (Camiré et al., 2019; Riley & Smith, 2011).
style did not moderate the indirect effect of supportive strategy These associations align with Mageau and Vallerand’s (2003)
on perceived competence. proposition, grounded in the hierarchical model of motivation
Finally, it was also found that a secure attachment style (see, Vallerand, 1997), that coaches’ interpersonal behaviours
moderates the indirect effects of a motivational focused strat­ can offer autonomy-support (e.g., involving athletes in deci­
egy between meta relationship quality and competence, sion-making). Additionally, these findings associate with
autonomy, and relatedness. The interactions suggest that Relationships Motivation Theory (RMT; Deci & Ryan, 2014),
with a high secure attachment, the indirect effect of motiva­ which highlights that people should not only experience the
tional strategy is stronger with regards to feelings of compe­ satisfaction of the need for relatedness but also experience the
tence, autonomy, and relatedness than under a low level of satisfaction of the need for autonomy and competence within
secure attachment The model including both mediation effects quality relationships. As such, the findings of the present study
and moderated mediation effects explained 49% of the var­ reaffirm the proposition that a coaching context characterized
iance in athletes perceived competence, 68% in autonomy, and by high levels of closeness (e.g., trust, respect), commitment
43% in relatedness. All significant moderated mediation effects (e.g., loyalty, promise), and complementarity (e.g., readiness,
are presented in Table 5. responsiveness), has the potential capacity to satisfy all three
basic psychological needs.
Perceptions of both direct and meta relationship quality
Discussion were also significantly related to all seven COMPASS strategies
supporting conceptual assumptions (Jowett & Poczwardowski,
The present study investigated the mediating role of commu­ 2007) and empirical evidence (Davis et al., 2019b; Rhind &
nication strategies (COMPASS) on the association between the Jowett, 2011). These findings indicate that good quality rela­
quality of the coach-athlete relationship and athletes’ basic tionships promote better communication, and within these
psychological needs satisfaction. Further, we investigated if relationships athletes engage in a range of diverse communica­
athletes’ attachment style moderates the mediational influence tion strategies. More specifically, good quality relationships can
of communication strategies on the relationship between the provide athletes a social situation that facilitates communica­
quality of the coach-athlete relationship and athletes basic tion which aims to stimulate (motivation), support, assure,
needs satisfaction. encourage (preventative), unlock (openness), connect (social
JOURNAL OF SPORTS SCIENCES 2433

network) and manage (conflict management) relationship been found to be a central psychological and interpersonal
dynamics. It should be noted however, that motivation and process (Rees et al., 2016). However, the provision of support
support were the only two communication strategies that from an athlete may induce perceptions that the recipient
directly associated with basic psychological needs satisfaction; (coach) is inadequate or ineffective. It may also suggest that
that is, motivation was positively associated with satisfaction of an athlete’s role is not to provide their coach support, hence
all three psychological needs and support was negatively asso­ when support is offered the athlete’s needs satisfaction is
ciated with competence and autonomy for both direct and impeded. A recent study found that coaches’ individualised
meta perspective models. As such, athletes who communicate support of athletes (one of 6 transformational leadership beha­
their motivation and passion to work hard, showcase their viours) was negatively linked with the commitment dimension
talent, and achieve their goals are more likely to feel in control of the 3Cs of relationship quality (Zhao & Jowett, 2022).
of their actions and be self-reliant as well as connected with Although contradicting research noted individualised support
others within the context of sport. to be positively linked to high quality coach-athlete relation­
The indirect effects of relationship quality on the three out­ ships overall (López de Subijana et al., 2021). It warrants con­
comes of psychological needs satisfaction showed that for both sideration that the nature of support (e.g., type, frequency,
direct relationship quality (model 1) and meta relationship intensity) can affect athletes’ perceptions of themselves (cf
quality (model 2) two of the seven COMPASS strategies (i.e., R. E. Smith et al., 1983) and their coaches both negatively and
motivation, support) transferred the effects of relationship positively. Consequently, the strategy of support within the
quality onto variations of athletes’ perceptions of needs satis­ context of coach-athlete relationships requires further
faction. Specifically, motivational strategies positively mediated examination.
indirect effects on all three needs, whereas support negatively The communication strategies of openness predicted ath­
mediated specific indirect effects on perceived competence lete competence, and assurance predicted athlete autonomy;
and autonomy. Additionally, openness positively mediated both strategies were mechanisms that potentially explained
indirect effects on competence, and assurance positively the direct link between relationship quality and psychological
mediated a specific indirect effect on autonomy for direct needs satisfaction. These findings suggest that when athletes
perceptions of the coach-athlete relationship; however, it is perceive open lines of communication (e.g., information is
worth noting that the links while significant, were relatively exchanged freely and frequently), and/or they are assured
small (model 1). and comforted during challenging times, they can experience
Research highlights good quality relationships are more a strong connection with their coach as well as free-will and
prevalent in environments that are motivational, engaging, self-sufficiency. When athletes can communicate to their coa­
stimulating, involving, energetic, and dynamic (Duda & ches that they are reliable, responsible, and trustworthy both
Appleton, 2016; N. Smith et al., 2016). Moreover, the findings their relationships and autonomy benefit.
of this study outline the importance for athletes to engage in
motivational communication strategies and interact with their
Moderated mediated effects on the associations between
coaches in ways that showcase their ability, passion, motiva­
relationship quality and basic psychological needs
tion, and commitment, as well as desire to work hard. Athletes
satisfaction
that engage in these aspects of motivational communication
can feel more personally fulfilled (Camiré et al., 2019) as well as The second hypothesis focused on the moderated mediated
develop and maintain functional relationships with their coa­ effects of athlete attachment styles, with the aim of better
ches (Davis et al., 2019b). understanding the mediating effects by determining under
Although motivational strategies can facilitate athletes’ what conditions or variables the mediation is significant
needs satisfaction, support strategies may undermine percep­ (Preacher et al., 2007). To this end, only the COMPASS strate­
tions of competence and autonomy fulfilment. In acknowled­ gies that indicated significant indirect effects were included in
ging the small indirect effects, these findings remain thought the model. For direct relationship quality (model 1) secure,
provoking as the support strategy appeared to potentially anxious, and avoidant attachment styles were included as
detract from both development of high quality coach-athlete moderators of the indirect effects of motivation, support,
relationships and satisfying athletes’ needs of autonomy and openness, and assurance strategies on the relationship
competence. This finding does not directly follow previous between direct relationship quality and athletes’ perceived
research that has found the support dimension of COMPASS competence, autonomy, and relatedness. Our findings were
to positively transfer the effects of coach-athlete relationship partially in line with our hypothesis and highlighted that
quality onto athlete satisfaction (Davis et al., 2019b). a secure attachment style to the coach moderates the indirect
Nonetheless,support, operationalized as giving coach support effect of a support-focused strategy between direct relation­
(e.g., “I give my coach support when things are not going well”; ship quality and perceived competence and autonomy. The
an item on the CARM-Q) appears to act as a mechanism negative indirect effect of a support-focused strategy for per­
between relationship quality and satisfaction of needs, ceived competence and autonomy was stronger for those
although in a negative direction. This suggests that within athletes with high levels of secure attachment than for those
a good quality relationship, giving support to coaches can with low secure attachment. The findings for meta relation­
thwart athletes’ satisfaction of the needs of competence and ship quality (model 2) were corresponding.
autonomy. The provision of support may communicate one’s In explanation of these findings, securely attached athletes
desire to be compassionate, caring, and helpful; support has can feel that within good quality relationships giving support
2434 L. DAVIS ET AL.

to their coaches is not required. Alternatively, athletes high in must possess knowledge and understanding of each unique
security may not view giving support to their coaches as facil­ athlete in the team or squad (Lorimer & Jowett, 2012).
itating their autonomy and/or competence (Felton & Jowett, Understanding athletes’ attachment styles can help coaches
2013a, 2013b). It may also be that a secure attachment style adjust aspects of communication and interaction to facilitate
and quality relationships promotes support provision when it is effective and ethical coaching.
actually required (e.g., during adversity). Beyond sport, secure The findings of this study highlight that motivational strate­
individuals report more positive relationship experiences and gies are an important vehicle to athletes with secure attach­
are satisfied with both giving and receiving support (Kafetsios & ments for developing and maintaining good quality
Sideridis, 2006). For highly secure individuals, relationship qual­ relationships and for satisfying their basic psychological
ity and psychological needs satisfaction may be at their peak needs. Therefore, coaches should pick up the cues when ath­
when there is mutuality (give and take) in support; yet this letes display their motivation, passion and hard work by
warrants investigation within the sport context. acknowledging their efforts. It may also be that coaches try to
It was also found that a secure attachment style moderates cultivate or even educate athletes with insecure attachment
the indirect effects of motivational-focused strategies between styles about the positive effects of using motivational strategies
relationship quality (direct and meta) and satisfaction of all of communication for both the benefits in regard to the rela­
three psychological needs. Interaction effects suggest that the tionship with their coach and the satisfaction of their psycho­
indirect effects of a motivational-focused strategy were some­ logical needs. Training coaches to recognise individual
what stronger on competence, autonomy, and relatedness for difference characteristics such as attachment styles can help
athletes with high levels of secure attachment. This finding coaches overcome interpersonal difficulties that may arise from
signifies that within the context of good quality relationships, athletes’ apparent lack of connection, unwillingness to invest in
secure athletes can readily communicate enthusiasm through the relationship, or overbearing need for coaches’ attention. An
working hard, showing their talent, and striving to achieve their awareness of the implications of diverse personalities may help
goals; in turn their three psychological needs are satisfied. As coaches support athletes to develop secure and positive rela­
such, facilitating a secure attachment within the coach-athlete tionships over time (Côté & Gilbert, 2009; Jowett & Felton,
relationship is paramount for healthy relational development 2014). This type of training is important to ensure coaches
and continued satisfaction. provide effective and fair provision to all athletes in their care.
Findings in the present study did not observe anxious and
avoidant attachment to be significant moderators. This partially
Strength, Limitations and Future Directions
supports previous research (Davis & Jowett, 2014; Davis et al.,
2013; Felton & Jowett, 2013b) that has found non-significant Overall, the findings of this study make an original contribution
findings for attachment anxiety on athlete outcomes including to the literature and extend previous research by demonstrat­
relationship satisfaction and well-being. The absence of any ing how engaging in different communication strategies
moderated mediation findings for the insecure attachment (COMPASS) can enhance coach-athlete relationship quality
styles of avoidance and anxiousness may be due to anxious and satisfaction of athletes’ psychological needs. Moreover,
individuals being rarely satisfied with the quality of their rela­ examining individual differences contributes to understanding
tionships with others, and that avoidant individuals are happy of the mediating effect of COMPASS by determining under
to remain emotionally and behaviourally disconnected which attachment styles the mediation maybe significant.
(Ainsworth et al., 1978; Mikulincer & Shaver, 2017). It is possible Taken together, the present study addresses all layers of the
that insecure athletes do not engage in communication strate­ integrated research model (Jowett & Poczwardowski, 2007)
gies either, because they have limited interest or capacity/ illustrating the various factors that are necessary to consider
ability to do so. in achieving effective relationships. That said, whilst the study
has conceptual and empirical strengths, there are also limita­
tions that need to be acknowledged. First, with any self-report
Practical significance of the findings
data there is concern about social desirability and the truthful­
The significance of this research lies largely on its practical ness of responses. In this study, questionnaires were anon­
significance for effective coaching. Understanding the role of ymous and were completed without the coach present to
communication strategies (COMPASS) is important as it pro­ facilitate a safe environment for participants to supply honest
vides tangible means for practitioners (e.g., coaches, coach and truthful responses. Second, cross-sectional data can intro­
developers, sport psychologists, lifestyle advisors) aiming to duce common method variance or bias and prevents causality
maximise the effectiveness of key interpersonal relationships from being established between the variables. Prospective
and help athletes fulfil their basic psychological needs; both of longitudinal or experimental studies examining an interven­
which are essential for growth and development (Deci & Ryan, tion-based programme informed by the 3Cs and COMPASS
2014). The applied implications of the research can promote could further establish predictive power relative to important
the development of positive coaching environments within performance and wellbeing outcomes. Third, data were gath­
which athletes fulfil their basic psychological needs while ered from only athletes; data from coaches would yield addi­
acknowledging the role of individual difference characteristics. tional insights and offer a more holistic picture of associations
A key aspect of effective coaching is to improve athletes’ out­ within coach-athlete dyads. This is an important consideration
comes: competence, confidence, connection, and character as athletes and coaches do not interact and communicate in
(Côté & Gilbert, 2009); to promote optimal outcomes, coaches isolation. Finally, the sample comprised both individual and
JOURNAL OF SPORTS SCIENCES 2435

team sports which potentially creates issues with biased esti­ attachment. Journal of Clinical Sport Psychology, 4(2), 112–132. https://
mates and generalizability of the findings. Future research doi.org/10.1123/jcsp.4.2.112
should consider further analytical techniques to determine Davis, L., & Jowett, S. (2013). Attachment styles within the coach athlete
dyad: Preliminary investigation and assessment development. Journal of
potential differences between individual and team sports.
Clinical Sport Psychology, 7(2), 120–145. https://doi.org/10.1123/jcsp.7.2.
120
Davis, L., & Jowett, S. (2014). Coach-athlete attachment and the quality of
Summary the coach-athlete relationship: Implications for athlete’s well-being.
Journal of Sports Science, 32(15), 1454–1464. https://doi.org/10.1080/
In conclusion, the present study adds to the growing body of 02640414.2014.898183
research on interpersonal relationships in sport by highlighting Davis, L., Jowett, S., & Lafrenière, M. A. K. (2013). An attachment theory
the ways by which a number of communication strategies from perspective in the examination of relational processes associated with
the COMPASS model (i.e., motivational, support, openness and coach-athlete dyads. Journal of Sport Exercise Psychology, 35(2), 156–167.
assurance) transferred the effects of coach-athlete relationship https://doi.org/10.1123/jsep.35.2.156
Davis, L., Jowett, S., & Tafvelin, S. (2019a). Communication strategies: The
quality onto athletes’ competence, autonomy and relatedness.
fuel for quality coach-athlete relationships and athlete satisfaction.
Furthermore, the findings demonstrated that athletes’ indivi­ Frontiers in Psychology, 10, 2156. https://doi.org/10.3389/fpsyg.2019.
dual difference in attachment played a significant role in the 02156
interplay between communication strategy, relationship qual­ Davis, L., Stenling, A., Gustafsson, H., Appleby, R., & Davis, P. (2019b).
ity and satisfaction of psychological needs. Subsequently, indi­ Reducing the risk of athlete burnout: Psychosocial, sociocultural, and
vidual differences (e.g., attachment styles) and social individual considerations for coaches. International Journal of Sports
Science & Coaching, 14(4), 444–452. http://doi.org/10.1177/
characteristics (e.g., communication) are important factors
1747954119861076
that provide a context from which to better understand the Deci, E. L., & Ryan, R. M. (2014). Autonomy and Need Satisfaction in Close
relationship and its outcomes. Relationships: Relationships Motivation Theory. In N. Weinstein (Ed.),
Human Motivation and Interpersonal Relationships (53–73). Springer.
https://doi.org/10.1007/978-94-017-8542-6_3
Disclosure statement Duda, J., & Appleton, P. (2016). Empowering and Disempowering Coaching
Behaviors: Conceptualization, Measurement Considerations, and
No potential conflict of interest was reported by the author(s). Intervention Implications. In M. Raab, P. Wylleman, R. Seiler, A.-M. Elbe,
& A. Hatzigeorgiadis (Eds.), Sport and Exercise Psychology Research: From
Theory to Practice (1st) ed., pp. 373–388). Elsevier.
Funding Felton, L., & Jowett, S. (2013a). The mediating role of social environmental
factors in the associations between attachment styles and basic needs
This research was funded and supported by the Swedish Research Council satisfaction. Journal of Sports Sciences, 31(6), 618–628. https://doi.org/10.
for Sports Science (centrumforidrottsforskning; CIF) and Umeå School of 1080/02640414.2012.744078
Sport Sciences (Idrottshögskolan; IH). Felton, L., & Jowett, S. (2013b). Attachment and well-being: The mediating
effects of psychological needs satisfaction within the coach-athlete and
parent-athlete relational contexts. Psychology of Sport and Exercise, 14(1),
References 57–65. https://doi.org/10.1016/j.psychsport.2012.07.006
Felton, L., & Jowett, S. (2017). A Self-Determination Theory Perspective on
Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of Attachment, Need Satisfaction, and Well-Being in a Sample of Athletes:
attachment: Assessed in the strange situation and at home. Lawrence A Longitudinal Study. Journal of Clinical Sport Psychology, 11(4), 304–323.
Erlbaum.
https://doi.org/10.1123/jcsp.2016-0013
Arbuckle, J. L. (2016). Amos (Version 26.0) [Computer Program]. IBM SPSS.
Finney, S. J., & DiStefano, C. (2006). Non-normal and Categorical data in
Bergin, C., & Bergin, D. (2009). Attachment in the classroom. Educational
structural equation modeling. In G. R. Hancock & M. R.O (Eds.), Structural
Psychology Review, 21, 141–170. https://doi.org/10.1007/s10648-009-
Equation Modeling: A Second Course (pp. 269–314). Greenwich,
9104-0
Connecticut: Information Age Publishing.
Camiré, M., Rathwell, S., Turgeon, S., & Kendellen, K. (2019). Coach-athlete
Gosai, J., Jowett, S., & Nascimento-Júnior, J. R. A. D. (2021). When leadership,
relationships, basic psychological need satisfaction and thwarting, and
relationships and psychological safety promote flourishing in sport and
the teaching of life skills in Canadian high school sport. International
life. Sports Coaching Review, 1–21. https://doi.org/10.1080/21640629.
Journal of Sports Science & Coaching, 14(5), 591–606. http://doi.org/10.
2021.1936960
1177/1747954119869542
Collins, N. L., & Read, S. J. (1990). Adult attachment, working models, and Hampson, R., & Jowett, S. (2014). Effects of coach leadership and
relationship quality in dating couples. Journal of Personality and Social coach-athlete relationship on collective efficacy. Scandinavian Journal
Psychology, 58(4), 644–663. https://doi.org/10.1037/0022-3514.58.4.644 of Medicine & Science in Sports, 24(2), 454–460. https://doi.org/10.1111/j.
Côté, J., & Gilbert, W. (2009). An Integrative Definition of Coaching 1600-0838.2012.01527.x
Effectiveness and Expertise. International Journal of Sports Science & Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional
Coaching, 4(3), 307–323. https://doi.org/10.1260/174795409789623892 process analysis: A regression-based approach. New York, NY: Guilford.
Davis, L., Appleby, R., Davis, P., Wetherell, M., & Gustafsson, H. (2018). The Hooper, D., Coughlan, J., & Mullen, M. R. (2008). Structural Equation
role of coach-athlete relationship quality in team sport athletes’ psycho­ Modelling: Guidelines for Determining Model Fit. Electronic Journal of
physiological exhaustion: Implications for physical and cognitive perfor­ Business Research Methods, 6(1), 53–60. https://doi.org/10.21427/D7CF7R
mance. Journal of Sports Science, 36(17), 1985–1992. https://doi.org/10. Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance
1080/02640414.2018.1429176 structure analysis: Conventional criteria versus new alternatives.
Davis, L., Brown, D., Arnold, R., & Gustafsson, H. (2021). Thriving through Structural Equation Modelling: A Multidisciplinary Journal, 6(1), 1–55.
relationships in sport: The role of the parent-child and coach-athlete https://doi.org/10.1080/10705519909540118
attachment relationship. Frontiers in Psychology, 12. https://doi.org/10. IBM Corp. (2019). IBM SPSS Statistics for Windows. Version Vol. 26.0.
3389/fpsyg.2021.694599 Armonk, NY: IBM Corp.
Davis, L., & Jowett, S. (2010). Investigating the interpersonal dynamics Jackson, B., Dimmock, J. A., Gucciardi, D. F., & Grove, J. R. (2011). Personality
between coaches and athletes based on fundamental principles of traits and relationship perceptions in coach-athlete dyads: Do opposites
2436 L. DAVIS ET AL.

really attract? Psychology of Sport and Exercise, 12(3), 222–230. https:// Preacher, K. J., Rucker, D. D., & Hayes, A. F. (2007). Addressing moderated
doi.org/10.1016/j.psychsport.2010.11.005 mediation hypothesis: Theory, methods and prescriptions. Multivariate
Jowett, S. (2007). Expanding the interpersonal dimension of coaching: Behavioural Research, 42(1), 185–227. https://doi.org/10.1080/
Closeness in the coach-athlete relationship: A commentary. 00273170701341316
International Journal of Sport Science and Coaching, 2(4), 513–517. Rees, T., Hardy, L., Güllich, A., Abernethy, B., Côté, J., Woodman, T.,
https://doi.org/10.1260/174795407783359632 Montgomery, H., Laing, S., & Warr, C. (2016). The Great British medalists
Jowett, S. (2009). Validating coach-athlete relationship measures with the project: A review of current knowledge into the development of the
nomological network. Measurement in Physical Education and Exercise world’s best sporting talent. Sports Medicine, 46, 1041–1058. https://doi.
Science, 13(1), 34–51. https://doi.org/10.1080/10913670802609136 org/10.1007/s40279-016-0476-2
Jowett, S., & Felton, L. (2014). Group dynamics in exercise and sport psychol­ Rhind, D. J. A., & Jowett, S. (2010). Relationship maintenance strategies in
ogy (pp. 73–90). (M. Beauchamp & M. Eys, Eds.). Taylor & Francis Group. the coach-athlete relationship: The COMPASS model. Journal of Applied
Jowett, S., & Nezlek, J. (2012). Relationship interdependence and satisfac­ Sport Psychology, 22(1), 106–121. https://doi.org/10.1080/
tion with important outcomes in coach–athlete dyads. Journal of Social 10413200903474472
and Personal Relationships, 29(3), 287–301. https://doi.org/10.1177/ Rhind, D. J. A., & Jowett, S. (2011). Linking maintenance strategies to the
0265407511420980 quality of coach-athlete relationships. International Journal Sport
Jowett, S., & Ntoumanis, N. (2004). The coach-athlete relationship question­ Psychology, 42(1), 55–68.
naire (CART – Q): Development and initial validation. Scandinavian Rhind, D. J. A., & Jowett, S. (2012). Development of the coach-athlete
Journal of Medicine & Science in Sports, 14(4), 245–257. https://doi.org/ relationship maintenance questionnaire (CARM-Q). International
10.1111/j.1600-0838.2003.00338.x Journal of Sports Science & Coaching, 7(1), 121–137. https://doi.org/10.
Jowett, S., & Poczwardowski, A. (2007). Understanding the coach-athlete 1260/1747-9541.7.1.121
relationship. In S. Jowett & D. Lavallee (Eds.), Social Psychology in Sport Riley, A., & Smith, A. L. (2011). Perceived coach-athlete and peer relation­
(pp. 3–14). Human Kinetics. ships of young athletes and self-determined motivation for sport.
Jowett, S., & Shanmugam, V. (2016). Relational Coaching in Sport: Its International Journal of Sport Psychology, 42(1), 115–133.
psychological underpinnings and practical effectiveness. In R. Schinke, Shanmugam, V., Jowett, S., & Meyer, C. (2014). Interpersonal difficulties as a risk
K. R. McGannon, & B. Smith (Eds.), Routledge International Handbook of factor for athletes’ eating psychopathology. Scandinavian Journal of Medicine
Sport Psychology (pp. 471–484). Routledge. & Science in Sports, 24(2), 469–476. https://doi.org/10.1111/sms.12109
Kafetsios, K., & Sideridis, G. D. (2006). Attachment, Social Support and Smith, N., Quested, E., Appleton, P., & Duda, J. (2016). Observing the
Well-being in Young and Older Adults. Journal of Health Psychology, 11 coach-created motivational environment across training and competi­
(6), 863–876. https://doi.org/10.1177/1359105306069084 tion in youth sport. Journal of Sports Sciences, 35(2), 149–158. https://doi.
Kline, R. B. (2010). Principles and Practice of Structural Equation Modeling org/10.1080/02640414.2016.1159714
(3rd) ed.). Guilford Press. Smith, R. E., Zane, N. W., Smoll, F. L., & Cappel, D. B. (1983). Behavioral
Li, J., Gao, H., Liu, P., & Zhong, C. (2020). Does distance produce beauty? The assessment in youth sports: Coaching behaviors and children’s attitudes.
influence of COVID-19 lockdown on the coach-athlete relationship in Medicine and Science in Sports and Exercise, 15(3), 208–214. https://doi.
a Chinese football school. Frontiers in Psychology, 11, 560–638. https:// org/10.1249/00005768-198315030-00005
doi.org/10.3389/fpsyg.2020.560638 Steadman, L. 2021 Words cannot describe how special it was to have my
López de Subijana, C., Martin, L. J., Ramos, J., & Côté, J. (2021). How coach coach by my side yesterday. He believed in me, encouraged me. Tweet.
leadership is related to the coach-athlete relationship in elite sport. https://twitter.com/LaurenSteadman/status/1432264948848238592?s=
International Journal of Sports Science & Coaching, 16(6), 1239–1246. 20&t=KaOo0DvHOXYKFxDRXUQZ8Q. Twitter.
https://doi.org/10.1177/17479541211021523 Sullivan, P., Rhind, D., & Jowett, S. (2014). Communication. In
Lorimer, R., & Jowett, S. (2012). Empathic understanding and accuracy in the A. Papaioannou & D. Hackfort (Eds.), Routledge Companion to Sport and
coach-athlete relationship. In J. Denison, W. Gilbert, & P. Potrac (Eds.), Exercise Psychology: Global Perspectives and Fundamental Concepts (pp.
Routledge Handbook of Sports Coaching (321–332). Routledge. 559–570). Routledge.
MacCallum, R. C., Browne, M. W., & Suguwara, H. M. (1996). Power analysis Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics (5th) ed.).
and determination of sample size for covariance structure modelling. Allyn & Bacon/Pearson Education.
Psychological Methods, 1(2), 130–149. https://doi.org/10.1037/1082- Vallerand, R. J. 1997. Toward a Hierarchical Model of Intrinsic and Extrinsic
989X.1.2.130 Motivation. In Advances in Experimental Social Psychology M. P. Zanna Ed.
Mageau, G. A., & Vallerand, R. J. (2003). The coach–athlete relationship: Academic Press 271–360. http://dx.doi.org/10.1016/s0065-2601(08)60019-2
A motivational model. Journal of Sports Sciences, 21(11), 883–904. Vella, S. A., Oades, L. G., & Crowe, T. P. (2013). The relationship between
https://doi.org/10.1080/0264041031000140374 coach leadership, the coach–athlete relationship, team success, and the
Marsh, H. W. (2007). Application of confirmatory factor analysis and struc­ positive developmental experiences of adolescent soccer players.
tural equation modelling in sport/exercise psychology. In Physical Education and Sport Pedagogy, 18(5), 549–561. https://doi.org/
(G. Tenenbaum & R. C. Eklund, Eds.). Handbook of sport psychology (3rd 10.1080/17408989.2012.726976
ed., pp. 774–798) Wiley. Wachsmuth, S., Jowett, S., & Harwood, C. (2018). Managing conflict in
Mayseless, O. (2010). Attachment and the leader-follower relationship. coach-athlete relationships. Sport, Exercise, and Performance
Journal of Social and Personal Relationships, 27(2), 271–280. https://doi. Psychology, 7(4), 371–390. https://doi.org/10.1037/spy0000129
org/10.1177/2F0265407509360904 Wheaton, B., Muthen, B., Alwin, D. F., & Summers, G. (1977). Assessing
Mikulincer, M., & Shaver, P. R. (2017). Attachment in Adulthood: Structure, Reliability and Stability in Panel Models. Sociological Methodology, 8,
dynamics and change (2nd) ed.). The Guildford press. 84–136. https://doi.org/10.2307/270754
Ng, J. Y. Y., Lonsdale, C., & Hodge, K. (2011). The Basic Needs Satisfaction in Yang, S. X., Jowett, S., & Chan, D. K. (2015). Effects of big-five personality
Sport Scale (BNSSS): Instrument development and initial validity traits on the quality of relationship and satisfaction in Chinese coach–
evidence. Psychology of Sport and Exercise, 12(3), 257–264. https://doi. athlete dyads. Scandinavian Journal of Medicine & Science in Sports, 25(4),
org/10.1016/j.psychsport.2010.10.006 568–580. https://doi.org/10.1111/sms.12329
Preacher, K. J., & Kelley, K. (2011). Effect size measures for mediation models: Zhao, C., & Jowett, S. (2022). Coach transformational leadership can fall flat
Quantitative strategies for communicating indirect effects. Psychological on its face without quality coach-athlete relationships. Manuscript under
Methods, 16(2), 93–115. http://dx.doi.org/10.1037/a0022658 review.

You might also like