Interpersonal Emotion Regulation Among Adolescent Athletes ......Athletes: A Bayesian Multilevel...

541

ORIGINAL RESEARCH

Journal of Sport & Exercise Psychology, 2016, 38, 541 -555http://dx.doi.org/10.1123/jsep.2015-0189

© 2016 The Authors. Published by Human Kinetics, Inc. This is an Open Access article distributed under the terms of the Creative Commons Attribution License CC BY NC 4.0, which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original work is properly cited, the new use includes a link to the license, and any changes are indicated. See http://creativecommons.org/licenses/by-nc/4.0. This license does not cover any third-party material which may appear with permission in the article. For commercial use, permission should be requested from Human Kinetics, Inc., through the Copyright Clearance Center (http://www.copyright.com).

*Katherine Tamminen and Patrick Gaudreau made equal contributions to the writing of this article. Katherine A. Tamminen is with the Faculty of Kinesiology & Physical Education, University of Toronto, Toronto, Ontario, Canada. Patrick Gaudreau is with the School of Psychology, University of Ottawa, Ottawa, Ontario, Canada. Carolyn E. McEwen and Peter R.E. Crocker are with the School of Kinesiology, University of British Columbia, Vancouver, British Columbia, Canada. Address author correspondence to Katherine A. Tamminen at [email protected].

Interpersonal Emotion Regulation Among Adolescent Athletes: A Bayesian Multilevel Model Predicting Sport

Enjoyment and Commitment

Katherine A. Tamminen,1* Patrick Gaudreau,2* Carolyn E. McEwen,3 and Peter R.E. Crocker3

1University of Toronto; 2University of Ottawa; 3University of British Columbia

Efforts to regulate emotions can influence others, and interpersonal emotion regulation within teams may affect athletes’ own affective and motivational outcomes. We examined adolescent athletes’ (N = 451, N teams = 38) self- and interpersonal emotion regulation, as well as associations with peer climate, sport enjoyment, and sport commitment within a multilevel model of emotion regulation in teams. Results of multilevel Bayesian structural equation modeling showed that athletes’ self-worsening emotion regulation strategies were negatively associated with enjoyment while other-improving emotion regulation strategies were positively associated enjoyment and commitment. The team-level interpersonal emotion regulation climate and peer motivational climates were also associated with enjoyment and commitment. Team-level factors moderated some of the relationships between athletes’ emotion regulation with enjoyment and commitment. These findings extend previous research by examining interpersonal emotion regulation within teams using a multilevel approach, and they demonstrate the importance of person- and team-level factors for athletes’ enjoyment and commitment.

Keywords: emotion regulation; peer motivational climate; youth sport; team sport; emotional climate, Bayes-ian structural equation model

Emotion regulation occurs in an interpersonal context in which individuals can influence one another to improve or worsen relationship and performance outcomes (Eisenberg, Fabes, Guthrie, & Reiser, 2000; Tamminen & Gaudreau, 2014). However, sport research-ers have mostly taken an intrapersonal perspective by focusing on how individual athletes appraise stressors (Thatcher & Day, 2008); how they experience, express, and regulate their own emotions (Martinent, Campo, & Ferrand, 2012); and how they cope with stress before and during competitions (Gaudreau, Nicholls, & Levy, 2010; Mellalieu, Hanton, & Shearer, 2008). This study adopted an interpersonal, multilevel approach to examine

the ways adolescent athletes try to regulate their own emotions and the emotions of their teammates, and the associations of these emotion regulation strategies (ERS) with enjoyment and commitment.

Emotion regulation refers broadly to the “processes by which individuals influence which emotions they have, when they have them, and how they experience and express these emotions” (Gross, 1998, p. 275). Recently, Niven and colleagues proposed a conceptual framework that differentiates the intrinsic ERS that indi-viduals may use to improve and worsen their own emo-tions and extrinsic ERS to try to improve or worsen the emotions of others (Niven, Totterdell, & Holman, 2009;

Unauthenticated | Downloaded 06/15/21 01:00 AM UTC

http://dx.doi.org/10.1123/jsep.2015-0189http://creativecommons.org/licenses/by-nc/4.0http://www.copyright.commailto:[email protected]

542 Tamminen et al.

JSEP Vol. 38, No. 6, 2016

Niven, Totterdell, Stride, & Holman, 2011). Actions to improve one’s own emotions may include strategies such as seeking support, thinking about positive things, or doing something enjoyable (self-improving ERS). In contrast, emotion regulation actions to worsen one’s own emotions could include strategies such as thinking about negative experiences or shortcomings, expressing cynicism, or looking for problems in the current situation (self-worsening ERS). Emotion regulation actions to try to improve the emotions of others may include strategies such as giving someone helpful advice, doing something nice with someone, or listening to their problems (other-improving ERS). Emotion regulation actions to try to worsen others’ emotions may include acting annoyed toward someone, telling someone about their shortcom-ings, and explaining to someone how they had hurt others to make them feel worse (other-worsening ERS). This conceptual framework moves beyond an intrapersonal perspective to propose that individuals play an active role in trying to shape the emotional experiences of others.

Coping and emotion regulation share many con-ceptual similarities (Compas et al., 2014) and, as such, they each provide a reasonable anchor point to inform the development of hypotheses in regard to intrinsic or intrapersonal emotion. Researchers examining coping and emotion regulation from an intrapersonal perspective have provided evidence that athletes perceive they are effective in regulating their emotions (Wagstaff, 2014), and that task-oriented or problem-focused coping and emotional self-regulation in sport are associated with positive outcomes such as improved performance (Balk, Adriaanse, de Ridder, & Evers, 2013; Gaudreau, Nicholls, & Levy, 2010), positive affect (Gaudreau & Blondin, 2004), motivation, and well-being (Smith, Ntoumanis, Duda, & Vansteenkiste, 2011). Beyond the sport domain, there is also evidence that self-regulation strategies such as problem solving, reappraisal, and acceptance are nega-tively associated with depression and anxiety (Aldao, Nolen-Hoeksema, & Schweizer, 2010), and such self-improving ERS are associated with increases in positive mood (Niven et al., 2011; Niven, Totterdell, Miles, Webb, & Sheeran, 2013). Conversely, self-regulation strategies such as rumination, avoidance, and emotion suppression have been associated with increased depression and anxi-ety (Aldao et al., 2010; Nolen-Hoeksema & Aldao, 2011), and similar self-worsening ERS have been associated with increases in negative mood (Niven et al., 2011; Niven et al., 2013). Overall, there is evidence that individuals can and do improve and worsen their own emotions across a variety of situations and that emotional self-regulation can be associated with a number of positive or negative affective and motivational outcomes.

Previous researchers have noted that much of the coping and emotion regulation research in sport has been conducted from an intrapersonal perspective (Crocker, Tamminen, & Gaudreau, 2015; Friesen, Lane, et al., 2013; Tamminen & Gaudreau, 2014). Despite the social proximity, interdependence, and ubiquitous interactions among teammates in competitive sports, only a handful

of studies have examined emotion regulation among athletes from an interpersonal perspective. Researchers examining mood linkage and emotional contagion have demonstrated that the positive moods of professional cricket players are linked to the collective happy moods of their teammates (Totterdell, 2000; Totterdell & Leach, 2001). Additional research has shown that athletes are aware of their influence on the emotions, cognitions, and behaviors of their teammates, and they also deliberately try to regulate the emotions of their teammates (Friesen, Devonport, Sellars, & Lane, 2013; Tamminen & Crocker, 2013). Research beyond the sport domain indicates that efforts to regulate the emotions of others can influence one’s own emotions. In field and laboratory studies, Niven, Totterdell, Holman, and Headley (2012) revealed that individuals who engaged in other-improving ERS reported greater emotional well-being whereas individu-als who engaged in other-worsening ERS reported worse emotional well-being 1 month later. Efforts to improve the emotions of a coworker have also been associated with interpersonal connections in which the coworkers each experience an increase in their perceptions of friendship (Niven, Holman, & Totterdell, 2012). However, sport researchers have yet to examine affective and motiva-tional outcomes associated with athletes’ use of other-improving and other-worsening ERS among teammates.

A Multilevel Model of Emotion Regulation of Athletes in Teams

In this study, our goal was to examine intrinsic (self-improving and self-worsening) and extrinsic (other-improving and other-worsening) ERS within a multilevel model of emotion regulation of athletes in teams (MERAT; Figure 1). In the MERAT, we propose that an emotion regulation climate emerges to create differ-ences across teams in terms of how teammates interact to regulate their own emotions and the emotions of others. On the one hand, athletes in some teams may tend to use more other-worsening ERS, thereby creating an emotion regulation climate in which it is socially acceptable to act in ways that worsen the emotions of others. Athletes inter-acting and competing in such teams are likely to experi-ence affective and motivational deficits. Conversely, some teams may develop an emotion regulation climate that is constituted by a high degree of extrinsic other-improving ERS. In such teams, athletes would typically engage in actions to try to improve the emotions of teammates. As such, it is expected that involvement in this emotion regulation climate would contribute to fostering better emotional and motivation adjustment of athletes.

The MERAT also incorporates some of the basic principles from achievement goal theory because it offers a fertile conceptual framework to delineate how distinct sets of values and priorities can set the tone for the emergence of qualitatively different dimensions of peer motivational climate in sport. The peer motiva-tional climate refers to perceptions of motivational cues


543JSEP Vol. 38, No. 6, 2016

Fig

ure

1 —

Mul

tilev

el m

odel

of

emot

ion

regu

latio

n of

ath

lete

s in

team

s (M

ER

AT

).


544 Tamminen et al.

JSEP Vol. 38, No. 6, 2016

and expectations that encourage a particular goal ori-entation and influence goal involvement states (Ames, 1992; Ntoumanis & Vazou, 2005). An ego-involving (or performance) climate emphasizes social compari-son and normative ability while a task-involving (or mastery) climate emphasizes effort and improvement. A recent systematic review demonstrated that an ego-involving peer motivational climate was positively associated with negative affect and negative cognitions among athletes while showing negative associations with positive affect; conversely, a task-involving climate showed positive associations with positive affect and negative associations with negative affect (Harwood, Keegan, Smith, & Raine, 2015). Researchers have argued that perceptions of the peer motivational climate appear to vary between teams (Vazou, 2010) and that the motivational climate should be studied as a team-level variable within sport settings (Papaioannou, Marsh, & Theodorakis, 2004). In a year-long study, Ntoumanis, Taylor, and Thogersen-Ntoumani (2012) reported that athletes on teams with higher average perceptions of a peer task climate showed greater positive and prosocial attitudes in sport, greater subjective vitality and com-mitment, and decreased burnout. In contrast, athletes on teams with higher perceptions of a peer ego climate displayed higher levels of antisocial attitudes (i.e., endorsement of cheating) and burnout, and an ego peer climate was negatively associated with commitment. Overall, findings suggest that the peer motivational climate is an important team-level factor which may influence the emotions, cognitions, and interpersonal behaviors of athletes.

In the MERAT, we hypothesized that the emo-tion regulation and peer motivational climates within teams have the potential to shape the relationship of interpersonal emotion regulation with emotional and motivational outcomes experienced by competitive ath-letes. More specifically, the peer motivational climate and emotion regulation climate were conceptualized as team-level factors that could moderate the associations between intrinsic and extrinsic emotion regulation and the aforementioned outcomes of athletes at the person level (see Figure 1). In a team with a high task-oriented motivational climate and a high other-improving emo-tion regulation climate, athletes are likely to be recep-tive to and supportive of their teammates’ extrinsic other-improving ERS. The collaborative nature of these climates should create the needed social nexus to bolster the positive effects and dampen the negative effects associated with extrinsic other-improving and other-worsening ERS, respectively. Conversely, athletes in teams with a high ego-involving peer motivational climate and a high other-worsening emotion regulation climate are likely to compete for limited resources, rewards, and privileges, thereby minimizing the likeli-hood that teammates will attempt to improve the emo-tions of others and to be supportive of their emotional needs. As a result, an ego-involving motivational climate

and other-worsening emotion regulation climate have the potential to reduce the positive effects and aggravate the negative effects of athletes’ extrinsic other-improving and other-worsening ERS.

This StudyWe selected sport enjoyment and commitment as our dependent variables to provide an empirical test of hypotheses from the MERAT because they can be viewed as indicators of pleasurable engagement—a positive psy-chological state that plays an important role in promoting self-regulation, learning, and achievement (Fredrickson, 2001). Sport enjoyment is defined as “a positive affective response to the sport experience that reflects generalized feelings such as pleasure, liking, and fun” (Scanlan, Car-penter, Simons, Schmidt, & Keeler, 1993, p. 18) whereas sport commitment represents a motivational desire and resolve to continue sport participation (Scanlan et al., 1993). We proposed that both individual differences and team-level factors need to be considered because enjoy-ment and commitment can be impacted by the actions of others, particularly during adolescence as interactions with peers are thought to be particularly salient during this developmental period (Chan, Lonsdale, & Fung, 2012). Positive affiliation with peers and perceived peer acceptance as well as many correlates of sport dropout (e.g., conflict with teammates, negative team atmosphere) are associated with young athletes’ sport enjoyment and commitment (e.g., Boiché & Sarrazin, 2009; MacDonald, Côté, Eys, & Deakin, 2011; Ullrich-French & Smith, 2006). Thus, enjoyment and commitment were appropri-ate dependent variables to provide an account of affective and motivational outcomes that are important in the lives of adolescent athletes.

We hypothesized that self-improving and self-worsening ERS of athletes would be associated with higher and lower enjoyment and commitment, respec-tively. We also hypothesized that other-improving and other-worsening ERS of athletes would be associated with athletes’ higher and lower sport enjoyment and commitment, respectively. Beyond these Level 1 effects, we expected that climates characterized by positive inter-actions (e.g., a task-involving peer motivational climate or an emotion regulation climate consisting of other-improving ERS) could strengthen the positive effects associated with person-level ERS aimed at improving our own and others’ emotions. These climates could also weaken the negative effects associated with ERS aimed at worsening our own and others’ emotions. In contrast, climates characterized by negative interactions (e.g., an ego-involving peer motivational climate or an emotion regulation climate consisting of other-worsening ERS) could strengthen the negative effects associated with person-level ERS aimed at worsening emotions while such climates could weaken the positive effects associated with athletes’ use of strategies to improve their own and their teammates’ emotions.


Athletes’ Self and Other Regulation 545

JSEP Vol. 38, No. 6, 2016

MethodParticipants and ProcedureThis study used a cross-sectional design with 465 team sport athletes recruited from 38 competitive sport teams in Ontario and British Columbia in Canada (i.e., hockey, volleyball, soccer, basketball, baseball, softball, lacrosse, rugby). The study was approved by University Research Ethics Boards in both provinces and all participants pro-vided informed consent before completing the survey. Researchers contacted coaches to arrange a time to meet with teams to inform athletes about the study; 1 week later the researchers returned to distribute surveys to athletes. Data were collected midway through their season to ensure teams had played together for a mini-mum of 1 month, and data were collected before a team practice to ensure recent successes or losses did not bias or unilaterally influence participants’ responses. Team sizes ranged from 10 to 30 athletes, and the response rates ranged from 35% to 100% (M = 83%). We had 15 teams in which all the athletes completed surveys; only two teams had less than a 50% response rate: These two teams had 11/30 and 7/20 athletes who completed the surveys. We included all the teams in the analyses because retaining only the teams with high participation rates would produce inflated effect size estimates and decreased statistical power (Hirschfeld, Cole, Berneth, & Rizzuto, 2013). Fourteen participants were removed as they did not complete the surveys or more than 50% of their data were missing, leaving 451 athletes in the final sample (247 males, 204 females, M age = 16.3, SD = 1.02) who had an average of 7.82 (SD = 3.17) years of competitive experience in their sport. Thirty-six of the teams competed at A, AA, or AAA levels of competition (elite/competitive at provincial or regional level) while the remaining two teams competed at B and C levels (competitive at regional or local level).

MeasuresIntrinsic and Extrinsic Emotion Regulation. Athletes completed the intrinsic and extrinsic self-regulation subscales of the Emotion Regulation of Others and Self scale (EROS; Niven et al., 2011). The intrinsic subscales measure individuals’ efforts to improve (e.g., “I thought about my positive characteristics to make myself feel better”; six items) and worsen their own emotions (e.g., “I looked for problems in my current situation to make myself feel worse”; four items). Athletes rated each item on a Likert scale to indicate the extent to which they used each strategy to influence the way they felt over the past 2 weeks (1 = not at all, 5 = a great deal). Previously reported Cronbach’s alphas indicate acceptable levels of internal consistency for scores using the intrinsic ERS subscales: self-improving ERS (α = .87) and self-worsening ERS (α = .83; Niven et al., 2011). Cronbach’s alphas for our sample were as follows: self-improving ER (α = .81) and self-worsening ERS (α = .77).

The extrinsic subscales of the EROS contained six items designed to measure efforts to try to improve the emotions of teammates (e.g., “I listened to a teammate’s problems to improve how they felt”) and three items to measure efforts to try to worsen teammates’ emotions (e.g., “I acted annoyed towards a teammate to try to make them feel worse”). The original measure (Niven et al., 2011) asks, “To what extent have you used the following strategies to influence the way someone else feels over the past two weeks?” which was adapted to ask “To what extent have you used the following strategies to influence the way one of your teammates feels over the past two weeks?” Neither the intrinsic nor the extrinsic subscale was specific to sport, and this decision was motivated to ensure that the scale could be widely used in group-based settings requiring team work. Athletes rated each item on a Likert scale to indicate the extent to which they used each strategy to influence the way a teammate felt over the past two weeks (1 = not at all, 5 = a great deal). The intrinsic and extrinsic subscales of the EROS measure have been tested among partici-pants ranging from 10 to 67 years of age (Niven et al., 2011). Previously reported Cronbach’s alphas indicate acceptable levels of internal consistency for scale scores: other-improving ERS (α = .89) and other-worsening ERS (α = .74; Niven et al., 2011). Cronbach’s alphas for our sample were as follows: other-improving ERS (α = .88) and other-worsening ERS (α = .77). A fuller examination of the factor structure obtained in this study is presented in a supplementary file.

Peer Motivational Climate. The Peer Motivational Climate in Youth Sport Questionnaire (PMCYSQ; Ntou-manis & Vazou, 2005) consists of 21 items measuring athletes’ perceptions of the motivational climate within the team. Consistent with Ntoumanis and Vazou (2005), the PMCYSQ contains two higher order factors that each consist of lower order factors: Task peer climate consists of the lower order factors improvement, relatedness sup-port, and effort, and ego peer climate consists of the lower order factors intrateam competition/ability and intrateam conflict. Following the statement “On this team, most athletes. . . ,” an example of items reflecting the task peer climate was “Help each other improve” while an example of items reflecting the ego peer climate was “Criticize their teammates when they make mistakes.” Athletes rated the items on a scale ranging from 1 = strongly disagree to 7 = strongly agree. Previously reported Cronbach’s alphas indicate acceptable levels of internal consistency for scale scores: task peer climate (α = . 92), ego peer climate (α = .75; Ntoumanis et al., 2012). Cronbach’s alphas for our sample were as follows: task peer climate (α = .90) and ego peer climate (α = .80).Sport Enjoyment and Commitment. Sport enjoyment and commitment were measured using four items from Scanlan et al.’s (1993) measure of enjoyment (e.g., “Do you have fun playing in this program this season?”) and four items to measure athletes’ commitment (e.g., “How


546 Tamminen et al.

JSEP Vol. 38, No. 6, 2016

dedicated are you to playing with this team?”) rated on a Likert scale (1 = not at all, 5 = very much). Scanlan et al. (1993) reported acceptable internal consistency levels for scale scores: sport enjoyment (α = .90) and sport commitment (α = .88). Cronbach’s alphas for our sample were as follows: sport enjoyment (α = .94) and sport commitment (α =.83).

Plan of Analyses

The data were screened for missing responses and outli-ers and examined to test the assumptions of normality, linearity, and homoscedasticity (Tabachnick & Fidell, 2014). Less than 5% of the data were missing, and miss-ing values were handled with median replacement (using the participant’s median values for the other items on each subscale; Tabachnick & Fidell, 2014). Two subscales (self-worsening and other-worsening ERS) were posi-tively skewed, and two subscales (sport enjoyment and commitment) were negatively skewed. The correlations were examined with the raw and transformed variables, and the results were comparable; the untransformed results are presented for ease of interpretation. Descrip-tive statistics and bivariate correlations between variables at Level 1 and at Level 2 are reported in Table 1.

A multilevel structural equation model (MSEM) was performed to simultaneously examine the relations of person-level (Level 1) and team-level (Level 2) predictors with sport enjoyment and commitment (see Figure 1). The MSEM approach was preferable to multilevel regression because it can easily accommodate multivariate models with correlated dependent variables (e.g., Preacher, Zyphur, & Zhang, 2010). A Bayesian estimator was selected over the traditional maximum likelihood estima-tor to facilitate convergence of the model and minimize the risk of producing untrustworthy estimates (e.g., nega-tive residual variance), particularly with small samples (e.g., Hox, van de Schoot, & Matthijsse, 2012; Muthén & Asparaouhov, 2012). The Bayesian MSEM assumes that effects are random, thus enabling the effects of the Level 1 predictors to vary across the Level 2 units (i.e., teams). Random effects with the maximum likelihood estimator are obtained with computationally demanding numerical integration that often fails to converge (e.g., Preacher et al., 2010). Not including the random effects in a model could potentially increase the risk of biasing the estimate of the fixed effects, their significance testing, and their effect size (e.g., Nezlek, 2012). Bayesian MSEM was deemed to offer a flexible approach to examine the Level 1, Level 2, and cross-level interaction effects proposed in the MERAT. The analyses were performed with Mplus 7.4 with the default specification to use uninformative priors to ensure that results are only a mirror of the current data. Analyses were conducted using manifest variables. All predictors were grand mean centered (Hofmann & Gavin, 1998). Intrinsic and extrinsic ERS were entered at Level 1 whereas teams’ average scores for extrinsic emotion regulation (e.g., the emotion regulation climate), team task peer climate, and team ego peer climate were

calculated and entered at Level 2. The covariances among the four Level 1 predictors and the four Level 2 predictors were freely estimated, respectively. Sport enjoyment and commitment were freely correlated at both levels. The residual variance terms representing the part of person- and team-level variance not explained by predictors in the model were used to calculate a series of pseudo R2 values to determine the percentage of variance explained by predictors in the model (Hofmann, Griffin, & Gavin, 2000). Simple intercepts and slopes for significant cross-level interactions were analyzed for all the cross-level interactions that reached p < .10. The simple intercepts and slopes of a Level 1 predictor were estimated at +1 SD (high) and –1 SD (low) of the mean of the Level 2 predictor (Gaudreau, Fecteau, & Perrault, 2010). The fit indices that are used in Bayesian structural equation modeling (e.g., posterior predictive probability value, the 95% confidence interval [CI] of the difference in observed and replicated chi-square) are currently not available for multilevel models. Hence, we only reported the deviance information criteria (DIC) for which smaller values indicate better fit (e.g., Asparouhov, Muthén, & Morin, 2015). Model convergence was closely followed using the potential scale reduction convergence criteria (PSR) with values lower than 1.10 taken as the indicator that convergence has been obtained (Muthén & Asparouhov, 2012). Syntax codes used in Mplus for this analysis are reported in the supplementary file.

Results

Preliminary Analyses

Descriptive statistics and bivariate correlations at Level 1 and Level 2 are reported in Table 1. At the person level, self-improving ERS was significantly correlated with higher sport enjoyment whereas other-improving ERS was significantly correlated with higher sport enjoyment and commitment. Self-worsening ERS was significantly correlated with lower sport enjoyment whereas other-worsening ERS was significantly correlated with lower sport enjoyment and commitment. At the team level, team other-improving and team-other worsening were respectively correlated with higher and lower sport commitment but not significantly associated with sport enjoyment. Team task peer climate was significantly positively correlated with both sport enjoyment and com-mitment whereas team ego peer climate was significantly negatively correlated with sport enjoyment.

The intraclass correlation (ICC) indicated that 8% and 19% of the variability in athletes’ sport enjoyment and commitment scores can be attributed to between-teams variance (see Table 1). According to Hox (2010), ICCs of .05, .10, and .15 can be considered small, medium, and large values. As expected, the intrinsic ERS variables (i.e., self-improving ERS and self-worsening ERS) showed a small amount of between-teams variance. In contrast, the extrinsic ERS variables (i.e., other-improving ERS and other-worsening ERS) and the peer climate variables


547JSEP Vol. 38, No. 6, 2016

Tab

le 1

D

escr

ipti

ve a

nd

Co

rrel

atio

nal

Res

ult

s fo

r L

evel

1 a

nd

Lev

el 2

Var

iab

les

Varia

ble

Ran

geM

SD

ICC

12

34

56

7

Lev

el 1

a

1. S

elf-

impr

ovin

g E

RS

1–5

3.21

0.85

.040

—

2. S

elf-

wor

seni

ng E

RS

1–5

1.75

0.74

.003

–.08

—

3. O

ther

-im

prov

ing

ER

S1–

53.

290.

93.1

20.5

6**

.05

—

4. O

ther

-wor

seni

ng E

RS

1–5

1.49

0.74

.140

–.01

.21*

*–.

09—

5. T

ask

peer

clim

ate

1–7

5.30

0.96

.240

.31*

*–.

06.4

2**

–.29

**—

6. E

go p

eer

clim

ate

1–7

3.91

1.09

.260

–.03

.09

–.08

.35*

*–.

17**

—

7. S

port

enj

oym

ent

1–5

4.42

0.78

.081

.12*

–.14

**.2

0**

–.14

**.5

0**

–.04

—

8. S

port

com

mitm

ent

1–5

4.32

0.69

.195

.08

–.07

.19*

*–.

14**

.38*

*–.

06.6

2**

Varia

ble

Ran

geM

SD

12

34

5

Lev

el 2

b

1. T

eam

oth

er-i

mpr

ovin

g E

RS

1–5

3.31

0.39

—

2. T

eam

oth

er-w

orse

ning

ER

S1–

51.

480.

34–.

52**

—

3. T

eam

task

pee

r cl

imat

e1–

75.

260.

55.4

2**

–.52

**—

4. T

eam

ego

pee

r cl

imat

e1–

73.

870.

62–.

54**

.61*

*–.

35*

—

5. S

port

enj

oym

ent

1–5

4.41

0.12

–.22

–.12

.35*

–.35

*—

6. S

port

com

mitm

ent

1–5

4.32

0.21

.61*

*–.

62**

.61*

*–.

14.1

4

Not

e. I

CC

= in

trac

lass

cor

rela

tion;

ER

S =

em

otio

n re

gula

tion

stra

tegi

es.

a N =

451

ath

lete

s. b

N =

38

team

s.

*p <

.05.

**p

< .0

1.


548 Tamminen et al.

JSEP Vol. 38, No. 6, 2016

exhibited larger between-teams variance. The findings were consistent with our decision to include the latter variables as team-level predictors because it can be assumed that reports of extrinsic ERS and perceptions of the peer motivational climate entail overt behaviors that may be observed, repeated, and shared among members of a sport team.

Bayesian Multilevel Structural Equation Model

Model convergence was attained with a PSR value smaller than 1.10 after 600 iterations and a value nearing 1.00 from the 12,000th to the 60,000th iterations. The deviance crite-rion of the model and the effective number of parameters (DIC = 5769.99, pD = 90.059) were smaller than a model that did not contain the cross-level interactions (DIC = 5991.30, pD = 130.39). The main effects of the Level 1 and the Level 2 predictors are reported in Figure 2, and the cross-level interactions (i.e., effect of a Level 2 predictor

on the effect of a Level 1 predictor) are reported in Table 2. At the person level, self-worsening ERS was negatively associated with sport enjoyment whereas other-improving ERS was positively associated with both sport enjoyment and commitment. The random effect of each of the Level 1 predictors was statistically significant, and the predic-tors explained 10.7% and 6.4% of the between-persons variance of sport enjoyment and commitment. At the team level, the team task peer climate was positively associated with both sport enjoyment and commitment whereas the team ego peer climate was positively associated with sport commitment. Team other-improving ERS was negatively associated with sport enjoyment and marginally signifi-cantly associated with higher sport commitment. Finally, team-other worsening ERS was marginally significantly associated with lower sport commitment. The predictors explained 71.7% and 82.4% of the between-teams variance of sport enjoyment and commitment.

As shown in Table 2, four Level 1 effects were sub-stantially (p < .10) predicted/moderated by one of the

Figure 2 — Bayesian multilevel structural equation model of sport enjoyment and commitment. All parameters were unstandard-ized. Random effects are reported in parentheses. Covariance between the predictors were freely estimated but not depicted. Level 1 effects in bold were moderated by a Level 2 team-level predictor, but the cross-level interactions are not depicted in this figure. †p < .10. *p < 01. **p < .01. All probability values are one-tailed.



JSEP Vol. 38, No. 6, 2016

Table 2 Cross-Level Interactions From the Bayesian Multilevel Structural Equation Model of Sport Enjoyment and Commitment

Level 1 effects/team moderators

Sport Enjoyment Sport Commitment

Coefficient SD p Coefficient SD p

Self-improving ERS

Team other-improving ERS 0.095 0.216 .331 0.060 0.184 .371

Team other-worsening ERS –0.230 0.266 .190 –0.108 0.222 .311

Team task peer climate –0.137 0.168 .207 –0.224 0.145 .054†

Team ego peer climate 0.016 0.141 .454 –0.129 0.118 .138

Self-worsening ERS

Team other-improving ERS –0.064 0.184 .361 –0.040 0.177 .410

Team other-worsening ERS 0.373 0.249 .069† 0.297 0.238 .106

Team task peer climate 0.146 0.134 .137 0.090 0.125 .234

Team ego peer climate –0.230 0.127 .039* –0.231 0.123 .035*

Other-improving ERS

Team other-improving ERS –0.062 0.176 .360 0.004 0.167 .489

Team other-worsening ERS –0.064 0.220 .385 0.064 0.205 .373

Team task peer climate 0.108 0.116 .169 0.083 0.111 .218

Team ego peer climate –0.023 0.118 .422 –0.020 0.113 .429

Other-worsening ERS

Team other-improving ERS –0.235 0.215 .132 –0.075 0.204 .355

Team other-worsening ERS 0.042 0.267 .436 –0.309 0.252 .105

Team task peer climate –0.114 0.167 .242 –0.183 0.156 .120

Team ego peer climate –0.178 0.146 .108 0.070 0.140 .305

Grand mean estimate (intercept) 4.378 0.050 .001** 4.297 0.049 .001**

Note. ERS = emotion regulation strategies.

†p < .10. *p < .05. **p < .01. All probability values are one-tailed.

Level 2 predictors. These four cross-level interactions were probed using simple intercepts and slopes shown in Figure 3.

First, self-worsening ERS was significantly nega-tively related to enjoyment in teams with low team other-worsening ERS, (intercept = 4.29; slope = –0.29, t[33] = –5.04, p < .01) but not in teams with high team other-worsening ERS (intercept = 4.47; slope = 0.01, t[33] = 0.02 p > .05; see Figure 3A). These results did not support our hypotheses, and they suggest that the use of self-worsening ERS was negatively associated with enjoyment within teams in which the team-level other-worsening ERS was low, and the association was nonsig-nificant when team-level other-worsening ERS was high.

Second, self-worsening ERS was significantly nega-tively related to enjoyment in teams with high team ego peer climates (intercept = 4.30; slope = –0.31, t[33] = –5.37, p < .01) but not in teams with low team ego peer climates (intercept = 4.46; slope = 0.02, t[33] = 0.37 p > .05; see Figure 3B). These results provide support for our hypotheses regarding the moderating role of the team ego peer climate and suggest that athletes’ use of regulation

strategies to make themselves feel worse was negatively associated with enjoyment within teams with a high ego-involving peer motivational climate that is competitive and driven by normative comparisons. However, this association was nonsignificant when the ego-involving peer motivational climate is low.

Third, self-improving ERS was significantly posi-tively related to commitment in teams with low team task peer climate (intercept = 4.06; coefficient = 0.15, t[33] = 2.76 p < .05) but negatively related to commitment in teams with high team task peer climate (intercept = 4.53; coefficient = –0.14, t[33] = –2.57, p < .05; see Figure 3C). These results were contrary to our hypotheses and suggest a compensatory mechanism whereby efforts to improve one’s own emotions were positively associated with commitment in teams that were not supportive of individual mastery and self-improvement (e.g., in teams with a low task peer climate). However, efforts to improve one’s own emotions were negatively associated with commitment in teams that were supportive of individual athletes’ mastery and self-improvement (e.g., in teams with a high task peer climate).


550 JSEP Vol. 38, No. 6, 2016

Fig

ure

3 —

Sim

ple

slop

es f

or c

ross

-lev

el i

nter

actio

ns. (

A)

Self

-wor

seni

ng e

mot

ion

regu

latio

n (E

R)

with

spo

rt e

njoy

men

t m

oder

ated

by

team

oth

er-w

orse

ning

ER

(A

) an

d te

am

ego

peer

clim

ate

(B).

Sel

f-im

prov

ing

ER

with

spo

rt c

omm

itmen

t mod

erat

ed b

y te

am ta

sk p

eer

clim

ate

(C).

Sel

f-w

orse

ning

ER

with

spo

rt c

omm

itmen

t mod

erat

ed b

y te

am e

go p

eer

clim

ate

(D).



JSEP Vol. 38, No. 6, 2016

Fourth, self-worsening ERS was significantly nega-tively related to commitment in teams with high team ego peer climate (intercept = 4.46; coefficient = –0.23, t[33] = –4.13, p < .01), but not in teams with low team ego peer climate (intercept = 4.14; coefficient = 0.11, t[33] = 1.95, p > .05; see Figure 3D). Similar to the findings for enjoyment and in support of our hypotheses, athletes’ use of regulation strategies to make themselves feel worse was negatively associated with commitment when used within teams in which the peer climate was competitive and driven by normative comparisons. Within teams with a low ego peer climate, athletes reported similar levels of commitment regardless of the extent to which they used self-worsening ERS.

DiscussionIn this study, we proposed a MERAT to distinguish how athletes try to regulate their own emotions and the emo-tions of their teammates. Our findings provided initial evidence that athletes’ intrinsic and extrinsic ERS were associated with sport commitment and enjoyment, and the results also provided initial evidence for the direct and moderating effects of team-level motivational and emotion regulation climates on sport enjoyment and commitment. These findings highlight the need to consider the interplay between person- and team-level factors to better explicate the role of intrapersonal and interpersonal ERS for the sport enjoyment and commitment of adolescent athletes.

Partially consistent with our hypotheses, athletes’ intrinsic self-worsening ERS appeared to be more important for (negatively) predicting enjoyment than self-improving ERS, which was not significantly associated with enjoyment. These results are consistent with the idea that efforts to worsen one’s own emotions are associated with negative affective responses in sport (Lane, Beedie, Devonport, & Stanley, 2011) and nonsport contexts (Niven et al., 2011; Niven et al., 2013). These findings also pro-vide support for the idea that efforts to worsen one’s own emotions are easier than efforts to improve one’s emotions (Niven et al., 2013; Webb, Miles, & Sheeran, 2012); thus, actions to worsen one’s own emotions may have a stronger impact on affective and motivational outcomes in sport than improving one’s own emotions. However, the concept of enjoyment offers a global and unspecific perspective about the feelings of pleasure and fun experienced in sport and should therefore be complemented with finer grained assessments of specific achievement-related emotions in sport in future research.

Consistent with our hypotheses, athletes’ extrinsic other-improving ERS was positively associated with their own enjoyment and commitment while other-worsening ERS was negatively associated with enjoyment. These findings support previous work by Niven, Totterdell, et al. (2012) indicating that trying to improve the emotions of others has consequences for one’s own emotional well-being. Theoretical propositions and empirical findings suggest that actions to try to improve the emo-tions of others contribute to positive affective outcomes

such as increased enthusiasm and lower anxiety (Niven, Totterdell, et al., 2012), possibly due to processes of emotional contagion (Hatfield, Cacioppo, & Rapson, 1994), vicarious affect (Zaki & Williams, 2013), or inferential processing (van Kleef & Fischer, 2016). Our findings were particularly interesting in light of the fact that athletes’ self-improving ERS was not directly asso-ciated with enjoyment or commitment, which did not support our hypotheses. These results support the value of considering emotion regulation from an interpersonal perspective as they demonstrate that efforts to try to improve the emotions of others could have affective and motivational effects that appeared to be more important than efforts to regulate their own emotions.

Our findings provide support for the direct positive effect of a peer task motivational climate at the team level on sport enjoyment and commitment, which was consistent with our hypotheses based on propositions that athletes report positive outcomes when they are on teams that have a high task peer climate (Ntoumanis et al., 2012). These findings further emphasize the importance of a team-level task peer climate as a possible protective factor for adolescent athletes’ sport enjoyment and com-mitment. The findings regarding the moderating effect of the team-level task peer climate on associations between emotion regulation and our outcomes partially supported our hypotheses. We anticipated that being in a team with high levels of task peer climate would strengthen the positive association between athletes’ self-improving ERS with commitment and enjoyment. The result for enjoyment was not significant, whereas the effect for commitment, at first glance, appeared to contradict our hypothesis. More specifically, the relation between self-improving ERS and sport commitment was positive in teams with lower task peer climate. In team with higher task peer climate, this relation was negative, but the level of sport commitment nonetheless remained higher. As shown in Figure 3C, it appears that athletes with lower use of self-improving ERS are less committed when they belong to a team in which task peer climate is low. In contrast, the disadvantage associated with being in such a team shrinks among the athletes who are making higher use of self-improving ERS. This result is interest-ing because it does offer some support for the idea that efforts to improve one’s emotions act as a compensatory mechanism to promote positive outcomes for athletes in teams that are less oriented toward the pursuit of task mastery. This finding is important because it also helps qualify the nonsignificant path from self-improving ERS and sport commitment (see Figure 2). Overall, it appears that self-improving ERS matters the most and should be promoted when athletes are participating in teams in which learning, improving, and effort are not emphasized, valued, and reinforced by teammates.

We did not expect that team-level ego peer climate at Level 2 would have a direct positive association with sport commitment while showing a null association with enjoyment; this finding may be attributable to the com-petitive level and age group of the athletes in the study.


552 Tamminen et al.

JSEP Vol. 38, No. 6, 2016

The majority of the teams were among the highest levels of competition within their sport, and older youth athletes on competitive teams have been reported to have a strong ego goal orientation (Vazou, Ntoumanis, & Duda, 2006). Our findings may also be attributable to a third-variable effect, wherein athletes on teams with higher ego peer climates could report higher commitment if they are experiencing success in achieving their goals. Further research is required to determine whether these findings would generalize to lower levels of competition or if they may be explained by team success. At a first glance, the findings suggest that the presence of a high ego peer climate may not necessarily be detrimental for sport enjoyment and commitment. However, when considering the moderating effect of the motivational climate at the team level, a high ego peer climate within the team may exacerbate the negative association between athletes’ self-worsening ERS with commitment and enjoyment. Thus, athletes who engage in greater self-worsening ERS within teams with a high ego peer climate may report lower enjoyment and commitment than athletes engaging in such behaviors within teams with a lower ego peer climate. These findings highlight the importance of examining both person- and team-level factors in studies examining youth athletes’ experiences in sport.

In terms of the emotion regulation climate within teams, we expected that teams characterized by interac-tions intended to improve the emotions of teammates would be associated with higher sport enjoyment and commitment. Contrary to our hypotheses, however, team-level other-improving ERS was only marginally associated with commitment and it was negatively associated with sport enjoyment, although person-level other-improving ERS was positively associated with enjoyment and commitment. These divergent results suggest that person-level and team-level emotion regula-tion potentially represent two distinct phenomena. At the person level, trying to improve the emotions of teammates appears to be a valuable strategy to improve one’s sport enjoyment and commitment. However, a team-level tendency to use extrinsic other-improving ERS while athletes are also reporting lower sport enjoyment could be explained by a third-variable effect—that is, teams char-acterized by an emotion regulation climate where athletes are trying to improve the emotions of one another might reflect a greater need for the team to cope with adversity and performance difficulties. As such, the performance difficulties of some teams could potentially explain their lower levels of sport enjoyment and also their greater reliance on extrinsic emotion regulation efforts to try to improve the emotions of teammates. Only one of the cross-level effects regarding team-level interpersonal emotion regulation was significant: On teams with low other-worsening ERS, athletes’ higher self-worsening ERS was associated with lower enjoyment. This sug-gests that athletes who use self-worsening ERS may be vulnerable to experiencing lower sport enjoyment even when they are on teams where their teammates are not trying to make them feel bad.

Although additional research is needed to test the direction and causality of these effects, some applied implications can nonetheless be provided based on the findings from this study. First, engaging in actions to improve the emotions of teammates may contribute to improved enjoyment and commitment among adolescent athletes. Hence, parents and coaches should emphasize the positive effects of engaging in these types of posi-tive interactions. The team-level task peer climate was also positively associated with athletes’ enjoyment and commitment, which provides further support for encour-aging athlete behaviors that promote effort and personal improvement among teammates. Furthermore, the team-level ego peer climate appeared to exacerbate the nega-tive associations between athletes’ emotion regulation with enjoyment and commitment. Thus, coaches should seek to minimize or forestall actions within teams that emphasize social comparisons among teammates.

Limitations and Future Research Directions

One limitation of this study was the sample size; while we had a large sample of athletes at Level 1, the Level 2 sample size of 38 teams may have been a limitation for this work. Our cross-sectional design limited our ability to draw causal inferences from our data, and it will be important for future research to adopt longitudinal and experimental designs to test the direction of the associa-tions identified in this study. Additional work with larger samples is needed to determine whether the interaction effects identified in this study can be reproduced in other samples and to assess the practical and theoretical value of the interactions. We also noted additional cross-level interactions that had probability values slightly higher than .10 (see Table 2), which would be valuable to exam-ine in future research. In addition, we sampled team sport athletes, and it would be important to examine whether interpersonal emotion regulation differs among athletes in sports that vary in their structural interdependence (Evans, Eys, & Bruner, 2012). We also acknowledge that the EROS scale (Niven et al., 2011) has yet to undergo full validation, but the results of the Bayesian confirmatory factor analyses reported in the supplementary file none-theless offered preliminary support for its factor structure.

As shown in previous research (Scanlan et al., 1993), enjoyment may be a mediator of commitment in sport. The present study was the first to demonstrate the effect self- and interpersonal emotion regulation of athletes on enjoyment and commitment as dependent variables. A larger and independent sample would be needed, ideally with a longitudinal panel design, to replicate the original effects reported in this study while offering a robust test of mediation. Multilevel mediated-moderation structural equation modeling would be desirable to offer an optimal test of such a complex model.

Finally, we did not explore the purpose or func-tion of interpersonal emotion regulation within teams. Expressions of anger, frustration, pride, and appreciation



JSEP Vol. 38, No. 6, 2016

can have adaptive functions for relationships within groups by strengthening social relationships and enhanc-ing commitment and belonging, and the expression of some emotions may communicate to team members that their behavior is not acceptable and that corrections are required (Fischer & Manstead, 2008; von Scheve & Ismer, 2013). Accordingly, efforts to try to regulate a teammate’s emotions may serve instrumental functions (Tamir, 2016) to reinforce behaviors among teammates intended to improve the team’s performance. Qualita-tive research has found evidence that athletes regulated their own and others’ emotions to improve the team’s performance and to maintain teammate relationships (e.g., Tamminen & Crocker, 2013). However, little research has systematically explored the factors that determine when and why people regulate others’ emo-tions and when each approach is more or less useful (Zaki & Williams, 2013). Identifying the conditions under which athletes regulate their teammates’ emotions will provide useful information for developing team emotion-regulation interventions.

Conclusion

This study examined athletes’ efforts to try to improve and worsen their own and their teammate’s emotions, capturing different types of interpersonal emotion regu-lation interactions that may occur in team sport settings. Examining athletes’ interpersonal emotion regulation within a team sport context is a novel addition to the sport psychology literature as it broadens the scope of emotion regulation research from an intrapersonal to an interpersonal perspective. At the person level, athletes’ attempts to regulate their own emotions and the emo-tions of their teammates were associated with their own enjoyment and commitment. Team-level factors such as the peer motivational climate and the team’s overall use of interpersonal ERS were also associated either directly (Level 2 effects) or indirectly (cross-level effects) with athletes’ own enjoyment and commitment. Taken together, these findings highlight the importance of taking into account the team-level emotion regulation climate and peer motivational climate as potential moderators of the associations between interpersonal emotion regula-tion with affective and motivational outcomes in sport.

Acknowledgments

This research was supported by a postdoctoral fellowship awarded to the first author by the Social Sciences and Humani-ties Research Council of Canada.

References

Aldao, A., Nolen-Hoeksema, S., & Schweizer, S. (2010). Emotion-regulation strategies across psychopathology: A meta-analytic review. Clinical Psychology Review, 30, 217–237. PubMed doi:10.1016/j.cpr.2009.11.004

Ames, C. (1992). Classrooms: Goals, structures, and student motivation. Journal of Educational Psychology, 84, 261–271. doi:10.1037/0022-0663.84.3.261

Asparouhov, T., Muthén, B., & Morin, A.J.S. (2015). Bayes-ian structural equation modeling with cross-loadings and residual covariances: Comments on Stromeyer et al. Journal of Management, 41(6), 1561–1577. doi:10.1177/0149206315591075

Balk, Y.A., Adriaanse, M.A., de Ridder, D.T.D., & Evers, C. (2013). Coping under pressure: Employing emotion regu-lation strategies to enhance performance under pressure. Journal of Sport & Exercise Psychology, 35, 408–418. PubMed doi:10.1123/jsep.35.4.408

Boiché, J.C.S., & Sarrazin, P.G. (2009). Proximal and distal factors associated with dropout versus maintained partici-pation in organized sport. Journal of Sports Science, and Medicine, 8, 9–16. PubMed

Chan, D.K., Lonsdale, C., & Fung, H.H. (2012). Influences of coaches, parents, and peers on the motivational patterns of child and adolescent athletes. Scandinavian Journal of Medicine & Science in Sports, 22, 558–568. PubMed doi:10.1111/j.1600-0838.2010.01277.x

Compas, B.E., Jaser, S.S., Dunbar, J.P., Watson, K.H., Bettis, A.H., Gruhn, M.A., & Williams, E.K. (2014). Coping and emotion regulation from childhood to early adulthood: Points of convergence and divergence. Australian Journal of Psychology, 66, 71–81. PubMed doi:10.1111/ajpy.12043

Crocker, P.R.E., Tamminen, K.A., & Gaudreau, P. (2015). Conceptual advances in stress and coping in sport. In S. Hanton & S. Mellalieu (Eds.), Contemporary advances in sport psychology: A review (pp. 28–67). New York, NY: Routledge.

Eisenberg, N., Fabes, R.A., Guthrie, I.K., & Reiser, M. (2000). Dispositional emotionality and regulation: Their role in predicting quality of social functioning. Journal of Per-sonality and Social Psychology, 78, 136–157. PubMed doi:10.1037/0022-3514.78.1.136

Evans, M.B., Eys, M.A., & Bruner, M.W. (2012). Seeing the “we” in “me” sports: The need to consider individual sport team environments. Canadian Psychology/Psychologie canadienne, 53, 301–308. doi:10.1037/a0030202

Fischer, A.H., & Manstead, A.S.R. (2008). Social functions of emotion. In M. Lewis, J.M. Haviland-Jones, & J. Feldman-Barrett (Eds.), Handbook of emotions (3rd ed., pp. 456–468). New York, NY: Guilford Press.

Fredrickson, B.L. (2001). The role of positive emotions in posi-tive psychology. The American Psychologist, 56, 218–226. PubMed doi:10.1037/0003-066X.56.3.218

Friesen, A.P., Devonport, T.J., Sellars, C.H., & Lane, A.M. (2013). A narrative account of decision-making and interpersonal emotion regulation using a social-functional approach to emotions. International Journal of Sport and Exercise Psychology, 11(2), 203–214. doi:10.1080/1612197X.2013.773664

Friesen, A.P., Lane, A.M., Devonport, T.J., Sellars, C.N., Stan-ley, D.N., & Beedie, C.J. (2013). Emotion in sport: Con-sidering interpersonal regulation strategies. International Review of Sport and Exercise Psychology, 6, 139–154. doi:10.1080/1750984X.2012.742921


http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=20015584&dopt=Abstracthttp://dx.doi.org/10.1016/j.cpr.2009.11.004http://dx.doi.org/10.1037/0022-0663.84.3.261http://dx.doi.org/10.1177/0149206315591075http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=23966450&dopt=Abstracthttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=23966450&dopt=Abstracthttp://dx.doi.org/10.1123/jsep.35.4.408http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=24150550&dopt=Abstracthttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=21410540&dopt=Abstracthttp://dx.doi.org/10.1111/j.1600-0838.2010.01277.xhttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=24895462&dopt=Abstracthttp://dx.doi.org/10.1111/ajpy.12043http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=10653511&dopt=Abstracthttp://dx.doi.org/10.1037/0022-3514.78.1.136http://dx.doi.org/10.1037/a0030202http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=11315248&dopt=Abstracthttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=11315248&dopt=Abstracthttp://dx.doi.org/10.1037/0003-066X.56.3.218http://dx.doi.org/10.1080/1612197X.2013.773664http://dx.doi.org/10.1080/1612197X.2013.773664http://dx.doi.org/10.1080/1750984X.2012.742921http://dx.doi.org/10.1080/1750984X.2012.742921

554 Tamminen et al.

JSEP Vol. 38, No. 6, 2016

Gaudreau, P., & Blondin, J.-P. (2004). Different athletes cope differently during a sport competition: A cluster analysis of coping. Personality and Individual Differences, 36, 1865–1877. doi:10.1016/j.paid.2003.08.017

Gaudreau, P., Fecteau, M.-C., & Perrault, S. (2010). Multi-level modeling of dyadic data in sport sciences: Con-ceptual, statistical, and practical issues. Measurement in Physical Education and Exercise Science, 14, 29–50. doi:10.1080/10913670903455017

Gaudreau, P., Nicholls, A., & Levy, A.R. (2010). The ups and downs of coping and sport achievement: An episodic process analysis of within-person associations. Journal of Sport & Exercise Psychology, 32, 298–311. PubMed doi:10.1123/jsep.32.3.298

Gross, J.J. (1998). The emerging field of emotion regulation: An integrative review. Review of General Psychology, 2(3), 271–299. doi:10.1037/1089-2680.2.3.271

Harwood, C.G., Keegan, R.J., Smith, J.M.J., & Raine, A.S. (2015). A systematic review of the intrapersonal correlates of motivational climate perceptions in sport and physical activity. Psychology of Sport and Exercise, 18, 9–25. doi:10.1016/j.psychsport.2014.11.005

Hatfield, E., Cacioppo, J.T., & Rapson, R.L. (1994). Emotional contagion. New York, NY: Cambridge University Press.

Hirschfeld, R.R., Cole, M.S., Berneth, J.B., & Rizzuto, T.E. (2013). Voluntary survey completion among team mem-bers: Implications of noncompliance and missing data for multilevel research. The Journal of Applied Psychology, 98, 454–468. PubMed doi:10.1037/a0031909

Hofmann, D.A., & Gavin, M.B. (1998). Centering decisions in hierarchical linear models: Implications for research in organizations. Journal of Management, 24, 623–641. doi:10.1177/014920639802400504

Hofmann, D.A., Griffin, M.A., & Gavin, M.B. (2000). The application of hierarchical linear modeling to organiza-tional research. In K.J. Klein & S.W.J. Kozlowski (Eds.), Multilevel theory, research, and methods in organiza-tions: Foundations, extensions, and new directions (pp. 467–511). San Francisco, CA: Jossey-Bass.

Hox, J. (2010). Multilevel analysis: Techniques and applica-tions. New York, NY: Routledge.

Hox, J.J.C.M., van de Schoot, R., & Matthijsse, S. (2012). How few countries will do? Comparative survey analysis from Bayesian perspective. Survey Research Methods, 6(2), 87–93.

Lane, A.M., Beedie, C.J., Devonport, T.J., & Stanley, D.M. (2011). Instrumental emotion regulation in sport: Relation-ships between beliefs about emotion and emotion regula-tion strategies used by athletes. Scandinavian Journal of Medicine & Science in Sports, 21, e445–e451. PubMed doi:10.1111/j.1600-0838.2011.01364.x

MacDonald, D.J., Côté, J., Eys, M., & Deakin, J. (2011). The role of enjoyment and motivational climate in relation to the personal development of team sport athletes. The Sport Psychologist, 25, 32–46. doi:10.1123/tsp.25.1.32

Martinent, G., Campo, M., & Ferrand, C. (2012). A descriptive study of emotional process during competition: Nature, frequency, direction, duration and co-occurrence of dis-crete emotions. Psychology of Sport and Exercise, 13, 142–151. doi:10.1016/j.psychsport.2011.10.006

Mellalieu, S.D., Hanton, S., & Shearer, D.A. (2008). Hearts in the fire, heads in the fridge: A qualitative investigation into the temporal patterning of the precompetitive psychologi-cal response in elite performers. Journal of Sports Sciences, 26, 811–824. PubMed doi:10.1080/02640410701790787

Muthén, B., & Asparouhov, T. (2012). Bayesian structural equation modeling: A more flexible representation of substantive theory. Psychological Methods, 17, 313–335. PubMed doi:10.1037/a0026802

Nezlek, J.B. (2012). Multilevel modeling analyses of diary-style data. In M.R. Mehl & T.S. Conner (Eds.), Handbook of research methods for studying daily life (pp. 357–383). New York, NY: Guilford Press.

Niven, K., Holman, D., & Totterdell, P. (2012). How to win friendship and trust by influencing people’s feelings: An investigation of interpersonal affect regulation and the quality of relationships. Human Relations, 65, 777–805. doi:10.1177/0018726712439909

Niven, K., Totterdell, P., & Holman, D. (2009). A classifica-tion of controlled interpersonal affect regulation strate-gies. Emotion (Washington, D.C.), 9, 498–509. PubMed doi:10.1037/a0015962

Niven, K., Totterdell, P., Holman, D., & Headley, T. (2012). Does regulating others’ feelings influence people’s own affec-tive well-being? The Journal of Social Psychology, 152, 246–260. PubMed doi:10.1080/00224545.2011.599823

Niven, K., Totterdell, P., Miles, E., Webb, L., & Sheeran, P. (2013). Achieving the same for less: Improving mood depletes blood glucose for people with poor (but not good) emotion control. Cognition and Emotion, 27, 133–140. PubMed doi:10.1080/02699931.2012.679916

Niven, K., Totterdell, P., Stride, C.B., & Holman, D. (2011). Emotion Regulation of Others and Self (EROS): The development and validation of a new individual difference measure. Current Psychology (New Brunswick, N.J.), 30, 53–73. doi:10.1007/s12144-011-9099-9

Nolen-Hoeksema, S., & Aldao, A. (2011). Gender and age dif-ferences in emotion regulation strategies and their relation-ship to depressive symptoms. Personality and Individual Differences, 51, 704–708. doi:10.1016/j.paid.2011.06.012

Ntoumanis, N., Taylor, I.M., & Thogersen-Ntoumani, C. (2012). A longitudinal examination of coach and peer motivational climates in youth sport: Implications for moral attitudes, well-being, and behavioral investment. Developmental Psychology, 48(1), 213–223. PubMed doi:10.1037/a0024934

Ntoumanis, N., & Vazou, S. (2005). Peer motivational climate in youth sport: Measurement development and validation. Journal of Sport & Exercise Psychology, 27, 432–455. doi:10.1123/jsep.27.4.432

Papaioannou, A., Marsh, H.W., & Theodorakis, Y. (2004). A multilevel approach to motivational climate in physical education and sport settings: An individual or a group level construct? Journal of Sport & Exercise Psychology, 26, 90–118. doi:10.1123/jsep.26.1.90

Preacher, K.J., Zyphur, M.J., & Zhang, Z. (2010). A general multilevel SEM framework for assessing multilevel mediation. Psychological Methods, 15, 209–233. PubMed doi:10.1037/a0020141


http://dx.doi.org/10.1016/j.paid.2003.08.017http://dx.doi.org/10.1080/10913670903455017http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=20587819&dopt=Abstracthttp://dx.doi.org/10.1123/jsep.32.3.298http://dx.doi.org/10.1037/1089-2680.2.3.271http://dx.doi.org/10.1016/j.psychsport.2014.11.005http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=23458334&dopt=Abstracthttp://dx.doi.org/10.1037/a0031909http://dx.doi.org/10.1177/014920639802400504http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=21819448&dopt=Abstracthttp://dx.doi.org/10.1111/j.1600-0838.2011.01364.xhttp://dx.doi.org/10.1123/tsp.25.1.32http://dx.doi.org/10.1016/j.psychsport.2011.10.006http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=18569547&dopt=Abstracthttp://dx.doi.org/10.1080/02640410701790787http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=22962886&dopt=Abstracthttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=22962886&dopt=Abstracthttp://dx.doi.org/10.1037/a0026802http://dx.doi.org/10.1177/0018726712439909http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=19653772&dopt=Abstracthttp://dx.doi.org/10.1037/a0015962http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=22468424&dopt=Abstracthttp://dx.doi.org/10.1080/00224545.2011.599823http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=22712512&dopt=Abstracthttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=22712512&dopt=Abstracthttp://dx.doi.org/10.1080/02699931.2012.679916http://dx.doi.org/10.1007/s12144-011-9099-9http://dx.doi.org/10.1016/j.paid.2011.06.012http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=21787071&dopt=Abstracthttp://dx.doi.org/10.1037/a0024934http://dx.doi.org/10.1123/jsep.27.4.432http://dx.doi.org/10.1123/jsep.26.1.90http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=20822249&dopt=Abstracthttp://dx.doi.org/10.1037/a0020141


JSEP Vol. 38, No. 6, 2016

Scanlan, T.K., Carpenter, P.J., Simons, J.P., Schmidt, G.W., & Keeler, B. (1993). The sport commitment model: Measure-ment development for the youth-sport domain. Journal of Sport & Exercise Psychology, 15, 16–38. doi:10.1123/jsep.15.1.16

Smith, A.L., Ntoumanis, N., Duda, J.L., & Vansteenkiste, M. (2011). Goal striving, coping, and well-being: A prospec-tive investigation of the self-concordance model in sport. Journal of Sport & Exercise Psychology, 33, 124–145. PubMed doi:10.1123/jsep.33.1.124

Tabachnick, B.G., & Fidell, L.S. (2014). Using multivariate statistics (6th ed.). New York, NY: Allyn & Bacon.

Tamir, M. (2016). Why do people regulate their emotions? A taxonomy of motives in emotion regulation. Per-sonality and Social Psychology Review, 20, 199–222. doi:10.1177/1088868315586325

Tamminen, K.A., & Crocker, P.R.E. (2013). “I control my own emotions for the sake of the team”: Emotional self-regulation and interpersonal emotion regulation among female high-performance curlers. Psychology of Sport and Exercise, 14, 737–747. doi:10.1016/j.psych-sport.2013.05.002

Tamminen, K.A., & Gaudreau, P. (2014). Coping, social sup-port, and emotion regulation in teams. In M. Beauchamp & M. Eys (Eds.), Group dynamics in exercise and sport psychology: Contemporary themes (2nd ed., pp. 222–239). New York, NY: Routledge.

Thatcher, J., & Day, M.C. (2008). Re-appraising stress apprais-als: The underlying properties of stress in sport. Psychol-ogy of Sport and Exercise, 9, 318–335. doi:10.1016/j.psychsport.2007.04.005

Totterdell, P. (2000). Catching moods and hitting runs: Mood linkage and subjective performance in professional sport teams. The Journal of Applied Psychology, 85, 848–859. PubMed doi:10.1037/0021-9010.85.6.848

Totterdell, P., & Leach, D. (2001). Negative mood regulation expectancies and sports performance: An investigation involving professional cricketers. Psychology of Sport and Exercise, 2, 249–265. doi:10.1016/S1469-0292(01)00016-4

Ullrich-French, S., & Smith, A.L. (2006). Perceptions of relationships with parents and peers in youth sport: Independent and combined prediction of motivational outcomes. Psychology of Sport and Exercise, 7, 193–214. doi:10.1016/j.psychsport.2005.08.006

van Kleef, G.A., & Fischer, A.H. (2016). Emotional collectives: How groups shape emotions and emotions shape groups. Cognition and Emotion, 30, 3–19. PubMed doi:10.1080/02699931.2015.1081349

Vazou, S. (2010). Variation in the perceptions of peer and coach motivational climate. Research Quarterly for Exercise and Sport, 81, 199–211. PubMed doi:10.1080/02701367.2010.10599667

Vazou, S., Ntoumanis, N., & Duda, J.L. (2006). Predicting young athletes’ motivational indices as a function of their perceptions of the coach- and peer-created climate. Psy-chology of Sport and Exercise, 7, 215–233. doi:10.1016/j.psychsport.2005.08.007

von Scheve, C., & Ismer, S. (2013). Towards a theory of collective emotions. Emotion Review, 5, 406–413. doi:10.1177/1754073913484170

Wagstaff, C.R.D. (2014). Emotion regulation and sport per-formance. Journal of Sport & Exercise Psychology, 36, 401–412. PubMed doi:10.1123/jsep.2013-0257

Webb, T.L., Miles, E., & Sheeran, P. (2012). Dealing with feeling: A meta-analysis of the effectiveness of strategies derived from the process model of emotion regulation. Psychologi-cal Bulletin, 138, 775–808. PubMed doi:10.1037/a0027600

Zaki, J., & Williams, W.C. (2013). Interpersonal emotion regula-tion. Emotion (Washington, D.C.), 13, 803–810. PubMed doi:10.1037/a0033839


http://dx.doi.org/10.1123/jsep.15.1.16http://dx.doi.org/10.1123/jsep.15.1.16http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=21451174&dopt=Abstracthttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=21451174&dopt=Abstracthttp://dx.doi.org/10.1123/jsep.33.1.124http://dx.doi.org/0.1177/1088868315586325http://dx.doi.org/10.1016/j.psychsport.2013.05.002http://dx.doi.org/10.1016/j.psychsport.2013.05.002http://dx.doi.org/10.1016/j.psychsport.2007.04.005http://dx.doi.org/10.1016/j.psychsport.2007.04.005http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=11125650&dopt=Abstracthttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=11125650&dopt=Abstracthttp://dx.doi.org/10.1037/0021-9010.85.6.848http://dx.doi.org/10.1016/S1469-0292(01)00016-4http://dx.doi.org/10.1016/j.psychsport.2005.08.006http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=26391957&dopt=Abstracthttp://dx.doi.org/10.1080/02699931.2015.1081349http://dx.doi.org/10.1080/02699931.2015.1081349http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=20527305&dopt=Abstracthttp://dx.doi.org/10.1080/02701367.2010.10599667http://dx.doi.org/10.1080/02701367.2010.10599667http://dx.doi.org/10.1016/j.psychsport.2005.08.007http://dx.doi.org/10.1016/j.psychsport.2005.08.007http://dx.doi.org/10.1177/1754073913484170http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=25226609&dopt=Abstracthttp://dx.doi.org/10.1123/jsep.2013-0257http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=22582737&dopt=Abstracthttp://dx.doi.org/10.1037/a0027600http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=24098929&dopt=Abstracthttp://dx.doi.org/10.1037/a0033839

1

Supplementary file

Interpersonal Emotional Regulation among Adolescent Athletes:

A Bayesian Multilevel Model Predicting Sport Enjoyment and Commitment

2

Main analyses

The MPLUS 7.4 syntax code for the Bayesian Multilevel Structural Equation Model are reported

in Table S1.

Complementary analyses

In this study, we adapted the Emotion Regulation of Others and Self scale (EROS) for the sport

context. We decided to investigate the factor structure of the EROS in order to ensure that each of the 19

items was significantly loading on one of the four intended latent variables. We evaluated model fit based

on traditional confirmatory factor analysis (CFA). We tested the model with the typical MLR estimator

and with the WLSMV for categorical variables (e.g., Likert scale). We also used the recently developed

Bayesian structural equation modeling (BSEM). All analyses were performed in MPLUS 7.4 and syntax

codes are available upon request.

A four-factor solution appeared defendable because it resulted in better absolute and relative fit

compared to alternative two-factor models using both MLR and WLSMV (see Table S2). However, the

four-factor solution was rejected (p < .05) on the basis of the χ2. At a first glance, this finding could

suggest the presence of substantial conceptual problems with the EROS. However, CFA imposes

restrictive constraints on parameters (i.e., cross-loadings are fixed to zero, correlations between the

residual of items are fixed to zero) that are often small but non-zero in the participants’ data (e.g., Marsh,

Morin, Parker, & Kaur, 2013). Bayesian structural equation modeling (BSEM) has recently been

proposed (Asparouhov, Muthen, & Morin, 2015; Muthén & Asparouhov, 2012) and used (e.g., De Bondt

& Van Petegem, 2015; Golay, Reverte, Rossier, Favez, & Lecerf, 2013; Jackson, Gucciardi, & Dimmock,

2014) to investigate the sources of model mispecifications in measurement models. In this study, we re-

examined the four-factor model of the EROS using the BAYES estimator in MPLUS 7.4. All analyses

were performed twice with the scores of the EROS treated as continuous and categorical (Liang & Yang,

2014; Natesan, 2015). As recommended by Muthén and Asparouhov (2012), we standardized the 19

variables prior to the analyses with continuous variables1. Variables were not standardized for the

3

analyses with categorical variables2. The loadings of the 19 items were freely estimated. The metric of the

latent variable was set by fixing the variance of each latent variable to 1. The four latent variables were

free to correlate. To facilitate convergence, we estimated the model with one Markov chain using 100000

iterations; the first half of the iterations were used as burn-in whereas the second half of the iterations

were used to create the posterior distributions needed to obtain the point estimate of the parameters (e.g.,

De Bondt & Van Petegem, 2015). Model fit was examined with the posterior predictive p value (PPP),

the 95% CI of the difference in observed and replicated χ2, and the Deviance Information Criteria (DIC).

Good fit is obtained when the PPP is non-significant, when the 95% CI of the difference in observed and

replicated χ2 has a negative lower bound, and when the DIC is the smallest (e.g., Asparouhov et al.,

2015). Model convergence of the model was closely followed using the potential scale reduction

convergence criteria (PSR) with values lower than 1.10 taken as indicator that convergence has been

obtained (Muthén & Asparouhov, 2012).

In model 1, we reproduced the restrictive constraints of a CFA model. More specifically, we

applied an informative prior distribution with a median of zero and a variance of zero on the cross-

loadings. In less technical words, this CFA-like model fixed the value of the 57 cross-loading to zero

without variance. Similarly, this CFA-like model zero-fixed the value of all the 171 correlations between

the residuals. This type of distribution – known as uninformative priors or defuse prior distribution – is

the MPLUS default in SEM model with a BAYES estimator. Therefore, it can be used to evaluate a

typical CFA model within the confines of BSEM. Consistent with the results obtained in the CFA, the fit

of model 1 was not adequate regardless of whether the variables were analyzed as continuous or

categorical variables (see Table S3). Fit indices were superior when analyzing categorical variables, thus

suggesting that it might be preferable to consider the scores of the EROS as categorical rather than

continuous.

In model 2, we applied an informative prior distribution in which the cross-loading followed a

distribution with a median of zero accompanied by a small amount of variance: N (0, .01). More

4

specifically, the center of the distribution of the cross-loading was expected to be zero with lower and

upper bounds of -.20 to +.20. In less technical words, this model allowed us to examine the tenability of a

model with small yet non-zero cross-loadings. We compared this less restrictive model with a CFA-like

model in which the cross-loading and correlated residuals were rigorously fixed at zero (model 1). All fit

indices indicated the superiority of this model using both continuous and categorical variables. Items

loaded strongly on their targeted latent variable. Using continuous variables, these primary loadings were

strong (average = .708, range = .432 to .907). In contrast, the secondary loadings were small (average =

.057, range = -.175 to .179) and six of the 57 secondary loadings were statistically significant. Using

categorical variables, these primary loadings were strong (average = .756, range = .470 to .898). In

contrast, the secondary loadings were small (average = .049, range = -.088 to .164) and seven of the 57

secondary loadings were statistically significant. However, the fit of model 2 was not adequate (see Table

S3).

Finally, we tested a third model in which the additional constraints on the zero-fixed correlated

residuals were released. Hence, this model allowed both small non-zero cross-loadings and small non-

zero correlations between the residuals of items. We followed recommendations of Muthén and

Asparouhov (2012, p. 317) and applied an Inverse Wishart (IW) prior distribution with a median of zero

and degrees of freedom that corresponded to the number of variables + 6: IW (0, 25). In less technical

words, the center of the distribution of the correlation between residuals was expected to be zero with

small variance characterized by lower and upper bounds of -.20 to +.20. We compared this less restrictive

model with previous model (model 2). All fit indices indicated the superiority of this model. Furthermore,

the fit of model 3 was adequate when analyzing both the continuous and categorical variables (see Table

S3). Using continuous variables, items loaded strongly on their targeted latent variable (see Table S4).

These primary loadings were strong (average = .731, range = .629 to .897). In contrast, the secondary

loadings were small (average = .073, range = -.177 to .180) and 41 of the 57 secondary loadings were

statistically significant. The correlation between the residuals of the items were small (average = .051,

range = -.183 to .124), but 99 of the 171 correlated residuals reached statistical significance. Using

5

categorical variables, items loaded strongly on their targeted latent variable. These primary loadings were

strong (average = .739, range = .542 to .953). In contrast, the secondary loadings were small (average =

.036, range = -.087 to .114) and 5 of the 57 secondary loadings were statistically significant. The

correlation between the residuals of the items were small (average = .061, range = -.196 to .241), but 10

of the 171 correlated residuals reached statistical significance.

Conclusion

It is not surprising to observe negligible cross-loadings and correlated residuals in the EROS

questionnaire because all items were designed to measure emotion regulation strategies. Failure to

incorporate these parameters in traditional CFA certainly explains that the four-factor solution was

rejected on the basis of the χ2 with both MLR and WLMSV. Despite the presence of cross-loadings and

correlations between the residuals, the 19 items of the EROS loaded on their targeted latent variables with

strong primary factor loadings. Composite reliability were acceptable (see Table S5). The inter-scale

correlations showed no evidence of linearity or redundancy among the four factors of the EROS (see

Table S5). Overall, the results of the BSEM with both continuous and categorical variables provided

evidence for the factor validity of the EROS data collected in the current study.

Notes

1 Results with the continuous variables were taken from the unstandardized output (model results)

because the variables were standardized before the analyses. Loadings should be interpreted as

standardized whereas correlated residuals are standardized covariance between residuals.

2 Results with the categorical variables were taken from the fully standardized output (STDYX) because

the variables were not standardized before the analyses. Loadings should be interpreted as standardized

whereas correlated residuals are standardized covariance between residuals.

6

References

Asparouhov, T., Muthen, B., & Morin, A. J. S. (2015). Bayesian Structural Equation Modeling With

Cross-Loadings and Residual Covariances: Comments on Stromeyer et al. Journal of

Management, 41(6), 1561-1577. doi:10.1177/0149206315591075

De Bondt, N., & Van Petegem, P. (2015). Psychometric evaluation of the Overexcitability Questionnaire-

Two applying Bayesian Structural Equation Modeling (BSEM) and multiple-group BSEM-based

alignment with approximate measurement invariance. Frontiers in Psychology, 6.

doi:10.3389/fpsyg.2015.01963

Golay, P., Reverte, I., Rossier, J., Favez, N., & Lecerf, T. (2013). Further insights on the French WISC–

IV factor structure through Bayesian structural equation modeling. Psychological Assessment,

25(2), 496-508. doi:10.1037/a0030676

Jackson, B., Gucciardi, D. F., & Dimmock, J. A. (2014). Toward a Multidimensional Model of Athletes'

Commitment to Coach-Athlete Relationships and Interdependent Sport Teams: A Substantive-

Methodological Synergy. Journal of Sport & Exercise Psychology, 36(1), 52-68.

Liang, X., & Yang, Y. (2014). An evaluation of WLSMV and Bayesian methods for confirmatory factor

analysis with categorical indicators. International Journal of Quantitative Research in Education,

2(1), 17-38. doi:10.1504/IJQRE.2014.060972

Marsh, H. W., Morin, A. J. S., Parker, P. D., & Kaur, G. (2013). Exploratory structural equation

modeling: An integration of the best features of exploratory and confirmatory factor analysis.

Annual Review of Clinical Psychology, 10, 85-110. doi:10.1146/annurev-clinpsy-032813-153700

Muthén, B., & Asparouhov, T. (2012). Bayesian structural equation modeling: A more flexible

representation of substantive theory. Psychological Methods, 17, 313-335. doi:10.1037/a0026802

Natesan, P. (2015). Comparing interval estimates for small sample ordinal CFA models. Frontiers in

Psychology, 6. doi:10.3389/fpsyg.2015.01599

7

Table S1

MPLUS syntax codes for the Bayesian Multilevel Structural Equation Model

TITLE: Multilevel SEM with two outcomes with Bayes estimator SIMPLE SLOPE CODING for the following cross-level interaction: Commit ON self-improve ERS @ Team task climate (see Figure 3C) VARIABLE: NAMES ARE ID Team Enjoyment Commitment ERS_Improve ERS_Worsen ERO_Improve ERO_Worsen EroImpAvg EroWorAvg TPCAvg EPCAvg; USEVARIABLES ARE team Enjoyment Commitment ERS_Improve ERS_Worsen ERO_Improve ERO_Worsen EroImpAvg EroWorAvg TPCAvg EPCAvg; WITHIN = ERS_Improve ERS_Worsen ERO_Improve ERO_Worsen; BETWEEN = EroImpAvg EroWorAvg TPCAvg EPCAvg; MISSING ARE ALL (999); CLUSTER IS Team; IDVARIABLE IS ID; DEFINE: CENTER ERS_Improve ERS_

Interpersonal Emotion Regulation Among Adolescent Athletes ......Athletes: A Bayesian Multilevel...

Documents

Transcript of Interpersonal Emotion Regulation Among Adolescent Athletes ......Athletes: A Bayesian Multilevel...