For Doc Daimler pt.5

ABarticles.docx

Home >Business & Finance homework help >Marketing homework help >For Doc Daimler pt.5

AB #1

In Norway, athletes often attend elite sport colleges facilitating the combination of education and elite sport development. These colleges are located in different parts of the country, and many athletes live far away from home with limited family support. As such, youth athletes’ ability to regulate behaviors and control thoughts, emotions, attention, and cognitive impulses is important to successfully reach long-term goals (Baumeister & Vohs, 2016b; Englert, 2016, 2017). This self-control strength will likely help athletes in their strenuous physical and mental exercises, as it makes them persistent and focused in the face of adversity and better at volitionally controlling their emotions and attention as they enter important competitions (Englert, 2017). In addition, athletes’ types of motivation stem from a variety of sources, ranging from intrinsic and autonomously motivated behaviors inspired by genuine interest and inherent satisfaction to amotivation and nonintentional behaviors performed without athletes’ control (Jordalen, Lemyre, & Durand-Bush, 2019; Ryan & Deci, 2000). This multidimensional motivation profile prompts youth athletes’ sport endeavors as they intrinsically enjoy hours of activity in their sport, while they simultaneously head toward increasing their capacity beating their own personal best as well as other opponents. These various types of motivation have been found to associate with more or less successful outcomes. The more autonomous types of motivation do energize self-regulatory behaviors, thus have been associated with energy maintenance and importantly lower levels of depletion (Jordalen, Lemyre, Solstad, & Ivarsson, 2018; Ryan & Deci, 2008). In addition, shifts in the quality of motivation across a competitive season have been found to reliably predict burnout susceptibility among elite athletes (Lemyre, Treasure, & Roberts, 2006). As such, a better understanding of the psychology behind youth athletes’ behaviors and actions requires an investigation of the interrelated processes of human motivation and cognition (Baumeister, 2016). In this study, we investigated the temporal ordering of motivation and self-control—key concepts that facilitate athletes’ development throughout a competitive season.

Many youth athletes enjoy their engagement in sport activities while simultaneously competing at high levels adapting their actions to facilitate development and perhaps, one day, become world champions. They may not only differ in quantity of motivation (i.e., total amount of motivation; Gould, 1996), but also in quality of motivation (i.e., type of motivation; Lemyre, Roberts, & Stray-Gundersen, 2007). Within self-determination theory (SDT; Ryan & Deci, 2000), human motivation is considered along a continuum composed of three types of autonomous motivation regulation (i.e., intrinsic, integrated, and identified), two types of controlled motivation regulation (i.e., introjected and external) and one type of nonregulated motivation (i.e., amotivation). Autonomously motivated athletes are driven by genuine interest that is fueled by intrinsically felt rewards, and their activities are experienced as meaningful (Ryan & Deci, 2000). Conversely, more controlled types of motivation differ in degree of self-determination, and athletes are fueled by the value and utility of workouts or are solely guided by their coaches. They may experience disinterest and no personal causation. As such, actions motivated by autonomous or controlled reasons will lead to qualitatively different experiences and performances, as motives typically guide direction, intensity, and persistence of youth athletes’ development (Cerasoli, Nicklin, & Ford, 2014). For example, athletes motivated by autonomous reasons will experience less inhibition and control, leading to increased psychological energy and prolonged engagement (Ryan & Deci, 2008). Thus, it seems that autonomous reasoning is key to healthy youth sport development (Ryan & Deci, 2000), even though athletes are likely driven by various types of motivation at once (Chantal, Guay, Dobreva-Martinova, & Vallerand, 1996). A pure autonomous motivation profile may not exist in sport (Gillet, Vallerand, & Rosnet, 2009), and research needs to go beyond the unique framework of motivation theories and emphasize cognitive psychological aspects of youth athletes’ sport performance to better understand what drives these athletes in a competitive context over time.

Self-control, the effortful subset of self-regulation, empowers cognitive competencies and directs attention away from tempting stimuli (e.g., stay up late with friends resulting in excessive fatigue negatively affecting their subsequent training) that interfere with youth athletes’ planned actions and long-term goals (e.g., training for finishing a long-distance race; Baumeister, Vohs, & Tice, 2007). This volitional ability is crucial in athletes’ everyday life, as they are confronted with challenging situations in training and competitions (e.g., maintain focus and concentration in high-pressure contexts; Englert, 2017). The strength model (Baumeister, Bratslavsky, Muraven, & Tice, 1998; Baumeister et al., 2007) has conceptually informed the majority of research on self-control. This model is based on the notion that both state and trait self-control depend on limited resources and become temporarily impaired when used sequentially. This leaves the individual in a state labeled ego depletion, where further acts of self-control are prone to failure (Baumeister et al., 2007).

Investigating self-control strength, studies have used a common research design where participants perform two separate, independent tasks, both requiring self-control (e.g., individuals who regulated their emotions to an upsetting movie experienced reduced physical stamina; Baumeister & Vohs, 2016b; see Muraven, Tice, & Baumeister, 1998). Typically, individuals spend effort on the first task, thus less remains for the second task and performance suffers. Depletion inhibits inhibition, and individuals’ top–down control is impaired allowing more bottom–up automatic responses without restraint or inhibition (i.e., this increases a range of impulsive, disinhibited behaviors, and individuals are not aware of their conscious attitudes). However, trait measures as opposed to state measures of self-control are more stable across situations and over time, and are rather individually determined (Anusic & Schimmack, 2016; de Ridder, Lensvelt-Mulders, Finkenauer, Stok, & Baumeister, 2012). As athletes’ state self-control is more susceptible to situational influences, it is assumed that high trait self-control is preferable as this makes athletes better at controlling impulses regardless of context and situation (e.g., whether in training or competitions, high- versus low-pressure situations; de Ridder et al., 2012; Englert, 2017; Tangney, Baumeister, & Boone, 2004). High trait self-control, as opposed to state self-control, may also reduce the risk of experiencing a state of ego depletion, even though both trait and state self-control are prone to depletion just as a muscle gets tired from exertion (Baumeister et al., 2007; de Ridder et al., 2012). Within the self-control literature, much attention has been given to the processes leading to depletion patterns and whether they actually cause depletion (Hagger & Chatzisarantis, 2016; Lee, Chatzisarantis, & Hagger, 2016; Tuk, Zhang, & Sweldens, 2015). For example, when depleted participants went through a brief period of mindfulness meditation or were provided incentives or choice, this helped sustain their self-control performance (Friese, Messner, & Schaffner, 2012; Moller, Deci, & Ryan, 2006; Muraven & Slessareva, 2003).

Analyzing arguments in defense or against the ego depletion effect, Friese, Loschelder, Gieh the School of Human Kinetics, University of Otta seler, Frankenbach, and Inzlicht (2019) concluded that “despite several hundred published studies, the available evidence is inconclusive” (p. 107). Attempting to explain alternative mechanisms behind self-control performance and gaining a more precise account of depletion processes, a revision of the strength model resulted in the process model of depletion (Inzlicht & Schmeichel, 2012). The process model proposes that shifts in individuals’ motivation, attention, and emotions, as well as an imbalance in internal and external motives, are associated with regulatory failures (Inzlicht, Schmeichel, & Macrae, 2014). Motivational shifts are explained in an evolutionary psychology perspective—they serve an adaptive function of redirecting behavior toward activities with increasing inherent utility. “Have to tasks” are motivated by a sense of duty or obligation and need energy to be sustained (cf. more controlled types of motivation), whereas “want to tasks” are motivated by interest and enjoyment and are more easily maintained (cf. more autonomous types of motivation; Inzlicht et al., 2014; Ryan & Deci, 2000). In this sense, motivation moderates depletion, and it seems that individuals’ motivation highly affects their self-control capacity and vice versa. According to the process model of depletion, shifts in motivation are caused by acts of self-control, and individuals subsequently prefer activities that are enjoyable and gratifying (i.e., autonomously motivated behaviors) over activities that require effort (i.e., controlled motivated behaviors). Thus, individuals’ self-control performance is impaired as a function of changed motivation in subsequent acts of self-control.

One explanation of self-control as a potential determinant of motivation quality is that individuals high in trait self-control may be more likely to possess higher autonomous reasons for their actions, as they find more interest and/or meaning in what they are doing even though they are navigating through both conflicting and sometimes tedious tasks (Converse, Juarez, & Hennecke, 2019; Holding, Hope, Verner-Filion, & Koestner, 2019). Converse et al. (2019) examined associations between trait self-control and autonomous and controlled motivation in a series of studies, and they favored the interpretation that self-control especially affects autonomous motivation. In the SDT literature, it is outlined that self-control may influence goal internalization processes, as “the types of behaviors and values that can be assimilated to the self increase with growing cognitive and ego capacities” (Deci & Ryan, 2000, p. 63). Furthermore, it is reasoned that self-control and autonomous motivation have mutual links to ease of goal pursuit and task construal (Holding et al., 2019). Longitudinally, trait self-control positively and negatively predicts autonomous and controlled motivation beyond other possible determinants of motivation quality, respectively (e.g., the Big Five personality traits; Holding et al., 2019). However, these recent studies did not provide direct causal evidence for one specific temporal ordering between these concepts, as they did not examine all possible causal associations (e.g., autonomous motivation on trait self-control; Jose, 2016).

In the SDT model of vitality, Ryan and Deci (2008) suggest that more autonomously driven self-control is less depleting as compared with when individuals are driven by more external forces, draining energy. Autonomously motivated acts of self-control are experienced as harmonious and efficient, require less inhibition, and are related to reduced temptations (e.g., see Milyavskaya, Inzlicht, Hope, & Koestner, 2015). More externally controlled acts of self-control are often associated with pressure and tension requiring greater resources by the individual. As a result, individuals will possess maintained vitality and less depletion, or contrary, significantly lower levels of vitality and higher risk of depletion, respectively. Summarized, motivational explanations of ego depletion suggest on the one hand, that youth athletes’ self-control performance is impaired as a function of changed motivation in subsequent acts of self-control (Inzlicht et al., 2014); and on the other hand, that self-control behaviors are associated with external motivational forces draining athletes’ psychological energy (Ryan & Deci, 2008). As such, these motivational explanations of self-control depletion suggest that cognition and youth athletes’ self-control direct motivation and vice versa, respectively.

Conceptually, it seems that the type rather than total amount of motivation reflected in self-control efforts is important. Though, the method of evaluating the various types of motivation has been debated, concerned to testing motivation regulations individually, using aggregates like autonomous and controlled motivation, or calculating a motivational index (e.g., Relative Autonomy Index, Self-Determination Index; Chemolli & Gagné, 2014). Calculating an index, this score may mask important results and the unique contribution of each individual regulation, thus results are less reliable (Chemolli & Gagné, 2014). Widely used, aggregates have also been described as oversimplifications that do not account for the nuanced perspective of each motivation regulation (Cerasoli et al., 2014). Finally, testing motivation regulations individually allows the different regulations yield different outcomes (Chemolli & Gagné, 2014). Autonomous versus controlling motivational incentives are likely associated with self-control behaviors and reduced or increased ego depletion, respectively. In this interaction with autonomous motivation, acts of self-control will lead to positive sport participation outcomes, such as increased well-being, whereas acts of controlled motivation will induce self-control depletion and even more severe experiences such as athlete burnout (Briki, 2016; Cresswell & Eklund, 2005; Jordalen et al., 2018; Muraven, 2008; Ryan & Deci, 2008). However, the individual contribution of each motivation regulation and the ordering of motivation and self-control constructs is unclear (e.g., Converse et al., 2019), and have not been studied explicitly in youth winter sport participants previously. Guided by the SDT and self-control literature theoretical frameworks (Baumeister et al., 1998, 2007; Inzlicht & Schmeichel, 2012; Ryan & Deci, 2000, 2008), we hypothesized that youth athletes’ self-control capacity will be more influenced by the type of motivation that inspires behavior than vice versa. We suggest that autonomous and controlled types of motivation positively and negatively predict trait self-control, respectively. The motivational regulations were evaluated individually in six different models, as this method allowed testing independent effects of each motivation regulation and the complexities of motivation associated with self-control competencies. As such, the current study investigates the reciprocal associations of various types of motivation and trait self-control over time.

Methods

Participants

A total of 321 youth winter sport athletes (173 males and 98 females; aged 16–20 years, M = 17.98, SD = 0.89) attending elite sport colleges in Norway consented to participate. Athletes competed in cross-country skiing (n = 122), biathlon (n = 64), ski jumping (n = 15), alpine skiing (n = 63), and Nordic combined (n = 7). Competitive experiences ranged from 1 to 16 years (M = 7.86 years, SD = 2.93), and athletes competed at international (n = 54), national (n = 193), or regional levels (n = 24). Descriptive information was collected at Time Point 1 (T1). Athletes who only participated at Time Points 2 (T2) and/or 3 (T3) did not report descriptive statistics (T1: n = 271; T2: n = 201; and T3: n = 197).

Measures

Motivation

A Norwegian version (Jordalen, Lemyre, & Durand-Bush, 2016) of the Sport Motivation Scale II (Pelletier, Rocchi, Vallerand, Deci, & Ryan, 2013) measured athletes’ motivational regulations, and response options ranged from 1 (does not correspond at all) to 7 (corresponds completely). Latent variable modeling was used to evaluate scale reliability (ρ; see Supplementary Table 1 [available online]; Raykov, 2009). This method offers scale reliability point estimates and identifies potentially weak components of a scale, inspecting loading, variance estimates, and SEs. Each motivation regulation included three items, and participants reported the extent to which the listed reasons for practicing their sport corresponded with their own personal reasons. The assessed regulations were intrinsic (e.g., “because it is very interesting to learn how I can improve”), integrated (e.g., “because participating in sport is an integral part of my life”), identified (e.g., “because I have chosen this sport as a way to develop myself”), introjected (e.g., “because I feel better about myself when I do”), external (e.g., “because people around me reward me when I do”), and amotivated (e.g., “it is not clear to me anymore; I don’t really think my place is in sport”).

Self-Control

A Norwegian version (Toering & Jordet, 2015) of the Brief Self-Control Scale (BSCS; Tangney et al., 2004) assessed athletes’ dispositional self-control competencies (13 items, e.g., “I am good at resisting temptations”). Items 6 and 8 were deleted due to low factor loadings (<.20; Kline, 2011). Response options ranged from 1 (not at all) to 5 (very much). Items 2, 3, 4, 5, 7, 9, 10, 12, and 13 were reverse scored (Tangney et al., 2004).

Procedures

Subsequent to approval by the Norwegian Center for Research Data, sports directors and coaches at elite sport colleges in Norway were contacted. Athletes were invited to partake if sports directors approved participation. The first author gave written and verbal presentations of the study, and visited colleges every fifth week for data collection, three times in total. That is, the author spent 1 week times 3 traveling to the various colleges in the country (Weeks 4, 9, and 14), this all athletes answered questionnaires during the same time point in their competitive season. Athletes who agreed to participate provided written informed consent prior to data collection. Answering questionnaires, athletes indicated the extent to which questions represented their thoughts or actions during practice sessions the last 5 weeks. The data collection was completed within the last 2 months of the competitive season, corresponding to a key time point where athletes competed in national and international competitions while also preparing for subsequent college exams scheduled in the off-season period. As such, athletes were challenged to demonstrate excellent competencies at different arenas simultaneously, and the resulting combination of social, psychological, and physiological demands when living far from home represented a great context to test athletes’ quality of motivation and self-control competencies (Martinent, Decret, Guillet-Descas, & Isoard-Gautheur, 2014). It is especially interesting to investigate the temporal ordering of motivation and self-control in this high-pressure context and to explore possible explanations of youth athletes’ ego depletion experiences and even more severe consequences such as athlete burnout. SurveyXact 8.0 (QuickQuest, Oslo, Norway) was used to collect data.

Analyses

First, descriptive statistical analyses were performed with JASP (version 0.8.0.0; Amsterdam, the Netherlands; see Table 1; JASP Team, 2016). Second, variables composition, model fit, and reliability were examined in Mplus (version 7.4; Los Angeles, CA; see Supplementary Table 1 [available online]; Muthén & Muthén, 1998–2016; Raykov, 2009). In addition, approximate measurement invariance (AMI) was tested between time points to ensure that respondents attribute the same meaning to, and understand, items at each data collection time point (van de Schoot, Lugtig, & Hox, 2012). Third, six Bayesian Structural Equation Modeling (Muthén & Asparouhov, 2012) cross-lagged panel model analyses were performed, each analysis including self-control and one motivation regulation represented as latent variables. In these six analyses, three time points were included (T1, T2, and T3; Figure 1). As peoples’ motivation seems to be driven by both internal and external motives (Ryan & Deci, 2000), and self-control in exercise contexts seems to be changeable (Hagger, Wood, Stiff, & Chatzisarantis, 2010), examination of stability and temporal causality within these concepts is relevant.

Table 1

Descriptive Statistics and Correlations of Study Variables at Time Points 1, 2, and 3

M (SD)

1. InT1

6.11 (0.71)

—

.63*

.59*

.66*

.44*

.54*

.40*

.45*

.20*

.11

.10

.07

−.07

.01

−.25*

−.39*

−.28*

.25*

.32*

.24*

2. InT2

6.07 (0.82)

—

.69*

.49*

.74*

.55*

.35*

.64*

.58*

.09

.23*

.19

−.02

.01

.07

−.44*

−.48*

−.56*

.34*

.26*

.36*

3. InT3

5.90 (0.88)

—

.39*

.51*

.75*

.26*

.42*

.65*

.10

.13

.32*

−.04

−.10

.01

−.34*

−.53*

−.51*

.20

.31*

.33*

4. IeT1

5.80 (0.81)

—

.65*

.53*

.49*

.45*

.43*

.45*

.35*

.31*

.17

.06

.11

.17

−.24*

−.12

.26*

.27*

.17

5. IeT2

5.79 (0.90)

—

.68*

.33*

.61*

.52*

.27*

.44*

.39*

.07

.12

.16

−.33*

−.37*

−.38*

.37*

.27*

.31*

6. IeT3

5.66 (0.93)

—

.22

.35*

.58*

.14

.20

.46*

−.05

−.10

.03

−.27*

−.45*

−.44*

.26*

.35*

7. IdT1

5.54 (0.96)

—

.60*

.43*

.36*

.22

.26*

.21*

.10

.20

−.06

−.14

.13

.26*

.13

8. IdT2

5.57 (1.04)

—

.62*

.22

.34*

.26*

.16

.19

.16

−.17

−26*

−.35*

.22

.29*

.26*

9. IdT3

5.52 (0.99)

—

.14

.21

.33*

.04

−.01

.10

−.21

−.41*

−.29*

.16

.30*

.17

10. IrT1

4.45 (1.26)

—

.67*

.57*

.44*

.30*

.32*

.19*

.12

.22

−.10

−.02

−.15

11. IrT2

4.37 (1.30)

—

.67*

.36*

.48*

.39*

.04

.01

.11

−.07

−.03

−.10

12. IrT3

4.46 (1.22)

—

.31*

.27*

.41*

.05

−.01

.03

.01

13. ExT1

2.76 (1.17)

—

.67*

.61*

.29*

.16

.25*

−.23*

−.19

14. ExT2

2.73 (1.24)

—

.61*

.22

.26*

−.30*

−.32*

−.28*

15. ExT3

2.86 (1.17)

—

.14

.20

.28*

−.11

−.19

−.22

16. AT1

2.36 (1.44)

—

.75*

.64*

−.30*

−.32*

−.36*

17. AT2

2.43 (1.55)

—

.81*

−.38*

−.45*

−.54*

18. AT3

2.61 (1.59)

—

−.29*

−.36*

−.56*

19. SCT1

3.56 (0.53)

—

.76*

.64*

20. SCT2

3.57 (0.54)

—

.72

21. SCT3

3.44 (0.52)

—

Note. In = intrinsic regulation; Ie = integrated regulation; Id = identified regulation; Ir = introjected regulation; Ex = external regulation; A = amotivation regulation; SC = self-control; T1, T2, and T3 = Time Points 1, 2, and 3.

*Bayes factor > 10.

Figure 1

· Download Figure

· Download figure as PowerPoint slide

—The cross-lagged panel model of motivation regulations and self-control. Note. T1, T2, and T3 = Time Points 1, 2, and 3.

Citation: Journal of Sport and Exercise Psychology 42, 2; 10.1123/jsep.2019-0143

Bayesian Structural Equation Modeling is based on Bayes’ theorem, and information (priors) from previous studies will, together with current data, generate the posterior distribution (Muthén & Asparouhov, 2012). This approach will, in comparison with more traditional maximum likelihood estimation, improve convergence issues, aid in model identification, and is especially helpful when researchers deal with small sample sizes (Depaoli & van de Schoot, 2015). Parameter specifications of exact zeros were replaced with approximate zeros by weakly informative priors (Muthén & Asparouhov, 2012), which influence posterior distributions to a lesser extent but contain useful information to model identification processes (Depaoli & van de Schoot, 2015). That is, priors allowed low cross loadings and variances within and between each latent variable at different time points, and their distributional form was defined as normal (0, 0.005) or inverse Wishart (0, 32). Sensitivity analyses were performed and investigated (i.e., varying residual correlation and cross-loading variance priors; De Bondt & Van Petegem, 2015), and are available upon request.

In model identification, we implemented two Markov Chain Monte Carlo simulation procedures with the Gibbs sampler method (Depaoli & van de Schoot, 2015). Here, the distribution of one set of parameters is used to make random draws of other parameter values, and missing values (current study item-level missing data < 36.8%) are treated as values to be estimated. In this study, some athletes were away due to competition travelling or practices scheduled at the time of data collection (see Table 2 for response rate). Therefore, no implicit factors affect missing data, and this procedure seems justifiable as analyses indicated data missing completely at random (Little’s Missing Completely at Random test, p = .443; Enders, 2011). Convergence of the Markov Chain Monte Carlo chains was based on the potential scale reduction factor (i.e., potential scale reduction factor close to 1; Muthén & Asparouhov, 2012), and convergence cutoff values were specified at 0.01 to reduce bias caused by precision (van de Schoot et al., 2013, 2014). The resulting model fit was based on the Bayesian posterior predictive p and the 95% confidence interval (CI; van de Schoot et al., 2014). A posterior predictive p close to .50 and a symmetric 95% CI centering on zero indicate excellent fit, although posterior predictive p > .01 is still acceptable (Muthén & Asparouhov, 2012; van de Schoot et al., 2014). To reduce autocorrelation between the Markov Chain Monte Carlo draws, every 10th iteration was used (De Bondt & Van Petegem, 2015), resulting in 200,000 (50,000 burn-in) iterations. Trace plots were visually inspected for chain convergence (Depaoli & van de Schoot, 2015).

Table 2

Overall Response Rate

Time point	Total (n)	Limited response rate (n)
T1	271
T2	201	17
T3	197	20
T1 and T2	184
T1, T2, and T3	136
T2 and T3		13
T1 and T3		17
Sum	321

Note. T1, T2, and T3 = Time Points 1, 2, and 3; limited response rate = athletes who only participated at one or two time points; sum = all athletes enrolled in the study (i.e., new athletes enrolled at T1, T2, and T3, as well as T2 and T3).

Testing cross-lagged panel models longitudinally, researchers should ensure measurement invariance (Little, 2013). This implies that constructs are equivalent over time, and respondents attribute the same meaning to the latent factor(s) and equality in the levels of underlying items at different time points (van de Schoot et al., 2012). However, AMI used in this study allows for some wiggle room for factor loading and intercept variance differences between time points, as the precision of priors may vary (van de Schoot et al., 2013). As such, this is an interesting alternative compared with the unrealistic assumption of exact zeros in more strict measurement invariance testing (van de Schoot et al., 2013). In AMI, zero mean, small variance priors (0.05, 0.01, and 0.005) for differences between estimates of the same parameters (factor loadings and intercepts) at T1–T3 were evaluated (Muthén & Asparouhov, 2012, 2013, January 11). The lowest deviance information criterion indicated the best-fitting model (Asparouhov, Muthén, & Morin, 2015). Then, Muthén and Asparouhov’s (2013, January 11) two-step approach testing AMI for factor loadings and intercepts simultaneously was performed, freeing eventually noninvariant parameters in the second step.

Autoregressive paths (e.g., T1 → T2 intrinsic motivation) and stability over time (e.g., T1 → T2 vs. T2 → T3 intrinsic motivation) were investigated in the cross-lagged panel models, as well as the temporal causality and cross-lagged paths (e.g., T1 intrinsic motivation → T2 self-control vs. T1 self-control → T2 intrinsic motivation; Little, 2013). The cross-lagged and autoregressive paths are both predictors onto another variable, and are reciprocally controlled for when leading to the same construct (Little, 2013). That is, autoregressive effects are uniquely controlled for estimating cross-lagged paths and vice versa. Therefore, the goal of investigating constructs in cross-lagged panel models is to find a reduced, more parsimonious, and theoretically meaningful set of structural paths that explains associations within data (Little, 2013). Interpreting structural paths, 95% credibility intervals not covering zero were considered credible (Muthén & Asparouhov, 2012; van de Schoot et al., 2014).

A simulation analysis, using the Monte Carlo framework, was conducted to evaluate the power and precision of the structural paths within the specified model (for more information about simulation analysis, see Muthén & Muthén, 2002). The coverage rates for the structural paths were above 93.8%, and the power (β) for the parameters ranged between 0.93 and 1.00.

Results

Descriptive statistics and correlations for motivation and self-control are presented in Table 1. Generally, self-control is positively and negatively correlated with autonomous and controlled types of motivation, respectively, and motivation regulations adjacent on the SDT continuum are positively related. Reliability analyses (ρ; Raykov, 2009) indicated acceptable reliability, except some motivation regulations (see Supplementary Table 1 [available online]). Testing AMI between time points, time point difference variances of .05, .01, and .005 were examined. Deviance information criterion values indicated that variances of .005 resulted in the best model reflecting the lowest deviance information criterion, and all factor loadings and intercepts were invariant (see Supplementary Tables 2–7 [available online]). As such, parameters were constrained to be approximately equal (see Table 3 for approximate measurement invariance model fit).

Table 3

Approximate Measurement-Invariance Model Fit for the Three Time-Point Models

Model	#fp	λ prior (μ, σ2)	ν prior (μ, σ2)	PPp	2.5% PP limit	97.5% PP limit	DIC
1, Step 1	785	.05	.05	.468	−105.038	125.416	24173.362
	785	.01	.01	.490	−109.306	114.096	24161.410
	785	.005	.005	.504	−113.570	113.134	24154.002
2, Step 1	785	.05	.05	.510	−116.999	138.414	24346.728
	785	.01	.01	.532	−118.225	134.054	24334.237
	785	.005	.005	.537	−117.666	134.229	24326.324
3, Step 1	785	.05	.05	.316	−87.323	146.343	24197.218
	785	.01	.01	.332	−91.724	145.464	24184.188
	785	.005	.005	.341	−93.047	147.852	24179.177
4, Step 1	785	.05	.05	.282	−80.283	160.897	24383.048
	785	.01	.01	.290	−87.769	159.203	24371.392
	785	.005	.005	.296	−87.600	159.075	24365.056
5, Step 1	785	.05	.05	.498	−121.697	132.471	24279.686
	785	.01	.01	.506	−124.903	123.466	24261.677
	785	.005	.005	.518	−127.241	119.482	24253.079
6, Step 1	785	.05	.05	.420	−107.741	139.817	23618.120
	785	.01	.01	.442	−115.779	134.142	23607.077
	785	.005	.005	.460	−120.078	133.579	23599.967

Note. Factor loading and intercept difference variances = .005 indicated the best model fit according to the PPp and DIC. Analyses with difference variances = 0.01 and = 0.05 were estimated, and are available on request. Note. Model 1 = intrinsic regulation—self-control; Model 2 = integrated regulation—self-control; Model 3 = identified regulation—self-control; Model 4 = introjected regulation—self-control; Model 5 = external regulation—self-control; Model 6 = amotivation—self-control; #fp = number of free parameters; PPp = posterior predictive p; DIC = deviance information criterion. Items are standardized.

Results from the Bayesian Structural Equation Modeling cross-lagged panel models are presented in Table 4. Strong effects were found for all autoregressive paths (0.449 ≥ β ≤ 0.742; e.g., T1 → T2 self-control, β = 0.631, 95% CI [0.517, 0.724]). Cross-lagged paths were weaker when controlling for autoregressive paths (−0.003 ≥β ≤ 0.278; e.g., intrinsic motivation T2→ self-control T3, β = 0.278, 95% CI [0.120, 0.436]). Furthermore, the stability over time within the various constructs displayed some instability. Motivation regulations showed higher T2 → T3 compared with T1 → T2 autoregressive paths, whereas the self-control construct reflected more complex patterns, as T2 → T3 compared with T1 → T2 paths both increased and decreased. However, we focused on self-control and motivation cross-lagged paths and temporal causality in this article, and these results are presented next.

Table 4

Cross-Lagged Three Time-Point Models

Model

T1Mot

→

T2Mot

[95% CI]

T2Mot

→

T3Mot

[95% CI]

T1SC

→

T2SC

[95% CI]

T2SC

→

T3SC

[95% CI]

T1Mot

→

T2SC

[95% CI]

T2Mot

→

T3SC

[95% CI]

T1SC

→

T2Mot

[95% CI]

T2SC

→

T3Mot

[95% CI]

0.449

[0.280, 0.584]

0.636

[0.497, 0.741]

0.626

[0.503, 0.717]

0.564

[0.387, 0.689]

0.056

[−0.067, 0.181]

0.278

[0.120, 0.436]

0.182

[0.028, 0.336]

0.120

[0.017, 0.258]

0.502

[0.348, 0.627]

0.653

[0.508, 0.759]

0.625

[0.488, 0.718]

0.577

[0.417, 0.696]

0.024

[−0.113, 0.165]

0.239

[0.081, 0.389]

0.211

[0.071, 0.347]

0.150

[0.009, 0.290]

0.511

[0.365, 0.624]

0.538

[0.369, 0.667]

0.632

[0.504, 0.720]

0.674

[0.495, 0.777]

0.105

[−0.046, 0.263]

−0.003

[−0.166, 0.169]

0.092

[−0.046, 0.233]

0.128

[−0.028, 0.296]

0.584

[0.448, 0.690]

0.629

[0.480, 0.736]

0.627

[0.497, 0.711]

0.680

[0.555, 0.765]

−0.114

[−0.253, 0.026]

−0.083

[−0.219, 0.057]

−0.042

[−0.177, 0.096]

0.006

[−0.143, 0.147]

0.537

[0.371, 0.664]

0.591

[0.405, 0.723]

0.618

[0.505, 0.713]

0.643

[0.505, 0.750]

−0.075

[−0.211, 0.064]

−0.084

[−0.228, 0.055]

−0.137

[−0.283, 0.008]

0.021

[−0.136, 0.177]

0.572

[0.432, 0.676]

0.742

[0.618, 0.835]

0.631

[0.517, 0.724]

0.551

[0.397, 0.678]

−0.069

[−0.205, 0.066]

−0.252

[−0.398, −0.097]

−0.146

[−0.271, −0.016]

−0.021

[−0.158, 0.119]

Note. The 95% CIs not covering zero are considered credible (Muthén & Asparouhov, 2012; van de Schoot et al., 2014). T1, T2, and T3 = Time Points 1, 2, and 3; Mot = motivation; CI = credibility interval; SC = self-control; Model 1 = intrinsic regulation—self-control; Model 2 = integrated regulation—self-control; Model 3 = identified regulation—self-control; Model 4 = introjected regulation—self-control; Model 5 = external regulation—self-control; Model 6 = amotivation—self-control.

In the cross-lagged panel models, we investigated the hypothesized associations that individuals’ self-control capacity will be more influenced by the type of motivation than vice versa, and that autonomous and controlled types of motivation positively and negatively predict trait self-control, respectively. In these three time-point models, self-control credibly predicted intrinsic motivation (β = 0.182, 95% CI [0.028, 0.336]), integrated regulation (β = 0.211, 95% CI [0.071, 0.347]), and amotivation (β = −0.146, 95% CI [−0.271, −0.016]) in the T1 → T2 cross paths, and integrated regulation (β = 0.150, 95% CI [0.009, 0.290]) in the T2 → T3 cross paths. Self-control was credibly predicted by intrinsic motivation (β = 0.278, 95% CI [0.120, 0.436]), integrated regulation (β = 0.239, 95% CI [0.081, 0.389]), and amotivation (β = −0.252, 95% CI [−0.398, −0.097]) in the T2 → T3 cross paths. Analyses revealed noncredible and weak cross paths between self-control and identified, introjected, and external regulations.

Discussion

Anchored in the frameworks of the SDT (Ryan & Deci, 2000) and theories of self-control (Baumeister et al., 1998; Inzlicht & Schmeichel, 2012; Ryan & Deci, 2008), we investigated the temporal ordering of motivation regulations and dispositional self-control in young, high-level winter sport athletes. Recent empirical evidence states that athletes are not always capable of dealing with the self-control demands they are constantly confronted with (Englert, 2017). Evidence also suggests that various types of motivation play a crucial role in the optimal functioning of self-control competencies among athletes (Jordalen et al., 2016; Jordalen et al., 2018). Self-control capacity has been conceptualized as limited by psychological and physiological resources, and sequential acts of self-control without adequate recovery will result in temporary shifts in motivation and depletion patterns followed by self-control failure (Baumeister et al., 2007; Inzlicht & Schmeichel, 2012). However, models explaining self-control depletion and self-control failure do not agree about the temporal ordering of these concepts.

Although no previous studies have examined the temporal ordering of motivation and self-control in the sport context, it is possible to infer from earlier findings that autonomous motivation positively directs acts of self-control in other domains (e.g., Muraven, Gagné, & Rosman, 2008). For example, autonomous types of motivation may protect athletes against temptations and thereby boost their self-control capacity, as they may experience fewer obstacles and tempting in-the-moment desires in the face of their goal pursuits (Milyavskaya et al., 2015). This direction of effects has previously been supported in exercise contexts when investigating the association between motivation and well-being mediated by trait self-control (Briki, 2016). In this study, Briki (2016) found that autonomous and controlled types of motivation positively and negatively predicted well-being via self-control competencies among regular exercisers. Conceptually consistent, previous study findings suggest that motivation directs acts of self-control (Baumeister, 2016; Ryan & Deci, 2008).

In the current study, the temporal ordering of high-level athletes’ motivation and self-control looked different than what was anticipated. Athletes’ initial self-control was a stronger predictor of motivation, than vice versa, in the three-wave cross-lagged models. As such, current study findings support Inzlicht and Schmeichel’s (2012) process model of self-control, as well as recent findings (see e.g., Converse et al., 2019; Holding et al., 2019), that self-control initially directs changes in motivation subsequently affecting acts of self-control. The process model questions whether individuals actually are depleted and suggest that there is a shift in motivation, attention, and emotion that causes self-control decreased performance (Inzlicht & Schmeichel, 2012). For example, an initial act of self-control leaves athletes less motivated to deliberately control their actions (e.g., persistently engage in alternative training when not fully recovered from injury) and more motivated to execute personally rewarding and enjoyable tasks (e.g., go for a favorite workout even though not fully recovered). This motivational shift leading to reduced motivation for have to tasks and the increased motivation for want to tasks corresponds to the shifts in types of motivation where the more externally motivated performances are difficult to maintain over time (Ryan & Deci, 2000). This may exemplify how self-control and high-level athletes’ strong work ethic override other tempting desires in their development of exceptional competencies.

High levels of trait self-control may help athletes move along the self-determination continuum and gain more autonomous reasons for goal striving (Holding et al., 2019). For example, athletes who autonomously control their efforts and persist in the face of adversity may experience their activities as more meaningful and interesting, leading to increased pleasure and fun within the activity (i.e., internalization of motivation; Ryan & Deci, 2000). Moreover, the fact that athletes’ self-control competencies initiated the causal paths between self-control and motivation reflect that athletes in the competitive nature of elite sport possess strong self-control and willpower competencies (Hoffer & Giddings, 2016). These mental characteristics have been found to help athletes stay focused on the task at hand and guide their performance toward goal achievements (Boes, Harung, Travis, & Pensgaard, 2014). However, these findings may reflect the relative stability of trait self-control over time (Tangney et al., 2004), underlined in strong autoregressive effects (Adachi & Willoughby, 2015), and a requirement to assess these associations longitudinally, for example, during competitive as well as off-season periods (Anusic & Schimmack, 2016; Jordalen et al., 2018).

Investigating how trait self-control predicted changes in motivation quality across the academic year, self-control positively and negatively predicted undergraduate students’ autonomous and controlled motivation above other personality traits (Holding et al., 2019). In their study, Holding et al. (2019) confirmed strong autoregressive effects of motivation, but did not investigate autoregressive effects of self-control. Accordingly, current study findings offer important information as autoregressive effects are controlled for investigating cross-lagged paths (Adachi & Willoughby, 2015). Athletes in the current study participated in different sports, but competed in national and international competitions within the same period of time. The first measurement time point was organized prior to these competitions, and one could speculate whether athletes’ self-control was especially important predicting their motivation already at this point. Elite athletes have previously been found to yield multidimensional motivation profiles including a combination of autonomous and more controlled types of motivation (Gillet, Berjot, Vallerand, Amoura, & Rosnet, 2012). Thus, an important consideration is whether athletes’ motivation was fluctuating throughout the season (Lemyre et al., 2006). Based on the strong autoregressive effects of motivation underlining stability over change, findings contradict this suggestion and showcase the power of self-control predicting motivation over time (Adachi & Willoughby, 2015; Holding et al., 2019).

Interestingly, trait self-control was specifically associated with motivation regulations (i.e., intrinsic, integrated, and amotivation) that are not characterized by volitional processes (Ryan & Deci, 2000). Rather, these types of motivation reflect natural motivational desires to act or a total lack of motivation to act, and refer to doing something because it is interesting, fun, and meaningful, or conversely represent a lack of intentionality and sense of personal causation. However, these motivation regulations were further associated with athletes’ trait self-control at the end of the season, reflecting how autonomous types of motivation may energize acts of self-control, whereas controlled forms of motivation will rather deplete these cognitive resources (Briki, 2016; Holding et al., 2019; Jordalen et al., 2018; Muraven, 2008; Ryan & Deci, 2008). At the later stages of the season, athletes likely shifted their competitive sport focus and deliberately focused on their academic efforts as college exams are typically scheduled during this period. In this transition period, they are especially challenged by educational as well as psychological and physiological demands related to the student–athlete life, and would typically benefit from the interaction of autonomous motivation and self-control competencies (e.g., thought control; Martinent & Decret, 2015). This may help athletes continue their process of goal pursuit, as the combination of self-control and more autonomous types of motivation results in decreased impulsive attraction to goal-disruptive temptations and individuals’ perceptions of obstacles (Milyavskaya et al., 2015). In addition, the direction of associations in the current study appears to be influenced by athletes’ type of motivation. In accordance with previous research findings (e.g., Briki, 2016; Holding et al., 2019; Muraven, 2008), associations between self-control and more autonomous types of motivation were positive, and conversely, athletes’ amotivation negatively associated with trait self-control. This type of motivation is characterized by a lack of control (Ryan & Deci, 2000), and therefore, athletes’ levels of T2 amotivation are negatively associated with end of season self-control. This association has been found maladaptive, as it likely leads to exhaustion and eventually burnout experiences (Jordalen et al., 2018).

Current study findings suggest that self-control initiate the causal paths between athletes’ various types of motivation and self-control competencies. These associations were conceptually consistent, as self-control was positively and negatively associated with autonomous and controlled types of motivation, respectively. In addition, findings suggest a need to examine these concepts longitudinally, during competitive season and off-season periods to better identify complex and interrelated psychological patterns.

Limitations

The current study makes a unique contribution to the literature concerning the temporal ordering of self-control and motivation in youth sport athletes. However, findings should be interpreted based on potential limitations. Although it is hard to assess their influence, factors such as the use of self-reported data and the first author’s presence when visiting colleges at every data collection time point are likely to have influenced athletes’ perception (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). Furthermore, possible limitations related to questionnaires, design, and type of analyses should be emphasized. For example, some motivation regulations’ reliability scores were low and two BSCS items were deleted due to low factor loadings. The translation of items needs to be further investigated as this may cause unseen linguistic or cultural gaps (Benítez, Padilla, Hidalgo Montesinos, & Sireci, 2016). However, validity analyses were performed with the Mplus robust maximum likelihood estimator, and even though this estimator is robust for nonnormal conditions and missing data (Enders, 2010), main analyses were performed with Bayesian Structural Equation Modeling where priors allowed for some wiggle room for differences between parameters (van de Schoot et al., 2013). This method, where exact zero constraints (e.g., for cross loadings) are replaced with approximate zero constraints, is an attractive and more realistic approach. The content validity of the Sport Motivation Scale II has been criticized as well (e.g., see Langan et al., 2016), and the wording of items may not necessarily apply to elite sport settings (e.g., asking athletes if they were engaged because they enjoyed learning more about their sport). In addition, self-control reverse-scored items may cause method bias (Marsh, Scalas, & Nagengast, 2010; Podsakoff et al., 2003), and various compositions of this construct have previously been investigated (Jordalen et al., 2016, 2018; Maloney, Grawitch, & Barber, 2012; Toering & Jordet, 2015). Finally, the current study’s 10-week data collection period may not be sufficient for some interaction effects to emerge (e.g., Martinent & Decret, 2015; Stenling, Ivarsson, & Lindwall, 2017). This data collection period may not reflect longitudinal associations between motivation and self-control.

Future Directions

Recently, it has been emphasized that there is a need to go beyond laboratory research settings when measuring self-control and cognitive competencies, as the validity of tasks used to manipulate or measure self-control capacity is generally unknown (Baumeister & Vohs, 2016a; Carter, Kofler, Forster, & McCullough, 2015). These competencies should be studied longitudinally in real-life settings, for example, in youth sport competitive and off-season periods (Holding et al., 2019; Stenling et al., 2017). Former research has advocated intervention strategies to improve self-control (see e.g., Jordalen et al., 2018; Milyavskaya & Inzlicht, 2017). These strategies could either act on autonomous or controlled self-control motives (Ryan & Deci, 2000, 2008). For example, individuals may internalize the reason for engaging in acts of self-control, thus they experience self-control behaviors as personally meaningful and interesting (cf., autonomous motive); or individuals are able to resist immediate temptations in favor of a distal goal to receive an extra bonus (cf., more controlled motive, delay of gratification; Mischel, 2014). The former motive is, according to SDT, preferable, as it helps individuals maintain behaviors over time without the necessity of separable consequences (Deci & Ryan, 2000). This, in line with current study findings as well as former research (e.g., Converse et al., 2019: Holding et al., 2019; Jordalen et al., 2016, 2018) suggests that autonomous self-control motives enhances self-control performance over time. For example, athletes can increase the value of engaging in acts of self-control by discussing important self-control processes (e.g., behavioral and emotional responses, self-management, enhanced focus, as well as thought and impulse control) with their coach and significant others (Dubuc-Charbonneau & Durand-Bush, 2015).

Based on limitations of the current study questionnaires, some future directions should be endorsed. For example, it is important to evaluate the potential consequences of deleting two BSCS items and why these items displayed low factor loadings. The various compositions of this questionnaire previously investigated (e.g., Jordalen et al., 2016, 2018; Maloney et al., 2012; Tangney et al., 2004; Toering & Jordet, 2015), as well as current study analyses, suggest that a thorough investigation of self-control items is needed. Do current BSCS items reflect athletes’ actual self-control, or do these items solely measure participants illusive self-control; is this a unidimensional construct as originally suggested (see Tangney et al., 2004), or a two-factor scale more recently investigated (Toering & Jordet, 2015); and do the BSCS items actually measure trait self-control, or do they inadequately measure stability over time (Fullerton, Lane, Nevill, & Devonport, 2018)? Future research should additionally evaluate the Norwegian version of the Sport Motivation Scale II and consider validation of a new motivation regulations questionnaire for high-level youth athletes. Finally, it is important that sport psychology research apply longitudinal designs and methods to evaluate causal processes in athletes’ everyday life (Preacher, 2015; Stenling et al., 2017), and account for threats of method bias using self-report measures (e.g., social desirability; Grossbard, Cumming, Standage, Smith, & Smoll, 2007; Podsakoff et al., 2003). For example, answering sport motivation questionnaires, youth athletes most likely report favorable scores consistent with their long-term agenda to achieve elite-level status.

Conclusion

Our study’s findings highlight interrelated associations between youth athletes’ dispositional self-control and various types of motivation. Investigating the temporal ordering of these concepts throughout athletes’ competitive winter sport season, findings challenge the established fact that inherent motivation initially and exclusively moves athletes to act. This belief disregards that psychological and cognitive competencies may energize and enable drives to be fulfilled. Our results suggest a multifaceted relationship between athletes’ motivation and trait self-control, and suggest that self-control capacity initially enables motivation desires to evolve. In a three-wave cross-lagged panel model, intrinsic regulation, integrated regulation, and amotivation were predicted at Time Point 2 by athletes’ self-control at the beginning of the competitive season. These motivation regulations further predicted trait self-control at the end of the season, and findings reflect that athletes’ self-control capacity is associated with types of motivation not specifically characterized by volitional processes. Noteworthy, the more autonomous and controlled forms of motivation were positively and negatively associated with trait self-control, respectively, likely important for the maintenance of self-control performance.

These findings have important applied implications for high-level youth athletes. As these athletes are constantly challenged with self-control demands in their strenuous everyday combining education and elite sport development with family and social activities, an emphasis on supporting the types of motivation that are positively associated with athletes’ volitional resources is important. Autonomous types of motivation protect athletes against tempting self-control dilemmas, as they experience an increased awareness of the value, meaning, and inherent satisfaction of their own developmental processes. This suggests that athletes’ with high-levels of trait self-control may better internalize external types of motivation and maintain their persistent goal striving. They avoid the debilitating effects of self-control depletion and exhaustion, experiencing increased feelings of well-being and other positive health outcomes. Finally, current study findings showed that motivation and self-control are stable constructs over a 10-week period of time. It is important to outline an extended time frame when coaches and significant others intend to facilitate positive changes in these mental dispositions. As such, research exploring the forces directing competitive athletes’ behaviors and performances needs to further integrate ideas from multiple lines of research and theory, and explore motivational and cognitive issues simultaneously.

Acknowledgments

This study is based on data used in a doctoral dissertation. The authors want to acknowledge Bård Erlend Solstad and Harald Solhaug Naess for valuable feedback on this paper. Additionally, the authors highly appreciate the cooperation of sports directors and coaches, and are thankful for athletes’ contribution.

References

· Adachi, P., & Willoughby, T. (2015). Interpreting effect sizes when controlling for stability effects in longitudinal autoregressive models: Implications for psychological science. European Journal of Developmental Psychology, 12(1), 116–128. doi:10.1080/17405629.2014.963549

· Crossref Search Google Scholar Export Citation

· Anusic, I., & Schimmack, U. (2016). Stability and change of personality traits, self-esteem, and well-being: Introducing the meta-analytic stability and change model of retest correlations. Journal of Personality and Social Psychology, 110(5), 766–781. PubMed ID: 26619304 doi:10.1037/pspp0000066

· Crossref Search Google Scholar Export Citation

· Asparouhov, T., Muthén, 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

· Crossref Search Google Scholar Export Citation

· Baumeister, R.F. (2016). Toward a general theory of motivation: Problems, challenges, opportunities, and the big picture. Motivation and Emotion, 40(1), 1–10. doi:10.1007/s11031-015-9521-y

· Crossref Search Google Scholar Export Citation

· Baumeister, R.F., Bratslavsky, E., Muraven, M., & Tice, D.M. (1998). Ego depletion: Is the active self a limited resource? Journal of Personality and Social Psychology, 74(5), 1252–1265. PubMed ID: 9599441 doi:10.1037/0022-3514.74.5.1252

· Crossref Search Google Scholar Export Citation

· Baumeister, R.F., & Vohs, K.D. (2016a). Misguided effort with elusive implications. Perspectives on Psychological Science, 11(4), 574–575. PubMed ID: 27474143 doi:10.1177/1745691616652878

· Crossref Search Google Scholar Export Citation

· Baumeister, R.F., & Vohs, K.D. (2016b). Strength model of self-regulation as limited resource: Assessment, controversies, update. In J.M. Olson & M.P. Zanna (Eds.), Advances in experimental social psychology (Vol. 54, pp. 67–127). Cambridge, MA: Academic Press.

· Search Google Scholar Export Citation

· Baumeister, R.F., Vohs, K.D., & Tice, D.M. (2007). The strength model of self-control. Current Directions in Psychological Science, 16(6), 351–355. doi:10.1111/j.1467-8721.2007.00534.x

· Crossref Search Google Scholar Export Citation

· Benítez, I., Padilla, J.-L., Hidalgo Montesinos, M.D., & Sireci, S.G. (2016). Using mixed methods to interpret differential item functioning. Applied Measurement in Education, 29(1), 1–16. doi:10.1080/08957347.2015.1102915

· Crossref Search Google Scholar Export Citation

· Boes, R., Harung, H.S., Travis, F., & Pensgaard, A.M. (2014). Mental and physical attributes defining world-class Norwegian athletes: Content analysis of interviews. Scandinavian Journal of Medicine & Science in Sports, 24(2), 422–427. PubMed ID: 22830416 doi:10.1111/j.1600-0838.2012.01498.x

· Crossref Search Google Scholar Export Citation

· Briki, W. (2016). Motivation toward physical exercise and subjective wellbeing: The mediating role of trait self-control. Frontiers in Psychology, 7, 1546. doi:10.3389/fpsyg.2016.01546

· Crossref Search Google Scholar Export Citation

· Carter, E.C., Kofler, L.M., Forster, D.E., & McCullough, M.E. (2015). A series of meta-analytic tests of the depletion effect: Self-control does not seem to rely on a limited resource. Journal of Experimental Psychology: General, 144(4), 796–815. doi:10.1037/xge0000083

· Crossref Search Google Scholar Export Citation

· Cerasoli, C.P., Nicklin, J.M., & Ford, M.T. (2014). Intrinsic motivation and extrinsic incentives jointly predict performance: A 40-year meta-analysis. Psychological Bulletin, 140(4), 980–1008. PubMed ID: 24491020 doi:10.1037/a0035661

· Crossref Search Google Scholar Export Citation

· Chantal, Y., Guay, F., Dobreva-Martinova, T., & Vallerand, R.J. (1996). Motivation and elite performance: An exploratory investigation with Bulgarian athletes. International Journal of Sport Psychology, 27(2), 173–182. Retrieved from www.ijsp-online.com

· Search Google Scholar Export Citation

· Chemolli, E., & Gagné, M. (2014). Evidence against the continuum structure underlying motivation measures derived from self-determination theory. Psychological Assessment, 26(2), 575–585. PubMed ID: 24611788 doi:10.1037/a0036212

· Crossref Search Google Scholar Export Citation

· Converse, B.A., Juarez, L., & Hennecke, M. (2019). Self-control and the reasons behind our goals. Journal of Personality and Social Psychology, 116(5), 860–883. PubMed ID: 30070566 doi:10.1037/pspp0000188

· Crossref Search Google Scholar Export Citation

· Cresswell, S.L., & Eklund, R.C. (2005). Motivation and burnout among top amateur rugby players. Medicine & Science in Sports & Exercise, 37(3), 469–477. PubMed ID: 15741847 doi:10.1249/01.MSS.0000155398.71387.C2

· Crossref Search Google Scholar Export Citation

· 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, 1963. doi:10.3389/fpsyg.2015.01963

· Crossref Search Google Scholar Export Citation

· Deci, E.L., & Ryan, R.M. (2000). The “what” and “why” of goal pursuits: Human needs and the self-determination of behavior. Psychological Inquiry, 11(4), 227–268. doi:10.1207/S15327965PLI1104_01

· Crossref Search Google Scholar Export Citation

· Depaoli, S., & van de Schoot, R. (2015). Improving transparency and replication in Bayesian statistics: The WAMBS-checklist. Psychological Methods, 22(2), 240–261. PubMed ID: 26690773 doi:10.1037/met0000065

· Crossref Search Google Scholar Export Citation

· de Ridder, D.T.D., Lensvelt-Mulders, G., Finkenauer, C., Stok, F.M., & Baumeister, R.F. (2012). Taking stock of self-control: A meta-analysis of how trait self-control relates to a wide range of behaviors. Personality and Social Psychology Review, 16(1), 76–99. PubMed ID: 21878607 doi:10.1177/1088868311418749

· Crossref Search Google Scholar Export Citation

· Dubuc-Charbonneau, N., & Durand-Bush, N. (2015). Moving to action: The effects of a self-regulation intervention on the stress, burnout, well-being, and self-regulation capacity levels of university student-athletes. Journal of Clinical Sport Psychology, 9(2), 173–192. doi:10.1123/jcsp.2014-0036

· Crossref Search Google Scholar Export Citation

· Enders, C.K. (2010). Applied missing data analysis. New York, NY: The Guilford Press.

· Search Google Scholar Export Citation

· Enders, C.K. (2011). Analyzing longitudinal data with missing values. Rehabilitation Psychology, 56(4), 267–288. PubMed ID: 21967118 doi:10.1037/a0025579

· Crossref Search Google Scholar Export Citation

· Englert, C. (2016). The strength model of self-control in sport and exercise psychology. Frontiers in Psychology, 7, 314. PubMed ID: 26973590 doi:10.3389/fpsyg.2016.00314

· Crossref Search Google Scholar Export Citation

· Englert, C. (2017). Ego depletion in sports: Highlighting the importance of self-control strength for high-level sport performance. Current Opinion in Psychology, 16(Suppl. C), 1–5. PubMed ID: 28813329 doi:10.1016/j.copsyc.2017.02.028

· Crossref Search Google Scholar Export Citation

· Friese, M., Loschelder, D.D., Gieseler, K., Frankenbach, J., & Inzlicht, M. (2019). Is ego depletion real? An analysis of arguments. Personality and Social Psychology Review, 23(2), 107–131. PubMed ID: 29591537 doi:10.1177/1088868318762183

· Crossref Search Google Scholar Export Citation

· Friese, M., Messner, C., & Schaffner, Y. (2012). Mindfulness meditation counteracts self-control depletion. Consciousness and Cognition, 21(2), 1016–1022. PubMed ID: 22309814 doi:10.1016/j.concog.2012.01.008

· Crossref Search Google Scholar Export Citation

· Fullerton, C.L., Lane, A.M., Nevill, A.M., Devonport, T.J. (2018). Does the brief self-control scale assess relatively stable individual differences in self-control among endurance athletes? Journal of Sport Behavior, 41(1), 27–39. Retrieved from https://www.questia.com/library/p2171/journal-of-sport-behavior

· Search Google Scholar Export Citation

· Gillet, N., Berjot, S., Vallerand, R.J., Amoura, S., & Rosnet, E. (2012). Examining the motivation-performance relationship in competitive sport: A cluster-analytic approach. International Journal of Sport Psychology, 43(2), 79–102. Retrieved from www.ijsp-online-com

· Search Google Scholar Export Citation

· Gillet, N., Vallerand, R.J., & Rosnet, E. (2009). Motivational clusters and performance in a real-life setting. Motivation and Emotion, 33(1), 49–62. doi:10.1007/s11031-008-9115-z

· Crossref Search Google Scholar Export Citation

· Gould, D. (1996). Personal motivation gone awry: Burnout in competitive athletes. Quest, 48(3), 275–289. doi:10.1080/00336297.1996.10484197

· Crossref Search Google Scholar Export Citation

· Grossbard, J.R., Cumming, S.P., Standage, M., Smith, R.E., & Smoll, F.L. (2007). Social desirability and relations between goal orientations and competitive trait anxiety in young athletes. Psychology of Sport and Exercise, 8(4), 491–505. doi:10.1016/j.psychsport.2006.07.009

· Crossref Search Google Scholar Export Citation

· Hagger, M.S., & Chatzisarantis, N.L. (2016). A multilab preregistered replication of the ego-depletion effect. Perspectives on Psychological Science, 11(4), 546–573. PubMed ID: 27474142 doi:10.1177/1745691616652873

· Crossref Search Google Scholar Export Citation

· Hagger, M.S., Wood, C.W., Stiff, C., & Chatzisarantis, N.L.D. (2010). Self-regulation and self-control in exercise: The strength-energy model. International Review of Sport and Exercise Psychology, 3(1), 62–86. doi:10.1080/17509840903322815

· Crossref Search Google Scholar Export Citation

· Hoffer, A., & Giddings, L. (2016). Exercising willpower: Differences in willpower depletion among athletes and nonathletes. Contemporary Economic Policy, 34(3), 463–474. doi:10.1111/coep.12150

· Crossref Search Google Scholar Export Citation

· Holding, A., Hope, N., Verner-Filion, J., & Koestner, R. (2019). In good time: A longitudinal investigation of trait self-control in determining changes in motivation quality. Personality and Individual Differences, 139, 132–137. doi:10.1016/j.paid.2018.11.001

· Crossref Search Google Scholar Export Citation

· Inzlicht, M., & Schmeichel, B.J. (2012). What is ego depletion? Toward a mechanistic revision of the resource model of self-control. Perspectives on Psychological Science, 7(5), 450–463. PubMed ID: 26168503 doi:10.1177/1745691612454134

· Crossref Search Google Scholar Export Citation

· Inzlicht, M., Schmeichel, B.J., & Macrae, C.N. (2014). Why self-control seems (but may not be) limited. Trends in Cognitive Sciences, 18(3), 127–133. PubMed ID: 24439530 doi:10.1016/j.tics.2013.12.009

· Crossref Search Google Scholar Export Citation

· JASP Team. (2016). JASP (Version 0.8.0.0) [Computer Software].

· Export Citation

· Jordalen, G., Lemyre, P.-N., & Durand-Bush, N. (2016). Exhaustion experiences in junior athletes: The importance of motivation and self-control competencies. Frontiers in Psychology, 7, 1867. doi:10.3389/fpsyg.2016.01867

· Crossref Search Google Scholar Export Citation

· Jordalen, G., Lemyre, P.-N., & Durand-Bush, N. (2019). Interplay of motivation and self-regulation throughout the development of elite athletes. Qualitative Research in Sport, Exercise and Health. Advance online publication. doi:10.1080/2159676X.2019.1585388

· Crossref Search Google Scholar Export Citation

· Jordalen, G., Lemyre, P.-N., Solstad, B.E., & Ivarsson, A. (2018). The role of self-control and motivation on exhaustion in youth athletes: A longitudinal perspective. Frontiers in Psychology, 9, 2449. doi:10.3389/fpsyg.2018.02449

· Crossref Search Google Scholar Export Citation

· Jose, P.E. (2016). The merits of using longitudinal mediation. Educational Psychologist, 51(3–4), 331–341. doi:10.1080/00461520.2016.1207175

· Crossref Search Google Scholar Export Citation

· Kline, R.B. (2011). Principles and practice of structural equation modeling (3rd ed.). New York, NY: The Guilford Press.

· Search Google Scholar Export Citation

· Langan, E., Hodge, K., McGowan, S., Carney, S., Saunders, V., & Lonsdale, C. (2016). The influence of controlled motivation alongside autonomous motivation: Maladaptive, buffering, or additive effects? International Journal of Sport & Exercise Psychology, 14(1), 57–71. doi:10.1080/1612197X.2015.1016084

· Crossref Search Google Scholar Export Citation

· Lee, N., Chatzisarantis, N., & Hagger, M.S. (2016). Adequacy of the sequential-task paradigm in evoking ego-depletion and how to improve detection of ego-depleting phenomena. Frontiers in Psychology, 7, 136. doi:10.3389/fpsyg.2016.00136

· Crossref Search Google Scholar Export Citation

· Lemyre, P.-N., Roberts, G.C., & Stray-Gundersen, J. (2007). Motivation, overtraining, and burnout: Can self-determined motivation predict overtraining and burnout in elite athletes? European Journal of Sport Science, 7(2), 115–126. doi:10.1080/17461390701302607

· Crossref Search Google Scholar Export Citation

· Lemyre, P.-N., Treasure, D.C., & Roberts, G.C. (2006). Influence of variability in motivation and affect on elite athlete burnout susceptibility. Journal of Sport & Exercise Psychology, 28(1), 32–48. Retrieved from http://journals.humankinetics.com.ezproxy.wpunj.edu/jsep

· Crossref Search Google Scholar Export Citation

· Little, T.D. (2013). Longitudinal structural equation modeling (1st ed.). New York, NY: The Guilford Press.

· Search Google Scholar Export Citation

· Maloney, P.W., Grawitch, M.J., & Barber, L.K. (2012). The multi-factor structure of the brief self-control scale: Discriminant validity of restraint and impulsivity. Journal of Research in Personality, 46(1), 111–115. doi:10.1016/j.jrp.2011.10.001

· Crossref Search Google Scholar Export Citation

· Marsh, H.W., Scalas, L.F., & Nagengast, B. (2010). Longitudinal tests of competing factor structures for the Rosenberg Self-Esteem Scale: Traits, ephemeral artifacts, and stable response styles. Psychological Assessment, 22(2), 366–381. PubMed ID: 20528064 doi:10.1037/a0019225

· Crossref Search Google Scholar Export Citation

· Martinent, G., & Decret, J.-C. (2015). Motivational profiles among young table–tennis players in intensive training settings: A latent profile transition analysis. Journal of Applied Sport Psychology, 27(3), 268–287. doi:10.1080/10413200.2014.993485

· Crossref Search Google Scholar Export Citation

· Martinent, G., Decret, J.-C., Guillet-Descas, E., & Isoard-Gautheur, S. (2014). A reciprocal effects model of the temporal ordering of motivation and burnout among youth table tennis players in intensive training settings. Journal of Sports Sciences, 32(17), 1648–1658. PubMed ID: 24802802 doi:10.1080/02640414.2014.912757

· Crossref Search Google Scholar Export Citation

· Milyavskaya, M., & Inzlicht, M. (2017). Attentional and motivational mechanisms of self-control. In D. de Ridder, M. Adriaanse, & K. Fujita (Eds.), Handbook of self-control in health and well-being (pp. 11–24). New York, NY: Routledge.

· Search Google Scholar Export Citation

· Milyavskaya, M., Inzlicht, M., Hope, N., & Koestner, R. (2015). Saying “no” to temptation: Want-to motivation improves self-regulation by reducing temptation rather than by increasing self-control. Journal of Personality and Social Psychology, 109(4), 677–693. PubMed ID: 25984785 doi:10.1037/pspp0000045

· Crossref Search Google Scholar Export Citation

· Mischel, W. (2014). The marshmallow test: Understanding self-control and how to master it. London, UK: Bantam Press, Transworld Publishers.

· Search Google Scholar Export Citation

· Moller, A.C., Deci, E.L., & Ryan, R.M. (2006). Choice and ego-depletion: The moderating role of autonomy. Personality and Social Psychology Bulletin, 32(8), 1024–1036. PubMed ID: 16861307 doi:10.1177/0146167206288008

· Crossref Search Google Scholar Export Citation

· Muraven, M. (2008). Autonomous self-control is less depleting. Journal of Research in Personality, 42(3), 763–770. PubMed ID: 18704202 doi:10.1016/j.jrp.2007.08.002

· Crossref Search Google Scholar Export Citation

· Muraven, M., Gagné, M., & Rosman, H. (2008). Helpful self-control: Autonomy support, vitality, and depletion. Journal of Experimental Social Psychology, 44(3), 573–585. PubMed ID: 18496610 doi:10.1016/j.jesp.2007.10.008

· Crossref Search Google Scholar Export Citation

· Muraven, M., & Slessareva, E. (2003). Mechanisms of self-control failure: Motivation and limited resources. Personality and Social Psychology Bulletin, 29(7), 894–906. PubMed ID: 15018677 doi:10.1177/0146167203029007008

· Crossref Search Google Scholar Export Citation

· Muraven, M., Tice, D.M., & Baumeister, R.F. (1998). Self-control as a limited resource: Regulatory depletion patterns. Journal of Personality and Social Psychology, 74(3), 774–789. PubMed ID: 9523419 doi:10.1037/0022-3514.74.3.774

· Crossref Search Google Scholar Export Citation

· Muthén, B., & Asparouhov, T. (2012). Bayesian structural equation modeling: A more flexible representation of substantive theory. Psychological Methods, 17(3), 313–335. doi:10.1037/a0026802

· Crossref Search Google Scholar Export Citation

· Muthén, B., & Asparouhov, T. (2013, January 11). BSEM measurement invariance analysis. Mplus web notes: No. 17. Retrieved from http://www.statmodel.com

· Search Google Scholar Export Citation

· Muthén, L.K., & Muthén, B.O. (1998–2016). Mplus user’s guide (7th ed.). Los Angeles, CA: Muthén & Muthén.

· Search Google Scholar Export Citation

· Muthén, L.K., & Muthén, B.O. (2002). How to use a Monte Calro study to decide on the sample size and determine power. Structural Equation Modeling, 9, 599–520. doi:10.1207/S15328007SEM0904_8

· Crossref Search Google Scholar Export Citation

· Pelletier, L.G., Rocchi, M.A., Vallerand, R.J., Deci, E.L., & Ryan, R.M. (2013). Validation of the revised sport motivation scale (SMS-II). Psychology of Sport and Exercise, 14(3), 329–341. doi:10.1016/j.psychsport.2012.12.002

· Crossref Search Google Scholar Export Citation

· Podsakoff, P.M., MacKenzie, S.B., Lee, J.Y., & Podsakoff, N.P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88(5), 879–903. PubMed ID: 14516251 doi:10.1037/0021-9010.88.5.879

· Crossref Search Google Scholar Export Citation

· Preacher, K.J. (2015). Advances in mediation analysis: A survey and synthesis of new developments. Annual Review of Psychology, 66(1), 825–852. doi:10.1146/annurev-psych-010814-015258

· Crossref Search Google Scholar Export Citation

· Raykov, T. (2009). Evaluation of scale reliability for unidimensional measures using latent variable modeling. Measurement and Evaluation in Counseling and Development, 42(3), 223–232. doi:10.1177/0748175609344096

· Crossref Search Google Scholar Export Citation

· Ryan, R.M., & Deci, E.L. (2000). Intrinsic and extrinsic motivations: Classic definitions and new directions. Contemporary Educational Psychology, 25(1), 54–67. PubMed ID: 10620381 doi:10.1006/ceps.1999.1020

· Crossref Search Google Scholar Export Citation

· Ryan, R.M., & Deci, E.L. (2008). From ego depletion to vitality: Theory and findings concerning the facilitation of energy available to the self. Social and Personality Psychology Compass, 2(2), 702–717. doi:10.1111/j.1751-9004.2008.00098.x

· Crossref Search Google Scholar Export Citation

· Stenling, A., Ivarsson, A., & Lindwall, M. (2017). The only constant is change: Analysing and understanding change in sport and exercise psychology research. International Review of Sport and Exercise Psychology, 10(1), 230–251. doi:10.1080/1750984X.2016.1216150

· Crossref Search Google Scholar Export Citation

· Tangney, J.P., Baumeister, R.F., & Boone, A.L. (2004). High self-control predicts good adjustment, less pathology, better grades, and interpersonal success. Journal of Personality, 72(2), 271–324. PubMed ID: 15016066 doi:10.1111/j.0022-3506.2004.00263.x

· Crossref Search Google Scholar Export Citation

· Toering, T.T., & Jordet, G. (2015). Self-control in professional soccer players. Journal of Applied Sport Psychology, 27(3), 335–350. doi:10.1080/10413200.2015.1010047

· Crossref Search Google Scholar Export Citation

· Tuk, M.A., Zhang, K., & Sweldens, S. (2015). The propagation of self-control: Self-control in one domain simultaneously improves self-control in other domains. Journal of Experimental Psychology: General, 144(3), 639–654. doi:10.1037/xge0000065

· Crossref Search Google Scholar Export Citation

· van de Schoot, R., Kaplan, D., Denissen, J., Asendorpf, J.B., Neyer, F.J., & van Aken, M.A.G. (2014). A gentle introduction to Bayesian analysis: Applications to developmental research. Child Development, 85(3), 842–860. PubMed ID: 24116396 doi:10.1111/cdev.12169

· Crossref Search Google Scholar Export Citation

· van de Schoot, R., Kluytmans, A., Tummers, L.G., Lugtig, P., Hox, J., & Muthén, B. (2013). Facing off with Scylla and Charybdis: A comparison of scalar, partial, and the novel possibility of approximate measurement invariance. Frontiers in Psychology, 4, 770. PubMed ID: 24167495 doi:10.3389/fpsyg.2013.00770

· Crossref Search Google Scholar Export Citation

van de Schoot, R., Lugtig, P., & Hox, J. (2012). A checklist for testing measurement invariance. European Journal of Developmental Psychology, 9(4), 486–492. doi:10.1080/17405629.2012.686740

AB #2

AB #3

Reducing child aggression through sports intervention: The role of self-control skills and emotions☆

Author links open overlay panel

KerenShacharaTammieRonen-RosenbaumaMichaelRosenbauma1HodOrkibibLiatHamamaa

Show more

https://doi.org/10.1016/j.childyouth.2016.11.012

Get rights and content

Highlights

•

Sports intervention linked with change in self-control skills (SCSs).

•

For girls, changes in SCSs linked directly to changes in physical aggression.

•

For boys, changes in SCSs linked indirectly to changes in physical aggression.

•

Boys indirect effect was through changes in positive and negative emotions.

Abstract

This study examined how sports intervention may reduce aggressive behaviors in children (Grades 3–6), focusing on the relations between acquisition of self-control skills (SCSs) and aggressive behavior through the mediation of thoughts (i.e., hostility) and emotions (i.e., positive and negative). In a sample of 649 Israeli children, 50% were assigned to an experimental group and the remainder to a waitlisted control group. As hypothesized, children in the experimental group reported significantly larger gains in SCSs and significantly larger decreases in physical aggression, hostile thoughts, and negative emotions. Results of structural equation modeling suggested that SCS gains were linked to changes in hostile thoughts, as mediated by changes in both positive and negative emotions. In addition, changes in hostile thoughts were linked to changes in physical aggression through the mediation of changes in anger. Among girls, changes in SCSs were linked directly to changes in physical aggression (with no indirect effect), whereas among boys, changes in SCSs were linked indirectly to changes in physical aggression, through changes in positive and negative emotions. Findings contribute to understanding of possible mechanisms underlying the associations between children's self-control and aggression, with particular implications for the roles of positive and negative emotions.

· Previous article

· Next article

Keywords

AggressionSelf-control skillsChildrenPositive and negative emotionHostilityAnger

1. Introduction

Aggression among children and adolescents is a serious social problem that has sharply escalated in recent years (Hoffman et al., 2011, World Health Organization, 2015). Studies show that nearly 40% of students in primary and secondary schools encounter incidences of taunting, harassment, or bullying at their schools (Benbenishty et al., 2002, Hoffman et al., 2011). Moreover, aggression also comprises one of the most common reasons underlying children's referrals to therapy (Kazdin & Weisz, 2010).

Researchers have noted that the decline of aggressive behavior in early childhood occurs during the same period when children's executive cognitive functions increase. Thus, gains in self-control skills (SCSs) enable better control over thoughts, emotions, and actions (Ellis, Weiss, & Lochman, 2009). Given this association, the current study aimed to examine whether changes in SCSs would reduce aggressive behavior and whether children's thoughts and emotions would play a mediating role.

Targeted SCS training programs have been implemented to help at-risk children control their anger and aggression (e.g., Ronen, 2004, Ronen and Rosenbaum, 2010). Yet, such explicit training opportunities require investment of resources that are often unavailable for general school populations. Thus, we examined whether children could be imparted with SCSs via a more generalizable and less stigmatic intervention, by involving them in sports activities that would better promote their ability to control aggressive tendencies, as well as to increase their positive emotions.

1.1. Aggression

Aggression is often defined as an overt social behavior that involves at least two people, in which aversive physical and/or verbal acts are delivered to others with the immediate intent to cause harm (Bushman & Huesmann, 2010). Tremblay and Nagin (2005) suggested that aggression is neither a behavior that is socially learned nor a drive that must be satisfied. Rather, it is an internal disposition that children need to learn to control. This view upholds the notion that aggression is and always will be part of the human behavioral repertoire, although it remains under control most of the time by most people (Tremblay & Nagin, 2005). In addition, this view highlights the shift from a focus on aggressive behavioral acts per se to a focus on aggressive traits and on the aggressive child as a whole (Hartup, 2005). Aggressive behavior, in particular among children, is thus seen as a social act aimed at achieving social goals such as gaining acceptance and recognition by others (Ayduk et al., 2000, Denson et al., 2012). Hence, considered in light of human beings' need to feel part of society and their fear of rejection, aggression is no longer seen as a mere behavior but rather as related to how people think and feel (Schwalbe, 2009).

In the current study, we adopt Buss and Perry's (1992) model of aggression, which depicts the tendency to respond aggressively when facing difficulties, rejection, or stress. This model of aggression combines cognitive, emotional, and behavioral components. The cognitive component refers to hostile thoughts, which are perceptions that the world is a menacing, unfair place where nobody can be trusted because everybody acts out of selfish motives. The emotional component refers to angry feelings, which comprise emotional responses to frustration, provocation, or occasionally anxiety – typically coupled with physiological arousal (Arsenio et al., 2000, Milaniak and Widom, 2015). The behavioral component refers to actual acts of physical aggression, which include inflicting injury on someone else with the intention of causing pain. Research has shown that these three components are interrelated in a sequence, so that hostile cognitions contribute to angry feelings, which in turn contribute to physical aggression; namely, anger acts as a mediator between hostile thoughts and acts of aggression (see Buss & Perry, 1992). In the present study, we examined this mediating chain – where anger mediates the link between hostile thoughts and physical aggression – within a larger chain that incorporated SCSs, positive emotions, and other negative emotions.

Regarding gender differences, there are consistent finding for higher levels of aggressive behavior in boys than in girls (Archer, 2004), including in Israeli adolescents (e.g., Gavriel-Fried et al., 2015).

1.2. Self-control

Our study draws on Rosenbaum's (1998) conceptualization of self-control as a set of goal-directed skills that enable human beings to act upon their aims; overcome difficulties relating to thoughts, emotions, and behaviors; delay gratification; and cope with distress. A considerable body of research has previously shown that adults, adolescents, and children who are high in self-control behaviors – such as postponing gratification, planning the future, and using cognitions to guide actions – are less likely to behave aggressively, and vice versa (e.g., Ayduk et al., 2000, Ronen and Rosenbaum, 2010, Weisbrod et al., 2007). Furthermore, past studies have shown that high SCSs are related to better coping abilities, better adjustment, higher general happiness, and higher levels of positive emotions like excitement and enthusiasm (Ronen and Seeman, 2007, Rosenbaum and Ronen, 2013). At the same time, research has pinpointed the link between low SCSs and high levels of negative emotions like irritability or guilt (Hamama and Ronen-Shenhav, 2012, Orkibi et al., 2014). Furthermore, prior controlled research conducted by Ronen and Rosenbaum (2010) indicated that imparting at-risk children and adolescents with SCSs in school was an effective training tool for reducing aggressive behavior.

With respect to gender differences, females have consistently been found to exhibit higher levels of self-control than males (Nolen-Hoeksema, 2012). A meta-analysis of research on children aged three months to 13 years reported large effect sizes for gender, with girls showing higher levels of self-control than boys (Else-Quest et al., 2012). Similar trends have been found regrading adolescents' emotional regulation (e.g., Calvete & Orue, 2012).

1.3. Positive and negative emotions

Past research pinpointed the link between low SCSs and high negative emotions (Hamama and Ronen-Shenhav, 2012, Orkibi et al., 2014) and positive associations between high SCSs and happiness and positive emotions (Gilbert, 2005, Hamama et al., 2013, Orkibi et al., 2014). In the current study, we examined if changes in SCSs would change emotions. Namely, we investigated whether promotion of higher positive emotions and lower negative emotions would in turn decrease aggressive behavior.

Positive emotions like excitement and negative emotions like despondency are components of subjective well-being and relate to the good and bad feelings, respectively, that people experience in daily life (Myers & Diener, 1995). Some researchers have suggested that positive and negative emotions may not be located on two ends of a single spectrum but rather may be distinct from one another, deriving from different physiological mechanisms (Busseri et al., 2012, Folkman, 2008, Fredrickson et al., 2003). This view implies that people can feel both at the same time; moreover, increasing positive emotions may promote happiness but may not help in decreasing negative emotions (Fredrickson, 2009).

In support of these distinctions, positive emotions were conceptualized as broadening thought-action repertoires, resulting in a higher likelihood of pursuing a wider range of thoughts and actions, because one can see more possibilities (Fredrickson, 2009). In contrast, negative emotions were shown to reduce recovery from adversity and to narrow people's momentary thought-action repertoires and perceived opportunities (Fredrickson, 2009, Magyar-Moe, 2009). Thus, in the current study, as illustrated by the bolded arrows in Fig. 1, we expected that changes in SCSs through a sports intervention would lead to changes (increases) in positive emotions and changes (decreases) in negative emotions, which in turn would broaden the participants' range of thoughts so as to reduce their hostile thoughts. Consequently, as illustrated with the double arrows in Fig. 1, changes in hostility was expected to lead to changes in the specific experience of relational-anger and aggressive behavior.

· Download : Download high-res image (299KB) Download : Download full-size image

Fig. 1. General model of the hypothesized mediators between changed self-control skills and changed physical aggression for the entire sample (N = 649), with standardized regression weights presented. In the Group paths, positive values indicate increases in intervention group (coded 1) compared to control group (coded 0), and negative values indicate decreases in intervention group compared to control group. Covariates and controlled variables (ethnicity and grade level) are omitted for clarity. Dashed paths are non-significant. Bolded and double arrows indicate the two hypothesized mediational chains. ⁎p < 0.05. ⁎⁎p < 0.01. ⁎⁎⁎p < 0.001.

Finally, with respect to gender differences, a meta-analysis of gender differences in children's and adolescents' emotion expression revealed significant but very small gender differences, with girls showing more positive emotions and internalizing emotions (e.g., sadness) than boys, and boys showing more externalizing emotions (e.g., anger) than girls (Chaplin & Aldao, 2013). In a sample of 185 Hebrew-speaking Jewish adolescents aged 15 to 17, girls reported higher levels of negative emotions than boys (Ben-Zur, 2003).

1.4. Sports intervention, self-control, and aggression

Many intervention programs have been developed for reducing aggressive behavior (e.g., Jimerson and Furlong, 2006, Kazdin, 2000), typically offering two approaches. The environmental approach focuses on guiding parents, educators, and caregivers to practice reinforcement, punishment, and boundary setting (Olweus, 2004, Rolider and Van Houten, 1995). The child-centered approach focuses on imparting the child with emotional, cognitive, and behavioral SCSs, social skills, and problem-solving skills (Crick and Dodge, 1994, Elias et al., 1997, Kazdin and Weisz, 2010), which were proven effective by numerous studies)Ronen, 1997, Ronen and Rosenbaum, 2010, Wilson et al., 2001). These child-centered interventions include efforts specifically designed to explicitly impart children with SCSs in order to reduce aggressive behavior (Ronen et al., 2007, Ronen and Rosenbaum, 2010). Yet, differently in the current study, SCSs were imparted implicitly, through child engagement in sports activities that require SCS application. Sports activities are organized, structured, and supervised, yet have the potential pleasurable benefits of leisure activities' situational context. Leisure activities have been associated with positive outcomes in diverse aspects of functioning among children and youth, such as intellectual development, psychological adjustment, and prevention of problem behaviors such as substance use (Wilson, Gottfredson, Cross, Rorie, & Connell, 2010).

The dictionary definition of sport is “a contest or game in which people do certain physical activities according to a specific set of rules and compete against each other” (Sport, 2015). Sports activities have been shown to cultivate SCSs and thereby better interpersonal relationships with peers and adults (e.g., Findlay and Coplan, 2008, Mahoney et al., 2005). The link between sports and SCSs was explained through cognitive mechanisms that include concentration, focus, and attention because sports games require compliance with laws and regulations as well as movements and actions that must be performed in a specific order with precise timing (Marsh and Kleitman, 2003, Miller et al., 2005). Other possible mechanisms underlying the link between sports and SCSs may include problem solving because while playing sport games children must make rapid decisions under pressure, as well as the development of intelligence because sports encourage the child to think quickly, consciously, and creatively (Marsh and Kleitman, 2003, Miller et al., 2005). Nevertheless, with regard to gender differences, boys and girls still do not participate equally in sport activities. Studies have shown that girls participate in a substantially fewer number of sport activities than boys. Moreover, they have different patterns of proclivity: compared to girls, boys are more likely to engage in sport-related activities that involve teams, complex rule systems and high levels of interdependency (Eccles and Harold, 1991, Dietz-Uler et al., 2000, Vilhjalmsson and Kristjansdottir, 2003).

Additionally, studies from a wide range of disciplines such as medicine, psychology, sociology, and education indicate that routine sports activity among children and adolescents has a positive effect on physical and mental health (Gaya et al., 2011) and well-being (Donaldson & Ronan, 2006). Research studies have also indicated that planned sports activities are effective in reducing participants' aggressive behaviors (Fleming et al., 2008, Lufi and Parish-Plass, 2011). Some studies, however, have found that male athletes report significantly more aggressive behavior than their non-athlete peers (e.g., Rhea & Lantz, 2004).

Studies demonstrated that sports activities are negatively linked with negative emotions, positively linked with positive emotions (Hyde, Conroy, Pincus, & Ram, 2011), and positively linked with better social functioning and higher social status in adolescents (Simpkins et al., 2008, Thirer and Wright, 2010). Thus, rather than developing the hostility and anger that may lead to aggression, children who participate in sports activities may experience positive emotions and happiness that may reduce aggression.

1.5. Study hypotheses

The current study examined how sports intervention may reduce aggressive behaviors among elementary school children, by studying the relationship between SCSs' acquisition and a reduction in aggressive behavior through mediating variables. We tested the following two hypotheses:

Intervention effects comparing a sports intervention (experimental) group and a waitlisted (control) group: Following the sports intervention, children in the experimental group will report significantly larger increases in SCSs and positive emotions and significantly larger decreases in aggressive behaviors, hostile thoughts, anger, and negative emotions, compared to children in the control group. These changes will emerge in both children's self-reports and teachers' reports.

Mediation model: Participation in the sports intervention will contribute to changes (increases) in SCSs, which in turn will decrease aggressive behavior through a chain of mediating variables. An increase in SCSs will reduce hostile thoughts, through the mediation of increased positive emotions and decreased negative emotions (see bolded arrows in Fig. 1). In turn, a decrease in hostile thoughts will reduce aggressive behavior, through the mediation of decreased anger (see double arrows in Fig. 1).

2. Method

2.1. The sports intervention

This study was part of the assessment process for one of several Israeli social-educational projects established by the Rashi Foundation (http://www.rashi.org.il/) in collaboration with and with funding from the Children and Youth at Risk Foundation (National Insurance Institute) and the education departments of participating local municipalities. The project administrators (not the research team) selected 39 typical schools from Israel's underprivileged peripheral geographical areas to participate in the project: 25 from the north and 14 from the south of Israel. The project administrators selected 21 of the schools to receive a sports intervention over one academic year (the experimental group) and placed the remaining 18 schools on a waiting list to receive the intervention the following year (the control group). The experimental group received a total of 120 h of extra afterschool sports activities delivered over 24 weeks, comprising two weekly hours of martial arts and three weekly hours of other group sports activities (e.g., soccer, basketball, volleyball, mini-football, capoeira). The trainers, parents, and children were aware of the aim of the project: to enhance self-control skills. In each school, two certified sports trainers (i.e., coaches), who were not part of the regular school staff, delivered the activities; one specialized in martial arts and the other in group sports activities. Fidelity of intervention delivery was monitored by a school coordinator who reported on a regular basis to the project's regional supervisor.

2.2. Sample

In accordance with the project's criteria, each elementary school's staff (e.g., principal, homeroom teacher, psychologist, counselor) selected students in Grades 3–6 to participate in the project. Inclusion criteria were: the child's demonstration of observed aggressive behavior, no therapeutic treatment inside or outside school, and the child's self-declaration that s/he liked sports. Of the 1047 children selected for participation in the project, the final sample included 649 children (61%) – after removing children who dropped out (due to time constraints, moving, or disinterest in the sports activities) and children who did not have complete data sets (self-reports and teacher reports at pretest and posttest). Of the final sample (n = 649), 70% children were from the north and 30% were from the south of Israel; 75.8% were boys; and 47% were Jews and 53% were Arabs.

Of this sample, 330 children were in the experimental group (82% boys, 40% Arabs, 40% fifth graders) and 319 children were in the waitlisted control group (70% boys, 60% Arabs, 35% fourth graders). To match these two groups, we divided the baseline ranges of physical aggression in the experimental group (using Buss & Perry's, 1992 Aggression Questionnaire) and matched the control group to these proportions, thereby creating two comparison groups with equal distributions of baseline physical aggression. Preliminary chi-square analyses revealed statistically significant differences between the experimental and control groups on the three aforementioned demographic variables: sex, χ2(1) = 14.59, Jewish or Arab ethnicity, χ2(1) = 37.48, and grade level, χ2(3) = 24.66 (all ps < 0.001). These differences were accounted for in further analyses as indicated in the Results section.

2.3. Procedure

The study received approval from the Chief of Research at the Israel Ministry of Education and the ethical committee of Tel-Aviv University. Recruitment procedures included several steps. First, after schools identified the children who met the aforementioned inclusion criteria, the first author invited children and their parents to personal meetings with a research team member, who explained the project's goals (i.e., to improve behavior) and structure. The children were assured of complete confidentiality and of withdrawal options from the research and from the sports activities at any time without negative consequences. Children gave their verbal consent to participate, and parents gave written consent for their child to participate in the project and the research.

Data were collected from all children and homeroom teachers in all 39 schools twice: at the beginning of the program and at its end 24 weeks later. Research assistants administered the three self-report questionnaires in groups of 2 to 5 children in regular classrooms (completion duration ranged from 20 to 45 min). Each child was assigned a numerical code that was securely guarded by the research team. For each child who participated, the homeroom teacher completed aggression ratings using the student's corresponding numerical code (completion duration of ~ 10 min per student).

2.4. Instruments

All measures were administered to children and teachers in their mother tongue. All instruments completed by the participants were Hebrew and Arabic versions of standardized measures that had previously been administered in research on Israeli Jewish and Arab children/teachers and that showed good psychometric properties.

2.5. Self-reported emotions

The 20-item Positive and Negative Affect Schedule (PANAS; Watson, Clark, & Tellegan, 1988), a self-report checklist of affect adjectives, was designed to provide independent measures of positive and negative affect. Respondents ranked each adjective on a 5-point scale ranging from 1 (very few times) to 5 (a lot of times). The subscales' reliabilities ranged from 0.83 to 0.90 for positive affect and from 0.85 to 0.90 for negative affect. The original correlation between positive and negative affect ranged from − 0.05 to − 0.35. Test–retest reliability (for Times 1 and 2 in the original study) was 0.88 (p < 0.01) for both the positive affect and negative affect subscales. Both subscales demonstrated good convergent and discriminant validity with existing self-report measures of childhood anxiety and depression (Laurent et al., 1999). In the present study, Cronbach alphas were 0.70 for positive affect and 0.78 for negative affect at pretest and 0.75 for positive affect and 0.82 for negative affect at posttest.

2.6. Self-reported SCSs

The 32-item Self-Control Skills Scale (SCS; Rosenbaum & Ronen, 1991) was designed to assess problem-solving skills, attentional control (i.e., distraction), cognitive reframing, delay of gratification, and use of self-talk and self-reinforcement. Participants rated items on a scale ranging from 1 (not characteristic of me at all) to 6 (very characteristic of me), with higher scores indicating higher SCSs. Evidence was established for convergent validity through significant positive correlations with a measure of positive automatic thoughts (r = 0.34, p < 0.001) and for divergent validity through significant negative correlations with depressive symptoms (r = − 0.37, p < 0.001; Zauszniewski, Bekhet, & Bonham, 2010) and with aggression (r = − 0.48, p < 0.01; Ronen & Rosenbaum, 2010). Test–retest reliability in Ronen and Rosenbaum (2010) was 0.46 (p < 0.01), and internal consistency (Cronbach α) in the present study was 0.73 at pretest and 0.81 at posttest. The scale revealed good validity in terms of correlations with a large range of other scales (see detailed report in Rosenbaum, 1980).

2.7. Children's self-reported aggression

The 29-item Aggression Questionnaire (Buss & Perry, 1992) includes four self-report subscales: hostile thoughts, anger, verbal aggression, and physical aggression, which can be aggregated for a general aggression score or calculated separately. In the present study, we only used the hostile thoughts, anger, and physical aggression subscales. The hostile thoughts subscale included 8 items (e.g., “I sometimes feel that people are laughing at me behind me back”), and the anger subscale included 7 items (e.g., “Some of my friends think I am a hothead”). The physical aggression subscale included 9 items (e.g., “If somebody hits me, I hit back”), but we omitted an item (#7) to increase the scale's internal consistency. Respondents rated items on a scale ranging from 1 (strongly disagree) to 6 (strongly agree), with higher scores indicating a greater tendency toward hostility, anger, or physical aggression. According to Buss and Perry (1992), the questionnaire's test-retest reliability was r = 0.80, and its internal consistency was α = 0.89. In the present study, reliabilities (Cronbach α) were 0.85 at pretest and 0.84 at posttest for overall aggression, 0.78 at pretest and 0.76 at posttest for physical aggression, 0.60 at pretest and 0.56 at posttest for anger, and 0.74 at pretest and 0.75 at posttest for hostility.

2.8. Teacher-rated child aggression

In the current study, homeroom teachers completed the Eyberg Child Behavior Inventory (ECBI; Robinson, Eyberg, & Ross, 1980) at pretest and posttest, to cross-validate students' self-reports on the Aggression Questionnaire. The 36-item ECBI was originally designed as a parent rating scale to measure disruptive behavior in children ages 2–16 years, including physical behavior (e.g., fights with friends) and verbal behavior (i.e., teases or provokes other children), and it underwent previous adaptation to teachers (Treves-Pelleg, 2003). For each item, teachers rated how often each child displayed that behavior on a scale ranging from 1 (never) to 5 (always), with the total score reflecting the severity of the disruptive behavior in terms of its frequency. The scale showed reliabilities (Cronbach α) of 0.95 at pretest and 0.94 at posttest.

2.9. Data analysis

First, we computed Pearson correlations among all study variables for the entire sample as well as separately for boys and for girls. Second, as preliminary analysis to assess differences in baseline scores of the experimental and control groups by sex and ethnicity (Arab/Jewish), we conducted a three-way multivariate analysis of covariance (MANCOVA), while controlling for grade level as a covariate. Third, to assess the intervention effects, we calculated the pretest-posttest differences using residualized gain scores – defined by the residual of the linear regression of pretest on posttest – in the experimental and control groups (Maris, 1998). To assess differences in the gain scores of boys and girls in the experimental and control groups, we conducted a two-way MANCOVA, while controlling for ethnicity (Jewish/Arab) and grade level as covariates. Fourth, using Amos 23 and structural equation modeling (SEM), we conducted path analysis to test the study's theoretical model. We used multigroup SEM to test this model for boys versus girls (Byrne, 2004). We evaluated the models' fit to the data using the criteria of χ2/df ≤ 3, comparative fit index (CFI) ≥ 0.95, Tucker-Lewis coefficient (TLI) ≥ 0.95, and root mean square error of approximation (RMSEA) < 0.80 (Schreiber, Nora, Stage, Barlow, & King, 2006). We used the bootstrap test method for indirect effects, with a confidence level (CI) set at 95% and bootstrap bias-corrected samples set at 5000 (Preacher & Hayes, 2004).

3. Results

3.1. Pearson correlations

As seen in Table 1, correlations were in the expected directions. SCSs significantly correlated positively with positive emotions and negatively with negative emotions and with physical aggression, for boy and for girls. However, whereas SCSs correlated negatively with anger only for boys, it correlated negatively with hostility only for girls.

Table 1. Intercorrelations among study variables in boys and girls.

	Self-control skills	Positive emotions	Negative emotions	Anger	Hostility	Physical aggression
		Girls (n = 157)
Self-control skills		0.26⁎⁎	− 0.16⁎⁎	− 0.09	− 0.12⁎	− 0.13⁎
Positive emotions	0.23⁎⁎		− 0.12⁎	− 0.09	− 0.08	− 0.12⁎
Negative emotions	− 0.26⁎⁎	− 0.24⁎⁎		0.17⁎⁎	0.13⁎	0.12⁎
Anger	− 0.22⁎⁎	− 0.06	0.41⁎⁎		0.41⁎⁎	0.36⁎⁎
Hostility	− 0.04	− 0.20⁎⁎	0.33⁎⁎	0.42⁎⁎		0.36⁎⁎
Physical aggression	− 0.24⁎⁎	− 0.14⁎	0.37⁎⁎	0.47⁎⁎	0.32⁎⁎
	Boys (n = 492)

N = 649.

⁎

p < 0.05.

⁎⁎

p < 0.01.

3.2. Preliminary analysis

A three-way MANCOVA was computed to examine intergroup differences in the baseline scores by sex and ethnicity, with grade level as a covariant. As can be seen in Table 2, a significant difference emerged only for SCSs. Post hoc Bonferroni correction test indicated three significant differences: Arab girls in the control group significantly differed from Jewish boys in the experimental group (p < 0.001), from Jewish boys in the control group (p < 0.01); and from Arab boys in the control group (p < 0.01). Given these findings, all further analyses controlled for ethnicity differences.

Table 2. Three-way MANCOVA results for baseline differences by group, ethnicity, and sex.

Experimental group (n = 330)

Control group (n = 319)

Variable

Arab

Jewish

Arab

Jewish

Group ×

Ethnicity ×

sex

Girls

Boys

Girls

Boys

Girls

Boys

Girls

Boys

(n = 19)

(SD)

(n = 117)

(SD)

(n = 40)

(SD)

(n = 154)

(SD)

(n = 59)

(SD)

(n = 149)

(SD)

(n = 39)

(SD)

(n = 72)

(SD)

Self-control

skills

20.74

(20.23)

22.28

(19.77)

22.22

(24.57)

17.49

(23.33)

31.95

(22.13)

17.08

(22.30)

18.08

(22.29)

15.46

(25.86)

4.91⁎

Positive

emotions

40.63

(5.70)

38.56

(6.48)

38.87

(6.63)

38.88

(7.42)

40.14

(5.61)

38.47

(7.12)

40.64

(5.65)

38.77

(5.60)

0.50

Negative

emotions

23.00

(6.11)

21.75

(7.18)

19.55

(8.58)

21.34

(8.68)

22.50

(8.15)

25.23

(6.89)

18.70

(7.40)

20.87

(6.93)

1.37

Anger

19.93

(6.40)

23.82

(6.18)

19.95

(6.85)

21.15

(7.55)

21.56

(6.91)

24.11

(6.81)

17.23

(8.25)

19.95

(6.94)

1.16

Hostility

21.05

(8.82)

27.51

(8.87)

25.12

(9.36)

25.35

(8.93)

25.40

(8.68)

25.37

(8.86)

25.80

(9.26)

25.09

(10.11)

1.36

Physical

aggression

23.11

(6.67)

28.69

(9.15)

16.85

(7.28)

23.90

(9.25)

22.72

(8.96)

27.68

(9.61)

18.70

(7.45)

23.97

(8.12)

0.41

Teacher-rated

child behavior

61.95

(25.20)

73.69

(30.61)

62.37

(24.42)

75.64

(27.32)

66.32

(25.09)

80.79

(27.95)

51.11

(22.96)

63.01

(25.41)

0.02

⁎

p < 0.05.

3.3. Intervention effects

Hypothesis 1

posited that children in the sports intervention group would report significantly larger increases in SCSs and positive emotions (indicated by positive mean gain scores) and significantly larger decreases in aggressive behaviors (both self-reported and teacher-reported), hostile thoughts, anger, and negative emotions (indicated by negative mean gain scores), compared to children in the control group. As shown in Table 3, the two-way MANCOVA controlling for ethnicity and grade as covariates, which we conducted to assess differences in these change scores of boys and girls in the experimental and control groups, yielded results that largely confirmed Hypothesis 1. Compared to children in the control group, the children in the experimental group reported significantly larger increases in SCSs (but not in positive emotions) and significantly larger decreases in physical aggression, hostile thoughts, anger, and negative emotions. These self-reported changes in aggression were corroborated by the teachers' reports of significantly less frequent behavioral problems for children in the intervention group compared to those in the control group.

Table 3. MANCOVA results for intergroup differences on change scores by sex.

Experimental group (N = 330)

Control group (N = 319)

F (ηp2)

Variable

Total

Girls

Boys

Bon-ferroni

Total

Girls

Boys

Bon-ferroni

Group

Sex

Group ×

sex

MGS

(SD)

MGS

(SD)

MGS

(SD)

MGS

(SD)

MGS

(SD)

MGS

(SD)

Self-control

skills

3.35

(22.80)

3.13

(22.70)

3.40

(22.84)

− 3.85

(22.67)

− 1.81

(24.84)

− 4.65

(21.76)

8.06⁎⁎

(0.012)

0.29

(0.000)

0.43

(0.001)

Positive

emotions

1.24

(5.90)

− 1.08

(6.20)

1.75

(5.70)

⁎⁎⁎

− 1.30

(6.60)

0.12

(6.90)

− 1.86

(6.40)

⁎

3.00

(0.005)

0.68

(0.001)

16.69⁎⁎⁎

(0.025)

Negative

emotions

− 1.68

(7.33)

− 2.41

(5.12)

− 1.53

(7.73)

2.00

(7.20)

1.30

(7.30)

2.26

(7.20)

17.04⁎⁎⁎

(0.026)

0.79

(0.001)

0.00

(0.000)

Anger

− 1.76

(5.80)

− 1.64

(5.66)

− 1.79

(5.90)

1.90

(6.20)

1.30

(6.10)

2.18

(6.22)

19.52⁎⁎⁎

(0.029)

0.00

(0.000)

0.77

(0.001)

Hostility

− 2.00

(7.85)

− 0.45

(7.60)

− 2.33

(7.90)

2.26

(8.50)

1.40

(8.10)

2.60

(8.60)

17.02⁎⁎⁎

(0.026)

0.28

(0.000)

3.70

(0.006)

Physical

aggression

− 2.86

(7.30)

− 2.70

(6.20)

− 2.89

(7.50)

3.10

(8.50)

0.60

(8.20)

4.10

(8.40)

⁎⁎⁎

45.00⁎⁎⁎

(0.065)

4.86⁎

(0.008)

5.90⁎

(0.009)

Teacher-rated child behavior

− 4.50

(17.70)

0.81

(16.50)

− 5.60

(17.80)

⁎

5.50

(21.05)

− 2.45

(15.60)

8.63

(22.10)

⁎⁎⁎

7.07⁎⁎

(0.011)

1.41

(0.002)

23.72⁎⁎⁎

(0.036)

Note. Controlling for ethnicity and grade as covariates. MGS = Mean gain score, where positive MGS indicates increase in value; negative MGS indicates decrease in value. ηp2 = partial eta square, indicating effect size, ns = non-significant.

⁎

p < 0.05.

⁎⁎

p < 0.01.

⁎⁎⁎

p < 0.001.

Post hoc analyses revealed significant sex differences for two change measures in the experimental group: Self-reports of positive emotions showed a decrease in girls but an increase in boys, and teacher-reported child behavior problems showed an increase for girls but a decrease for boys. As shown in Table 3, with respect to the interaction of group and sex, intergroup differences as a function of sex emerged for the change scores in positive emotions, physical aggression, and teacher-reported child behavior problems.

3.4. Mediation model

Hypothesis 2

posited that participation in the yearlong sports intervention would decrease aggressive behavior through a chain of mediating variables. We hypothesized that gains in SCSs would produce changes (reductions) in hostile thoughts, through the mediation of changes (increases) in positive emotions and changes (reductions) in negative emotions (see bolded arrows in Fig. 1). In turn, we posited that changes (reductions) in hostile thoughts would produce changes (reductions) in physical aggression, through the mediation of changes (reductions) in anger (see double arrows in Fig. 1).

3.4. Model for the entire sample

We first tested Hypothesis 2 on the entire sample (N = 649) using SEM analysis of intervention effect on mean change scores and the theoretical mediated model of mechanisms underlying the intervention effects. Ethnicity (Jewish/Arab) and grade level (3–6) were included in the model to control for their potential effect. Analysis revealed that the mediational model depicted in Fig. 1 provided a good fit to the observed data on the following fit indices: CFI = 0.996, TLI = 0.979, RMSEA = 0.027 (90% CI = [0.000, 0.060]), χ2(7) = 10.42, and χ2/df = 1.488 (p = 0.166).

Note that the paths between group (intervention/control) and the model's variables that show positive values suggest increases in SCS and positive emotions for the intervention group (coded 1) compared to the control group (coded 0). The negative values suggest decreases in negative emotions, hostile thoughts, and physical aggression for the intervention group compared to the control group. As indicated by the dashed path, the direct effect of the intervention on anger was not significant.

Regarding indirect effects (mediation), the bias-corrected bootstrap analysis of indirect effects indicated two full mediations in the model. As seen in the bolded arrows in Fig. 1, gains in SCSs were linked to changes in hostile thoughts, through the mediation of changes in both positive emotions and negative emotions (95% CI = − 0.136, − 0.059, p < 0.001). As the confidence intervals (CI) did not include zero, the null hypothesis of no mediation was rejected. The direct path between SCSs and hostile thoughts was not significant, indicating full mediation (p = 0.636). In addition, as seen in the double arrows in Fig. 1, the changes in hostile thoughts were linked to changes in physical aggression through the mediation of changes in anger (95% CI = 0.270, 0.433, p < 0.001). As the intervals did not include zero, the null hypothesis of no mediation was rejected. The direct path between hostile thoughts and physical aggression was not significant, indicating full mediation (p = 0.771).

3.5. Sex differences

We conducted multigroup SEM to test whether the model differed for girls and boys, constructing two models for comparison: an unconstrained model that posited variance across sex groups and a fully constrained model that posited equality (i.e., invariance in parameter estimates) across the groups. Chi-square test results confirmed that the constrained and unconstrained models differed significantly: χ2(32) = 77.83, p < 0.001 (see Fig. S1 and S2 in the online supplemental material for distinct models for girls and boys).

Fit indices of the separate model for girls (n = 149) were: CFI = 0.999, TLI = 0.992, RMSEA = 0.016 (90% CI = 0.000, 0.103), χ2(7) = 7.26, and χ2/df = 1.037 (p = 0.402). Close examination of this model indicated that the findings for girls did not support the hypothesized mediation model. Namely, the girls' increases in SCSs failed to reduce their hostile thoughts indirectly, through the mediation of increased positive emotions and decreased negative emotions (95% CI = − 0.201, 0.088, p = 0.428). Also, the direct path between changes in SCSs and changes in hostile thoughts was non-significant (p = 0.724). In addition, girls' decreases in hostile thoughts failed to reduce their physical aggression indirectly, through the mediation of decreased anger (95% CI = − 1.276, 0.150, p = 0.350). The direct path between changes in hostile thoughts and changes in physical aggression was also non-significant (p = 0.190). However, a significant direct path did emerge between girls' changes in SCSs and changes in physical aggression, suggesting that the former reduce the latter directly for girls (β = − 0.29, p = 0.012).

Fit indices of the separate model for boys (n = 500) were: CFI = 0.988, TLI = 0.939, RMSEA = 0.049 (90% CI = 0.014, 0.083), χ2(7) = 15.40, and χ2/df = 2.200 (p = 0.031). Close examination of this model indicated that the findings for boys lent partial support to the hypothesized mediation model, only for the first chain and not for the second. Specifically, boys' increases in SCSs did reduce their hostile thoughts as hypothesized, through the mediation of increased positive emotions and decreased negative emotions (95% CI = − 0.183, − 0.063, p < 0.001). The non-significance of the direct path between SCSs and hostile thoughts (p = 0.502) indicates full mediation. However, like for the girls, boys' decreases in hostile thoughts failed to reduce their physical aggression indirectly, through the hypothesized mediation of decreased anger (95% CI = − 1.073, 0.075, p = 0.141). Interestingly, unlike for the girls, the boys' direct path between changes in hostile thoughts and changes in physical aggression was significant (β = 0.62, p = 0.003). Finally, an overall significant indirect effect emerged between changes in SCSs and changes in physical aggression in boys (95% CI = − 0.247, − 0.045, p = 0.004) but, unlike for the girls, their direct path was not significant (β = − 0.15, p = 0.755), suggesting that the underlying mechanism may differ for boys versus girls.

In sum, girls' changes in SCSs emerged as linked directly to changes in physical aggression – there was no indirect (mediation) effect. Among boys, changes in SCSs linked indirectly to changes in physical aggression, through the parallel mediators of changes in positive and negative emotions – there was no direct link.

4. Discussion

Focusing on elementary school children with aggressive behavior, the current study attempted to achieve two goals. First, we aimed to demonstrate the effectiveness of sports activity intervention for reducing children's aggressive behavior. Second, we sought to empirically test a model explaining underlying indirect mechanisms for the reduction of aggressive behavior. Specifically, our model proposed, first, that gains in SCSs would reduce hostile thoughts through two parallel mediators (decreased general negative emotions and increased positive emotions) and, second, that those decreased hostile thoughts would eventually reduce aggressive behavior through the mediation of reduced angry feelings.

In line with our expectations, we found that after experiencing an academic year of afterschool sports activities, five times per week, children in the experimental group reported larger increases in SCSs and larger reductions in physical aggression, hostile thoughts, anger, and negative emotions in comparison to children from the waitlisted control group. This finding is of great importance, indicating that SCSs can be enhanced in at-risk children through cost-effective afterschool sports programs, to enable children to better act upon their aims and overcome difficulties related to thoughts, emotions, and behaviors (Rosenbaum, 1998, Rosenbaum, 2000). Moreover, our findings demonstrate that these significant gains in SCSs contributed to a chain of mediating mechanisms that led to reductions in aggressive behavior. Therefore, it appears that reducing children's aggressive behavior can be accomplished without explicit targeted training – by engaging children in active participation in generalizable, non-stigmatic sports activities, which in turn increase their SCSs.

The effects found for the current sports intervention are generally consistent with past findings on the positive correlations between sports activities and SCSs (Eldar et al., 2006, Findlay and Coplan, 2008, Reynes and Lorant, 2004). Possible explanations for the impact of sports on SCSs may include their shared underlying cognitive mechanisms like focus and concentration, and their requirements for rule compliance, sequenced body movements, problem-solving skills, and decision-making under pressure) Marsh and Kleitman, 2003, Miller et al., 2005). Future researchers would do well to investigate the relative effectiveness of additional possibilities for reducing aggression in children, such as active engagement in creative arts programs (e.g., music, drama, art; see Azoulay & Orkibi, 2015). Such interventions may likewise enhance SCS development (and the chain of mediated effects) without the stigma often associated with explicit therapeutic interventions. However, further research is necessary to pinpoint the structural and content characteristics of such implicit SCS-promoting interventions in order to determine possible underlying change factors, particularly because the changes appeared to differ for girls and boys. The sex differences highlight the importance of conducting separate analyses for males and females to enable cross-study comparisons and to scrutinize how sex differences may influence the links between children's SCS gains and their reductions in aggressive behavior.

The current finding of mediation for boys but nor for girls contradicts that of Weisbrod, Rosenbaum, and Ronen (2007), who reported mediation for adolescent girls but not for boys. Namely, in their study aggression among adolescent girls derived from hostile thoughts and anger such that when girls experienced hostile thoughts or anger they responded in verbal and physical aggression. However, this link did not emerge for adolescent boys, who were physically aggressive even if they did not report anger. A possible explanation for the different outcomes might relate to participants' ages – in the current study the participants were younger (Grades 3–6) than those in the abovementioned study. Hence, it might be that for younger boys the nature of the interventions appealed more than to girls; namely sport activities which included martial arts and group sports. This speculation echoes Lakes and Hoyt's (2004) study where the effects of martial arts for boys were larger than the effects for girls on many scales (e.g., focus, concentration, attention, obedience to adults, aggression toward other children, and anger). Moreover they suggested that martial arts intervention is especially well suited to promoting boys' self-regulation in cognitive and behavioral areas. Thus, in order to better understand the sex differences outcomes further comparative research is warranted with possibly different age trajectories.

Finally, we found that girls' changes in SCSs linked directly to their changes in physical aggression (without a mediation effect), whereas boys' changes in SCSs linked indirectly to changes in physical aggression, through the parallel mediators of changes in positive and negative emotions. This finding may be due to the empirically-based notion that girls tend to use more indirect (social and verbal) forms of aggression (Lansford et al., 2012) – which were not measured in this study but should be considered in the future.

4.1. Study limitations

Several limitations of the present study should be considered. First, given this study's quasi-experimental nature, the allocation of children to the experimental and control groups was based on the school teams' referrals rather than on randomization, although we matched the scores on the main targeted variable of physical aggression. Thus, the current sample was a non-probabilistic convenience sample because the project administrators purposefully selected the participating schools. This lack of a representative sample limits our findings' generalizability. Second, the 39% dropouts and noncompleters of post-test is another limitation. Third, due to the lack of follow-up data it is not clear whether the sports intervention had a short or long lasting effect. Fourth, it is possible that unmeasured process variables may have also affected the outcome. Based on psychotherapy process-outcome studies, these may include the working alliance between the student and the trainer, student level of engagement, the group cohesion level, and parental support, to name a few (see Kazdin, 2009). It would also be interesting to examine differences in outcomes between abused versus non- abused children given the documented links between exposure to parent-child aggression and child peer-directed aggression (Agbaria, Hamama, Orkibi, Gabriel-Fried, & Ronen, 2016). Finally, while self-reports are often used to assess children's subjective emotional experiences, additional sources are recommended for outcome data on overt physical aggression. Whereas in this study we collected self-reported and teacher reported data on children's physical aggression at the beginning and end of the intervention, future studies would do well to include parental reports and peer ratings, as well as physiological data, to provide valuable additional information. Such methodology may help account for the social desirability bias that may influence children's responses.

5. Conclusion

In conclusion, although the role of SCSs in overcoming difficulties has generally been documented for youth (Gavriel-Fried et al., 2015, Hamama et al., 2013, Orkibi and Ronen, 2015, Ronen et al., 2013, Ronen and Rosenbaum, 2010), our findings regarding the direct and indirect links between SCSs and aggressive behavior are of particular importance. Theoretically, this study contributes to understanding of possible mechanisms underlying the associations between boys' self-control and aggression, with particular implications for the role played by positive and negative emotions. In other words, it seems that for boys, SCS gains can help to reduce negative emotion and to increased positive emotion, which can lead to broadening of those thought-action repertoires (Fredrickson, 2009) that may consequently bring about reductions in physically aggressive behavior.

On the practical side, the present study points to the feasibility of implementing sports programs within the regular school curriculum, as an effective non-stigmatic means for reducing students' aggressive behavior. Given the typical availability of sports within the school setting, applying such a program can be more economical than applying individual or group interventions with specialized professionals. The sport activities program can be viewed as interesting, enjoyable, challenging, and rewarding, while still accomplishing the objective of enhancing SCSs in at-risk children.

Finally, the current findings hold implications related to sex differences. Namely, it is important to consider sex preferences in sport activities when designing a unified intervention and to examine differences in manifestations of SCSs between girls and boys. Using a broad range of assessment instruments appears to be necessary to better understand the effects of sport activities on boys' and girls' SCSs, emotions (positive and negative), and aggressive behavior.

Appendix A. Supplementary data

Download : Download Acrobat PDF file (118KB)

Supplementary figures

References

Agbaria et al., 2016

Q. Agbaria, L. Hamama, H. Orkibi, B. Gabriel-Fried, T. Ronen

Multiple mediators for peer-directed aggression and happiness in Arab adolescents exposed to parent–child aggression

Child Indicators Research, 9 (3) (2016), pp. 785-803, 10.1007/s12187-015-9346-y

CrossRef View Record in Scopus Google Scholar

Archer, 2004

J. Archer

Sex differences in aggression in real-world settings: A meta-analytic review

Review of General Psychology, 8 (4) (2004), p. 291, 10.1037/1089-2680.8.4.291

CrossRef View Record in Scopus Google Scholar

Arsenio et al., 2000

W.F. Arsenio, S. Cooperman, A. Lover

Affective predictors of preschoolers' aggression and peer acceptance: Direct and indirect effects

Developmental Psychology, 36 (2000), pp. 438-448

View Record in Scopus Google Scholar

Ayduk et al., 2000

O. Ayduk, R. Mendoza-Denon, W. Mischel, G. Downey

Regulating the interpersonal self: Strategic self-regulation for coping with rejection sensitivity

Journal of Personality and Social Psychology, 79 (5) (2000), pp. 776-792

View Record in Scopus Google Scholar

Azoulay and Orkibi, 2015

B. Azoulay, H. Orkibi

The four-phase CBN psychodrama model: A manualized approach for practice and research

The Arts in Psychotherapy, 42 (2015), pp. 10-18, 10.1016/j.aip.2014.12.012

Article Download PDF View Record in Scopus Google Scholar

Benbenishty et al., 2002

R. Benbenishty, R.A. Astor, A. Zeira, A.D. Vinokur

Perceptions of violence and fear of school attendance among junior high school students in Israel

Social Work Research, 26 (2) (2002), pp. 71-88

CrossRef View Record in Scopus Google Scholar

Ben-Zur, 2003

H. Ben-Zur

Happy adolescents: The link between subjective well-being, internal resources, and parental factors

Journal of Youth and Adolescence, 32 (2) (2003), pp. 67-79, 10.1023/a:1021864432505

View Record in Scopus Google Scholar

Bushman and Huesmann, 2010

B.J. Bushman, L.R. Huesmann

Aggression

S.T. Fiske, D.T. Gilbert, G. Lindzey (Eds.), Handbook of social psychology (5th ed.), John Wiley, New York (2010), pp. 833-863

View Record in Scopus Google Scholar

Buss and Perry, 1992

A.H. Buss, M. Perry

The aggression questionnaire

Journal of Personality and Social Psychology, 63 (1992), pp. 452-459

View Record in Scopus Google Scholar

Busseri et al., 2012

M.A. Busseri, B.L. Choma, S.W. Sadava

Subjective temporal trajectories for subjective wellbeing

The Journal of Positive Psychology, 7 (2012), pp. 1-15, 10.1080/17439760.2011.565784

CrossRef View Record in Scopus Google Scholar

Byrne, 2004

B.M. Byrne

Testing for multigroup invariance using AMOS graphics: A road less traveled

Structural Equation Modeling: A Multidisciplinary Journal, 11 (2) (2004), pp. 272-300, 10.1207/s15328007sem1102_8

CrossRef View Record in Scopus Google Scholar

Calvete and Orue, 2012

E. Calvete, I. Orue

The role of emotion regulation in the predictive association between social information processing and aggressive behavior in adolescents

International Journal of Behavioral Development, 36 (5) (2012), pp. 338-347, 10.1177/0165025412444079

CrossRef View Record in Scopus Google Scholar

Chaplin and Aldao, 2013

T.M. Chaplin, A. Aldao

Gender differences in emotion expression in children: A meta-analytic review

Psychological Bulletin, 139 (4) (2013), pp. 735-765, 10.1037/a0030737

CrossRef View Record in Scopus Google Scholar

Crick and Dodge, 1994

N.R. Crick, K.A. Dodge

A review and reformulation of social information-processing mechanisms in children's social adjustment

Psychological Bulletin, 115 (1994), pp. 74-101, 10.1037/0033-2909.115.1.74

View Record in Scopus Google Scholar

Denson et al., 2012

T.F. Denson, C.N. DeWall, E.J. Finkel

Self-control and aggression

Current Directions in Psychological Science, 21 (2012), pp. 20-25, 10.1177/0963721411429451

CrossRef View Record in Scopus Google Scholar

Dietz-Uler et al., 2000

B. Dietz-Uler, E. Harrick, C. End, L. Jacquemotte

Sex differences in sport fan behavior and reasons for being a sport fan

Journal of Sport Behavior, 23 (3) (2000), p. 219

Google Scholar

Donaldson and Ronan, 2006

S.J. Donaldson, K.R. Ronan

The effect of sport participation on young adolescents' well-being

Adolescence, 41 (2006), pp. 370-389

Google Scholar

Eccles and Harold, 1991

J.S. Eccles, R.D. Harold

Gender differences in sport involvement: Applying the eccles' expectancy-value model

Journal of Applied Sport Psychology, 3 (1) (1991), pp. 7-35

CrossRef View Record in Scopus Google Scholar

Eldar et al., 2006

E. Eldar, D. Morris, R. Da Costa, T. Wolf

Are you square? A game for developing self-control and social skills

Strategies, 19 (2006), pp. 17-21

CrossRef View Record in Scopus Google Scholar

Elias et al., 1997

M. Elias, J.E. Zins, R.P. Weissberg, K.S. Frey, M.T. Greenberg, N.M. Haynes, ..., T.P. Shriver

Promoting social and emotional learning: A guide for educators

Association for Supervision and Curriculum Development, Alexandria, VA (1997)

Google Scholar

Ellis et al., 2009

M.L. Ellis, B. Weiss, J.E. Lochman

Executive functions in children: Associations with aggressive behavior and appraisal processing

Journal of Abnormal Child Psychology, 37 (7) (2009), pp. 945-956

CrossRef View Record in Scopus Google Scholar

Else-Quest et al., 2012

N.M. Else-Quest, A. Higgins, C. Allison, L.C. Morton

Gender differences in self-conscious emotional experience: A meta-analysis

Psychological Bulletin, 138 (5) (2012), pp. 947-981, 10.1037/a0027930

CrossRef View Record in Scopus Google Scholar

Findlay and Coplan, 2008

C.F. Findlay, R.J. Coplan

Come out and play: Shyness in childhood and the benefits of organized sports participation

Canadian Journal of Behavioural Science, 40 (2008), pp. 153-161

CrossRef View Record in Scopus Google Scholar

Fleming et al., 2008

C.B. Fleming, R.F. Catalano, J.J. Mazza, E.C. Brown, K.P. Haggerty, T.W. Arachi

After-school activities, misbehavior in school, and delinquency from the end of elementary school through the beginning of high school: A test of social development model hypotheses

Journal of Early Adolescence, 28 (2008), pp. 277-303

CrossRef View Record in Scopus Google Scholar

Folkman, 2008

S. Folkman

The case for positive emotions in the stress process

Anxiety, Stress and Coping, 21 (2008), pp. 3-14

CrossRef View Record in Scopus Google Scholar

Fredrickson, 2009

B. Fredrickson

Positivity: Groundbreaking research reveals how to embrace the hidden strength of positive emotions, overcome negativity, and thrive

Crown, New York (2009)

Google Scholar

Fredrickson et al., 2003

B.L. Fredrickson, M.M. Tugade, C.E. Waugh, G.R. Larkin

What good are positive emotions in crisis? A prospective study of resilience and emotions following the terrorist attacks on the United States on September 11th, 2001

Journal of Personality and Social Psychology, 84 (2003), pp. 365-376

View Record in Scopus Google Scholar

Gavriel-Fried et al., 2015

B. Gavriel-Fried, T. Ronen, Q. Agbaria, H. Orkibi, L. Hamama

Multiple facets of self-control in Arab adolescents: Parallel pathways to greater happiness and less physical aggression

Youth & Society (2015), 10.1177/0044118x15606157

Google Scholar

Gaya et al., 2011

A.R. Gaya, P. Silva, C. Martins, A. Gaya, J.C. Ribeiro, J. Mota

Association of leisure time physical activity and sports competition activities with high blood pressure levels: Study carried out in a sample of Portuguese children and adolescents

Child: Care, Health and Development, 37 (2011), pp. 329-334

CrossRef View Record in Scopus Google Scholar

Gilbert, 2005

D. Gilbert

Stumbling on happiness

Vintage Books, New York (2005)

Google Scholar

Hamama and Ronen-Shenhav, 2012

L. Hamama, A. Ronen-Shenhav

Self-control, social support, and aggression among adolescents in divorced and two-parent families

Children and Youth Services Review, 34 (2012), pp. 1042-1049

Article Download PDF View Record in Scopus Google Scholar

Hamama et al., 2013

L. Hamama, T. Ronen, K. Shachar, M. Rosenbaum

Links between stress, positive and negative affect, and life satisfaction among teachers in special education schools

Journal of Happiness Studies, 14 (2013), pp. 731-751, 10.1007/s10902-012-9352-4

CrossRef View Record in Scopus Google Scholar

Hartup, 2005

W.W. Hartup

The development of aggression: Where do we stand?

R.E. Tremblay, W.W. Hatup, J. Archer (Eds.), Developmental origins of aggression, Guilford Press, New York (2005), pp. 3-22

Google Scholar

Hoffman et al., 2011

J.S. Hoffman, L.M. Knox, R. Cohen

Beyond suppression: Global perspectives on youth violence

Praeger, Santa Barbara, CA (2011)

Google Scholar

Hyde et al., 2011

A.L. Hyde, D.E. Conroy, A.L. Pincus, N. Ram

Unpacking the feel-good effect of free-time physical activity: Between- and within-person associations with pleasant–activated feeling states

Journal of Sport & Exercise Psychology, 33 (2011), pp. 884-902

CrossRef View Record in Scopus Google Scholar

Jimerson and Furlong, 2006

S.R. Jimerson, M.J. Furlong (Eds.), Handbook of school violence and school safety: From research to practice, Lawrence Erlbaum, Mahwah, NJ (2006)

Google Scholar

Kazdin, 2000

A.E. Kazdin

Psychotherapy for children and adolescents: Directions for research and practice

Oxford University Press, New York (2000)

Google Scholar

Kazdin, 2009

A.E. Kazdin

Understanding how and why psychotherapy leads to change

Psychotherapy Research, 19 (4–5) (2009), pp. 418-428, 10.1080/10503300802448899

CrossRef View Record in Scopus Google Scholar

Kazdin and Weisz, 2010

A.E. Kazdin, J.R. Weisz (Eds.), Evidence-based psychotherapies for children and adolescents (2nd ed.), Guilford, New York (2010)

Google Scholar

Lakes and Hoyt, 2004

K.D. Lakes, W.T. Hoyt

Promoting self-regulation through school-based martial arts training

Journal of Applied Developmental Psychology, 25 (3) (2004), pp. 283-302

Article Download PDF View Record in Scopus Google Scholar

Lansford et al., 2012

J.E. Lansford, A.T. Skinner, E. Sorbring, L.D. Giunta, K. Deater-Deckard, K.A. Dodge, et al., L. Chang

Boys' and girls' relational and physical aggression in nine countries

Aggressive Behavior, 38 (4) (2012), pp. 298-308, 10.1002/ab.21433

CrossRef View Record in Scopus Google Scholar

Laurent et al., 1999

J. Laurent, S.J. Catanzaro, T.E. Joiner Jr., K.D. Rudolph, K.I. Potter, S. Lambert, ..., T. Gathright

A measure of positive and negative affect for children: Scale development and preliminary validation

Psychological Assessment, 11 (3) (1999), pp. 326-338, 10.1037/1040-3590.11.3.326

View Record in Scopus Google Scholar

Lufi and Parish-Plass, 2011

D. Lufi, J. Parish-Plass

Sport-based group therapy program for boys with ADHD or with other behavioral disorders

Child and Family Behavior Therapy, 33 (2011), pp. 217-230

CrossRef View Record in Scopus Google Scholar

Magyar-Moe, 2009

J.L. Magyar-Moe

Therapist's guide to positive psychological interventions

Elsevier Academic Press, New York (2009)

Google Scholar

Mahoney et al., 2005

J.L. Mahoney, R.W. Larson, J.S. Eccles, H. Lord

Organized activities as developmental contexts for children and adolescents

J.L. Mahoney, R.W. Larson, J.S. Eccles (Eds.), Organized activities as contexts of development: Extracurricular activities, after-school, and community programs, Lawrence Erlbaum, Mahwah, NJ (2005), pp. 3-22

View Record in Scopus Google Scholar

Maris, 1998

E. Maris

Covariance adjustment versus gain scores-revisited

Psychological Methods, 3 (1998), pp. 309-327

View Record in Scopus Google Scholar

Marsh and Kleitman, 2003

H.W. Marsh, S. Kleitman

School athletic participation: Mostly gain with little pain

Journal of Sport & Exercise Psychology, 25 (2003), pp. 205-228

CrossRef View Record in Scopus Google Scholar

Milaniak and Widom, 2015

I. Milaniak, C.S. Widom

Does child abuse and neglect increase risk for perpetration of violence inside and outside the home?

Psychology of Violence, 5 (3) (2015), pp. 246-255, 10.1037/a0037956

CrossRef View Record in Scopus Google Scholar

Miller et al., 2005

K.E. Miller, M.J. Melnick, G.M. Barnes, M.P. Farrell, D. Sabo

Untangling the links among athletic involvement, gender, race, and adolescent academic outcomes

Sociology of Sport Journal, 22 (2005), pp. 178-193

CrossRef View Record in Scopus Google Scholar

Myers and Diener, 1995

D.G. Myers, E. Diener

Who is happy?

Psychological Science, 6 (1995), pp. 10-17

CrossRef Google Scholar

Nolen-Hoeksema, 2012

S. Nolen-Hoeksema

Emotion regulation and psychopathology: The role of gender

Annual Review of Clinical Psychology, 8 (1) (2012), pp. 161-187, 10.1146/annurev-clinpsy-032511-143109

CrossRef View Record in Scopus Google Scholar

Olweus, 2004

D. Olweus

The Olweus bullying prevention program: Design and implementation issues and a new initiative in Norway

P.K. Smith, D. Peter, K. Rigby (Eds.), Bullying in schools: How successful can interventions be?, Cambridge University Press, Cambridge, UK (2004), pp. 13-16

View Record in Scopus Google Scholar

Orkibi and Ronen, 2015

H. Orkibi, T. Ronen

High self-control protects the link between social support and positivity ratio for Israeli students exposed to contextual risk

Journal of School Psychology, 53 (4) (2015), pp. 283-293, 10.1016/j.jsp.2015.06.001

Article Download PDF View Record in Scopus Google Scholar

Orkibi et al., 2014

H. Orkibi, T. Ronen, N. Assoulin

The subjective well-being of Israeli adolescents attending specialized school classes

Journal of Educational Psychology, 106 (2) (2014), pp. 515-526, 10.1037/a0035428

CrossRef View Record in Scopus Google Scholar

Preacher and Hayes, 2004

K. Preacher, A. Hayes

SPSS and SAS procedures for estimating indirect effects in simple mediation models

Behavior Research Methods, Instruments, & Computers, 36 (4) (2004), pp. 717-731, 10.3758/bf03206553

View Record in Scopus Google Scholar

Reynes and Lorant, 2004

E. Reynes, J. Lorant

Competitive martial arts and aggressiveness: A 2-yr. longitudinal study among young boys

Perceptual and Motor Skills, 98 (1) (2004), pp. 103-115

View Record in Scopus Google Scholar

Rhea and Lantz, 2004

D.J. Rhea, C.D. Lantz

Violent, delinquent, and aggressive behaviors of rural high school athletes and non-athletes

The Physical Educator, 61 (4) (2004), pp. 170-174

Google Scholar

Robinson et al., 1980

E.A. Robinson, S.M. Eyberg, A.W. Ross

The standardization of an inventory of child conduct problem behaviors

Journal of Clinical Child Psychology, 9 (1980), pp. 22-29

CrossRef Google Scholar

Rolider and Van Houten, 1995

A. Rolider, R. Van Houten

The interpersonal treatment model: Teaching appropriate social inhibitions through the development of personal stimulus control by the systematic introduction of antecedent stimuli

R. Van Houten, S. Axelrod (Eds.), Effective behavioral treatment: Issues and implementation, Plenum Press, New York (1995), pp. 127-168

Google Scholar

Ronen, 1997

T. Ronen

Cognitive development therapy with children

John Wiley, Chichester (1997)

Google Scholar

Ronen, 2004

T. Ronen

Imparting self-control skills to decrease aggressive behavior in a 12-year-old boy: A case study

Journal of Social Work, 4 (2004), pp. 269-288

CrossRef View Record in Scopus Google Scholar

Ronen and Rosenbaum, 2010

T. Ronen, M. Rosenbaum

Developing learned resourcefulness in adolescents to help them reduce their aggressive behavior: Preliminary findings

Research on Social Work Practice, 20 (4) (2010), pp. 410-426, 10.1177/1049731509331875

CrossRef View Record in Scopus Google Scholar

Ronen and Seeman, 2007

T. Ronen, A. Seeman

Subjective well-being of adolescents in boarding schools under threat of war

Journal of Traumatic Stress, 20 (2007), pp. 1053-1062

CrossRef View Record in Scopus Google Scholar

Ronen et al., 2013

T. Ronen, I. Abuelaish, M. Rosenbaum, Q. Agbaria, L. Hamama

Predictors of aggression among Palestinians in Israel and Gaza: Happiness, need to belong, and self- control

Children and Youth Services Review, 35 (1) (2013), pp. 47-55

Article Download PDF View Record in Scopus Google Scholar

Ronen et al., 2007

T. Ronen, G. Rahav, A. Moldawsky

Aggressive behavior among Israeli elementary school students and associated emotional/behavioral problems and self-control

School Psychology Quarterly, 22 (3) (2007), pp. 407-431

CrossRef Google Scholar

Rosenbaum, 1980

M. Rosenbaum

A schedule for assessing self-control behavior: Preliminary finding

Behavior Therapy, 11 (1980), pp. 109-121

Article Download PDF View Record in Scopus Google Scholar

Rosenbaum, 1998

M. Rosenbaum

Learned resourcefulness, stress, and self-regulation

S. Fisher, J. Reason (Eds.), Handbook of life stress, cognition and health, Wiley, Chichester, UK (1998), pp. 483-496

Google Scholar

Rosenbaum, 2000

M. Rosenbaum

The self-regulation of experience: Openness and construction

P. Dewe, T. Cox, M. Leiter (Eds.), Coping, health, and organizations, Taylor & Francis, London (2000), pp. 51-67

Google Scholar

Rosenbaum and Ronen, 1991

M. Rosenbaum, T. Ronen

Self-control and education for self-control: Development of rating scale

Paper presented at the 25th anniversary meeting of the Association for Advancement of behavior therapy, York, New (1991)

Google Scholar

Rosenbaum and Ronen, 2013

M. Rosenbaum, T. Ronen

Emotional well-being and self-control skills of children and adolescents: The Israeli perspective

C.L.M. Keyes (Ed.), Mental well-being, Springer Netherlands, New York (2013), pp. 209-229

CrossRef View Record in Scopus Google Scholar

Schreiber et al., 2006

J.B. Schreiber, A. Nora, F.K. Stage, E.A. Barlow, J. King

Reporting structural equation modeling and confirmatory factor analysis results: A review

Journal of Educational Research, 99 (6) (2006), pp. 323-338

CrossRef View Record in Scopus Google Scholar

Schwalbe, 2009

C.S. Schwalbe

Reducing risk for violence and aggression in youth

T. Andrade (Ed.), Handbook of violence risk assessment and treatment: New approaches for mental health professionals, Springer, New York (2009), pp. 425-449

View Record in Scopus Google Scholar

Simpkins et al., 2008

S.D. Simpkins, J.S. Eccles, J.N. Becnel

The mediational role of adolescents' friends in relations between activity breadth and adjustment

Developmental Psychology, 44 (2008), pp. 1081-1094

CrossRef View Record in Scopus Google Scholar

Sport, 2015

Sport

In Merriam-Webster's online dictionary

Retrieved from

http://www.m-w.com/dictionary/sport (2015)

Google Scholar

Thirer and Wright, 2010

J. Thirer, S.D. Wright

Research note sport and social status for adolescent males and females

Sociology of Sport Journal, 2 (2) (2010), pp. 164-171

Google Scholar

Tremblay and Nagin, 2005

R.E. Tremblay, D.S. Nagin

The developmental origins of physical aggression in humans

R.E. Tremblay, W.W. Hartup, J. Archer (Eds.), Developmental origins of aggression, Guilford Press, New York (2005), pp. 83-106

View Record in Scopus Google Scholar

Treves-Pelleg, 2003

Treves-Pelleg, M. (2003). Emotion and cognition among children with conduct problems: the relationship between emotional and cognitive abilities and conduct problems among children (Unpublished dissertation). Tel-Aviv University, Israel.

Google Scholar

Vilhjalmsson and Kristjansdottir, 2003

R. Vilhjalmsson, G. Kristjansdottir

Gender differences in physical activity in older children and adolescents: The central role of organized sport

Social Science & Medicine, 56 (2) (2003), pp. 363-374

Article Download PDF View Record in Scopus Google Scholar

Watson et al., 1988

D. Watson, L.A. Clark, A. Tellegen

Development and validation of brief measures of positive and negative affect: The PANAS scales

Journal of Personality and Social Psychology, 54 (1988), pp. 1063-1070

CrossRef View Record in Scopus Google Scholar

Weisbrod et al., 2007

N. Weisbrod, M. Rosenbaum, T. Ronen

Rejection sensitivity as an aggression enhancing cognitive–affective disposition

Schwartz research report, William S. Schwartz Laboratory for Health Behavior Research, Tel-Aviv University, Tel-Aviv (2007), pp. 16-17

View Record in Scopus Google Scholar

Wilson et al., 2010

D.M. Wilson, D.C. Gottfredson, A.B. Cross, M. Rorie, N. Connell

Youth development in after-school leisure activities

Journal of Early Adolescence, 30 (5) (2010), pp. 668-690, 10.1177/0272431609341048

CrossRef View Record in Scopus Google Scholar

Wilson et al., 2001

D.B. Wilson, D.C. Gottfredson, S.S. Najaka

School-based prevention of problem behaviors: A meta-analysis

Journal of Quantitative Criminology, 17 (2001), pp. 247-272

View Record in Scopus Google Scholar

World Health Organization, 2015

World Health Organization

Adolescent mental health

Retrieved from

http://www.who.int/topics/adolescent_health/en/ (2015)

Google Scholar

Zauszniewski et al., 2010

J.A. Zauszniewski, A.K. Bekhet, E. Bonham

Psychometric testing of the children's resourcefulness scale

Journal of Child and Adolescent Psychiatric Nursing, 23 (3) (2010), pp. 181-188, 10.1111/j.1744-6171.2010.00239.x

AB #4

Research Quarterly for Exercise and Sport

ISSN: 0270-1367 (Print) 2168-3824 (Online) Journal homepage: https://www.tandfonline.com/loi/urqe20

Associations Between Self-Control, Practice, and Skill Level in Sport Expertise Development

Rafael A. B. Tedesqui & Bradley W. Young

To cite this article: Rafael A. B. Tedesqui & Bradley W. Young (2017) Associations Between Self- Control, Practice, and Skill Level in Sport Expertise Development, Research Quarterly for Exercise and Sport, 88:1, 108-113, DOI: 10.1080/02701367.2016.1267836

To link to this article: https://doi.org/10.1080/02701367.2016.1267836

Published online: 27 Jan 2017.

Submit your article to this journal Article views: 800

View related articles View Crossmark data Citing articles: 10 View citing articles

Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=urqe20

RESEARCH QUARTERLY FOR EXERCISE AND SPORT 2017, VOL. 88, NO. 1, 108–113 http://dx.doi.org/10.1080/02701367.2016.1267836

Associations Between Self-Control, Practice, and Skill Level in Sport Expertise Development

Rafael A. B. Tedesqui and Bradley W. Young

University of Ottawa

ABSTRACT

Purpose: The purpose of this study was to test the association between self-control (SC) variables and (a) sport-specific practice amounts, (b) engagement in various practice contexts, (c) threats to commitment to one’s sport, and (d) skill development using the Brief Self-Control Scale (BSCS) in a diverse sport sample. Method: Two hundred forty-four athletes (47% female; Mage 1⁄4 21.96 years, SD 1⁄4 6.98 years; 68.8% individual sports and 31.2% team sports; 13.77 [SD 1⁄4 8.12] hr/week of sport- specific practice) completed a survey composed of the BSCS and practice-related measures. Three skill groups (basic/intermediate, advanced, expert) were informed by athletes’ self-reported highest level of competition. Separate analyses were conducted for juniors (aged 12–17 years) and seniors (aged 18 – 43 years). Results: A 2-factor model (self-discipline and impulse control) fit the BSCS data. Fewer thoughts of quitting from one’s sport were associated with higher self-discipline in juniors and seniors and were also related to higher impulse control in seniors. Greater practice amounts were associated with higher self-discipline; however, only seniors showed such associations in voluntary practice contexts. For juniors and seniors, impulse control was associated with more voluntary practicing. There were, however, no skill-group differences for levels of self-discipline or impulse control. Conclusions: Self-discipline and impulse control may be dispositional characteristics associated with how athletes engage in practice and avert conditions that threaten their sport commitment. SC dispositions may relate to practice amounts differently in juniors and seniors, depending on the requirements for self-regulation in a practice context.

Aspiring experts in any domain have to show a prolonged engagement in high amounts of deliberate practice (DP; Ericsson, Krampe, & Tesch-Römer, 1993), comprising highly effortful activities specifically designed to improve performance. Expert athletes accrue more DP than less- expert athletes over successive points in career develop- ment (Baker & Young, 2014), thus requiring great discipline and commitment to effortful practice. As athletes go from early to late adolescence, they increase commitment to DP in a single sport and decrease engagement in other sport-related activities played for fun (Côté, Baker, & Abernethy, 2003). By the midteen years, committing to high amounts of practice in one sport entails the ability and discipline to resist attractive alternative activities (Young & Medic, 2008). To keep engaged in escalating amounts of DP over time, athletes must constantly overcome effort and motiva- tional constraints, give up momentary pleasures daily (Tedesqui & Young, 2015), and delay the need for gratification (Côté et al., 2003). For example, athletes may

have to resist the temptation to accept social invitations to attend an exhausting, less attractive, and less immediately rewarding practice session. This is under- scored by the fact that professional soccer players invest far less time in social encounters (Toering & Jordet, 2015), suggesting that high-level athletes steer themselves away from gratifying contexts that compete for valuable practice time.

The premise of this study is that developing expertise through challenging sport-specific practice requires a great deal of self-control (SC). SC is required during arduous training conditions or when faced with tempting decisions that might threaten one’s commitment to practice in their sport (Tedesqui & Young, 2015). SC is the ability to control thoughts, regulate emotions, and resist temptations (Tangney, Baumeister, & Boone, 2004). In keeping with conceptualizations of SC as both a state and a trait (Tangney et al., 2004), this study concerned the latter aspect—that is, the dispositional tendency to control one’s impulses and remain

ARTICLE HISTORY

Received 21 June 2016 Accepted 23 November 2016

KEYWORDS

Deliberate practice; impulse control; personality; self-discipline

CONTACT Rafael A. B. Tedesqui rafael.tedesqui@uottawa.ca School of Human Kinetics, University of Ottawa, 125 University Street, Montpetit Hall, MNT 416B, Ottawa, ON K1N 6N5, Canada.

q 2017 SHAPE America

SELF-CONTROL AND EXPERTISE 109

disciplined while pursuing valued long-term goals (Duckworth & Steinberg, 2015). The notion of SC fits within the broader research on self-regulated learning and its value in expertise development (e.g., Zimmerman, 1989). Baker and Young (2014) contended that individual difference variables pertaining to how an athlete self-regulates may predispose them toward DP. Moreover, individuals with higher levels of SC tend to follow their workout routines and more capably translate personal intentions into exercise participation (Allom, Panetta, Mullan, & Hagger, 2016). Outside sport, high trait SC has been linked to higher achievement (e.g., Tangney et al., 2004), including better school and work performance, likely because SC fosters regular practice and good working habits (De Ridder, Lensvelt-Mulders, Finkenauer, Stok, & Baumeister, 2012). Altogether, these works suggest that individual differences in SC may have a bearing on sport practice and commitment; however, no studies have specifically addressed these associations.

Only Toering and Jordet (2015) have examined SC within the realm of sport expertise. They found that one aspect of SC (i.e., impulse control) significantly explained whether male Norwegian professional soccer players had been chosen for the national team. Although there were no mean group differences between more- and less- skilled players for impulse control, more-skilled players did score higher for a second aspect of SC (i.e., levels of restraint). Amounts of soccer practice were positively associated only with levels of restraint. Although the investigators used one measure that assessed average practice hours per day, they foremost focused on the associations that SC had with 11 lifestyle activities (e.g., sleeping, watching TV, gaming) and not on its associations with practice in various contexts. This is a limitation as not all practice contexts are equal; for example, unscheduled practice may require athletes to recruit more self-regulated effort, or more discipline may be required to initiate optional practice compared with when it is prescribed by a coach. Altogether, the literature suggests that trait SC may be key in the initiation and sustenance of instrumental practice habits. However, there are no studies on dispositional SC as it relates to conditions of practice and various training contexts (Englert, 2016), nor are there any studies that have examined SC according to an expert performance approach.

The expert performance approach is predicated on verification of key assumptions. First, researchers must isolate tasks or variables (e.g., SC) for which more-expert individuals perform consistently better compared with less-skilled individuals (Ericsson & Smith, 1991). Secondly, to infer expert development, such variables (e.g., SC) should also be positively associated with

amounts of intensive sport-specific training (Ericsson et al., 1993). Accordingly, we tested four hypotheses: (a) higher SC is related to greater amounts of sport- specific practice (H1); (b) higher SC is associated with more instrumental patterns of engagement in mandatory and voluntary practice contexts (H2); (c) higher SC is associated with a reduced vulnerability to threats to commitment (H3); and (d) more-skilled athletes score higher than less-skilled athletes on dispositional SC (H4). We also examined associations between SC and engagement in playful activities, but no specific hypotheses were posited due to the exploratory nature of these analyses.

Method

Participants and procedure

Participants were 244 athletes (47% female; Mage 1⁄4 21.96 years, range 1⁄4 12–43 years) mainly from Canada (85.9%) and the United States (5%) who were involved in individual (68.8%; e.g., swimming) and team sports (31.2%; e.g., soccer). On average, they reported 13.77 (SD 1⁄4 8.12) weekly hr of sport-specific training and had 8.1 (SD 1⁄4 5.53) years of involvement in their main sport. Informed/parental consent was obtained for all partici- pants. All procedures received institutional ethics approval.

Instruments

All athletes completed an online survey that included demographic questions, questions about athletes’ skill level and practice-related measures, and a scale assessing dispositional SC.

Skill level

Athletes reported their highest level of competition ever achieved at junior (i.e., younger than 18 years) and senior (i.e., aged 18 years and older) age groups, by indicating an appropriate category ranging from (a) local, (b) city, (c) regional, (d) provincial, (e) national, to (f) international level.

Practice-related measures

To measure quantity of weekly sport-specific practice, we referred participants to a typical midseason week and asked them to report how many hours per week they spent on individual or team sport-specific activities deliberately designed to improve performance in their main sport such as technical and tactical training (Hopwood, 2013). This estimate of weekly practice was considered a proxy measure for DP (Hopwood, 2013).

110 R. A. B. TEDESQUI AND B. W. YOUNG

We also asked participants about their weekly amount of play activities in their main sport so that sport-specific practice could be delimited from other sport-related activities.

Athletes also responded to five questions on a Likert scale ranging from 1 (never) to 7 (always). Three questions asked how often they: (a) attended mandatory practice sessions, (b) attended optional practice sessions, and (c) practiced outside scheduled training hours. Because both (b) and (c) share similar assumptions on the need for self-direction and the recruitment of personal resources to engage in these forms of practice, we collapsed them to derive a score for voluntary practice. These questions were considered measures of engagement in practice contexts. The last two questions asked how often athletes considered either playing a different sport or quitting their main sport. Higher responses for these final items might indicate athletes’ likelihood of giving up an expertise pursuit; thus, we refer to these measures as threats to commitment.

Brief Self-Control Scale

The Brief Self-Control Scale (BSCS; Tangney et al., 2004) uses 13 items (e.g., “I am good at resisting temptation”; “I wish I had more self-discipline”) to assess dispositional SC. Athletes rated the degree to which items reflected their typical behavior on a Likert scale ranging from 1 (not at all) to 5 (very much). Tangney et al. (2004) found the BSCS had good internal reliability (a 1⁄4 .83 2 .85), good test–retest reliability (r 1⁄4 .87), and concurrent criterion validity (e.g., high correlation with academic grade point average, r 1⁄4 .39). The BSCS is widely used in domains outside sport (e.g., education and work; Duckworth & Steinberg, 2015; Maloney, Grawitch, & Barber, 2012). Although the BSCS has been treated as a single SC construct, there has been support for two SC factors (e.g., Maloney et al., 2012), including the lone study to inspect the BSCS in sport (i.e., Toering & Jordet, 2015).

Data analysis

Through confirmatory factor analyses using AMOS, we tested both a one-factor and two-factor structure of the BSCS. Fit indices were far better for the two-factor (Comparative Fit Index [CFI] 1⁄4 .92, standardized root mean square residual [SRMR] 1⁄4 .05, root mean square error of approximation [RMSEA] 1⁄4 .06, 90% CI [.04, .08]) compared with the one-factor model (CFI 1⁄4 .85, SRMR 1⁄4 .07, RMSEA 1⁄4 .09, 90% CI [.07, .10]). Thus, in our main analyses, we used two BSCS subscales (correlated at r1⁄4.71): (a) Self-Discipline (SD; seven items; a 1⁄4 .80; factor loadings 1⁄4 .53 2 .69), one’s ability to be self-disciplined and work toward goals; and (b) Impulse Control (IC; six items; a 1⁄4 .71; factor loadings 1⁄4 .41 2 .68), an individual’s ability to control impulses and resist temptations.

Primary analyses

Partial correlations (controlling for age) assessed the associations that SD and IC had with each of the practice- related measures. To test group differences, we collapsed Competitive Levels 1 (local) to 4 (provincial) into a basic/ intermediate group (B/I; n 1⁄4 40 junior athletes, n 1⁄4 33 senior athletes); Level 5 (national) formed the advanced group (Adv; n 1⁄4 15 juniors, n 1⁄4 31 seniors), and Level 6 (international) formed the expert group (Exp; n 1⁄4 17 juniors, n 1⁄4 80 seniors). We tested whether these groups differed in SC through two separate one-way analyses of covariance, while controlling for age: one for SD and another for IC. Data were analyzed separately for junior

Table 1. Means, standard deviations, and partial correlations between self-discipline (SD), impulse control (IC), and practice-related variables, while controlling for age.

(ages 12 – 17 athletes.

Results

years) and

senior

(ages 18 – 43

years)

Descriptive statistics are presented in Table 1 along with associations of SD and IC with sport-specific practice, engagement in practice contexts, and threats to commitment. In terms of group differences among

Variable Sport-specific practicea

Play activitiesa Mandatory practiceb Voluntary practiceb,c Think switchingb Think quittingb

M (SD) 14.38 (8.80)

2.61 (4.34) 6.57 (0.72) 5.08 (1.34) 2.82 (1.55) 2.11 (1.20)

Seniors SD

.09 2 .01

.27*

.26* 2 .24* 2 .27*

IC 2 .12

2 .25* .10

.17* 2 .15

2 .32*

M (SD) 13.17 (6.78)

3.76 (5.21) 6.79 (0.41) 5.30 (1.36) 2.53 (1.42) 1.68 (1.04)

Juniors SD

.27* 2 .12 .06 .05

2 .26* 2 .26*

IC 2 .06

2 .11 2 .04

.31* 2 .03 2 .11

a Hours per week. b Measured on a 7-point Likert scale ranging from 1 (never) to 7 c Composite score of optional practice and unscheduled practice measures. *p , .05. Small, medium, and large effect sizes are respectively r 1⁄4 .10, r 1⁄4 .30, r 1⁄4 .50 (Cohen, 1992).

(always).

SELF-CONTROL AND EXPERTISE 111

seniors, skill groups did not significantly differ in levels

junior sport appears to involve more socially regulated than self-regulated contexts (Bandura, 1986). For juniors, commitment may depend more heavily on external regulation (e.g., parents taking athletes to practice or coaches deciding when, what, and how much athletes should practice).

Considering that practice attendance and activity among juniors may be mostly socially prescribed, this may also explain why we found no associations between their SD and measures of frequent engagement in the practice contexts. As athletes grow older, as performance improves, and as they are afforded more autonomy in their decision making, there may be a change in the agency for learning with athletes gradually internalizing self- regulation processes that allow them to increasingly control the learning situation and skill refinement (Bandura, 1986). As a result, SD may be more instrumental to seniors who no longer depend on the presence of external agents (e.g., coach or family) to regulate their attendance at practice or activity during practice.

Another nuanced finding was the positive association between IC and engagement in voluntary practice among both juniors and seniors, whereas no association was found between IC and mandatory practice in either cohort. Compared with mandatory practice, voluntary practice presumably requires more self-regulation; it is less socially prescribed, with less strict mandates/super- vision from a coach. When there are little or no social expectations for athletes to practice (i.e., during voluntary practice), athletes’ efforts to engage in this type of training will rely more on self-regulation, and our results suggest that junior and senior athletes’ ability to resist temptations to engage in attractive alternative activities may be valuable in this regard. Although SD remains valuable in voluntary contexts among seniors, the results suggest a possible complementary role for IC in these self-regulated contexts, thereby corroborating the notion that inhibitive self-regulation may help athletes complete large amounts of personally selected training (Tedesqui & Young, 2015).

Associations with threats to commitment

Overall, our results suggest that we can accept H3. SD showed an inverse association with threats to commit- ment for both juniors and seniors of at least a small effect size (Zhu, 2012) approaching a medium effect size (Cohen, 1992). Athletes scoring higher on SD tended to think less about quitting or switching out of their main sport, suggesting that SD may help athletes keep themselves on their developmental path in one sport. SD may aid athletes’ commitment to their chosen sport by helping them avoid instead of resist temptations related to switching or quitting their main sport.

(M 1⁄4 3.23, SD 1⁄4 0.70), Adv (M 1⁄4 3.38, Exp (M 1⁄4 3.64, SD 1⁄4 0.69), F(2, p . .05, partial h 2 1⁄4 .02. Senior skill

of SD: B/I SD 1⁄4 0.64), 140) 1⁄4 1.65, groups did not differ in IC levels: B/I (M 1⁄4 3.38, SD 1⁄4 0.61), Adv (M 1⁄4 3.42, SD 1⁄4 0.65), Exp (M 1⁄4 3.59, SD 1⁄4 0.75), F(2, 140) 1⁄4 0.11, p . .05, partial h 2 1⁄4 .00. Similar results were obtained among juniors. Skill groups did not differ in SD levels: B/I (M 1⁄4 3.43, SD 1⁄4 0.75), Adv (M 1⁄4 3.45, SD 1⁄4 0.89), Exp (M 1⁄4 3.70, SD 1⁄4 0.75), F(2, 68) 1⁄4 0.72, p . .05, partial h 2 1⁄4 .02; nor did they differ in IC: B/I (M 1⁄4 3.78, SD 1⁄4 0.64), Adv (M 1⁄4 3.62, SD 1⁄4 0.65), Exp (M 1⁄4 3.48, SD 1⁄4 0.71), F(2, 68) 1⁄4 0.52, p . .05, partial h 2 1⁄4 .02.

Discussion

Associations with quantity of weekly sport-specific practice and play

We found a positive association between SD and DP, with a small effect size (Zhu, 2012) approaching a medium effect size (Cohen, 1992) for juniors. Thus, we accept H1 for juniors. SD could be valuable for young athletes aiming for the top, especially if it prompts training and considering how younger athletes who become adult experts amass more DP than less-skilled peers from early stages (Baker & Young, 2014). Although H1 was not supported for seniors, IC and weekly play were negatively associated, suggesting an indirect benefit to seniors’ training. Playful activities are not directly aimed at skill improvement and might be considered inefficient investments of time for skill acquisition, especially at later developmental stages (Côté et al., 2003). Aspiring elite athletes transitioning from adolescence to adulthood should gradually reduce their engagement in play (Côté et al., 2003). If higher IC levels help senior athletes resist temptations to play, then IC may help seniors steer themselves toward more structured DP activities.

Associations with measures of engagement in practice contexts

Nuanced associations in the expected directions between SC variables and attendance at mandatory and voluntary practice suggest that we can accept H2, especially among seniors. While SD seems relevant for seniors’ engagement in both mandatory and voluntary practice, it does not seem important in either context for juniors. Differences in self-regulatory requirements imposed on juniors and seniors may help explain why. Consistent with the changing nature of social influences as athletes develop through stages of sport involvement (Côté et al., 2003),

112 R. A. B. TEDESQUI AND B. W. YOUNG

If athletes are able to avoid temptations, they might not need to actually resist them. De Ridder et al. (2012) found that individuals high in trait SC tended to form habits that prevented them from needing to resist temptations. They suggested that the application of SC may be most effective by “establishing and maintaining stable patterns of behavior rather than by performing single acts of self- denial” (De Ridder et al., 2012, p. 91). Athletes high in trait SD may be better able to control their routines and their daily activities such that they circumvent situations that tempt them to think about quitting or withdrawing from their main sport. This situation-management ability may be what helps athletes, seniors and juniors alike, to remain committed to their sport.

Seniors high in IC thought less about quitting their main sport, suggesting that the ability to resist temptations may help seniors keep themselves on their developmental path. The lack of a similar association among juniors may be explained by differences in the context of junior and senior sport. Threats to commit- ment may be more salient as competitive athletes age— they reduce involvement in playful activities (Côté et al., 2003), training demands increase, and they meet greater motivational constraints as the rate of skill acquisition declines and learning plateaus are met (Ericsson et al., 1993). Thus, as athletes transition out of adolescence, the ability to resist the temptation to pursue attractive nonsport activities over one’s sport practice may assume a more pivotal role in helping seniors remain committed to their development (Young & Medic, 2008).

Between-group differences in levels of self-control

We found no skill-group differences in SC levels among juniors or seniors; thus, we cannot accept H4. To fully substantiate SC within a sport expertise framework (Ericsson & Smith, 1991), it would have been important to show mean level differences in SC between multiple skill groups. The current findings do not satisfy this criterion. Considering the significant associations between SC and many of the practice-related measures, however, the lack of skill-group differences may indicate that SC variables do not necessarily distinguish skilled from less-skilled athletes, but instead may help all athletes alike develop skill by allowing them to engage in multiple practice contexts and to successfully overcome threats to their commitment.

Limitations and future research

Although we used Hopwood’s (2013) protocol for measuring sport-specific practice, which is arguably an acceptable proxy for DP, we were limited in drawing

conclusions regarding SC and DP. The debate about how to measure DP in sport continues, and more associations may be uncovered with better DP measures. Further, studies might consider whether sport type or regulatory requirements (self vs. social) moderate the link between SC and practice. Perhaps our varied sample (both sexes, several sports) diluted skill-group differences; when Toering and Jordet (2015) applied SC to discriminate performance groups in a large sample from one sport, group differences were found. Finally, future studies might benefit from considering the impact of SC on practice quality in addition to practice quantity and engagement in practice contexts.

Conclusion

SD may help juniors stay committed to their sport and accrue more DP. Among seniors, both SD and IC show positive associations with engagement in voluntary practice and may help them avoid threats to commitment. Given the cross-sectional nature of our data, any comparisons across age groups have to be made cautiously, and we refrain from interpretations that imply directionality between SC and our practice measures. Yet, the observed associations with SC variables warrant further work examining how these traits influence DP and sport achievement.

What does this article add?

First, our results corroborate previous findings outside sport on the influential role of SD on good work/practice habits (De Ridder et al., 2012). Second, they extend Toering and Jordet’s (2015) work on SC and lifestyle activities in soccer to more squarely focus associations between SC and various practice contexts (e.g., DP and contexts requiring degrees of self-regulation) in a mixed sample of athletes (e.g., various genders, ages, sports). Third, this article sensitizes sport expertise researchers to (a) important associations between SC and engagement in practice contexts, as well as threats to commitment to specialized practice, and (b) the necessity for future research to substantiate skill-group differences on these SC variables to more firmly position SC within a sport expertise framework. Finally, this article illustrates how studies might integrate personality research to further understand the role of individual differences on sport expertise development.

Funding

This work was supported by the Social Sciences and Humanities Research Council of Canada through a Joseph-Armand Bombardier Canada Graduate Scholarship (767-2013-2136).

SELF-CONTROL AND EXPERTISE 113

References

Allom, V., Panetta, G., Mullan, B., & Hagger, M. S. (2016). Self- report and behavioural approaches to the measurement of self-control: Are we assessing the same construct? Personality & Individual Differences, 90, 137–142. doi:10.1016/j.paid. 2015.10.051

Baker, J., & Young, B. W. (2014). 20 years later: Deliberate practice and the development of expertise in sport. International Review of Sport & Exercise Psychology, 7, 135 – 157.

Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall. Cohen, J. (1992). A power primer. Psychological Bulletin, 112,

155 – 159. Côté, J., Baker, J., & Abernethy, B. (2003). From play to

practice: A developmental framework for the acquisition of expertise in team sports. In J. L. Starkes & K. A. Ericsson (Eds.), Expert performance in sports: Advances in research on sport expertise (pp. 89–110). Champaign, IL: Human Kinetics.

De Ridder, D. T. D., Lensvelt-Mulders, G., Finkenauer, C., Stok, F. M., & Baumeister, R. F. (2012). Taking stock of self- control: A meta-analysis of how trait self-control relates to a wide range of behaviors. Personality & Social Psychology Review, 16, 76–99.

Duckworth, A. L., & Steinberg, L. (2015). Unpacking self- control. Child Development Perspectives, 9, 32–37. doi:10. 1111/cdep.12107

Englert, C. (2016). The strength model of self-control in sport and exercise psychology. Frontiers in Psychology, 7:314. doi:10.3389/fpsyg.2016.00314

Ericsson, K. A., Krampe, R. T., & Tesch-Römer, C. (1993). The role of deliberate practice in the acquisition of expert performance. Psychological Review, 100, 363–406.

Ericsson, K. A., & Smith, J. (1991). Toward a general theory of expertise: Prospects and limits. New York, NY: Cambridge University Press.

Hopwood, M. J. (2013). The Developmental History of Athletes Questionnaire: Towards a comprehensive understanding of the development of sport expertise (Doctoral dissertation). Retrieved from http://vuir.vu.edu.au/22353

Maloney, P. W., Grawitch, M. J., & Barber, L. K. (2012). The multi-factor structure of the Brief Self-Control Scale: Discriminant validity of restraint and impulsivity. Journal of Research in Personality, 46, 111–115. doi:10.1016/j.jrp. 2011.10.001

Tangney, J. P., Baumeister, R. F., & Boone, A. L. (2004). High self-control predicts good adjustment, less pathology, better grades, and interpersonal success. Journal of Personality, 72, 271 – 324.

Tedesqui, R. A. B., & Young, B. W. (2015). Perspectives on active and inhibitive self-regulation relating to the deliberate practice activities of sport experts. Talent Development & Excellence, 7, 29–39.

Toering, T., & Jordet, G. (2015). Self-control in professional soccer players. Journal of Applied Sport Psychology, 27, 335–350. doi:10.1080/10413200.2015.1010047

Young, B. W., & Medic, N. (2008). The motivation to become an expert athlete: How coaches can promote long-term commitment. In D. Farrow, J. Baker, & C. MacMahon (Eds.), Developing elite sports performers: Lessons from theory and practice (pp. 43–59). New York, NY: Routledge.

Zhu, W. (2012). Sadly, the earth is still round (p , 0.05). Journal of Sport & Health Science, 1, 9–11. doi:10.1016/j.jshs. 2012.02.002

Zimmerman, B. J. (1989). A social cognitive view of self- regulated academic learning. Journal of Educational Psychol- ogy, 81, 329–339.

AB #5

On the importance of self-control strength for regular physical activity

Author links open overlay panel

EmilyFinnea1ChrisEnglertbc1DarkoJekaucc

Show more

https://doi.org/10.1016/j.psychsport.2019.02.007

Get rights and content

Highlights

•

259 participants in university exercise courses were followed for 13 weeks.

•

Course participation declined significantly over 3 months despite high intentions.

•

Initial and weekly exercise intentions predicted course participation.

•

High Self-control promotes the enactment of short-term exercise intentions.

•

Weekly fluctuations in intentions are more important than absolute intention levels.

Abstract

Objectives

Physical activity intentions do not automatically lead to physical activity behavior, indicating that there are other (psychological) factors involved. In the present study, we tested the assumption that self-control strength is required to bridge the intention-behavior gap in terms of initial and weekly intentions.

Design

A total of N = 259 individuals, who registered for a weekly university sports course, were followed for 13 weeks.

Methods

At baseline, trait self-control strength as well as the intention to take part in the respective sports course on a regular basis were measured. Then, we registered weekly participation and asked the participants after each training session to report their intention to show up at next week’s training session.

Results

Despite very high baseline as well as weekly exercise intentions, the participation rate dropped considerably over time, indicating a large intention-behavior gap. The association of within-person fluctuations in weekly intentions and actual participation was moderated by self-control: Only in individuals with high levels of trait self-control strength stronger intentions were associated with a higher chance of actual re-attendance. Baseline intention was also associated with participation but not moderated by self-control.

Conclusions

These findings indicate that high levels of self-control strength are beneficial in order to translate short-term intentions into actual behavior. Practical implications on how to improve self-control are discussed.

· Previous article in issue

· Next article in issue

Keywords

Intention-behavior gapPhysical activitySelf-controlSelf-regulationSport

1. Introduction

Previous research has repeatedly demonstrated that regular physical activity has beneficial effects on several health-related (e.g., overweight, cardiovascular diseases; e.g., Hamilton, Healy, Dunstan, Zderic, & Owen, 2008) and psychological variables (e.g., reduced levels of stress and anxiety, subjective well-being; e.g., Bhui & Fletcher, 2000; Biddle & Mutrie, 2007; Goodwin, 2003; Ströhle, 2009). Regardless of positive health effects, which are well recognized by many people, Hallal et al. (2012) found out that 31% of adults worldwide are physically inactive, meaning that they do not match the World Health Organization’s (WHO, 2010) recommendation of at least 150 min of moderate physical activity per week.

These findings seem to indicate that there is an intention-behavior gap, as individuals often have physical activity intentions, but fail to translate them into actual exercise behaviors (e.g., Rhodes, Plotnikoff, & Courneya, 2008). A meta-analysis by Rhodes and de Bruijn (2013) showed that only about half (54%) of adults classified as physical activity (PA) ‘intenders’ acted on their intention and reached public health guidelines for moderate-to-vigorous physical activity.

Traditional social-cognitive models of physical activity almost exclusively focus on the importance of intentions and how they are built (e.g., Theory of Planned Behavior, Ajzen, 1991). However, these models have been heavily criticized, as intentions alone cannot explain how and why intentions are translated into physical activity behavior (e.g., Rhodes & de Bruijn, 2013; Rhodes & Yao, 2015; Sniehotta, Presseau, & Araújo-Soares, 2014). Interestingly, as Rhodes and Yao (2015) point out in their review, scientific discussions regarding the importance of post-intentional volitional processes can be traced back to the early and mid-20th century (Ach, 1905; Lewin, 1951). More recently developed models of physical activity do not only focus on the importance of intentions, but also investigate the role of volitional variables for regular physical activity (e.g., Hagger, 2010; Schwarzer & Luszczynska, 2008). One important aspect of volition is the ability to control oneself (e.g., Hagger, Wood, Stiff, & Chatzisarantis, 2009).

Self-control enables human beings to focus on long-term goals, while resisting immediate gratifications along the way (e.g., Baumeister, Heatherton, & Tice, 1994). In general, self-control helps individuals to suppress dominant impulses and urges, to avoid distraction, or to focus on long-term goals (e.g., de Ridder, Lensvelt-Mulders, Finkenauer, Stok, & Baumeister, 2012). One of the most prominent self-control theories is the strength model of self-control (Baumeister, Bratslavsky, Muraven, & Tice, 1998). It is assumed that there is a central energy resource on which all self-control acts are based. However, this metaphorical self-control strength has limited capacity, meaning that self-control cannot be exerted limitlessly. Some individuals’ self-control strength possesses a larger capacity than that of others, meaning that the former ones are better at regulating themselves than the latter ones. In this conceptualization, self-control strength can be viewed as a psychological trait (Tangney, Baumeister, & Boone, 2004), which helps an individual inhibit certain behavioral tendencies in order to achieve a more desirable long-term goal. Higher levels of trait self-control strength have been linked to several positive outcomes, for instance better grades, better social relationships, or a healthier lifestyle (for an overview, see de Ridder et al., 2012; Hagger, Wood, Stiff, & Chatzisarantis, 2010).

The strength model is also suited to explain the intention-behavior gap (e.g., Englert, 2016; Englert & Rummel, 2016; Martin Ginis & Bray, 2010). If individuals are not fully committed to their physical activity program, they may sense external and/or internal hurdles regarding the execution of their intentions. For instance, it might be more gratifying, at any given moment, to relax on the sofa at home, instead of going out for a run in the cold, as originally intended. In these cases, individuals need to resist the immediate temptation of relaxing at home, in order to follow their long-term goal of getting in better shape. Some individuals do not seem to have any problem in controlling themselves, while it seems to be a little more difficult for others (for an overview, see Englert, 2016). Several studies have demonstrated the beneficial effects of high levels of trait self-control strength (de Ridder et al., 2012; Hagger et al., 2010). Bertrams and Englert (2013) asked university students to specify their physical activity intentions for the upcoming weeks. The results revealed that individuals with higher levels of trait self-control strength were more likely to meet their physical activity intentions. In the same vein, a recent study by Stork and colleagues found empirical support for the assumption that trait self-control strength predicts exercise adherence over a 4-week period (Stork, Graham, Bray, & Martin Ginis, 2016). Viewed together, higher levels of trait self-control strength seem to increase the likelihood of translating physical exercise intentions into actual exercise behavior, as it helps an individual to resist immediate gratifications in favor of achieving a long-term goal.

However, the empirical basis for this assumption seems rather weak. In their systematic review on moderating effects on the relationship between intentions and actual physical activity behavior, Rhodes and Dickau (2013) found that intention stability was the most consistent moderator in this context. Moderating effects of self-control strength, however, were not analyzed in the included studies and changes in intentions were mostly considered as changes between two occasions, as in most studies so far (Conroy, Elavsky, Hyde, & Doerksen, 2011). Thus far it has been primarily investigated how self-control moderates the translation of global physical activity intentions, without consideration of within-person variation, into action (e.g., intention to participate in a two-week physical training; Englert & Rummel, 2016). However, as previous research (Conroy et al., 2011, 2013) has shown, there is a large amount of within-person variation in physical activity intentions with only about half of the overall variability in PA intentions between persons. These short-term fluctuations in intentions are related to fluctuations in actual PA behavior (Conroy et al., 2011, 2013). Therefore, in the present study we not only focus on how the translation of global physical activity intentions into action is affected by self-control strength, but furthermore, how self-control strength interplays with weekly fluctuations in physical activity intentions.

The aim of the present study, therefore, was to build on previous findings and to test two hypotheses about how self-control strength moderates intention-behavior relations. We hypothesized that a) higher self-control strength would lead to a higher chance to enact the initial intention to participate in an exercise class regularly, i.e. for persons with higher self-control strength the relation between initial intention and actual participation behavior should be stronger, and b) higher self-control strength would lead to a higher chance to enact weekly intentions, i.e. for persons with higher self-control strength a closer relation between weekly intentions and actual participation following this intentions should be observed.

2. Method

2.1. Participants

Questionnaire data were collected on a weekly basis in 10 sports and gym courses, offered by two German universities during the winter term 2015/16 (6th of October 2015 until 6th of February 2016 for the one university and 13th of October 2015 until 6th of February for the other university). These kind of courses are offered by the departments of collegiate sports to all students and employees of the universities. Courses are offered each term and run from the beginning of the lecture period to the end. To participate, students as well as employees have to register for specific courses in time, as available places are assigned on a ‘first come first serve’ basis. Only courses of medium size (about 15–30 participants) and weekly training durations of 60–90 min were chosen for the current study, in agreement with the head of collegial sports. This approach led to a total sample of 177 female und 82 male (N = 259) university students and employees, who volunteered to participate in the present study, and provided sufficient complete data, which was included in the presented analyses. Course instructors were informed about the study and had to give their consent. The study was carried out in accordance with the Helsinki Declaration of 1975 and was approved by the Data Security Commissioner and the Ethics committee of the respective universities.

The courses encompassed various kinds of sports: martial arts (Kickboxing, Taekwondo, Capoeira), aerobic exercise, including dance steps (Zumba, Bokwa), basketball training and Freeletics (a specific set of endurance and strength exercises). The duration of the courses depended on the length of the semester at the respective university (i.e. 13 weekly sessions at one and 14 to 15 sessions at the other university), for comparability only the first 13 sessions were included in our analyses. There was a Christmas holiday break (20th of December 2015 until 7th of January 2016) at both universities, during which no courses were offered.

2.2. Procedure

Individuals who agreed to participate in the study signed a consent form, when attending the course for the first time, and then filled in a baseline questionnaire. The baseline questionnaire and the other questionnaires will be explained in greater detail below.

Student assistants attended all selected courses on a weekly basis. They documented participation and distributed, at the beginning of the course, a short weekly questionnaire to all attending study participants. This questionnaire was recollected at the end of the respective course session. To be able to match the different questionnaires to each participant, each individual generated an anonymous code composed of letters and numbers from family names, year and place of birth.

2.3. Measures

2.3.1. Baseline measures

2.3.1.1. Self-control

After assessing sociodemographic information (gender, age, student status), we applied the German adaption of the short form of the Self-Control Scale (SCS-K-D; Bertrams & Dickhäuser, 2009), which measures trait self-control strength. The SCS-K-D contains 13 items (e.g., ‘I am good at resisting temptations’) which are answered on 5-point Likert-type scales (1 - not at all to 5 - very much). The internal consistency of the scale was sufficient in the present study (Cronbach’s α = .80). For this scale and all the other measures applied in the current study, we calculated mean scores, with higher scores always indicating higher values in the respective variable.

2.3.1.2. Intention

The intention towards attending the specific university sports course during the winter term was measured via two items (for this procedure, see de Bruijn, Gardner, van Osch, & Sniehotta, 2014). Participants were asked to rate how strongly they agree with the following statements on 7-point bipolar scales (totally disagree – totally agree): “I intend to participate in this course on a regular (weekly) basis”, and, “I am sure that I will participate in this course regularly (weekly)”. This conceptualization of a continuum of intention strength relates to the intensity of the commitment to enact a behavior which can be seen as the result of motivational processes (Rhodes & Rebar, 2017).

2.3.1.3. Past exercise

Participants were asked, if they already were regularly exercising (yes/no) and if so, for how long (months).

2.3.2. Weekly questionnaire

2.3.2.1. Attendance

A research student attended every session and recorded weekly participation of each participant (1 = yes, 0 = no).

2.3.2.2. Weekly intention

At the end of each session, every participant was asked to respond to the intention item: “Do you intend to participate in this course again next week (next time)?” (1 = absolutely not to 10 = at any rate).

2.4. Statistical analyses

Descriptive statistics were calculated as percentages for categorical data and means and standard deviations for continuous variables. For severely skewed variables, the median is additionally reported.

For the attendance description over time, weekly participation rates were calculated as proportion of all study participants who were still under observation attending a specific week. Some individuals entered the study later in the term and, therefore, had fewer opportunities to participate. We started counting the weeks for every individual, beginning with the first attendance (participation week 0) and from then on numbering the opportunities to re-attend an exercise class (starting with participation week 1).

To predict participation in exercise classes within the term, from baseline self-control and intention (time-invariant) as well as weekly intention values (time-varying) and the interaction of self-control strength with intentions, mixed effect Cox proportional hazard models were estimated as a variant of survival analysis.

These models analyze time-to-event-data, more precisely, the probability of an event is deduced from the length of time until it happens and characteristics of covariates are analyzed in their effect on this probability. Survival analysis was chosen because information on covariates which were measured on a weekly basis was available only for weeks when the event re-attendance occurred. Therefore, the instances that were analyzed were not the weekly events but these time-to-event periods.

Unlike in standard survival analysis, the event ‘re-attendance’, we were interested in, was not terminal, but recurrent and, unlike in most application of survival or failure time analysis, the event was positively defined. With every attended class, a new interval started, which would then either end with the desirable event ‘re-attendance’ or with the end of the observation period (i.e. case was censored). Interval length was counted in opportunities to participate and constituted the time to the event that was the analyzed outcome of our models. Since most individuals contributed more than one interval (i.e. attended the class more than once), several intervals were nested within the individuals and therefore not independent. Mixed effect models account for these dependencies by allowing for random effects of each person. Additionally, the specific exercise class an individual was participating in was included as another level of nesting (i.e. random effect).

The Cox proportional hazard model predicts the hazard rate (or chance in this case) to participate in the exercise class again (event), compared to a reference group, considering the time duration until this event occurs (time to event) and starting with the time of the preceding attendance. The resulting hazard ratio (HR) can be interpreted as the ratio of re-attendance rates at any given point in time and therefore approximately expresses the probability to re-attend the exercise class, compared to a reference group. Every person was “at risk” of coming back after each attended (or intermittently cancelled) class again, i.e. a new time-to-event interval started. We used the Andersen-Gill (counting process) approach to define the risk set for each interval (Andersen & Gill, 1982). That means that for a given week, all participants for whom the course was still ongoing were part of the risk set, had they attended the subsequent week or any week before. All models included time aspects (participation week, number of attended classes so far, and history of past exercising) as additional predictors to adjust for changes in event rates over time and for the possibility that past successes in translating an intention into action would enhance the chance to act on this intention again.

We followed a stepwise approach of model fitting, starting with a model adjusting for time-effects and exercise history, testing for potential confounders, then adding baseline intention and self-control, as well as their interaction to examine our first hypothesis, and finally including weekly intention values, as well as the interaction with self-control to examine our second hypothesis. Models were compared by likelihood ratio tests. The R-package ‘coxme’ was used for estimation of the models (Therneau, 2018).

For the weekly measured intention, the variation was decomposed into a between-person and a within-person component. The between component is the mean of the grand-mean centered variable per person (time-invariant), and the within-component is the deviation of the weekly score from this overall mean per person (time-varying) (Bolger & Laurenceau, 2013).

The Cox proportional hazard model assumes that hazard functions for survival curves at different values of a predictor are proportional over time. This assumption was tested by using a standard survival model with an added frailty term to account for non-independence of observations of the same individual.

3. Results

3.1. Descriptive results

259 individuals provided enough information from the baseline questionnaire and allowed for at least one of the weekly questionnaires to be included in the analyses. The sample is described in Table 1, Table 2 in terms of sociodemographic information, past exercising, frequency of attendance, as well as baseline intention and self-control. The sample predominantly consisted of students complemented by a small group of university employees. About two thirds were female, and the mean age was 24.5 years (SD = 5.7). Although a majority reported regular exercise or sport participation in the past and a very high intention to regularly participate in the chosen exercise class over the entire term, mean frequency of participation was only 5.16 times (SD = 3.47), corresponding to a participation in 44% of offered classes. Baseline intention was highly skewed, resulting in a ‘J-shaped’ distribution with 38% declaring the highest possible intention. The variable was therefore transformed as inverse of the reflected value and then scaled to have the same range of 1–7 as the original scale. This resulted in a reduction of skewness from −1.57 to 0.02 although no normal distribution could be achieved. Self-control values approximately followed a normal distribution (Shapiro-Wilk-test: p > 0.20), and the mean self-control capacity was comparable in size to that of the validation sample of German students (Bertrams & Dickhäuser, 2009) (see Table 2).

Table 1. Description of the sample (N = 259; categorial variables).

Measures	N	%
Gender (% female)	177	68.3
Student status (% students)	233	90.0
Type of sport:
Basketball	73	28.2
Martial arts	115	44.4
Dancing/funsport	71	27.4
Regular exercising in the past (yes)	190	73.36

Table 2. Description of the sample (N = 259; continous variables).

	M	SD	range
Age (years)	24.51	5.69	16–57
Regular exercising in the past (in months)	91.85	96.58	0–384
Number of weeks attended	5.16	3.47	1–13
Baseline intention	5.95	1.29	1–7
Baseline self-control capacity	3.06	0.59	1.38–4.77

The development of participation rates over time is depicted in Figure 1. Starting with participation week 0 (i.e. the first week an individual attended the course), participation decreased continually over the term with a participation rate as low as 19% in the last observed week.

· Download : Download high-res image (252KB) Download : Download full-size image

Figure 1. Participation rate over time. Participant weeks start with the first attendance of a participant (week 0). For every week, the participation rate is the proportion of participants attending the class (of all participants who had the opportunity to do so in the given week).

The weekly intention showed slight fluctuations but no clear trend over time with the stated intention to re-attend the following week constantly being very high. Since the weekly intention was also highly skewed (weekly means ranged from 8.84 to 9.24 on a scale from 1 to 10) the same transformation, like for the baseline intention, was applied in the regression models. Over the weeks, between 28% and 58% of individuals, who stated the highest possible intention to re-attend the following week, did not actually enact this intention.

3.2. Survival analysis

In the mixed effects Cox proportional hazard models, 1,243 time-to-event intervals from 259 individuals were analyzed of which 982 intervals resulted in the event re-attendance (see Table 3, Table 4). Preparatory analyses revealed that time was best modelled as week plus a quadratic term of week, while the Christmas holiday break had no additional effect. To arrive at a meaningful baseline model, including potential confounders, in a first step potentially relevant background variables, as well as time aspects (participation week, accumulated number of attended classes so far, and past exercising in months), were entered as predictors of time-to-event data. All time aspects, except for self-reported past exercising, significantly predicted participation in the exercise class. Particularly, the chance to re-attend increased by 19% with each time the class had been attended before. None of the sociodemographic background variables was significantly related to attendance but females exhibited slightly lower attendance rates (HR = 0.923, 95% CI: 0.80–1.06). Gender, however, violated the proportional hazard assumption and was, therefore, used as a stratification variable in our baseline model (model 0, used as a baseline for our hypothesis tests) and subsequent models, allowing for different baseline hazards for males and females. In model 1, stratified by gender and adjusted for time course and past exercise history, baseline self-control and intention values as well as their interaction were entered as predictors to test our first hypothesis (stating a moderating effect of self-control strength on the baseline intention - participation relation). Intention significantly predicted higher rates of re-attendance. Neither a meaningful effect of self-control, nor an interaction with baseline intention, was revealed. Hence, our first hypothesis could not be corroborated. The interaction term was therefore removed (model 1b) without impairing model fit (see Table 3). In a final step (model 2), to test our second hypothesis (that self-control strength would moderate the weekly intention – participation relation), weekly intention was entered, decomposed into a (time-invariant) between-person and a (time-varying) within-person component plus an interaction term of the within component with self-control. While the main effect of baseline self-control remained insignificant, there was a significant positive interaction of intra-individual intention fluctuations with self-control, verifying our second hypothesis. The chance of re-attendance was higher after weeks when an individual expressed a higher intention to re-attend than usual, but this effect was dependent on self-control. The mean level of weekly intention (between-component) of a person, in contrast, showed no significant effect beyond baseline intention.

Table 3. Comparison of Log-likelihoods (LL) of the sequence of tested models.

	LL	χ2	df	p-value
Model 0: baseline model including time aspects (participation week, weekˆ2, accumulated number of attended classes, past exercising (months), stratified by gender)	−3902.81	640.0a	6	<0.001
Model 1: adding baseline values for intention and self-control plus interaction	−3895.83	13.95	3	0.003
Model 1b: removing nonsignificant interaction term	−3895.84	0.01	1	0.907
Model 2: adding values for weekly intention plus interaction with self-control	−3887.44	16.80	3	<0.001
Model 2b: like 2, but self-control grouped into quintiles for illustration of interaction	−3883.30	8.28	6	0.218

Note. compared against null (empty) model

Table 4. Mixed effect Cox proportional hazard models for the prediction of time to the recurrent event re-attendance at exercise class (259 individuals with 982 events in 1243 intervals).

	Model 0 (stratified by gender, including only time aspects)	Model 1 (+baseline intention and self-control)	Model 2 (+weekly intention and interaction with self-control)
Fixed effects of predictors: HR (95% CI)
Past exercising (in months)	1.000 (0.999–1.001)	1.000 (0.999–1.001)	1.000 (0.999–1.001)
Accumulated number of attended classes	1.190 (1.112–1.273) ***	1.169 (1.093–1.250)***	1.158 (1.083–1.238)***
Intention (baseline)		1.064 (1.029–1.099)***	1.054 (1.015–1.095)**
Self-control (baseline)		0.947 (0.844–1.062)	0.951 (0.847–1.067)
Intention × self-control (baseline)		1.003 (0.950–1.060)	/
Weekly intention between-persons			1.016 (0.985–1.047)
Weekly intention within-person			1.050 (1.018–1.084)**
Intention within × self-control			1.076 (1.021–1.134)**
Random effects (variance intercept):
Individual	0.000	0.000	0.000
Exercise class	0.061	0.055	0.055

Note. All models are stratified by gender and adjusted for time (participant week and participant week squared).

*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001; HR = hazard ratio

To further explore and illustrate the interaction effect, self-control values were divided into quintiles. The effect of within-person intention fluctuations was then examined separately for the five groups with the lowest to the highest self-control, by including the interaction of self-control quintiles with weekly intention (model 2b). This model showed a similar model fit, like model 2. It revealed that the intention effect increased from an HR = 0.986 (95% CI: 0.923–1.052, p = 0.660) in the group with the lowest self-control over an HR = 1.024 (95% CI: 0.933–1.124, p = 0.620) and HR = 1.056 (CI: 0.968–1.153, p = 0.220) for medium self-control to an HR = 1.218 (CI: 1.072–1.384, p = 0.002) for the fourth self-control quintile, and a HR = 1.135 (CI: 1.030–1.250, p = 0.010) for those with the highest self-control values. That is to say, higher weekly intention values were only positively related to higher re-attendance rates in those two groups with the highest trait self-control. Although a higher baseline intention was also positively related to re-attendance, in terms of weekly intentions, not the absolute mean intention level of a person was of importance, but the deviation from that mean level, with higher re-attendance after weeks with higher intention scores only in those with high trait self-control. For those two groups with the highest self-control scores, a one point increase in weekly intention was associated with about 22% and 13.5% higher hazard rates to re-attend, adjusted for time course and baseline intention values, and stratified by gender.

4. Discussion

The intention-behavior gap describes a discrepancy between what an individual intends to do in a given situation (e.g., going to the gym) and what the individual is actually doing (e.g., meeting with friends; Rhodes et al., 2008). It has not been yet empirically tested whether self-control moderates the intention-behavior gap in terms of short-term intentions, long-term intentions, or both. In the present study, we addressed this shortcoming by analyzing both moderator effects of trait self-control, on the effects of the initial intention to participate in an exercise class regularly over the whole term, as well as on the weekly intentions to re-attend the class in the upcoming week, and particularly within-person fluctuations in this intention. We hypothesized that self-control strength would moderate both effects in the same line, namely by strengthening the observed positive associations with attendance as the behavior in question.

Our results show that, in a sample with overall typical self-control strength, individuals with higher levels of trait self-control strength were more likely to translate their short-term intentions into action. This result was in line with our second hypothesis. In this relation, not the absolute magnitude of intention was crucial, but the magnitude of the deviation from the intra-individual mean level. The result that within-person fluctuations are more important than between-person differences, even for short-term intentions measured on repeated occasions, is in line with other studies on the relation between short-term intentions and physical activity (Conroy et al., 2011, 2013). Baseline intention was also positively related to attendance but, contradicting our first hypothesis, this effect was not moderated by self-control.

Even though baseline as well as weekly intentions were better predictors of participation in the exercise classes than self-control, there was a remarkable inconsistency between intentions and actual behavior. Although nearly all individuals expressed extremely high intentions at baseline, as well as on a weekly basis, participation rates showed a steep decline. Mean intention scores for the two-item baseline intention were considerably higher than in most other studies that used this scale with college students (Conroy, Elavsky, Doerksen, & Maher, 2013; de Bruijn et al., 2014), although one smaller study also reported a rather high mean score for a similar two-item measure of weekly intentions (Conroy et al, 2011). The result that most participants did not transfer their (short- or long-term) intentions into action, further underpins the need to look for moderators of the intention-behavior relationship, not only in terms of long-term intentions measured once, but also for short-term intentions. Although within-person fluctuations in intentions did predict behavior in our study, a great deal of attendance was not explained in view of the constantly high intention scores. Particularly, between 28% and 58% of those individuals who stated the highest possible intention to re-attend the following week actually did not enact this intention. The number of accumulated training sessions during the term significantly increased the chance to re-attend the class. This finding can be interpreted as indicating a self-enhancing effect of successfully enacting an intended behavior. However, no interaction with intention was revealed (data not shown), i.e. we could not find that successfully enacting the behavior increased the intention-behavior relation.

The fact that we did not find a statistically significant main effect of self-control is to be expected, since as a moderator self-control is primarily needed for the translation of motivation into actions, i.e. intentions are required (Rhodes & Dickau, 2013). If an individual does not have the intention to work out regularly, it does not matter if he/she has high levels of self-control strength or not as there is no intention which needs to be shielded from distracting or tempting alternative behaviors (e.g., Englert, 2016).

4.1. Strengths and limitations

This study has several merits. First, PA behavior was observed for a period of several months in a relatively large sample. Secondly, the predicted criterion participation in the exercise class was measured quasi-objectively by weekly registration of attendance. Thirdly, we included intention measured at baseline, as well as within-person intention fluctuations, which constitute a large portion of PA intention variability that seems to be important for PA. Fourthly, the timeframe of our intention measures directly corresponded to the observed behavior as advised for intention measures.

However, we would also like to discuss the potential shortcomings of the present study. First, a high number of individuals dropped out during the course of the semester and did no longer attend the exercise classes. We neither have any information regarding the reasons for the high attrition, nor on the current intentions of those not attending at a specific week. A weekly online-survey would be a possibility to monitor those not attending the course, although the motivation to regularly participate in such an online-survey after leaving the exercise course seems questionable.

Secondly, the study was an observational study, and even though we predicted future events from preceding characteristics, causality of the observed effects cannot be determined.

Thirdly, the quality of the employed one-item intention measure remains unclear and the skewed distribution of the (baseline and weekly) intention scores points to a ceiling effect. Social desirability may possibly have also played a role. However, since the intention effects were large, even when considering these limitations, they may very well be underestimated in our study. Two-item-intention measures were used in many studies before and we are not aware of others reporting problems with highly skewed scores. Nevertheless, the development of short/one-item intention scales with proven measurement properties is highly necessary.

Fourthly, only intentions one week before the actual behavior were measured, with possible change in intentions during the week until the actual participation not being observed. Since other studies (Conroy et al, 2013) show considerable day-to-day-variation in PA intentions, one possible explanation for the low rate of enactment of the generally high weekly intentions is a decrease in motivation during the week until participation was able to take place. Since intention was only measured in those participants attending the exercise class in any given week, intentions moreover may seem higher than they were, because those people not attending the class, owing to a currently low intention on class day, did not provide data in these weeks.

Furthermore, past exercise history was self-reported and the item has not been validated. It is therefore possible that the nonsignificant effect was due to low reliability of the item.

We would also like to mention, that it appears as if self-control and conscientiousness are similar psychological constructs. Conscientiousness can also be considered a personality trait which describes a tendency to be orderly, achievement striving, self-disciplined, and deliberate. A systematic review revealed, that conscientiousness moderates the intention-behavior gap (Rhodes & Dickau, 2013). Future research should try to investigate the similarities and differences between these psychological constructs and how they contribute to the intention-behavior gap.

4.2. Implications

According to the strength model, individuals do not only differ in their trait self-control strength but also in their state self-control strength (Baumeister et al., 1994). If individuals had to control themselves during the course of the day, their self-control resources might be depleted in the evening, when they had originally intended to be physically active (Englert & Rummel, 2016). This temporary loss of self-control strength is termed ego depletion (Baumeister et al., 1994). In a state of ego depletion individuals are less capable at controlling themselves (e.g., Hagger et al., 2010). In the present study, we focused on trait self-control strength, which is why future research should also investigate the effects of state self-control strength on the intention-behavior gap. Especially, since we found evidence that only the effect of weekly intention fluctuations was moderated by trait self-control, it is possible that weekly fluctuations in self-control show even stronger effects on the intention-behavior gap. However, we would also like to mention that a recent replication study did not find empirical evidence for the ego depletion effect (Hagger et al., 2016), which is why alternative theoretical models might be better suited to explain temporary self-control impairments. For instance, the process model of self-control proposes that self-control impairments are not caused by ego depletion but rather by shifts in attention, motivation and emotions after previous self-control demands (Inzlicht & Schmeichel, 2012). Therefore, future research should also focus on how attentional, motivational, and emotional process might influence the intention-behavior gap (cf. Rhodes & Dickau, 2013).

As it has been reliably shown that trait self-control strength is associated with several positive outcomes, it should be investigated how self-control strength can be improved. Baumeister et al. (1994) compare self-control strength to a muscle, which can be trained. It has been empirically shown, that practicing self-control regularly improves self-control performance in the long-run (Baumeister, Gailliot, DeWall, & Oaten, 2006). For example, individuals who had been instructed to control certain behavioral tendencies over a 2-week period (e.g., use of the non-dominant hand in everyday life) displayed better self-control performances than participants from a control group afterwards (e.g., Gailliot, Plant, Butz, & Baumeister, 2007). Interestingly, interventions to improve self-control strength have not been tested rigorously in sport and exercise contexts, which is a shortcoming that needs to be addressed in future research.

4.3. Conclusion

In conclusion, the present study underlines the importance of trait self-control strength in physical exercise contexts. Interventions aimed at improving self-control strength might help to reduce the intention-behavior gap and enable individuals to exercise on a regular basis, which is required for achieving lasting health benefits.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflicts of interest

The authors hereby declare that there is no conflict of interest.

Declarations of interest

None.

Acknowledgements

We thank Florian Loetz, Sabrina Tonn and Percy Marks for their help with data collection.

Recommended articlesCiting articles (0)

References

Ach, 1905

N. Ach

Ueber die Willenstaetigkeit und das Denken

Vandenhoeck and Ruprecht, Goettingen, Germany (1905)

Google Scholar

Ajzen, 1991

I. Ajzen

The theory of planned behavior

Organizational Behavior and Human Decision Processes, 50 (1991), pp. 179-211, 10.1016/0749-5978(91)90020-T

Article Download PDF View Record in Scopus Google Scholar

Andersen and Gill, 1982

P.K. Andersen, R.D. Gill

Cox's regression model for counting processes: A large sample study

Annals of Statistics, 10 (4) (1982), pp. 1100-1120

CrossRef Google Scholar

Baumeister et al., 1998

R.F. Baumeister, E. Bratslavsky, M. Muraven, D.M. Tice

Ego depletion: Is the active self a limited resource?

Journal of Personality and Social Psychology, 74 (1998), pp. 1252-1265, 10.1037/0022-3514.74.5.1252

View Record in Scopus Google Scholar

Baumeister et al., 2006

R.F. Baumeister, M. Gailliot, C.N. DeWall, M. Oaten

Self-regulation and personality: How interventions increase regulatory success, and how depletion moderates the effects of traits on behavior

Journal of Personality, 74 (2006), pp. 1773-1801, 10.1111/j.1467-6494.2006.00428.x

CrossRef View Record in Scopus Google Scholar

Baumeister et al., 1994

R.F. Baumeister, T.F. Heatherton, D.M. Tice

Losing control: How and why people fail at self-regulation

Academic Press, San Diego, CA (1994)

Google Scholar

Bertrams and Dickhäuser, 2009

A. Bertrams, O. Dickhäuser

Diagnostica, 55 (2009), pp. 2-10, 10.1026/0012-1924.55.1.2

CrossRef View Record in Scopus Google Scholar

Bertrams and Englert, 2013

A. Bertrams, C. Englert

Sportwissenschaft, 43 (2013), pp. 276-282, 10.1007/s12662-013-0304-x

CrossRef View Record in Scopus Google Scholar

Bhui and Fletcher, 2000

K. Bhui, A. Fletcher

Common mood and anxiety states: Gender differences in the protective effect of physical activity

Social Psychiatry and Psychiatric Epidemiology, 35 (2000), pp. 28-35, 10.1007/s001270050005

View Record in Scopus Google Scholar

Biddle and Mutrie, 2007

S.J. Biddle, N. Mutrie

Psychology of physical activity: Determinants, well-being and interventions

Routledge, New York (2007)

Google Scholar

Bolger and Laurenceau, 2013

N. Bolger, J.-P. Laurenceau

Intensive longitudinal methods: An introduction to diary and experience sampling research

Methodology in the social sciences

Guilford Press, New York (2013)

Google Scholar

de Bruijn et al., 2014

G.-J. de Bruijn, B. Gardner, L. van Osch, F.F. Sniehotta

Predicting automaticity in exercise behaviour: The role of perceived behavioural control, affect, intention, action planning, and behaviour

International Journal of Behavioral Medicine, 21 (5) (2014), pp. 767-774, 10.1007/s12529-013-9348-4

CrossRef View Record in Scopus Google Scholar

Conroy et al., 2013

D.E. Conroy, S. Elavsky, S.E. Doerksen, J.P. Maher

A daily process analysis of intentions and physical activity in college students

Journal of Sport & Exercise Psychology, 35 (5) (2013), pp. 493-502, 10.1123/jsep.35.5.493

CrossRef View Record in Scopus Google Scholar

Conroy et al., 2011

D.E. Conroy, S. Elavsky, A.L. Hyde, S.E. Doerksen

The dynamic nature of physical activity intentions: A within-person perspective on intention-behavior coupling

Journal of Sport & Exercise Psychology, 33 (6) (2011), pp. 807-827, 10.1123/jsep.33.6.807

CrossRef View Record in Scopus Google Scholar

Englert, 2016

C. Englert

The strength model of self-control in sport and exercise psychology

Frontiers in Psychology, 7 (2016), p. 314, 10.3389/fpsyg.2016.00314

Google Scholar

Englert and Rummel, 2016

C. Englert, J. Rummel

I want to keep on exercising but I don't: The negative impact of momentary lack of self-control on exercise adherence

Psychology of Sport and Exercise, 26 (2016), pp. 24-31, 10.1016/j.psychsport.2016.06.001

Article Download PDF View Record in Scopus Google Scholar

Gailliot et al., 2007

M.T. Gailliot, E.A. Plant, D.A. Butz, R.F. Baumeister

Increasing self-regulatory strength can reduce the depleting effect of suppressing stereotypes

Personality and Social Psychology Bulletin, 33 (2007), pp. 281-294, 10.1177/0146167206296101

CrossRef View Record in Scopus Google Scholar

Goodwin, 2003

R.D. Goodwin

Association between physical activity and mental disorders among adults in the United States

Preventive Medicine, 36 (2003), pp. 698-703, 10.1016/S0091-7435(03)00042-2

Article Download PDF View Record in Scopus Google Scholar

Hagger, 2010

M.S. Hagger

Self-regulation: An important construct in health psychology research and practice

Health Psychology Review, 4 (2010), pp. 57-65, 10.1080/17437199.2010.503594

CrossRef View Record in Scopus Google Scholar

Hagger et al., 2016

M.S. Hagger, N.L.D. Chatzisarantis, H. Alberts, C.O. Anggono, A. Birt, R. Brand, ..., T. Cannon

A multi-lab pre-registered replication of the ego depletion effect

Perspectives on Psychological Science, 11 (2016), pp. 546-573, 10.1177/1745691616652873

CrossRef View Record in Scopus Google Scholar

Hagger et al., 2009

M.S. Hagger, C. Wood, C. Stiff, N.L. Chatzisarantis

The strength model of self-regulation failure and health-related behaviour

Health Psychology Review, 3 (2009), pp. 208-238, 10.1080/17437190903414387

CrossRef View Record in Scopus Google Scholar

Hagger et al., 2010

M.S. Hagger, C. Wood, C. Stiff, N.L.D. Chatzisarantis

Ego depletion and the strength model of self-control: A meta-analysis

Psychological Bulletin, 136 (2010), pp. 495-525, 10.1037/a0019486

CrossRef View Record in Scopus Google Scholar

Hallal et al., 2012

P.C. Hallal, L.B. Andersen, F.C. Bull, R. Guthold, W. Haskell, U. Ekelund

Global physical activity levels: Surveillance progress, pitfalls, and prospects

The Lancet, 380 (2012), pp. 247-257, 10.1016/S0140-6736(12)60646-1

Article Download PDF View Record in Scopus Google Scholar

Hamilton et al., 2008

M.T. Hamilton, G.N. Healy, D.W. Dunstan, T.W. Zderic, N. Owen

Too little exercise and too much sitting: Inactivity physiology and the need for new recommendations on sedentary behavior

Current Cardiovascular Risk Reports, 2 (2008), pp. 292-298, 10.1007/s12170-008-0054-8

CrossRef View Record in Scopus Google Scholar

Inzlicht and Schmeichel, 2012

M. Inzlicht, B.J. Schmeichel

What is ego depletion? Toward a mechanistic revision of the resource model of self-control

Perspectives on Psychological Science, 7 (2012), pp. 450-463, 10.1177/1745691612454134

CrossRef View Record in Scopus Google Scholar

Lewin, 1951

K. Lewin

Intention, will, and need

D. Rapaport (Ed.), Organization and pathology of human thought, Columbia University Press, New York (1951), pp. 95-153

CrossRef Google Scholar

Martin Ginis and Bray, 2010

K.A. Martin Ginis, S.R. Bray

Application of the limited strength model of self-regulation to understanding exercise effort, planning and adherence

Psychology and Health, 25 (2010), pp. 1147-1160, 10.1080/08870440903111696

CrossRef View Record in Scopus Google Scholar

Rhodes and Bruijn, 2013

R.E. Rhodes, G.-J de Bruijn

How big is the physical activity intention-behaviour gap? A meta-analysis using the action control framework

British Journal of Health Psychology, 18 (2) (2013), pp. 296-309, 10.1111/bjhp.12032

CrossRef View Record in Scopus Google Scholar

Rhodes and Dickau, 2013

R.E. Rhodes, L. Dickau

Moderators of the intention-behaviour relationship in the physical activity domain: A systematic review

British Journal of Sports Medicine, 47 (4) (2013), pp. 215-225, 10.1136/bjsports-2011-090411

CrossRef View Record in Scopus Google Scholar

Rhodes et al., 2008

R.E. Rhodes, R.C. Plotnikoff, K.S. Courneya

Predicting the physical activity intention–behavior profiles of adopters and maintainers using three social cognition models

Annals of Behavioral Medicine, 36 (2008), pp. 244-252, 10.1007/s12160-008-9071-6

CrossRef View Record in Scopus Google Scholar

Rhodes and Rebar, 2017

R.E. Rhodes, A.L. Rebar

Conceptualizing and defining the intention construct for future physical activity research

Exercise and Sport Sciences Reviews, 45 (2017), pp. 209-216, 10.1249/JES.0000000000000127

View Record in Scopus Google Scholar

Rhodes and Yao, 2015

R.E. Rhodes, C.A. Yao

Models accounting for intention-behavior discordance in the physical activity domain: A user's guide, content overview, and review of current evidence

International Journal of Behavioral Nutrition and Physical Activity, 12 (2015), p. 9, 10.1186/s12966-015-0168-6

CrossRef View Record in Scopus Google Scholar

de Ridder et al., 2012

D.T. de Ridder, G. Lensvelt-Mulders, C. Finkenauer, F.M. Stok, R.F. Baumeister

Taking stock of self-control: A meta-analysis of how trait self-control relates to a wide range of behaviors

Personality and Social Psychology Review, 16 (2012), pp. 76-99, 10.1177/1088868311418749

CrossRef View Record in Scopus Google Scholar

Schwarzer and Luszczynska, 2008

R. Schwarzer, A. Luszczynska

How to overcome health-compromising behaviors: The health action process approach

European Psychologist, 13 (2008), pp. 141-151, 10.1027/1016-9040.13.2.141

CrossRef View Record in Scopus Google Scholar

Sniehotta et al., 2014

F.F. Sniehotta, J. Presseau, V. Araújo-Soares

Time to retire the theory of planned behaviour

Health Psychology Review, 8 (2014), pp. 1-7, 10.1080/17437199.2013.869710

CrossRef View Record in Scopus Google Scholar

Stork et al., 2016

M.J. Stork, J.D. Graham, S.R. Bray, K.A. Martin Ginis

Using self-reported and objective measures of self-control to predict exercise and academic behaviors among first-year university students

Journal of Health Psychology (2016), 10.1177/1359105315623627

Advance online publication

Google Scholar

Ströhle, 2009

A. Ströhle

Physical activity, exercise, depression and anxiety disorders

Journal of Neural Transmission, 116 (2009), pp. 777-784, 10.1007/s00702-008-0092-x

CrossRef View Record in Scopus Google Scholar

Tangney et al., 2004

J.P. Tangney, R.F. Baumeister, A.L. Boone

High self-control predicts good adjustment, less pathology, better grades, and interpersonal success

Journal of Personality, 72 (2004), pp. 271-324, 10.1111/j.0022-3506.2004.00263.x

View Record in Scopus Google Scholar

Therneau, 2018

T.M. Therneau

coxme: Mixed effects Cox models. R package version 2.2-7

Retrieved from

https://CRAN.R-project.org/package=coxme (2018, March 15)

Google Scholar

World Health Organization, 2010

World Health Organization

Global recommendations on physical activity for health

World Health Organization, Geneva, Switzerland (2010)

Google Scholar

AB #7

How reliable are the effects of self-control training?: A re-examination using self-report and physical measures

Brian M. Lee☯*, Markus Kemmelmeier☯ Interdisciplinary Social Psychology Ph.D. Program, University of Nevada, Reno, Nevada, United States of

America ☯ These authors contributed equally to this work.

* blee2@unr.edu

Abstract

In light of recent challenges to the strength model of self-control, our study re-examines the effects of self-control training on established physical and self-report measures of self-con- trol. We also examined whether beliefs about the malleability of self-control qualify any train- ing effects. Participants in the training condition were assigned to increase use of their non- dominant hand for two weeks, and did comply mainly if they held high-malleability beliefs; yet, compared to a control condition, the physical measure of self-control did not improve. This was also evident in a secondary objective measure of self-control, a Stroop task, as well as in self-reported self-control. The discussion focuses on the lack of replication of train- ing effects on self-control.

Introduction

Self-control is one of the most important human endowments, as it allows people to limit impulsive behaviors [1]. Poor self-control has been found to be related to numerous problems, such as obesity, criminality, risky sexual behavior, drug and alcohol use, as well as other nega- tive outcomes [2–6]. Conversely, high self-control has been found to be related to better grades, less psychopathology, better relationships, better interpersonal skills, healthier eating habits, better emotional control, as well as other positive outcomes [4, 7–9].

The strength model of self-control argues that self-control behaves like a muscle in that it becomes weakened from active use and people are less successful at using self-control on sub- sequent tasks which require its use [4]. Similarly, repeated use can strengthen self-control, much like exercising a muscle, making it less susceptible to becoming weakened from active use and leading to better subsequent outcomes on tasks which also require self-control. Thus, self-control seems to increase following training. A meta-analysis by Hagger, Wood, Stiff, and Chatzisarantis [10] reported an overall positive effect of training on improving self-control, though few studies have examined how long effects persist after the end of the training. Hui, Wright, Stewart, Simmons, Eaton, and Nolte [11] documented some persistence of effects, but findings were complicated by the specific domain of the research (dental hygiene). Muraven

PLOS ONE | https://doi.org/10.1371/journal.pone.0178814 June 8, 2017 1 / 14

Competing interests: The authors have declared that no competing interests exist.

[12] tracked participants for a month following training, but again, the persistence of effects was domain specific (smoking cessation). Recently, Bertrams and Schmeichel [13] found that training effects were not sustained after the end of training, contradicting the previous findings.

With the persistence of self-control training (SCT) effects being unclear, the first goal of the current study was to examine the robustness of SCT effects. Such an examination is pertinent because a recent meta–analysis by Carter and McCullough [14] showed that there is a publica- tion bias with regard to research on self-control depletion. This raises potential doubt with regard to other self-control research, especially because informal communications with self- control researchers revealed that unsuccessful and unpublished training studies exist, pointing to a file-drawer problem [15]. Moreover, a meta-analysis using alternative procedures from those by Hagger et al. [10] obtained somewhat smaller effect sizes than reported by these authors [16].

A second goal of this study was to examine the role of implicit beliefs in SCT. Research on self-control highlights the importance of implicit beliefs about self-control. Job, Dweck, and Walton [17] found that whether participants’ believed willpower to be a limited resource or not moderated depletion effects. Participants who did not believe willpower to be limited did not demonstrate reduced self-control after having engaged in a depleting task [17, 18]. Simi- larly, it may be possible that implicit beliefs about the trainability of self-control impact whether or not SCT effects manifest following training. Thus, we developed a scale which mea- sures implicit beliefs about the malleability of self-control. We hypothesized that beliefs in the malleability of self-control would be linked with (a) greater compliance with the training instructions, and ultimately (b) greater success of the SCT.

Participants attended three study sessions, each two weeks apart. In the first session they were randomly assigned to an established SCT condition or a control condition and asked to follow instructions during the period between Session 1 and 2. During this time they also reported their compliance with instructions via a website. Participants did not receive any instructions for the period between Session 2 and 3. At all three sessions participants com- pleted a set of self-report measures and completed the physical self-control task.

Notably, we used a SCT task and a physical self-control measure which are both well-estab- lished in the literature, but which are similar in nature. Our SCT asked participants to increase the use of their non-dominant hand. This procedure has been demonstrated to increase self- control, such that following this procedure participants were better able to resist the tempta- tion to smoke [12], less likely to entertain thoughts about and engage in interpersonal aggres- sion [19] and more resistant to self-control depletion [20]. To assess self-control, our participants were asked to squeeze a handgrip exerciser for as long as they could, with the time serving as a physical measure of self-control. The meta-analysis by Hagger et al. [10] revealed this to be a popular task when assessing self-control strength [1, 21, 22, 23]. Although muscle strength contributes to the time that a participant can hold a handgrip exerciser, previous research has documented that it is best understood as a measure of self-control as participants are required to withstand the discomfort resulting from continuing to hold the device [24, 25].

Whereas participants in the SCT condition were told to increase the use of their non-domi- nant hand during the training period, the physical measure was applied to both dominant and non-dominant hands. If participants use their non-dominant hand as part of the training regime, it should not be surprising if this increased use of the non-dominant hand also improves subsequent muscle strength in this particular hand. However, there is no reason to believe that such increased muscle strength in the non-dominant hand would translate to improved performance in the dominant hand unless the training also improved self-control. Muraven et al. [1] used participants’ ability to squeeze a handgrip as a dependent variable in

Reliability of self-control training effects

PLOS ONE | https://doi.org/10.1371/journal.pone.0178814 June 8, 2017 2 / 14

Reliability of self-control training effects

assessing changes in self-control, but did not mention whether participants used their domi- nant or their non-dominant hand to do so [21, 23]. Note that Muraven et al.’s [1] training regime did not involve use of any particular hand; hence, these authors seemed to consider this issue unimportant enough to not even report which hand was used, nor whether partici- pants used the same hand before and after the training. In contrast, the present investigation does make use of a SCT procedure focused on the non-dominant hand, thus making it impera- tive to differentiate possible effects on muscle-strength improvement from self-control improvements by assessing performance on the handgrip task with each hand in all sessions.

Additionally, a subset of participants completed a second objective measure of self-control, namely, a Stroop task. This secondary measure was included to confirm potential findings per- taining to the physical measure. The Stroop task, in which participants are shown color words in various font colors and indicate the color of the word while ignoring what the word actually says, has also been used in previous self-control research, and is another popular measure for self-control [10, 26, 27].

Method

Participants

All participants provided written informed consent at the initial session, prior to engaging in the research activities. The consent forms and procedures were approved by the University of Nevada, Reno Social Behavior and Education IRB (approval number 508754–5). Across four different semesters, 147 students (74.8% female; 67.6% White) participated in the study in exchange for course credit and $15 (USD). The study consisted of three lab sessions.

Assuming the estimate for training effects reported by Hagger et al. [10], d = 1.07 (p. 510, averaged corrected standardized difference effect size), and assuming α = .05 and β = .99, the required sample size was 68 to detect an effect of this magnitude. This estimate pertains to the comparison between a training condition and a control condition following the training. Only after the study was completed and the present manuscript submitted for publication did the authors learn about two recent meta-analyses that estimated the effects to be smaller, Hedges’ g = .30 [28], and Hedges’ g = .36 [29] (note that Hedges’ g and Cohen’s d only differ in that g includes a correction factor for sample size).

One can question whether a simple comparison between a control group and a self-control training group does provide the correct basis for a sample size estimate because we employed a mixed-model design which was predicated on the notion that the control group and the train- ing group would diverge over time. Indeed, ours was a mixed-model design, in which we assigned each participant to an experimental condition (between-groups factor) but assessed each participant in three different sessions (repeated-measures factor) on both their dominant and nondominant hand (repeated-measures factor), for which we expected differential effects over time. Hence, it was important to determine the sensitivity of this design for a Condition x Session x Hand interaction. Assuming α = .05 and β = .80, as well as an average correlation of r = .60 between measurements, our planned sample size of 68 was able to detect a Condition x Hand x Session interaction effect with an effect size that corresponded to a Cohen’s d of .23— arguably a small effect size. Note that, other than the estimates reported by previous authors [10, 28, 29] we do not focus on a simple comparison between a training group and a control group post-training. Rather, our estimate pertains to the differential change that emerges between these two groups over the course of the training.

The data were collected in two batches. In the first set of data, 108 students participated across three successive semesters. After removal of partial completers and ambidextrous stu- dents, the number of usable cases reached 65, thus making it necessary to collect additional

PLOS ONE | https://doi.org/10.1371/journal.pone.0178814 June 8, 2017 3 / 14

data. Because preliminary analyses of the incomplete data set revealed ambiguous results, in the final semester of data collection (resulting in 39 more participants), we added one depen- dent variable to the very end of the procedure to examine the consistency of potential findings across measures.

Procedure

Session 1. At the first session, participants provided written consent, and then completed all of the self-control measures (discussed below). Participants were given the self-report mea- sures first, then the physical measure, and lastly, for the second batch, the Stroop task. Partici- pants were randomly assigned to conditions via a coin toss and received corresponding instructions.

SCT condition. Participants in the training condition were instructed to use their non- dominant hand for mundane, daily tasks (e.g., carrying objects, brushing their teeth, stirring beverages) between the hours of 8AM and 6PM. This training program is modeled after the exercises in Gailliot et al. [20] (see Study 2 & 4); also [19], which was previously demonstrated to bolster self-control.

Control condition. Participants assigned to the control condition were instructed to keep a journal of the temptations they encountered daily (e.g., not doing homework when they ought to, making poor eating choices, hanging out with friends instead of studying) between the hours of 8AM and 6PM, and to do so whether they resisted the temptation or capitulated. This control condition was modeled after Muraven [12], and served the purpose of making self-control behaviors salient, without requiring participants to practice self-control. Notably, keeping a journal has been occasionally used in the context of a self-control manipulation [1, 8, 9, 30]. However, diaries were either used to record participants’ activities [1, 8, 9, 30], or to record participants’ food intake [1]. Even though in Muraven et al. [1] there were no explicit instructions to modify food intake while keeping a food diary, keeping a food journal may lead to a spontaneous modification of behavior, though effects are small [31, 32]. Nevertheless, the task of recording one’s food intake might be reactive and increase the likelihood of people exercising self-control. However, a similar process is much less likely to occur for the record- ing of daily temptations, which are experienced often as spontaneous and outside of the indi- vidual’s volition. Consistent with this notion, Hufford et al. [33] found little evidence for problem drinkers tasked with recording their alcohol-related temptations to alter their behav- ior. Moreover, Muraven [12] demonstrated that the journaling task was no different in its con- sequences on self-control improvement from an alternative control condition (performing 3–5 minutes of simple math problems twice a day).

Participants in both conditions were told to follow instructions for a period of two weeks ending with Session 2.

Interim between Session 1 and Session 2. All participants regardless of condition received text messages every three days reminding them to follow condition-specific study instructions. Participants were also reminded to complete an online survey which inquired about their levels of compliance with instructions. For participants in the SCT condition the survey asked how often participants engaged in using their non-dominant hand, and for par- ticipants in the control condition how often they recorded temptations in their journal. Rat- ings were made on a 7-point scale (1 Not At All, 7 Very Often). Each time they accessed the online survey, participants listed examples of their use of their non-dominant hand or the temptations they encountered.

Session 2. Two weeks after Session 1, participants returned to the lab to complete all mea- sures previously assessed in Session 1. Moreover, at the end of Session 2, participants were

Reliability of self-control training effects

PLOS ONE | https://doi.org/10.1371/journal.pone.0178814 June 8, 2017 4 / 14

instructed to no longer engage in the condition-specific instructions. Additionally, no further text message reminders were sent to participants between Session 2 and Session 3.

Session 3. Again, participants completed all measures from Session 1 and Session 2. At the end of Session 3, participants were debriefed.

Measures

Across three sessions, participants received the same set of self-report measures, with partici- pants indicating how much an item applied to them or how much they agreed with it (1 Not At All, 7 Very Much).

Self-reported self-control. Levels of self-control were assessed using the 13-item Brief Self-Control Scale [7] (α = .81, .85 and .84 for Sessions 1, 2, and 3, respectively) as well as the 24-item Low Self-Control Scale [34], a popular scale in criminology with reliabilities (α = .82, .87, and .89). Scales were scored such that higher values always indicated greater levels of self- control. The authors of these scales generally conceive of one’s capacity to exercise self-control as a disposition [7, 34, 35]. However, with training being able to induce changes in self-control over time [1, 8, 9, 12, 20, 30], it should not be surprising that authors have frequently reported changes in self-reported self-control over time [36–38].

Implicit beliefs about depletability of self-control. Participants completed the scale by Job et al. [17], which assessed whether participants believed that self-control was a limited resource and would be depleted whenever participants engaged in the exercise of self-control. We used the 12-item version employed in their Study 4, which included both subscales per- taining to the effects of strenuous mental activity as well as the effects of resisting temptation (α = .83, .88 and .89). (Results did not change if we only employed the subset of six items, which represented the scale used by Job et al. [17] in their Study 1.) Higher scores indicate a greater belief in self-control being a non-depletable resource.

Implicit beliefs about malleability of self-control. We generated a novel scale that mea- sured participants’ beliefs as to whether their capacity to exercise self-control is malleable, and whether their capacity to resist temptation is malleable. Parallel to Job et al. [17], we generated 10 items addressing malleability of self-control and 10 items pertaining to resisting impulsivity. Based on principal component analyses, we reduced the number of items in the malleability of self-control scale to nine, and the number of items of the resisting impulsivity scale to six (see Fig 1 for final items). Reliabilities of the two scales were satisfactory, all α > .73, across the three points in time. Because all items loaded on the same factor, focusing on Session 1 data we performed a confirmatory factor analysis to find that a two-factor solution, distinguishing malleability of self-control and resisting impulsivity, reveal a slightly better model fit than a single-factor solution, AIC = 7269.63 vs. 7272.75, -2 LL = 7177.64 vs. 7182.76, likelihood ratio test χ2(1) = 5.12, p = .024. The same pattern was obtained for Session 2 and 3 data. However, the subscales were highly correlated (r = .68, .66 and .78), and for results presented here they were combined into a single scale, α = .87, .90 and .91 (Sessions 1, 2 and 3, respectively), as sub- sequent analyses did not reveal any discernable difference in findings using a single scale or two parallel scales.

Physical measure of self-control. Following completion of the self-report measures, par- ticipants were given the handgrip task previously employed by other researchers [1, 21, 23]. Specifically, participants were asked to hold a handgrip exerciser closed as long as they could, which was timed. To determine when the participant’s hand was no longer applying the neces- sary amount of force to maintain closure of the handgrip exerciser, a small foam square was inserted between the handles. When the foam square fell out, the timer was stopped. The length of time that participants were able to squeeze the handgrip exerciser served as a physical

Reliability of self-control training effects

PLOS ONE | https://doi.org/10.1371/journal.pone.0178814 June 8, 2017 5 / 14

Fig 1. Malleability of self-control items and loadings from principle component analysis. Analysis was based on Session 1 data (n = 147). The dominant factor explained 40.4% of the variance.

https://doi.org/10.1371/journal.pone.0178814.g001

measure of self-control [1, 21, 23]. Other than previous authors who did not report which hand participants used to squeeze the handgrip exerciser [1, 21, 23], we recorded participants’ times separately for both their dominant hand and their non-dominant hand.

Stroop. After completing the handgrip exerciser task a subset of participants (n = 39), all of whom participated in the fourth of the four semesters during which this study was run, were asked to complete a color-word Stroop task (available through millisecond.com) in English. Participants were shown words (“red”, “blue”, “green”, or “black”) which were pre- sented in various font colors. In some instances, the word and its font color were congruent. In others, they were incongruent. Participants were to indicate the color of the font while ignoring what the word actually said. Also, solid blocks of color were displayed, which served as a control. Participants’ response times were recorded in milliseconds, which served as a

Reliability of self-control training effects

PLOS ONE | https://doi.org/10.1371/journal.pone.0178814 June 8, 2017 6 / 14

secondary physical measure of self-control. A total of 84 trials are given to each participant (4 colors x 3 color-stimulus congruency (congruent, incongruent, control) x 7 repetitions).

Results

Participant signup, compliance and retention

Time of signup. Participants signed up for the study anytime between the first and the twelfth week of a 16-week semester. Because higher levels of self-control in participants might be linked to completing course-related responsibilities sooner rather than later, we examined whether individual differences related to self-control, malleability beliefs and non-depletability beliefs predicted when in the semester participants joined the study [39, 40]. Zero-order corre- lations suggested that people who believed that self-control was malleable (as measured at Session 1) were more likely to sign up earlier in the semester (see Fig 2). When week of the semester of participants’ first session of the study was regressed onto these three predictors (assessed at Session 1, using Tangney et al. [7] as the measure of self-control) there was only a trend for individuals with higher malleability beliefs to sign up earlier in the semester, b = -0.78, se = 0.46, p = .093 (controlling gender and age). When this model was run again using Grasmick et al. [34] as the measure of self-control, very similar results emerged. No predictors were significant, and again, there was only a trend for individuals with higher malleability beliefs to sign up earlier in the semester, b = -0.76, se = 0.46, p = .105. This is consistent with the notion that believing in the malleability of self-control was part of the motivation of signing up for the study early in the term, as the study was advertised as a self-control training study. Importantly, a separate analysis of variance showed that there was no difference in the week of the semester during which participants in the training condition and the control condition participated in the study, F < 1.

Retention. There was considerable loss of research participants over the course of the study. Whereas 147 students participated in Session 1, 113 attended Session 2, and 87 attended Session 3. Although a substantial decline, our retention rate was comparable to other SCT studies using student samples [11, 41, 42], though other than in [41], we did not observe any differential attrition by experimental condition. However, most studies using student samples do not report any information concerning participants terminating a multi-week study prema- turely [1, 12, 19, 43]. Because dropout might reflect lack of self-control, we predicted retention using logistic regression based on trait self-control (as before, one model using Tangney et al. [7] and another model using Grasmick et al. [34]), malleability beliefs (at Session 1), non- depletability beliefs (at Session 1), age, and gender. No effects emerged. Dropout between Ses- sion 2 and Session 3 was also examined with logistic regression using the same predictors, and again, no effects emerged.

Compliance. We computed average compliance ratings (M = 4.30, SD = 1.23; range 1.33– 7.00) based on all ratings that participants provided between Session 1 and Session 2, regard- less of how many such ratings participants had provided. Participants were expected to submit four compliance ratings (median = 3), though a total of 16.3% of participants did not submit any compliance ratings, whereas 11.6% participants reported more than the expected number of four (maximum of 6). Notably, the SCT and control conditions did not differ with regard to the average compliance ratings nor in the number of such ratings that were received, both

F < 1. Compliance did not change over time either, meaning that participants’ first, second, third and fourth self-reported compliance scores were compared, F(3, 256.5) = 0.36, p = .78.

Malleability beliefs at Session 1 correlated with the average compliance ratings during the subsequent two weeks, though non-depletability beliefs showed a similar correlation (see Fig 2). An analysis in which average compliance was regressed on Session 1 malleability beliefs,

Reliability of self-control training effects

PLOS ONE | https://doi.org/10.1371/journal.pone.0178814 June 8, 2017 7 / 14

Reliability of self-control training effects

Fig 2. Descriptive statistics and correlations of measures, Session 1 through Session 3. For Tangney et al. higher values indicate higher levels of self-control. Dropout scores at Session 1 reflect whether a participant discontinued participation following Session 1. Dropout scores at Session 2 reflect whether a participant discontinued participation following Time 2 after having participated at both Time 1 and Time 2. Week is referring to week of the semester for which participants’ first session occurred. SCT is self-control training. Handtimes are non-transformed, raw scores.

https://doi.org/10.1371/journal.pone.0178814.g002

non-depletability beliefs, self-reported self-control (Tangney et al. [7] measure), age, and gen- der showed that malleability beliefs, b = .61, se = .16, p < .001, as well as non-depletability beliefs, b = .26, se = .13, p = .050, predicted compliance. The same model using Grasmick et al. [34] self-reported self-control revealed very similar results. Malleability beliefs predicted com- pliance, b = .62, se = .16, p < .001, and non-depletability beliefs was nearing significance as well, b = .24, se = .12, p = .054.

PLOS ONE | https://doi.org/10.1371/journal.pone.0178814 June 8, 2017 8 / 14

Self-report measures: Stability over time and correlations

We report analyses including participants who contributed to all three sessions of the study; though we also conducted comparisons including all participants who contributed to Session 1 and Session 2. Results were consistent. Fig 2 provides means and standard deviations for all measures, displayed for each point in time in the bottom row of each panel. It also shows those for each experimental condition separately for additional information.

We used a series of general linear models to examine changes over time in each variable by condition over time. Means for Job et al.’s non-depletability measure [17] increased over time, F(2, 82) = 7.95, p = .001, ηp2 = .16. Examination of simple slopes showed that there was a sub- stantive change between Session 1 and Session 2, p = .003, and between Session 1 and Session 3, p < .001, even when there was no change between Session 2 and Session 3, p = .10. Mallea- bility means did not vary over time, F < 1. However, there was an increase in self-reported self-control with the Tangney measure, F(2, 83) = 10.04, p < .001, ηp2 = .20, with scores increasing between Session 1 and Session 2, p = .001, and Session 1 and Session 3, p < .001, but no change between Session 2 and Session 3, p = .080. No change over time occurred with the Grasmick measure of self-reported self-control though, F(2, 83) = 0.77, p = .47, ηp2 = .02. Across sessions, all three scores were highly correlated, non-depletability all r > .75, p < .001, malleability beliefs all r > .79, p < .001, Tangney self-reported self-control all r > .84, p < .001, and Grasmick self-reported self-control all r > .83, p < .001. Fig 2 also displays the correlations of measures with one another at all three sessions. The pattern was remarkably consistent over time, but experimental condition never produced any significant effect. Notably, Job et al.’s [17] non-depletability beliefs were related to both measures of self-reported self-control [7, 34].

Of particular interest were the correlations pertaining to the new measures of malleability beliefs. This scale did not correlate with self-reported self-control. There was, however, an association between malleability beliefs and non-depletability beliefs, which emerged at all three points in time. This pattern is consistent with the notion that malleability beliefs are not redundant with trait self-control, and that they are conceptually distinct from, even though empirically related to, non-depletability beliefs.

Effects of SCT

Physical self-control. Times of how long participants squeezed the handgrip exerciser with their dominant and non-dominant hand at the three different sessions were analyzed using a multilevel analysis. Specifically, because of the non-independent data structure, we used a generalized linear mixed model that took account of the fact that both times for the dominant and non-dominant hands across three sessions were nested within the same partici- pant. This model allowed for performance comparisons between the two hands and, more importantly, an analysis of differential changes in the performance of each hand over time. The error covariance matrix assumed that errors at different points in time were independent from one another. We expected to observe a differential change for the dominant and non- dominant hand over time for participants in the control and the training condition. Our initial generalized linear mixed model included Hand (dominant vs. non-dominant), Session (1–3), Condition (SCT vs. control) as categorical predictors. Condition was modeled as a between- participant fixed effect, whereas Hand and Session were modeled as within-participant ran- dom effects. This analysis was based on those participants who had contributed data in all three sessions. One ambidextrous person was excluded because it was not possible to distin- guish a dominant and a non-dominant hand, leaving 79 participants—a number greater than

Reliability of self-control training effects

PLOS ONE | https://doi.org/10.1371/journal.pone.0178814 June 8, 2017 9 / 14

Reliability of self-control training effects

the sample size requirement discussed earlier. Prior to analysis, all hand-times were submitted to a log-transformation to normalize distributions.

This model (-2 LL = 666.92, AIC = 685.32) revealed a main effect for Hand showing individ- uals had longer hand-times for their dominant hand than their non-dominant hand (Mlog(s) = 3.94 vs. 3.81), F(1, 52) = 7.60, p = .008, which was qualified by interactions with Session, F(2, 29) = 3.96, p = .03, and condition, F(1, 52) = 4.93, p = .031. However, the anticipated three- way interaction showing a training effect on the dominant hand nor any other effect did not approach significance, Fs < 1.19, ps > .31. With 79 participants and assuming α = .05, β = .80, and with an observed average r = .62, this analysis should have been able to detect a small inter- action effect of d = .21.

As discussed earlier, we were interested if implicit beliefs about the malleability of self-con- trol would moderate such changes over time. Therefore, we added to our model malleability beliefs (at Session 1) and average compliance ratings as continuous predictors to test for possi- ble moderator effects. All interactions involving the three factors critical to our experimental design, Hand, Session, Condition, and either malleability beliefs or average compliance were included. That is, we generated all possible interactions, but never allowed malleability beliefs and average compliance to interact with each other. To control for other extraneous effects, we also controlled for gender, week of sign up, and depletability beliefs at Session 1 (all main effects only). The resulting model fit the data worse than the above simpler model (-2 LL = 685.42, AIC = 703.88). In addition to the Hand main effect, F(1,8) = 8.65, p = .018, the Hand by Session interaction, F(2,8) = 4.97, p = .039, and the Hand by Condition interaction, F(1,8) = 6.83, p = .033, we only found a tendency for men to squeeze the handgrip exerciser longer than women, Mlog(s) = 3.63 vs. 4.21), F(1, 2) = 14.92, p = .071. No other significant effects were found, all ps > .15.

Because multilevel models are based on maximum likelihood estimation, they are quite tol- erant for missing values, which allowed us to relax inclusion criteria and add participants’ data into the analysis who had completed two of the three sessions in addition to those that had completed all three sessions (total n = 90). Despite this larger sample size and increased statisti- cal power, this analysis produced identical findings to the model above.

The above analyses were also run as a three-way model in which hand times were treated as nested within each of the three sessions, which in turn were all nested within the same par- ticipants. This model yielded essentially the same result, except that the above gender effect emerged more strongly, F(1, 9) = 15.37, p = .003, and that it also revealed an effect for week of the semester of participants’ first session, F(11, 401) = 3.971, p < .001. Thus, the present find- ings do not support that training one’s non-dominant hand produces increased performance in either hand, i.e., increased performance in self-control.

Self-report measure of self-control. Self-report measures of self-control were entered into a two-level mixed model to account for three consecutive self-reports of self-control being nested within participants. We were interested in any differential change between the control and SCT conditions, whether malleability beliefs would moderate such changes over time, and if compliance with the instructions qualified these changes. The mixed model used the Tang- ney et al. [7] self-reported self-control scale as the dependent variable, Session (1–3), and Con- dition (training vs. control) as factors, and malleability beliefs and compliance as continuous predictors. Gender, week of semester of participants’ first session, and depletability beliefs were added as main effects only to control for their potential effects. Analyses were based only on those participants who had contributed data at all three sessions, leaving 82 participants. Depletability beliefs as a main effect approached significance, F(1, 63.00) = 2.62, p = .080; how- ever, no other effects emerged, all ps > .31. The same mixed model using the Grasmick et al. [34] measure as the dependent variable yielded no significant effects, all ps > .26. The present

PLOS ONE | https://doi.org/10.1371/journal.pone.0178814 June 8, 2017 10 / 14

Reliability of self-control training effects

findings do not support that training self-control through increased use of the non-dominant hand leads to increased self-reported self-control.

Stroop. A subset of participants (n = 38) had taken a Stroop task as a secondary objective measure of self-control. Similar to the other tasks, because of participant attrition, only 20 of these participants had completed all three sessions. One participant had extreme scores. Another participant was not a native English speaker, thus making the Stroop task (which was performed in English) difficult to interpret for that participant. These two participants were removed from subsequent analyses for these reasons. For participants for whom there were two sessions completed of the three, multiple imputation through use of Stata [44] was utilized to generate data for the missing data points for Session 3. The resultant number of participants for the subsequent analyses was 23.

We calculated Stroop scores by subtracting congruent times from incongruent times, as well as calculating an adjusted Stroop score by dividing the above Stroop score by control times to adjust for overall speed of responses. Both Stroop scores as well as adjusted Stroop scores were entered into two-level mixed models to account for three consecutive Stroop tasks being nested within participants. As in the previous models, we were interested in any differential change between experimental conditions, whether malleability beliefs would moderate such changes over time, and if compliance with the instructions qualified these changes. The mixed model included the Stroop scores and then the adjusted Stroop

scores as the dependent variable, Session (1–3) and Condition (training vs. control) as factors, and malleability beliefs and compliance as continuous predictors. Gender, week of participation, and depletability beliefs were controlled for (i.e., added as main effects only). No significant effects or interactions emerged with either model, all ps > .086. The findings do not support that training with the non-dominant hand leads to improved Stroop task performance. However, this result must be interpreted with caution because only a sub- group of our participants completed the Stroop task. This particular aspect adds to the pres- ent study’s pattern of non-supportive findings, albeit remains inconclusive, as it is likely underpowered.

Discussion

The present investigation yielded only null findings. On the one hand, our study provided behavioral evidence that participants in the SCT condition adhered to experimental instruc- tions. This was evident in self-reported compliance, which has been used in previous research [45], as well as the act of reporting their compliance online, i.e., number of compliance checks completed. This finding is surprising in light of much of the published evidence and appears consistent with the criticism leveraged against research on the effects of self-control training, which has been suspected to be compromised [14, 16].

Similarly, our attempt to examine whether subjective beliefs about the malleability of one’s self-control capacity moderate the expected effects ultimately failed. Although our novel scale revealed good psychometric qualities and showed satisfactory convergent and divergent valid- ity, malleability scores did not moderate training effects, as the latter did not materialize.

A possible criticism of the present study is that the SCT instructions might have drawn attention to the non-dominant hand, thus signaling to participants that experimenters were interested in the non-dominant hand, but not the dominant hand. However, there was no change in the performance of the dominant hand that would support the claim, there were no changes in Stroop performance, nor were there changes in self-reported self-control. Overall, the present research found little evidence that self-control training successfully improved indi- viduals’ physical measures of self-control or self-reported self-control. Whereas our findings

PLOS ONE | https://doi.org/10.1371/journal.pone.0178814 June 8, 2017 11 / 14

Reliability of self-control training effects

cannot refute that improving self-control is possible, the underlying processes are likely com- plex and difficult to translate into a simple training procedure.

Our study in the light of recent developments

Only after completion of the present research did we became aware of two recent meta-analy- ses, which were unpublished at the time of submission [28, 29]. Both analyses reported small but positive effects of training on self-control. Whereas this overlap is unfortunate, the inclu- sion of our study would have further reduced their already low effect size estimates. Con- versely, the fact that we looked toward the large effect size estimate published by Hagger et al. [10] for sample size requirements, yet to detect the type of effects emerging from the recent meta-analyses [28, 29], a larger sample size would have been required had our observed aver- age correlation between measurements been smaller. We believe that Hagger et al. [10] only synthesized the evidence available to them and cannot be blamed for their inflated estimate (but see [16]); however, this instance may serve as evidence that publication bias does have real consequences and can mislead members of the scientific community.

Author Contributions

Conceptualization: BML MK. Data curation: BML MK. Formal analysis: BML MK. Funding acquisition: BML MK. Investigation: BML MK. Methodology: BML MK. Project administration: BML MK. Resources: BML MK.

Software: BML MK. Supervision: MK. Validation: BML MK. Visualization: BML MK. Writing – original draft: BML MK. Writing – review & editing: BML MK.

References

1. Muraven M., Baumeister RF, Tice DM. Longitudinal improvement of self-regulation through practice: building self-control strength through repeated exercise. Journal of Social Psychology. 1999; 139: 446– 457. https://doi.org/10.1080/00224549909598404 PMID: 10457761

2. Baumeister RF, Vohs KD, Tice DM. The strength model of self-control. Current Directions in Psycholog- ical Science. 2007; 16: 351–355.

3. Conor BT, Stein JA, Longshore D. Examining self-control as a multi-dimensional predictor of crime and drug use in adolescents with criminal histories. Journal of Behavioral Health Services and Research. 2009; 36: 137–149. https://doi.org/10.1007/s11414-008-9121-7 PMID: 18575984

PLOS ONE | https://doi.org/10.1371/journal.pone.0178814 June 8, 2017 12 / 14

Reliability of self-control training effects

4. Gailliot MT, Baumeister RF. Self-regulation and sexual restraint: Dispositionally and temporarily poor self-regulatory abilities contribute to failures at restraining sexual behavior. Personality and Social Psy- chology Bulletin. 2007; 33: 173–186. https://doi.org/10.1177/0146167206293472 PMID: 17259579

5. Gottfredson MR, Hirschi T. A general theory of crime. Stanford, CA: Stanford University Press; 1990.

6. Muraven M, Collins RL, Nienhaus K. Self-control and alcohol restraint: An initial application of the self-

control strength model. Psychology of Addictive Behaviors. 2002; 16: 113–120. PMID: 12079249

7. Tangney JP, Baumeister RF, Boone AL. High self-control predicts good adjustment, less pathology,

better grades, and interpersonal success. Journal of Personality. 2004; 72: 271–324. PMID: 15016066

8. Oaten M, Cheng K. Improved self-control: The benefits of a regular program of academic study. Basic

Appl Soc Psych. 2006; 28: 1–16.

9. Oaten M, Cheng K. Longitudinal gains in self-regulation from regular physical exercise. British Journal of Health Psychology. 2006; 11: 717–733. https://doi.org/10.1348/135910706X96481 PMID: 17032494

10. Hagger MS, Wood C, Stiff C, Chatzisarantis NL. Ego depletion and the strength model of self-control: A meta-analysis. Psychological Bulletin. 2010; 136: 495–525. https://doi.org/10.1037/a0019486 PMID: 20565167

11. Hui SA, Wright RA, Stewart CC, Simmons A, Eaton B, Nolte RN. Performance, cardiovascular, and health behavior effects of an inhibitory strength training intervention. Motivation and Emotion. 2009; 33: 419–434.

12. Muraven M. Building self-control strength: Practicing self-control leads to improved self-control perfor- mance. Journal of Experimental Social Psychology. 2010; 46: 465–468. https://doi.org/10.1016/j.jesp. 2009.12.011 PMID: 20401323

13. Bertrams A, Schmeichel BJ. Improving self-control by practicing logical reasoning. Self and Identity. 2013; 13: 419–431.

14. Carter EC, McCullough ME. Publication bias and the limited strength model of self-control: Has the evi- dence for ego-depletion been overestimated?. Frontiers in Psychology. 2014; 5. https://doi.org/10. 3389/fpsyg.2014.00823

15. Rosenthal R. The “file drawer” problem and tolerance for null results. Psychological Bulletin. 1979; 86: 638–641.

16. Inzlicht M., Berkman E. Six questions for the resource model of control (and some answers). Social and Personality Psychology Compass. 2016; 9: 511–524.

17. Job V, Dweck CS, Walton GM. Ego depletion- Is it all in your head? Implicit theories about willpower affect self-regulation. Psychological Science. 2010; 21: 1686–1693. https://doi.org/10.1177/ 0956797610384745 PMID: 20876879

18. Job V, Walton GM, Bernecker K, Dweck CS. Beliefs about willpower determine the impact of glucose on self-control. Proceedings of the National Academy of Sciences. 2013; 110: 14837–14842

19. Finkel EJ, DeWall CN, Slotter EB, Oaten M, Foshee VA. Self-regulatory failure and intimate partner vio- lence perpetration. Journal of Personality and Social Psychology. 2009; 97: 483–499. https://doi.org/ 10.1037/a0015433 PMID: 19686003

20. Gailliot MT, Plant EA, Butz DA, Baumeister RF. Increasing self-regulatory strength can reduce the depleting effect of suppressing stereotypes. Personality and Social Psychology Bulletin. 2007; 33: 281–294. https://doi.org/10.1177/0146167206296101 PMID: 17259587

21. Hagger MS, Leaver E, Esser K, Leung CM, Pas NT, Keatley DA, et al. Cue-induced smoking urges deplete cigarette smokers’ self-control resources. Annals of Behavioral Medicine. 2013; 46: 394–400. https://doi.org/10.1007/s12160-013-9520-8 PMID: 23720186

22. Martijn C, Tenbu ̈lt P, Merckelbach H, Dreezens E, de Vries NK. Getting a grip on ourselves: Challenging expectancies about loss of energy after self-control. Social Cognition. 2002; 20: 441–460.

23. Muraven M, Tice DM, Baumeister RF. Self-control as limited resource: Regulatory depletion patterns. Journal of Personality and Social Psychology. 1998; 74: 774–789. PMID: 9523419

24. Bray SR, Ginis KA, Hicks AL, Woodgate J. Effects of self-regulatory strength depletion on muscular per- formance and EMG activation. Psychophysiology. 2008; 45: 337–343. https://doi.org/10.1111/j.1469- 8986.2007.00625.x PMID: 17995906

25. Rethlingshafer D. Relationship of tests of persistence to other measures of continuance of activities. Journal of Abnormal and Social Psychology. 1942; 37: 71–82.

26. DeWall CN, Baumeister RF, Vohs KD. Satiated with belongingness? Effects of acceptance, rejection, and task framing on self-regulatory performance. Journal of Personality and Social Psychology; 2008; 95: 1367–1382 https://doi.org/10.1037/a0012632 PMID: 19025289

PLOS ONE | https://doi.org/10.1371/journal.pone.0178814 June 8, 2017 13 / 14

Reliability of self-control training effects

27. Gino F, Schweitzer ME, Mead NL, Ariely D. Unable to resist temptation: How self-control depletion pro- motes unethical behavior. Organizational Behavior and Human Decision Processes. 2011; 115: 191– 203.

28. Friese M, Frankenbach J, Job V, Loschelder DD. Does self-control training improve self-control?: A meta-analysis. Perspectives on Psychological Science. In Press.

29. Beames J, Schofield TP, Denson TF. A Meta-analysis of Improving Self-control with Practice. In Hand- book of self-control in health and well-being. In Press.

30. Oaten M, Cheng K. Improvements in self-control from financial monitoring. Journal of Economic Psy- chology. 2007; 28: 487–501.

31. Rebro SM, Patterson RE, Kristal AR, Cheney CL. The effect of keeping food records on eating patterns. Journal of the American Dietetic Association. 1998; 98: 1163–1165. https://doi.org/10.1016/S0002- 8223(98)00269-7 PMID: 9787724

32. Kirkpatrick SI, Midthune D, Dodd KW, Potischman N, Subar AF, Thompson FE. Reactivity and its asso- ciation with body mass index across days on food checklists. Journal of the Academy of Nutrition and Dietetics. 2012; 112: 110–118. https://doi.org/10.1016/j.jada.2011.10.004 PMID: 22308230

33. Hufford MR, Shields AL, Shiffman S, Paty J, Balabanis M. (Reactivity to ecological momentary assess- ment: an example using undergraduate problem drinkers. Psychology of Addictive Behaviors. 2002; 16: 205. PMID: 12236455

34. Grasmick HG, Tittle CR, Bursik RJ Jr, Arneklev BJ. Testing the core empirical implications of Gottfred- son and Hirschi’s General Theory of Crime. Journal of Research in Crime and Delinquency. 1993; 30: 5–29.

35. Arneklev BJ, Grasmick HG, Bursik RJ Jr. Evaluating the dimensionality and invariance of” low self-con- trol”. Journal of Quantitative Criminology. 1999; 15: 307–331.

36. MacKenzie DL, Bierie D, Mitchell O. An experimental study of a therapeutic boot camp: Impact on impulses, attitudes and recidivism. Journal of Experimental Criminology. 2007; 3: 221–246.

37. Winfree LT, Taylor TJ, He N, Esbensen FA. Self-control and variability over time: Multivariate results using a 5-year, multisite panel of youths. Crime & Delinquency. 2006; 52: 253–286.

38. Burt CH, Simons RL, Simons LG. A longitudinal test of the effects of parenting and the stability of self- control: negative evidence for the general theory of crime. Criminology. 2006; 44: 353–396.

39. Harber KD, Zimbardo PG, Boyd JN. Participant self-selection biases as a function of individual differ- ences in time perspective. Basic and Applied Social Psychology. 2003; 25: 255–264.

40. Roman RJ, Moskowitz GB, Stein MI, Eisenberg RF. Individual differences in experiment participation: Structure, autonomy, and the time of the semester. Journal of Personality. 1995; 63: 113–138.

41. Sultan AJ, Joireman J, Sprott DE. Building consumer self-control: The effect of self-control exercises on impulse buying urges. Marketing Letters: A Journal of Research in Marketing. 2012; 23: 61–72.

42. Job V, Friese M, Bernecker K. Effects of practicing self-control on academic performance. Motivation Science. 2015; 1: 219–232.

43. Denson TF, Capper MM, Oaten M, Friese M, Schofield TP. Self-control training decreases aggression in response to provocation in aggressive individuals. Journal of Research in Personality. 2011; 45: 252–256.

44. StataCorp. Stata Statistical Software. Release 13. College Station, TX: StataCorp LP, 2013.

45. Denson TF, Pedersen WC, Friese M, Hahm A, Roberts L. Understanding impulsive aggression: Angry rumination and reduced self-control capacity are mechanisms underlying the provocation-aggression relationship. Personality and Social Psychology Bulletin. 2011; 37: 850–862. https://doi.org/10.1177/ 0146167211401420 PMID: 21421767

AB #8

The Effect of Gradual Self-Control of Task Difficulty and Feedback on Learning Golf Putting

Mojtaba Jalalvand, Abbas Bahram, Afkham Daneshfar & Saeed Arsham

To cite this article: Mojtaba Jalalvand, Abbas Bahram, Afkham Daneshfar & Saeed Arsham (2019) The Effect of Gradual Self-Control of Task Difficulty and Feedback on Learning Golf Putting, Research Quarterly for Exercise and Sport, 90:4, 429-439, DOI: 10.1080/02701367.2019.1612510

To link to this article: https://doi.org/10.1080/02701367.2019.1612510

Published online: 22 Jul 2019.

Submit your article to this journal Article views: 291

View related articles View Crossmark data

Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=urqe20

RESEARCH QUARTERLY FOR EXERCISE AND SPORT 2019, VOL. 90, NO. 4, 429–439 https://doi.org/10.1080/02701367.2019.1612510

The Effect of Gradual Self-Control of Task Difficulty and Feedback on Learning Golf Putting

Mojtaba Jalalvanda, Abbas Bahrama, Afkham Daneshfarb, and Saeed Arshama aKharazmi University; bAlzahra University

ABSTRACT

Purpose: This study aimed to examine the effect of gradual self-control of task difficulty and feedback on accuracy and movement pattern of the golf putting sport skill. Method: Sixty students were quasi-randomly assigned to four groups under a varying combination of the two factors of task difficulty control (self-controlled or yoked) and feedback control (self-controlled or yoked). The participants in the two groups (dual-factor gradual self-control and self-controlled task difficulty group) that granted control over task difficulty in the acquisition phase were told they could choose any of the pre-set distances from the target. All groups were given 100% feedback in the first half of the acquisition phase, but the participants in the two groups (dual- factor gradual self-control group and self-controlled feedback) that could control their feedback were told that in the second half of the acquisition phase they would be able to ask for feedback when needed. The practice schedule of each member of the dual-factor gradual self-control group was used as a basis to plan the practice of predetermined distances and feedback presentation to the corresponding participants in the yoked conditions. Results: ANOVA with repeated measures showed that the practice method involving gradual self-control of two factors had a positive impact on accuracy and movement pattern of golf putting in the retention and transfer tests compared to other methods (ps < .05). Conclusion: The advantages of self-control practice presumably come from better adjustment of challenge points by the learner in the course of the practice.

Self-control is one of the variables that has to be taken into consideration in organizing practice sessions. Practice conditions in which participants allow for the control of certain aspects of the practice, in motor learning literature, are typically termed self-controlled practice (Wulf, Clauss, Shea, & Whitacre, 2001). Moreover, motor learning depends on variables such as feedback, which refers to sensory information about the performance of the skill received by the individual during or after the performance (Schmidt & Lee, 2013). Feedback examined in the context of motor learning research usually involves information about the out- come (termed “knowledge of results” [KR]) or the quality of the movement (termed “knowledge of per- formance” [KP]). The latter corresponds more to the feedback given by an instructor (Wulf, Shea, & Lewthwaite, 2010). There are many methods for pre- senting feedback. One is to offer feedback on a learner’s request. Janelle, Kim, and Singer (1995) in one of the first attempts at studying the effects of self-control on learning a motor task found that participants who could decide when to receive feedback had a better

result compared to the control and yoked groups. Following this research, self-control manipulation was carried with regard to different types of instructional support (Post, Fairbrother, & Barros, 2011). Task diffi- culty is one of the instructional supports that can be controlled by the learner during practice. Self-control of task difficulty and its benefits have been studied and observed in different laboratory tasks such as ski simu- lation in two difficulty levels (Wulf et al., 2001; Wulf & Toole, 1999), coincidence-anticipation tasks in three difficulty levels (Andrieux, Boutin, & Thon, 2016; Andrieux, Danna, & Thon, 2012), and balancing tasks in two task difficulty levels (Hartman, 2007). While studies on self-control of task difficulty (Andrieux et al., 2016, 2012; Hartman, 2007; Wulf et al., 2001; Wulf & Toole, 1999) have mostly been done using laboratory tasks, there is an interest in applying the results to research on sports skills.

Studies on self-control of task difficulty (Andrieux et al., 2016, 2012; Hartman, 2007; Wulf & Toole, 1999) are replete with the usage of performance outcome measures to track the effect of this variable. But few

ARTICLE HISTORY

Received 25 December 2017 Accepted 23 April 2019

KEYWORDS

Challenge point; knowledge of performance; motor learning; self-regulation

CONTACT Mojtaba Jalalvand Jalalvand_mojtaba@yahoo.com Department of Physical Education and Sport Science, Kharazmi University, Tehran, Iran. Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/urqe. © 2019 SHAPE America

430 M. JALALVAND ET AL.

studies (Wulf et al., 2001) have employed performance production measures to survey the effects of self- controlled task difficulty. A wide range of methods are used to measure the performance of motor skills. The criteria for measuring motor skills are divided into two categories. One, criteria for measuring perfor- mance outcome, which indicate the outcome or conse- quence of the skill performance. These criteria do not provide the researcher with any information as to the limb or body behavior that produces the outcome. To learn about these features, researchers use other criteria known as a performance production measure (Magill & Anderson, 2014). Therefore, the study of the effects of self-controlled task difficulty on the movement pattern of a complex sport skill can provide useful information about the effect of self-controlled practice.

The existing body of research on self-control also seems to suffer from another limitation, in that in most such studies, the learners have controlled only one aspect of the practice (Chiviacowsky & Wulf, 2002, 2005; Grand et al., 2015; Janelle, Barba, Frehlich, Tennant, & Cauraugh, 1997; Janelle et al., 1995; Post et al., 2011), or even if they have been able to control multiple aspects of the practice, such control has been simultaneous at the beginning of the practice (Brydges, Carnahan, Rose, & Dubrowski, 2010; Hodges, Edwards, Luttin, & Bowcock, 2011; Jowett, LeBlanc, Xeroulis, MacRae, & Dubrowski, 2007). But research in which the learner has gradual and increasing control over different aspects of the practice seems to be missing. Early into the skill acquisition phase, the cognitive effort to individualize a practice context is expected to interact less than optimally with cognitive demands required to perform the task (Patterson, Carter, & Sanli, 2011). Therefore, if the learner’s partici- pation in determining different aspects of the practice is gradually increased, a better interaction can be expected between cognitive effort for individualizing practice con- text and cognitive demands for performing the task in the skill acquisition period, leading to increased learning.

In general, it is assumed that effectiveness of self- controlled practice, compared to externally imposed practice conditions, is due to more active involvement of the learner, enhanced motivation, and increased effort invested in the practice (Wulf et al., 2010). In recent studies on self-controlled practice (Andrieux et al., 2016, 2012; Hemayattalab, 2014; Patterson et al., 2011; Post et al., 2011), the challenge point framework (CPF) has been used to explain the benefits of this variable. Based on this framework, the effectiveness of practice conditions depends on a number of factors, including task difficulty and individual characteristics (e.g., skill level). In CPF, information is seen as a challenge for the performer and there are certain

limits to using the potential available information. These limitations are due to the individual’s informa- tion-processing capacity. As the task’s functional diffi- culty increases to an optimal challenge point, so does the information load that provides a potential learning benefit. But from that point onward, the increased difficulty results in diminished learning due to infor- mation volume exceeding the individual’s processing capacity (Guadagnoli & Lee, 2004). There are no well- known ways to adjust practice conditions such that functional difficulty of the task would align with opti- mal challenge point. Even if the examiner (instructor) can identify the gap between a task’s functional diffi- culty and optimal challenge point during the practice, there still is no known efficient way to adjust the practice conditions such that functional difficulty of the task can match the optimal challenge point (Akizuki & Ohashi, 2015). However, evidence from self-control research suggests that the learners them- selves can determine the optimal challenge point dur- ing the practice (Andrieux et al., 2016, 2012; Keetch & Lee, 2007; Patterson et al., 2011). For example, Andrieux et al. (2012) demonstrated that the learner in each trial can determine the optimal level of task difficulty based on ability to perform the task.

Based on CPF, the amount of learning in every trial is determined by the amount of information processing before (response planning), during (internal feedback), and after (external feedback) the performance (Magill, 2011). Patterson, Carter, and Hansen (2013) enhanced learning through challenging the motor planning and response interpretation phases by combining the two variables of experimenter-controlled task practice sche- dule (random and blocked schedule) and self- controlled feedback. The question is how learning would be affected if both phases of motor planning and response interpretation were to be challenged by a self-controlled variable. Findings of Andrieux et al. (2012) show that learners can decide the difficulty level of the task before response initiation, based on the optimal amount of interpretable information during the practice. Also, Grand et al. (2015) found that self- control of feedback enhances its processing (more negative FRN1 amplitudes). Therefore, based on CPF, it is predicted that the application of a self-controlled variable should increase the information processing before and after performing the trial and, consequently, learning should increase as well. Given the above, per- haps it can be assumed that the benefits of self- controlled task difficulty in a sport skill such as golf putting are due to which allow for a higher freedom in

1Electroencephalography-derived feedback-related negativity (FRN).

determining the difficulty level in proportion to the rise in skill level. Existing laboratory research (Andrieux et al., 2016, 2012; Hartman, 2007; Wulf et al., 2001; Wulf & Toole, 1999) offers learners limited options from two or three levels of task difficulty. Moreover, feedback control after performance for receiving infor- mation in accordance with information-processing capacity would enable the learner to adjust the chal- lenge point in the course of the practice. Meanwhile, since a learner’s control over a part of the learning environment is exceptionally effortful (Kanfer & Ackerman, 1989), the gradual self-control method based on CPF during the acquisition phase creates new challenges that will likely result in increased learning.

The current study has been designed and conducted to expand knowledge and examine applicability of the effect of gradual self-control of the two factors of task difficulty and feedback on learning a sport task (golf putting). Also the study has strived to improve on the limitations of the previous studies, which had provided the learner with limited options (two or three task difficulty levels), by offering more options (eight task difficulty levels) so that the learners could have more freedom to decide task difficulty level in proportion to skill development. Moreover, the effect of self-control practice was evaluated by performance outcome mea- sures as well as the performance production measure.

Method

Participants

Sixty undergraduate students (28 girls, 32 boys) of physical education (average age = 20.92 years, standard

deviation = 1.59) from the university voluntarily took part in the research. The participants were quasi- randomly (equal number based on sex, dominant hand, and eye) were assigned to four groups (each group consisting of seven girls and eight boys). The participants did not have any prior experience in prac- ticing golf and mini golf. They were unaware of the research objectives. A signed informed consent was obtained from each participant. This research was approved by Ethics Committee of the university.

Apparatus and task

The current study employed similar tasks as used by Maxwell, Masters, Kerr, and Weedon (2001) and Poolton, Masters, and Maxwell (2005), which involved golf putting at distances of 25, 50, 70, 100, 125, 150, 175, 200 cm from a U-shaped device with a diameter of 11.50 cm (instead of a hole) (see Figure 1). Each dis- tance was marked on the carpet with a thickness of approximately 15 mm by tape. The zero end of a measuring tape was fixed to the center of the U-shaped device. For measuring the changes in perfor- mance in the acquisition and test phases, the radial error of each trial was recorded. The radial error was calculated such that scoring the ball would result in zero errors and, otherwise, the distance between the center of the ball and the center of the U-shaped device would be considered as the performance error. Putter (Iron No. 6) and standard white golf balls were used in the practice and test sessions.

For measuring the mental difficulty of learning golf putting in different practice methods employed in this research, the specific single-question questionnaire for

RESEARCH QUARTERLY FOR EXERCISE AND SPORT 431

12345678

Figure 1. The U-shaped device.

11.50 cm

20 cm

432 M. JALALVAND ET AL.

learning motor tasks as suggested by Sweller (personal communication, November 11, 2016) was used for the first time, which included nine difficulty levels (the question was: How much learning golf putting was mentally difficult for you? Answer by choosing one of the options below: very very low, very low, low, rather low, neither high nor low, rather high, high, very high, very very high). The questionnaire’s reliability was cal- culated at 0.822 using the test–retest method in a pilot study the next day. The results of the questionnaire were used for discussion and reaching a better conclu- sion as to the challenge point framework.

For producing sounds used in the dual-task transfer test, similar to Poolton et al. (2005), a computer was programmed to produce randomly sequenced high and low pitched tones with a one-second frequency. The secondary task required the participants to count and report only the number of high-pitched tones from the moment of the beginning of the golf putt to the end.

To evaluate the golf putting movement pattern dur- ing the acquisition phase (last block), retention, and transfer tests, a W810 Sony camera was used with 20.1-megapixel resolution and shutter speed of 30 frames per second. In accordance with Porter, Landin, Hebert, and Baum (2007), filming took place from the front of participants. The golf putting movement pat- tern was analyzed based on three factors of pendulum movement, length of swing, and head stability (using the participant’s performance footage). The golf putting movement pattern of the each trial can be obtained between zero and four scores. The specific scoring procedures for each of these parameters are provided in Table 1. This tool was devised based on the criteria of correct performance as seen in skilled golfers and has been reported to have a reliability of above 0.8.

Procedure

The current experiment, similar to the study by Maxwell et al. (2001), consists of two phases of acquisi- tion and test. The acquisition phase included the per- formance of eight blocks of 10 trials (80 trials in total)

Table 1. Scoring criteria for the movement pattern of the golf putt with permission from Porter et al. (2007).

Pendulum: 0 = all wrist, no pendulum motion of hands, arms, shoulders 1 = some evidence of pendulum, but wrist break 2 = all pendulum motion Swing length: 0 = beyond restraining line 1 = with restraining line Head motion: 0 = head lifted at, or prior to, moment of club/ball contact 1 = definite in head lift

in the first day. The participants were put in four experiment groups based on a different combination of the two factors of task difficulty control (self- controlled or yoked) and feedback control (self- controlled or yoked). These groups were as follows: (a) gradual self-control of the two factors of task diffi- culty and feedback (dual-factor gradual self-control group), (b) self-controlled task difficulty and yoked feedback (self-controlled task difficulty group), (c) yoked task difficulty and self-controlled feedback (self- controlled feedback group), and (d) yoked task diffi- culty and feedback (dual-factor yoked group).

In the acquisition phase, day one, the participants of the two groups that allowed for the control of the task’s difficulty (dual-factor gradual self-control and self- controlled task difficulty groups) were told that before each trial, they could choose any of the preset distances (25, 50, 75, 100,125, 150, 175, 200 cm). The dual-factor gradual self-control group was the first to do their practice trials. The practice schedule of each member of the dual-factor gradual self-control group was recorded by the researcher and used as a basis for planning the practice distances for the participants in the two corresponding groups that had yoked task difficulty practice (self-controlled feedback and dual- factor yoked groups). Based on Andrieux et al. (2016) and Andrieux et al. (2012), the members of the dual- factor gradual self-control and self-control task diffi- culty groups were told that the retention test would be conducted from 200 cm distance, therefore the aim would be to learn to perform from the 200 cm distance. Also, the members of the two groups yoked in task difficulty factor were told “the retention test will be from 200 cm distance.”

The acquisition phase in terms of KP, based on Andrieux et al. (2016) and Patterson et al. (2011), was divided into two halves. In the first half (1 to 40 trials), learners in all groups were given 100% KP. But the second half of the acquisition phase (41 to 80 trials) had a different KP control schedule. An instructor, an accomplished golfer, provided verbal feedback correct- ing major errors related to the stance, grips, arm actions, and head stability. Based on Patterson et al. (2011), the participants of the two groups capable of controlling KP (dual-factor gradual self-control and self-controlled feedback) were told in the second half of the acquisition phase that they could ask for feed- back on movement pattern whenever needed, and that final tests would be conducted without feedback (Hemayattalab, 2014). The feedback schedule for each member of the dual-factor gradual self-control group was registered by the researcher and used as a basis for scheduling feedback for participants in the two groups

with yoked feedback (dual-factor yoked and self- controlled task difficulty groups). Based on Patterson et al. (2011), the participants of these groups were told that in the second half of the acquisition phase, a random KP would be assigned to them.

The mean scores for movement pattern in the final block of the acquisition phase, which included perfor- mance from 200 cm distance, were used as the move- ment pattern data for the acquisition phase. The participants in all groups completed the mental diffi- culty questionnaire and the verbal protocol at the end of the acquisition phase.

According to Maxwell et al. (2001) one day after the acquisition phase, the retention test, single-task transfer test from 300 cm distance, and dual-task transfer test from 200 cm (each consisting of a 10-trial block) were carried out. For controlling the order bias of transfer tests, half of the participants chosen at random per- formed the single-task transfer test and then the dual- task transfer test. For the rest of the participants, this procedure was applied in reverse.

Data analysis

For each participant, the mean (and standard deviation) of the radial error (the dependent variable of the perfor- mance outcome) of eight blocks of the acquisition phase and learning tests was calculated and recorded as the performance of those phases. To evaluate changes in radial error from beginning to end of the practice, a mixed design analysis of variance 2 (task difficulty control: self-controlled, yoked) × 2 (feedback control: self- controlled, yoked) × 8 (blocks of 10 trials) with repeated measurement on the last factor was used. In this part of the study, after the block effect became significant, planned contrasts were used in a polynomial method for multiple comparisons. To analyze the radial error in the testing phase, a mixed variance analysis design 2 (task difficulty control: self-controlled, yoked) × 2 (feedback control: self-controlled, yoked) × 3 (test: retention, single- task transfer, dual-task transfer) with repeated measures on the last factor was used.

For each participant, the mean (and standard devia- tion) of the movement pattern score (dependent vari- ables of performance production) at the last block of the acquisition phase and the learning tests were calcu- lated and recorded as its movement pattern. The move- ment pattern in the acquisition phase was analyzed using two-way ANOVA 2 (task difficulty control: self- controlled, yoked) × 2 (feedback control: self- controlled, yoked). For movement pattern analysis in the testing phase, a mixed design ANOVA 2 (task difficulty control: self-controlled, yoked) × 2 (feedback

control: self-controlled, yoked) × 3 (test: retention, single-task transfer, dual-task transfer) with repeated measures on the last factor was used. Other dependent variables including mental difficulty (questionnaire), the number of recalled rules (verbal protocol), and counting tone (secondary task) were each analyzed using two-way ANOVA 2 (task difficulty control: self- controlled, yoked) × 2 (feedback control: self- controlled, yoked). In these sections, wherever needed, planned contrasts were used with a simple method to perform multiple comparisons. The statistical signifi- cance level was considered to be p ≤ .05.

Results

Choosing task difficulty and requesting KP

Figure 2 shows the method of choosing the task diffi- culty level (schedule of distance practice) in dual-factor gradual self-control and self-controlled task difficulty groups. Learners in these two groups had somewhat different practice schedules. About 87% of the partici- pants in these two groups chose the shortest distance (25 cm) to the target and gradually increased the dis- tance. The average distance to the target in the fourth block of the acquisition phase for the two groups (175 cm) was almost equal to the distance used during retention test (200 cm). About 73% of the participants in the dual-factor gradual self-control group asked for 200 cm distance (most difficult level) from the start of the fifth block. Therefore, a new challenge (feedback self- control) seemed necessary from fifth block onward. The frequency (percent) with which the participants in the dual-factor gradual self-control group and self-controlled feedback group requested feedback is shown in right- hand side of Figure 2. The frequency of asking for KP in these two groups had a gradual decline.

Radial error

Acquisition phase: The left side of Figure 3 shows the mean radial error of practice blocks in the acquisition phase for different groups. In the analysis of variance, the Mauchly’s test of sphericity showed that the sphericity assumption is not valid for the main effect of the block, x2 (27) = 68.27, p < .001. Therefore, the degrees of freedom were corrected using the Greenhouse-Geisser estimates of sphericity (ε = .736). Based on the results of ANOVA, the main effect of the block was significant, F(5.15, 288.54) = 12.15, p < .001, ɳp2 = .178. All other main effects and interactions were not significant (ps > .05).The planned contrasts with polynomial showed that during the practice blocks of the acquisition phase, the change in

RESEARCH QUARTERLY FOR EXERCISE AND SPORT 433

434 M. JALALVAND ET AL.

200 175 150 125 100

75 50 25

Dual-factor gradual self-control

Self-controlled task difficulty

Self-controlled feedback

Block 1

Block 2

Block 3

Block 4

Block 5

Block 6

Block 7

Block 8

Figure 2. Mean task difficulty level (distance to U-shaped Device) asked by the participants in the dual-factor gradual self-control and self-controlled task difficulty groups. Mean frequency (percent) feedback requested by the participants in the dual-factor gradual self-control group and self-controlled feedback group (right side).

Dual-factor gradual self-control

Self-controlled task difficulty

Self-controlled feedback

Dual-factor yoked

Block 1

Block 2

Block 3

Block 4

Block 5

Block 6

Block 7

Block 8

Retention Singl-task Dual-task transrer transfer

Figure 3. Mean radial error (cm) as a function of the trial block during acquisition session and tests.

the mean error of the groups was quadratic, F(1, 56) = 32.94, p < .001, ɳp2 = .370. On the whole, in can be concluded that radial errors across different groups in the acquisition phase did not vary significantly.

Test phase (retention and transfer): The mean radial error during the test phase for different groups is presented on the right side of Figure 3. According to the results of ANOVA, the main effect of task difficulty control, F(1, 56) = 45.55, p < .001, ɳp2 = .449, the main effect of feedback control, F(1, 56) = 6.64, p = .013, ɳp2 = .106, the main effect of the test, F(2, 112) = 30.38, p < .001, ɳp2 = .352, and interaction effect of test × task difficulty control, F(2, 112) = 4.10,p = .019, ɳp2 = .068, were significant. All other interactions were not significant (ps > .05). Planned contrasts with a simple method showed that the radial error of the dual-factor gradual self-control group in the

retention test was significantly lower compared to the three groups of self-controlled task difficulty, t(56) = 2.77, p = .008, r2 = .12, self-controlled feedback, t(56) = 4.28, p < .001, r2 = .25, and dual-factor yoked, t (56) = 5.33, p < .001, r2 = .34. According to the results, in the single-task transfer test, the radial error of the dual-factor gradual self- control group did not significantly vary from that of self- controlled task difficulty group, t(56) = 1.70, p = .095, r2 = .05. But it was significantly lower than that of self- controlledfeedback group, t(56) = 4.74, p < .001, r2 = .29, and the dual-factor yoked group, t(56) = 5.77, p < .001, r2 = .37. The results also showed that in the dual-task transfer test, the radial error of the dual-factor gradual self- control group did not vary significantly from that of self- controlled task difficulty group, t(56) = 1.61, p = .113, r2 = .04, but it was significantly more precise compared to

Radial Error (cm)

Chosen distance to target (cm)

Frequency (percent) equested KP

the self-controlled feedback, t(56) = 3.83, p < .001, r2 = .17, and dual-factor yoked groups, t(56) = 4.44, p < .001, r2 = .26.

Movement pattern

Acquisition phase: The mean movement pattern scores of the golf putting skill across groups in the last block of the acquisition phase are presented on the left side of Figure 4. Based on the results of ANOVA, the main effect of task difficulty control, F(1, 56) = 7.93, p = .007, ɳp2 = .124, and the main effect of feedback control, F(1, 56) = 7.17, p = .010, ɳp2 = .113, were significant. But interaction effect of task difficulty control × feedback control, F(1, 56) = .06, p = .809, ɳp2 = .001, was not significant. Planned contrasts with the simple method showed that in the acquisition phase the golf putting movement pattern was significantly better in the dual-factor gradual self-control group com- pared with the dual-factor yoked group, t(56) = 4.05, p = .001, r2 = .23, but did not significantly differ from self- controlled task difficulty group, t(56) = 1.45, p = .158, r2 = .04, and self-controlled feedback group, t(56) = 1.76, p = .091, r2 = .05.

Test phase (retention and transfer): The mean move- ment pattern scores of golf putting in the retention, single- task transfer, and dual-task transfer tests are presented on the right side of Figure 4. In the analysis of variance, the Mauchly’s test of sphericity showed that sphericity assump- tion does not hold for the main effect of the test, x2(2) = 8.06, p = .018. Therefore, the degrees of freedom were corrected using the Huynh-Feldt estimates of spheri- city (ε = .822). Based on the results of ANOVA, the main effect of the task difficulty control, F(1, 56) = 12.77, p = .001, ɳp2 = .186, the main effect of feedback control, F(1, 56) = 10.04, p = .002, ɳp2 = .152, the main effect of test, F(1.91, 106.97) = 5.08, p = .009, ɳp2 = .083, and the interaction effect of test × feedback control, F(1.91, 106.97) = 3.76, p = .028, ɳp2 = .063, were significant. All other interactions were not significant (ps > .05). Planned contrasts with a simple method showed that in retention test, the golf putting movement pattern was not

significantly better in the dual-factor gradual self-control group compared to the self-controlled task difficulty, t(56) = 1.60, p = .122, r2 = .04, and self-controlled feedback groups, t(56) = 2.05, p = .053, r2 = .07, but was significantly better than the dual-factor yoked group, t(56) = 4.23, p < .001, r2 = .24. According to the results, in the single- task transfer test, the movement pattern of the dual-factor gradual self-control group did not significantly vary from that of the self-controlled task difficulty, t(56) = 1.54, p = .134, r2 = .04, and self-controlled feedback groups, t(56) = 1.85, p = .079, r2 = .06, but it was significantly better compared to the dual-factor yoked group, t(56) = 5.28, p < .001, r2 = .32. The results also showed that in the dual- task transfer test, the movement pattern was not signifi- cantly different in the dual-factor gradual self-control group compared to the self-controlled task difficulty group, t(56) = 1.68, p = .099, r2 = .05, but it was significantly better than that of the self-controlled feedback group, t(56) = 2.36, p = .022, r2 = .08, and dual-factor yoked group, t(56) = 3.81, p < .001, r2 = .19.

Mental difficulty questionnaire

The mental difficulty of learning golf putting in dif- ferent groups is presented in Table 2. Based on the results of ANOVA, the main effect of the task diffi- culty control, F(1, 56) = 16.21, p < .001, ɳp2 = .224, and the main effect of feedback control, F(1, 56) = 9.70, p = .003, ɳp2 = .148, were significant, but the interaction effect of the task difficulty control × feed- back control, F(1, 56) = .006, p = .94, ɳp2 = .000, was not significant. The results of planned contrasts with the simple method indicated that the mental diffi- culty of learning golf putting by the method of gra- dual self-control of two factors was significantly higher than the three groups of self-controlled task difficulty, t(56) = 2.15, p = .036, r2 = .07, self- controlled feedback, t(56) = 2.79, p = .007, r2 = .10, and dual-factor yoked, t(56) = 5.05, p < .001, r2 = .28.

4 3.5 3 2.5 2 1.5 1 0.5 0

Dual-factor gradual self-control

Self-controlled task difficulty

Self-controlled feedback

Dual-factor yoked

RESEARCH QUARTERLY FOR EXERCISE AND SPORT 435

Figure 4. Mean movement pattern scores in the last block acquisition and tests.

Block 8

Retention

Singl-task transfer

Dual-task transfer

Mean movement pattern scores

436 M. JALALVAND ET AL.

Table 2. Results of the questionnaire and verbal protocol com- pleted at the end of acquisition phase (means and standard deviations).

cognitive load necessary to individualize practice context should come into better interaction with cognitive demands for performing the task in the initial phases of acquisition (Andrieux et al., 2016; Patterson et al., 2011). To achieve this goal, the learner was granted control over task difficulty from the beginning, and feedback control from the onset of the second half of the acquisition phase. The participants were quasi-randomly assigned to four groups under a varying combination of the two factors of task difficulty control (self-controlled or yoked) and feed- back control (self-controlled or yoked). One day after 80 trials in the acquisition phase, the learners took part in retention, single transfer, and dual-task transfer tests.

The study’s findings show that, in the acquisition phase, the groups (practice methods) did not differ in precision of putting (radial error); however, the practice method of dual-factor gradual self-control prevailed over other practice methods in terms of precision in the retention test. Furthermore, the group’s practice method in transfer tests resulted in more precise put- ting compared to the self-controlled feedback and dual- factor yoked groups. Also the study’s results indicate that the dual-factor gradual self-control practice method yields better movement patterns for golf put- ting compared to the dual-factor yoked practice method in the acquisition, retention, and transfer phases. Nonetheless, the advantages of the dual-factor gradual self-control practice method in improving golf putting movement patterns over the feedback control method could be observed only in the dual-task transfer test. These results are consistent with the findings of Andrieux et al. (2016), Andrieux et al. (2012), Fairbrother, Laughlin, and Nguyen (2012), Hartman (2007), Hemayattalab (2014), Lessa and Chiviacowsky (2015), and Wulf, Chiviacowsky, and Drews (2015).

The results of this study are consistent with the pre- dictions of the challenge point framework (Guadagnoli & Lee, 2004). According to CPF, the amount of learning in each trial is determined by the amount of information processing before (response planning), during (internal feedback), and after (external feedback) the performance (Patterson et al., 2013). It seems in this study, the added cognitive load of the self-control of the two factors has increased information processing in the motor planning and action interpretation phases. The results of verbal protocol implementation showed that in remembering the rules and principles of golf putting, the learners of dual-factor gradual self-control practice had better per- formances over only the dual-factor yoked group. Nevertheless, their higher scores in verbal protocol com- pared to other groups can indicate a higher involvement level in the learning process. Therefore, it can be assumed that the benefits of the gradual dual-factor self-control

Groups

Dual-factor gradual self-control Self-controlled task difficulty Self-controlled feedback Dual-factor yoked

Verbal protocol

Questionnaire

M SD

6.1 1.9 4.8 1.4 4.4 2.1 3.0 1.3

Verbal Protocol

M SD

3.5 2.0 3.7 1.1 2.9 1.5 2.3 1.1

The number of rules and principles remembered by lear- ners in performing golf putting in different groups is presented in Table 2. Based on the results of ANOVA, the main effect of task difficulty control, F(1, 56) = 7.39, p = .009, ɳp2 = .117, was significant. Other main effects and interactions were not significant (ps > .05). Planned con- trasts with the simple method showed that the dual-factor gradual self-control group was significantly higher than the dual-factor yoked group, t(56) = 2.23, p = .030, r2 = .07, in terms of the number of rules remembered, but it had no significant difference compared to the self-controlled task difficulty group, t(56) = .50, p = .622, r2 < .001, and self- controlled feedback group, t(56) = 1.12, p = .269, r2 = .02.

Secondary task (tone counting)

Based on the results of ANOVA, main effects and interactions were not significant (ps > .05). Mean accu- racy percentage was 91.5%, 93.3%, 93.8%, and 94.1% for the dual-factor gradual self-control group, self- controlled task difficulty, self-controlled feedback, and dual-factor yoked groups, respectively. The analysis suggests participants in the different groups were equally involved in performance of the secondary task.

Discussion

This study aimed to determine the effect of gradual self- control of the two factors of task difficulty and feedback on accuracy and movement pattern of golf putting. In previous self-control studies (Chiviacowsky & Wulf, 2002, 2005; Grand et al., 2015; Janelle et al., 1997, 1995; Post et al., 2011), the learners have mainly had control over only one aspect of the practice or, even if they could control multiple aspects, such control has been simulta- neous (Brydges et al., 2010; Hodges et al., 2011; Jowett et al., 2007). But to our knowledge, there is no previous research that presents learners with gradual and increas- ing control over different aspects of the practice. As the learner gradually becomes more engaged in determining different aspects of the practice, it is expected that the

method are attributable to a higher level of information processing related to performing the task. Meanwhile, according to CPF, the practice context, which is used to challenge the learner’s information processing capability, is an important factor to facilitate the learning of motor skills (Guadagnoli & Lee, 2004). The questionnaire’s results support this CPF prediction. The results show that the dual-factor gradual self-control practice method poses a higher level of mental difficulty in learning golf putting compared to other methods. This is further con- firmed by the fact that putting precision was not different in varying practice methods in the acquisition phase. Perhaps the challenging nature of the dual-factor gradual self-control practice method has acted as a negative per- formance variable to temporarily lower performance level in the acquisition phase, resulting in no difference preci- sion across varying methods. But since this challenge was removed in the retention test for the learners of this method, the conditions became equal for all and the advantages of the method emerged. This is consistent with the findings of Patterson et al. (2013) who had challenged the motor planning and response interpreta- tion phases by combining the two factors of experimen- ter-controlled practice schedule (random and blocked schedule) and feedback self-control. This study gave more freedom to learners to choose from eight levels of task difficulty (in contrast with previous studies that pro- vided two or three options), resulting in better compat- ibility between task difficulty level and individual skill. The study showed that the participants in the dual-factor gradual self-control method gradually requested more difficult levels in the acquisition phase. Most participants in the dual-factor gradual self-control group asked for the highest level of task difficulty at the beginning of the second half of the acquisition phase. It seems further along the practice; a new challenge was created by giving self-control to feedback. In the second half of the acquisi- tion phase, the learners gradually decreased frequency of feedback requests (see the right side of Figure 2). Therefore, the advantages of gradual self-control practice presumably come from its challenging nature.

This study expands on the findings of Patterson et al. (2013) and further suggests that the gradual self-control method provides probably an optimal challenge for learner’s cognitive processes before and after skill per- formance, resulting in a facilitated learning process. According to Schmidt’s schema theory (1975), a series of cognitive events are related to motor planning (recall schema) and response performance result interpreta- tion (recognition schema), and are chronologically separate and consecutive. Self-controlled task difficulty perhaps affects cognitive events related to parameteriz- ing the movement and self-controlled feedback control

affects cognitive events related to motor planning; therefore, it is not unlikely that these two variables have different effects on learning. Since the task used in this research (golf putting) did not have a very complex movement pattern, it is possible that the lear- ner would not need to depend much on KP for inter- preting movement pattern and skill performance results and that self-controlled feedback would not optimally affect postperformance cognitive processes.The low fre- quency of feedback requests supports this claim (see Figure 2). This might explain the better movement patterns and higher precision of learners in the gradual dual-factor self-control group compared to the feed- back self-control method and dual-factor yoked method groups in the transfer phase. Even though the task used in this research is ecologically valid, research- ers are advised to use tasks with more complex move- ment patterns in future studies so that the learners will need KP for performance result interpretation and movement pattern. Alternatively, instead of KP, they could use KR in a task whose performance requires knowledge of the performance outcome.

In general, this study shows that the preplanned schedule for task difficulty control and feedback reduces the participants’ capability to learn optimally. Furthermore, the study supports the idea that gradual self-control provides the learner with an opportunity to improve processing of information related to task per- formance and organize the learning environment as needed for the performance. Given the fact that with an increased sense of autonomy and responsibility the learner will produce more mental effort during practice sessions, it is suggested that sports coaches and physical education teachers should shift from trainer-centered practice sessions to learner-centered sessions based on an ample knowledge about the individual, the task, and practice conditions. These features help the learner to rely on his or her personal effort and cognitive process to have a better learning experience.

What does this article add?

This study showed that the practice method involving learners with gradual and increasing control over differ- ent aspects of the practice had a positive impact on learn- ing the sport skill. The gradual self-control practice is expected that the cognitive load necessary to individualize practice context come into better interaction with cogni- tive demands for performing the task in the initial phases of acquisition. The study showed that the participants in the gradual self-control method gradually requested more difficult levels in the acquisition phase. Most participants in the dual-factor gradual self-control group asked for the

RESEARCH QUARTERLY FOR EXERCISE AND SPORT 437

438 M. JALALVAND ET AL.

highest level of task difficulty at the beginning of the second half of the acquisition phase. It seems further along the practice, a new challenge was created by giving self-control to feedback. Therefore, according to CPF, it is expected that the advantages of gradual self-control prac- tice presumably come from better adjustment of challenge points by the learner in the course of the practice and its challenging nature. This study showed that golf putting movement pattern was better in the gradual self-control practice method. Therefore, the effect of self-control prac- tice could be observed by the performance production measure.

Acknowledgments

The authors thank all students who participated in this study. They also thank John Sweller of the University of New South Wales for his suggestions regarding the mental difficulty ques- tionnaire, Jared Porter of Southern Illinois University for the instrument of the movement pattern of the putt, and Hamid Salehi of Isfahan University for statistical data analysis.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

References

Akizuki, K., & Ohashi, Y. (2015). Measurement of functional task difficulty during motor learning: What level of diffi- culty corresponds to the optimal challenge point? Human Movement Science, 43, 107–117. doi:10.1016/j. humov.2015.07.007

Andrieux, M., Boutin, A., & Thon, B. (2016). Self-control of task difficulty during early practice promotes motor skill learning. Journal of Motor Behavior, 48(1), 57–65. doi:10.1080/00222895.2015.1037879

Andrieux, M., Danna, J., & Thon, B. (2012). Self-control of task difficulty during training enhances motor learning of a complex coincidence-anticipation task. Research Quarterly for Exercise and Sport, 83(1), 27–35. doi:10.1080/02701367.2012.10599822

Brydges, R., Carnahan, H., Rose, D., & Dubrowski, A. (2010). Comparing self-guided learning and educator-guided learning formats for simulation-based clinical training. Journal of Advanced Nursing, 66(8), 1832–1844. doi:10.1111/j.1365-2648.2010.05338.x

Chiviacowsky, S., & Wulf, G. (2002). Self-controlled feed- back: Does it enhance learning because performers get feedback when they need it? Research Quarterly for Exercise and Sport, 73(4), 408–415. doi:10.1080/ 02701367.2002.10609040

Chiviacowsky, S., & Wulf, G. (2005). Self-controlled feedback is effective if it is based on the learner’s performance. Research Quarterly for Exercise and Sport, 76(1), 42–48. doi:10.1080/02701367.2005.10599260

Fairbrother, J. T., Laughlin, D. D., & Nguyen, T. V. (2012). Self-controlled feedback facilitates motor learning in both high and low activity individuals. Frontiers in Psychology, 3, 323. doi:10.3389/fpsyg.2012.00323

Grand, K. F., Bruzi, A. T., Dyke, F. B., Godwin, M. M., Leiker, A. M., Thompson, A. G., ... Miller, M. W. (2015). Why self-controlled feedback enhances motor learning: Answers from electroencephalography and indices of motivation. Human Movement Science, 43, 23–32. doi:10.1016/j.humov.2015.06.013

Guadagnoli, M. A., & Lee, T. D. (2004). Challenge point: A framework for conceptualizing the effects of various practice conditions in motor learning. Journal of Motor Behavior, 36(2), 212–224. doi:10.3200/JMBR.36.2.212-224

Hartman, J. M. (2007). Self-controlled use of a perceived physical assistance device during a balancing task. Perceptual and Motor Skills, 104(3), 1005–1016. doi:10.2466/pms.104.3.1005-1016

Hemayattalab, R. (2014). Effects of self-control and instructor-control feedback on motor learning in individuals with cerebral palsy. Research in Developmental Disabilities, 35 (11), 2766–2772. doi:10.1016/j.ridd.2014.07.006

Hodges, N. J., Edwards, C., Luttin, S., & Bowcock, A. (2011). Learning from the experts: Gaining insights into best practice during the acquisition of three novel motor skills. Research Quarterly for Exercise and Sport, 82(2), 178–187. doi:10.1080/02701367.2011.10599745

Janelle, C. M., Barba, D. A., Frehlich, S. G., Tennant, L. K., & Cauraugh, J. H. (1997). Maximizing performance feedback effectiveness through videotape replay and a self-controlled learning environment. Research Quarterly for Exercise and Sport, 68(4), 269–279. doi:10.1080/ 02701367.1997.10608008

Janelle, C. M., Kim, J., & Singer, R. N. (1995). Subject- controlled performance feedback and learning of a closed motor skill. Perceptual and Motor Skills, 81(2), 627–634. doi:10.2466/pms.1995.81.2.627

Jowett, N., LeBlanc, V., Xeroulis, G., MacRae, H., & Dubrowski, A. (2007). Surgical skill acquisition with self-directed practice using computer-based video training. The American Journal of Surgery, 193(2), 237–242. doi:10.1016/j.amjsurg.2006.11.003

Kanfer, R., & Ackerman, P. L. (1989). Motivation and cogni- tive abilities: An integrative/aptitude-treatment interaction approach to skill acquisition. Journal of Applied Psychology, 74(4), 657. doi:10.1037/0021-9010.74.4.657

Keetch, K. M., & Lee, T. D. (2007). The effect of self-regulated and experimenter-imposed practice schedules on motor learning for tasks of varying difficulty. Research Quarterly for Exercise and Sport, 78(5), 476–486. doi:10.1080/ 02701367.2007.10599447

Lessa, H. T., & Chiviacowsky, S. (2015). Self-controlled prac- tice benefits motor learning in older adults. Human Movement Science, 40, 372–380. doi:10.1016/j. humov.2015.01.013

Magill, R. (2011). Motor control and learning: Concepts and applications (9th ed.). New York, NY: McGraw-Hill.

Magill, R., & Anderson, D. (2014). Motor control and learn- ing: Concepts and applications (10th ed.). New York, NY: McGraw-Hill.

Maxwell, J. P., Masters, R. S. W., Kerr, E., & Weedon, E. (2001). The implicit benefit of learning without errors. The

Quarterly Journal of Experimental Psychology: Section A, 54

(4), 1049–1068. doi:10.1080/713756014 Patterson, J. T., Carter, M. J., & Hansen, S. (2013). Self-

controlled KR schedules: Does repetition order matter? Human Movement Science, 32(4), 567–579. doi:10.1016/j. humov.2013.03.005

Patterson, J. T., Carter, M., & Sanli, E. (2011). Decreasing the proportion of self-control trials during the acquisition period does not compromise the learning advantages in a self-controlled context. Research Quarterly for Exercise and Sport, 82(4), 624–633. doi:10.1080/02701367.2011.10599799

Poolton, J. M., Masters, R. S. W., & Maxwell, J. P. (2005). The relationship between initial errorless learning conditions and subsequent performance. Human Movement Science, 24(3), 362–378. doi:10.1016/j.humov.2005.06.006

Porter, J. M., Landin, D., Hebert, E. P., & Baum, B. (2007). The effects of three levels of contextual interference on performance outcomes and movement patterns in golf skills. International Journal of Sports Science & Coaching, 2(3), 243–255. doi:10.1260/174795407782233100

Post, P. G., Fairbrother, J. T., & Barros, J. A. (2011). Self- controlled amount of practice benefits learning of a motor skill. Research Quarterly for Exercise and Sport, 82(3), 474–481. doi:10.1080/02701367.2011.10599780

Schmidt, R., & Lee, T. (2013). Motor learning and perfor- mance, 5E with web study guide: From principles to appli- cation. Champaign, Illinois: Human Kinetics.

Schmidt, R. A. (1975). A schema theory of discrete motor skill learning. Psychological Review, 82(4), 225. doi:10.1037/ h0076770

Wulf, G., Chiviacowsky, S., & Drews, R. (2015). External focus and autonomy support: Two important factors in motor learning have additive benefits. Human Movement Science, 40, 176–184. doi:10.1016/j. humov.2014.11.015

Wulf, G., Clauss, A., Shea, C. H., & Whitacre, C. A. (2001). Benefits of self-control in dyad practice. Research Quarterly for Exercise and Sport, 72(3), 299–303. doi:10.1080/02701367.2001.10608964

Wulf, G., Shea, C., & Lewthwaite, R. (2010). Motor skill learning and performance: A review of influential factors. Medical Education, 44(1), 75–84. doi:10.1111/j.1365- 2923.2009.03421.x

Wulf, G., & Toole, T. (1999). Physical assistance devices in complex motor skill learning: Benefits of a self-controlled practice schedule. Research Quarterly for Exercise and Sport, 70(3), 265–272. doi:10.1080/ 02701367.1999.10608045