Mental toughness & perseverance 1
Running head: Mental toughness & perseverance
When the going gets tough: Mental toughness and its relationship with behavioural
perseverance
1Daniel F. Gucciardi*, 2Peter Peeling, 1Kagan J. Ducker, and 2Brian Dawson
1School of Physiotherapy and Exercise Science, Curtin University
2School of Sport Science, Exercise & Health, The University of Western Australia
Author Notes
*Address correspondence to Daniel Gucciardi, School of Physiotherapy and Exercise
Science, Curtin University, GPO Box U1987, Perth, Australia, 6845. Email:
Word Count: 3060
Abstract Word Count: 208
Number of Tables: 3
Number of Figures: 0
Supplementary Material: yes
Manuscript accepted for publication in Journal of Science and Medicine in Sport on
December 6th, 2014
Mental toughness & perseverance 2
Abstract 1
Objectives: This study examined the association between self-reported mental toughness and 2
behavioural perseverance among a sample of male Australian footballers in a naturalistic 3
context. 4
Design: Cross-sectional study, with the multistage 20 m shuttle run test (MST) employed as a 5
proxy for behavioural perseverance. 6
Methods: 330 male Australian footballers aged between 15 and 18 years (M = 16.86; SD = 7
.71) with between 2 and 14 years playing experience (M = 9.32; SD = 2.51) participated. 8
Initially, footballers completed a mental toughness questionnaire, before having their height 9
and body mass measurements taken. Subsequently, a performance testing session was 10
completed, which included the 20 m sprint, Australian football-specific agility run, vertical 11
jump, and the MST. Bayesian estimation was employed to allow for the simultaneous 12
examination of existing findings with our new data in a way that provides an automatic meta-13
analysis of evidence in this area. 14
Results: The analysis indicated a 95% probability that the association between mental 15
toughness and behavioural perseverance lies between .14 and .34, even when controlling for 16
other factors known to influence MST performance, including age, height, body mass, and 17
years playing experience. 18
Conclusions: Taken together with previous research, these findings support the theoretical 19
proposition that persistence, effort or perseverance represents a behavioural signature of 20
mental toughness. 21
22
Keywords: Bayesian structural equation modelling; beep test; mentally tough; substantive-23
methodological synergy 24
Mental toughness & perseverance 3
Introduction 1
Coaches, athletes, and sport scientists agree that psychological characteristics – 2
alongside physical, technical, and tactical skills – are essential for optimal sport 3
performance1. Over the past decade, mental toughness has gained considerable attention as 4
the umbrella construct that appears to encapsulate several of the key psychological attributes 5
central to achieving sporting success. Initial research on this concept was founded in 6
professional practice knowledge, that is, information derived largely from the experiences 7
and perceptions of practitioners working in the field. The unsystematic nature of the 8
accumulation and communication of this professional practice knowledge was addressed in 9
the next wave of research, whereby scholars identified and described unobservable personal 10
attributes considered central to mental toughness (e.g., self-belief, emotion regulation) 11
through systematic investigations. As a collective, the work completed thus far suggests that 12
mental toughness represents as a psychological capacity to deliver high performance on a 13
regular basis despite varying degrees of situational demands2,3. 14
In an attempt to advance conceptualisations of mental toughness, researchers have 15
turned their attention to the behavioural features of this psychological concept3,4, that is, 16
observable behaviours or actions that are typically demonstrated in challenging or demanding 17
situations (i.e., person-situation interaction). Persistence, effort, or perseverance is often 18
reported as a behavioural signature of mentally tough individuals, akin to the psychological 19
concept of grit5; for example, pushing through challenging situations, refusing to give up or 20
quit, and sustaining high effort levels over time6,7,8. Recent research on mental toughness in 21
non-athlete samples has supported these initial contentions, particularly as it pertains to 22
sustained efforts towards longer-term goals. Among tertiary students, for example, self-23
reported mental toughness has been associated with higher levels of academic and social goal 24
progress over a university semester2. Furthermore, in a six-week selection course designed to 25
Mental toughness & perseverance 4
assess candidate’s suitability for entry into elite military training (i.e., Special Forces), results 1
indicated that for a one unit increase in self-reported mental toughness, the odds of passing 2
the selection test (versus failing) increased by a factor of 3.48, that is, just over three times as 3
likely to have passed the selection course2. 4
Only one study to date has examined persistence, effort, or perseverance as a 5
behavioural signature of mentally tough athletes. Compared with the aforementioned studies 6
in non-athlete samples2, behavioural perseverance was examined using a discrete task rather 7
than over an extended period of time among a sample of young elite cricketers9. In this study, 8
an intervention that systematically exposed cricketers to punishment-conditioned stimuli (i.e., 9
consequences for behaviour) using a multidisciplinary team and transformational delivery 10
approach (e.g., coaches, medical staff, and administrators expressed belief in the cricketers’ 11
ability to achieve the vision) improved the experimental groups’ behavioural perseverance as 12
measured by the 20-m multistage shuttle run test (MST) when compared with the control 13
group. Additionally, these improvements in behavioural perseverance over the 12 month 14
testing period were accompanied by increases in coach-rated mental toughness for the 15
experimental group, but not for the control group. 16
The current study was designed to extend this foundational work by examining the 17
extent to which these findings obtained with cricketers in a controlled setting would 18
generalise to Australian footballers in a naturalistic context. The inclusion of an objective 19
measure of behaviour also addressed one of the major criticisms of the mental toughness 20
literature, that is, the reliance on arbitrary metrics (e.g., correlating self-reported mental 21
toughness with self-reported stress). Arbitrary metrics provide little insight into the meaning 22
of self-reported psychological concepts for real-world behaviour and are therefore inadequate 23
for testing and informing psychological theories10. The Australian Football League (AFL) 24
National Draft Combine testing protocol is completed annually by potential draftees and 25
Mental toughness & perseverance 5
development players to assess a wide range of physical capacities. This testing battery 1
contains numerous assessments of physical qualities such as anthropometry (∑7 skinfolds, 2
height, body mass, arm length, hand span), speed (20 m sprint), speed-endurance (6 x 30 m 3
repeat sprint), agility (AFL-specific agility), power (vertical jump and running vertical jump) 4
and aerobic capacity (MST). Given the relationship between the MST and aerobic capacity as 5
measured by VO2max in elite and sub-elite athletes is imperfect (r = .61-78)11,12, psychological 6
factors may provide insight into the unexplained variance in this test (between 39% and 7
63%). Previously, the MST has been employed as a proxy of behavioural perseverance9, as it 8
requires individuals to sustain high levels of effort and push through physical and mental 9
demands in a single testing session. Consistent with experimental8 and prospective evidence2, 10
we hypothesised that mental toughness would be positively associated with behavioural 11
perseverance as measured by the MST. 12
This study also offers a methodological extension to previous research on mental 13
toughness, and more broadly the sport and exercise sciences. In contrast to traditionally 14
employed frequentist approaches (e.g., maximum likelihood, p values), we utilise Bayesian 15
estimation to explicitly integrate the prior findings of Bell and colleagues9 into our analytical 16
framework, thereby providing an automatic meta-analysis13. Other differences also exist 17
between Bayesian and frequentist approaches14. For example, the frequentist interpretation of 18
the 95% confidence interval is based on long run frequency, such that the 95% confidence 19
intervals of an infinite number of replications of the same experiment or study will capture 20
the fixed but unknown parameter estimate under the null hypothesis (i.e., true population 21
estimate). This interpretation differs within Bayesian statistics whereby the 95% confidence 22
interval (referred to as a credibility interval) indicates the probability that the unknown and 23
therefore random parameter estimate lies between the lower and upper values of the interval. 24
The parameter estimate is considered substantively important when the 95% credibility 25
Mental toughness & perseverance 6
interval does not contain zero, and therefore the null hypothesis is rejected. Additionally, 1
Bayesian analysis offers intuitive interpretations because of the direct focus on the probability 2
distribution of parameters, such that one can make statements about the probability of a 3
hypothesised theoretical model, given the data13. In contrast, frequentist approaches permit 4
statements about the probability of the data, given the hypothesised theoretical model. 5
Methods 6
Participants were 330 male Australian footballers aged between 15 and 18 years (M = 7
16.86; SD = .71). Footballers had played competitive football for between 2 and 14 years (M 8
= 9.32; SD = 2.51) at the time of participating, and were playing in the highest competition 9
within the state for their age group. Institutional ethics approval was obtained prior to the 10
commencement of this study. 11
Footballers were tested in groups of approximately 40 players during a 2-3 hour 12
session, which occurred over a 3-day period at the conclusion of the pre-season phase of the 13
Western Australian Football League colts training program. Upon arrival, participants first 14
completed a brief questionnaire including demographic information (age, years playing) and 15
self-reported mental toughness in a large lecture theatre (seating capacity of 150). A semantic 16
differential approach was employed to directly measure mental toughness, using an 17
adaptation of the revised performance profile technique15. This approach contrasts with an 18
indirect approach whereby mental toughness is inferred from participants’ scores on 19
individual facets such as self-efficacy, optimism, and emotional regulation. There is evidence 20
to support the superiority of a direct approach to the measurement of mental toughness2. 21
Participants responded to 11 items on a 7-point scale anchored by bipolar construct 22
descriptions of the 11 key mental toughness components in Australian football5. The 7-point 23
scale included anchors of “100% of the time” for both poles (1 = contrast pole of mental 24
toughness and 7 = mentally tough pole) and “50% of the time” at the midpoint (see Table 2). 25
Mental toughness & perseverance 7
After participants completed the survey, their height and body mass was assessed. 1
Height was measured using a wall mounted stadiometer (Harpenden stadiometer, Holtain 2
Ltd. UK) and body mass recorded on a platform scale (UC300, A & D Mercury Pty Ltd. SA, 3
AUS). Following the completion of these anthropometric measures, a team-led warm-up was 4
completed before a performance testing session was commenced. The performance tests used 5
here were a subset of those implemented at the AFL National Draft Combine, which included 6
the 20 m sprint, AFL-specific agility run, vertical jump, and the MST16. The MST was 7
completed as per the standardised instructions published by Leger et al.17 in groups of 8
approximately 20 footballers. All tests were performed in an indoor gymnasium on a sprung 9
wooden floor. Only the MST data is reported in this study, given our focus on behavioural 10
perseverance rather than skill-related movement (e.g., vertical jump). 11
As correlation and multiple regression analyses do not take measurement error into 12
consideration (i.e., observed variables are assumed to be measured without error), the 13
relationship between mental toughness and behavioural perseverance (as measured by MST 14
performance) was tested within a Bayesian structural equation modelling (BSEM) 15
framework18 using Mplus 7.219. Age, height, body mass, and football experience were 16
included as covariates in the model, thereby permitting a test of the contribution of mental 17
toughness to behavioural perseverance while accounting for the influence of these variables. 18
The covariates were modelled as observed variables, whereas mental toughness was modelled 19
as a latent variable including those item indicators detailed in Table 2 and their error terms. 20
We drew from statistical recommendations regarding the quality of factor loadings to guide 21
the specification of priors for the mental toughness measurement model20, such that items 22
were specified to have a normal prior of .80 and a standard deviation + .34. The prior for the 23
structural path between mental toughness and behavioural perseverance was guided by Bell 24
and colleagues’ study with elite youth cricketers9, such that we converted their interaction 25
Mental toughness & perseverance 8
effect size into an unstandardised correlation coefficient (i.e., μ =.40, σ2 = .02, SD = + .28)21. 1
Weakly informative priors incorporate prior knowledge regarding the population parameter in 2
the model (e.g., specific value of the mean most likely) but with some degree of uncertainty 3
(i.e., moderate amount of variance around a specific value of the mean), and therefore do not 4
substantially influence the final parameter estimate in the posterior distribution once 5
combined with the data14. The Mplus syntax is provided in Table 3. 6
The posterior predictive p value (PPP) is computed to provide an indication of model 7
fit within BSEM; this index compares the deviation between the real and replicated data to 8
produce a 95% confidence interval for this discrepancy function22. A small positive PPP 9
value (e.g., 0.05) is indicative of poor fit, and a value around 0.5 and above is suggestive of 10
excellent fit18. Statistical criteria (i.e., potential scale reduction factor value is < 1.123) 11
alongside visual inspection of the trace plots (i.e., multiple chains converged to a similar 12
target distribution11) provided an indication of model convergence. See Appendix A of the 13
online supplementary material for further information on Bayesian estimation. The reliability 14
of the latent mental toughness factor was computed using McDonald’s24 omega coefficient 15
(ω). 16
Results 17
The probability of the hypothesised model, given the data, was acceptable (PPP = 18
.228, Δobserved and replicated 2 95% CI [-29.21, 61.31]). Four chains were estimated and 19
in 23300 iterations reached an appropriate convergence criterion. Visual inspection of trace 20
plots (e.g., see Figures S1, S2, S3, and S4 of Supplementary Material) and an examination of 21
the PSR development over iterations (i.e., smooth decrease in PSR, last few thousand 22
iterations were close to 1) verified support for model convergence18. Standardised factor 23
loadings for the mental toughness items (λ > .635; M = .72, SD = .06) and its reliability were 24
excellent (ω = .92). Mental toughness was positively associated with football experience (β = 25
Mental toughness & perseverance 9
.19; 95% CI = .08, .30), and body mass was positively associated with height (β = .73; 95% 1
CI = .67, .77) and age (β = .12; 95% CI = .01, .23); all other correlations among the 2
covariates and mental toughness were not significant (see Table 1). Body mass (β = -.35; 3
95% CI = -.49, -.20) and mental toughness (β = .24; 95% CI = .14, .34) were substantively 4
important antecedents of MST performance (see Table 1). Mental toughness and the 5
covariates accounted for 17.2% of the variance in behavioural perseverance, whereby the 6
inclusion of mental toughness to the model accounted for an additional 5.4% of the explained 7
variance. Couched within a Bayesian framework, these findings indicated that there is a 95% 8
probability that the parameter value falls between the lower and upper limits of these 9
intervals. For example, the relationship between mental toughness and behavioural 10
perseverance lies between .14 and .34. 11
Discussion 12
Consistent with previous research2,9, we hypothesised that mental toughness would be 13
positively associated with behavioural perseverance, and this relationship would hold while 14
controlling for covariates including age, body mass, height, and years playing experience. In 15
testing this hypothesis, we employed an emerging methodology – Bayesian structural 16
equation modelling – to formally integrate prior findings into our analytical framework, 17
thereby providing an automatic meta-analysis13. Our hypothesis regarding the relationship 18
between mental toughness and behavioural perseverance was supported, indicating that this 19
psychological capacity accounted for an additional 5.4% of the variance in MST performance 20
beyond key covariates known to influence performance on this test (age, height, playing 21
experience, body mass). This result provides evidence that previous findings obtained in a 22
controlled setting with cricketers – whereby mental toughness was manipulated via a psycho-23
social intervention – generalise to a naturalistic, field-testing situation with Australian 24
footballers9. Taken together with previous research in non-athlete samples2 and descriptions 25
Mental toughness & perseverance 10
of mentally tough athletes6,7,8, there appears to be emerging evidence to support the notion 1
that persistence, effort, or perseverance represents a behavioural signature of mentally tough 2
individuals. From a practical standpoint, the MST is a key predictor of being selected for elite 3
football (AFL), with only small differences between those players who are ‘drafted’ and 4
those who compete in the National championships25. Nevertheless, additional work is 5
required to ascertain whether or not mental toughness provides incremental validity over and 6
above other pertinent individual difference variables (e.g., emotion regulation). 7
Mental toughness is a complex psychological concept that captures one’s personal 8
capacity to deliver high performance on a regular basis despite varying degrees of situational 9
demands2,3. From a transactional perspective of stress, mental toughness may play a role in 10
facilitating behavioural perseverance in performance situations that are goal-relevant and 11
therefore task-engaging through an appraisal system where situational demands are perceived 12
as either challenging or threatening for valued outcomes such as performance and well-13
being26. Previous research on mental toughness in sport has supported the importance of 14
distinguishing between negative and positive stressors5. The extent to which individuals 15
perceive features of the situation as challenging or threatening depends on their relative 16
evaluations of the demands of the situation alongside their available resources. When the 17
available resources are perceived as outweighing the situational demands, individuals 18
appraise the situation as challenging; this perception contrasts with threatening appraisals in 19
which individuals perceive the demands of the situation as exceeding their available 20
resources26,27. Challenge appraisals typically foster positive outcomes such as mastery, 21
growth, or gains, whereas hindrance stressors typically thwart the attainment of such positive 22
outcomes28. An empirical test of this theoretical speculation represents one way in which the 23
current study may be extended in future research. 24
Mental toughness & perseverance 11
A key strength of this study was the use of Bayesian estimation, which allowed us to 1
directly test our informative hypothesis that self-reported mental toughness would be 2
associated with behavioural perseverance, as well as integrate prior knowledge into the 3
analysis to facilitate estimation in the model with our new data. In this way, existing 4
knowledge is updated with new data to produce results that are an automatic meta-analysis13. 5
Bayesian estimation contrasts with the traditional frequentist approach in which one aims to 6
examine such informative hypotheses yet tests the null hypothesis. For example, the null 7
hypothesis testing approach would permit the following statement as it applies in the current 8
study; the probability of making an error with our finding that the relationship between 9
mental toughness and behavioural perseverance is not zero is below 5% (when p < .05). 10
Bayesian estimation has been advocated29 and demonstrated30 as an important 11
methodological consideration for knowledge development and accumulation in the sport and 12
exercise sciences. 13
Despite this methodological strength, it is important to acknowledge additional 14
avenues of future research that are grounded in the limitations of the current study. First, 15
although a casual direction is implied in a structural equation modelling framework of cross-16
sectional data, the non-experimental nature of this study limits our ability to infer causality 17
from the current findings. Longitudinal examinations and experimental manipulations of 18
mental toughness (or appraisals of the situational demands) would also assist interpretations 19
on the causal nature of these relations. For example, it would be interesting to ascertain 20
whether or not mental toughness moderates the relationship between behavioural 21
perseverance in a low and high pressure testing situation. Second, it should be considered that 22
as a result of the large number of athletes sampled in this investigation, it was not possible to 23
test all footballers at the same time of day. As a result, it is possible that the performance 24
outcomes may be biased by circadian influences. Third, our methodological design may have 25
Mental toughness & perseverance 12
introduced a ‘mere measurement effect’ whereby completing a mental toughness survey 1
immediately prior to the MST enhanced the accessibility of beliefs, attitudes, or motivations 2
about the target behaviour and therefore increased the likelihood of actioning that behaviour. 3
Fourth, we were unable to account for potential non-independence in the MST data (i.e., 4
footballers who performed the test as a group, e.g., social facilitation effects), as this 5
information was not recorded during data collection. Finally, given the field-based nature of 6
our measure of behavioural perseverance, we are unable to ascertain whether or not this 7
psychological capacity to deliver high performance despite varying degrees of situational 8
demands is related to one’s underlying fitness levels. A baseline measure of physical fitness 9
(i.e., maximal aerobic capacity) would need to be included in any statistical model alongside 10
mental toughness to account for this variable in the prediction of behavioural perseverance. 11
Conclusions 12
In this study, we provided cross-sectional support for the theoretical proposition that 13
perseverance is a behavioural signature of mentally tough athletes. By incorporating 14
empirical2,9 and subjective knowledge6,7,8 into our analysis, Bayesian estimation allowed for 15
the simultaneous examination of these ‘old’ results with our new data to provide an automatic 16
update on this aspect of mental toughness theory13. Delineating the mechanisms by which 17
mental toughness translates into behavioural perseverance represents an important 18
consideration for advancing theory and practice in this area (e.g., appraisal systems). 19
Practical Implications 20
• Persistence, effort or perseverance is a behavioural signature of mental toughness 21
• Physical and psychological attributes are both important considerations for MST 22
performance 23
• When athletes deliver high performances or even fail, reinforce aspects of their behaviour 24
that relate to persistence, effort, or perseverance 25
Mental toughness & perseverance 13
Acknowledgements
We wish to acknowledge the assistance of Jon Haines from the Western Australian
Football Commission. Appreciation is also extended to the following individuals for their
assistance with data collection: Kym Guelfi, Ray Davey, Chee Yong Low, Marcus Lee and
Bradley Kelly. No financial assistance was received for this study.
Mental toughness & perseverance 14
References
1. Gulbin JP, Croser MJ, Morley EJ, et al. An integrated framework for the optimisation
of sport and athlete development: a practitioner approach. J Sports Sci 2013; 31(12):
1319-1331. doi: 10.1080/02640414.2013.781661
2. Gucciardi DF, Hanton S, Gordon S, et al. The concept of mental toughness: tests of
dimensionality, nomological network, and traitness. J Pers in press. doi:
10.1111/jopy.12079
3. Hardy L, Bell J, Beattie S. Preliminary evidence for a neuropsychological model of
mentally tough behaviour. J Pers 2014; 82(1): 69-81. doi: 10.1111/jopy.12034
4. Gucciardi DF, Jackson B, Hanton S, et al. Motivational correlates of mentally tough
behaviors in tennis. J Sci Med Sport in press. doi: 10.1016/j.jsams.2013.11.009
5. Duckworth AL, Peterson C, Matthews MD, et al. Grit: perseverance and passion for
long-term goals. J Pers Soc Psychol 2007; 92(6): 1087-1101.
6. Gucciardi DF, Gordon S, Dimmock JA. Towards an understanding of mental
toughness in Australian football. J App Sport Psychol 2008; 20(3): 261-281.
7. Jones G, Hanton S, Connaughton D. What is this thing called mental toughness? an
investigation of elite sport performers. J App Sport Psychol 2002; 14(3): 205-218. doi:
10.1080/10413200290103509
8. Weinberg R, Butt J, Culp, B. Coaches’ views of mental toughness and how it is built.
Int J Sport Exercise Psychol 2011; 9(2): 156-172. doi:
10.1080/1612197X.2011.567106
9. Bell JJ, Hardy L, Beattie S. Enhancing mental toughness and performance under
pressure in elite young cricketers: a 2-year longitudinal intervention. Sport, Exercise
Perform Psychol 2014; 2(2): 281-297. doi: 10.1037/a0033129
Mental toughness & perseverance 15
10. Andersen MB, McCullagh P, Wilson GJ. But what do the numbers really tell us?:
Arbitrary metrics and effect size reporting in sport psychology research. J Sport
Exercise Psychol 2007; 29(5): 664-672.
11. Gibson AS, Broomhead S, Lambert MI, et al. Prediction of maximal oxygen uptake
from a 20-m shuttle run as measured directly in runners and squash players. J Sport
Sci 1998; 16(4): 331-335.
12. Thomas A, Dawson B, Goodman C. The yo-yo test: reliability and association with a
20-m shuttle run and VO2max. Int J Sports Physiol Perform 2006; 1(2): 137-149.
13. Zyphur MJ, Oswald FL. Bayesian estimation and inference: A user’s guide. J Manage
in press. doi: 10.1177/0149206313501200
14. van de Schoot R, Kaplan D, Denissen J, et al. A gentle introduction to Bayesian
analysis: applications to developmental research. Child Dev 2014; 85(3): 842-860.
doi: 10.1111/cdev.12169
15. Gucciardi DF, Gordon, S. Revisiting the performance profile technique: theoretical
underpinnings and application. Sport Psychol 2009; 23(1): 93-117.
16. Pyne DB, Gardner AS., Sheehan K, et al. Positional differences in fitness and
anthropometric characteristics in Australian football. J Sci Med Sport 2006; 9(1-2);
143-150. Doi: 10.1016/j.jsams.2005.10.001
17. Léger LA, Mercier D, Gadoury C, et al. The multistage 20 metre shuttle run test for
aerobic fitness. J Sports Sci 1988; 6(2): 93–101. doi: 10.1080/02640418808729800
18. Muthén BO, Asparouhov T. Bayesian structural equation modelling: a more flexible
representation of substantive theory. Psychol Methods 2012; 17(3): 313-335. doi:
10.1037/a0026802
19. Muthén, L. K., & Muthén, B. Mplus users guide (7th ed.). Los Angeles CA: Muthén &
Muthén. 2012.
Mental toughness & perseverance 16
20. Comrey AL, Lee HB. A first course in factor analysis, New Jersey, Erlbaum. 1992.
21. Lakens D. Calculating and reporting effect sizes to facilitate cumulative science: A
practical primer for t-tests and ANOVAs. Front Psychol 2013; 4: 863.
doi:10.3389/fpsyg.2013.00863.
22. Gelman A, Meng XL, Stern, H. Posterior predictive assessment of model fitness via
realized discrepancies. Stat Sinica 1996; 6(4): 733–759.
23. Asparouhov T, Muthén B. Bayesian analysis using Mplus: Technical implementation
2010. Retrieved from http://www.statmodel.com/download/Bayes3.pdf
24. McDonald RP. The theoretical foundations of principal factor analysis, canonical
factor analysis and alpha factor analysis. Brit J Math Psychol 1970; 23: 1–21.
doi:10.1111/j.2044-8317.1970.tb00432.x
25. Robertson S, Woods C, Gaston P. Predicting higher selection in elite junior Australian
rules football: the influence of physical performance anthropometric attributes. J Sci
Med Sport in press. doi: 10.1016/j.jsams.2014.07.019
26. Lazarus RS, Folkman S. Stress, appraisals, and coping. New York, Springer, 1984.
27. Jones MV, Meijen C, McCarthy PJ, et al. (2009). A theory of challenge and threat
states in athletes. Int Rev Sport Exercise Psych 2009; 2(2): 161–180. doi:
10.1080/17509840902829331.
28. Cavanaugh MA, Boswell WR, Roehling MV, et al. An empirical examination of self-
reported work stress among U.S. managers. J App Psychol 2000; 85(1): 65–74. doi:
10.1037/0021-9010.85.1.65
29. Welsh AH, Knight EJ. “Magnitude-based inference”: a statistical review. Med Sci
Sport Exercise in press. doi: 10.1249/MSS.0000000000000451
Mental toughness & perseverance 17
30. Gucciardi DF, Jackson B. Understanding sport continuation: an integration of the
theories of planned behaviour and basic psychological needs. J Sci Med Sport in
press. doi: 10.1016/j.jsams.2013.11.011
Mental toughness & perseverance 18
Table 1. Descriptive statistics and standardised parameter estimates for structural equation model including mental toughness and covariates as
predictors of behavioural perseverance (Note: 95% credibility intervals are presented in parentheses; MST = multistage 20 m shuttle run test).
Correlations Structural
Path
M SD 1 2 3 4 5 MST
1 Age 16.86 .71 - -.07 (-.17, .04)
2 Playing experience 9.32 2.51 .18 (.07, .28) - .04 (-.06, .15)
3 Height 180.55 6.46 .02 (-.09, .13) -.02 (-.13, .09) - .08 (-.07, .22)
4 Body mass 74.11 8.37 .12 (.01, .23) -.05 (-.16, .06) .73 (.67, .78) - -.35 (-.49, -.20)
5 Mental toughness 5.67 .69 -.02 (-.14, .09) .19 (.08, .30) .02 (-.09, .13) .03 (-.09, .14) - .24 (.14, .34)
6 MST 12.61 1.02 - - - - - -
Mental toughness & perseverance 19
19
Table 2. Semantic differential approach to measuring mental toughness in Australian football.
Instructions: The items below capture various beliefs, attitudes, and behaviours that people hold or display. Using the scale provided, please indicate the extent to which these
attributes provide a reflection of you as footballer. For example, if you always believe in your physical and mental abilities, mark the box on the right-hand side (X). If you sometimes
believe in your ability, and in other circumstances you doubt your abilities, then mark yourself within the middle sections of the continuum based on approximate percentages; for
example, you believe in yourself approx. 50% of the time (O) or only 25-30% of the time (*). If you always have doubts in yourself, then mark the box on the left-hand side (#).
100% 50/50 100%
100% doubt in my physical my physical and mental ability
under pressure.
# * O X 100% belief in my physical and mental ability under pressure.
Succumb to physical fatigue and niggly injuries, and dislike
50/50 contests.
100% ability to play through physical fatigue, play whilst carrying a
niggly injury, and enjoy 50/50 situations.
100% lack of motivation and an attitude of only doing the
basics and enough to get by.
A philosophy characterized by always working hard and pushing yourself
through (physically and mentally) demanding situations in competition
and training.
Lack “footy smarts” or an understanding of the game and the
way it is played.
100% awareness of the pressures, adversities, and challenges in football
and an understanding of the game.
Poor integrity and personal philosophy on life and football,
and generally conform without any care in the world.
I place great importance and significance on personal values relevant to
one becoming a better person and footballer.
Lack an awareness of my emotions, how they facilitate my
performance, and how to regulate them to perform well.
100% awareness of my emotions, and ability to manage my emotions to
enhance performance across all situations.
Extrinsically and/or unmotivated thereby requiring an
external source to get me going 100% of the time.
Internal motivation and desire for competitive challenges and team
success, and having the desire to put the necessary things into practice to
achieve my goals 100% of the time.
Suboptimal performance under pressure as a result of
anxiety, nervousness, and threat.
I am 100% able to execute skills and procedures under pressure and
stress.
A weak attitude characterized by laziness, easily intimidated
and giving in too easily, and succumbing to pressure
A 100% unshakeable, tough attitude directed towards becoming a
champion of the game.
Easily distracted and unfocused on the job at hand 100% of
the time.
100% focus and concentrate on the job at hand and what you want to
achieve despite internal or external pressures, obstacles, or adversities.
An inability to adapt to pressure, being easily broken, and
not putting up a fight in the face of adversity.
100% ability to overcome adversities with an exceptional work ethic and
persevering determination to showcase my mental and physical ability.
Mental toughness & perseverance 20
20
Table 3. Overview of Mplus specifications for Bayesian analysis (Note: text in green and preceded
by an exclamation mark is not read by Mplus when executing the analysis).
MODEL:
! measurement model of mental toughness
MT BY mt1* mt2 mt3 mt4 mt5 mt6 mt7 mt8 mt9 mt10 mt11 (f1l1-f1l11);
MT@1;
! structural model
shuttle ON MT (b1);
shuttle ON height weight age yrsplay;
height weight age yrsplay MT WITH height weight age yrsplay MT (cf1-cf10);
! name residual variances
mt1 mt2 mt3 mt4 mt5 mt6 mt7 mt8 mt9 mt10 mt11 (rv1-rv11);
! name the correlated residuals
mt1-mt11 WITH mt1-mt11 (cr1-cr55);
ANALYSIS:
! tells Mplus to use Bayesian estimation
ESTIMATOR = BAYES;
! specify the value of the Gelman-Rubin convergence criterion to be .01
BCONVERGENCE = .01;
! specify the use of 4 processors (to speed up computations when available)
PROCESSORS = 4;
! specify 4 independent chains of the MCMC procedure
CHAINS = 4;
MODEL PRIORS:
! informative prior for structural parameter
! where the mean is set at 0.40 and the variance is .02
! see DOI: 10.1037/a0033129
b1~N(.40,.02);
! informative priors for factor loadings of mental toughness model
! where the mean is set at 0.80 and the variance is .03
f1l1-f1l11~N(.80,.03);
! residual variances
rv1-rv11~IW(1,17);
! correlated residuals for mental toughness items
cr1-cr55~IW(0,17);
OUTPUT:
STDYX CINTERVAL TECH1 TECH8;
PLOT:
TYPE = PLOT2 PLOT3;
Mental toughness & perseverance 21
21
Supplementary Material
Appendix A – Additional Information on Bayesian Structural Equation Modelling
Broadly speaking, there are three key components to a Bayesian analysis (for reviews, see
van de Schoot & Depaoli, 2014; van de Schoot et al., 2014; Zyphur & Oswald, in press). First,
analysts must define (un)certainty about the parameters of the model in the prior distribution. This
distribution reflects background knowledge prior to collecting new data, which can be derived from
sources such as meta-analyses, related empirical research, theoretical expectations, or experts’
subjective views. As detailed in the methods section of the main document, our prior distribution
was informed by previous research on similar data (Bell, Hardy, & Beattie, 2014). The variance of a
parameter estimate captures the degree of (un)certainty in this background knowledge, whereby a
small prior variance reflects a high degree of certainty that the estimate reflects the population mean
(i.e., informative prior), and a large prior variance reflects a high degree of uncertainty in prior
beliefs of the estimate (i.e., non-informative prior). Priors can occur anywhere along this continuum
of (un)certainty from low to high precision. Second, new data is collected so that it can be
compared against these prior beliefs; in other words, what is the likelihood or probability of the
observed data given existing beliefs as captured in the parameters of the prior distribution? Third,
new data is mixed or combined with existing beliefs via Bayes theorem thereby producing a
posterior distribution. For interested readers, Zyphur and Oswald (in press) have provided a
technical but accessible description of Bayes’ rule. Essentially, the posterior distribution represents
an update in the current state of affairs that encompasses a compromise between prior beliefs and
new data. The median of the posterior distribution reflects the most likely estimate of a parameter,
whereby the shape of the peak provides an indication of the (un)certainty of this estimate (e.g.,
shallow peak indicates a large 95% range of possible estimates).
Markov chain Monte Carlo (MCMC) methods are typically performed to obtain the
posterior distribution (van de Schoot et al., 2014; Zyphur & Oswald, in press). These simulations
involve an iterative process whereby “the conditional distribution of one set of parameters given
Mental toughness & perseverance 22
22
other sets can be used to make random draws of parameter values, ultimately resulting in an
approximation of the joint distribution of all the parameters” (Muthén & Asparouhov, 2012, p.
334). In other words, all parameters of the model are repeatedly estimated to generate a distribution
that is built up from the interaction of prior beliefs with new data. As can be seen in Table 3 of the
main document, we specified the MCMC convergence criterion using the Gelman-Rubin potential
scale reduction value of .01, rather than the default Mplus formula based on the value of .05.
Interested readers are referred elsewhere for a detailed review and application of Bayesian
estimation in the sport and exercise sciences (Gucciardi & Zyphur, in press).
Mental toughness & perseverance 23
23
References
Bell, J. J., Hardy, L., & Beattie, S. (2014). Enhancing mental toughness and performance under
pressure in elite young cricketers: A 2-year longitudinal intervention. Sport, Exercise and
Performance Psychology, 2, 281-297. doi: 10.1037/a0033129
Gucciardi, D.F., & Zyphur, M.J. (in press). Exploratory structural equation modelling and Bayesian
estimation. In N. Ntoumanis & N.D. Myers (Eds.), An introduction to intermediate and
advanced statistical analyses for sport and exercise scientists. Wiley.
Muthén, B. O., & Asparouhov, T. (2012). Bayesian structural equation modeling: A more flexible
representation of substantive theory. Psychological Methods, 17, 313-335. doi:
10.1037/a0026802
van de Schoot, R., & Depaoli, S. (2014). Bayesian analyses: Where to start and what to report. The
European Health Psychologist, 16, 75-84.
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, 842-860. doi: 10.1111/cdev.12169
Zyphur, M. J., & Oswald, F. L. (in press). Bayesian estimation and inference: A user’s guide.
Journal of Management. doi: 10.1177/0149206313501200
Mental toughness & perseverance 24
24
Appendix B – Figures of Trace Plots
Figure S1. Four chains specified for the Gibbs sampler of the regression of shuttle run performance
on mental toughness.
Figure S2. Four chains specified for the Gibbs sampler of the regression of shuttle run performance
on weight.