Moderating variables in the relationship between mental toughness and performance in basketball

Aurey Newlnd,Mri Newton,Lur Finh,Colin R.Hrke,Leslie Podlog

aDepartment of Exercise and Sport Science,University of Utah,Salt Lake City,UT 84112,USA

bDepartment of Kinesiology,Health&Physical Education,St Cloud State University,St.Cloud,MN 56301,USA

cDepartment of Psychology,Western Illinois University,Macomb,IL 61455,USA

Moderating variables in the relationship between mental toughness and performance in basketball

Aubrey Newlanda,*,Maria Newtona,Laura Finchb,Colin R.Harbkec,Leslie Podloga

aDepartment of Exercise and Sport Science,University of Utah,Salt Lake City,UT 84112,USA

bDepartment of Kinesiology,Health&Physical Education,St Cloud State University,St.Cloud,MN 56301,USA

cDepartment of Psychology,Western Illinois University,Macomb,IL 61455,USA

Purpose:This study explored the relationship between mental toughness and college basketball performance,specifically examining possible moderating variables(gender and starting status).

Methods:Male and female(n=197)college basketball players completed the Psychological Performance Inventory-Alternative(PPI-A), a measure of characteristics and skills consistent with mental toughness,and the PERF,an objective measure of basketball performance.

Results:Findings suggest that basketball performance can be partially predicted by mental toughness and starting status.Males reported greater mental toughness than females.Starters and nonstarters did not differ in mental toughness.Moderated hierarchical regression analysis indicated that mental toughness was related to performance for male players as both a main effect and interaction with starter status.For female players,in contrast,starter status was the only significant predictor of performance.Practitioners are encouraged to foster the psychological skills associated with mental toughness in females and non-starters in basketball.

Conclusion:Discussion of the PPI-A as a measure of mental toughness and suggestions for its improvement are explored.A need exists for additional research on mental toughness and objective performance,as performance enhancement is a major impetus for research on mental toughness.

Copyright?2012,Shanghai University of Sport.Production and hosting by Elsevier B.V.All rights reserved.

Basketball;Mental toughness;Moderating variables;Performance;Psychological Performance Inventory-Alternative;Psychological skills

1.Introduction

Early research on mental toughness indicated that 82%of wrestling coaches identified mental toughness as the most important prerequisite to competitive success.1In support of this research,athletes,media personnel,and coaches regularly characterize successful performers as being mentally tough. While the term mental toughness is frequently used colloquially to describe athletic success,empirical support for the relationship between mental toughness and performance success has not been fully established.Given the pointed focus of this study on the relationship between mental toughness and performance,a full review of mental toughness literature is beyond the scope of this article.However,relevant conceptual discussion and measurement issues will be addressed.

Among the issues in forming a solid construct are theoretical basis,definition and conceptualization,and measurement.A social-cognitive theoretical approach has been suggested when investigating the topic of mental toughness given the social and cognitive processes involved in achieving mentaltoughness.Scholarshaveconceptualized mentaltoughness in a variety of ways and,in general,the definitions lack consistency.A focused and precise definition of mental toughness has been a topic of much debate among researchers. Jones et al.2defined mental toughness as a psychological quality that helps in coping with sport pressures and allows athletes to be consistently resolute in demonstrating psychological skills such as focus,motivation,confidence,and control.Middleton et al.3interviewed athletes from a variety of sports and proposed that mental toughness is strong determination in the face of adversity.Clough et al.4used four C’s to describe mental toughness.They have suggested that challenge,commitment,control,and confidence are central to mental toughness.While variability among the definitions exists,mental toughness appears to relate to the skillful demonstration of a collection of psychological skills.

Bull et al.5have suggested that the characteristics of mental toughness in a global sense might be distinct from how it is understood in a particular sport.Similarly,it is possible that the constituents of mental toughness differ in particular sports. For instance,mental toughness in rugby may be distinct from mental toughness in swimming.Both Bull et al.5and Thelwell et al.6in studying mental toughness in cricket and soccer respectively,found that the sport-specific definitions proposed by coaches and athletes were in line with the global definition of mental toughness created by Jones and colleagues.2Gucciardi et al.7created a sport-specific definition of mental toughness for Australian football after interviews with 11 male coaches in the Western Australian Football League.Their definition also implies that several psychological skills are necessary for an athlete to be mentally tough.Specifically, according to Gucciardi et al.7mental toughness encompasses one’scollective beliefs(encompassing attitudes,values, behaviors,and emotions)which help in overcoming barriers to success.While proposed definitions of mental toughness differ slightly in syntax,the essence of mental toughness appears to be the same:psychological resoluteness and positive coping in the face of the demands of the sporting context.The current study was informed by the general definition provided by Jones et al.2as well as the work of Gucciardi et al.7who added specificity in relation to the psychological attributes of mental toughness.

A multitude of measures have been proposed to measure mental toughness in sport.In recent years several general (Psychological Performance Inventory,PPI;8Psychological Performance Inventory-Alternative,PPI-A;9Mental Toughness Questionnaire,MTQ;4Sports Mental Toughness Questionnaire,SMTQ10)and sport-specific(Australian football MentalToughnessInventory,AfMTI;11CricketMental Toughness Inventory,CMTI12)instruments have been employed in the literature.However,none of these measures purporting to examine mental toughness have satisfied all the rigorous construct validation principles deemed necessary for the development of sound instrumentation.13For example, a measure may display satisfactory factorial validity,but lack face or external validity.Even when validity criteria have been satisfied,a measure that is not grounded in theory may fail to adequately inform researchers why particular components are included(or not)in the measure.Given the paucity of adequate measures in the current literature14the PPI-A was selected for use in this study based on its item brevity(14-items),heuristic appeal,and factorial validity.15

The PPI-A is a revised version of Loehr’s8original 42-item Psychological Performance Inventory,which has been used in a number of studies related to psychological performance skills and mental toughness.16—20Loehr8described mental toughness with characteristics such as refusal to be intimidated,unyielding attitude when beaten,retaining optimal arousal,and eagerness to compete.Based on interviews Loehr devised a 7-factor structure which included self-confidence, negative energy control,attention control,visualization and imagery control,motivation,positive energy,and attitude control.Loehr used the PPI to both evaluate mental skills and discern levels of mental toughness.Commonalities exist between Loehr’s ideas and the definition proposed by Jones and colleagues.2For instance,an unyielding attitude when beaten8suggests that coping strategies2following stressful events are used by an athlete.Also,the inclusion of factors dealing with control,motivation,and self-belief8is similar to remaining in control,determined,and confident.2Subsequent studies exploring the factor structure of the PPI9,21indicated a need for re-evaluation of the measure.Therefore,Golby et al.9explored the factor structure of the PPI and created an alternative measure called the PPI-A.

Gucciardi et al.22have suggested that,rather than being a measure of mental toughness,the PPI-A may be better conceptualized as an assessment of characteristics and skills consonant with mental toughness.The factor structure of the PPI-A captures many of the core elements of mental toughness as articulated by Jones et al.2and Gucciardi et al.7Specifically, it measures psychological skills that are indicative of mental toughness such as,the athlete’s resolve and commitment to their sport,sustained confidence in themselves,control of their energy and attitude,ability to regulate their thoughts and energy,and use of visualization skills in practice and competition.Although this factor structure does not contain an element specifically named control,the essence of control is captured by several factors.For example,positive cognition includes controlling negative thoughts and redirecting focus. Intuitively,controlencompassesthe ability to manage emotions and attitudes.Also,Golby et al.9have suggested that self-belief includes self-regulatory feelings,which is a part of control.

While it is presumed that athletes who perform better are more mentally tough,research has tended to tangentially address athlete performance,and findings have been somewhat equivocal.For example,using the Mental Toughness Questionnaire 48(MTQ48),Crust and Clough23reported a positive association between mental toughness and the time a weight could be suspended in a sample of undergraduate students. Golby et al.17along with Kuan and Roy18reported that rugby and Wushu athletes,respectively,at higher levels of competition were more mentally tough than their less skilled peers. However,these findings are limited because of measurement issues,the noted differences in mental toughness betweenskilled and less skilled athletes were quite narrow in scope, and a proxy(level of competition)was used to assess performance.In order to address these limitations and further examine the link between psychological skills associated with mental toughness and performance,we employed the PPI-A and a metric of basketball performance that more directly assesses athlete’s performance while competing.

Aforementioned performance findings17,18have suggested an association between skill level and mental toughness.It may be that relative skill level moderates the relationship of mental toughness to performance.In other words,the skill level of an athlete,relative to their level of competition,may influence mental toughness but not account for it.In a very basic way the skill level of athletes on teams are often classified by the terms starter and non-starter.Starters typically receive more playing time than non-starters and are assumed to have more skill and to be better performers than nonstarters.In order to extend this line of inquiry we chose to examine the construct of starting status in basketball as a moderating variable in the relationship between psychological skills indicative of mental toughness and performance.

As yet,few studies have examined gender differences in mental toughness.In a study of athletes at various levels of expertise Nicholls et al.24reported that males were more mentally tough than females based on the use of the MTQ48. Related literature also points to possible gender differences. Gender differences have been reported on self-confidence and anxiety management.25,26Specifically,females scored lower on both constructs.Collectively,this literature suggests that gender differences are likely to exist in self-reported mental toughness.

The purpose of this study was to explore the relationship between psychological skills consonant with mental toughness and collegiatebasketball performance.Furthermore,we examined possible gender and starting status differences in mental toughness.Moreover,we explored the extent to which gender,starting status,and psychological skills indicative of mental toughness predict basketball performance.It was hypothesized that:(1)there would be a positive relationship between mental toughness and performance;(2)male college basketball players would score higher on mental toughness than female college basketball players,and starters would score higher on mental toughness than non-starters;and finally,(3)gender and starting status,would be moderating variables in the relationship between psychological skills associated with mental toughness and basketball performance.

2.Methods

2.1.Participants

Participants were recruited from men’s and women’s basketball teams in National Association of Intercollegiate Athletes(NAIA)and National Collegiate Athletic Association (NCAA)Division I,II,and III universities across the USA.Of the 20 college basketball teams that were contacted,16 teams (7 men’s and 9 women’s)from 13 universities participated(9 NCAA Divisionischools,3 NCAA Division III schools,and 1 NAIA school).Although 206 participants returned surveys, nine submitted incomplete data or their basketball statistics were not reported on the university’s website.Therefore, a total of 197 participants were included in the analyses.Fiftythree percent of the sample were females(n=105)and 47% were male(n=92).Of the entire sample,37%(n=72)were considered starters,having started in at least half of the games up until that point in the season.

2.2.Data collection

After obtaining approval for the human subjects protocol from the university institutional review board,prospective team coaches were contacted either by email or phone about the possibility of participating.After the purpose of the study was explained to the coach,questionnaires were mailed to each coach.Prior to administering the questionnaire,informed consent was obtained from the athletes.Coaches were instructed to have an assistant coach,manager,or athletic trainer administer the questionnaire.Data collection occurred at coaches’convenience at a team meeting or practice session. Having individuals other than the head coaches facilitate the data collection was used as a means of promoting greater disclosure on the questionnaire by the athletes.

The person administering the questionnaire read the provided instructions verbatim and appropriate steps were taken to assure anonymity.Participants sealed their questionnaires in individual envelopes prior to returning them to the survey administrator.The questionnaires were placed in a larger envelope with pre-paid postage and returned to the researcher.Athletes identified themselves on the questionnaire by university and jersey number for purposes of obtaining performance statistics only.

2.3.Measures

2.3.1.Psychological performance inventory-alternative

The PPI-A was created by Golby et al.9through an exploration of the factor structure of the original PPI with a sample of 408 sport performers.Based on these findings a 14-item measure comprised of four factors was created(self-belief, determination,positive cognition,and visualization).Confirmatory factor analysis of this model indicated a psychometrically acceptable factor structure and reliability.9The stability of the PPI-A’s psychometric properties provides support for its use in this study.

The PPI-A uses a 5-point Likert scale ranging from 1 (almost never)to 5(almost always).Participants are instructed to select whichever response“best fits your interpretation of the item as it relates to you in sport.Be as open and honest with yourself as you can and respond to each item as it pertains to you now.”Two items were reverse scored to attenuate possible response set bias.An example of a reverse scored item was“I lose my confidence very quickly.”An example of a regularly scored item was“My self-talk during competition is negative.”Higher overall scores suggest greatermental toughness.Item responses were totaled yielding an overall mental toughness score for analysis in this study.Prior research9reported acceptable reliability for overall mental toughness(α =0.81).Overall,the PPI-A demonstrated acceptable internal reliability(α=0.81)in this study.

2.3.2.Basketball performance

Measuring objective individual performance in a multifaceted sport such as basketball is a challenge because many factors contribute to success.Sonstroem and Bernardo27recognized that a measure focused only on the offensive end of the game is not a comprehensive indicator of performance success or failure.Therefore,they developed a measure of individual basketball performance based on a variety of skills necessary for success in basketball based on shot percentage, points,rebounds,assists,steals,turnovers,and personal fouls. The PERF statistic has been widely used in prior research as an indicator of basketball performance.27—30It is computed in the following manner:

The components of the equation include:SHOT%=field goal(percentages were calculated as decimals in the equation);TP=total points in the game;REB=rebounds; AS = assists;ST = steals;PF = personalfouls; TO=turnovers;“10”=a constant to assure positive scores. Because the current study focused on overall mental toughness and not mental toughness in a single game,an adjustment was made to the original PERF measure.All of the statistics collected were current season averages to the point at which data were collected.For example,the TP statistic was an average of the player’s total points scored in each game,rather than in a single game.

In the current study,performance statistics were collected via the university website at the time each team’s questionnaires were returned.The statistics collected were averages up to that point in time.The mean numbers of games played per team and per player when statistics were collected were 26.75 and 19.80,respectively.However,the median number of games played per player was 21.From these season statistics PERF was calculated.

2.4.Data analysis

All statistical analyses were conducted using SPSS 18.0 (SPSS Inc.,Chicago,IL,USA).Missing data were excluded from the analysis listwise.Less than 5%of the data were missing.Statistical assumptions were tested prior to using parametric statistics.Data were checked for normality,skewness,and homogeneity of variance.Scatterplots and histograms revealed normal distribution of mental toughness scores across gender and within gender.Data were not significantly skewed and Levene’s test for equality of variances revealed that equal variances could be assumed.

Analysis of variance was used to examine gender and starting status differences in mental toughness.Moderated hierarchical regression analyses were used to examine the relationshipsbetween mentaltoughness,starting status, gender,and basketball performance.Models were created which tested the influence of starting status and mental toughnesson basketballperformance foreach gender separately.

3.Results

3.1.Descriptive statistics

The grand mental toughness score for all participants was 54.87±6.33(mean±SD).The scores for male(n=92)and female(n=105)participants were 56.28± 6.33 and 53.46±6.33 respectively.The grand mean for PERF scores was 11.28±2.52.The PERF means for males and females were 11.42±2.46 and 11.13±2.58 respectively.

3.2.Correlation between mental toughness and performance

A Pearson product—moment correlation examining the relationship between overall PPI-A scores and basketball performance revealed a weak,non-significant relationship,r=0.12,p=0.09.Our first hypothesis was not supported.

3.3.Comparison of mental toughness between genders and starting status

Results of a 2× 2 between-subjects analysis of variance (ANOVA),with the PPI-A scores(i.e.,mental toughness)as the dependent variable and gender(male/female)and starting status(starter/non-starter)as independent variables,revealed a significant main effect for gender,F(1,193)=10.55,p=0.001,partial η2=0.05.Males reported significantly higher PPI-A scores than females.However,there was no significant main effect for starting status,F(1,193)=0.02,p=0.89,partial η2＜0.001,indicating that starters were not more skilled at using psychological strategies that are consonant with mental toughness than non-starters.Lastly,the interaction between PPI-A scores and gender was also nonsignificant,F(1,193)=0.86,p=0.35,η2=0.004.Our second hypothesis included predictions for both gender and starting status.Given the support for gender differences and lack of support for starting status differences,our second hypothesis was partially supported.

3.4.Moderating variables in the relationship between mental toughness and basketball performance

This study’s main purpose was to examine the relationship between psychological skills indicative of mental toughness and basketball performance.Using a simple linear regression analysis,these psychological skills,as measured by the PPI-A alone,explained only 1.5%of the variance in performance,b=0.05,SEofb=0.03,t(195)=1.72,p=0.09.Preliminary analyses indicated no differences in PPI-A scores between starters and non-starters.However,the noted difference in PPI-A scores between the genders after holding starting status constant suggested that it would be logical to consider an interaction between the PPI-A scores and starting status as a predictor of basketball performance.An interaction term accounts for the multiplicative effect of two variables, allowing for conditional effects,such as moderation,to occur. In other words,under certain conditions(e.g.,when an athlete is a starter)the magnitude of the relationship between psychological skills indicative of mental toughness and performance may be stronger than others(e.g.,when the athlete is a non-starter).

Two separate moderated regression analyses were completed to accommodatetheaforementioned gender differences in PPI-A scores.The potential predictors were entered hierarchically with PPI-A scores entered first(Step 1), followed by starting status(Step 2),and finally,the interaction term was entered last(i.e.,PPI-A scores× starting status) (Step 3).Prior to construction of the moderated hierarchical regression models,scores from the PPI-A were centered. Centering is a technique used to reduce the potential for multicollinearity and can aid in interpretation of an interaction.31,32The interpretive advantage of centering is most clear when the predictor in question does not have a meaningful zero value,such as the PPI-A with its lowest possible value of 14.To center PPI-A scores,the mean PPI-A score was subtracted from the individual PPI-A scores using the appropriate mean value for each gender.With use of this new,centered PPI-A score predictor,the intercept for the moderated hierarchical regression analyses represents the mean performance score among non-starters with an average level of psychological skills indicative of mental toughness,relative to their same-gender peers.

3.4.1.Predictors of basketball performance among males

Table 1 displays the results of a parallel moderated hierarchical regression model predicting male basketball performance(n=92),with standardized(β)and unstandardized(b)coefficients,including confidence intervals.PPI-A scores explained 5.6%of the variance in male basketball performance in Step 1,F(1,90)=5.37,p=0.02.An additional 27% of variance in performance was explained by including starting status in Step 2,ΔF(1,89)=35.57,p＜0.001.Lastly,the interaction between PPI-A scores and starting status was also associated with a significantimprovementin variance explained,ΔR2=0.08,ΔF(1,88)=12.05,p=0.001.The final model in Step 3 accounted for a total of 41%of variance in male basketball performance,and supported our hypothesis that these factors would significantly predict performance,F(3,88)=74.66,p＜0.001.Fig.1A displays the predicted male college basketball performance for a range of PPI-A scores(i.e.,±3SDs)for starters and non-starters.

Table 1 Hierarchical moderated regression model predicting male(n=92)basketball performance.

3.4.2.Predictors of basketball performance among females

Table 2 displays the results of a parallel moderated hierarchicalregression modelpredicting female basketball performance(n=105).In contrast to the male basketball players,PPI-A scores(Step 1)were not related to performance among females,R2＜0.001,F(1,103)=0.01,p=0.94. Starter status,however,accounted for 34%of the variance in female basketball performance in Step 2,ΔF(1,102)=54.53,p＜ 0.001.The variance explained for female basketball players did not change when the interaction variable was included in the prediction equation(Step 3),ΔR2=0.001,ΔF(1,101)=0.03,p=0.86.The lack of a significant increase in ΔR2in Steps 1 and 3 indicate that psychological skills consonant with mental toughness were unrelated to performance for female college basketball players and that starter status was the sole significant predictor of performance in the final model,R2=0.34,F(3,101)=17.36,p＜0.001.Fig.1B displays the predicted female college basketball performance for a range of PPI-A scores for starters and non-starters.

Our third hypothesis that gender and starting status would be moderating variables in the relationship between psychological skills use and performance was supported.For males, an interaction term(PPI-A scores×starting status)explained a larger portion of the variance in basketball performance than just starting status alone.For females,starting status also explained a significant portion of the variance in basketball performance,but psychological skills indicative of mental toughness were unrelated to performance as both a main effect and an interaction with starting status.

4.Discussion

This study sought to examine the relationship between psychological skills indicative of mental toughness and performance in college basketball players.Of additional interest were the variables of gender and starting status and their influence on PPI-A scores and performance.The results of the current study demonstrate that mental toughness skills (as measured by the PPI-A),starting status,and gender were able to significantly explain a meaningful proportion of the variance in basketball performance for college basketball players.

Fig.1.Performance for male(A)and female(B)college basketball players by mental toughness and starter status.Mental toughness was related to performance for male players as both a main effect and interaction with starter status,totalR2=0.41.For female players,in contrast,starter status was the only significant predictor of performance,totalR2=0.34.Graphs are based on the final model after all three predictors were entered in mental toughness(MT),starting status(SS),and MT×SS interaction.

Overall,therelationship between PPI-A scoresand basketball performance was weak,r=0.12.This finding contradicts prior research17,18,23and leads us to believe that there are likely additional moderating variables beyond those examined here that influence under what conditions or for whom these psychological skills affect basketball performance,such as game situations,motivational dispositions of athletes,and coaching style.

In general,mental toughness seems to be most evident in situations where adversity is present.Pressure-filled game situations may warrant more mental toughness from athletes than games that are considered routine and not as meaningful. It is impossible to know whether some of the games used in this study were of greater importance or carried more weight than others.The varying levels of adversity and pressure in games may partially account for the non-significant relationship between mental toughness and basketball performance in this study.

Kuan and Roy18suggested that athletes with certain goal orientations perform better and are considered more mentally tough.Because there are likely a multitude of factors influencing mental toughness and basketball performance,future studies should consider additional variables in the relationship between mental toughness and performance.

Table 2 Hierarchical moderated regression model predicting female(n=105) basketball performance.

Another possible explanation for this finding is the use of the PPI-A as a measure of mental toughness despite some obvious limitations.Selection of an instrument to measure mental toughness is no simple task given that at least ten different instruments have been created in the past 10 years that purport to measure the construct.22The PPI-A has received criticism for its lack of concurrent validity and lack of solid a conceptual foundation,22but has been used based on its practical utility(only 14 items)and on the strength of its face validity.14However,recent support for the convergent validity of the PPI-A(with mastery approach goals and global mental toughness)22indicated that further research is necessary to explore these measurement issues.Gucciardi et al.22recently reported support for the factorial validity of the PPI-A in a group of adolescent Australian football players. Nevertheless,the PPI-A has its weaknesses,as do the other measures of mental toughness that are currently available.13

Because the PPI-A is based on the same conceptual foundation as its predecessor(the PPI),valid concerns exist about its ability to measure mentaltoughness.Ratherthan a systematic research process to develop the PPI,Loehr8used his practical experiences working with athletes.He did not offer any methodological discussion or empirical evidence to support his original seven factor model.Some have suggested that perhaps the PPI-A is essentially a product of psychological skills use by athletes.

4.1.Gender differences in mental toughness of college basketball players

Female college basketball players scored significantly lower on mental toughness than male college basketball players, which supports our second hypothesis and prior literature.24Few studies have specifically examined the effect that gender may have on mental toughness.Although using the MTQ48 as a measure of mental toughness,Nicholls et al.24noted a significant difference between genders in overall mental toughness and its subscales.Males scored higher on all measures.Some studies have noted differences between genders on self-confidence and anxiety management,25,26noting that females scored lower on both constructs.While these studies did not directly test the construct of mental toughness,self-confidence and anxiety management are similar to the PPI-A subscales of self-belief and positive cognition.

In a more general way,the current study supports broad indications of previous literature that gender differences exist in the sources and types of confidence in sport.33Because the source of confidence seems to differ between gender,females may experience confidence differently than males.While confidence is not the sole indicator of mental toughness,the less confident group(females)scored significantly lower on mental toughness than the more confident group(males). Future research should explore a possible correlation between mental toughness and confidence.

Pragmatically,the influence of the coach on psychological skills and mental toughness is an intriguing area for future research.In addition,these findings indicate a need in future research to carefully consider genderdifferences when examining mental toughness in sports.From a practical standpoint,these findings also suggest that coaches who work with female athletes may need to implement mental toughness enhancing strategies.A specific focus on self-belief elements may be warranted because it has traditionally been a lower scoring area for females.

4.2.The influence of starting status on mental toughness

The hypothesis that starters would be more mentally tough than non-starters was not supported.After accounting for the difference between genders,the male and female starters did not report greater use of psychological skills indicative of mental toughness than their respective non-starter peers.This finding contradicts prior literature suggesting that better performers are more mentally tough.16,18,22One study evaluating physiological and psychological pre-competition states discovered that both starters and non-starters reacted physiologically and psychologically similar in game preparation.34These results indicated that perhaps similar stress levels and pressures occur for starters and non-starters alike.Thus,the label of starter or non-starter may not be meaningful in psychological processes.

Another explanation for the results of the current study could be how the concept of“starter/non-starter”was assessed. The definition of starter may have been too narrow to reveal differences in skills related to mental toughness.Athletes starting 50%or more of games were classified as starters.In basketball,usually the top six to eight players contribute significant time on the court to the team’s success.In this study,the degree to which players contributed meaningfully on the court was not used in defining starters.In the aforementioned study34evaluating psychophysiological differences between starters and non-starters,this status was simply determined by the coach.However,the study occurred over a period of three games rather than the entire season.Future research should consider classifying players according to playing time or allowing the coach to label players as starters or non-starters when assessing mental toughness.

4.3.Predicting basketball performance based on mental toughness and starting status

In support of our third hypothesis,the results suggested that for both male and female college basketball players,starting status explained a significant portion of the variance in basketball performance.The relationship between psychological skills indicative of mental toughness and performance, however,was more complex.For females,mental toughness was largely unrelated to basketball performance.For males,on the other hand,the relationship between mental toughness and performance was stronger overall.Moreover,the strength of the relationship between psychological skills indicative of mental toughness and performance was most pronounced for male starters.

In total,these findings suggest that performance is maximized for male college basketball players possess the mental skills associated with mental toughness and get opportunities to meaningfully contribute on the court.Logically,those who receive the most playing time are usually starters.Those who spend the most time on the basketball court are those who are the most skilled and have captured the confidence of their coaches.In this sense then it is logical that mental toughness and good performance covary.Indeed,the results indicate that for a non-starter,mental toughness has a smaller effect on basketball performance.Some of the characteristics of mental toughness(e.g.,self-belief and determination)seem to also define the qualities of a skilled player so it may appear that being mentally tough and skilled are similar.

4.4.Limitations and future research

4.4.1.PPI-A

Lack of previous research using the PPI-A may be viewed as a limitation in the current study.Although the psychometric evaluationsofthe PPI-A have demonstrated adequate results,9,22the PPI-A has some limitations beyond previous discussion.First,several components of the questionnaire could be modified to enhance its efficacy.One of these changes is the formatting of the questionnaire.On the PPI-A, participants rated themselves on items using a scale from 1(almost never)to 5(almost always).Two negatively worded items are designed to be reverse-scored and stand out differently than the others.Instead of going from 1 to 5,the response choices on these items went from 5 to 1.This format is unlike most questionnaires and may have confused the participants.The PPI-A could also be modified by reordering the items.Items on the PPI-A are grouped together by subscales.Grouping similarly themed items may reduce the amount of reflection a participant invests in each question because of the proximity of like items.Randomly ordering the items would allow for greater response validity and reliability.

Second,the PPI-A is based on the original PPI,which was not based on a conceptualization of mental toughness norrigorous and systematic scientific research.While the original PPI and the subsequent PPI-A are intuitively appealing in regard to the manner in which mental toughness is described, there is no scientific justification for the conceptualization of the construct.

Third,the statistical method used to create the PPI-A was principle components analysis(PCA),which does not require an existing model of the construct,but searches for structure within the data that are already collected.14PCA can be a valuable data exploration technique,but the use of a confirmatory technique after a model has been supported is more appropriate for validation of a measure.

Although there is not one perfect measure of mental toughness in sport,all of the currently available measures have both strengths and weaknesses.Gucciardi22suggested that the brevity of the PPI-A allows for practical use in sport.He also explained that the generality of the PPI-A for a variety of cultures is one of its strengths.Recent literature examining other measures also described strengths and weaknesses.For example,the MTQ48 has a strong conceptual basis,but some argued that the construct it touches is too similar to hardiness.14Additionally,recent research using a confirmatory factor analysis failed to support the four C’s proposed model.15A more recently created measure of mental toughness,the SMTQ,also exhibits strengths and weaknesses.It has a number of key correlates and follows some of the construct validation principles advocated by experts.However,the SMTQ lacks some of the major components of mental toughness which have been identified in qualitative research.14

4.4.2.Objective basketball performance measure

The PERF statistic,27used in this study,includes indicators of a player’s productivity in basketball,yet needs continued examination to determine if it should be used as an accurate measure of basketball performance.The PERF statistic does not account for individual playing time.Several participants in the current study received minimal playing time(i.e.,an average of 3 min per game)and took only two shots,but made one of them,equating to a shooting percentage of 50%.This limitation skews the statistics when compared to an individual who has more playing time(i.e.,an average of 20 min per game)and shoots more often,leading to a lower shooting percentage(i.e.,33%shooting average for the season,making 25 of 75 shots).Therefore,the overall performance of players who are not playing many minutes per game may not be captured by the PERF statistic.Although it was not a part of this study,consideration for playing time will be essential in future research using an objective basketball performance measure such as the PERF statistic.

4.5.Future research

Mental toughness in athletes is one of the most popular research topics in sport psychology and interest in the topic continues to grow.The findings from this study suggest several potential directions for future research.First,the measurement of mental toughness needs to be continually investigated. Several proposed measures are available,yet none have adequatelyconceptualizedtheconstruct,beenvalidatedthrough statistical procedures,and been brief enough to accommodate use in real sport settings.Thus,continued evaluation of the PPIA is necessary.Implementing the changes previously suggested and continued scrutiny of all mental toughness measures will lead to improvement of mental toughness instrumentation.

Second,there is a need for more research on the relationship between mental toughness and performance.Our underlying assumption is that mental toughness enhances performance,but supporting research is scarce.The findings of the present study suggest that the relationship between mental toughness and performance is more complex than we initially hypothesized. This relationship should continue to be examined in light of other potential moderators and tested within contexts or sports. Additionally,there is a need for more research using objective measures of performance.Gucciardi et al.35suggested that although self-report measures of performance and other constructs are valuable,there is a need for more research using objectively measured data.Correlating subjective data with a performance variable that is measured objectively provides greater evidence for performance enhancement.

Finally,future research should more carefully examine gender differences in mental toughness.Recognition of these differences is essential for coaches and practitioners who implement techniques to foster psychological skills that are consonant with mental toughness in athletes of either gender.

In summary,this research on college basketball players focused on mental skills that are indicative of mental toughness as measured by the PPI-A,whereas previous research focused almost exclusively on world champion athletes and international competitors.Few studies have used college athletes as participants in mental toughness research.Yet, a specific need for focusing on the development of mental toughness exists in this population.Rather than highlighting athletes who have consistently demonstrated mental toughness qualities,it would be beneficial to examine the development and maintenance of mental toughness in college athletes.

Among the most significant issues facing mental toughness research is the need for a clear line of research to follow. Specifically,a single definition of mental toughness and a general conceptualization of mental toughness would be helpful in guiding future research.The foundation of any line of inquiry should be a conceptual framework.Continued development of a universal mental toughness framework will aid in the creation of more solid instrumentation,which will then allow for more valid investigations into the overall construct of mental toughness in athletes.

1.Gould D,Hodge K,Peterson K,Petlitchkoff L.Psychological foundations of coaching:similarities and differences among intercollegiate wrestling coaches.Sport Psychol1987;1:293—308.

2.Jones G,Hanton S,Connaughton D.What is this thing called mental toughness?An investigation of elite sports performers.J Appl Sport Psychol2002;14:205—18.

3.Middleton SC,Marsh HW,Martin AJ,Richards GE,Savis J,Perry C. Discovering mental toughness:a qualitative study of mental toughness in elite athletes.Available at:www.sectiononewrestling.com/discovering_ mental_toughness.pdf[accessed:08.08.2008].

4.Clough PJ,Earle K,Sewell D.Mental toughness:the concept and its measurement.In:Cockerill I,editor.Solutions in sport psychology.London:Thomson;2002.p.32—43.

5.Bull SJ,Shambrook CJ,James W,Brooks J.Towards an understanding of mental toughness in elite cricketers.JApplSportPsychol2005;17:209—27.

6.Thelwell R,Weston N,Greenlees I.Defining and understanding mental toughness within soccer.J Appl Sport Psychol2005;17:326—32.

7.GucciardiDF,GordonS,DimmockJA.Towardsanunderstandingofmental toughness in Australian football.J Appl Sport Psychol2008;20:261—81.

8.Loehr JE.Mental toughness training for sports:achieving athletic excellence.Lexington,MA:Stephen Greene Press;1986.

9.Golby J,Sheard M,van Wersch A.Evaluating the factor structure of the Psychological Performance Inventory.PerceptMotSkills2007;105:309—25.

10.Sheard M,Golby J,van Wersch A.Progress toward construct validation of the Sports Mental Toughness Questionnaire(SMTQ).Eur J Psychol Assess2009;25:186—93.

11.Gucciardi DF,Gordon S,Dimmock JA.Development and preliminary validation of a mental toughness inventory for Austrailian football.Psychol Sport Exerc2009;10:201—9.

12.Gucciardi DF,Gordon S.Development and preliminary validation of the cricket mental toughness inventory.J Sports Sci2009;27:1293—310.

13.Middleton CS,Martin AJ,Marsh H.Development and validation of the mental toughness inventory:a construct validation approach.In:Gucciardi DF,Gordon S,editors.Mental toughness in Sports:developments in theory and research.1st ed.New York,NY:Routledge;2011.p.91—107. 14.Gucciardi DF,Mallett CJ,Hanrahan SJ,Gordon S.Measuring mental toughness in sport.In:Gucciardi DF,Gordon S,editors.Mental toughness in sports:developments in theory and research.1st ed.New York,NY: Routledge;2011.p.108—32.

15.Gucciardi DF,Hanton S,Mallet CJ.Progressing measurement in mental toughness:a case example of the mental toughness questionnaire 48.Sport Exerc Perform Psychol2012;1:194—214.

16.Golby J,Sheard M.Mental toughness and hardiness at different levels of rugby league.Pers Individ Differ2004;37:933—42.

17.Golby J,Sheard M,Lavallee D.A cognitive-behavioral analysis of mental toughness in national rugby league football teams.Percept Mot Skills2003;96:455—62.

18.Kuan G,Roy J.Goal profiles,mental toughness and its influence on performance outcomes among Wushu athletes.J Sports Sci Med2007;6:28—33.

19.Rana MS.Assessment of mental toughness among high and low achievers Indian wrestlers:a comparative study.Res J Phy Educ Sports Sci2009;4:60—70.

20.Shin DS,Lee K.A comparative study of mental toughness between elite and non-elite female athletes.Kor J Sport Sci1994;6:85—102.

21.Middleton SC,Marsh HW,Martin AJ,Richards GE,Savis J,Perry C, et al.The psychological performance inventory:is the mental toughness test tough enough?Int J Sport Psychol2004;35:91—108.

22.Gucciardi DF.Measuring mental toughness in sport:a psychometric evaluation of the Psychological Performance Inventory-A and its predecessor.J Pers Assess2012;94:393—403.

23.Crust L,Clough PJ.Relationship between mental toughness and physical endurance.Percept Mot Skills2005;100:192—4.

24.Nicholls AR,Polman R,Levy AR,Backhouse SH.Mental toughness in sport:achievement level,age,gender,experience,and sport type differences.Pers Individ Differ2009;47:73—5.

25.Mahoney MJ,Gabriel TJ,Perkins TS.Psychological skills and exceptional athletic performance.Sport Psychol1987;1:181—99.

26.Meyers MC,Bourgeois AE,LeUnes A,Murray NG.Mood and psychological skills of elite and sub-elite equestrian athletes.J Sport Behav1999;22:399—409.

27.Sonstroem R,Bernardo P.Intraindividual pregame state anxiety and basketball performance:a re-examination of the Inverted-U Curve.J Sport Psychol1982;4:235—45.

28.Bray CD,Whaley DE.Team cohesion,effort,and objective individual performance of high school basketball players.SportPsychol2001;15:260—75.

29.Perreault S,Marisi DQ.A test of multidimensional anxiety theory with male wheelchair basketball players.Adapt Phys Activ Q1997;14:108—18.

30.Swain A,Jones G.Explaining performance variance:the relative contribution of intensity and direction dimensions of competitive state anxiety.Anxiety Stress Copin1996;9:1—18.

31.Aiken LS,West SG.Multiple regression:testing and interpreting interactions.Thousand Oaks,CA:Sage;1991.

32.Enders CK,Tofighi D.Centering predictor variables in cross-sectional multi-level models:a new look at an old issue.Psychol Methods2007;12:121—38.

33.Hays K,Maynard I,Thomas O,Bawden M.Sources and types of conf idence identified by world class sport performers.J Appl Sport Psychol2007;19:434—56.

34.Alix-Sy D,Le Scanff C,Filaire E.Psychophysiological responses in the pre-competition period in elite soccer players.J Sports Sci Med2008;7:446—54.

35.Gucciardi DF,Gordon S,Dimmock JA.Evaluation of a mental toughness training program for youth-aged Australian footballers:I.a quantitative analysis.J Appl Sport Psychol2009;21:307—23.

2 May 2012;revised 15 July 2012;accepted 15 August 2012

*Corresponding author.

E-mail address:aubrey.newland@utah.edu(A.Newland)

Peer review under responsibility of Shanghai University of Sport

2095-2546/$-see front matter Copyright?2012,Shanghai University of Sport.Production and hosting by Elsevier B.V.All rights reserved. http://dx.doi.org/10.1016/j.jshs.2012.09.002