AcceptedVersionCredeetal 1


1 GRIT META 1 - ANALYSIS Much Ado about Grit: A Meta - Analytic Synthesis of the Grit Literature Marcus Credé Department of Psychology, Iowa State University [email protected] Michael C. Tynan Department of Psychology, Iowa State University [email protected] Peter D. Harms School of Management, University of Alabama [email protected] ACCEPTED FOR PUBLICATION: JOURNAL OF PERSONALITY AND SOCIAL PSYCHOLOGY (Not the Paper of Record) Corresponding Author: Marcus Credé, Department of Psychology, W112 Lagomarcino Hall, 901 Stange Road, Iowa State University, Ames, 50011 - 1041, [email protected]

2 GRIT META - 2 ANALYSIS Abstract order personality trait that is highly predictive of Grit has been presented as a higher both - success and performance and distinct from other traits such as conscientiousness. This paper analytic review of the provides a meta - grit literature with a particular focus on the structure of relation between grit and performance, retention, conscientiousness, cognitive grit and the based on 5 84 ability, and demographic variables. Our results 8 8 independent effect sizes from samples representing 6 6 , 807 individuals indicate that the higher - order structure of grit is not , confirmed is that grit is only moderately correlated with performance and retention , and that grit very strongly correlated with conscientiousness . We also find that the perseverance of effort facet has significantly stronger criterion validities than the consistency of interest facet and that perseverance of effort explains variance in academic performance even after controlling for conscientiousne ss . In aggregate our results suggest that interventions designed to enhance grit may only have weak effects on performance and success , that the construct validity of grit is in question , and that the primary utility of the grit construct may lie in the pe rseverance facet . analysis Keywords: grit performance ; meta - ; ; perseverance of effort ; consistency of interest

3 GRIT META - 3 ANALYSIS Interest in non academic - cognitive variables as potential predictors and determinants of - variables such as study habits and spurred by meta analytic findings that performance has been ) , personality traits study skills (Credé & Kuncel, 2008; Robbins et al., 200 such as 4 (Porapat, 2009, 2014), test anxiety (Hembree, 1988 ; Seipp, 1991 ), adjustment conscientiousness Credé & Niehorster, 2012), emotional intelligence (Perera & DiGiacomo, 2013), ( and learning strategies (Credé & Phillips, 201 ; Richardson, Abraham & Bond, 201 2 ) , exhibit relation s with 1 s between adm relation academic performance issions test scores and that are often comparable to Hezlett et al., 2001 - . Many of these non academic performance (e.g., cognitive characteristics ) also appear to responsive to interventions. For example, meta - analytic reviews have be demonstrated that interventions can reduce anxiety (Hembree, 1988), and improve study skills (Hattie, Biggs, & Purdie , 1996 ) as well as social and personal skills (Durlak, Weissberg, & Pachan, 2010). One additional - cognitive variable that has received widespread attention and that has non been widely touted as an important predictor of success and performance is a personality trait referred to as g rit (Duckworth, Peterson, Matthews, & Kelly, 2007; Duckworth & Quinn, 2009). Grit is d - efined as “perseverance and passion for long term goals” (Duckworth, Peterson, Matthews, & Kelly, 2007 ) and as “... not just resilience in the face of failure, but also , p. 1087 having deep commitments that you remain loyal to over many years” (Duckworth as quoted in Perkins - Gough, 2013, p. 16). Duckworth et al. a rgue that g rit can help to explain why some individuals perform better than their scores on ability tests might predict and that grit was a core contributor to the success of highly accomplished individuals such as Albert Einstein. Recently, Duckworth (201 3a) has even argued that grit is as good or even a better predictor of success than - analytic findings that cognitive ability correlates cognitive ability ; a strong claim given meta

4 GRIT META - 4 ANALYSIS about ., 20 12 ; Schmidt & ρ = .50 with performance in academic and work settings (Sackett et al . Hunter, 1998) for grit as a potentially novel predictor and Despite the widespread enthusiasm there are sound empirical and theoretical reasons why a critical determinant of performance nature of the , its contribution to our understanding of reappraisal of the grit construct performance, and its general position within the nomological network may be warranted. I t is the to p resent findings from a goal of this paper - analytic synthesis of the rapidly growing meta empirical literature on g rit in order to help shed light on the nature and construct validity of grit , and to highlight potentially new areas of inquiry for grit researchers. We begin by reviewing the five core theoretical features of literature relating to rit: 1) the proposed h ierarchical structure of g rit with performance, conscientiousness, 4) g rit, 2) the relation of g 3) the distinction of g rit from the distinction of rit from cognitive ability and 5) the lack of group differences on grit . g Grit as a Hierarchical Construct Grit is typically operationalized as a higher - order construct with two lower - order facets : “perseverance of effort” and “consistency of interest” . These two facets (hereafter referred to as erseverance and p , c onsistency) respectively refer to the tendency work hard even in the face of setback and the tendency to not frequently change goals and interests. Both are thought to s contribute to success: persistence because the process of attaining mastery in a field often involves initial f ailures that the individual must persist through, and consistency because many hours of deliberate practice are normally required to achieve mastery (Ericcson, Krampe, & Tesch - Römer, 1993) . That is, individuals who either disengage their efforts in the fac e of obstacles or who constantly change their interests are unlikely to ever engage in enough distinction between the two facets deliberate practice to achieve high levels of performance. Th e

5 GRIT META ANALYSIS 5 - is reflected in the two primary self - report inventories used to measure grit: the subscales of the – Grit Scale (Duckworth et al., 2007) and the Short Grit Scale (Duckworth & Quinn, 2009) both of which can be found on Angela Duckworth’s homepage ( h ttps:// ) . Although some researchers examine p erseverance and c onsistency as two separate constructs , most research on grit only report s findings at the level of the overall rit score. g First, ) This practice appears to be informed by two factors. in Duckworth et al. (2007 their discussion of the two facets state that “...neither factor was consistently more predictive of outcomes than the other, and in most cases, the two together were more predictive than either alone” (p. 1091) . Second, Duckworth and Quinn (2009) reported examin ing th e theoretical higher - order factor structure of grit using confirmatory factor analysis (CFA) , and claim ed order model support for the higher - order structure b ased on thei r finding that the higher - ( comprised of two first order factors and one second - order factor ) exhibits significantly better fit - - factor model. However, t h e reported analysis is problematic because a model with than a single one second - order factor and two first - order factor s is not identified at the higher - order level ( ); this means that no unique loadings onto the higher - order factor can be computed Kline, 2011 kworth and . However, it does not appear that Duc without additional constraints being imposed equality constraints on Quinn imposed the type of the loadings of the first - order factors onto the higher - achieve identification because they report non - order factor that would be required to identical - order factors onto th e second - order factor. Importantly, even if an loadings of the first equality constrain t had been imposed at the higher - order level the resultant second - order model would have exhibited identical fit to a model with two correlated first - order factors and no the That is, a model in which second - order factor (see Cred é & Harms, 2015 for a discussion).

6 GRIT META ANALYSIS 6 - c onsistency facets p are simply two correlated ersistence and constructs would exhibit identical fit to the higher - order model. Interestingly, Duckworth and colleagues had tested the fit of such a two - factor model in an earlier paper (Duckworth et al., 2007) , and reported relatively poor fit for the model (i.e., CFI = .83, RMSEA = .11). higher - The CFA strategy for determining whether a particularly order grit construct exists is therefore not meaningful because standard indexes of model fit cannot be used to distinguish between a higher - order model and a model with two correlated factors . A potentially more useful approach would involve examining the correlation between the order construct is - two theoretical facets of grit; high correlations would suggest that a higher plausible. However, e mpirical estimates of the strength of this relation exhibit substantial some reporting correlations that are close to zero (e.g., Chang, 2014; Datu, variability, with Valdez, & King; 2015; Jordan, Gabriel, Teasley, Walker, 2015) , while others have & Schraeder, reported very strong correlations (e.g., Arslan, Akin, & Ç î temel , 2013; Meriac, Slifk a, & LaBat, 2015). A meta - analytic synthesis will help to establish a population estimate of the correlation between the two facets and thereby allow readers to make a more informed judgment about - whether or not grit exhibits the hypothesized higher structure order . This will in turn help to determine whether the practice of simply summing across the perseverance and persistence items rth et al., 2007) is reasonable, or to compute an overall grit score (as recommended by Duckwo whether the two facets should be considered separately. Grit as a Predictor of Success and Performance Proponents of g rit as a predictor of performance have argued that between - person differences in g rit can help to explain why two individuals with the same level of ability in a , particular domain are often observed to perform at substantially different levels. Specifically

7 GRIT META - 7 ANALYSIS individuals with high levels of g rit are thought to be able to better utilize their capabilities - because they a failures and term goals and less discouraged by re less distracted by short the that are commonly encountered in many performance domains Indeed, Duckworth et setbacks . al. (2007) that the importance of grit for success had long been noted by prior research described ( e.g., Howe, 1999 ) . Arguments for the importance of grit into highly accomplished individuals are also in line with work on the development of expertise that has highlighted the importance of s son, - R ӧ mer, 1993; Krampe & Krampe, & Tesch sustained deliberative practice (e.g., Eric . Indeed, recent work by Duckworth, Kirby, Tsukayama, Berstein and Ericsson Ericcson, 1996) (2011) has explicitly tied g rit to success in spelling bees via the mediating mechanism of That is, indi deliberative practice. viduals who are high on grit are more likely to engage in the amount of deliberative practice that is required to achieve expertise. a number of At the same time there are ly plausible moderators of the g rit - theoretical - p erformance rel ation t hat suggest that the relation may not be strong in all - or even most settings. First, the grit - may be moderated by the nature of the performance performance relation domain. Specifically, high levels of grit may be most useful when the task is difficult but well defined; that is, high levels of sustained effort and deliberative practice are required to succeed (see MacNamara, and the manner in which performance is to be attained is relatively clear . Thus Hambrick & Oswald, 2014) grit may be an excellent predictor of an individual’s ability to , well - defined academic tasks, but be less well complete military basic training or succeed in related to performance on tasks that are very easy (thus not requiring grit) or performance on that are tasks novel and il l - defined and th at therefore require both creativity and the willingness to abandon unsuccessful strategies (i.e., tasks on which grit may be counterproductive) . Second, individual differences such as ability the grit - performance relation may be moderated by other

8 GRIT META - 8 ANALYSIS & Phillips, 2011) ( Credé cognition . That is, high levels of grit may not necessarily be and meta - general potential or ability to succeed in a domain and adaptive unless it is accompanied by the the ability to engage i n the type of reflection and self monitoring that the self - regulated learning - - literature (e.g., Zimmerman, 1990) and the social cognitive view of learning (e.g., Ryan & minants Pintrich, 1997; Zimmerman, 1994, see also Kohn, 2014) has identified as important deter . of learning and performance adets who lack some minimum level of physical For example, c ability to pass the highly strenuous test s of physical ability in a military academy are unlikely to rit. Similarly, a cadet who is benefit substantially from g unable to recognize that a particular approach to studying for class material is not working is to perform well in academic unlikely courses. - performance may be moderated by the level of grit itself. Very Third, the grit relation high levels of grit may become dysfunctional if they reduc e the likelihood of help - seeking behaviors that have themselves been linked to performance (e.g., Karabenick, 2003) or if they increase the likeli hood that an individual persists too long in attempting to solve a problem that is particularly difficult rather than spending their time on other, more solvable problems (see Lucas, Gratch, Cheng, & Marsella, 2015) ventions . This would, in turn, suggest that inter designed to enhance grit levels may not benefit all individual equally. A recent large - scale evaluation of the impact of resiliency interventions (Paunesku et al., 2015) found that such interventions benefit primarily those students who were most at risk of dropping out of high schools and provided less benefit for other students. Meta - analytic evidence suggests similar effects for resilience interventions in working populations (Vanhove, Herian, Perez, Harms, & may also exist for grit. Lester, 2015 ). A similar non - linear relation

9 GRIT META ANALYSIS 9 - A more nuanced conceptualization of grit’s contribution to performance also seems re considering the biographi cal details of some of the highly accomplished warranted when scientific figures referenced by Duckworth et al. (2007) to highlight the importance of grit . For Einstein persisted for many years in his attempt to develop the example, i t is true that Albert field equations that represent the , but it is General Theory of Relativity mathematical description of also true that Einstein persisted for years in pursuing an avenue of investigation that was based . Einstein also on an earlier mathematical error appears to have only resolved some of the challenging mathematical obstacles after consulting with the mathematicians Marcel Grossman and (later) David Hilbert (Earman & Glymour, 1978) who, according to some accounts (e.g., Parker, 2004), almost scooped Einstein in the development of mathematical framework for the General Theory of Relativity because of Einstein’s delay in seeking assistance. That is, and Einstein only “succeeded” in this persistence in this narrow case almost resulted in “ failure ” particular endeavor once he recognized his mathematical limitations and sought the help of more accomplishe d mathematicians. The possibility that the grit - performance relation is not uniformly strong is also strongly ome have found that g suggested by an examination of the empirical literature. S rit scores are relatively strongly related to success as suggested by the initial findings by Duckworth and colleagues (e.g., Strayhorn, 2013), but many others (e.g., Chang, 2014; Cross, 2013; Davidson, 2014; Hogan, 2014; Sheehan, 2014) have failed to find strong relation s between grit scores and indicators of success. T his is particularly the case for studies examining academic success. Indeed, many of the reported s between grit and academic success are weaker or equal to relation the relation of ρ = .23 that has been reported between conscientiousness and = .2 1 and ρ - analytic reviews (e.g., Porapat, 2009; ac ademic performance in two recent large - scale meta

10 GRIT META ANALYSIS 10 - Richardson et al., 2013). Meta analytic synthesis will not only help to clarify the strength of the - relation between grit and success but will also help to clarify whe ther the observed variability in relation s is simply a function of sampling error and other study artifacts (e.g., differences in the r , or if this variability reflects the presence of eliability in the measurement of variables) meaningful moderators. A meta - analytic summary should also help to address possible confusion among readers of the grit literature about the ability of grit to predict the successful completion of rigorous programs. This confusion may have arisen because the authors of both of the foundational papers appear to confuse odds ratios with (Duckworth et al., 2007; Duckworth & Quinn, 2009) probabilities in their discussion of logistic regression results resulting in incorrect inferences , about the size of observed effects. This misunders tanding may have led readers to infer a much greater predictive power for grit scores than is warranted. For example, Duckworth and Quinn (2 009) discussed the ability of grit scores to predict the successful completion a summer program for cadets from the United States Military Academy at West Point and interpret an odds ratio of 1.99 to mean that “Cadets who scored a standard deviation higher than average on the Grit - S were 99% more likely to complete summer training” (p. 171). This interpretation is incor rect b A relatively small ecause approximately 94% of all cadets successfully completed the program. increase in the completion rate from, say, 95% to 97.5% associated with a one point increase in grit scores would, of course, represent an odds ratio of 2 , but this is only a 2.6% increase in the likelihood of completing the program. A meta - analytic synthesis will also help to establish whether either of the two facets of grit exhibit higher levels of criterion validity than the other or whether the two are l argely with important outcomes as suggested by Duckworth and colleagues equivalent in their relation

11 GRIT META - 11 ANALYSIS (2007, p. 1091). This might, in turn, change the manner in which grit scores are presented and s by Duckworth et al. and examine interpreted. Many grit researchers follow the recommendation only an overall grit score . However, facets are often better predictors than broad traits (e.g., , and substantial differences between the grit facets to predict Paunonen & Ashton, 2001) important criteria might suggest that this scoring strategy should be revisited. - analytic synthesis of the strength of the relation between g A meta rit and success will also help to inform judgments about whether interventions designed t o enhance g rit are likely to have an impact on performance. Initial reports of the high predictive validity of g rit scores and g their relative independence from indicators of cognitive ability, combined with claims that rit can be taught (Perkins Gough, 20 13), has resulted in some schools implementing interventions - designed to increase students’ level s of g rit. For example, The Knowledge is Power Program (KIPP) network of public charter schools is training its teachers to foster grit in their pupils (Shech tman, DeBarger, Dornsife, Rosier, & Yarnall, 2013), while many school districts across the US are reportedly considering integrating the teaching of g rit into curricula (Cohen, 2015). Grit was even highlighted as a promising focus of school interventions i n a US Department of Education report (Shechtman et al., 2013). The time and resources that are likely to be devoted to rit - based interventions in schools are likely to be non - trivial and should therefore only be based g on the best available knowledge abou t the role of grit in predicting and determining performance. The Distinction of Grit from Conscientiousness An interest in what Duckworth and colleagues refer to as grit, perseverance, and consistency is not new to psychology. Studies of attributes such as will power, tenacity, years (see determination, persistence of motives, and volitional perseveration date back over 80 Ryans, 1939 for an early review). More recently, researchers have investigated a variety of other

12 GRIT META ANALYSIS 12 - trait like constructs that are characterized by persistence and consistency including proactivity - (e.g., Crant, 1995), persistence (e.g., De Fruyt , Van de Wiele, & Van Heeringen, 2000), industriousness ( e.g., Eisenberger, 1992; Jackson & Paunonen, & Tremblay, 2000), need for achievement (McClelland, 1985), conscientiousness, and some of the facets of conscientiousness such as industriousness, self - c ontrol, and order (Roberts, Chernyshenko, Stark & Goldberg, 2005). The conceptual similarities between these constructs and grit raises the possibility that the victim to what Kelley (1927) referred to as the “jangle proponents of grit may have fallen fallacy” – the belief that two things are different simply because they have different names. The contribution of grit to the psychological literature would, of course, be severely limited if the construct was simply a case of “old wine in new bottles” and it would therefore appear to be important to formally establish the discriminant validity of grit relative to these related constructs . There has however been almost no empirical investigation of the discr iminant using the types of methodologies (e.g., Mulitrait - validity of grit from these other constructs multimethod matrices, confirmatory factor analyses) commonly employed to determine espite the fact that grit has been explicitly presented as a construct that is discriminant validity, d distinct from these previously examined constructs – particularly conscientiousness and need for achievement (Duckworth et al., 2007; Duckworth & Quinn, 2009; Perkins - Gough, 2013). The bivariate relation of grit with consci entiousness has often been reported by researchers but even conscientiousness for ( and its facets ) there are both empirical and theoretical reasons for suspecting that the overlap with grit may be stronger than is widely assumed. For example, the definitio n of grit as “perseverance and passion for long - term goals” (Duckworth et al., 2007) is highly similar to the definitions given by Costa and McCrae ( 1992 ) - discipline facet (“capacity to begin tasks and follow through to completion despite for the self

13 GRIT META - 13 ANALYSIS bore (“need for dom or distractions”) and the achievement striving facet of conscientiousness personal achievement and sense of direction”). This theoretical similarity is also reflected in the item Duckwort h et al. (2007) perseverance considerable similarity in the s that are found in the and items in widely used inventories of conscientiousness subscale the such as those provided by IPIP, Goldberg, 1999). For example, perseverance items International Personality Item Pool ( very similar to IPIP items used such as “I finish whatever I begin” and “I am a hard worker” are to measure achievement striving such as “I carry out my plans” and “I work hard”. Items from the consistency scale overlap less strongly with items from conscientiousness inventories and are IPIP to instead more similar to items used to measure the adventurousness facet of openness experience and the IPIP measure of planfulness that is modeled after the Achievement via scale found in the C alifornia Personality Inventory (Gough, 199 6 ) . Conformance A cursory examination of the grit literature also empirical that the grit – suggests W . conscientiousness relation may be much stronger than is commonly assumed hile some (e.g., Cooper, 201 , 4) have presented evidence the grit is largely distinct from conscientiousness numerous others have reported correlations between grit and conscientiousness that approach unity when correcting the observed correlations for unreliability. Reed, Pritschet a nd Cutton (2012), for example, report a correlation of ρ = .92 based on 1165 college students, Engel (2013) reports a correlation of = .95 based on a smaller sample of 88, and Meriac, Slifka, and LaBat ρ (2015) report a disattenuated correlation of ρ = .98 based on a sample of 322 students. Even Duckworth et al. (2007, 2009) report correlations between conscientiousness and grit scores that = rise to ρ = .97 (N 308) after 1 , 554); ρ = .90 (N = 706), ρ = .83 (N = 190), and ρ = .80 (N = 1 , rit scores. High correlations correcting for the un reliability of both c onscientiousness scores and g such as these have led some (e.g., MacCann & Roberts, 2010) to suggest that grit should be

14 GRIT META - 14 ANALYSIS considered a facet of conscientiousness a position that seems theoretically plausib le when – control or self - discipline facet of conscientiousness focus considering that both grit and the self - - term gain for long - term goals (see Costa & McCrae, 1992; Roberts, on the deferment of short Bogg, Walton, Chernyshenko, & Stark, 2004). d correlations between grit and High observe conscientiousness are also of concern when considering that a concurrent assessment of the same personality trait using different scales typically yields much lower correlations = .50 of around r Pace & Brannick, 2010 ; Miller, Price & Campbell, 2012 ). Meta - analytic synthesis of the (e.g., conscientiousness will help clarify whether the strength of the literature on the grit - relation relation is such that grit might be a case of the “old wine i n new bottles” phenomenon . The Distinction of Grit from Cognitive Ability Grit is typically described as being largely distinct from cognitive ability (Duckworth et al., 2007; Duckworth & Quinn, 2009; Perkins - Gough, 2013) , although Duckworth (2013a, relation 2013b) has also suggested a negative between grit and cognitive ability in noting that “... gritty people, on average, tend to be slightly less talented” . This distinction, if correct, suggests that g rit might explain unique variance in performance over and above the substantial varian ce in performance accounted for by cognitive ability (e.g., Kuncel, Hezlett, Ones, 2004). Further, a finding that grit is largely orthogonal from general cognitive ability would also suggest that interventions designed to enhance grit levels might result i n substantial increases in performance. Primary research findings have found broad support for the assertion that cognitive ability and grit are largely distinct. Many of the studies in this domain utilize admissions test scores as a proxy for cognitive ab ility test scores but this approach seems reasonable when considering that c ognitive ability tests administered in research settings have questionable because many test takers will not be motivated to exert maximal effort on ability tests in validity

15 GRIT META - 15 ANALYSIS a low ber, & Stouthamer - Loeber, 2011). Most stake setting (Duckworth, Quinn, Lynam, Loe commonly used college admissions test scores are highly correlated with cognitive ability test nce (i.e., high scores (e.g., Frey & Dotterman, 2004) and are also taken under maximal performa stakes) conditions and are therefore likely to represent a sound indicator of cognitive ability . relation between grit scores and admissions test scores have largely Research examining the relation s (e.g., Chang, 2014; Duckw orth et al., 2007; Eskreis - Winkler, found very weak Kelly, Matthews, & Bartone, 2014). Shulman, Duckworth, & Beal, 2014; Group Differences in Grit Scores Concerns about the reliance on cognitive ability tests for the prediction of success and performance have often revolved around the persistent finding that groups exhibit non - trivial mean score differences on such tests (e.g., Camara & Schmidt, 1999; Davis et al., 2013). A finding that grit exhibits smaller differences between groups is likely to make the construct more stakes - attractive in settings where scores are used for selection purposes or for making other high is reduced decisions because the likelihood of adv erse impact on legally - protected group s . Prior research suggests only one type of group difference. Duckworth et al. suggested that grit might increase with age – a phenomenon that has also been observed for conscientiousness (Roberts, Walton, & Viechtbauer, 2006) – but prior findings from the personality literature (e.g., Costa, Terracciano, & McCrae, 2001; Foldes, Duehr, & Ones, 2008) suggest that differences across lity measures. ethnicities and gender are likely to be more modest than those observed for abi Empirical findings on the direction and strength of the relation between grit scores and demographic variables such as age (e.g., Engel, 2013, Eskreis - Winkler , Duckworth, Shulman & (e.g., Chang, 2014, Beal. , 2014) , gender (e.g., Allen, 2014; Davidson, 2014) , and ethnicity

16 GRIT META - 16 ANALYSIS Eskreis have been mixed and a meta - analytic synth esis will help to clarify Winkler, 2014) - whether average grit scores are largely similar across groups. The General Position of Grit within the Nomological Network been not only been related to performance, cognitive ability, and Grit has variables reflecting either states or traits. conscientiousness but also to a wide array of other et al., Eskreis - Winkler et al., 2014), optimism (e.g., These include Big Five traits (e.g., Lovering 2015) , psychological well - being (e.g., McCann & Roberts, 2010); suicide ideation (Blalock, Young, & Kleinman, 2015), intended persistence in academic programs (e.g., Bowman et al., relation 2015), and life satisfaction (e.g., Samson et al., 2011 ) . A meta - a nalytic synthesis of the of grit with these other variables further help to clarify the general position of grit within the will broader nomological network. Interpreting Criterion - Related Validity Estimates Cohen’s (1988) guidelines for what constitutes small (r = .10) , medium (r = .30), and large (r = .50) effect sizes are widely used to make describe the size of the relation between a However, t predictor variable and a criterion variable . he American Psychological Association (Wilkinso n & Task Force on Statistical Inference, 1999) has also encouraged researchers to place effect sizes in a practical and theoretical context. To this end we briefly discuss meta - analytic estimates of the criterion - related validity of various widely studied predictors of academic performance and retention. We use these meta - analytic estimates to inform our assessment of the relative ability of grit to predict academic performance and retention , although it is also – important to note that even a relatively low criterion validity can be practically very important

17 GRIT META - 17 ANALYSIS especially when the predictor provides information about the criterion that is not provided by other predictors and when the criterion is important. analyses of predictors of academic perfor mance have identified two variables - Prior meta that correlate at approximately = .50 with academic performance in college : 1) indicators of ρ 2) prior academic performance such as high cognitive ability such as scores on the SAT and (Sackett et al., 2012). Other predictors that correlate approximately at ρ = .40 with school GPA academic performance include study skills and study habits (Credé & Kuncel, 2008), 4) academic adjustment (Credé & Niehorster, 2012), - efficacy (Robbins et al., 2004) academic self 6 ) class and attendance (Credé et al., 2010). These appear to be the best known predictors of academic performance in college. Other variables that meta - analyses have shown to exhibit = .20 with academi relation weaker but practically still very meaningful c s of around ρ performance include specific learning strategies (Credé & Phillips, 2011), 2) emotional : 1) intelligence (Perera & DiGiacomo, 2013), 3) conscientiousness (Porapat, 2009), and 4) test anxiety (Hembree, 1988). The ability to predict retention is typically weaker. Meta - analytic findings indicate that ρ the best predictors are: academic self - efficacy ( = .36) and academic - related skills ( ρ = .37, Robbins et al., 2004), institutional attachment ( = .29) and social adjustment ( ρ = .25, Credé & ρ Niehorster, 2012) , high school grades ( ρ = .20, Robbins, Allen, Casillas, Peterson, & Le, 2006), and SAT and ACT scores ( ρ = .17, Mattern & Patterson, 2009; Robbins et al., 2006) . Method Search Strategy

18 GRIT META - 18 ANALYSIS Potential sources for inclusion in our review were identified using keyword, abstract, and title searches of the PsycINFO, Dissertations Abstracts, and ERIC databases using the search . These search results were This yielded a total of 778 potential data sources. term “grit” ists of identified sources supplemented by an examination of the referen ce l . We also examined oogle search engine and the search term the first 500 search results of the internet using the G to identify additional unpublished sources “grit” . Potential sources for inclusion of information were first screen of the source and all possible sources were ed by examining the abstract and title or exclusion then examined more closely to determine if the reported data met the inclusion criteria. Inclusion and Exclusion Criteria Sources were included in our review i f they reported on the Pearson correlation between – scores on any of the Duckworth et al. (2007 , 2009) measures of grit and other variables or if they reported information that could be used to estimate the size of such a correlation (e.g., dard deviations for two criteria groups) means and stan . The year of publication, source of the material, and country of origin of the data were not used to exclude any sources, although non - also English sources were excluded. Sources were lations for excluded if they reported corre individuals below a middle school age because personality is still highly fluid at earlier ages , and because prior meta - analyses on personality as a predictor of achievement (Porapat, 2014) found that the strength of the at younger ages was very different to the relation at older ages. relation We also one study that only reported on significant correlations (and excluded non - significant correlations) because the inclusion of this data would have resulted in an upwardly biased eff ect size estimate. When studies did not report data in a format that could be coded and when these studies had been published in the last five years we attempted to contact the authors to request

19 GRIT META - 19 ANALYSIS the necessary information. epresenting data from 88 unique Data from a total of 73 studies r samples and 66,807 individuals was ultimately included in the analyses. Coding Procedure All articles were coded by using a systematic coding procedure , one of two of the authors articles for meta - analytic analyses . An accuracy check which has extensive experience coding 98. 8 % agreement in coding across the four most important coding categories . 60% of revealed (e.g., incorrect coding of an effect size) were errors of commission errors and 40% were coding . A (e.g., an effect size that could have been computed was not coded) errors of omi ssion ll disagreements were resolved via discussion. Each correlation that was included in our review was described using ten coding categories: 1) the size of the correlation, 2) the sample size, 3) the reliability of the grit scores, 4) the reliability of the correlate scores, 5) the name of the correlate, 6) the source of grit ratings (self - ratings or other - ratings), 7) the source of the correlate data (self - ratings, other - ratings, records), 8) whether the grit scores reflected overall grit or either of the tw o facets: c onsistency and perseverance , 9) the source of the publication (peer - reviewed reviewed), and 10) the year of publication. versus not peer - Self - reported grades are very highly correlated with actual grades (Kuncel, Credé & Thomas, 2005) and we the refore included correlations with grades irrespective of whether the grade information was based on self reports - or were obtained from records. When sources reported correlations involving both self - reported grades and grades obtained from records we coded the correlations involving grades obtained from records. A summary of the coding of the most important variables is included in the Appendix . Transformations

20 GRIT META - 20 ANALYSIS Our coding process involved transformations of data. First, we used formulas three presented by Hunter and Schmidt (2004) to calculate estimates of the correlation between grit d and correlate variables when the original sources ha artificially dichotomized the correlate (e.g., pr esenting grit scores for “low” and “high” scoring students) . The artificial variable dichotomization of data results in downwardly biased estimates of the population correlation if such a correction is not made. Second, we mpbell and used the formula presented by Ghiselli, Ca calculate composite correlations when the original source only presented Zedeck (1981) to . For example, correlations involving the facets of grit and/or facets of the correlate variable (2015), presents correlations among the two facets of grit, Fall GPA, and Spring Bowman et al. six correlations among these four variables were used to arrive at an estimate of the GPA; the correlation between overall grit and overall GPA. Mosier reliability estimates f or composite - Finally, we computed point variables ( Mosier, 1943 ) were also calculated whenever possible. biserial correlations between retention and grit when the mean and standard deviation of grit - retain ed group. scores were reported for both the retained and non Criterion Categories Grit researchers have examined the relation of grit with a wide variety of indicators of success. Meta - analytic synthesis requires a grouping of similar criteria with each other but because success criteria can be grouped in a wide variety of way s we present meta - analytic estimates for ten criterio n categories. First, we present separate meta - analytic estimates of the Second, we relation between grit and high school GPA, college GPA, and post - graduate GPA. aggregate these into a broader general GPA criterion category (i.e., GPA across all three educ ational levels) . For this general GPA criterion category we relied on the correlations for the when authors reported correlations involving both high school GPA more recent college GPA

21 GRIT META - 21 ANALYSIS and college GPA. We then also combined this general GPA category furthe r with correlations criterion category. cademic Performance involving grades in individual courses to form an A Third, we present meta analytic estimates of the point - biserial correlation between grit and - an academic or military setting (e.g., degree retention. Most studies examining retention do so in completion, completion of basic training) but one study also examined marital status as an indicator of retention (i.e., staying married versus getting a divorce or separating) and we , - a therefore present meta nalytic estimates both with and without the study on marital retention. We also present findings for a criterion representing a collection of non - academic criteria comprised of performance in spelling bees, military settings, and athletics. Finally, we als o - between grit and the intent to persist in both present meta analytic estimates of the relation college and with a particular employer. State , Trait , and Demographic Categories Grit researchers have examined the relation between grit and a variety of other variables that represent both relatively stable personality traits such as the Big Five personality traits and cognitive ability, and variables that have a strong mood and emotion component and that could escribed as falling somewhere along the state therefore be d - trait continuum (e.g., happiness, depression, positive affect). We use the descriptors of these variables as given in our source puted meta - articles to group these various state and trait variables into categories and com analytic estimates for those categories for which at least three effect sizes were reported. We also report meta - analytic estimates of the relation between grit and four demographic variables (gender, age, year in school and ethnicity). cal Method Statisti

22 GRIT META - 22 ANALYSIS We used the Hunter and Schmidt (2004) - analytic method based on a interactive meta effects model to arrive at population estimates of the size of the s between grit - relation random used and other variables. t o compute meta - analytic The Schmidt and Le (2004) software was relation s involving grit estimates of the and we corrected for unreliability in the measurement of the dependent variable and unreliability in the measurement of the independent variable . Grit level of range restriction in many samples but the absence of scores are likely to be exhibit some normative data on grit scores and variance in how grit is measured (e.g., number of items, number of response options) did not allow us to correct for range restriction. u Corrections for bility. In order to correct for the attenuating effect of nrelia measurement error on the size of the observed correlations we constructed reliability artifact distributions from the reliability information that was described in the included studies. These ility distributions are described in Table . The included studies did not report information reliab 1 on the reliability of grades but in order to - to - apples comparison with the facilitate an apples recent meta - analytic summary of the relation between conscientious ness and academic performance by Porapat (2009) we corrected for the unreliability of grades using a distribution of reliability estimates for that was largely similar to the reliability estimates used by Porapat rds we used the average of reliability estimates (2009). For GPA information taken from reco (alpha = .90) for all courses across four years as reported by Bacon and Bean (2006) while f or self - reported GPA we used the operational validity estimates of .90 for college GPA and .82 for high school GPA as reported in the meta - analysis by Kuncel, Credé and Thomas (2005). [INSERT TABLE 1 HERE] Other Analytic Decisions. For the meta - analysis involving the retention criteria we took a dual between grit and analytical approach. More than half of the studies examining the relation

23 GRIT META ANALYSIS 23 - retention did not report means and standard deviations for those individuals who dropped out of a program and for those individual who stayed in the program. Instead these authors reported , but odds ratios cannot be di odds ratios - biserial correlation when rectly transformed into a point the independent variable is treated as a continuous variable. We therefore present two meta - analytic estimates for the grit - retention relation . We provide one estimate based purely on those studies - biserial correlation, and then that report data that could be transformed into a point provide another estimate that includes correlation values computed by taking the root of the 2 fr values reported for those studies that reported odds ratios Nagelkerke R om bivariate logistic 2 regression models . Nagelkerke R values tend to be too high as an estimate of the strength of the but we include these values in order to provide readers with an bivariate relation (Allison, 2014) estimate based on the most complete d ata remind readers that this estimate is likely to be . We . upwardly biased - analytic estimates of the relation s of overall grit (or the two grit facets) We present meta with other variables whenever at least three studies reported on such a relation . We summarize our meta - analytic findings for each relation using six pieces of information: 1) k refers to the number of studies used to compute the estimates, 2) N refers to the total sample size used to compute the estimate, 3) r refers to the sample - s ize weighted average observed correlation, 4) obs refers to the estimate of the population correlation, 5) SD refers to the estimate of the standard ρ ρ deviation of effect sizes after taking into account the variability that is due to sampling error and differ and 6 ) 10%CV and 90%CV ences in the reliability of measurement between studies, represent the upper and lower bounds of the 80% credibility interval. The width of the credibility interval is indicative of the presence of undetected moderators. That is, wid e credibility intervals . indicate that the correlation can be expected to vary widely across settings

24 GRIT META ANALYSIS 24 - In order to examine whether grit scores explain incremental variance in academic usness we performance outcomes over and above the variance explained by conscientio - analytic intercorrelations matrix between grit, conscientiousness, and constructed a full meta conscientiousness correlations of ρ = .21 for high school academic performance by importing the GPA and ρ = .23 for college GPA as reported by Porapat (2009) , and using the average of these correlation matri ( = .2 2) for overall academic performance . Th ese ρ ces were then used to perform hierarchical regression analyses based on the harmonic mean of sample sizes. Results Before proceeding with our primary analyses we first examined the data from the studies included in this meta analytic review for publication and source bias. - Publication and Source Bias We examine the possibility that the literature included in this meta - analytic review represents a biased sample of the research on grit in two ways. First, in order to examine whether the published and unpublished literature report grit - performance relation s of different magnitudes we report separate meta analytic estimates based on those studies that were - p - reviewed journals and all other studies (e.g., dissertations, conference ublished in peer presentations). Because of the limited number of total studies that report correlations at the facet level we only perform this analysis for overall grit. Results are presented in Table 2 . In general the evidence for source bias is weak with only small differences in correlations reported for overall academic performance, the overall GPA criterion, and undergraduate GPA. None of the differences in correlations were sig nificant at alpha = .05. HERE] [INSERT TABLE 2

25 GRIT META - 25 ANALYSIS Second, we use Egger’s Test of funnel - plot asymmetry (Eggers, Smith, Schneider, & analysis for the exclusion of - Minder, 1997) to examine whether there is evidence in the meta Studies with small sample sizes that find weak effects may not small studies with weak effects. be published and not found via literature searches and their exclusion may result in an overestimate of the strength of an effect. Egger’s Test regresses the standard normal deviate of the effe ct size for each sample onto the precision of the effect size estimate. The intercept of the statistically significant regression line provides information about the size of any asymmetry; negative intercepts suggesting that small studies with weak effects may have been suppressed from the literature. Because of the relatively small number of studies that examined any one we only perform a single test of asymmetry for the relation based on the largest number relation relation of studies: the between grit an d overall academic performance. For this relation Egger’s not Test indicated no significant asymmetry with the intercept being negative but weak and - . 15 , p=. significantly different from zero (a= ). 85 The absence of evidence for strong source bias and publication bias suggests that our meta - analytic estimates are unlikely to be substantially biased in either a positive or negative rit g direction by missing studies . We therefore present meta - analytic estimates of the relation of in Tables with criteria, trait variables, and demographic variables state and 3 - 5 . [INSERT TABLES 3 – 5 ABOUT HERE] Relation between Perseverance and Consistency between perseverance and consistency (k Our meta - analytic estimate of the relation = 1 7 , .6 N = 2 2 , 048 , ρ = 0 , SD = . 21 ), indicates a generally strong relation although the width of the ρ As an is substantially moderated. credibility interval suggests that the strength of this relation

26 GRIT META - 26 ANALYSIS exploratory follow relation observed for the two different gri t - up analysis we compared the , N = 11 (k = scales and found a stronger relation when researchers relied on the short grit scale , ρ = .6 6 , SD ) than when the original grit scale was used (k = .1 5 18, = 996 , N = 3 , 052 , ρ = . 27 , 6 ρ = .1 7 ) SD . ρ Relation s with Criteria relation with overall academic performance of ρ Overall grit exhibits a 8 (k = 3 9 , N = = .1 13 141 , SD ). = .1 1 ) and ρ = .1 7 with the overall GPA criterion (k = 3 7 , N = 1 2 , 601 , SD 10 , . = ρ ρ approximately strongly related to college Among the academic performance criteria grit was as ρ = .17, SD = .10) as it was to high school GPA (k = GPA (k = 30, N = 10,526, 17 , N = 6 , 364 , ρ ρ ) that both facets Duckworth et al. (2007 = . 16 , SD Contrary to early assertions by = . 14 ). ρ predicted success outcomes equally well, t he perseverance facet of grit exhibited much stronger s with all academic performance criteria than the consistency facet. For example, relation ρ = .2 6 (k = 11 perseverance correlated at = 5 , 221 , SD .12) with overall academic = , N ρ performance while cons istency correlated at only ρ = . 10 (k = 11 , N = 5 , 221 , SD ). A = .0 2 ρ comparison of the correlations of perseverance and consistency with four academic performance criteria using the procedure for comparing correlated correlation coefficients described by Meng, Rosenthal, and Rubin (1992) showed that the correlations di ffered significantly (p<.001) in all four cases. = .12 when Grit c orrelated with retention at ρ = the marital success study is included (k ) , N = 17, 525 , SD cluded = . 09 1 , at ρ = .1 8 (k = 10 , N = 11, 163 SD = .03) when it was ex 1 , and ρ ρ = .16 (k = at, ρ 5 , N = 2, 705 , SD if the upwardly biased correlations estimated from = .06) ρ 2 = .00) values are excluded. Grit was correlated at Nagelkerke R = .2 1 (k = 7, N = 4,116, SD ρ ρ persist in with performance in n on - academic domains . The relation between grit and the intent to

27 GRIT META - 27 ANALYSIS college and with the current employer was 8 (k = 5, N = 3,967, SD ρ = .00), and ρ = .1 5 (k = = .1 ρ = .00) respectively. 4, N = 519, SD ρ Relation with State and Trait Variables orthogonal , g rit Consistent with the claim that grit and cognitive ability are largely 513 cognitive ability (k = relation 21 , N = 1 1 , with , ρ = .0 5 , SD exhibited only a very weak = ρ 2 ) . Similarly weak relation s with cognitive ability were also observed for both the perseverance .1 0 5 , N = 2, 204 , ρ = - .01, SD = .04) and consistency facet (k = 5 , N = 2, 204 , ρ = .0 = , SD facet (k ρ ρ = .00). Grit exhibited much stronger relation s with other trait variables. Conscientiousness was = 22 , N = 1 very strongly correlated with overall grit (k , 826 , ρ = .84, SD ) and also with = .0 7 8 ρ , 1 perseverance (k = 8 , N = 4 , 967 , ρ = .8 3 , SD .6 = .14) and consistency (k = 8 , N = 4 , 967 , ρ = ρ SD = .1 7 ). Grit also exhibited a very strong relation = .72, with self - control (k = 4, N = 2,615, ρ ρ SD = .05), a variable that is often seen to be a facet of conscientiousness (e.g., Roberts et al., ρ 2005). Grit also exhibited a relatively strong relation with emotional stability (k = 1 4 , N = 1 , 501 , ρ = .4 1 , SD n considering that low levels of 4 = .04) but this should not be surprising whe ρ emotional stability are likely to be associated with an inconsistency of interest because negative affect states may be interpreted as a signal that the activity being engaged in is no longer interesting. Grit also exhibite d relatively strong relation s with a number of other variables that are sometimes presented as having a causal influence on success and performance, including: self - control (k = 4, N = 2,615, ρ = .72, SD ρ = .05), generalized self - efficacy (k = 3, N = , 1,908 ρ 3,817, .43 , SD 3, = . 11 ), mental toughness (k = 6, N = = ρ = .4 6 , SD = . 08 ), positive affect (k = ρ ρ N .12). = 670, ρ = .4 6 , SD = .0 3 ), and depression (k = 5, N = 3,86 5 , ρ = - .48, SD = ρ ρ

28 GRIT META - 28 ANALYSIS Relation with Demographic Variables relation he between grit and demographic variables such as gender (k = As expected, t s 5 , N = 18 , 750 , ρ =.0 5 , SD ), and = .07), year in school (k = 4, N = 2,961, ρ =.05, SD 5 2 = .0 ρ ρ .01, SD 9 , N = 1 5 , 261 , ρ = ethnic minority status (k = = .0 1 ) were all very weak with the ρ exception of age which exhibited a slight positive correlation with overall grit (k = , N = 22 ) in line with the prediction by Duckworth et al. (2007) that grit would 4 1 2 , 349 , ρ = .1 2 , SD .0 = ρ increase with age. This increase is similar to the general increase in cons cientiousness observed with age (Roberts et al., 2006). 6 HERE] [INSERT TABLE Incremental Validity The incremental validity estimates from these regression results are summarized in Table 6 . Results for Model 1 indicate that o verall grit explains no variance in either overall academic performance or high school GPA after controlli ng for conscientiousness, and explains only a i.e., Importantly, very small amount of incremental variance in college GPA ( Δ R = .00 4 ). conscientiousness explains incremental variance in these outcomes if first controlling for overall grit. Results for Model 2 show that p erseverance explained a substantial amount of incremental variance in overall academic performance ( , and a R = .040), high school GPA ( Δ R = . 08 5 ) Δ somewhat lo wer amount for college GPA ( Δ R = .0 23 ). Consistency explained almost no unique variance in the three criteria after controlling for either conscientiousness (see Model 3) or both conscientiousness and persistence (Model 4) and the negative sign of the regression coefficients for consistency for the overall academic performance and college GPA criteria also suggests a Perseverance explained relatively large amounts of unique variance in possible suppressor effect.

29 GRIT META - 29 ANALYSIS three crite ria even after controlling for both conscientiousness and consistency (Model 5). the consistency facet both overall grit and Overall, the incremental validity findings suggest that , while the ance of grit add little to our ability to understand or predict academic perform perseverance facet does offer an important improvement in explanatory power . Discussion grit is a higher - order construct composed of a Proponents of grit have asserted that perseverance facet and a consistency facet, grit scores are h ighly predictive of success (and a that and that grit scores provide information about better predictor than cognitive ability) , of success individuals that is meaningfully distinct from conscientiousness. Three primary findings from the may our meta - analy tic review of the grit literature suggest that assertions validity of these need to be revisited our findings although also suggest that a revised approach to the study of ; . grit may still hold value for our understanding of the determinants of performance First, our findings indicate that the current evidence does not support the claim that grit is a higher - order construct that is characterized by two lower - order facets. The original factor analytic studies could not speak to the presence of a higher - order factor structure because of methodological limitations , and our results indicate that the practice of combining perseverance scores and consistency scores into an overall grit score appears to result in a significant loss in nce. That is, perseverance is a much better predictor of the ability to predict performa performance than either consistency or overall grit and should therefore probably be treated as a construct that is largely distinct from consistency in order to maximize its utility . Second, overal l grit exhibits relation s with academic performance and retention that are known predictors of academic - only modest and that do not compare favorably with other well

30 GRIT META ANALYSIS 30 - performance such as cognitive ability (Sackett et al., 2012) , study habits and skills (Cre de & , and academic adjustment Kuncel, 2008) (Crede & Niehorster, 20 12 ) . Indeed, m eta - analytic reviews of the literatures for some of these other predictors report correlations with academic performance that are more than twice as big as those observed for overall grit in and retention this review. At the same time it should be remembered that variables that exhibit s mall to moderate effect sizes can still be very useful in high - stakes settings because even marginal improvements in individuals’ performance - or organizations ability to predict this performance – can have very meaningful positive effects. For example, a grit intervention that increasing the retention rate in college by even a single percentage point would potentially benefit thousands of future performance in a colle ge students . Similarly, even a small increase in the ability to predict selection setting may yield very substantial financial benefits for an organization (see Hunter & Such a benefit would be partic ularly large if the variable in Hunter, 1984 for a discussion). question reflected information about individuals that was distinct from the information reflected by other well - known predictors of performance and retention. ur third primary finding suggests that the incremental value of grit for the prediction of O likely to be limited. Grit scores exhibited very strong correlations with performance is conscientiousness and with self - control – a facet of conscientiousness . Indeed, the size of the correlation ( ρ = .84) with overall conscien tiousness is so strong as to not only limit the incremental value of grit scores for the prediction of performance over and above conscientiousness but also suggest that grit may be redundant with conscientiousness. Indeed, the correlation between overall grit and conscientiousness , and between persistence and conscientiousness ( ρ = .89) is much stronger than what is typically found between scores on two Pace & Brannick, 2010) . This, in turn, different global measures of conscientiousness ( ρ = .63;

31 GRIT META ANALYSIS 31 - suggests that grit research may have fallen victim to the jangle fallacy and that grit as currently . measured is simply a repackaging of conscientiousness or one of the facets of conscientiousness that two variables can be cC (1956) , of course, illustrated M very strongly correlated but ornack still exhibit very different correlations with a third variable but the meta - analytic estimates of the relation between overall grit and GPA in middle/ high school ( ρ = . 16 ) and college ( ρ = . 1 7 ) are largely identical to those reported for conscientiousness in the recent meta - (if somewhat weaker) analytic review by Porapat (2009): ρ = .21 for middle/high school GPA and ρ = .23 for college GPA. appraisal of the grit construct is Although our findings indicate that a critical re - - warranted t hree meta analytic findings reported in this paper do hold some promise for , proponents of grit as a predictor of success and as a potential focus of interventions. First, grit predicts retention approximately as well as many more t raditional predictors of retention such as cognitive cognitive ability and high school grades although not as well as some other non - – predictors. This suggests that the assessment of grit may be useful in settings in which retention is problematic (e.g., higher education) because it may allow researchers to identify individuals who might benefit the most from interventions that target grit or offer assistance in some other fashion. Second - as noted earlier - - our meta analytic results show that the perseverance of effort facet of grit for the grade criteria than the exhibits substantially higher criterion validity facet . Indeed, the observed criterion validity of perseverance for the hig h consistency of interest school GPA criterion is also significantly higher than the criterion validity observed for overall This suggests that the focus of the grit scores and also for conscientiousness (Poropat, 2009) . grit researchers should shift to perseverance as the most promising avenue of future research. Third , - analytic estimates suggest that perseverance of our hierarchical regression results based on meta

32 GRIT META - 32 ANALYSIS effort scores explain incremental variance over and above conscientiousness in t he various grade This is, of course, encouraging criteria. , but the only moderately high correlations among scores (see Pace & Brannick, 2010) on most personality traits assessed via two different inventories even if grit was simply a mean that such incremental validity findings wo uld be observed different manifestation of conscientiousness as our other results suggest. - criterion relation Even modest predictor s can be very important in applied settings, especially when individuals’ standing on the predictor can be impacted by simple interventions. Whether it is possible to enhance grit via interventions is not yet clear although evidence that social and personal skills as well as resiliency are responsive to interventions (Durlak et al., ., 2015) suggest that grit interventions may have some positive effect. 2010; Paunesku et al Although we do believe that our results regarding the validity of the perseverance facet offer - evaluation of t some promise we also believe that our overall results should lead to a re he appropriateness of planned or existing grit interventions. Schools and colleges have limited resources t o devote to interventions and are likely to be best served by focusing those resources y related to performance and strongl on variables that have been demonstrated to be 1) most Fortunately there persistence/retention and 2) responsive to interventions. are a number of For example, s variables that meet both of those requirements. tudy skills and habits have been ρ = .40 shown to correlate approximately with college GPA ( Credé & Kuncel, 2008), while Hattie et al. (1996) showed that study skills interventions can have moderate positive effects on study skills. College students’ adjustment to college has been shown to be similarly predictive of rformance ( academic pe ρ = .39 for academic adjustment) , is also one of the best predictors of retention in college ( ρ = .29 for institutional attachment ) and can be slightly improved by simple Simple interventions such as orientation programs (see Credé & Niehorster, 2012 f or a review).

33 GRIT META ANALYSIS 33 - class attendance is also very strongly related to academic performance, and making class attendance compulsory appears to dramatically reduce the proportion of students who fail a class (Credé et al., 2010). S tudy skills and study habi ts, adjustment to college , and class attendance are thus far more strongly related to academic performance and retention than g rit , and there is sound evidence that interventions can improve students’ standing on these constructs (especially for study skil ls and habits). Limitations and Future Research Meta - analyses are limited by the nature and quality of the data present in a literature. As such this meta - analysis of the grit literature has some notable limitations. First, the literature relating grit to academic performance is primarily based on concurrent designs. This, in turn, because may have resulted in inflated estimates of the grit - academic performance relation individuals’ knowledge of their academic performance may influence their responses to the the grit literature may be Second, there are at least three reasons why measure of grit. characterized by a non - trivial amount of range restriction. A ll of t he examined studies relied on grit items self - reports of grit and the social desirability of may have resulted in range restriction reported grit scores . Individuals may also generally not be aware of their true level of grit - in self . Further, and unintentionally re port inflated levels of grit ( Kruger & Dunning, 1999) samples drawn from populations that have been selected based on prior performance may exhibit some range restriction on grit. For example, cadets at the US Military Academy at West Point have likely exhibited outstanding academic performance in high school and may therefore have a lower range of grit scores than the range found in the general population . We were unable to normative data and the correct for range restriction in our meta - analysis because of the lack of variability in how the grit scales were used by researchers but future research may be able to

34 GRIT META ANALYSIS 34 - estimate the level of range restriction that is present in samples. , many of the studies Third examining th relation between grit and re tention were characterized by very high base rates of e retention (i.e., low rates of dropout). Duckworth et al. (2007) for example report data on one sample from the United States Military Academy in which 94.2% of the sample were retained through the exami ned period. Such low base rates severely attenuate the size of the correlation that can be observed. In such circumstances meta - analyses could make corrections for range restriction , but such a correction would require information about the size of the sta ndard deviation for the retention criteria in the general population and we are not aware of a reasonable estimate for this value. Finally, although the empirical grit literature is sufficiently large to allow yet us to comment with relative confidence on average population effect sizes the literature is not large enough to allow moderator analyses characterized by high power (Hunter & Schmidt, 200 4 ). An exploration of the reasons for the occasionally wide credibility intervals will require the accumulation of further data. We believe that future research in this domain should consider f ive broad issues . First, researchers should atte mpt to examine whether grit exhibit the type of stability that is associated Grit interventions will with other personality traits or whether it is responsive to interventions. need to be tested to evaluate the malleability of grit but there are sound theo retical reasons why such interventions may be effective. Eisenberger (1992) argued that industriousness – a construct that is similar to grit – can be acquired via reinforcement and that repeated reinforcement for high effort (i.e., grit) can eventually re sult in a generalized increase in effort across tasks even when these tasks are not extrinsically reinforced. This work will not only require long - term experimental manipulations in the form of reinforcements for high effort but should help to - like and also help to clarify the type of grit establish th e degree to which grit is truly trait

35 GRIT META 35 - ANALYSIS interventions that are likely to be most effective. Second, grit researchers should consider ater range of examining criteria that span to different domains (e.g., work settings), a gre difficulty and a greater variety of task types (e.g., intellective tasks versus creative tasks). This may help to establish boundary conditions for the influence of grit on success and performance. Third , grit researchers should consider exam ining the potential moderators of the grit - performance relation discussed earlier: the moderating role of the performance domain; the moderating role of individual differences such as ability and meta - cognition; and the moderating , it may also be useful to examine the degree to which scores on role of the level of grit . Fourth measures of grit are related to scores on measures of motivation. A popular definition of motivation is that it reflects “an individual’s intensity, direction, and persistence of effo rt toward achieving a goal” (e.g., Robbins, Judge, & Campbell, 2010) and as such bears clear conceptual similarities to grit. Finally, it is possible that the grit literature may benefit from a refinement of the grit scale using method based on Item Respon se Theory. It is unlikely that the relatively short measures of grit are equally good at assessing low, medium, and high levels of grit . This lack of depth and breadth in item content could lead to attenuation of the reported effects (see Credé, Harms, Nie horster, & Gaye - Valentine, 2012) . Better measures of grit would not only help to clarify the nature of the grit - performance relation but would also be important for the evaluation of future grit interventions. Conclusion Grit as a predictor of performance and success and as a focus of interventions holds much intuitive appeal , but grit as it is currently measured does not appear to be particularly predictive of success and performance and also does not appear to be all that different to conscientiousness. e hope that greater rigor in scale development, a greater focus on the perseverance facet, and a W

36 GRIT META ANALYSIS 36 - more nuanced approach in study design will help future grit researchers to develop boundary conditions for grit in its role as in influence on performance and success.

37 GRIT META - 37 ANALYSIS References analysis) - (* denotes inclusion in meta *Allen, B. R. (2014). An analysis of student success predictors for technical college persistence . (Unpublished doctoral thesis). Edgewood College, Madison, Wisconsin. - 2014. SAS Institute. Allison, P.D. (2014). Measuring fit for logistic regression. Research paper 1485 - suicidal *Anestis, M. D., & Shelby, E. A. (2015). Grit and perseverance in suicidal behavior and non self - Death Studies, 39 (4), 211 - 218. doi:10.1 080/07481187.2014.946629 injury. *Anestis, M. D., Kleiman, E. M., Lavender, J. M., Tull, M. T., & Gratz, K. L. (2014). The pursuit of death versus escape from negative affect: An examination of the nature of the relationship suicidal behavior and non - suicidal self - injury. between emotion dysregulation and both , 1820 - 1830. doi:10.1016/j.comppsych.2014.07.007 Comprehensive Psychiatry, 55 Does *Arouty, D. (2015). academic character matter? The role of grit and resilience in predicting success (Unpublished doctoral dissertation). Hofstra University, Hempstead, New York. . *Arslan, S., Akin, A., Ç î temel, N. (2013). The predictive role of grit on metacognition in Turkish university students. Studia Psychologia, 55 (4), 311 - 320. Bacon, D. R., & Bean., B. (2006). GPA in re search studies: An invaluable but neglected opportunity. Journal of Advanced Academics, 26 - 293. doi:10.1177/0273475305284638 , 274 *Batres, I. (2011). The relationship of grit, subjective happiness, and meaning in life on alternative education students’ GPA and attendance. (Unpublished doctoral thesis). University of La Verne, La Verne, California. *Black, K. R. (2014). Grit in college students: Associations with achievement and parental influences. (Unpublished masters thesis). West Virginia University, Morgantown, West Virginia.

38 GRIT META - 38 ANALYSIS Blalock, D.V., Young, K.C., & Kleiman, E.M. (2015). Stability amidst turmoil: Grit buffers the effects , 781 - 784. Psychiatry Research, 228 of negative life events on suicide ideation. doi:10.1016/j.psychres.2015.04.041 *Bowman, N. A., Hill, P. L., Denson, N., & Bronkema, R. (2015). Keep on truckin’ or stay the course? Exploring grit dimensions as differential predictors of educational achievement, satisfaction, and . Social Psychology and Personality Science, 6 (6), 639 - intentions 654. doi:10.1177/1948550615574300 The relationship between grit and academic, military and physical performance at *Buller, E. F. (2012). (Unpublished doctoral disse rtation). University of Kansas, the United States Military Academy. Lawrence, Kansas. Camara, W.J., & Schmidt, A.E. (1999). Group differences in standardized testing and social stratification. College Entrance Examination Board . New York. Tubré, T. (2012). *Chambers, R., Brawley, A. M., Castille, C. M., & Examining the validity of the . Poster presented at the 26th National Conference on Undergraduate Resilience Quotient Research, Ogden, Utah, March 29 - 31, 2012. Premuzic, T., Furnham, A., & Ackerman, P. L. (2006). Incremental val Chamorro - idity of the Typical Intellectual Engagement Scale as predictor of different academic performance measures. Journal of Personality Assessment, 87 , 261 – 268. *Chang, W. (2014). Grit and academic performance: Is being grittier better? (Unpublished doctoral di ssertation). University of Miami, Miami, Florida. Cohen, R.M. (2015). Will students soon be tested for “grit”? The American Prospect. Retrieved from grit - students - soon - be - tested -

39 GRIT META - 39 ANALYSIS *Cooper, A. D. (2014). of non - cognitive factors in predicting college academic Exploring the use (Unpublished masters thesis). University of Tennessee, Chattanooga, Tennessee. outcomes. - R professional manual. Odessa, FL: Psychological Costa, P.T., & McCrae, R.R. (1992). The NEO PI Assessment Resources. Crant, J.M. (1995). The proactive personality scale and objective job performance among real Journal of Applied Psychology 80, 532 − estate agents. 537. Credé, M., & Harms, P.D. (2015). 25 years of higher order confirmatory factor analysis in the - organizational sciences: A critical review and development of reporting recommendations. - 872 . Journal of Organizational Behavior, 36, doi:10.1002/job.2008 845 Credé, M., Harms, P.D., Niehorster, S., & Gaye - Valentine, A. (2012). An evaluation of the conse quences of using short measures of the Big Five personality traits. Journal of Personality doi:10.1037/a0027403 - and Social Psychology, 102, 874 888. Credé, M., & Kuncel, N.R. (2008). S tudy habits, study skills, and study attitudes: The third pillar suppor ting collegiate academic performance. Perspectives on Psychological Science, 3, 415 - 453. d oi :10.1111/j.1745 - 6924.2008.00089.x Credé, M., & Niehorster, S. (2012). Adjustment to college as measured by the Student Adaptation to College Questionnaire: A quantitative review of its structure and relationships with correlates and consequences. Educational Psychology Review, 24, - 133 165. doi:10.1007 /s10648 - 011 - 9184 - 5 Credé, M., & Phillips, L.A. (2011). A meta - analytic review of the Motivated Strategies for Learning Questionnaire. 21, 337 - 346 . Learning and Individual Differences, 2011.03.002 doi:10.1016/j.lindif.

40 GRIT META - 40 ANALYSIS Cohen, J. (1988). (2nd ed.). Hillsdale, NJ: Statistical power ana lysis for the behavioral sciences Lawrence Earlbaum Associates. Costa, P. T., & McCrae, R. R. (1992). - PI - R) and NEO Five - Revised NEO Personality Inventory (NEO - FFI) Factor Inventory (NEO M anual . Odessa, FL: Psychologi cal Assessment Professional Resources. Costa, P. T., Terracciano, A., & McCrae, R. R. (2001). Gender differences in personality traits across cultures: Robust and surprising findings. Journal of Personality and Social Psychology, 81 , 322 - 331. doi:10.1037//0022 - 3514.81.2.322 Staying the course: Grit, academic success and non traditional doctoral students. *Cross, T. M. (2013). - (Unpublished doctoral dissertation). Pepperdine University, Malibu, California. *Datu, J. A. D., Valdez, J. P. M., & Ki ng, R. B. (2015). Perseverance counts but consistency does not! Validating the Short Grit Scale in a collectivist setting. Current Psychology . doi:10.1007/s12144 - 015 - 9374 - 2 *Davidson, B.A. (2014). Examining the relationship between non - cognitive skills and leadership: The influence of hope and grit on transformational leadership behavior. (Unpublished doctoral dissertation). University of Kansas, Lawrence, Kansas. Davis, D., Dorsey, J.K., Franks, R.D., Sackett, P.R, Cynthia, A., Zhao, X. (2013). Do racial and ethnic group differences in performance on the MCAT exam reflect test bias? Academic Medicine, 88 , 593 - 602. doi:10.1097/ACM.0b013e318286803a *DeCandia, G. M. (2014). Relationships between academic identity and academic achievement in low - income urban a dolescents. (Unpublished doctoral dissertation). State University of New Jersey, New Brunswick, New Jersey.

41 GRIT META ANALYSIS 41 - De Fruyt, F.; Van De Wiele, L. & Van Heeringen, C. (2000). Cloninger's psychobiological model of emperament and c haracter and the f t - f actor m odel of p ersonality . Personality and Individual ive Differences 29, 441 – 452. *DiMenichi, B. C. & Richmond, L. L. (2015). Reflecting on past failures leads to increa s ed Journal of Cognitive Psychology, 27 (2), 180 - perseverance and sustained attention. 193. doi: 10.1080/20445911.2014.995104 Duckworth, A.L. (2013a). The key to success? Grit. Retrieved from - 9644 Duckworth, A.L. (2013 b ). Research Psychologist Angela Duckworth, 2013 MacArthur Fellow | MacArthur Foundation. Retri e ved from: S). - *Duckworth, A. L., & Quinn, P. D. (2009). Development and validation of the short grit scale (Grit Journal of Personality Assessment, 91 174. d oi:10.1080/00223890802634290 - (2), 166 *Duckworth, A. L., Kirby, T. A., Tsukayama, E., Berstein, H., & Ericsson, K. A. (2011). Deliberate practice spells success: Why grittier competitors triumph at the national spelling bee. Social e, 2 (2), 174 181. doi:10.1177/1948550610385872 Psychology and Personality Scienc - *Duckworth, A. L., Peterson, C., Matthews, M. D., & Kelly, D. R. (2007). Grit: Perseverance and passion for long term goals. Journal of Personality and Social Psychology, 92 (6), 1087 - 1101. doi:10.1037/0022 - 351 - Duckworth, A.L., Quinn, P.D., Lynam, D.R., Loeber, R., & Stouthamer Loeber, M. (2011). Role of test motivation in intelligence testing , 7716 - 7720. doi:10.1073/pnas.1018601108 . PNAS, 198 Durlak, J. A., Weissberg, R. P., & Pachan, M. (2010). A me ta - analysis of After - School Programs that seek to promote personal and social skills in children and adolescents. American Journal of 6 Community Psychology, 45 , 294 - 309. doi:10.1007/s10464 - 010 - 9300 -

42 GRIT META - 42 ANALYSIS Earman, J., & Glymour, C. (1978). Lost in the tensors: Einstein’s struggles with covariance principles 1916. 278. doi:10.1016/0039 , 251 - - - Studies in History and Philosophy of Science, 9 1912 0 - 3681(78)90008 s in meta - analysis detected by a Eggers, M., Smith, G.D., Schneider, M., & Minder, C. (1997). Bia British Medical Journal, 315 simple, graphical test. - 634. doi:10.1136/bmj.315.7109.629 , 629 What predicts first semester college performance? Cognitive ability, SAT, *Engel, L. I. (2013). (Unpub lished doctoral dissertation). Hofstra University, Hempstead, conscientiousness, and grit. New York. Ericsson, K. A., Krampe, R. T., & Tesch - Römer, C. (1993). The role of deliberate practice in the , 363 - acquisition of expert performance. Psychological Review, 100 doi:10.1037/0033 - 406. 295X.100.3.363 *Eskreis - Winkler, L., Duckworth, A. L., Shulman, E. P., & Beal, S. (2014). The grit effect: Predicting retention in the military, the workplace, school and marriage. Frontiers in Psychology, 5 . doi:10.3389/fpsyg.2014.00036 *Eskreis - Winkler, L., Shulman, E. P., & Duckworth, A. L. (2014). Survivor mission: Do those who survive have a drive to thrive at work? The Journal of Positive Psychology, 9 (3), 209 - 218. doi:10.1080/17439760.2014.888579 Field, A. & Gillett, R. (2010). How to do meta - analysis. British Journal of Mathematical and Statistical Psychology, 63 , 665 - 694. doi:10.1348/000711010X502733 *Fillmore, E. P. (2015). Grit and beliefs about intelligence: The relationship and the role these factors pl ay in the self - regulatory processes involved in medical students learning gross anatomy. (Unpublished doctoral dissertation). Indiana University, Bloomington, Indiana.

43 GRIT META ANALYSIS 43 - Foldes, H.J., & Duehr, E.E., & Ones, D.S. (2008). Group differences in personality: Met - analysis a Personnel Psychology, 61 , 579 - comparing five U.S. racial groups. doi:10.1111/j.1744 - 616. 6570.2008.00123.x Frey, M.C., & Dotterman, D.K. (2004). Scholastic assessment or g? The relationship between the Scholastic Assessment Test and General Cogn itive Ability. Psychological Science, 15 , 373 - 378. doi:10.1111/j.0956 7976.2004.00687.x - Ghiselli, E. E., Campbell, J. P., & Zedeck, S. (1981). Measurement theory for the behavioral sciences. San Francisco, CA: W.H. Freeman & Company. Goldberg, L. R. (1999) . A broad - bandwidth, public domain, personality inventory measuring the lower - level facets of several five - factor models. In I. Mervielde, I. Deary, F. De Fruyt, & F. Ostendorf 28). Tilburg, The Netherlands: Tilburg (Eds.), Personality Psychology in Europe , Vol. 7 (pp. 7 - University Press. rd CPI manual (3 Gough, H.G. (1996). ed.). Palo Alto, CA: Consulting Psychologists Press. *Hammer, J. H., & Good, G. E. (2010). Positive psychology: An empirical examination of beneficial aspects of endorsement of mas culine norms. Psychology of Men & Masculinity, 11 (4), 303 - 318. doi:10.1037/a0019056 *Haran, U., Ritov, I., & Mellers, B. A. (2013). The role of actively open - minded thinking in information g, 8 acquisition, accuracy, and calibration. Judgment and Decision Makin (3), 188 - 201. Hattie, J., Biggs, J., & Purdie, N. (1996). Effects of learning skills interventions on student learning: A , 99 - analysis. Review of Educational Research , 66 meta – 136. doi:10.3102/00346543066002099 Hembree, R. (1988). Correlates, causes, effects, and treatment of test anxiety. Review of Educational doi:10.3102/00346543058001047 Research, 58 , 47 - 77.

44 GRIT META - 44 ANALYSIS Hezlett, S. A., Kuncel, N. R., Vey, M. A., Ahart, A., Ones, D. S., Campbell, J. P., & Camara, W. (2001) . analysis. In D. S. Ones & S. A. The predictive validity of the SAT: A comprehensive meta - Hezlett (Eds.), Predicting performance: The interface of I/O psychology and educational research. San Diego: Symposia presented at the annual conference of the Socie ty for Industrial and Organizational Psychology. Hill, P. L., Burrow, A. L., & Bronk, K. C. (2014). Persevering with positivity and purpose: An * examination of purpose commitment and positive affect as predictors of grit. Journal of doi :10.1007/s10902 Happiness Studies. 014 - 9593 - 5 - *Hogan, M. L. (2013). - cognitive traits that impact female success in BigLaw. (Unpublished Non doctoral dissertation). University of Pennsylvania, Philadelphia, Pennsylvania. doi: 10.1037/a0019056 Howe, M. J. A. (1999). Genius ex plained. New York: Cambridge University Press. Hunter, J.E., & Hunter, R.F. (1984). Validity and utility of alternative predictors of job performance. Psychological Bulletin, 96 - 98. , 72 Methods of meta Correcting error and bias in research analysis: Hunter, J. E., & Schmidt, F. L. (2004). - nd findings ed.). Thousand Oaks, CA: Sage (2 IBM Corp. Released 2013. IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY: IBM Corp. *Incantalupo - Kuhner, J. (2015). Teacher dispositions and perceived environment: The relationship among grit, resiliency, and perceptions of school climate (Unpublished doctoral dissertation). . Hofstra University, Hempstead, New York. *Ivcevic, Z., & Brackett, M. (2014). Predicting school success: Comparing conscientiousness, grit, an d emotion regulation ability. Journal of Research in Personality, 52 , 29 - 36. 10.1016/j.jrp.2014.06.005 doi:

45 GRIT META ANALYSIS 45 - Jackson, D. N., Paunonen, S. V., & Tremblay, P. F. (2000). . Six factor personality questionnaire manual Port Huron, MI: Sigma Assessment Systems. *Jordan, M. H., Gabriel, T. J., Teasley, R., Walker, W.J., & Schra der, M. (2015). An integrative e approach to identifying factors related to long - term career commitments: A military example. Career Development International, 20 (2), 163 - 178. doi:10.1108/CDI - 05 - 2013 - 0071 *Joseph, A. I. (2009). The role of grit in prediciting performance in collegiate athletes. (Unpublished doctoral dissertation). Capella University, Minneapolis, Minnesota. Karabenick, S. A. (2003). Seeking help in large college classes: A pe rson - centered approach. 7 Contemporary Educational Psychology, 28 (1), 37 - 58. doi:10.1016/S0361 - 476X(02)00012 - *Karlson, R. W. (2014). Perry’s scheme of intellectual and ethical development influence on Duckworth’s Short Grit Score in interior design student s. (Unpublished masters thesis). University of Minnesota, Minneapolis, Minnesota. Kelley, T.L. (1927). Interpretation of Educational Measurement . World Books: Yonkers - on - Hudson, NY. *Kelly, D. R., Matthews, M, D., & Bartone, P. T. (2014). Grit and hardiness as predictors of performance among West Point cadets. Military Psychology, 26 (4), 327 - 342. doi: 10.1037/mil0000050 *Khaler, K. L. (2014). On beating the odds: A study of successful community college students . (Unpu blished doctoral dissertation). Michigan State University, East Lansing, Michigan. *Kleiman, E. M., Adams, L. M., Kashdan, T. B., & Riskind, J. H. (2013). Gratitude and grit indirectly reduce risk of suicidal ideations by enhancing meaning in life: Eviden ce for a mediated moderation model. Journal of Research in Personality, 47 , 539 - 546. 10.1016/j.jrp.2013.04.007 doi:

46 GRIT META ANALYSIS 46 - rd Kline, R.B. (2011). Ed.). Principles and Practices of Structural Equation Modeling (3 The Guildford Press: New York. Kohn, A. (2014, April 6). The downside of “grit”. Washington Post. Retrieved from: - grit/ Krampe, R. T., & Ericsson, K. A. (1996). Maintaining excellence: Deliberate practice and elite Journal of Experimental Psychology: General, 125 , performance in young and older pianists. 331 - 359. doi:10.1037/0096 - 3445.125.4.331 Kruger, J., & Dunning, D. (1999). Unskilled and unaware of it: How difficulties in recognizing one’s own incompetence lead to inflated self - assessments. Journal of Personality and Social 3514.77.6.1121 Psychology, 77 , 1121 - 1134. doi:10.1037/0022 - Kuncel, N.R., Crede, M., & Thomas, L.L. (2005). The reliability of self reported grade point averages: - A meta - analysis and summary of the literature. Review of Educational Research, 75, 63 - 82 . doi:10.3102/00346543075001063 Kuncel, N. R., Hezlett, S.A, & Ones, D.S. (2004). Academic performance, career potential, creativity, and job performance: Can one construct predict them all ? Journal of Personality and Social – Psychology, 86 , 148 doi:10.1037/0022 161. - 3514.86.1.148 * Lovering, M. E., Heaton, K. J., Banderet, L. E., Neises, K., Andrews, J, & Cohen, B. S. (2015). Psychological and physical characteristics of U.S. marine recruits. Military Psychology, 27 (5), 261 - 275. doi: 10.1037/mil000008 2 Lucas, G.M., Gratch, J., Cheng, L., & Marsella, S. (2015). When the going gets tough: Grit predicts 22. costly perseverance. Journal of Research in Personality, 59 , 15 -

47 GRIT META - 47 ANALYSIS *MacCann, C., & Roberts, R. D. (2010). Do time management, grit, and self - control relate to academic achievement independently of conscientiousness? In R. Hicks (Ed.), Personality and individual (pp. 79 - 90). Queensland, Australia: Australi an Academic Press. differences: Current directions MacNamara, B.N., Hambrick, D.Z., & Oswald, F.L. (2014). Deliberate practice and performance in music, games, sports, education, and professions: A meta analysis. Psychological Science, 25 , - - 1618. 1608 doi:10.1177/0956797614535810 *Maddi, S. R., Erwin, L. M., Carmody, C. L., Villareal, B. J., White, M., & Gundersen, K. K. (2013). Relationship of hardiness, grit, and emotional intelligence to internet addiction, excessive 134. consumer spending, and gambling. The Journal of Positive Psychology, 8 (2), 128 - doi:10.1080/17439760.2012.758306 *Maddi, S. R., Matthews, M. D., Kelly, D. R., Villareal, B., & White, M. (2012). The role of hardiness and grit in predicting performance and retention of USMA cadets. Military Psychology, 24 , 19 - 28. doi:10.1080/08995605.2012.639672 Mattern, K.D., & Patterson, B.F. (2009). Is performance on the SAT related to college retention? Research Report 2009 - 7. The College Board, New York, NY. thods. McCornack, R.L. (1956). A criticism of studies comparing weighting me Journal of Applied Psychology, 40 - 345. doi:10.1037/h0045635 , 343 McClelland, D. C. (1985). How motives, skills, and values determine what people do. American Psychologist, 40 , 812 – 825. *McCutcheon, C. A. (2014). Perseverance and persistence in achie ving educational goals: A case study of an adult charter high school examining perseverance of students utilizing a mixed methods approach. (Unpublished doctoral dissertation). Ball State University, Muncie, Indiana.

48 GRIT META - 48 ANALYSIS Meng, X.L., Rosenthal, R., & Rubin, D. B. (1992). Comparing correlated correlation coefficients. - 175. Psychological Bulletin, 111 - 2909.111.1.172 , 172 doi:10.1037/0033 *Meriac, J. P., Slifka, J. S, & LaBat, L. R. (2015). Work ethic and grit: An examination of empirical Personality and Individual Differences, 86 , 401 - redundancy. 405. doi: 10.1016/j.paid.2015.07.009 Miller, J.D., Price, J., & Campbell, W.K. (2012). Is the Narcissistic Personality Inventory still relevant? A test of independent grandiosity and entitlement scales in the assessment o f narcissism. , 8 - 13. Assessment, 19 doi:10.1177/1073191111429390 *Morris, R. B. (2011). Can chameleons lead change? The effect of resistance to change on high self - (Unpublished doctoral dissertation). Columbia monitoring leaders’ strength of purpose. Unive rsity, New York, New York. Mosier, C.I. (1943). On the reliability of a weighted composite. Psychometrika, 8 , 161 - 168. doi:10.1007/BF02288700 *Office of Institutional Research & Assessment (2012). Social Science Survey: GRIT & AMBITION Scales. Unpublished manuscript, Metropolitan Community College - Blue River, Independence, Missouri. Pace, V.L., & Brannick, M.T. (2010). How similar are personality scales of the “same” construct? A meta - analytic investigation. Personality and Individual Differences, 49 , 669 - 676. doi:10.1016/j.paid.2010.06.014 Parker, B. (2004). Albert Einstein’s Vision: Remarkable Discoveries that Shaped Modern Science. Prometheus Books: Amherst: NY.

49 GRIT META ANALYSIS 49 - Paunesku, D., Walton, G.M., Romero, C., Smith, E.N., Yeager, D.S., & Dweck, C.S. (2015). Mind set Psychological Science . interventions are a scalable treatment for academic underachievement. Advance Publication Online. doi:10.1177/0956797615571017 Paunonen, S. V. & Ashton, M.C. (2001). Big Five factors and facets and the prediction of behavior. Journal of Personality and Social Psychology, 81 , 524 - 539. doi:10.1037/0022 - 3514.81.3.524 *Pennings, S. M., Law, K. C., Green, B. A., & Anestis, M. D. (2015). The impact of grit on the International Journal of Cognitive Therapy, relationship between hopelessness and suicidiality. 8 (2), 130 - 142. doi:10.1521/ijct.2015.8.2.130 Perkins - Gough, D. (2 013). The significance of grit: A conversation with Angela Lee Duckworth. Educational Leadership, 71 (1), 14 - 20. The relationship of trait emotional intelligence with academic (2013). Perera, H.N. & DiGiacomo, M. performance: a meta Learning and Individual Differences, 28, 20 - 33. analytic review. - doi:10.1016/j.lindif.2013.08.002 - Watson, B., López, B., Ojeda, L., & Rodriguez, K. M. (2015). Cultural and cognitive predictors of *Piña academic motivation among Mexican American adolescents: Caution agains t discounting the impact of cultural processes. Journal of Multicultural Counseling and Development, 43 , 109 - 121. doi:10.1002/j.2161 - 1912.2015.00068.x - Porapat, A.E. (2009). A meta analysis of the five - factor model of personality and academic doi:10.1037/a0014996 performance. sychological Bulletin, 135 , 322 - 338. P Porapat, A. E. (2014). A meta - analysis of adult - rated personality and academic performance in primary doi:10.1111/bjep.12019 education. British Journal of Educational Psychology, 84 , 239 - 252.

50 GRIT META - 50 ANALYSIS * J., Pritschet, B. L., & Cutton, D. M. (2012). Grit, conscientiousness, and the transtheoretical Reed, 619. (5), 612 - Journal of Health Psychology, model of change for exercise behavior. 18 doi:10.1177/1359105312451866 Richardson, M., Abraham, C., & Bond, R. (2012). Psychological correlates of university students' - analysis. Psychological Bulletin, 138, academic performance: a systematic review and meta 353 - 387. doi:10.1037/a0026838 g and non Whites: A structural equation g influences on GPA for Hispanics and *Richmond, M. (2015). - . modeling (SEM) approach to Spearman’s law of diminishing returns (SLDOR) ublished (Unp doctoral dissertation). The University of Texas at San Antonio. Roberts, B. W., Bogg, T., Walton, K. E., Chernyshenko, O. S., & Stark, S. E. (2004). A lexical Journal of Research in order structure of conscientiousness. investigation of the lower - Personality, 38 - 178. doi:10.1016/S0092 - 6566(03)00065 - 5 , 164 Roberts, B.W., Chernyshenko, O.S., Stark, S., & Goldberg, L.R. (2005). The structure of conscientiousness: An empirical investigation based on seven major personality questionnaires. Personnel Psychology, 58 , 103 - 139. Robbins, S.B., Lauver, K., Le, H., Davis, D., Langley, R., & Carlstrom, A. (2004). Do psychosocial and analysis. study skills factors predict college outcomes? A meta - Psychological Bulletin, 130 , 261 - 288. - 2909.130.2.261 doi:10.1037/0033 Robbins, S.B., ,Allen, J., Casillas, A., Peterson, C.H., & Le, H. (2006). Unravelling the differential e ffects of motivational and skills, social, and self - management measures from traditional predictors of college outcomes. Journal of Educational Psychology, 98 , 598 - 616. Robbins, S.P., Judge, T.A., & Campbell, T.T. (2010). Organizational Behavior . Financial Times Prentice Hall: Harlow, England.

51 GRIT META - 51 ANALYSIS Roberts, B. W., Walton, K. & Viechtbauer, W. (2006). Patterns of mean - level change in personality analysis of longitudinal studies. Psychological Bulletin , 132 , traits across the life course: A meta - - 25 . doi:10.1037 1 - 2909.132.1.1 /0033 * - Kraft, C. (2014). Teachers’ motivational responses to new teacher performance Robertson management systems: An evaluation of the pilot of Aldine ISD’s INVEST system. (Unpublished delphia, Pennsylvania. doctoral dissertation). University of Pennsylvania, Phila * - Kraft, C., & Duckworth, A. L. (2014). True grit: Trait - level perseverance and passion for Robertson Teachers College long - term goals predicts effectiveness and retention among novice teachers. Record, 116, 030302. *Rojas, J.P. (2015). The relationship among creativity, grit, academic motivation, and academic success Theses and Dissertations -- Educational, School, and Counseling Psychology. in college students. Paper 39. Ryan, A.M. & Pintrich, P.R. (1997). “Should I ask for help?” The role of motivation and attitudes in adolescents’ help seeking in math class . Journal of Educational P sychology, 89 , 329 - 341. 0663.89.2.329 doi:10.1037/0022 - Ryans, D.G. (1939). The measurement of persistence: An historical review. Psychological Bulletin, 36, 715 - 739. Sackett, P.R., Kuncel, N.R., Beatty, A.S., Rigdon, J.L., Shen, W., & Kiger, T.B. (2012). The role of socioeconomic status in SAT - grade relationships and in college admissions decisions. Psychological Science, 23 , 1000 - 1007. . Schmidt, F.L. & Le, H.A. (2004). Software for the Hunter - Schmidt meta - analysis method: version 2.0

52 GRIT META - 52 ANALYSIS * Samson, A. C., Proyer, R. T., Ceschi, G., Pedrini, P. P., & Ruch, W. (2011). The fear of being laughed at in Switzerland: Regional differences and the role of positive psychology. Swiss Journal of doi:10.1024/1421 62. Psychology, 70 (2), 53 - 0185/a000039 - Sc hrepfer - Tarter, A. (2013). An exploratory study of the academic optimism of principals. * (Unpublished doctoral dissertation). The Ohio State University, Columbus, Ohio. - Anxiety Research, 4, Seipp, B. (1991). Anxiety and academic performance: A meta analysis of findings. - 41. doi:10.1080/08917779108248762 27 Shechtman, N., DeBarger, A. H., Dornsife, C., Rosier, S., & Yarnall, L. (2013). Promoting grit, tenacity, st and perseverance: Critical factors for success in the 21 century. U.S. Department of Educa tion, 13.pdf Office of Education Technology . - Draft - Grit - Report - 2 - 17 - * Storm clouds in the mind: A comparison of hope, grit, happiness, and life Sheehan, K. (2014). satisfaction in traditional and alternative high school students. (Unpublished doctoral dissertation). Hofstra University, Hempstead, New York. * Shishim, M. D. (2012). The relationship between college student success and well - being determinants: An explor atory study of measures . (Unpublished doctoral dissertation). University of California, Santa Barbara, California. California Polytechnic University, San Luis Obispo, California. * Singh, K. & Jha, S. D. (2008). Positive and negative affect, and grit as pre dictors of happiness and life satisfaction. - Journal of the Indian Academy of Applied Psychology, 34 , 40 45. *Stewart, S.B. (2015). - control as predictors of first - year student success. (Unpublished Grit and self doctoral dissertation). The University of So uthern Maine, Portland, Maine. * Strayhorn, T. L. (2013). What role does grit play in the academic success of black male collegians at predominantly white institutions? Journal of African American Studies. doi:10.1007/s12111 - 012 - . 9243 - 0

53 GRIT META - 53 ANALYSIS *Suzuki, Y., Tamesue, D., Asahi, K., & Ishikawa, Y. (2015). Grit and work engagement: A cross - : e0137501. DOI: 10.1371/journal.pone.0137501 sectional study. PLOS One 10(9) (2015). [Grit, Conscientiousness, Performance, & Satisfaction]. Unpublish ed raw data. * Third Author 2015 ). Can resilience be developed at Vanhove, A., Herian, M., Perez, A., Harms, P.D. & Lester, P. ( - work? A meta of resilience - building program effectiveness. Journal of analytic review Advance Occupational and Organizational Psychology. online publication. 10.1111/joop.12123 doi: Von Culin, K. R., Tsukayama, E., & Duckworth, A. L. (2014). Unpacking grit: Motivational correlates * - term goals. The Journal of Positive Psychology, 9 (4), 306 - of perseverance and passion for long 17439760.2014.898320 10.1080/ 312. doi: * Academic Achievement of Tradition and Nontraditional Warden, D., Myers, C., & Harrell, B. (2015). st College Students . Poster presented at the 61 Annual Meeting of the Southeastern Psychological th st Association, – 21 March 18 , 2015. *Waring, A. (2015). The influence of attachment and grit on life satisfaction and romantic relationship satisfaction . (Unpublished doctoral dissertation). University of La Verne, La Verne, California. Predicting college students’ positi * Watson, H. (2013). ve psychology attributes with dimensions of executive functioning. (Unpublished masters thesis). Middle Tennessee State University, Murfreesboro, Tennessee. *Wenner, J.R. (2015). Predictors of prosocial behavior and civic involvement: Differences in middle aged and older adults . (Unpublished masters thesis). North Dakota State University, Fargo, North Dakota. *Weston, L.C. (2015). A replication and extension of psychometric research on the grit scale. nd, College Park, Maryland. (Unpublished masters thesis). University of Maryla

54 GRIT META - 54 ANALYSIS Wilkinson, L. & Task Force on Statistical Inference (1999). Statistical methods in psychology journals: American Psychologist, 54 , 594 - 604. Guidelines and explanations. d its relation with college students’ self - *Wolters, C.A., & Hussain, M. (2015). Investigating grit an Metacognition and Learning, 10 , 293 - 311. regulated learning and academic achievement. Zimmerman, B.J. (1990). Self - regulated learning and academic achievement: An overview. Educational - Psychology, 25 , 3 .1207/s15326985ep2501_2 17. doi:10 - regulation: A conceptual framework for Zimmerman, B.J. (1994). Dimensions of academic self education. In D. Schunk & B. Zimmerman (Eds.), Self - regulation of learning and performance: 301). Hillsdale, NJ: Erlbaum. Issues and educational applications (pp. 283 - * Grit and legal education . Pace Law Review, Forthcoming; Zimmerman, E., & Brogan, L. (2015). Drexel University Thomas R. Kline School of Law Research Paper No. 2015 - A - 06. Available at SSRN:

55 GRIT META ANALYSIS 55 - Table 1 Analytic Computations - Artifact Distributions used for Meta k Variable α SD Mean α α 46 .79 0.07 Overall Grit Perseverance .71 0.13 10 .74 Consistency 0.11 11 Overall Academic 4 .88 0.04 Performance Overall GPA 4 .88 0.04 College GPA .9 0 0 .00 2 0 .9 Graduate School GPA 0 .00 2 2 .86 0.06 High School GPA Intent to Persist at Current Employer 4 .93 0.05 Intent to Persist in College 2 .7 0 0.25 Cognitive Ability 2 .87 0.07 Agreeableness 11 .75 0.14 0.09 Conscientiousness 17 .79 Emotional Stability .81 0.07 11 .83 Extroversion 0.1 0 11 12 .76 0.06 Openness to Experience Generalized Self - Efficacy 2 .9 0 0.06 Optimism 1 .86 Gratitude 3 .81 0.08 Mental Toughness 3 .84 0.12 0.17 Hope 4 .77 Positive Affect .89 1 Life Satisfaction 5 .87 0.03 1 Work Satisfaction (College and Job) .89 Depression 5 .86 0.05 Happiness 3 .69 0.14 Resiliency 3 .88 0.07 Self - Control 3 .84 0.04 Note: k is the number of reliability estimates in distribution, Mean α is the mean of the reliability estimates, SD is α α the standard deviation of the reliability estimates.

Related documents