GRIT META-ANALYSIS
1
Much Ado about Grit: A Meta-Analytic Synthesis of the Grit Literature
Marcus Credé
Department of Psychology, Iowa State University
mcrede@iastate.edu
Michael C. Tynan
Department of Psychology, Iowa State University
mtynan@iastate.edu
Peter D. Harms
School of Management, University of Alabama
pdharms@cba.ua.edu
ACCEPTED FOR PUBLICATION AT:
JOURNAL OF PERSONALITY AND SOCIAL PSYCHOLOGY
(Not the version of record)
PLEASE DO NOT CITE OR DISTRIBUTE WITHOUT PERMISSION
Corresponding Author: Marcus Credé, Department of Psychology, W112 Lagomarcino Hall, 901
Stange Road, Iowa State University, Ames, 50011-1041, mcrede@iastate.edu.
GRIT META-ANALYSIS
2
Abstract
Abstract
Grit has been presented as a higher-order personality trait that is highly predictive of both
success and performance and distinct from other traits such as conscientiousness. This paper
provides a meta-analytic review of the grit literature with a particular focus on the structure of
grit and the relation between grit and performance, retention, conscientiousness, cognitive
ability, and demographic variables. Our results based on 584 effect sizes from 88 independent
samples representing 66,807 individuals indicate that the higher-order structure of grit is not
confirmed, that grit is only moderately correlated with performance and retention, and that grit is
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
conscientiousness. 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 perseverance facet.
Keywords: grit; performance; meta-analysis; perseverance of effort; consistency of interest
GRIT META-ANALYSIS
3
Interest in non-cognitive variables as potential predictors and determinants of academic
performance has been spurred by meta-analytic findings that variables such as study habits and
study skills (Credé & Kuncel, 2008; Robbins et al., 2004), personality traits such as
conscientiousness (Porapat, 2009, 2014), test anxiety (Hembree, 1988; Seipp, 1991), adjustment
(Credé & Niehorster, 2012), emotional intelligence (Perera & DiGiacomo, 2013), and learning
strategies (Credé & Phillips, 2011; Richardson, Abraham & Bond, 2012), exhibit relations with
academic performance that are often comparable to relations between admissions test scores and
academic performance (e.g., Hezlett et al., 2001). Many of these non-cognitive characteristics
also appear to be responsive to interventions. For example, meta-analytic reviews have
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 non-cognitive variable that has received widespread attention and that has
been widely touted as an important predictor of success and performance is a personality trait
referred to as grit (Duckworth, Peterson, Matthews, & Kelly, 2007; Duckworth & Quinn, 2009).
Grit is defined as “perseverance and passion for long-term goals” (Duckworth, Peterson,
Matthews, & Kelly, 2007, p. 1087) and as “… not just resilience in the face of failure, but also
having deep commitments that you remain loyal to over many years” (Duckworth as quoted in
Perkins-Gough, 2013, p. 16). Duckworth et al. argue that grit 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 (2013a) has even argued that grit is as good or even a better predictor of success than
cognitive ability; a strong claim given meta-analytic findings that cognitive ability correlates
GRIT META-ANALYSIS
4
about ρ = .50 with performance in academic and work settings (Sackett et al., 2012; Schmidt &
Hunter, 1998).
Despite the widespread enthusiasm for grit as a potentially novel predictor and
determinant of performance there are sound empirical and theoretical reasons why a critical
reappraisal of the nature of the grit construct, its contribution to our understanding of
performance, and its general position within the nomological network may be warranted. It is the
goal of this paper to present findings from a meta-analytic synthesis of the rapidly growing
empirical literature on grit 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
literature relating to five core theoretical features of grit: 1) the proposed hierarchical structure of
grit, 2) the relation of grit with performance, 3) the distinction of grit from conscientiousness, 4)
the distinction of grit from cognitive ability and 5) the lack of group differences on grit.
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
perseverance and consistency), respectively refer to the tendency work hard even in the face of
setbacks and the tendency to not frequently change goals and interests. Both are thought to
contribute to success: persistence because the process of attaining mastery in a field often
involves initial failures 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 face of
obstacles or who constantly change their interests are unlikely to ever engage in enough
deliberate practice to achieve high levels of performance. The distinction between the two facets
GRIT META-ANALYSIS
5
is reflected in the subscales of the two primary self-report inventories used to measure grit: 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
(https://sites.sas.upenn.edu/duckworth/pages/research). Although some researchers examine
perseverance and consistency as two separate constructs, most research on grit only reports
findings at the level of the overall grit score.
This practice appears to be informed by two factors. First, Duckworth et al. (2007) in
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 examining the theoretical
higher-order factor structure of grit using confirmatory factor analysis (CFA), and claimed
support for the higher-order structure based on their finding that the higher-order model
(comprised of two first-order factors and one second-order factor) exhibits significantly better fit
than a single-factor model. However, the reported analysis is problematic because a model with
one second-order factor and two first-order factors is not identified at the higher-order level
(Kline, 2011); this means that no unique loadings onto the higher-order factor can be computed
without additional constraints being imposed. However, it does not appear that Duckworth and
Quinn imposed the type of equality constraints on the loadings of the first-order factors onto the
higher-order factor that would be required to achieve identification because they report nonidentical loadings of the first-order factors onto the second-order factor. Importantly, even if an
equality constraint 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
second-order factor (see Credé & Harms, 2015 for a discussion). That is, a model in which the
GRIT META-ANALYSIS
6
persistence and consistency facets are simply two correlated 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). The CFA strategy for determining whether a higherorder grit construct exists is therefore not particularly 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
two theoretical facets of grit; high correlations would suggest that a higher-order construct is
plausible. However, empirical estimates of the strength of this relation exhibit substantial
variability, with some reporting correlations that are close to zero (e.g., Chang, 2014; Datu,
Valdez, & King; 2015; Jordan, Gabriel, Teasley, Walker, & Schraeder, 2015), while others have
reported very strong correlations (e.g., Arslan, Akin, & Çîtemel, 2013; Meriac, Slifka, & 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-order structure. This will in turn help to
determine whether the practice of simply summing across the perseverance and persistence items
to compute an overall grit score (as recommended by Duckworth et al., 2007) is reasonable, or
whether the two facets should be considered separately.
Grit as a Predictor of Success and Performance
Proponents of grit as a predictor of performance have argued that between-person
differences in grit 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,
GRIT META-ANALYSIS
7
individuals with high levels of grit are thought to be able to better utilize their capabilities
because they are less distracted by short-term goals and less discouraged by the failures and
setbacks that are commonly encountered in many performance domains. Indeed, Duckworth et
al. (2007) described that the importance of grit for success had long been noted by prior research
into highly accomplished individuals (e.g., Howe, 1999). Arguments for the importance of grit
are also in line with work on the development of expertise that has highlighted the importance of
sustained deliberative practice (e.g., Ericsson, Krampe, & Tesch-Rӧmer, 1993; Krampe &
Ericcson, 1996). Indeed, recent work by Duckworth, Kirby, Tsukayama, Berstein and Ericsson
(2011) has explicitly tied grit to success in spelling bees via the mediating mechanism of
deliberative practice. That is, individuals who are high on grit are more likely to engage in the
amount of deliberative practice that is required to achieve expertise.
At the same time there are a number of theoretically plausible moderators of the gritperformance relation that suggest that the relation may not be strong in all - or even most settings. First, the grit-performance relation may be moderated by the nature of the performance
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
and the manner in which performance is to be attained is relatively clear (see MacNamara,
Hambrick & Oswald, 2014). Thus, grit may be an excellent predictor of an individual’s ability to
complete military basic training or succeed in well-defined academic tasks, but be less well
related to performance on tasks that are very easy (thus not requiring grit) or performance on
tasks that are novel and ill-defined and that therefore require both creativity and the willingness
to abandon unsuccessful strategies (i.e., tasks on which grit may be counterproductive). Second,
the grit-performance relation may be moderated by other individual differences such as ability
GRIT META-ANALYSIS
8
and meta-cognition (Credé & Phillips, 2011). That is, high levels of grit may not necessarily be
adaptive unless it is accompanied by the general potential or ability to succeed in a domain and
the ability to engage in 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 &
Pintrich, 1997; Zimmerman, 1994, see also Kohn, 2014) has identified as important determinants
of learning and performance. For example, cadets who lack some minimum level of physical
ability to pass the highly strenuous tests of physical ability in a military academy are unlikely to
benefit substantially from grit. Similarly, a cadet who is unable to recognize that a particular
approach to studying for class material is not working is unlikely to perform well in academic
courses.
Third, the grit-performance relation may be moderated by the level of grit itself. Very
high levels of grit may become dysfunctional if they reduce the likelihood of help-seeking
behaviors that have themselves been linked to performance (e.g., Karabenick, 2003) or if they
increase the likelihood 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). This would, in turn, suggest that interventions
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, &
Lester, 2015). A similar non-linear relation may also exist for grit.
GRIT META-ANALYSIS
9
A more nuanced conceptualization of grit’s contribution to performance also seems
warranted when reconsidering the biographical details of some of the highly accomplished
scientific figures referenced by Duckworth et al. (2007) to highlight the importance of grit. For
example, it is true that Albert Einstein persisted for many years in his attempt to develop the field
equations that represent the mathematical description of General Theory of Relativity, but it is
also true that Einstein persisted for years in pursuing an avenue of investigation that was based
on an earlier mathematical error. Einstein also 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,
persistence in this narrow case almost resulted in “failure” and Einstein only “succeeded” in this
particular endeavor once he recognized his mathematical limitations and sought the help of more
accomplished mathematicians.
The possibility that the grit-performance relation is not uniformly strong is also strongly
suggested by an examination of the empirical literature. Some have found that grit 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 relations between grit scores and
indicators of success. This is particularly the case for studies examining academic success.
Indeed, many of the reported relations between grit and academic success are weaker or equal to
the relation of ρ = .21 and ρ = .23 that has been reported between conscientiousness and
academic performance in two recent large-scale meta-analytic reviews (e.g., Porapat, 2009;
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 whether the observed variability in
relations is simply a function of sampling error and other study artifacts (e.g., differences in the
reliability in the measurement of variables), or if this variability reflects the presence of
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
(Duckworth et al., 2007; Duckworth & Quinn, 2009) appear to confuse odds ratios with
probabilities in their discussion of logistic regression results, resulting in incorrect inferences
about the size of observed effects. This misunderstanding may have led readers to infer a much
greater predictive power for grit scores than is warranted. For example, Duckworth and Quinn
(2009) 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 incorrect
because approximately 94% of all cadets successfully completed the program. A relatively small
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 largely
equivalent in their relation with important outcomes as suggested by Duckworth and colleagues
GRIT META-ANALYSIS
11
(2007, p. 1091). This might, in turn, change the manner in which grit scores are presented and
interpreted. Many grit researchers follow the recommendations by Duckworth et al. and examine
only an overall grit score. However, facets are often better predictors than broad traits (e.g.,
Paunonen & Ashton, 2001), and substantial differences between the grit facets to predict
important criteria might suggest that this scoring strategy should be revisited.
A meta-analytic synthesis of the strength of the relation between grit and success will
also help to inform judgments about whether interventions designed to enhance grit are likely to
have an impact on performance. Initial reports of the high predictive validity of grit scores and
their relative independence from indicators of cognitive ability, combined with claims that grit
can be taught (Perkins-Gough, 2013), has resulted in some schools implementing interventions
designed to increase students’ levels of grit. For example, The Knowledge is Power Program
(KIPP) network of public charter schools is training its teachers to foster grit in their pupils
(Shechtman, DeBarger, Dornsife, Rosier, & Yarnall, 2013), while many school districts across
the US are reportedly considering integrating the teaching of grit into curricula (Cohen, 2015).
Grit was even highlighted as a promising focus of school interventions in a US Department of
Education report (Shechtman et al., 2013). The time and resources that are likely to be devoted to
grit-based interventions in schools are likely to be non-trivial and should therefore only be based
on the best available knowledge about 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,
determination, persistence of motives, and volitional perseveration date back over 80 years (see
Ryans, 1939 for an early review). More recently, researchers have investigated a variety of other
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-control, and order (Roberts, Chernyshenko, Stark & Goldberg,
2005). The conceptual similarities between these constructs and grit raises the possibility that the
proponents of grit may have fallen victim to what Kelley (1927) referred to as the “jangle
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 discriminant
validity of grit from these other constructs using the types of methodologies (e.g., Mulitraitmultimethod matrices, confirmatory factor analyses) commonly employed to determine
discriminant validity, despite the fact that grit has been explicitly presented as a construct that is
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 conscientiousness has often been reported by researchers but even
for conscientiousness (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 definition 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)
for the self-discipline facet (“capacity to begin tasks and follow through to completion despite
GRIT META-ANALYSIS
13
boredom or distractions”) and the achievement striving facet of conscientiousness (“need for
personal achievement and sense of direction”). This theoretical similarity is also reflected in the
considerable similarity in the items that are found in the Duckworth et al. (2007) perseverance
subscale and items in widely used inventories of conscientiousness such as those provided by the
International Personality Item Pool (IPIP, Goldberg, 1999). For example, perseverance items
such as “I finish whatever I begin” and “I am a hard worker” are very similar to IPIP items used
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
instead more similar to IPIP items used to measure the adventurousness facet of openness to
experience and the IPIP measure of planfulness that is modeled after the Achievement via
Conformance scale found in the California Personality Inventory (Gough, 1996).
A cursory examination of the empirical grit literature also suggests that the grit –
conscientiousness relation may be much stronger than is commonly assumed. While some (e.g.,
Cooper, 2014) 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 and Cutton
(2012), for example, report a correlation of ρ = .92 based on 1165 college students, Engel (2013)
reports a correlation of ρ = .95 based on a smaller sample of 88, and Meriac, Slifka, and LaBat
(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 = 1,554); ρ = .90 (N = 706), ρ = .83 (N = 190), and ρ = .80 (N = 1,308) after
correcting for the unreliability of both conscientiousness scores and grit scores. High correlations
such as these have led some (e.g., MacCann & Roberts, 2010) to suggest that grit should be
GRIT META-ANALYSIS
14
considered a facet of conscientiousness – a position that seems theoretically plausible when
considering that both grit and the self-control or self-discipline facet of conscientiousness focus
on the deferment of short-term gain for long-term goals (see Costa & McCrae, 1992; Roberts,
Bogg, Walton, Chernyshenko, & Stark, 2004). High observed correlations between grit and
conscientiousness are also of concern when considering that a concurrent assessment of the same
personality trait using different scales typically yields much lower correlations of around r = .50
(e.g., Pace & Brannick, 2010; Miller, Price & Campbell, 2012). Meta-analytic synthesis of the
literature on the grit-conscientiousness relation will help clarify whether the strength of the
relation is such that grit might be a case of the “old wine in 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,
2013b) has also suggested a negative relation between grit and cognitive ability in noting that
“… gritty people, on average, tend to be slightly less talented”. This distinction, if correct,
suggests that grit might explain unique variance in performance over and above the substantial
variance 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 in 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 ability test scores but this approach seems reasonable when
considering that cognitive ability tests administered in research settings have questionable
validity because many test takers will not be motivated to exert maximal effort on ability tests in
GRIT META-ANALYSIS
15
a low stake setting (Duckworth, Quinn, Lynam, Loeber, & Stouthamer-Loeber, 2011). Most
commonly used college admissions test scores are highly correlated with cognitive ability test
scores (e.g., Frey & Dotterman, 2004) and are also taken under maximal performance (i.e., high
stakes) conditions and are therefore likely to represent a sound indicator of cognitive ability.
Research examining the relation between grit scores and admissions test scores have largely
found very weak relations (e.g., Chang, 2014; Duckworth et al., 2007; Eskreis-Winkler,
Shulman, Duckworth, & Beal, 2014; Kelly, Matthews, & Bartone, 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
attractive in settings where scores are used for selection purposes or for making other high-stakes
decisions because the likelihood of adverse impact on legally-protected groups is reduced. 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
ethnicities and gender are likely to be more modest than those observed for ability measures.
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 &
Beal., 2014), gender (e.g., Allen, 2014; Davidson, 2014), and ethnicity (e.g., Chang, 2014,
GRIT META-ANALYSIS
16
Eskreis-Winkler, 2014) have been mixed and a meta-analytic synthesis will help to clarify
whether average grit scores are largely similar across groups.
The General Position of Grit within the Nomological Network
Grit has been not only been related to performance, cognitive ability, and
conscientiousness but also to a wide array of other variables reflecting either states or traits.
These include Big Five traits (e.g., Eskreis-Winkler et al., 2014), optimism (e.g., Lovering et al.,
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.,
2015), and life satisfaction (e.g., Samson et al., 2011). A meta-analytic synthesis of the relation
of grit with these other variables will further help to clarify the general position of grit within the
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
predictor variable and a criterion variable. However, the American Psychological Association
(Wilkinson & 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 –
GRIT META-ANALYSIS
17
especially when the predictor provides information about the criterion that is not provided by
other predictors and when the criterion is important.
Prior meta-analyses of predictors of academic performance have identified two variables
that correlate at approximately ρ = .50 with academic performance in college: 1) indicators of
cognitive ability such as scores on the SAT and 2) prior academic performance such as high
school GPA (Sackett et al., 2012). Other predictors that correlate approximately at ρ = .40 with
academic performance include study skills and study habits (Credé & Kuncel, 2008), 4)
academic adjustment (Credé & Niehorster, 2012), academic self-efficacy (Robbins et al., 2004)
and 6) class 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
weaker but practically still very meaningful relations of around ρ = .20 with academic
performance include: 1) specific learning strategies (Credé & Phillips, 2011), 2) emotional
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
the best predictors are: academic self-efficacy (ρ = .36) and academic-related skills (ρ = .37,
Robbins et al., 2004), institutional attachment (ρ = .29) and social adjustment (ρ = .25, Credé &
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
GRIT META-ANALYSIS
18
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
term “grit”. This yielded a total of 778 potential data sources. These search results were
supplemented by an examination of the reference lists of identified sources. We also examined
the first 500 search results of the internet using the Google search engine and the search term
“grit” to identify additional unpublished sources of information. Potential sources for inclusion
were first screened by examining the abstract and title of the source and all possible sources were
then examined more closely to determine if the reported data met the inclusion or exclusion
criteria.
Inclusion and Exclusion Criteria
Sources were included in our review if 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.,
means and standard deviations for two criteria groups). The year of publication, source of the
material, and country of origin of the data were not used to exclude any sources, although nonEnglish sources were excluded. Sources were also excluded if they reported correlations for
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 relation at younger ages was very different to the relation at older ages.
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 effect
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
GRIT META-ANALYSIS
19
the necessary information. Data from a total of 73 studies representing data from 88 unique
samples and 66,807 individuals was ultimately included in the analyses.
Coding Procedure
All articles were coded by two of the authors using a systematic coding procedure, one of
which has extensive experience coding articles for meta-analytic analyses. An accuracy check
revealed 98.8% agreement in coding across the four most important coding categories. 60% of
coding errors were errors of commission (e.g., incorrect coding of an effect size) and 40% were
errors of omission (e.g., an effect size that could have been computed was not coded). All
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 two facets: consistency and perseverance, 9) the source of the publication (peer-reviewed
versus not peer-reviewed), and 10) the year of publication. Self-reported grades are very highly
correlated with actual grades (Kuncel, Credé & Thomas, 2005) and we therefore 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
GRIT META-ANALYSIS
20
Our coding process involved three transformations of data. First, we used formulas
presented by Hunter and Schmidt (2004) to calculate estimates of the correlation between grit
and correlate variables when the original sources had artificially dichotomized the correlate
variable (e.g., presenting grit scores for “low” and “high” scoring students). The artificial
dichotomization of data results in downwardly biased estimates of the population correlation if
such a correction is not made. Second, we used the formula presented by Ghiselli, Campbell and
Zedeck (1981) to calculate composite correlations when the original source only presented
correlations involving the facets of grit and/or facets of the correlate variable. For example,
Bowman et al. (2015), presents correlations among the two facets of grit, Fall GPA, and Spring
GPA; the six correlations among these four variables were used to arrive at an estimate of the
correlation between overall grit and overall GPA. Mosier reliability estimates for composite
variables (Mosier, 1943) were also calculated whenever possible. Finally, we computed pointbiserial correlations between retention and grit when the mean and standard deviation of grit
scores were reported for both the retained and non-retained group.
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 ways we present meta-analytic
estimates for ten criterion categories. First, we present separate meta-analytic estimates of the
relation between grit and high school GPA, college GPA, and post-graduate GPA. Second, we
aggregate these into a broader general GPA criterion category (i.e., GPA across all three
educational levels). For this general GPA criterion category we relied on the correlations for the
more recent college GPA when authors reported correlations involving both high school GPA
GRIT META-ANALYSIS
21
and college GPA. We then also combined this general GPA category further with correlations
involving grades in individual courses to form an Academic Performance criterion category.
Third, we present meta-analytic estimates of the point-biserial correlation between grit and
retention. Most studies examining retention do so in an academic or military setting (e.g., degree
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
therefore present meta-analytic 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 also
present meta-analytic estimates of the relation between grit and the intent to persist in both
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
therefore be described as falling somewhere along the state-trait continuum (e.g., happiness,
depression, positive affect). We use the descriptors of these variables as given in our source
articles to group these various state and trait variables into categories and computed metaanalytic 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).
Statistical Method
GRIT META-ANALYSIS
22
We used the Hunter and Schmidt (2004) interactive meta-analytic method based on a
random-effects model to arrive at population estimates of the size of the relations between grit
and other variables. The Schmidt and Le (2004) software was used to compute meta-analytic
estimates of the relations involving grit and we corrected for unreliability in the measurement of
the dependent variable and unreliability in the measurement of the independent variable. Grit
scores are likely to be exhibit some level of range restriction in many samples but the absence of
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.
Corrections for unreliability. In order to correct for the attenuating effect of
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
reliability distributions are described in Table 1. The included studies did not report information
on the reliability of grades but in order to facilitate an apples-to-apples comparison with the
recent meta-analytic summary of the relation between conscientiousness 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
(2009). For GPA information taken from records we used the average of reliability estimates
(alpha = .90) for all courses across four years as reported by Bacon and Bean (2006) while for
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
analytical approach. More than half of the studies examining the relation between grit and
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
odds ratios, but odds ratios cannot be directly transformed into a point-biserial correlation when
the independent variable is treated as a continuous variable. We therefore present two metaanalytic estimates for the grit-retention relation. We provide one estimate based purely on those
studies that report data that could be transformed into a point-biserial correlation, and then
provide another estimate that includes correlation values computed by taking the root of the
Nagelkerke R2 values reported for those studies that reported odds ratios from bivariate logistic
regression models. Nagelkerke R2 values tend to be too high as an estimate of the strength of the
bivariate relation (Allison, 2014) but we include these values in order to provide readers with an
estimate based on the most complete data. We remind readers that this estimate is likely to be
upwardly biased.
We present meta-analytic estimates of the relations of overall grit (or the two grit facets)
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) robs refers to the sample-size weighted average observed correlation, 4)
ρ refers to the estimate of the population correlation, 5) SDρ refers to the estimate of the standard
deviation of effect sizes after taking into account the variability that is due to sampling error and
differences in the reliability of measurement between studies, and 6) 10%CV and 90%CV
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, wide credibility intervals
indicate that the correlation can be expected to vary widely across settings.
GRIT META-ANALYSIS
24
In order to examine whether grit scores explain incremental variance in academic
performance outcomes over and above the variance explained by conscientiousness we
constructed a full meta-analytic intercorrelations matrix between grit, conscientiousness, and
academic performance by importing the conscientiousness correlations of ρ = .21 for high school
GPA and ρ = .23 for college GPA as reported by Porapat (2009), and using the average of these
(ρ = .22) for overall academic performance. These correlation matrices 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 relations of different
magnitudes we report separate meta-analytic estimates based on those studies that were
published in peer-reviewed journals and all other studies (e.g., dissertations, conference
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 significant at alpha = .05.
[INSERT TABLE 2 HERE]
GRIT META-ANALYSIS
25
Second, we use Egger’s Test of funnel-plot asymmetry (Eggers, Smith, Schneider, &
Minder, 1997) to examine whether there is evidence in the meta-analysis for the exclusion of
small studies with weak effects. Studies with small sample sizes that find weak effects may not
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 effect size for each sample onto the precision of the effect size estimate. The intercept of the
regression line provides information about the size of any asymmetry; statistically significant
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
relation we only perform a single test of asymmetry for the relation based on the largest number
of studies: the relation between grit and overall academic performance. For this relation Egger’s
Test indicated no significant asymmetry with the intercept being negative but weak and not
significantly different from zero (a= -.15, p=.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
direction by missing studies. We therefore present meta-analytic estimates of the relation of grit
with criteria, state and trait variables, and demographic variables in Tables 3-5.
[INSERT TABLES 3 – 5 ABOUT HERE]
Relation between Perseverance and Consistency
Our meta-analytic estimate of the relation between perseverance and consistency (k = 17,
N = 22,048, ρ = .60, SDρ = .21), indicates a generally strong relation although the width of the
credibility interval suggests that the strength of this relation is substantially moderated. As an
GRIT META-ANALYSIS
26
exploratory follow-up analysis we compared the relation observed for the two different grit
scales and found a stronger relation when researchers relied on the short grit scale (k = 11, N =
18,996, ρ = .66, SDρ = .15) than when the original grit scale was used (k = 6, N = 3,052, ρ = .27,
SDρ = .17).
Relations with Criteria
Overall grit exhibits a relation with overall academic performance of ρ = .18 (k = 39, N =
13,141, SDρ = .11) and ρ = .17 with the overall GPA criterion (k = 37, N = 12,601, SDρ = .10).
Among the academic performance criteria grit was approximately as strongly related to college
GPA (k = 30, N = 10,526, ρ = .17, SDρ = .10) as it was to high school GPA (k = 17, N = 6,364, ρ
= .16, SDρ = .14). Contrary to early assertions by Duckworth et al. (2007) that both facets
predicted success outcomes equally well, the perseverance facet of grit exhibited much stronger
relations with all academic performance criteria than the consistency facet. For example,
perseverance correlated at ρ = .26 (k = 11, N = 5,221, SDρ = .12) with overall academic
performance while consistency correlated at only ρ = .10 (k = 11, N = 5,221, SDρ = .02). A
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 differed significantly (p<.001) in all
four cases.
Grit correlated with retention at ρ = .12 when the marital success study is included (k =
11, N = 17,525, SDρ = .09), at ρ = .18 (k = 10, N = 11,163 SDρ = .03) when it was excluded, and
at, ρ = .16 (k = 5, N = 2,705, SDρ = .06) if the upwardly biased correlations estimated from
Nagelkerke R2 values are excluded. Grit was correlated at ρ = .21 (k = 7, N = 4,116, SDρ = .00)
with performance in non-academic domains. The relation between grit and the intent to persist in
GRIT META-ANALYSIS
27
college and with the current employer was ρ = .18 (k = 5, N = 3,967, SDρ = .00), and ρ = .15 (k =
4, N = 519, SDρ = .00) respectively.
Relation with State and Trait Variables
Consistent with the claim that grit and cognitive ability are largely orthogonal, grit
exhibited only a very weak relation with cognitive ability (k = 21, N = 11,513, ρ = .05, SDρ =
.12). Similarly weak relations with cognitive ability were also observed for both the perseverance
facet (k = 5, N = 2,204, ρ = -.01, SDρ = .04) and consistency facet (k = 5, N = 2,204, ρ = .00, SDρ
= .00).
Grit exhibited much stronger relations with other trait variables. Conscientiousness was
very strongly correlated with overall grit (k = 22, N = 18,826, ρ = .84, SDρ = .07) and also with
perseverance (k = 8, N = 4,967, ρ = .83, SDρ = .14) and consistency (k = 8, N = 4,967, ρ = .61,
SDρ = .17). Grit also exhibited a very strong relation with self-control (k = 4, N = 2,615, ρ = .72,
SDρ = .05), a variable that is often seen to be a facet of conscientiousness (e.g., Roberts et al.,
2005). Grit also exhibited a relatively strong relation with emotional stability (k = 14, N =
14,501, ρ = .41, SDρ = .04) but this should not be surprising when considering that low levels of
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 exhibited relatively strong relations 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, SDρ = .05), generalized self-efficacy (k = 3, N = 1,908, ρ
= .43, SDρ = .11), mental toughness (k = 6, N = 3,817, ρ = .46, SDρ = .08), positive affect (k = 3,
N = 670, ρ = .46, SDρ = .03), and depression (k = 5, N = 3,865, ρ = -.48, SDρ = .12).
GRIT META-ANALYSIS
28
Relation with Demographic Variables
As expected, the relations between grit and demographic variables such as gender (k =
25, N = 18,750, ρ =.05, SDρ = .07), year in school (k = 4, N = 2,961, ρ =.05, SDρ = .05), and
ethnic minority status (k = 9, N = 15,261, ρ = .01, SDρ = .01) were all very weak with the
exception of age which exhibited a slight positive correlation with overall grit (k = 22, N =
12,349, ρ = .12, SDρ = .04) in line with the prediction by Duckworth et al. (2007) that grit would
increase with age. This increase is similar to the general increase in conscientiousness observed
with age (Roberts et al., 2006).
[INSERT TABLE 6 HERE]
Incremental Validity
The incremental validity estimates from these regression results are summarized in Table
6. Results for Model 1 indicate that overall grit explains no variance in either overall academic
performance or high school GPA after controlling for conscientiousness, and explains only a
very small amount of incremental variance in college GPA (i.e., ΔR = .004). Importantly,
conscientiousness explains incremental variance in these outcomes if first controlling for overall
grit. Results for Model 2 show that perseverance explained a substantial amount of incremental
variance in overall academic performance (ΔR = .040), high school GPA (ΔR = .085), and a
somewhat lower amount for college GPA (ΔR = .023). 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
possible suppressor effect. Perseverance explained relatively large amounts of unique variance in
GRIT META-ANALYSIS
29
three criteria even after controlling for both conscientiousness and consistency (Model 5).
Overall, the incremental validity findings suggest that both overall grit and the consistency facet
of grit add little to our ability to understand or predict academic performance, while the
perseverance facet does offer an important improvement in explanatory power.
Discussion
Proponents of grit have asserted that grit is a higher-order construct composed of a
perseverance facet and a consistency facet, that grit scores are highly predictive of success (and a
better predictor of success than cognitive ability), and that grit scores provide information about
individuals that is meaningfully distinct from conscientiousness. Three primary findings from
our meta-analytic review of the grit literature suggest that the validity of these assertions may
need to be revisited; although our findings 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
the ability to predict performance. That is, perseverance is a much better predictor of
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, overall grit exhibits relations with academic performance and retention that are
only modest and that do not compare favorably with other well-known predictors of academic
GRIT META-ANALYSIS
30
performance such as cognitive ability (Sackett et al., 2012), study habits and skills (Crede &
Kuncel, 2008), and academic adjustment (Crede & Niehorster, 2012). Indeed, meta-analytic
reviews of the literatures for some of these other predictors report correlations with academic
performance and retention that are more than twice as big as those observed for overall grit in
this review. At the same time it should be remembered that variables that exhibit small 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
college students. Similarly, even a small increase in the ability to predict future performance in a
selection setting may yield very substantial financial benefits for an organization (see Hunter &
Hunter, 1984 for a discussion). Such a benefit would be particularly large if the variable in
question reflected information about individuals that was distinct from the information reflected
by other well-known predictors of performance and retention.
Our third primary finding suggests that the incremental value of grit for the prediction of
performance is likely to be limited. Grit scores exhibited very strong correlations with
conscientiousness and with self-control – a facet of conscientiousness. Indeed, the size of the
correlation (ρ = .84) with overall conscientiousness 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
different global measures of conscientiousness (ρ = .63; Pace & Brannick, 2010). This, in turn,
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.
McCornack (1956), of course, illustrated that two variables can be very strongly correlated but
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 = .17) are
largely identical (if somewhat weaker) to those reported for conscientiousness in the recent metaanalytic review by Porapat (2009): ρ = .21 for middle/high school GPA and ρ = .23 for college
GPA.
Although our findings indicate that a critical re-appraisal of the grit construct is
warranted, three 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 traditional predictors of retention such as
cognitive ability and high school grades – although not as well as some other non-cognitive
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 exhibits substantially higher criterion validity for the grade criteria than the
consistency of interest facet. Indeed, the observed criterion validity of perseverance for the high
school GPA criterion is also significantly higher than the criterion validity observed for overall
grit scores and also for conscientiousness (Poropat, 2009). This suggests that the focus of the grit
researchers should shift to perseverance as the most promising avenue of future research. Third,
our hierarchical regression results based on meta-analytic estimates suggest that perseverance of
GRIT META-ANALYSIS
32
effort scores explain incremental variance over and above conscientiousness in the various grade
criteria. This is, of course, encouraging, but the only moderately high correlations among scores
on most personality traits assessed via two different inventories (see Pace & Brannick, 2010)
mean that such incremental validity findings would be observed even if grit was simply a
different manifestation of conscientiousness as our other results suggest.
Even modest predictor-criterion relations 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.,
2010; Paunesku et al., 2015) suggest that grit interventions may have some positive effect.
Although we do believe that our results regarding the validity of the perseverance facet offer
some promise we also believe that our overall results should lead to a re-evaluation of the
appropriateness of planned or existing grit interventions. Schools and colleges have limited
resources to devote to interventions and are likely to be best served by focusing those resources
on variables that have been demonstrated to be 1) most strongly related to performance and
persistence/retention and 2) responsive to interventions. Fortunately there are a number of
variables that meet both of those requirements. For example, study skills and habits have been
shown to correlate approximately ρ = .40 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
academic performance (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
interventions such as orientation programs (see Credé & Niehorster, 2012 for a review). Simple
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). Study skills and study habits, adjustment to college, and class attendance are
thus far more strongly related to academic performance and retention than grit, and there is
sound evidence that interventions can improve students’ standing on these constructs (especially
for study skills 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,
may have resulted in inflated estimates of the grit-academic performance relation because
individuals’ knowledge of their academic performance may influence their responses to the
measure of grit. Second, there are at least three reasons why the grit literature may be
characterized by a non-trivial amount of range restriction. All of the examined studies relied on
self-reports of grit and the social desirability of grit items may have resulted in range restriction
in self-reported grit scores. Individuals may also generally not be aware of their true level of grit
and unintentionally report inflated levels of grit (Kruger & Dunning, 1999). Further, 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
correct for range restriction in our meta-analysis because of the lack of normative data and the
variability in how the grit scales were used by researchers but future research may be able to
GRIT META-ANALYSIS
34
estimate the level of range restriction that is present in samples. Third, many of the studies
examining the relation between grit and retention were characterized by very high base rates of
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 examined 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 standard
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
us to comment with relative confidence on average population effect sizes the literature is not yet
large enough to allow moderator analyses characterized by high power (Hunter & Schmidt,
2004). 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 five broad issues. First,
researchers should attempt to examine whether grit exhibit the type of stability that is associated
with other personality traits or whether it is responsive to interventions. Grit interventions will
need to be tested to evaluate the malleability of grit but there are sound theoretical 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 result 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
establish the degree to which grit is truly trait-like and also help to clarify the type of grit
GRIT META-ANALYSIS
35
interventions that are likely to be most effective. Second, grit researchers should consider
examining criteria that span to different domains (e.g., work settings), a greater range of
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 examining the potential moderators of the gritperformance 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
role of the level of grit. Fourth, it may also be useful to examine the degree to which scores on
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 effort 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 Response 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, Niehorster, & 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.
We hope that greater rigor in scale development, a greater focus on the perseverance facet, and a
GRIT META-ANALYSIS
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.
36
GRIT META-ANALYSIS
37
References
(* denotes inclusion in meta-analysis)
*Allen, B. R. (2014). An analysis of student success predictors for technical college persistence.
(Unpublished doctoral thesis). Edgewood College, Madison, Wisconsin.
Allison, P.D. (2014). Measuring fit for logistic regression. Research paper 1485-2014. SAS Institute.
*Anestis, M. D., & Shelby, E. A. (2015). Grit and perseverance in suicidal behavior and non-suicidal
self-injury. Death Studies, 39(4), 211-218. doi:10.1080/07481187.2014.946629
*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
between emotion dysregulation and both suicidal behavior and non-suicidal self-injury.
Comprehensive Psychiatry, 55, 1820-1830. doi:10.1016/j.comppsych.2014.07.007
*Arouty, D. (2015). Does character matter? The role of grit and resilience in predicting academic
success. (Unpublished doctoral dissertation). Hofstra University, Hempstead, New York.
*Arslan, S., Akin, A., Çîtemel, 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 research studies: An invaluable but neglected opportunity.
Journal of Advanced Academics, 26, 274-293. doi:10.1177/0273475305284638
*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.
GRIT META-ANALYSIS
38
Blalock, D.V., Young, K.C., & Kleiman, E.M. (2015). Stability amidst turmoil: Grit buffers the effects
of negative life events on suicide ideation. Psychiatry Research, 228, 781-784.
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
intentions. Social Psychology and Personality Science, 6(6), 639-654.
doi:10.1177/1948550615574300
*Buller, E. F. (2012). The relationship between grit and academic, military and physical performance at
the United States Military Academy. (Unpublished doctoral dissertation). University of Kansas,
Lawrence, Kansas.
Camara, W.J., & Schmidt, A.E. (1999). Group differences in standardized testing and social
stratification. College Entrance Examination Board. New York.
*Chambers, R., Brawley, A. M., Castille, C. M., & Tubré, T. (2012). Examining the validity of the
Resilience Quotient. Poster presented at the 26th National Conference on Undergraduate
Research, Ogden, Utah, March 29-31, 2012.
Chamorro-Premuzic, T., Furnham, A., & Ackerman, P. L. (2006). Incremental validity 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
dissertation). University of Miami, Miami, Florida.
Cohen, R.M. (2015). Will students soon be tested for “grit”? The American Prospect. Retrieved from
http://prospect.org/blog/tapped/will-students-soon-be-tested-grit
GRIT META-ANALYSIS
39
*Cooper, A. D. (2014). Exploring the use of non-cognitive factors in predicting college academic
outcomes. (Unpublished masters thesis). University of Tennessee, Chattanooga, Tennessee.
Costa, P.T., & McCrae, R.R. (1992). The NEO PI-R professional manual. Odessa, FL: Psychological
Assessment Resources.
Crant, J.M. (1995). The proactive personality scale and objective job performance among real
estate agents. Journal of Applied Psychology 80, 532−537.
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.
Journal of Organizational Behavior, 36, 845-872. doi:10.1002/job.2008
Credé, M., Harms, P.D., Niehorster, S., & Gaye-Valentine, A. (2012). An evaluation of the
consequences of using short measures of the Big Five personality traits. Journal of Personality
and Social Psychology, 102, 874-888. doi:10.1037/a0027403
Credé, M., & Kuncel, N.R. (2008). Study habits, study skills, and study attitudes: The third pillar
supporting collegiate academic performance. Perspectives on Psychological Science, 3, 415-453.
doi: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/s10648-011-91845
Credé, M., & Phillips, L.A. (2011). A meta-analytic review of the Motivated Strategies for Learning
Questionnaire. Learning and Individual Differences, 21, 337-346.
doi:10.1016/j.lindif.2011.03.002
GRIT META-ANALYSIS
40
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ:
Lawrence Earlbaum Associates.
Costa, P. T., & McCrae, R. R. (1992). Revised NEO Personality Inventory (NEO-PI-R) and NEO FiveFactor Inventory (NEO-FFI) Professional Manual. Odessa, FL: Psychological Assessment
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, 322331. doi:10.1037//0022-3514.81.2.322
*Cross, T. M. (2013). Staying the course: Grit, academic success and non-traditional doctoral students.
(Unpublished doctoral dissertation). Pepperdine University, Malibu, California.
*Datu, J. A. D., Valdez, J. P. M., & King, R. B. (2015). Perseverance counts but consistency does not!
Validating the Short Grit Scale in a collectivist setting. Current Psychology. doi:10.1007/s12144015-9374-2
*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 lowincome urban adolescents. (Unpublished doctoral dissertation). State University of New Jersey,
New Brunswick, New Jersey.
GRIT META-ANALYSIS
41
De Fruyt, F.; Van De Wiele, L. & Van Heeringen, C. (2000). Cloninger's psychobiological model of
temperament and character and the five-factor model of personality. Personality and Individual
Differences 29, 441–452.
*DiMenichi, B. C. & Richmond, L. L. (2015). Reflecting on past failures leads to increased
perseverance and sustained attention. Journal of Cognitive Psychology, 27(2), 180-193.
doi:10.1080/20445911.2014.995104
Duckworth, A.L. (2013a). The key to success? Grit. Retrieved from
https://www.ted.com/talks/angela_lee_duckworth_the_key_to_success_grit?language=en#t-9644
Duckworth, A.L. (2013b). Research Psychologist Angela Duckworth, 2013 MacArthur Fellow |
MacArthur Foundation. Retrieved from: https://www.youtube.com/watch?v=_hAwj1eGvDw
*Duckworth, A. L., & Quinn, P. D. (2009). Development and validation of the short grit scale (Grit-S).
Journal of Personality Assessment, 91(2), 166-174. doi:10.1080/00223890802634290
*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
Psychology and Personality Science, 2(2), 174-181. doi:10.1177/1948550610385872
*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:10.1037/0022-3514.92.6.1087
Duckworth, A.L., Quinn, P.D., Lynam, D.R., Loeber, R., & Stouthamer-Loeber, M. (2011). Role of test
motivation in intelligence testing. PNAS, 198, 7716-7720. doi:10.1073/pnas.1018601108
Durlak, J. A., Weissberg, R. P., & Pachan, M. (2010). A meta-analysis of After-School Programs that
seek to promote personal and social skills in children and adolescents. American Journal of
Community Psychology, 45, 294-309. doi:10.1007/s10464-010-9300-6
GRIT META-ANALYSIS
42
Earman, J., & Glymour, C. (1978). Lost in the tensors: Einstein’s struggles with covariance principles
1912-1916. Studies in History and Philosophy of Science, 9, 251-278. doi:10.1016/00393681(78)90008-0
Eggers, M., Smith, G.D., Schneider, M., & Minder, C. (1997). Bias in meta-analysis detected by a
simple, graphical test. British Medical Journal, 315, 629-634. doi:10.1136/bmj.315.7109.629
*Engel, L. I. (2013). What predicts first semester college performance? Cognitive ability, SAT,
conscientiousness, and grit. (Unpublished doctoral dissertation). Hofstra University, Hempstead,
New York.
Ericsson, K. A., Krampe, R. T., & Tesch-Römer, C. (1993). The role of deliberate practice in the
acquisition of expert performance. Psychological Review, 100, 363-406. doi:10.1037/0033295X.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
play in the self-regulatory processes involved in medical students learning gross anatomy.
(Unpublished doctoral dissertation). Indiana University, Bloomington, Indiana.
GRIT META-ANALYSIS
43
Foldes, H.J., & Duehr, E.E., & Ones, D.S. (2008). Group differences in personality: Meta-analysis
comparing five U.S. racial groups. Personnel Psychology, 61, 579-616. doi:10.1111/j.17446570.2008.00123.x
Frey, M.C., & Dotterman, D.K. (2004). Scholastic assessment or g? The relationship between the
Scholastic Assessment Test and General Cognitive 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 lowerlevel facets of several five-factor models. In I. Mervielde, I. Deary, F. De Fruyt, & F. Ostendorf
(Eds.), Personality Psychology in Europe, Vol. 7 (pp. 7-28). Tilburg, The Netherlands: Tilburg
University Press.
Gough, H.G. (1996). CPI manual (3rd 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 masculine 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
acquisition, accuracy, and calibration. Judgment and Decision Making, 8(3), 188-201.
Hattie, J., Biggs, J., & Purdie, N. (1996). Effects of learning skills interventions on student learning: A
meta-analysis. Review of Educational Research, 66, 99–136. doi:10.3102/00346543066002099
Hembree, R. (1988). Correlates, causes, effects, and treatment of test anxiety. Review of Educational
Research, 58, 47-77. doi:10.3102/00346543058001047
GRIT META-ANALYSIS
44
Hezlett, S. A., Kuncel, N. R., Vey, M. A., Ahart, A., Ones, D. S., Campbell, J. P., & Camara, W. (2001).
The predictive validity of the SAT: A comprehensive meta-analysis. In D. S. Ones & S. A.
Hezlett (Eds.), Predicting performance: The interface of I/O psychology and educational
research. San Diego: Symposia presented at the annual conference of the Society 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
Happiness Studies. doi:10.1007/s10902-014-9593-5
*Hogan, M. L. (2013). Non-cognitive traits that impact female success in BigLaw. (Unpublished
doctoral dissertation). University of Pennsylvania, Philadelphia, Pennsylvania.
doi:10.1037/a0019056
Howe, M. J. A. (1999). Genius explained. New York: Cambridge University Press.
Hunter, J.E., & Hunter, R.F. (1984). Validity and utility of alternative predictors of job performance.
Psychological Bulletin, 96, 72-98.
Hunter, J. E., & Schmidt, F. L. (2004). Methods of meta-analysis: Correcting error and bias in research
findings (2nd ed.). Thousand Oaks, CA: Sage
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, and
emotion regulation ability. Journal of Research in Personality, 52, 29-36.
doi:10.1016/j.jrp.2014.06.005
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., & Schraeder, M. (2015). An integrative
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 person-centered approach.
Contemporary Educational Psychology, 28(1), 37-58. doi:10.1016/S0361-476X(02)00012-7
*Karlson, R. W. (2014). Perry’s scheme of intellectual and ethical development influence on
Duckworth’s Short Grit Score in interior design students. (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.
(Unpublished 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: Evidence for a mediated
moderation model. Journal of Research in Personality, 47, 539-546.
doi:10.1016/j.jrp.2013.04.007
GRIT META-ANALYSIS
46
Kline, R.B. (2011). Principles and Practices of Structural Equation Modeling (3rd Ed.). The Guildford
Press: New York.
Kohn, A. (2014, April 6). The downside of “grit”. Washington Post. Retrieved from:
http://www.alfiekohn.org/article/downside-grit/
Krampe, R. T., & Ericsson, K. A. (1996). Maintaining excellence: Deliberate practice and elite
performance in young and older pianists. Journal of Experimental Psychology: General, 125,
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
Psychology, 77, 1121-1134. doi:10.1037/0022-3514.77.6.1121
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, 148–161. doi:10.1037/0022-3514.86.1.148
*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/mil0000082
Lucas, G.M., Gratch, J., Cheng, L., & Marsella, S. (2015). When the going gets tough: Grit predicts
costly perseverance. Journal of Research in Personality, 59, 15-22.
GRIT META-ANALYSIS
47
*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
differences: Current directions (pp. 79-90). Queensland, Australia: Australian Academic Press.
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,
1608-1618. 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
consumer spending, and gambling. The Journal of Positive Psychology, 8(2), 128-134.
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, 1928. 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.
McCornack, R.L. (1956). A criticism of studies comparing weighting methods. Journal of Applied
Psychology, 40, 343-345. doi:10.1037/h0045635
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 achieving 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.
GRIT META-ANALYSIS
48
Meng, X.L., Rosenthal, R., & Rubin, D.B. (1992). Comparing correlated correlation coefficients.
Psychological Bulletin, 111, 172-175. doi:10.1037/0033-2909.111.1.172
*Meriac, J. P., Slifka, J. S, & LaBat, L. R. (2015). Work ethic and grit: An examination of empirical
redundancy. Personality and Individual Differences, 86, 401-405.
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 of narcissism.
Assessment, 19, 8-13. doi:10.1177/1073191111429390
*Morris, R. B. (2011). Can chameleons lead change? The effect of resistance to change on high selfmonitoring leaders’ strength of purpose. (Unpublished doctoral dissertation). Columbia
University, 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. http://www.mcckc.edu/services/research/files/GRIT%20Report%20S12.pdf
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.
GRIT META-ANALYSIS
49
Paunesku, D., Walton, G.M., Romero, C., Smith, E.N., Yeager, D.S., & Dweck, C.S. (2015). Mind set
interventions are a scalable treatment for academic underachievement. Psychological Science.
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
relationship between hopelessness and suicidiality. International Journal of Cognitive Therapy,
8(2), 130-142. doi:10.1521/ijct.2015.8.2.130
Perkins-Gough, D. (2013). The significance of grit: A conversation with Angela Lee Duckworth.
Educational Leadership, 71(1), 14-20.
Perera, H.N. & DiGiacomo, M. (2013). The relationship of trait emotional intelligence with academic
performance: a meta-analytic review. Learning and Individual Differences, 28, 20-33.
doi:10.1016/j.lindif.2013.08.002
*Piña-Watson, B., López, B., Ojeda, L., & Rodriguez, K. M. (2015). Cultural and cognitive predictors of
academic motivation among Mexican American adolescents: Caution against discounting the
impact of cultural processes. Journal of Multicultural Counseling and Development, 43, 109121. 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
performance. Psychological Bulletin, 135, 322-338. doi:10.1037/a0014996
Porapat, A. E. (2014). A meta-analysis of adult-rated personality and academic performance in primary
education. British Journal of Educational Psychology, 84, 239-252. doi:10.1111/bjep.12019
GRIT META-ANALYSIS
50
*Reed, J., Pritschet, B. L., & Cutton, D. M. (2012). Grit, conscientiousness, and the transtheoretical
model of change for exercise behavior. Journal of Health Psychology, 18(5), 612-619.
doi:10.1177/1359105312451866
Richardson, M., Abraham, C., & Bond, R. (2012). Psychological correlates of university students'
academic performance: a systematic review and meta-analysis. Psychological Bulletin, 138, 353387. doi:10.1037/a0026838
*Richmond, M. (2015). g and non-g influences on GPA for Hispanics and Whites: A structural equation
modeling (SEM) approach to Spearman’s law of diminishing returns (SLDOR). (Unpublished
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
investigation of the lower-order structure of conscientiousness. Journal of Research in
Personality, 38, 164-178. doi:10.1016/S0092-6566(03)00065-5
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
study skills factors predict college outcomes? A meta-analysis. Psychological Bulletin, 130, 261288. doi:10.1037/0033-2909.130.2.261
Robbins, S.B., ,Allen, J., Casillas, A., Peterson, C.H., & Le, H. (2006). Unravelling the differential
effects 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.
GRIT META-ANALYSIS
51
Roberts, B. W., Walton, K. & Viechtbauer, W. (2006). Patterns of mean-level change in personality
traits across the life course: A meta-analysis of longitudinal studies. Psychological Bulletin, 132,
1-25. doi:10.1037/0033-2909.132.1.1
*Robertson-Kraft, C. (2014). Teachers’ motivational responses to new teacher performance
management systems: An evaluation of the pilot of Aldine ISD’s INVEST system. (Unpublished
doctoral dissertation). University of Pennsylvania, Philadelphia, Pennsylvania.
*Robertson-Kraft, C., & Duckworth, A. L. (2014). True grit: Trait-level perseverance and passion for
long-term goals predicts effectiveness and retention among novice teachers. Teachers College
Record, 116, 030302.
*Rojas, J.P. (2015). The relationship among creativity, grit, academic motivation, and academic success
in college students. Theses and Dissertations--Educational, School, and Counseling Psychology.
Paper 39. http://uknowledge.uky.edu/edp_etds/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 Psychology, 89, 329-341.
doi:10.1037/0022-0663.89.2.329
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.
GRIT META-ANALYSIS
52
*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
Psychology, 70(2), 53-62. doi:10.1024/1421-0185/a000039
*Schrepfer-Tarter, A. (2013). An exploratory study of the academic optimism of principals.
(Unpublished doctoral dissertation). The Ohio State University, Columbus, Ohio.
Seipp, B. (1991). Anxiety and academic performance: A meta-analysis of findings. Anxiety Research, 4,
27-41. doi:10.1080/08917779108248762
Shechtman, N., DeBarger, A. H., Dornsife, C., Rosier, S., & Yarnall, L. (2013). Promoting grit, tenacity,
and perseverance: Critical factors for success in the 21st century. U.S. Department of Education,
Office of Education Technology. http://pgbovine.net/OET-Draft-Grit-Report-2-17-13.pdf
*Sheehan, K. (2014). Storm clouds in the mind: A comparison of hope, grit, happiness, and life
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 exploratory 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 predictors of happiness and life
satisfaction. Journal of the Indian Academy of Applied Psychology, 34, 40-45.
*Stewart, S.B. (2015). Grit and self-control as predictors of first-year student success. (Unpublished
doctoral dissertation). The University of Southern 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/s12111-0129243-0.
GRIT META-ANALYSIS
53
*Suzuki, Y., Tamesue, D., Asahi, K., & Ishikawa, Y. (2015). Grit and work engagement: A crosssectional study. PLOS One 10(9): e0137501.DOI:10.1371/journal.pone.0137501
*Third Author (2015). [Grit, Conscientiousness, Performance, & Satisfaction]. Unpublished raw data.
Vanhove, A., Herian, M., Perez, A., Harms, P.D. & Lester, P. (2015). Can resilience be developed at
work? A meta-analytic review of resilience-building program effectiveness. Journal of
Occupational and Organizational Psychology. Advance online publication.
doi:10.1111/joop.12123
*Von Culin, K. R., Tsukayama, E., & Duckworth, A. L. (2014). Unpacking grit: Motivational correlates
of perseverance and passion for long-term goals. The Journal of Positive Psychology, 9(4), 306312. doi:10.1080/17439760.2014.898320
*Warden, D., Myers, C., & Harrell, B. (2015). Academic Achievement of Tradition and Nontraditional
College Students. Poster presented at the 61st Annual Meeting of the Southeastern Psychological
Association, March 18th – 21st, 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.
*Watson, H. (2013). Predicting college students’ positive 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.
(Unpublished masters thesis). University of Maryland, College Park, Maryland.
GRIT META-ANALYSIS
54
Wilkinson, L. & Task Force on Statistical Inference (1999). Statistical methods in psychology journals:
Guidelines and explanations. American Psychologist, 54, 594-604.
*Wolters, C.A., & Hussain, M. (2015). Investigating grit and its relation with college students’ selfregulated learning and academic achievement. Metacognition and Learning, 10, 293-311.
Zimmerman, B.J. (1990). Self-regulated learning and academic achievement: An overview. Educational
Psychology, 25, 3-17. doi:10.1207/s15326985ep2501_2
Zimmerman, B.J. (1994). Dimensions of academic self-regulation: A conceptual framework for
education. In D. Schunk & B. Zimmerman (Eds.), Self-regulation of learning and performance:
Issues and educational applications (pp. 283-301). Hillsdale, NJ: Erlbaum.
*Zimmerman, E., & Brogan, L. (2015). Grit and legal education. Pace Law Review, Forthcoming;
Drexel University Thomas R. Kline School of Law Research Paper No. 2015-A-06. Available at
SSRN: http://ssrn.com/abstract=2646851
GRIT META-ANALYSIS
55
Table 1
Artifact Distributions used for Meta-Analytic Computations
Variable
Overall Grit
Perseverance
Consistency
Overall Academic Performance
Overall GPA
College GPA
Graduate School GPA
High School GPA
Intent to Persist at Current Employer
Intent to Persist in College
Cognitive Ability
Agreeableness
Conscientiousness
Emotional Stability
Extroversion
Openness to Experience
Generalized Self-Efficacy
Optimism
Gratitude
Mental Toughness
Hope
Positive Affect
Life Satisfaction
Work Satisfaction (College and Job)
Depression
Happiness
Resiliency
Self-Control
kα
46
10
11
4
4
2
2
2
4
2
2
11
17
11
11
12
2
1
3
3
4
1
5
1
5
3
3
3
Mean α
.79
.71
.74
.88
.88
.90
.90
.86
.93
.70
.87
.75
.79
.81
.83
.76
.90
.86
.81
.84
.77
.89
.87
.89
.86
.69
.88
.84
SDα
0.07
0.13
0.11
0.04
0.04
0.00
0.00
0.06
0.05
0.25
0.07
0.14
0.09
0.07
0.10
0.06
0.06
0.08
0.12
0.17
0.03
0.05
0.14
0.07
0.04
Note: kα is the number of reliability estimates in distribution, Mean α is the mean of the reliability estimates, SD α is
the standard deviation of the reliability estimates.
GRIT META-ANALYSIS
56
Table 2
Separate Meta-Analytic Estimates of Grit-Criteria Relations for Peer-Reviewed and Not-Peer-Reviewed Publications
Criterion
k
Academic Performance
GPA (All Levels)
Undergraduate GPA
High School GPA
14
14
11
8
Peer-Reviewed Publications
10%
N
robs
ρ
SDρ
CV
6,440 .15 .18 0.12
.03
6,440 .15 .18 0.12
.03
5,657 .15 .17 0.12
.02
4,381 .14 .17 0.15
-.02
90%
CV
.34
.34
.33
.36
k
25
23
19
9
Not-Peer-Reviewed Publications
10% 90%
N
robs
ρ
SDρ
CV CV
6,701 .15 .17 0.09 .07 .28
6,159 .13 .16 0.07 .07 .25
4,869 .13 .16 0.07 .07 .25
1,983 .11 .14 0.10 .01 .27
Note. k = number of studies, N =number of subjects, r obs = sample size weighted mean observed correlation, =true score correlation, SD = standard deviation of true
score correlation, 10%CV and 90%CV = lower and upper bound of 80% credibility intervals. Confidence intervals are not shown here due to space limitations but can be
obtained from first author by request.
GRIT META-ANALYSIS
57
Table 3
Meta-Analytic Estimates of the Relations between Grit and Performance Criteria
Overall Grit
Perseverance of Effort
Consistency of Interest
k
N
robs
ρ
SDρ
10%
CV
90%
CV
K
N
robs
ρ
SDρ
10%
CV
90%
CV
k
N
robs
ρ
SDρ
10%
CV
90%
CV
Academic Performance
GPA (All Levels)
Undergraduate GPA
39
37
30
13,141
12,601
10,526
.15
.14
.14
.18
.17
.17
0.11
0.10
0.10
.04
.04
.04
.31
.30
.30
11
10
8
5,221
5,065
4,595
.20
.20
.20
.26
.26
.25
0.12
0.12
0.12
.11
.11
.11
.41
.41
.40
11
10
8
5,221
5,065
4,595
.08
.08
.08
.10
.10
.09
0.02
0.03
0.03
.07
.06
.05
.13
.13
.14
High School GPA
17
6,364
.13
.16
0.14
-.02
.34
7
3,313
.22
.29
0.08
.19
.39
7
3,313
.11
.13
0.06
.05
.21
Graduate School GPA
Retention (General)
Retention (General)*
3
5
11
1,141
2,705
17,525
.07
.14
.10
.08
.16
.12
0.00
0.06
0.09
.08
.08
.00
.08
.23
.24
Retention (w/o Marriage)*
10
11,163
.16
.18
0.03
.14
.23
Non-Academic Performance
Intent to Persist in College
Intent to Persist with Current
Employer
7
5
4,116
3,967
.19
.13
.21
.18
0.00
0.00
.21
.18
.21
.18
4
2,959
.15
.22
0.00
.22
.22
4
2,959
.09
.12
0.05
.06
.19
4
519
.13
.15
0.00
.15
.15
17
22,048
.44
.60
0.21
.34
.87
Correlate/Criterion
Grit – Perseverance
Note. k = number of studies, N =number of subjects, r obs = sample size weighted mean observed correlation, =true score correlation, SD = standard deviation of true
score correlation, 10%CV and 90%CV = lower and upper bound of 80% credibility intervals. * denotes that these estimates include estimates of point-biserial correlations
computed by taking root of Nagelkerke R2 values. Confidence intervals are not shown here due to space limitations but can be obtained from first author by request.
GRIT META-ANALYSIS
58
Table 4
Meta-Analytic Estimates of the Relations between Grit and Trait and State Variables
Overall Grit
Perseverance of Effort
Consistency of Interest
Correlate
k
N
robs
ρ
SDρ
10%
CV
90%
CV
k
N
robs
ρ
SDρ
10%
CV
90%
CV
k
N
robs
ρ
SDρ
10%
CV
90%
CV
Agreeableness
13
14,395
.25
.33
0.07
.23
.42
3
3,203
.28
.39
0.03
.36
.43
3
3,203
.23
.17
0.00
.22
.23
Conscientiousness
22
18,826
.66
.84
0.07
.75
.92
8
4,967
.61
.83
0.14
.65
1.00
8
4,967
.47
.61
0.17
.39
.84
14
14,501
.33
.41
0.04
.36
.46
3
3,203
.30
.40
0.14
.22
.58
3
3,203
.29
.37
0.00
.37
.37
Emotional
Stability
Extraversion
13
14,395
.19
.23
0.09
.12
.35
3
3,203
.32
.41
0.09
.29
.53
3
3,203
.07
.09
0.05
.02
.15
Openness
14
14,585
.15
.19
0.14
.01
.37
3
3,203
.32
.45
0.07
.11
.78
3
3,203
-.02
-.03
0.05
-.09
.03
Cognitive Ability
Whole Candidate
Score (Military)
Generalized SelfEfficacy
Optimism
21
11,513
.04
.05
0.12
-.11
.20
5
2,204
-.01
-.01
0.04
-.06
.04
5
2,204
-.00
-.00
0.00
-.00
-.00
4
4,770
.02
.02
0.00
.02
.02
3
1,908
.36
.43
0.11
.28
.57
3
2,059
-.04
-.04
0.12
-.20
.11
Gratitude
Mental
Toughness
Hope
3
1,415
.24
.30
0.09
.19
.41
6
3,817
.37
.46
0.08
.35
.56
5
2,378
.33
.42
0.05
.35
.49
Positive Affect
3
670
.38
.46
0.03
.42
.49
Life Satisfaction
Work Satisfaction
(College and Job)
Depression
9
2,266
.25
.30
0.08
.20
.40
3
975
.41
.54
0.00
.54
.54
3
975
.16
.20
0.00
.20
.20
6
3,439
.10
.11
0.02
.09
.13
5
3,243
.19
.25
0.08
.15
.34
5
3,243
.04
.05
0.09
-.07
.17
5
3,865
-.40
-.48
0.10
-.61
-.35
Happiness
4
726
.22
.30
0.00
.30
.30
Resiliency
3
480
.08
.09
0.09
-.02
.21
Self-Control
4
2,615
.59
.72
0.05
.66
.78
GRIT META-ANALYSIS
59
Note. k = number of studies, N =number of subjects, r obs = sample size weighted mean observed correlation, =true score correlation, SD = standard deviation of true
score correlation, 10%CV and 90%CV = lower and upper bound of 80% credibility intervals. Confidence intervals are not shown here due to space limitations but can be
obtained from first author by request.
GRIT META-ANALYSIS
60
Table 5
Meta-Analytic Estimates of the Relations between Grit and Demographic Variables
Overall Grit
Perseverance of Effort
Consistency of Interest
Correlate
k
N
robs
ρ
SDρ
10%
CV
90%
CV
k
N
robs
ρ
SDρ
10%
CV
90%
CV
k
N
robs
ρ
SDρ
10%
CV
90%
CV
Gender
(M=0, F=1)
25
18,750
.05
.05
0.07
-.03
.14
8
5,040
.02
.03
0.05
-.04
.09
8
5,040
.03
.03
0.05
-.03
.09
Age
Year in
School
Ethnic
Minority
Status
African
American
Asian
22
12,349
.11
.12
0.04
.07
.18
4
2,961
.05
.05
0.05
-.02
.12
4
2,961
.04
.06
0.05
-.01
.11
4
2,961
.03
.04
0.02
.01
.07
9
15,261
.01
.01
0.01
.00
.02
5
3,301
-.02
-.02
0.00
-.02
-.02
5
3,301
.02
.02
0.01
.01
.04
7
14,403
.03
.03
0.04
-.02
.08
4
2,786
-.02
-.02
0.00
-.02
-.02
4
2,786
.00
-.00
0.05
-.07
.06
4
11,959
.01
.02
0.00
.02
.02
Hispanic
5
12,302
-.02
-.02
0.05
-.09
.05
Note. k = number of studies, N =number of subjects, r obs = sample size weighted mean observed correlation, =true score correlation, SD = standard deviation of true
score correlation, 10%CV and 90%CV = lower and upper bound of 80% credibility intervals. Confidence intervals are not shown here due to space limitations but can be
obtained from first author by request.
GRIT META-ANALYSIS
61
Table 6
Incremental Validity Estimates based on Hierarchical Regression Analysis using Meta-Analytic
Intercorrelations
Criteria
Predictor
Model 1
Model 2
Model 3
Model 4
Model 5
Overall Academic
Performance
HSGPA
College GPA
β
ΔR
β
ΔR
β
ΔR
Step 1
Conscientiousness
.23
.220
(.040)
.26
.210
(.052)
.33
.230
(.064)
Step 2
Overall Grit
-.02
.000
(.180)
-.06
.002
(.160)
-.11
.004
(.170)
Step 1
Conscientiousness
.01
.220
(.000)
-.10
.210
(.005)
.07
.230
(.003)
Step 2
Perseverance
.25
.040
(.260)
.37
.085
(.290)
.19
.023
(.250)
Step 1
Conscientiousness
.25
.220
(.114)
.21
.210
(.080)
.28
.230
(.149)
Step 2
Consistency
-.05
. 004
(.100)
.00
.000
(.130)
-.08
.009
(.090)
Step 1
Conscientiousness
.05
-.08
.28
.295
(.168)
.11
Perseverance
.260
(.171)
.253
(.178)
Step 2
Consistency
-.10
.011
(.100)
.39
.003
(.130)
-.11
.015
(.090)
Step 1
Conscientiousness
.05
-.08
-.10
.210
(.008)
.11
Consistency
.224
(.011)
.239
(.016)
Perseverance
.28
.047
(.260)
.39
Step 2
-.06
-.06
.088
(.290)
.22
-.11
.22
.027
(.250)
Note: ΔR values in table are all adjusted values based on the harmonic mean of sample sizes. Standardized Beta are those
observed at step 2 for all variables. Incremental R values in parentheses are those observed if the order of variables for the
hierarchical regression is reversed. Harmonic means ranged from 5,540 to 18,774.
GRIT META-ANALYSIS
62
Appendix
Summary of Studies that Reported Relations between Grit and Performance, Retention, Big Five Personality Traits, and Cognitive Ability.
Authors
Allen (2014)
Arouty (2015) - Sample 1
Arouty (2015) - Sample 1
Arouty (2015) - Sample 2
Black (2014)
Bowman et al. (2015) - Sample 1
Bowman et al. (2015) - Sample 1
Bowman et al. (2015) - Sample 1
Bowman et al. (2015) - Sample 2
Bowman et al. (2015) - Sample 2
Bowman et al. (2015) - Sample 2
Bowman et al. (2015) - Sample 2
Bowman et al. (2015) - Sample 2
Bowman et al. (2015) - Sample 2
Bowman et al. (2015) - Sample 3
Bowman et al. (2015) - Sample 3
Bowman et al. (2015) - Sample 3
Bowman et al. (2015) - Sample 3
Bowman et al. (2015) - Sample 3
Bowman et al. (2015) - Sample 3
Buller (2012)
Buller (2012)
Buller (2012)
Chambers et al. (2012)
Chang (2014)
Chang (2014)
Chang (2014)
Cooper (2014)
Cooper (2014)
Cooper (2014)
Cooper (2014)
Cooper (2014)
Cooper (2014)
N
30
124
124
110
97
417
417
417
938
938
938
938
938
938
1,089
1,089
1,089
1,089
1,089
1,089
968
968
1,123
106
342
342
342
515
515
515
515
515
515
Grit Type
Overall
Overall
Overall
Overall
Overall
Consistency
Overall
Perseverance
Consistency
Consistency
Overall
Overall
Perseverance
Perseverance
Consistency
Consistency
Overall
Overall
Perseverance
Perseverance
Overall
Overall
Overall
Overall
Consistency
Overall
Perseverance
Consistency
Consistency
Overall
Overall
Perseverance
Perseverance
Grit α
.75
.75
.82
.77
.79
.79
.81
.81
.79
.79
r
-.02
-.06
-.08
.02
.22
.14
.31
.26
.13
.10
.26
.23
.30
.29
.11
.09
.20
.21
.22
.25
.10
.22
Performance
Type
College GPA
College GPA
HSGPA
HSGPA
College GPA
College GPA
College GPA
College GPA
HSGPA
College GPA
HSGPA
College GPA
HSGPA
College GPA
HSGPA
College GPA
HSGPA
College GPA
HSGPA
College GPA
College GPA
Non-Academic
.18
-.08
.09
.21
.14
.07
.18
.16
.16
.19
College GPA
College GPA
College GPA
College GPA
HSGPA
College GPA
HSGPA
College GPA
HSGPA
College GPA
Retention
r
Type
.07 General
C
r
A
α
r
E
α
r
ES
α
r
O
α
r
α
g
r
-.15
.25
General
.08
General
-.07
.74
.50
.38
.79
.10
.82
.04
.89
.24
.84
.07
.77
.08
.03
-.07
-.03
.49
.79
.20
.82
.17
.89
.28
.84
.24
.77
.00
.42
.79
.23
.82
.25
.89
.21
.84
.34
.77
.04
GRIT META-ANALYSIS
63
Appendix (cont.)
Performance
Authors
Cross (2013)
Davidson (2014)
Davidson (2014)
DeCandia (2014)
Duckworth & Quinn (2009) - Study 2
Duckworth & Quinn (2009) - Study 2
Duckworth & Quinn (2009) - Study 2
Duckworth & Quinn (2009) - Study 4
Duckworth & Quinn (2009) - Study 6
Duckworth et al. (2011)
Duckworth et al (2007) - Study 2
Duckworth et al (2007) - Study 3
Duckworth et al (2007) - Study 4
Duckworth et al (2007) - Study 4
Duckworth et al (2007) - Study 4
N
669
116
116
413
1,554
1,554
1,554
279
190
190
706
139
1,218
1,218
1,218
Grit Type
Overall
Overall
Overall
Overall
Consistency
Overall
Perseverance
Overall
Overall
Overall
Overall
Overall
Overall
Overall
Overall
Grit
α
r
.09
-.03
.05
.18
Type
Retention
r
Type
.79
.79
.79
Duckworth et al (2007) - Study 5
Duckworth et al (2007) - Study 5
Duckworth et al (2007) - Study 6
Engel (2013)
Eskreis-Winkler et al. (2014) - Study 1
1,308
1,308
79
88
677
Overall
Overall
Overall
Overall
Overall
.79
.79
.80
.82
.77
Eskreis-Winkler et al. (2014) - Study 2
Eskreis-Winkler et al. (2014) - Study 3
Eskreis-Winkler et al. (2014) - Study 4
Fillmore (2015)
Hill et al. (2014)
Hogan (2013)
Hogan (2013)
Hogan (2013)
Hogan (2013)
Ivcevic & Brackett (2014)
Ivcevic & Brackett (2014)
Joseph (2009)
Karlson (2014)
442
4,813
6,362
384
337
405
425
423
477
213
213
57
15
Overall
Overall
Overall
Overall
Overall
Overall
Overall
Overall
Overall
Overall
Overall
Overall
Overall
.79
.90
.79
.32
HSGPA
.17
Non-Academic
.25
.06
.19
College GPA
College GPA
Non-Academic
.02
E
ES
O
g
r
α
r
α
r
α
r
α
r
α
.64
.73
.74
.84
.84
.84
.18
.23
.25
.80
.80
.80
.12
.19
.26
.88
.88
.88
.32
.37
.42
.87
.87
.87
-.02
.07
.14
.82
.82
.82
.70
.86
.44
.84
.12
.84
.28
.79
.77
.86
.24
.82
.22
.89
.38
.85
.18
.17
.14
.68
.68
.84
r
-.20
-.05
0.20a
General
0.12a
General
College GPA
.35
Grade in Class
-.01
.07
.04
.24
.14
.18
.01
HSGPA
College GPA
Graduate GPA
Non-Academic
HSGPA
Non-Academic
Non-Academic
.71
.72
.72
.80
A
Graduate GPA
HSGPA
College GPA
HSGPA
.77
.70
.86
.82
.82
.85
C
0.14a
General
0.18a
.18
.00a
General
General
Marriage
.64
.82
.77
.80
.64
-.08
.02
-.09
-.07
.39
.25
.48
.19
.15
.71
.60
.20
.81
.27
.21
.80
.16
.33
.86
.37
.08
.83
.05
.77
-.27
GRIT META-ANALYSIS
64
Appendix (cont.)
Performance
Authors
Kelly et al. (2014)
Kelly et al. (2014)
Kelly et al. (2014)
Kelly et al. (2014)
Kelly et al. (2014)
Khaler (2014)
MacCann & Roberts (2010)
MacCann & Roberts (2010)
MacCann & Roberts (2010)
McCutcheon (2014)
Meriac et al. (2015)
Meriac et al. (2015)
Meriac et al. (2015)
Metropolitan Community College (2012)
Morris (2011) - Sample 1
Morris (2011) - Sample 2
Reed et al. (2012)
Richmond (2015) - Sample 1
Richmond (2015) - Sample 2
Robertson-Kraft (2014)
Robertson-Kraft & Duckworth (2014) - Study 1
Robertson-Kraft & Duckworth (2014) - Study 2
Robertson-Kraft & Duckworth (2014) - Study 2
Rojas (2015) - Sample 1
Rojas (2015) - Sample2
Sheehan (2014)
Sheehan (2014)
Sheehan (2014)
Stewart (2015)
Stewart (2015)
Stewart (2015)
Stewart (2015)
Stewart (2015)
Stewart (2015)
N
993
993
993
1,308
993
21
291
291
291
134
322
322
322
470
74
76
1,165
144
199
2,662
154
307
127
187
817
179
179
179
88
88
88
88
88
88
Grit Type
Consistency
Overall
Overall
Overall
Perseverance
Overall
Consistency
Overall
Perseverance
Overall
Consistency
Overall
Perseverance
Overall
Overall
Overall
Overall
Overall
Overall
Overall
Overall
Overall
Overall
Overall
Overall
Consistency
Overall
Perseverance
Overall
Overall
Consistency
Consistency
Perseverance
Perseverance
Grit
α
r
Type
.06
.05
.14
College GPA
College GPA
Non-Academic
.02
-.15
.12
.14
.15
College GPA
College GPA
HSGPA
HSGPA
HSGPA
Retention
r
.80
.20
.77
.77
.79
.75
.64
.68
.82
.85
.85
.84
.84
.73
.73
.70
α
r
E
α
r
ES
α
r
O
α
r
g
α
r
General
-.04
.01
.75
.78
.65
r
A
-.02
-.03
.20
.81
Type
C
.46
.58
.70
.84
.69
.77
.63
.79
.79
.79
.48
.40
.72
.72
.63
.58
.85
.85
.78
.72
.80
.59
.00
.05
.10
.84
General
College GPA
.23
-.06
College GPA
College GPA
.02
-.09
College GPA
College GPA
.11
.22
.02
.25
.39
.19
.22
.17
.26
.11
.08
College GPA
College GPA
HSGPA
HSGPA
HSGPA
College GPA
HSGPA
College GPA
HSGPA
College GPA
HSGPA
.15
.12
.80
.80
.31
.19
.88
.88
.31
.20
.81
.81
.34
-.06
.78
.78
.13
.16
.30
.81
.75
.59
.08
.10
.06
.89
.87
.71
.34
.25
.29
.79
.81
.60
-.03
.01
.24
.76
.73
.65
-.22
-.11
.16
.06
.05
.09
-.02
GRIT META-ANALYSIS
65
Appendix (cont.)
Performance
Authors
Strayhorn (2013)
Strayhorn (2013)
Suzuki et al (2015)
Suzuki et al (2015)
Suzuki et al (2015)
Third Author (2015) - Sample 1
Third Author (2015) - Sample 1
Third Author (2015) - Sample 1
Third Author (2015) - Sample 2
Third Author (2015) - Sample 2
Third Author (2015) - Sample 2
Warden et al. (2015) - Sample 1
Warden et al. (2015) - Sample 2
Weston (2015)
Weston (2015)
Weston (2015)
Wolters (2015)
Wolters (2015)
Wolters (2015)
Wolters (2015)
Wolters (2015)
Wolters (2015)
Zimmerman & Brogan (2015)
Wolters (2015)
N
140
140
1,134
1,134
1,134
156
156
156
971
971
971
67
72
33
33
33
213
213
213
213
213
213
49
49
Grit Type
Overall
Overall
Overall
Consistency
Perseverance
Consistency
Overall
Perseverance
Consistency
Overall
Perseverance
Overall
Overall
Overall
Consistency
Perseverance
Overall
Overall
Consistency
Consistency
Perseverance
Perseverance
Overall
Overall
Grit
α
.87
.87
.87
.76
.89
.55
.74
.60
.86
.85
.81
r
.35
.38
Type
Retention
r
Type
C
r
A
α
r
E
α
r
ES
α
r
O
α
r
g
α
HSGPA
College GPA
.08
.10
.09
Grade in Class
Grade in Class
Grade in Class
.31
.33
College GPA
College GPA
r
.23
.56
.35
.47
.22
.42
.52
.35
.60
.68
.52
.52
.52
.80
.80
.80
.86
.86
.86
.39
.20
.35
.37
.37
.37
.37
.01
.43
.56
.56
.56
.28
.26
.18
.86
.86
.86
.45
-.07
.57
.85
.85
.85
.65
.52
.51
.72
.72
.74
.74
.64
.64
.26
-.02
0.1
-.12
.29
.09
.17
.00
College GPA
HSGPA
College GPA
HSGPA
College GPA
HSGPA
College GPA
Graduate GPA
-.27
Note: N= sample size, C=conscientiousness, A = agreeableness, E = extraversion, ES = emotional stability, O = openness, g = cognitive ability. No local reliability estimates for
cognitive ability were reported. a denotes correlations estimate from Nagelkerke R2 values. This appendix only includes studies reporting on the relations between grit and grit
facets on one hand and performance, retention, Big Five personality, and cognitive ability on the other hand. Coded results from all other studies are available from the first author
on request.