Personality Predictors of Academic Outcomes: Big Five Correlates of GPA and SAT Scores

PERSONALITY PROCESSES AND INDIVIDUAL DIFFERENCES
Personality Predictors of Academic Outcomes: Big Five Correlates of GPA

and SAT Scores
Erik E. Noftle and Richard W. Robins
University of California, Davis
The authors examined relations between the Big Five personality traits and academic outcomes,
specifically SAT scores and grade-point average (GPA). Openness was the strongest predictor of SAT
verbal scores, and Conscientiousness was the strongest predictor of both high school and college GPA.
These relations replicated across 4 independent samples and across 4 different personality inventories.
Further analyses showed that Conscientiousness predicted college GPA, even after controlling for high
school GPA and SAT scores, and that the relation between Conscientiousness and college GPA was
mediated, both concurrently and longitudinally, by increased academic effort and higher levels of
perceived academic ability. The relation between Openness and SAT verbal scores was independent of
academic achievement and was mediated, both concurrently and longitudinally, by perceived verbal
intelligence. Together, these findings show that personality traits have independent and incremental
effects on academic outcomes, even after controlling for traditional predictors of those outcomes.
Keywords: Conscientiousness, Openness to Experience, achievement, SAT scores, GPA
Personality has important influences on success in school (De amines the ability of the Big Five personality traits to predict
Raad & Schouwenburg, 1996; Digman & Takemoto-Chock, academic outcomes, specifically SAT scores and grade-point
1981) and work (Barrick & Mount, 1991; Ozer & Benet- average (GPA).
Martı́nez, 2006; Roberts & Hogan, 2001). It is important to note
that the predictive power of personality has little to do with Previous Research on the Personality Correlates of SAT
intelligence or other aspects of cognitive ability. Early trait Scores
researchers made a clear distinction between intelligence and
personality traits (Allport & Odbert, 1936). This distinction In contrast to the abundance of research on personality and
persists to this day and is reflected in the exclusion of explicit grades, there has been virtually no research on the personality
intelligence content from most contemporary personality inven- correlates of SAT scores and other standardized measures of
tories (McCrae & Costa, 1985, 1997). Recent studies have academic aptitude and achievement. The SAT is by far the most
shown that personality predicts academic performance (e.g., widely used measure of academic potential, and it plays a
Duckworth & Seligman, 2005; Wagerman & Funder, 2007) and central role in admissions decisions at most universities in the
occupational success (Hogan, 2005), even when intelligence United States. SAT scores have been interpreted in a number of
and cognitive ability are controlled. The current research ex- different ways, both by the test’s designers themselves (Edu-
cational Testing Service) and by college administrators, high
school counselors, the popular press, and researchers in fields
such as education and psychology. Indeed, even the name of the
Erik E. Noftle and Richard W. Robins, Department of Psychology, test has been repeatedly changed and reinterpreted over the
University of California, Davis. years. It was introduced in 1901 as the Scholastic Achievement
This project was supported by National Institute of Mental Health Grant Test, purporting to measure the level of achievement attained
MH-20006 to Erik E. Noftle and National Institute of Aging Grant by prospective college students. After considerable develop-
AG022057-01 to Richard W. Robins. The authors acknowledge and thank ment (and growing popularity), it was renamed the Scholastic
Michael Ashton and Kibeom Lee for their helpful comments on earlier
Aptitude Test in 1941 to emphasize the fact that it measures the
versions of the manuscript. A portion of these data was presented at the
annual meeting of the Association for Research in Personality, Memphis,
ability to succeed in college. After the rise of “coaching
TN, January 2007. courses,” which demonstrated that students could successfully
Correspondence concerning this article should be addressed to Erik E. increase their test scores, the test was renamed the Scholastic
Noftle, Department of Psychology, University of California, One Shields Assessment Test in 1991. Finally, in 1994, the test was reduced
Avenue, Davis, CA 95616-8686. E-mail: eenoftle@ucdavis.edu to its initials: “Please note that SAT is not an initialism. It does
Journal of Personality and Social Psychology, 2007, Vol. 93, No. 1, 116 –130
Copyright 2007 by the American Psychological Association 0022-3514/07/$12.00 DOI: 10.1037/0022-3514.93.1.116
116
PERSONALITY AND ACADEMIC OUTCOMES 117
not stand for anything” (College Board, 1994, as cited in as Block’s construct of ego undercontrol, and SAT scores
Harper, 2002). As of 2005, the current version of the SAT was (Letzring, Block, & Funder, 2005).
labeled the SAT Reasoning Test, which, according to the Ed- We believe that the inconsistent results may be due, at least
ucational Testing Service, assesses “reasoning ability” and not in part, to previous researchers’ failure to separately examine
intelligence. the correlates of SAT verbal and SAT math scores. Previous
Despite the test maker’s claim that the SAT is not an intel- research suggests that verbal and quantitative abilities have
ligence test, recent research suggests that the SAT measures different personality correlates (Schuerger, Kepner, & Lawler,
something very close to general mental ability. For example, 1979). One possible reason for their divergent correlates is that
Frey and Detterman (2004) found that the SAT correlated .82 the SAT verbal test may be more strongly related to crystallized
with a measure of “g” (or general intelligence) extracted from intelligence because of its vocabulary-related content, whereas
the Armed Services Vocational Aptitude Battery in a large the SAT math test may be more strongly related to fluid
sample from the National Longitudinal Survey of Youth (see intelligence because of its numerical and spatial reasoning-
also Brodnick & Ree, 1995). related content (e.g., Rohde & Thompson, 2007). In a meta-
Given the link between SAT scores and intelligence, research analysis of the overlap between intelligence and personality,
on the personality correlates of intelligence can provide one Ackerman and Heggestad (1997) found that crystallized intel-
window into the possible relation between personality and SAT ligence was related to three Openness-related personality con-
scores. The overlap between personality and intelligence is the structs, whereas fluid intelligence was not consistently related
subject of some controversy. Some researchers have argued that to any personality constructs. Similarly, Ashton, Lee, Vernon,
certain personality dimensions, particularly Openness to Expe- and Jang (2000) found moderate to strong relations between
rience, overlap substantially with intelligence (Ackerman & Openness and aspects of crystallized intelligence and only weak
Heggestad, 1997; Eysenck, 1991; see also Collis & Messick, relations between Openness and aspects of fluid intelligence.
2001). In contrast, other researchers have maintained that per- Thus, in the present research, we examined SAT verbal and
sonality and intelligence are conceptually and empirically dis- math scores separately and expected to find that Openness
tinct (Demetriou, Kyriakides, & Avraamidou, 2003; McCrae & predicts SAT verbal scores, whereas we made no specific
Costa, 1997). Taking somewhat of a middle ground in this predictions about SAT math scores.
debate are Goldberg and Saucier (Goldberg, 1990; Saucier,
1992, 1994; Saucier & Goldberg, 1996), who, drawing from
Previous Research on Personality Correlates of Academic
their lexical research on the Big Five trait domains, conceptu-
Performance
alized the Openness domain as “Intellect,” emphasizing its
connection to creativity, abstract thinking, depth of thought, In contrast to the paucity of personality research on standard-
and other intellective qualities. In general, the research litera- ized tests such as the SAT, research linking personality traits to
ture suggests that measures of intelligence and other aspects of academic achievement has a long history in psychology. Early
cognitive ability are modestly but consistently related to Open- studies by Harrison Gough and his collaborators showed that
ness but are not consistently related to the other four Big Five California Psychological Inventory scales related to Conscien-
domains (Moutafi, Furnham, & Crump, 2003, 2006). A recent tiousness predicted higher levels of achievement in both high
study by Bischel and Baker (2006) suggested that the relation school and college (Gough, 1964; Gough & Hall; 1964; Gough
between Openness and intelligence is strong in young adult- & Lanning, 1986). Similarly, Hogan and Weiss (1974) found
hood but weak later in adulthood, which may explain some that college students elected to Phi Beta Kappa (an academic
inconsistencies in previous findings. honor for high achievers) tended to score higher on the Cali-
It is surprising that we could identify only two published fornia Psychological Inventory scales of responsibility, self-
studies that correlated an established measure of the Big Five control, and socialization than did students who were not
dimensions with SAT scores. Wolfe and Johnson (1995) found elected to Phi Beta Kappa. The link between personality and
that low Agreeableness (assessed via the Big Five Inventory; achievement has also been demonstrated with non-self-report
John & Srivastava, 1999) was the only significant predictor of measures of personality. For example, John, Caspi, Robins,
total SAT scores. Conard (2006) found that Openness (assessed Moffitt, and Stouthamer-Loeber (1994) found that mother re-
via the NEO Five Factor Inventory; Costa & McCrae, 1992) ports of Conscientiousness and Openness in an ethnically di-
was the only significant predictor of total SAT scores. How- verse sample of middle school boys predicted teacher ratings of
ever, past research with non-Big Five measures has suggested school performance in reading, writing, spelling, and math.
that SAT scores are related to Openness-related traits, such as These cross-method correlations replicate and extend other
need for cognition, but also Conscientiousness-related traits, studies showing a link between teacher ratings of Conscien-
such as achievement and (work-oriented) resiliency (Tross, tiousness and teacher ratings of school performance (Digman,
Harper, Osher, & Kneidinger, 2000). Mischel, Shoda, and 1989; Graziano & Ward, 1992), as well as between self-
Peake (1988; Shoda, Mischel, and Peake, 1990) found a relation reported Openness and school grades (Lounsbury, Sundstrom,
between delay of gratification at age 4 (which reflects the Loveland, & Gibson, 2003). Finally, a few studies have found
self-control aspect of Conscientiousness) and higher SAT small to medium associations between grades and other person-
scores in late adolescence. However, other studies have found ality traits, including Agreeableness (E. C. Hair & Graziano,
only weak relations between aspects of Conscientiousness, such 2003) and Emotional Stability (Ridgell & Lounsbury, 2004).
118 NOFTLE AND ROBINS
Table 1
Previous Findings on Personality and Academic Outcomes in College
Criterion N Measure E A C N O
Barchard (2003) GPA 150 IPIP NEO-PI 0 0 ⫹⫹⫹ 0 ⫹

Busato et al. (2000)a GPA 409 5PFT 0 0 ⫹ 0 0
Conard (2006) GPA 289 NEO-FFI 0 0 ⫹⫹⫹ 0 0
de Fruyt & Mervielde (1996) GPA 714 NEO-PI-R 0 0 ⫹⫹ 0 0
Duff et al. (2004) GPA 146 16PFi 0 0 0 0 0
Farsides & Woodfield (2003)a GPA 432 NEO-FFI 0 0 0 0 ⫹⫹
Furnham et al. (2003)a GPA 93 NEO-PI-R ⫺⫺ 0 ⫹⫹⫹ 0 0
Gray & Watson (2002) GPA 300 NEO-FFIb 0 ⫹ ⫹⫹⫹ 0 ⫹
Langford (2003) GPA 203 BFM 0 0 ⫹⫹⫹ 0 0
Oswald et al. (2004) GPA 636 IPIP BFM 0 0 0 0 0
Ridgell & Lounsbury (2004) GPA 140 PSI 0 0 0 0 0
Wolfe & Johnson (1995) GPA 201 BFI 0 0 ⫹⫹⫹ 0 0
Conard (2006) Course grade 186 NEO-FFI 0 ⫹ ⫹⫹⫹ 0 0
Chamorro-Premuzic & Furnham (2003a)a Course grade 70 NEO-FFI 0 0 ⫹⫹⫹ ⫺⫺⫺ 0
Chamorro-Premuzic & Furnham (2003b) Course grade 247 NEO-PI-R 0 0 ⫹⫹⫹ ⫺ 0
Furnham & Chamorro-Premuzic (2004) Course grade 91 NEO-FFI ⫺⫺ 0 ⫹⫹ 0 0
P. Hair & Hampson (2006) Course grade 236 BFI ⫺ 0 ⫹ 0 0
Lounsbury et al. (2003) Course grade 175 PSI 0 0 ⫹ 0 ⫹
Lounsbury et al. (2005)a Course grade 434 APSI ⫹ ⫹ ⫹ ⫺⫺ ⫹
Ridgell & Lounsbury (2004) Course grade 140 PSI 0 0 0 ⫺ 0
Mean effect size 5,292 ⫺.04 .09 .26 ⫺.07 .05
Note. E ⫽ Extraversion; A ⫽ Agreeableness; C ⫽ Conscientiousness; N ⫽ Neuroticism; O ⫽ Openness; IPIP NEO-PI ⫽ abbreviated version of the IPIP
version of the Revised NEO Personality Inventory (NEO-PI-R; Costa & McCrae, 1992; Goldberg, 1999; see also Goldberg, et al., 2006); 5PFT ⫽ Vijf
Persoonlijkheids-Factoren Test (Elshout and Akkerman, 1975); NEO-FFI ⫽ NEO Five Factor Inventory (Costa & McCrae, 1992); 16PFi ⫽ 16PFi Form
A with second order factors for the Big Five (Cattell, 2000); BFM ⫽ Shafer’s (1999) Bipolar Big Five Markers; IPIP BFM ⫽ IPIP version of the 50-item
Big Five Marker measure; PSI ⫽ Personal Style Inventory (Lounsbury & Gibson, 2004); BFI ⫽ Big Five Inventory (John & Srivastava, 1999); APSI ⫽
Adolescent Personal Style Inventory (Lounsbury & Gibson, 2004). 0 refers to a nonsignificant correlation; ⫺ refers to a correlation between ⫺.10 and ⫺.19;
⫺⫺ refers to a correlation between ⫺.20 and ⫺.29; ⫺⫺⫺ refers to a correlation between ⫺.30 and ⫺1.00; similarly, the ⫹ signs refer to the parallel ranges
of positive correlation coefficients; for all correlations marked with ⫺ or ⫹ signs, p ⬍ .05.
a
A longitudinal study.
b
NEO-FFI for all domains except Conscientiousness, which is measured with the full NEO-PI-R.
To help readers better understand the overall pattern of relations

between the Big Five dimensions and academic achievement, we
present in Table 1 a summary of previous studies on the Big Five
correlates of academic performance in college.1 Conscientiousness
emerges as the most robust predictor of college grades (mean r ⫽
1
.26). The other four Big Five factors were not consistently related We used three methods to locate relevant studies. First, we reviewed
to academic performance (mean rs ⫽ ⫺.04, .09, ⫺.07, and .05 for reference lists from previously published articles and chapters on the
Extraversion, Agreeableness, Neuroticism, and Openness, respec- relation between personality and academic achievement. Second, we
searched the PsycINFO database (http://www.apa.org/psycinfo/) for arti-
tively), although Openness had significant positive effects in one
cles published between 1887 (the earliest entry in the PsycINFO database)
fourth of the studies. and June 2006, using the keywords “Big Five” and “Five Factor Model”
Although Conscientiousness was the best predictor of academic paired with each of the following keywords: GPA, grade*, academic
success in college, the magnitude of the effect varied substantially performance, and academic success. Third, we searched for relevant arti-
across studies. This pattern may reflect, at least in part, the differential cles by reviewing the reference lists of the articles identified in the
predictive validity of different facets of Conscientiousness (Ashton, PsychINFO searches that met the inclusion criteria described in the next
1998; Mershon & Gorsuch, 1988; Paunonen, 1998). For example, sentences. We included studies if they fulfilled five criteria. First, the study
Paunonen and Ashton (2001) found that some Conscientiousness- had to be published in a peer-reviewed journal. Second, the study had to
related facets are more closely linked to grades than others (see also include a self-report measure of the Big Five or the five-factor model;
Wolfe & Johnson, 1995); specifically, the Personality Research Form measures that assessed related traits (but not the actual domains) or single
Big Five/five-factor-model traits were not included. Third, the study had to
(Jackson, 1984) Achievement and Endurance scales were moderate
report correlations or regression (beta) coefficients for all five factors and
predictors of grades, whereas the Personality Research Form Order the criterion variable. Fourth, the study had to include either overall GPA
scale was almost completely unrelated to grades. In addition, re- or some aggregated measure of academic performance (a grade for a single
searchers who have assessed self-control, an aspect of Conscientious- course, summed exam scores in a course). Fifth, the sample had to be
ness (Roberts, Chernyshenko, Stark, & Goldberg, 2005), have found drawn from a college population. Twenty studies satisfied these criteria,
sizeable relations with grades (e.g., Tangney, Baumeister, & Boone, resulting in a total of 100 correlations on the basis of 5,292 participants.
2004). Using Costa and McCrae’s (1992) NEO PI–R, Gray and participants, allowing us to examine their differential personality
Watson (2002) found that GPA was most closely related to the correlates and whether the effects of personality on college GPA are
Conscientiousness facets of achievement striving (r ⫽ .39) and self- independent of high school GPA. Sixth, one of our studies included
discipline (r ⫽ .36) but only weakly related to the Conscientiousness GPA and SAT scores obtained from university records as well as
facet of order (r ⫽ .15).2 self-reported scores, allowing us to determine the degree to which
Gray and Watson (2002) found a divergent pattern of correlates for college students accurately report their GPA and SAT scores. Sev-
high school GPA and college GPA. When they simultaneously en- enth, we report findings from a 4-year longitudinal study, which
tered all of the Conscientious facets into multiple regression analyses, allowed us to test the long-term effects of personality, and changes in
dutifulness emerged as the best predictor of high school grades, personality, on academic outcomes. Eighth, we assessed several
whereas the achievement-striving facet emerged as the only signifi- achievement-related variables, including perceived verbal intelli-
cant predictor of college grades. These findings seem to fit with gence, perceived academic ability, and academic effort, to better
Gough’s (1957) distinction between achievement via conformance understand the processes that mediate any observed relations between
and achievement via independence, which have been linked, respec- personality and academic outcomes.
tively, to high school and college achievement (see Gough & Lan- On the basis of previous research, we expected that Conscientious-
ning, 1986). Achievement via conformance reflects the capacity to ness would be the best predictor of academic performance (both high
work effectively in highly structured educational contexts, whereas school and college GPA)— especially Conscientiousness facets that
achievement via independence reflects the ability to be successful in have to do with achievement motivation and self-control. In addition,
relatively open and unstructured settings (Gough & Lanning, 1986). we expected Openness to also be a predictor of academic performance
On the basis of Gough’s conceptualization of these two variables, we in college, in line with its conceptual link to achievement via inde-
would expect Openness—in addition to Conscientiousness—to be pendence. We expected that Openness would be related to SAT
linked to achievement via independence, and thus higher college verbal scores, but we made no predictions about personality correlates
GPA, because those who are high in Openness tend to have an of SAT math scores, given the inconsistencies in the literature. More-
intellectual style that is well-suited to contexts in which intellectual over, we expected that these effects would be independent; that is,
autonomy and creativity are rewarded. This prediction is consistent Conscientiousness and Openness would have independent effects on
with Gray and Watson’s finding that Openness was a significant GPA and SAT scores. We also expected that the effects of personality
positive predictor of college GPA but not high school GPA. on GPA would be independent of SAT and vice versa. Finally, we
The weak and inconsistent relation between Openness and GPA expected that all of these effects would replicate across the three
might also reflect the differential predictive validity of the facets of personality inventories used in the present research.
Openness. For example, although Gray and Watson (2002) did not
assess Openness facets, one might expect aspects of Openness that Method
have to do with creativity, imagination, and engagement in and
appreciation of abstract ideas to be positive predictors of college To test our basic research questions, we conducted four studies
GPA, whereas aspects that have to do with the unconventional with data collected on four independent samples, all of which
nature of the dimension may be negative predictors. In the present included measures of the Big Five dimensions, GPA, and SAT
research, we used multiple measures of the Big Five personality scores (as well as other variables). However, to simplify presen-
domains, including two that have facet-level scales, to gain a more tation of the findings, and to facilitate comparison of findings
nuanced understanding of connections between personality and across studies, we report the results from all four samples together.
academic achievement in college.
Samples
The Present Research Sample 1. Participants were drawn from the psychology depart-
Our primary goal was to examine the Big Five correlates of college ment subject pool at the University of California, Davis from 2003–
GPA and SAT scores. Our research builds on previous research in 2005. The sample included 10,497 undergraduate students (63%
several ways. First, we replicated our findings across four independent female, 37% male) from a diverse range of ethnic backgrounds (2%
samples, including one sample of over 10,000 college students. Sec- African American, 42% Asian, 34% Caucasian, 8% Hispanic/Latino,
ond, we replicated our findings across multiple personality invento- 2% Middle Eastern, 11% “Other/Multicultural,” and less than 1%
ries, including the Big Five Inventory (BFI; John & Srivastava, 1999), Native American). Participants ranged in age from 18 years to 30
the NEO Five-Factor Inventory (NEO-FFI), the Revised NEO Per- years (Mdn ⫽ 19 years; SD ⫽ 1.51).
sonality Inventory (NEO-PI–R; Costa & McCrae, 1992), and the Sample 2. Sample participants were drawn from the Berkeley
HEXACO Personality Inventory (HEXACO-PI; Lee & Ashton, Longitudinal Study, an ongoing study designed to examine personal-
2004). Third, we used the Big Five facet scales of the HEXACO-PI, ity, achievement motivation, and self-concept development during
as well as the Conscientiousness and Openness facet scales of the college and early adulthood (for further details about the study, see
NEO-PI–R, to examine how specific facets of the five broad person-
ality dimensions relate to indicators of academic aptitude and achieve- 2
de Fruyt and Mervielde (1996) previously reported relations between
ment. Fourth, we examined both GPA and SAT scores in the same NEO-PI–R facets and comprehensive exams in a Belgian sample. Their
samples of participants, allowing us to test whether the effects of findings for Conscientiousness facets were remarkably similar to Gray and
personality on GPA are independent of SAT, and vice versa. Fifth, we Watson’s (2002) findings—strongest relations for self-discipline and
examined both high school and college GPA in the same samples of achievement striving and weakest for order.
Robins, Fraley, Roberts, & Trzesniewski, 2001; Robins, Noftle, Trz- In Sample 4, the Conscientiousness and Openness to Experience
esniewski, & Roberts, 2005). The initial sample included 508 under- facet scales of the NEO-PI–R (96 items; Costa & McCrae, 1992)
graduate students who entered the University of California at Berke- were administered to participants. Alpha reliabilities for the 12
ley in 1992. Participants were recruited during the 1st week of their facet scales ranged from .53 (Openness to Actions)5 to .80 (Open-
1st year of college and then assessed annually throughout college. The ness to Aesthetics) and were .90 and .89 for the full-scale Consci-
original sample was diverse in terms of ethnicity (7% African Amer- entiousness and Openness dimensions, respectively. We did not
ican, 43% Asian, 36% Caucasian, 13% Hispanic/Latino, 1% Native administer facet scales related to Extraversion, Agreeableness, and
American), sex (56% female, 44% male), and socioeconomic status Neuroticism because of time constraints and because these three
(20% came from families with 1992 household incomes below domains are less conceptually and empirically linked to academic
$25,000 and 20% from families with household incomes above outcomes than are Conscientiousness and Openness.
$100,000). Participants ranged in age from 17 years to 30 years SAT scores. In all four samples, participants self-reported their
(Mdn ⫽ 18 years; SD ⫽ 1.10). The present study focused on a
verbal and math SAT scores. In Sample 2, SAT scores were also
subsample of 475 participants who completed the NEO-FFI at the
obtained from university records. Students’ self-reported total SAT
beginning of the 1st year of college and at the end of the 4th year of
scores correlated very highly with scores obtained from university
college.
records (r ⫽ .92), which is consistent with previous research (e.g.,
Sample 3. Participants were University of California, Davis stu-
Cassady, 2001).
dents taking introductory psychology classes who were part of the
subject pool in spring 2006. The sample included 470 undergraduate College and high school GPA. In all four samples, participants
students (78% female, 22% male) from a diverse range of ethnic reported their high school GPA and their current college GPA. In
backgrounds (2% African American, 43% Asian, 35% Caucasian, 7% Sample 2, the participants’ high school GPAs and current college
Hispanic/Latino, 4% Pacific Islander, 9% “Other/Multicultural,” and GPAs were also obtained from university records. As with SAT
less than 1% Native American). Participants ranged in age from 17 scores, we found that GPA obtained from university records cor-
years to 29 years (Mdn ⫽ 19 years; SD ⫽ 1.51). related very highly with self-reported GPA (r ⫽ .89). Similarly,
Sample 4. Participants were University of California, Davis stu- past research has found close associations between self-reported
dents taking introductory psychology classes who were part of the GPA and GPA obtained from university records (e.g., Gray &
subject pool in fall 2006. The sample included 425 undergraduate Watson, 2002, p. 190).
students (61% female, 39% male) from a diverse range of ethnic Perceived verbal intelligence, perceived academic ability, and ac-
backgrounds (1% African American, 41% Asian, 32% Caucasian, ademic effort. In Samples 2 and 3, participants were asked to rate
11% Hispanic/Latino, 3% Pacific Islander, 12% “Other/Multicul- their verbal intelligence relative to the general population ( perceived
tural,” and less than 1% Native American). Participants ranged in age verbal intelligence) using the following scale: 1 (bottom 5%), 2
from 17 years to 31 years (Mdn ⫽ 19 years; SD ⫽ 1.70). (bottom 10%), 3 (bottom 20%), 4 (bottom 30%), 5 (bottom 50%), 6
(top 50%), 7 (top 30%), 8 (top 20%), 9 (top 10%), 10 (top 5%). In
Measures Sample 2, perceived verbal intelligence was assessed at the end of the
2nd, 3rd, and 4th years of college, so a composite of all three
Personality. In Sample 1, the 44-item BFI (John & Srivastava, assessments was used for the analyses (except where noted below).
1999) was administered to participants in the quarterly prescreen- The correlation between perceived verbal intelligence and SAT verbal
ing questionnaire; alpha reliabilities were .87 for Extraversion, .79 was .34 ( p ⬍ .05) in Sample 2 and .42 ( p ⬍ .05) in Sample 3; the
for Agreeableness, .81 for Conscientiousness, .82 for Neuroticism, correlation between perceived verbal intelligence and SAT math was
and .79 for Openness to Experience. Items were rated on a 5-point
.10 (ns) in Sample 2 and .11 ( p ⬍ .05) in Sample 3. These correlations
scale ranging from 1 (strongly disagree) to 5 (strongly agree).
between self-reported intelligence and SAT scores are similar to those
In Sample 2, The 60-item NEO-FFI (Costa & McCrae, 1992)
was administered during the 1st week of college (Week 1) and at
the end of the 4th year (Year 4); alpha reliabilities were .83 and .82 3
Lee and Ashton (2004) refer to the Neuroticism dimension as “Emo-
for Extraversion, .76 and .77 for Agreeableness, .81 and .83 for tionality” in their HEXACO model. Although we refer to this domain as
Conscientiousness, .84 and .85 for Neuroticism, and .77 and .75 “Neuroticism” for ease of presentation of data across the three samples,
for Openness to Experience, respectively, for the two assessments. some conceptual differences between Lee and Ashton’s model and Big
Items were rated on a 5-point scale ranging from 1 (not very true Five Neuroticism should be noted. First, HEXACO Emotionality does not
of me) to 5 (very true of me). include the irritability or angry hostility content usually associated with
In Sample 3, the 208-item HEXACO-PI (Lee & Ashton, 2004, Neuroticism (this content is instead relegated to the low pole of Agree-
ableness). Second, HEXACO Emotionality includes content related to
2006) was administered to participants. Alpha reliabilities were .93
sentimentality and dependence (vs. toughness and bravery), which is
for Extraversion, .91 for Agreeableness, .89 for Conscientiousness, usually not associated with Neuroticism.
.87 for Emotionality (hereafter referred to as “Neuroticism” for ease 4
We do not report findings for the Honesty-Humility (H-H) domain,
of comparison across the three samples),3 and .87 for Openness to which is the sixth dimension of personality in Lee and Ashton’s (2004)
Experience.4 Each of the HEXACO-PI domain scales includes four model.
facet scales; alpha reliabilities for these facet scales ranged from .67 5
Although the Openness to Actions facet had by far the lowest alpha
(Openness:Unconventionality) to .88 (Conscientiousness:Organiza- reliability (the next lowest reliability was .63 for Competence), its reliabil-
tion). All items were rated on a 5-point scale ranging from 1 (strongly ity was comparable to that found for this scale in previous research (e.g.,
disagree) to 5 (strongly agree). ␣ ⫽ .58; Costa & McCrae, 1992, p. 44).
Table 2
Means (and Standard Deviations) for Personality and Academic Variables
Sample 3 Sample 4
Sample 1 (BFI) Sample 2 (NEO-FFI) (HEXACO) (NEO-PI-R)
Variable M SD M SD M SD M SD
Extraversion 3.26 0.75 3.47 0.57 3.21 0.56 — —

Agreeableness 3.71 0.60 3.62 0.51 3.05 0.49 — —
Conscientiousness 3.49 0.62 3.48 0.56 3.39 0.47 3.28 0.38
Neuroticism 2.95 0.72 2.87 0.64 3.45 0.44 — —
Openness 3.55 0.59 3.66 0.54 3.23 0.46 3.41 0.37
SAT verbal 567 94 546 100 567 99 567 99
SAT math 598 91 638 98 601 99 593 98
College GPA 2.92 0.55 3.14 0.54 3.00 0.56 2.89 0.51
High school GPA 3.73 0.35 3.91 0.48 3.76 0.33 3.70 0.36
Note. BFI ⫽ Big Five Inventory; NEO-FFI ⫽ NEO Five Factor Inventory; NEO-PI-R ⫽ Revised NEO Personality Inventory. Ns ⫽ 10,497 (BFI), 475
(NEO-FFI), 470 (HEXACO), and 425 (NEO-PI-R). Sample sizes for high school GPA are generally lower. Ns ⫽ 7,218 (BFI), 465 (NEO-FFI), 464
(HEXACO), and 414 (NEO-PI-R). The sample size for college GPA in Sample 4 is significantly lower (N ⫽ 256), because the study was conducted in
the fall quarter when the 1st-year students in the sample did not yet have college GPAs. — indicates that the data were not collected.
found in previous research between self-reported intelligence and IQ is quite substantial, despite the fact that the participants came from
tests (e.g., Paulhus, Lysy, & Yik, 1998). highly selective universities. The small mean differences between
In Samples 2 and 3, perceived academic ability was measured the UC Berkeley and UC Davis samples in GPA and SAT scores
with a standardized composite of six items (␣s ⫽ .91 and .84 for are probably due to cohort differences in SAT norming and the
the two samples, respectively), tapping two facets of perceived higher threshold of academic performance needed to gain admis-
ability: (a) direct self-reports of ability (“I am confident of my sion to UC Berkeley.
ability to do well in school,” “Compared to the average UC Across samples, women tended to score higher on Neuroticism
Berkeley [UC Davis] student, how would you rate your academic (rs ⫽ .23, .25, .20, and .23 for Samples 1, 2, 3, and 4, respectively;
ability?” “Do you think you have the ability to compete in col- ps ⬍ .01), which is consistent with past research (Costa, Terracciano,
lege?”) and (b) expected performance in college (“What overall & McCrae, 2001). The only other consistent gender difference was a
GPA do you think you are capable of attaining?” “Realistically, tendency for women to score lower on the SAT math test (rs ⫽ .23,
what overall GPA do you think you will attain?” “What is the .25, and .20 for Samples 1, 2, and 3, respectively; ps ⬍ .01), which
lowest overall GPA you would be satisfied attaining?”). In Sample also replicates past research (Stricker, Rock, & Burton, 1991). Given
2, perceived academic ability was assessed at the end of the 1st, that gender is correlated with two of our key variables, in subsequent
2nd, 3rd, and 4th years of college, so a composite of all assess- analyses we tested whether gender accounts for any of the observed
ments was used for the analyses (except where noted below). The personality effects.
correlation between perceived academic ability and college GPA
was .60 ( p ⬍ .05) in Sample 2 and .61 ( p ⬍ .05) in Sample 3. Big Five Correlates of SAT Scores
In Samples 2 and 3, academic effort was measured with a
standardized composite of two items, tapping into behaviors and Table 3 shows correlations between the Big Five dimensions
self-perceptions related to the effort participants put into school and SAT verbal and math scores. As predicted, Openness was
work: “On average, how many hours a week (outside of class time) consistently related to higher SAT verbal scores (rs ranged from
have you spent on school work the current semester [quarter]?” .20 to .26 across the samples);6 the Openness effect held when all
and “How much effort have you been devoting to your school five personality dimensions were entered simultaneously in a
work this current semester [quarter] in order to achieve the grades multiple regression analysis predicting SAT scores. The other four
you hope for?” In Sample 2, academic effort was assessed at the Big Five dimensions were not consistently related to SAT verbal
end of the 1st, 2nd, 3rd, and 4th years of college, so a composite scores, although we found a weak and inconsistent tendency for
of all assessments was used for the analyses (except where noted). individuals who were low in Extraversion and Agreeableness to
The correlation between academic effort and college GPA was .25 have higher SAT verbal scores. None of the Big Five dimensions
( p ⬍ .05) in Sample 2 and .17 ( p ⬍ .05) in Sample 3. was consistently related to SAT math scores across samples,
although significant but weak (rs ranging from .05 to .07) corre-
Results
6
Descriptive Statistics We examined differences between native English speakers and non-
native English speakers for our main findings and found little difference in
Table 2 shows means and standard deviations for the SAT, the magnitude of correlations between the two groups (for example, the
GPA, and the Big Five separately in each of the four samples. In correlations between Openness and SAT verbal scores were .17 and .16 for
all four samples, the range and variance in SAT scores and GPA native English and nonnative English speakers, respectively).
Table 3
Big Five Correlates of SAT Verbal and Math Scores
SAT verbal SAT math
Sample 1 Sample 2 Sample 3 Sample 4 Sample 1 Sample 2 Sample 3 Sample 4

Variable (BFI) (NEO-FFI) (HEXACO) (NEO-PI-R) (BFI) (NEO-FFI) (HEXACO) (NEO-PI-R)
Extraversion .02 (⫺.04*) ⫺.15* (⫺.17*) .07 (⫺.02) — ⫺.06* (⫺.09*) ⫺.08 (⫺.08) ⫺.04 (⫺.05) —
Agreeableness ⫺.03* (⫺.08*) ⫺.05 (⫺.03) ⫺.10 (⫺.16*) — ⫺.06* (⫺.08*) ⫺.06 (⫺.05) ⫺.03 (⫺.04) —
Conscientiousness ⫺.01* (⫺.02) ⫺.09 (⫺.05) .05 (.05) .00 (⫺.02) ⫺.07* (⫺.06*) ⫺.03 (.00) ⫺.03 (⫺.00) ⫺.06 (⫺.06)
Neuroticism ⫺.05* (⫺.07*) .03 (⫺.03) ⫺.02 (⫺.01) — ⫺.07* (⫺.12*) .03 (⫺.01) ⫺.08 (⫺.08) —
Openness .20* (.21*) .20* (.22*) .26* (.28*) .26* (.27*) .05* (.08*) .02 (.03) .04 (.06) .05 (.06)
Multiple R .22* .27* .30* .26* .17* .10 .11 .08
Note. BFI ⫽ Big Five Inventory; NEO-FFI ⫽ NEO Five Factor Inventory; NEO-PI–R ⫽ Revised NEO Personality Inventory. Ns ⫽ 10,497 (BFI), 475
(NEO-FFI), 448 (HEXACO), and 407 (NEO-PI-R). Values in the table are correlations (with standardized beta weights in parentheses). — indicates that
the data were not collected.
*
p ⬍ .01.
lations were found in the BFI sample (Sample 1), because even correlations, all four HEXACO Openness facets were related to
very small correlations will be significant in such a very large SAT verbal scores, with correlations ranging from .15 for Uncon-
sample. ventionality to .23 for Aesthetic Appreciation. In addition, five of
the six NEO-PI–R facets were related to SAT verbal scores, with
Facet-Level Correlates of SAT Scores correlations ranging from .12 for Openness to Aesthetics to .26 for
Openness to Values. Of all the Openness facets, only Openness to
To gain a more detailed understanding of the personality effects, Actions was unrelated to SAT verbal scores (r ⫽ .01; ns).
we examined relations between the facets of Conscientiousness In general, the Conscientiousness facets were not related to SAT
and Openness, as measured by the HEXACO-PI and NEO-PI–R, scores. There were two exceptions: HEXACO Perfectionism,
and SAT scores (see Table 4). Consistent with the domain-level which reflects a tendency toward thoroughness and attention to
detail, was positively related to SAT verbal scores (r ⫽ .12),
Table 4 whereas HEXACO Organization, which reflects orderliness and
Facet-Level Correlates of SAT Verbal and Math Scores in neatness, was negatively related to SAT math scores (r ⫽ ⫺.14).
Samples 3 and 4 An examination of the Extraversion, Agreeableness, and Neu-
roticism facets of the HEXACO revealed few significant effects,
Personality facet scale (␣ reliability) SAT verbal SAT math consistent with the findings at the level of the broad domains.
There was one exception: the Flexibility facet of Agreeableness
HEXACO Conscientiousness (.89) .05 ⫺.03
had a negative relation with SAT verbal scores (r ⫽ ⫺.14).
Organization (.88) ⫺.09 ⫺.14*
Diligence (.80) .08 .02
Perfectionism (.77) .12* .04 Big Five Correlates of GPA
Prudence (.77) .05 .03
NEO-PI-R Conscientiousness (.90) .00 ⫺.06 Table 5 shows correlations between the Big Five dimensions
Competence (.63) .12 ⫺.02 and college and high school GPA. The strongest and most consis-
Order (.66) ⫺.04 ⫺.03 tent finding is that Conscientiousness was positively related to
Dutifulness (.67) .00 .02 both college and high school GPA (seven of the eight zero-order
Achievement-Striving (.70) .00 ⫺.09
correlations were significant and ranged from .18 to .26). These
Self-Discipline (.77) ⫺.05 ⫺.04
Deliberation (.69) ⫺.04 ⫺.11 relations held when all five personality dimensions were entered
HEXACO Openness (.87) .26* .04 simultaneously in a multiple regression analysis predicting GPA.
Aesthetic Appreciation (.81) .23* .00 We found some support for our prediction that Openness and
Inquisitiveness (.72) .20* .03 college GPA (but not high school GPA) would be correlated.
Creativity (.78) .20* .07
Unconventionality (.67) .15* .04 Openness was weakly, but significantly, related to college GPA in
NEO-PI-R Openness (.89) .26* .05 three of four samples but was never significantly correlated with
Fantasy (.76) .24* .02 high school GPA. Finally, we found a weak negative relation
Aesthetics (.80) .12* ⫺.05 between Extraversion and college GPA and a weak positive rela-
Feelings (.76) .18* .00
tion between Agreeableness and high school GPA, but neither of
Actions (.53) .01 .03
Ideas (.75) .23* .12 these relations replicated across all four samples.
Values (.68) .26* .10
Facet-Level Correlates of GPA
Note. NEO-PI–R ⫽ Revised NEO Personality Inventory. Ns ⫽ 444
(HEXACO) and 407 (NEO-PI-R). To better understand the effects of the broad personality do-
*
p ⬍ .01. mains, we examined relations between the facets of Conscientious-
Table 5
Big Five Correlates of GPA
College GPA High school GPA
Sample 1 Sample 2 Sample 3 Sample 4 Sample 1 Sample 2 Sample 3 Sample 4

Variable (BFI) (NEO-FFI) (HEXACO) (NEO-PI-R) (BFI) (NEO-FFI) (HEXACO) (NEO-PI-R)
Extraversion ⫺.02 (⫺.06*) .02 (⫺.05) ⫺.11 (⫺.17*) — .03 (.01) ⫺.09 (⫺.11) .03 (⫺.01) —
Agreeableness .03* (⫺.03*) .10 (.05) ⫺.03 (⫺.07) — .10* (.05*) .06 (.06) .11* (.07) —
Conscientiousness .22* (.25*) .19* (.19*) .20* (.22*) .18 (17*)
*
.22* (.23*) .10 (.12) .26* (.24*) .25 (.25*)
*
Neuroticism .04* (.07*) ⫺.08 (⫺.02) .10 (.06) — .03 (.09*) .04 (.05) .05 (.01) —
Openness .06* (.06*) .13* (.14*) .05 (.10) .13* (.11) .01 (⫺.02) .03 (.05) .02 (⫺.01) .04 (.03)
Multiple R .25* .25* .28* .21* .24* .17 .26* .25*
Note. BFI ⫽ Big Five Inventory; NEO-FFI ⫽ NEO Five Factor Inventory; NEO-PI–R ⫽ Revised NEO Personality Inventory. Ns (college GPA) ⫽ 10,497
(BFI), 475 (NEO-FFI), 470 (HEXACO), and 256 (NEO-PI-R); Ns (high school GPA) ⫽ 7,218 (BFI), 465 (NEO-FFI), 464 (HEXACO), and 414
(NEO-PI-R). Values in the table are correlations (with standardized beta weights in parentheses). — indicates that the data were not collected.
*
p ⬍ .01.
ness and Openness, assessed by the HEXACO-PI and NEO-PI–R High school GPA was predicted by the same three HEXACO
scales, and GPA (see Table 6). Three HEXACO Conscientious- Conscientiousness facets (rs ranged from .20 to .24) as college
ness facets were significantly associated with college GPA: Dili- GPA; in addition, HEXACO Organization also predicted high
gence (r ⫽ .24), Prudence (r ⫽ .21), and Perfectionism (r ⫽ .19). school GPA (r ⫽ .11), albeit modestly. High school GPA had a
Three NEO-PI–R facets were significantly associated with college somewhat different pattern of correlations with the NEO-PI–R
GPA: Achievement-Striving (r ⫽ .21), Competence (r ⫽ .19), and facets. We found that all six facets of the NEO-PI–R were asso-
Self-Discipline (r ⫽ .15), which replicates Gray and Watson’s ciated with high school GPA, ranging from .13 for Order to .22 for
(2000) results, although they also found significant correlations Competence and Achievement-Striving. These relations demon-
with some of the other facets. strate that it is the goal-driven (HEXACO Diligence and NEO-
PI–R Achievement-Striving) and self-controlled (HEXACO Pru-
dence and NEO-PI–R Self-Discipline) aspects of
Table 6 Conscientiousness that are most important for academic achieve-
Facet-Level Personality Correlates of GPA in Samples 3 and 4 ment, whether at the high school or the college level. The weak or
Personality facet scale nonsignificant associations between NEO-PI–R Order and
(␣ reliability) College GPA High school GPA HEXACO Organization and college and high school GPA is a
point to which we return in the Discussion section.
HEXACO Conscientiousness (.89) .20* .26* In general, the Openness facets were not related to either high
Organization (.88) .00 .11*
Diligence (.80) .24* .24*
school or college GPA, which is consistent with the weak domain-
Perfectionism (.77) .19* .20* level effects. The one exception is that NEO-PI–R Openness to
Prudence (.77) .21* .21* Values was positively related to college GPA (r ⫽ .18).
NEO-PI-R Conscientiousness (.90) .18* .25* As with the SAT, an examination of the Extraversion, Agree-
Competence (.63) .19* .22* ableness, and Neuroticism facets of the HEXACO revealed few
Order (.66) .08 .13*
Dutifulness (.67) .13 .17* significant effects. College GPA was associated with (low) Socia-
Achievement-Striving (.70) .21* .22* bility (r ⫽ ⫺.15). High school GPA was associated with the
Self-Discipline (.77) .15* .18* Agreeableness facet of Patience and the Neuroticism/Emotionality
Deliberation (.69) .07 .17* facet of Sentimentality (both rs ⫽ .14).
HEXACO Openness (.87) .05 .03
Aesthetic Appreciation (.81) .08 .07
Inquisitiveness (.72) .07 .04
Creativity (.78) .00 .01 Are the Personality Correlates of GPA Independent of
Unconventionality (.67) ⫺.02 ⫺.06 SAT Scores and Vice Versa?
NEO-PI-R Openness (.89) .13 .04
Fantasy (.76) .05 ⫺.04 Table 7 shows the results of multiple regression analyses predicting
Aesthetics (.80) .10 .08 GPA, in which gender and SAT scores were entered at Step 1 and the
Feelings (.76) .12 .10
Actions (.53) ⫺.04 ⫺.10 Big Five dimensions were entered at Step 2. Adding the Big Five
Ideas (.75) .09 .07 dimensions at Step 2 produced a significant increase in R2 in all three
Values (.68) .17* .02 samples for college and high school GPA. Consistent with the zero-
order correlations, Conscientiousness was a significant positive pre-
Note. NEO-PI–R ⫽ Revised NEO Personality Inventory. Ns ⫽ 464 – 470
(HEXACO), 256 (college GPA and NEO-PI-R), and 414 (high school GPA
dictor of GPA, even controlling for gender and SAT scores, and this
and NEO-PI-R). finding replicated across all three samples. Thus, personality, in par-
*
p ⬍ .01. ticular the Conscientiousness dimension, and SAT scores have inde-
Table 7
Independent Effects of Big Five and SAT Scores on GPA
College GPA High school GPA
Sample 1 Sample 2 Sample 3 Sample 1 Sample 2 Sample 3

Variable (BFI) (NEO-FFI) (HEXACO) (BFI) (NEO-FFI) (HEXACO)
Step 1
Gender .17* .16* .10 .18* .09 .12
SAT verbal .20* .27* .19* .09* .15* .11
SAT math .15* .28* .24* .12* .54* .07
Multiple R .31* .49* .38* .22* .63* .19*
Step 2
Gender .13* .13* .06 .13* .05 .11
SAT verbal .19* .28* .18* .10* .16* .12
SAT math .16* .28* .25* .13* .53* .07
Extraversion ⫺.05* .01 ⫺.14* .00 ⫺.05 .00
Agreeableness ⫺.02 .04 ⫺.04 .05* .08 .07
Conscientiousness .24* .18* .22* .21* .12* .24*
Neuroticism .06* ⫺.07 .07 .07* .04 ⫺.03
Openness .01 .06 .03 ⫺.05* ⫺.01 ⫺.06
Multiple R .39* .54* .45* .31* .64* .32*
Note. BFI ⫽ Big Five Inventory; NEO-FFI ⫽ NEO Five Factor Inventory. Ns (college GPA) ⫽ 10,472 (BFI), 465 (NEO-FFI), and 444 (HEXACO); Ns
(high school GPA) ⫽ 7,214 (BFI), 465 (NEO-FFI), and 446 (HEXACO). Values in table are standardized beta weights. All R2 changes from Step 1 to Step
2 were significant. Gender is keyed toward female, so positive beta weights indicate that women have higher levels of the criterion variable.
*
p ⬍ .01.
pendent effects on both high school and college grades. Indeed, in Table 8 shows the results of multiple regression analyses
several cases, Conscientiousness was a slightly stronger predictor of predicting SAT scores, in which gender and high school GPA
GPA than were SAT scores. The overall multiple correlations, with all scores were entered at Step 1 and the Big Five dimensions were
variables entered in the equation, were moderate to large, ranging entered at Step 2. Adding the Big Five dimensions at Step 2
from .31 to .64. produced a significant increase in R2 for both SAT verbal and
Table 8
Independent Effects of Big Five and High School GPA on SAT Scores
SAT verbal SAT math
Sample 1 Sample 2 Sample 3 Sample 1 Sample 2 Sample 3

Variable (BFI) (NEO-FFI) (HEXACO) (BFI) (NEO-FFI) (HEXACO)
Step 1
Gender ⫺.12* ⫺.10 ⫺.04 ⫺.24* ⫺.21* ⫺.21*
High school GPA .14* .45* .16* .15* .60* .13*
Multiple R .17* .47* .16* .26* .64* .24*
Step 2
Gender ⫺.08* ⫺.10 ⫺.03 ⫺.20* ⫺.21* ⫺.20*
High School GPA .15* .45* .16* .17* .60* .14*
Extraversion ⫺.03 ⫺.12* .00 ⫺.06* .00 ⫺.04
Agreeableness ⫺.07* ⫺.03 ⫺.16* ⫺.05* ⫺.03 ⫺.03
Conscientiousness ⫺.03 ⫺.09 .00 ⫺.06* ⫺.04 ⫺.04
Neuroticism ⫺.05* ⫺.02 .00 ⫺.08* .04 .00
Openness .22* .21* .28* .07* .02 .07
Multiple R .27* .54* .34* .29* .65* .25*
Note. BFI ⫽ Big Five Inventory; NEO-FFI ⫽ NEO Five Factor Inventory. Ns ⫽ 7,214 (BFI), 475 (NEO-FFI), and 442 (HEXACO). Values in table are
standardized beta weights. All R2 changes from Step 1 to Step 2 were significant except for SAT math analyses for the NEO-FFI and HEXACO samples.
Gender is keyed toward female, so positive beta weights indicate that women have higher levels of the criterion variable.
*
p ⬍ .01.
SAT math in all three samples. Consistent with the zero-order Longitudinal Relations Between the Big Five and GPA
correlations, Openness was a significant predictor of SAT ver-
bal scores across all samples, even after controlling for the (in We followed the participants in Sample 2 from the beginning to
one case substantial) relation between high school GPA and the end of college, which allowed us to examine the relation
SAT verbal. Thus, Openness and high school GPA were inde- between change in personality and academic performance. To
pendently associated with SAT verbal scores. There were no assess change in the Big Five dimensions, we regressed each of the
consistent personality predictors of SAT math scores. The over- Year-4 NEO scores on the corresponding Week-1 NEO scores and
all multiple correlations, with all variables entered in the equa- saved the standardized residuals. These residual-change scores
tion, were moderate to large, ranging from .25 to .65. provide an individual-level measure of how much a person
Table 9 shows the results of multiple regression analyses changed and in which direction; they adjust for differences in
predicting college GPA, in which gender, high school GPA, and initial status and thus estimate how much individuals would have
SAT scores were entered at Step 1 and the Big Five dimensions changed had they all started out at the same level. Positive scores
were entered at Step 2. Adding the Big Five dimensions at Step indicate relative increases over time, and negative scores indicate
2 produced a significant increase in R2 in all three samples. relative decreases.
Consistent with the zero-order correlations, Conscientiousness We found that individuals who increased in Conscientiousness
was the strongest and most consistent predictor of college GPA, over the course of college tended to have higher GPAs (r ⫽ .22,
even after controlling for both high school GPA and SAT p ⬍ .01). This relation held even after controlling for mean levels
scores. Thus, Conscientiousness, prior academic performance, of Conscientiousness (partial r ⫽ .16, p ⬍ .01). Changes in
and standardized test scores independently predicted college Extraversion, Agreeableness, Neuroticism, and Openness were not
performance. The overall multiple correlations, with all vari- significantly related to GPA (rs ranged from ⫺.06 to .07, ns).
ables entered in the equation, were moderate to large, ranging
from .40 to .60. Testing Mediators of the Effects of Personality on
Together, these analyses show that personality traits have inde- Academic Outcomes
pendent and incremental effects on academic outcomes, even after
controlling for traditional predictors of those outcomes. We found two basic effects in our analyses in which we used
personality to predict academic outcomes: Conscientiousness was
related to higher grades, and Openness was related to higher SAT
verbal scores. In this section, we explore potential mediators of
Table 9 these two effects. First, we tested the hypothesis that Conscien-
Independent Effects of Big Five, SAT Scores, and High School tiousness is associated with higher grades because it is associated
GPA on College GPA with increased effort and higher levels of perceived academic
ability. Consistent with social-cognitive views of personality and
College GPA achievement, we assumed that the personality trait of Conscien-
Sample 1 Sample 2 Sample 3
tiousness would be expressed and exert its effects on real-world
Variable (BFI) (NEO-FFI) (HEXACO) outcomes by shaping a person’s thoughts and feelings (e.g., their
self-perceptions) and by influencing their actual behavior (e.g.,
Step 1 their level of effort). Second, we tested the hypothesis that Open-
Gender .12* .13* .08
ness is associated with higher SAT verbal scores because it is
High school GPA .19* .36* .13*
SAT verbal .18* .21* .18* associated with higher levels of perceived verbal intelligence. In
SAT math .13* .09 .23* this case, we are assuming that openness might influence a per-
son’s identity as a verbally intelligent person (e.g., as an “intel-
Multiple R .36* .56* .40* lectual”) and that this self-perception might translate into the
Step 2 pursuit of activities (e.g., reading, attending the theater, engaging
Gender .10* .12* .06 in philosophical discussions, interest in the arts) that ultimately
High school GPA .15* .33* .09 impact actual verbal intelligence, as assessed by the SAT.
SAT verbal .18* .22* .17* Academic effort. Academic effort was significantly associated
SAT math .14* .11 .24*
Extraversion ⫺.04* .03 ⫺.15*
with both Conscientiousness (r ⫽ .37 and .35 for Samples 2 and 3,
Agreeableness ⫺.04* .01 ⫺.05 respectively), and college GPA (r ⫽ .22 and .17 for Samples 2 and
Conscientiousness .21* .14* .19* 3, respectively). We conducted multiple regression analyses to test
Neuroticism .05* ⫺.08 .06 whether effort mediates the relation between Conscientiousness
Openness .02 .06 .04 and college GPA. We found evidence of significant mediation in
Multiple R .40* .60* .46* both Samples 2 (t ⫽ 3.02; p ⬍ .05) and 3 (t ⫽ 2.23; p ⬍ .05), on
the basis of a Sobel test. The relation between Conscientiousness
Note. BFI ⫽ Big Five Inventory; NEO-FFI ⫽ NEO Five Factor Inven- and college GPA was reduced from .19 to .10 (Sample 2) and from
tory. Ns ⫽ 7,214 (BFI), 465 (NEO-FFI), and 442 (HEXACO). Values in .20 to .16 (Sample 3) when effort was entered in the regression
table are standardized beta weights. All R2 changes from Step 1 to Step 2
were significant. Gender is keyed toward female, so positive beta weights
equation.
indicate that women have higher levels of the criterion variable. As a more stringent test of mediation, we examined these
*
p ⬍ .01. relations longitudinally, using the data from Sample 2. Specifi-
cally, we conducted a multiple regression analysis to test whether controlling for gender, SAT scores, and high school GPA. More-
effort assessed in the middle of college (i.e., in Years 1, 2, and 3), over, individuals who increased in Conscientiousness over the
mediated the effect of Conscientiousness, assessed at the begin- course of college tended to attain higher GPAs, even after con-
ning of college (Week 1), on cumulative college GPA, assessed at trolling for their average level of Conscientiousness during col-
the end of college (Year 4). Again, effort emerged as a significant lege. Further analyses showed that the effects of Conscientiousness
mediator (t ⫽ 3.67; p ⬍ .05); the effect of Conscientiousness was on college achievement were mediated by increased effort and
reduced from .13 to ⫺.01 when effort was entered in the regression more positive perceptions of one’s academic ability.
equation. The link between Conscientiousness and GPA is consistent with
Perceived academic ability. Perceived academic ability was the extant literature. However, there have been few studies that
significantly associated with both Conscientiousness (r ⫽ .19 and have demonstrated the somewhat surprising robust predictive
.25 for Samples 2 and 3, respectively) and college GPA (r ⫽ .62 power of the trait—that it remains a significant predictor even
and .61 for Samples 2 and 3, respectively). Consistent with the when controlling for traditional indicators of academic achieve-
findings for effort, perceived ability was a significant mediator of ment, such as SAT scores and high school GPA. A recent excep-
the relation between Conscientiousness and college GPA in both tion is Wagerman and Funder (2007), who showed that this pre-
Sample 2 (t ⫽ 3.66; p ⬍ .05) and Sample 3 (t ⫽ 5.17; p ⬍ .05). dictive power of Conscientiousness even extends to informant
The relation between Conscientiousness and college GPA was reports of the trait. In addition, analyses of the HEXACO and
reduced from .19 to .05 (Sample 2) and from .20 to .06 (Sample 3) NEO-PI–R facets demonstrate that it is the achievement-striving,
when perceived ability was entered in the regression equation. persevering, and self-controlled aspects of Conscientiousness, and
Moreover, this mediator effect replicated when we examined lon- not the orderly or organized aspects of Conscientiousness, that are
gitudinal relations in Sample 2 (t ⫽ 2.08; p ⬍ .05); the effect of most important to both high school and college achievement.
Conscientiousness on GPA was reduced from .13 to .00 when The finding that Openness was generally weakly related to
perceived ability was entered in the regression equation. academic performance was somewhat surprising. Although we
Perceived verbal intelligence. Perceived verbal intelligence expected Conscientiousness to be the major predictor of both high
was significantly associated with both Openness (r ⫽ .27 and .28
school and college GPA, we also expected Openness to be a
for Samples 2 and 3, respectively; ps ⬍ .01) and SAT verbal scores
predictor, at least of college GPA. Gough’s (1957) model of
(r ⫽ .42 for both samples, ps ⬍ .01). As expected, perceived
achievement distinguishes between achievement via conformance
verbal intelligence was a significant mediator of the relation be-
and achievement via independence, which correspond conceptu-
tween Openness and SAT verbal scores in both Sample 2 (t ⫽
ally to Conscientiousness and Conscientiousness plus Openness,
4.26; p ⬍ .05) and Sample 3 (t ⫽ 5.01; p ⬍ .05). The effect of
respectively. This distinction is also related to Hough’s (1992)
Openness on SAT verbal scores was reduced from .20 to .12
distinction between dependability and achievement, or conscien-
(Sample 2) and from .26 to .19 (Sample 3) when perceived verbal
tiousness in reference to others’ aims versus conscientiousness in
intelligence was entered in the regression equation.
reference to one’s own aims. Drawing from Gough and Lanning
In sum, we found that all three variables—perceived verbal
(1986), and from intuitive notions of college achievement as
intelligence, perceived academic ability, and academic effort—
were significant mediators. It is important to note that the mediator necessitating more independent approaches to schoolwork, we
effects held up when we analyzed the relations longitudinally from hypothesized that success in high school would entail achievement
the beginning to the end of college. Together, these results help to via conformance and thus be linked to Conscientiousness, whereas
explain our two basic findings and provide a preliminary glimpse success in college would entail both achievement via conformance
into the processes underlying the personality–academic-outcome and via independence and thus be linked to both Conscientiousness
relations. and Openness. Although the findings are consistent with our
expectations for Conscientiousness, the general lack of Openness
effects on GPA was surprising. We did find a relation between
Discussion Openness and college GPA in three of the four samples, but the
The present findings shed new light on the relation between relations were very small and disappeared when we controlled for
personality and academic aptitude and achievement. Data from gender and SAT scores. The facet analyses, instead of revealing
four samples and four different personality inventories demon- associations between GPA and aspects of Openness that were
strate a positive relation between Conscientiousness and college obscured in the domain-level analyses, further confirmed the lack
GPA and a positive relation between Openness to Experience and of a clear association between GPA and Openness. It is possible
SAT verbal scores. These relations were similar regardless of that if we had included scales that are combinations of high
whether we analyzed GPA and SAT scores from self-reports or Openness and Conscientiousness, such as traits like industrious-
university records. We discuss each of these findings in turn ness and foresightedness represented in the ABC5 structure (Hof-
below. stee, De Raad & Goldberg, 1992), which fit more closely with
Gough’s conception of achievement via independence, we may
have found that these “interstitial” traits were more strongly cor-
Conscientiousness and GPA
related with college GPA. However, if we had included more
Conscientiousness, whether assessed by the NEO-FFI, the BFI, lexically based measures of the Big Five, which conceptualize
the HEXACO, or the NEO-PI–R, was associated with higher Openness as Intellect (Goldberg, 1990), we may have found stron-
college grades, a robust association that persisted even when ger relations with college GPA.
Unlike the other facets of Conscientious, the HEXACO organi- fits with the findings of Ashton et al. (2000), who found moderate
zation and NEO-PI–R order facets were completely unrelated to to strong relations between Openness and aspects of crystallized
college GPA and were the most weakly linked out of each group intelligence and only a weak (or no) relation with aspects of fluid
of facets to high school GPA. This replicates the findings of intelligence. The differential personality facet correlates of SAT
Paunonen and Ashton (2001) and also fits well with Roberts et verbal and math scores demonstrate the problematic trend in past
al.’s (2005) study of facets of Conscientiousness in an adult research to identify personality correlates of summed SAT scores
sample. Roberts et al. found that different facets of Conscientious- and highlight the benefits of considering the two scores separately.
ness differentially predicted a set of behaviors comprising work These findings also support the claim that intelligence and person-
dedication, which was predicted positively by industriousness and ality are slightly overlapping but largely differentiable individual
self-control, and negatively by order, when all Conscientiousness differences (McCrae & Costa, 1997).
facets were entered simultaneously into a regression equation.
Overall, these findings support the idea that organization, order,
and generally keeping one’s ducks all in a row, has little to do with Implications, Conclusions, and Future Directions
achievement, and at high levels, such a rigidity (as in obsessive-
compulsive disorder) may actually hinder one’s steady progress In a review of the literature on predicting academic success,
towards valued goals. These findings might also help to explain Mouw and Khanna (1993) bemoaned the relatively weak predic-
past research that has found weaker relations between Conscien- tive validity of ability tests and suggested that willingness to work
tiousness and academic achievement. For example, using Gold- hard is “quite probably, at least half the battle” (Mouw & Khanna,
berg’s (1992) markers, E. C. Hair and Graziano (2003) found 1993, p. 334). Because the resurgence of interest in personality
stronger longitudinal relations between high school GPA and traits as predictors of important life outcomes (Hogan, 2005; Ozer
middle school self-rated Openness and Agreeableness than they & Benet-Martı́nez, 2006; Roberts et al., in press), studies such as
found for Conscientiousness. However, Goldberg’s markers of this one have since attempted to address the role of personality in
Conscientiousness are most strongly related to the organization academic achievement. Although willingness to work hard, as
and orderliness aspects of Conscientiousness and less strongly captured by Conscientiousness, was not quite as strong of a pre-
related to being hard working and persistent, traits that we found dictor as Mouw and Khanna might have expected (i.e., not quite
to be most closely related to earning good grades. “half the battle”), we found it to be a robust predictor of academic
Our results, together with those of Paunonen, Ashton, Goldberg, success. Overall, the present findings reveal highly replicable
and others, indicate that it is important to measure traits both relations between personality traits and the two most commonly
broadly and deeply and that studying only broad personality di- used indicators of academic aptitude and achievement: the SAT
mensions will not suffice in properly predicting these sorts of and GPA.
outcomes. It is important to neither overestimate nor underestimate the
practical significance of these findings. Although replicable, the
personality effects were generally small in magnitude. Several
Openness and SAT Scores points are worth considering. First, although the effect sizes found
We found a robust relation between Openness to Experience and in the present study suggest that personality explains only a small
SAT verbal scores. Individuals who are high in Openness, whether proportion of the variance in academic aptitude and achievement,
assessed by the NEO-FFI, the BFI, the HEXACO-PI, or the the magnitude of these effects is nonetheless comparable to other
NEO-PI–R, tended to score higher on the SAT verbal test, even widely used predictive tests in the biomedical sciences, such as the
after we controlled for their gender and their prior and concurrent prediction of heart disease by electrocardiogram stress tests, preg-
academic achievement. Facet-level analyses revealed that SAT nancy outcomes by ultrasound exams, and breast cancer by screen-
verbal scores were related to virtually every Openness facet on the ing mammograms (Meyer et al., 2001). Moreover, they are similar
HEXACO and the NEO-PI–R; the only exception was the Open- to the effects of personality on other life domains, such as rela-
ness to Actions facet of the NEO-PI–R, suggesting that verbal tionship, health, and work outcomes (Ozer & Benet-Martı́nez,
ability is unrelated to novelty seeking, In contrast, Conscientious- 2006).
ness (and its facets) were almost completely unrelated to SAT Second, small effects are to be expected when predicting a
scores, at least in our four samples. Thus, being a verbally intel- multiply determined outcome (Ahadi & Diener, 1989), and aca-
ligent individual has more to do with being creative, imaginative, demic achievement is a quintessential example of such an out-
and inquisitive than it does with being hard working, organized, come. In our own data, we saw that when personality and SAT test
and industrious. In contrast, mathematical aptitude, as measured by scores are combined to predict college GPA, the predictive validity
the SAT math test, does not seem to be reliably associated with any can reach moderate to high levels. Similarly, our mediational
of the Big Five personality traits. analyses suggest that exerting effort toward schoolwork, as well as
One interpretation of these findings might have to do with the having perceptions of one’s self as academically competent and
differential relation of the two sections of the SAT to intelligence, verbally intelligent, play an important role in predicting academic
which would explain why Openness is related to verbal but not outcomes. Other factors, such as values, motives, study and test-
math scores. The SAT verbal section may be related more strongly taking skills, attributional style, self-efficacy, and even pragmatic
to crystallized intelligence because of its vocabulary-related con- factors related to financial resources (e.g., not needing to work
tent, whereas the SAT math section may be related more strongly outside of school), are also important ingredients for academic
to fluid intelligence because of its reasoning-related content. This success. Thus, a wide range of predictors are needed to fully
account for individual differences in academic aptitude and Caspi, A., Bem, D. J., & Elder, G. H. (1989). Continuities and conse-
achievement. quences of interactional styles across the life course. Journal of Person-
Third, it is often overlooked that small effect sizes can have a ality, 57, 375– 406.
major impact on outcomes over time (Abelson, 1985; Rosenthal & Cassady, J. C. (2001). Self-reported GPA and SAT: A methodological
note. Practical Assessment, Research, and Evaluation, 7(12). Retrieved
Rubin, 1982). Over the course of a lived life, even relatively small
January 8, 2007, from http://PAREonline.net/getvn.asp?v⫽7&n⫽12
differences in achievement between individuals who are high and
Cattell, R. B. (2000). The 16PF Industrial Personality technical manual.
those who are low in Conscientiousness and Openness might Oxford, United Kingdom: Industrial Psychology, Psytech International,
produce large cumulative differences in lifetime achievements. and The Test Agency.
This may take place through the mechanism of cumulative conti- Chamorro-Premuzic, T., & Furnham, A. (2003a). Personality predicts
nuity, or Caspi and colleagues’ idea that behaviors are reinforced academic performance: Evidence from two longitudinal university sam-
and maintained through the snowballing accrual of their outcomes ples. Journal of Research in Personality, 37, 319 –338.
(Caspi, Bem, & Elder, 1989; Roberts & Caspi, 2003). Chamorro-Premuzic, T., & Furnham, A. (2003b). Personality traits and
With personality traits established as important predictors of academic examination performance. European Journal of Personality,
academic outcomes and some preliminary insights into the medi- 17, 237–250.
ating processes, future researchers should be able examine the Collis, J. M., & Messick, S. (Eds.). (2001). Intelligence and personality:
interplay between traits and other factors in academic achieve- Bridging the gap in theory and measurement. Mahwah, NJ: Erlbaum.
Conard, M. A. (2006). Aptitude is not enough: How personality and
ment. To the extent that personality is more malleable in childhood
behavior predict academic performance. Journal of Research in Person-
and adolescence than in adulthood (e.g., Roberts & DelVecchio,
ality, 40, 339 –346.
2000), our findings indicate that it may be useful for educators to Costa, P. T., Jr., & McCrae, R. R. (1992). NEO PI-R and NEO-FFI
foster and facilitate optimal personality development in their stu- professional manual. Odessa, FL: Psychological Assessment Resources.
dents, in addition to teaching the standard curriculum. However, Costa, P. T., Jr., Terracciano, A., & McCrae, R. R. (2001). Gender
before investing in interventions to modify personality in the hope differences in personality traits across cultures: Robust and surprising
of promoting academic achievement or using personality tests as findings. Journal of Personality and Social Psychology, 81, 322–331.
predictors in a selection battery, we believe that further research is de Fruyt, F., & Mervielde, I. (1996). Personality and interests as predictors
needed to establish the causal direction of the effects, to clarify the of educational streaming and achievement. European Journal of Per-
mediating processes, and to better specify the particular facets sonality, 10, 405– 425.
involved. Demetriou, A., Kyriakides, L., & Avraamidou, C. (2003). The missing link
in the relations between intelligence and personality. Journal of Re-
search in Personality, 37, 547–581.
References De Raad, B., & Schouwenburg, H. C. (1996). Personality in learning and
Abelson, R. P. (1985). A variance explanation paradox: When a little is a education: A review. European Journal of Personality, 10, 303–336.
lot. Psychological Bulletin, 97, 129 –133. Digman, J. M. (1989). Five robust trait dimensions: Development, stability,
Ackerman, P. L., & Heggestad, E. D. (1997). Intelligence, personality, and and utility. Journal of Personality, 57, 195–214.
interests: Evidence for overlapping traits. Psychological Bulletin, 121, Digman, J. M., & Takemoto-Chock, N. K. (1981). Factors in the natural
219 –245. language of personality: Re-analysis, comparison, and interpretation of
Ahadi, S., & Diener, E. (1989). Multiple determinants and effect size. six major studies. Multivariate Behavioral Research, 16, 149 –170.
Journal of Personality and Social Psychology, 56, 398 – 406. Duckworth, A. L., & Seligman, M. E. P. (2005). Self-discipline outdoes IQ
Allport, G. W., & Odbert, H. S. (1936). Trait-names: A psycho-lexical in predicting academic performance of adolescents. Psychological Sci-
study. Psychological Monographs, 47(211). ence, 16, 939 –944.
Ashton, M. C. (1998). Personality and job performance: The importance of Duff, A., Boyle, E., Dunleavy, K., & Ferguson, J. (2004). The relationship
narrow traits. Journal of Organizational Behavior, 19, 289 –303. between personality, approach to learning and academic performance.
Ashton, M. C., Lee, K., Vernon, P. A., & Jang, K. L. (2000). Fluid Personality and Individual Differences, 36, 1907–1920.
intelligence, crystallized intelligence, and the openness/intellect factor. Elshout, J. J., & Akkerman, A. E. (1975). Vijf persoonlijkheids-faktoren
Journal of Research in Personality, 34, 198 –207. test 5PFT [Five Personality-Factor Test]. Nijmegen, The Netherlands:
Barchard, K. A. (2003). Does emotional intelligence assist in the prediction Berkhout Nijmegen.
of academic success? Educational and Psychological Measurement, 63, Eysenck, H. J. (1991, July). Award address for distinguished contribution
840 – 858. to the study of individual differences. Lecture presented at the 5th
Barrick, M. R., & Mount, M. K. (1991). The Big Five personality dimen- biennial meeting of the International Society for the Study of Individual
sions and job performance: A meta-analysis. Personnel Psychology, 44, Differences, Oxford, United Kingdom.
1–26. Farsides, T., & Woodfield, R. (2003). Individual differences and under-
Bischel, J., & Baker, T. (2006, August). Personality predictors of intelli- graduate academic success: The roles of personality, intelligence, and
gence: Differences between young and cognitively healthy older adults. application. Personality and Individual Differences, 34, 1225–1243.
Poster presented at the 114th annual meeting of the American Psycho- Frey, M. C., & Detterman, D. K. (2004). Scholastic assessment or g? The
logical Association, New Orleans, LA. relationship between the scholastic assessment test and general cognitive
Brodnick, R. J., & Ree, M. J. (1995). A structural model of academic ability. Psychological Science, 15, 373–378.
performance, socioeconomic status, and Spearman’s g. Educational and Furnham, A., & Chamorro-Premuzic, T. (2004). Personality and intelli-
Psychological Measurement, 55, 583–594. gence as predictors of statistics examination grades. Personality and
Busato, V. V., Prins, F. J., Elshout, J. J., & Hamaker, C. (2000). Intellectual Individual Differences, 37, 943–955.
ability, learning style, personality, achievement motivation and aca- Furnham, A., Chamorro-Premuzic, T., & McDougall, F. (2003). Personal-
demic success of psychology students in higher education. Personality ity, cognitive ability, and beliefs about intelligence as predictors of
and Individual Differences, 29, 1057–1068. academic performance. Learning and Individual Differences, 14, 49 – 66.
Goldberg, L. R. (1990). An alternative “description of personality”: The and abridging Shafer’s (1999a) Big Five markers. Personality and
Big-Five factor structure. Journal of Personality and Social Psychology, Individual Differences, 35, 1127–1140.
59, 1216 –1229. Lee, K., & Ashton, M. C. (2004). Psychometric properties of the HEXACO
Goldberg, L. R. (1992). The development of markers for the Big-Five Personality Inventory. Multivariate Behavioral Research, 39, 329 –358.
factor structure. Psychological Assessment, 4, 26 – 42. Lee, K., & Ashton, M. C. (2006). Further assessment of the HEXACO
Goldberg, L. R. (1999). A broad-bandwidth, public-domain, personality Personality Inventory: Two new facet scales and an observer report
inventory measuring the lower-level facets of several five-factor models. form. Psychological Assessment, 18, 182–191.
In I. Mervielde, I. J. Deary, F. de Fruyt, & F. Ostendorf (Eds.), Person- Letzring, T. D., Block, J., & Funder, D. C. (2005). Ego-control and
ality psychology in Europe (Vol. 7, pp. 7–28). Tilburg, the Netherlands: ego-resiliency: Generalization of self-report scales based on personality
Tilburg University Press. descriptions for acquaintances, clinicians, and the self. Journal of Re-
Goldberg, L. R., Johnson, J. A., Eber, H. W., Hogan, R., Ashton, M. C., search in Personality, 39, 395– 422.
Cloninger, C. R., & Gough, H. G. (2006). The international personality Lounsbury, J. W., & Gibson, L. W. (2004). Technical manual for the
item pool and the future of public-domain personality measures. Journal Resource Associates Personal Style Inventory and Adolescent Personal
of Research in Personality, 40, 84 –96. Style Inventory. Knoxville, TN: Resource Associates.
Gough, H. G. (1957). Manual for the California Personality Inventory. Lounsbury, J. W., Huffstetler, B. C., Leong, F. T., & Gibson, L. W. (2005).
Palo Alto, CA: Consulting Psychologists Press. Sense of identity and collegiate academic achievement. Journal of
Gough, H. G. (1964). Academic achievement in high school as predicted College Student Development, 46, 501–514.
from the California Psychological Inventory. Journal of Educational Lounsbury, J. W., Sundstrom, E., Loveland, J. M., & Gibson, L. W. (2003).
Psychology, 55, 174 –180. Intelligence, “Big Five” personality traits, and work drive as predictors
Gough, H. G., & Hall, W. B. (1964). Prediction of performance in medical of course grade. Personality and Individual Differences, 35, 1231–1239.
school from the California Psychological Inventory. Journal of Applied McCrae, R. R., & Costa, P. T., Jr. (1985). Updating Norman’s “adequacy
Psychology, 48, 218 –226. taxonomy”: Intelligence and personality dimensions in natural language
Gough, H. G., & Lanning, K. (1986). Predicting grades in college from the and in questionnaires. Journal of Personality and Social Psychology, 49,
California Psychological Inventory. Educational and Psychological 710 –721.
Measurement, 46, 205–213. McCrae, R. R., & Costa, P. T. Jr. (1997). Conceptions and correlates of
Gray, E. K., & Watson, D. (2002). General and specific traits of personality openness to experience. In R. Hogan, J. A. Johnson, & S. R. Briggs
and their relation to sleep and academic performance. Journal of Per- (Eds.), Handbook of personality psychology (pp. 825– 847). San Diego,
sonality, 70, 177–206. CA: Academic Press.
Graziano, W. G., & Ward, D. (1992). Probing the Big Five in adolescence: Mershon, B., & Gorsuch, R. L. (1988). Number of factors in personality
Personality and adjustment during a developmental transition. Journal of sphere: Does increase in factors increase predictability of real life
Personality, 60, 425– 439. criteria? Journal of Personality and Social Psychology, 55, 675– 680.
Hair, E. C., & Graziano, W. G. (2003). Self-esteem, personality and Meyer, G. J., Finn, S. E., Eyde, L. D., Kay, G. G., Moreland, K. L., Dies,
achievement in high school: A prospective longitudinal study in Texas. R. R., et al. (2001). Psychological testing and psychological assessment:
Journal of Personality, 71, 971–994. A review of evidence and issues. American Psychologist, 56, 128 –165.
Hair, P., & Hampson, S. E. (2006). The role of impulsivity in predicting Mischel, W., Shoda, Y., & Peake, P. K. (1988). The nature of adolescent
maladaptive behaviour among female students. Personality and Individ- competencies predicted by preschool delay of gratification. Journal of
ual Differences, 40, 943–952. Personality and Social Psychology, 54, 687– 696.
Harper, J. W. (2002). The new, improved SAT: Surprisingly, the revised Moutafi, J., Furnham, A., & Crump, J. (2003). Demographic and person-
college admissions test is better than the old one. The Weekly Standard, ality predictors of intelligence: A study using the NEO Personality
7, 47–53. Inventory and the Myers-Briggs Type Indicator. European Journal of
Hofstee, W. K., De Raad, B., & Goldberg, L. R. (1992). Integration of the Personality, 17, 79 –94.
Big Five and circumplex approaches to trait structure. Journal of Per- Moutafi, J., Furnham, A., & Crump, J. (2006). What facets of openness and
sonality and Social Psychology, 63, 146 –163. conscientiousness predict fluid intelligence score? Learning and Indi-
Hogan, R. (2005). In defense of personality measurement: New wine for vidual Differences, 16, 31– 42.
old whiners. Human Performance, 18, 331–341. Mouw, J. T., & Khanna, R. K. (1993). Prediction of academic success: A
Hogan, R., & Weiss, D. S. (1974). Personality correlates of superior review of the literature and some recommendations. College Student
academic achievement. Journal of Counseling Psychology, 21, 144 – Journal, 27, 328 –336.
149. Oswald, F. L., Schmitt, N., Kim, B. H., Ramsay, L. J., & Gillespie, M. A.
Hough, L. M. (1992). The “Big Five” personality variables– construct (2004). Developing a biodata measure and situational judgment inven-
confusion: Description versus prediction. Human Performance, 5, 139 – tory as predictors of college student performance. Journal of Applied
155. Psychology, 89, 187–207.
Jackson, D. N. (1984). Personality Research Form manual. Port Huron, Ozer, D. J., & Benet-Martı́nez, V. (2006). Personality and the prediction of
MI: Research Psychologists Press. consequential outcomes. Annual Review of Psychology, 57, 401– 421.
John, O. P., Caspi, A., Robins, R. W., Moffitt, T. E., & Stouthamer-Loeber, Paulhus, D. L., Lysy, D. C., & Yik, M. S. M. (1998). Self-report measures
M. (1994). The “little five”: Exploring the nomological network of the of intelligence: Are they useful as proxy IQ tests? Journal of Person-
five-factor model of personality in adolescent boys. Child Development, ality, 66, 525–554.
65, 160 –178. Paunonen, S. V. (1998). Hierarchical organization of personality and the
John, O. P., & Srivastava, S. (1999). The Big Five trait taxonomy: History, prediction of behavior. Journal of Personality and Social Psychology,
measurement, and theoretical perspectives. In L. A. Pervin & O. P. John 74, 538 –556.
(Eds.), Handbook of personality: Theory and research (2nd ed., pp. Paunonen, S. V., & Ashton, M. C. (2001). Big Five predictors of academic
102–138). New York: Guilford. achievement. Journal of Research in Personality, 35, 78 –90.
Langford, P. H. (2003). A one-minute measure of the Big Five? Evaluating Ridgell, S. D., & Lounsbury, J. W. (2004). Predicting academic success:
General intelligence, “Big Five” personality traits, and work drive. Saucier, G. (1994). Trapnell versus the lexical factor: More ado about
College Student Journal, 38, 607– 619. nothing? European Journal of Personality, 8, 291–298.
Roberts, B. W., & Caspi, A. (2003). The cumulative continuity model of Saucier, G., & Goldberg, L. R. (1996). The language of personality:
personality development: Striking a balance between continuity and Lexical perspectives on the five-factor model. In J. S. Wiggins (Ed.), The
change in personality traits across the life course. In U. M. Staudinger & five-factor model of personality: Theoretical perspectives (pp. 21–50).
U. Lindenberger (Eds.), Understanding human development: Dialogues New York: Guilford.
with lifespan psychology (pp. 183–214). Dordrecht, The Netherlands: Schuerger, J. M., Kepner, J., & Lawler, B. (1979). Verbal-quantitative
Kluwer Academic. differential as indicator of temperamental differences. Multivariate Ex-
Roberts, B. W., Chernyshenko, O. S., Stark, S., & Goldberg, L. R. (2005). The perimental Clinical Research, 4, 57– 66.
structure of conscientiousness: An empirical investigation based on seven Shafer, A. B. (1999). Brief bipolar markers for the five factor model of
major personality questionnaires. Personnel Psychology, 58, 103–139. personality. Psychological Reports, 84, 1173–1179.
Roberts, B. W., & DelVecchio, W. F. (2000). The rank-order consistency Shoda, Y., Mischel, W., & Peake, P. K. (1990). Predicting adolescent
of personality traits from childhood to old age: A quantitative review of cognitive and self-regulatory competencies from preschool delay of
longitudinal studies. Psychological Bulletin, 126, 3–25. gratification: Identifying diagnostic conditions. Developmental Psychol-
Roberts, B. W., & Hogan, R. (Eds.). (2001). Personality psychology in the ogy, 26, 978 –986.
workplace: Decade of behavior. Washington, DC: American Psycho- Stricker, L. J., Rock, D. A., & Burton, N. W. (1991). Sex differences in SAT
logical Association. predictions of college grades (College Board Rep. No. 91–92). New
Roberts, B. W., Kuncel, N. R., Shiner, R., Caspi, A., & Goldberg, L. R. (in York: College Entrance Examination Board.
press). The comparative predictive validity of personality traits, SES, Tangney, J. P., Baumeister, R. F., & Boone, A. L. (2004). High self-control
and cognitive ability. Perspectives in Psychological Science. predicts good adjustment, less pathology, better grades, and interper-
Robins, R. W., Fraley, R. C., Roberts, B. W., & Trzesniewski, K. H. sonal success. Journal of Personality, 72, 271–322.
(2001). A longitudinal study of personality change in young adulthood. Tross, S. A., Harper, J. P., Osher, L. W., & Kneidinger, L. M. (2000). Not
Journal of Personality, 69, 617– 640. just the usual cast of characteristics: Using personality to predict college
Robins, R. W., Noftle, E. E., Trzesniewski, K. H., & Roberts, B. W. performance and retention. Journal of College Student Development, 41,
(2005). Do people know how their personality has changed? Correlates 323–334.
of perceived and actual personality change in young adulthood. Journal Wagerman, S. A., & Funder, D. C. (2007). Acquaintance reports of
of Personality, 73, 489 –521. personality and academic achievement: A case for conscientiousness.
Rohde, T. E., & Thompson, L. A. (2007). Predicting academic achieve- Journal of Research in Personality, 41, 221–229.
ment with cognitive ability. Intelligence, 35, 83–92. Wolfe, R. N., & Johnson, S. D. (1995). Personality as a predictor of college
Rosenthal, R., & Rubin, D. B. (1982). A simple, general purpose display of performance. Educational and Psychological Measurement, 55, 177–185.
magnitude of experimental effect. Journal of Educational Psychology,
74, 166 –169. Received September 19, 2006
Saucier, G. (1992). Openness versus intellect: Much ado about nothing? Revision received February 15, 2007
European Journal of Personality, 6, 381–386. Accepted February 19, 2007 䡲

Personality Predictors of Academic Outcomes: Big Five Correlates of GPA and SAT Scores

Uploaded by

Copyright:

Available Formats

Personality Predictors of Academic Outcomes: Big Five Correlates of GPA and SAT Scores

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Personality Predictors of Academic Outcomes: Big Five Correlates of GPA and SAT Scores

Uploaded by

Copyright:

Available Formats

PERSONALITY PROCESSES AND INDIVIDUAL DIFFERENCES

Personality Predictors of Academic Outcomes: Big Five Correlates of GPA

Keywords: Conscientiousness, Openness to Experience, achievement, SAT scores, GPA

Barchard (2003) GPA 150 IPIP NEO-PI 0 0 ⫹⫹⫹ 0 ⫹

Mean effect size 5,292 ⫺.04 .09 .26 ⫺.07 .05

To help readers better understand the overall pattern of relations

Extraversion 3.26 0.75 3.47 0.57 3.21 0.56 — —

SAT verbal SAT math

Sample 1 Sample 2 Sample 3 Sample 4 Sample 1 Sample 2 Sample 3 Sample 4

Multiple R .22* .27* .30* .26* .17* .10 .11 .08

College GPA High school GPA

Sample 1 Sample 2 Sample 3 Sample 4 Sample 1 Sample 2 Sample 3 Sample 4

Multiple R .25* .25* .28* .21* .24* .17 .26* .25*

College GPA High school GPA

Sample 1 Sample 2 Sample 3 Sample 1 Sample 2 Sample 3

Multiple R .31* .49* .38* .22* .63* .19*

Multiple R .39* .54* .45* .31* .64* .32*

SAT verbal SAT math

Sample 1 Sample 2 Sample 3 Sample 1 Sample 2 Sample 3

Multiple R .17* .47* .16* .26* .64* .24*

Multiple R .27* .54* .34* .29* .65* .25*

You might also like