individual differences and undergraduate academic success,the roles of personality, intelligence,...

www.elsevier.com/locate/paid

Individual differences and undergraduate academic success:the roles of personality, intelligence, and application

Tom Farsidesa,*, Ruth Woodfieldb

aSocial Psychology Group, School of Social Sciences, University of Sussex, East Sussex BN1 9SN, UKbSociology Group, School of Social Sciences, University of Sussex, East Sussex BN1 9SN, UK

Received 21 September 2001; received in revised form 15 March 2002; accepted 15 April 2002

Abstract

The roles of intelligence and motivation in predicting academic success are well established. Evidence is,however, mixed concerning the role of personality traits in predicting such success. The current study attemp-ted to overcome various methodological limitations associated with many previous studies to examine thepotency of the traits of the ‘five factor model of personality’ in predicting academic success up to 3 years later,both directly and when controlling for intelligence and ‘application’ (used as a proxy for motivation). Only twotraits yielded significant zero-order correlations with eventual undergraduate success, with both Openness toexperience and Agreeableness being positively associated with Final Grades. Openness to experienceexplained unique variance in Final Grades even when predicting in the company of intellect and applica-tion measures. The impact of Agreeableness on Final Grades was wholly mediated by the main applicationmeasure; namely, not missing seminars. Less than one fifth of Final Grade variance was explained by allthe individual difference variables in combination. Several practical, theoretical, and future researchimplications are explored. # 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Personality; Intelligence; Application; Motivation; Academic success

1. Introduction

Reviewing the educational research available at the time, Harris (1940) claimed that the mostessential determinants of academic success were intelligence and motivation. As Busato, Prins,Elshout, and Hamaker (2000) note, few today would disagree about the continued importance ofsuch factors. In the American Psychological Association ‘Task Force’ review of what is and whatis not known about intelligence, Neisser et al. (1996) are unanimous in accepting that intelligencetest scores predict a wide range of indicators of academic success (cf. Ackerman & Heggestad, 1997;

0191-8869/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.

PI I : S0191-8869(02 )00111-3

Personality and Individual Differences 34 (2003) 1225–1243

* Corresponding author. Tel.: +44-1273-678-886; fax: +44-1273-673-563.

E-mail address: [email protected] (T. Farsides).

http://www.elsevier.com/locate/paid/a4.3d

mailto:[email protected]

Busato et al., 2000; Hirschberg & Itkin, 1978; Mellanby, Martin, & O’Doherty, 2000). Similarly,positive relationships between motivation and academic success seem almost beyond doubt (Pintrich& Schunk, 1996: cf. Busato et al., 2000; Furnham&Mitchell, 1991; Mellanby et al., 2000). However,whilst accepting the roles of intelligence and motivation in determining academic success, con-temporary researchers are also interested in whether or not other individual differences may be usedto predict academic performance. Personality dispositions are one class of individual differencescurrently enjoying a considerable amount of attention (Paunonen & Ashton, 2001).There is growing acceptance that many of the most important personality dispositions may be

considered as collectively comprising the ‘Big Five’ traits of the Five Factor Model (FFM) ofpersonality. Currently popular labels for these traits enable the acronym ‘OCEAN’: Openness toexperience, Conscientiousness, Extraversion-introversion, Agreeableness, and Neuroticism-Emo-tional stability. Empirical evidence is mixed concerning the role each of these traits play indetermining academic success. One reason for this appears to be that the relationship betweencertain traits and academic success is age-specific (see De Raad & Schouwenburg, 1996; Furnham& Mitchell, 1991). In reviewing recent investigations of these relationships, therefore, it is neces-sary to order findings by both trait and age/educational level. (Relevant studies conducted before1970 will not be reviewed as this has already been done capably elsewhere: see De Raad &Schouwenburg, 1996; Entwistle, 1972; Kline & Gale, 1971. In addition, as this paper focuses ondispositional predictors of academic success, readers interested in relationships between the pre-dictor variables used are referred to Ackerman & Heggestad, 1997; Collis & Messick, 2001; Goff& Ackerman, 1992; Matthews, 1997; Rolfhus & Ackerman, 1999; Saklofske & Zeidner, 1995;Sternberg & Ruzgis, 1994; Stough et al., 1996; Zeidner & Matthews, 2000.)

1.1. Openness to experience

Positive correlations between openness to experience and academic success have been found bySchuerger and Kuna (1987) among school children, De Fruyt and Mervielde (1996) among firstyear undergraduates, and by Rothstein, Paunonen, Rush, and King (1994) among one of twosamples of Business School graduate students (using graduate GPA as the criterion variable). Ack-erman and Heggestad’s meta-analysis (1997) also reveals a positive relationship across two studiesbetween openness to experience and ‘‘knowledge and achievement.’’ Against this, no significantcorrelation between openness to experience and academic success was found amongRothstein et al.’s(1994) second sample of graduate students, nor among Wolfe and Johnson’s (1995) college studentsample, nor among Busato et al.’s (2000) sample of first year psychology undergraduates.

1.2. Conscientiousness

De Raad and Schouwenburg (1996) suggest that ‘‘there is an impressive list emphasizing theimportance of conscientiousness or related factors in learning and education’’ (p. 325), withconscientiousness and learning outcome variables showing ‘‘substantial zero-order correlations’’(p. 327). Consistent with this suggestion, positive correlations between conscientiousness andacademic success have been found at the school level by Heaven, Mak, Barry, and Ciarrochi(2002); at the college level by Wolfe and Johnson (1995); at the university level by Busato et al.(2000), De Fruyt and Mervielde (1996), and Goff and Ackerman (1992); and at the graduate level

1226 T. Farsides, R. Woodfield / Personality and Individual Differences 34 (2003) 1225–1243

by Hirschberg and Itkin (1978), Wiggins et al. (1969) and—for one of two samples using GPA asthe criterion variable—Rothstein et al. (1994). Against this, Goff and Ackerman (1992) found nosignificant correlation between conscientiousness and academic success at the high school leveland Rothstein et al. (1994) failed to find a significant correlation for the second of their graduatestudent samples when using GPA as the criterion variable. Ackerman and Heggestad’s (1997)meta-analysis includes one study showing a significant negative correlation between con-scientiousness and ‘‘knowledge and achievement’’ (p. 231).

1.3. Extraversion

De Barbenza and Montoya (1974, cited in Halamandaris & Power, 1999) found that, acade-mically, extraverted university students slightly outperformed introverted ones. Halamandarisand Power (1999) on the other hand found no significant correlation between extraversion andundergraduate academic success. Similarly, Heaven et al. (2002) found no significant relationshipbetween extraversion and academic performance at school. With a single exception (on a secondyear clinical practice assessment), Furnham and Mitchell (1991) similarly found no significantcorrelations between extraversion and any of a wide range of measures of academic success (overfour years) among their sample of occupational therapy students. Ackerman and Heggestad’s(1997) meta-analysis included seven relevant studies which again revealed no significant relation-ship between extraversion and ‘‘knowledge and achievement.’’ Mixed evidence about the rela-tionship between extraversion and undergraduate success has been found by other studies (cf.Kline & Gale, 1971; Roberts, 2002). In each of these, some criterion variables have revealed nosignificant correlations between extraversion and academic success, while others have suggestedeither a positive (De Fruyt & Mervielde, 1996) or a negative (Busato et al., 2000) association.Similarly, Rothstein et al. (1994) found mixed evidence about the relationship between extraver-sion and academic success at the graduate level, with some criterion variables suggesting a posi-tive relationship and others suggesting no such relationship. Goff and Ackerman (1992) foundextraversion to negatively correlate with both high school and undergraduate GPA.

1.4. Agreeableness

Heaven et al. (2002) found a positive zero-order correlation between agreeableness and self-reported academic performance among school children. De Fruyt and Mervielde (1996) found noassociation between agreeableness and either final grades or performance during the first exami-nation period of the final year among undergraduates. Busato et al. (2000) found no significantcorrelation between either agreeableness and the first examination scores of first year psychologyundergraduates or between agreeableness and any of the ‘study points’ these students earned foreach undergraduate year of study. Similarly, Rothstein et al. (1994) found no association betweenagreeableness and the written component of GPA among either of their two samples of BusinessSchool graduates (even when considered in combination). The single relevant study included inAckerman and Heggestad’s (1997) meta-analysis also found no significant association betweenagreeableness and ‘‘knowledge and achievement.’’ Finally, Rothstein et al. (1994) found agree-ableness to negatively correlate with both in-class performance and with overall GPA among theirtwo samples of Business School Graduates.

T. Farsides, R. Woodfield / Personality and Individual Differences 34 (2003) 1225–1243 1227

1.5. Neuroticism

De Raad and Schouwenberg (1996) conclude their review of the relationship between emotionalstability (the polar opposite of neuroticism) and academic achievement by saying that ‘‘particu-larly at the university level, highly neurotic students are probably handicapped as compared tolow neurotics’’ (p. 326). Ackerman and Heggestad (1997) consider neuroticism to be equivalent towhat they call ‘‘stress reaction.’’ Consistent with De Raad and Schouwenberg’s conclusion,Ackerman and Heggestad’s (1997) meta-analysis of 11 relevant studies revealed a significantnegative relationship between ‘‘stress reaction’’ and ‘‘knowledge and achievement.’’ Similarly, DeBarbenza and Montoya (1974, cited in Halamandaris & Power, 1999) found a negative corre-lation between neuroticism and academic achievement specifically among university students.However, Busato et al. (2000) and Halamandaris and Power (1999) find no such associationamong their undergraduate samples, while Heaven et al. (2002) find no such association amongschool children. De Fruyt and Mervielde (1996) and Furnham and Mitchell (1991) each foundmixed evidence about the relationship between neuroticism and undergraduate academic success,finding significant negative correlations using some criterion variables and no such relationshipsusing others (cf. Kline & Gale, 1971). Similarly mixed results were obtained at the graduate levelby Rothstein et al. (1994).

1.6. Rationale for the present study

Despite carefully ordering the review above by age/educational level, as well as by trait, it isobvious that the relevant literature paints no clear picture about the relationships between the keypersonality dispositions and academic success. There are several possible reasons for this, at leastsome of which may already be apparent from the review so far.First, in addition to employing samples of different ages, different studies have drawn samples

from different disciplines, e.g. Occupational Therapy and Business Studies in addition to themore common Psychology. Research evidence suggests that trait-performance links may differacross subjects (De Fruyt & Mervielde, 1996).Secondly, different studies have used potentially radically different criteria for academic success,

from course-specific evaluations (Halamandaris & Power, 1999), to first-year examination scores(e.g. Busato et al., 2000), to final year examination scores (e.g. De Fruyt & Mervielde, 1996), toin-class performance (e.g. Rothstein et al., 1994), to grade point average (e.g. Goff & Ackerman,1992), to assessment whilst on ‘placements’ (e.g. Furnham & Mitchell, 1991). It seems entirelypossible that certain traits predict academic success in certain domains but not in others. To givejust one example, extraversion may be positively correlated to academic success in terms of con-tributions in seminars but zero or even negatively correlated to performance in certain types ofexamination (cf. Rothstein et al., 1994).Thirdly, different studies have permitted considerably different time lapses between collection of

predictor and criterion data. At one extreme, final degree results have been predicted from measuresobtained up to 4 years previously (e.g. Furnham &Mitchell, 1991). At the other extreme, secondaryschool GPA has been ‘postdicted’ from measures obtained whilst participants were at college(Blickle, 1996). For a variety of reasons, ‘‘it may be that personality variables predict academicperformance in the short term but not the long term’’ (Furnham & Mitchell, 1991, p. 1068).


Relationships between individual differences and academic performance may be affected bymethodological considerations beyond temporality of measurement. At least some of the relevantpapers in the field report sample sizes smaller than their authors would have liked (e.g. Hala-mandaris & Power, 1999, despite having a relatively large number of individual difference mea-sures, had academic performance data for only 81 students). In addition, several researchers (e.g.Busato et al., 2000) have noted the ‘restriction of range’ problem, whereby individual differencescores are likely to become less variable at later stages in the academic process (but indicators ofacademic success within stages are not). Intelligence Quotient scores, for example, are likely to bemuch more variable among (unselected) primary school children than among university students(selected, among other things, because of their indicators of high intelligence).Despite the difficulties of reaching clear conclusions about possible relationships between indi-

vidual differences and academic performance, the use of individual difference measures for aca-demic selection continues to be seriously considered by, for example, university administrators(e.g. Wolchover, 2000). For this reason, the current study was designed to investigate the role ofindividual differences (and other variables) in predicting academic success specifically amongstudents at our university, the University of Sussex, England.A number of steps were taken to ensure augmentation of the existing literature. First, and most

obviously, we ensured a reasonably large sample size and employment of psychometrically defensiblemeasures. Secondly, we attempted to predict the most important measure of academic achievement(the percentage from which final degree classification was derived) from measures including disposi-tional ones completed at the very start of students’ undergraduate careers. Nevertheless, we supple-mented this by obtaining more specific measures of academic success from each year of study.Finally, rather than simply replicating one of the more clear and robust findings in the relevant lit-erature, namely the positive association between trait motivation (e.g. Cassidy & Lynn, 1989;McClelland, Atkinson, Clark, & Lowell, 1953) and academic success, we investigated the possiblerelationship between academic success and important practical manifestations of motivation whichwe term ‘application,’ e.g., seminar attendance and completion of set work the marks for whichdid not contribute to degree classification (i.e. work not ‘‘formally assessed’’).As the review above attempts to show, hypotheses about the relationships between personality

traits and academic success must of necessity be tentative. Significant relationships between per-sonality variables and academic performance in previous studies have been both erratic and,where present, modest. De Fruyt and Mervielde (1996), for example, investigated correlationsbetween each of the five ‘OCEAN’ traits and each of three separate educational outcome vari-ables. The largest significant correlation was 0.35 (cf. Ackerman & Heggestad, 1997; Busato et al.,2000; Goff & Ackerman, 1992; Rothstein et al., 1994). Similarly, Paunonen and Ashton (2001)found a multiple correlation of all five ‘OCEAN’ traits and university GPA of only 0.33 (therebyleaving almost 90% of the GPA variance unexplained).Bearing such concerns in mind, our best interpretation of the relevant literature as it bears upon

our undergraduate sample nevertheless lead us to hypothesise that there would be modest positivecorrelations between both Openness to experience (H1) and Conscientiousness (H2) and aca-demic success. We expected no significant correlations between academic success and eitherExtraversion (H3) or Agreeableness (H4). A modest negative relationship between Neuroticismand academic success was anticipated (H5). In addition, we predicted that ‘application’ whilst atuniversity (e.g. seminar attendance) would be positively correlated with academic success (H6).


Finally, we predicted that personality measures would be able to account for significant, if mod-est, variation in academic success scores, even once the effects of intelligence and ‘application’were accounted for (H7).

2. Method

2.1. Context

It is important to explain that at the time the data was collected for this study, most undergraduatestudents at Sussex University took both ‘Major’ and ‘School’ courses. To give an example, studentsstudying the Major subject of Geography would do so within one of several possible Schools, e.g.Social Sciences, African and Asian Studies, Culture and Community Studies. The modal degreestructure at the time was such that approximately half of students’ courses would have been withintheir Major subject, with the remainder being (typically multi- or inter-disciplinary) School courses.

2.2. Participants

Participants were 432 University of Sussex students who successfully completed their 3-yearundergraduate degrees in the summer of 2000. Of these, 205 were male, 226 were female, and onedid not specify their gender. Mean age at time of entry was 21.30 years old (S.D.=6.29). Parti-cipants read for a wide variety (N=48) of degree subjects within nine different schools of study.

2.3. Procedure

Self-completion questionnaire packs were administered in variety of mass testing sessions dur-ing students’ first week at university. Each questionnaire pack took approximately 1 hour tocomplete. Permission to collect data was obtained from all appropriate university authorities andindividual participation was entirely voluntary. The nature of the data to be collected wasexplained to participants and they were advised that their progress throughout their time at uni-versity would be tracked in terms of their seminar attendance, grades, etc. Only a handful ofstudents declined to take part by leaving testing sessions when given an explicit opportunity to doso. No incentive for participation was offered beyond the possibility of obtaining personal resultson the dispositional measures completed.

2.4. Measures

2.4.1. Personality traitsThe five traits of Costa and McCrae’s (1992) Five Factor Model (FFM) of personality were

measured using the short form of the NEO Five-Factor Inventory (NEO-FFI; Costa & McCrae,1989). This Inventory uses 12 items to measure each of the following traits: Neuroticism, Extra-version, Openness to experience, Agreeableness and Conscientiousness. Each response scale has arange of 0 (‘strongly disagree’) to 4 (‘strongly agree’).

2.4.2. IntelligenceIntelligence was measured using the AH5 Group Test of High Intelligence (Heim, 1968). To

overcome restriction of range problems associated with the use of standard intelligence measures


with samples such as those drawn from university students, the AH5 test was designed specificallyfor high intelligence participants. The test comprises of two sub-scales, one measuring pre-dominantly verbal intelligence and the other predominantly spatial (geometric) intelligence. In thecurrent sample, the correlation between verbal and spatial intelligence scores was moderate butnot sufficiently strong to warrant collapsing into a single ‘Intelligence’ scale (r=.45, P<0.001).

2.4.3. Demographic variablesWith permission from both university authorities and participants, the researchers had access to

participating students’ administrative records. It was therefore possible to obtain a considerableamount of demographic information. Much of this will not be reported here, primarily eitherbecause of suspicions about the data’s quality (e.g. self-reported ethnicity on UCAS forms, a fullthird of which were left blank) or because information consisted of categorical data unsuited to theanalytic strategy employed below. However, in addition to obtaining information about age uponentry, data was gathered on the ‘A level’ points (or equivalent) students entered university withand on students’ socio-economic status (SES, measured using the Registrar General’s classificationof descending SES from 1 to 5).

2.4.4. Tutorial report gradesEach term, tutors for each undergraduate course at Sussex University write ‘tutorial reports’ for

each of their students. In addition to providing other information, tutors give each student agrade on a scale from 1 (equivalent to a 1st classification) to 6 (equivalent to a ‘fail’ classification).These grades are intended primarily to be diagnostic and predictive of final degree classification.Other than in very exceptional circumstances, they play no role in formal assessment.Tutorial grades for all participating students were obtained for all the Autumn and Spring

courses they completed in Years 1 and 2. Data were restricted to these terms and years because ofthe difficulty of data collection in the Summer term (largely given over to examining) and in thethird year (when fewer ‘didactic’ courses occur).

2.4.5. Tutorial absencesTutorial reports are also used to register student absences from seminars. This data was also

obtained for all participating students in all their Year 1 and Year 2 Autumn and Spring courses.Because absolute number of absences is partly a reflection on the number of seminars given on acourse, absence data were converted to a ratio reflecting the proportion of time-tabled seminarsattended each term (across all courses), thus having a possible range of 0 (no absences) to 1(absent from all seminars).

2.4.6. Submission of non-assessed workRecords were obtained of participants’ submission of non-assessed work. In this context, non-

assessed work is work which is set and marked for diagnostic and formative purposes, but forwhich the marks obtained do not contribute to formal final assessment. This measure was derivedas a ratio of work set to work submitted each term (across all courses), thus having a possiblerange of 0 (no non-assessed work submitted) to 1 (all non-assessed work submitted). For com-parability with the previous two measures, submission data was collected in the Autumn andSpring terms of Years 1 and 2.


2.4.7. Academic successAt the time data was collected for this study, University of Sussex students were formally assessed

each year, separately for both Major and School courses. In all cases, assessments were recorded aspercentages. Year assessments were then calculated (where appropriate) by averaging the percen-tages achieved in the Major and School yearly assessments, usually weighted equally or approxi-mately equally. Final assessment was then calculated by averaging Year 2 and Year 3 assessments.Degree classification (for all subjects) was achieved by formulae approximately designating ‘firsts’ tostudents obtaining final assessment percentages of above 70%, 2.1s to students obtaining finalassessment percentages between 60 and 69%, 2.2s to students with final percentages of 50–59%, 3rdsto students with final percentages of 40–49%, pass degrees to students with final percentagesbetween 30 and 39%, and fail degrees to students with final percentages below 30%. It can be notedthat some academics at British Universities are suspicious that there may be widespread violations ofthe assumptions of ‘parametric data’ when percentages are to be awarded near degree classificationborderlines. Thus, for example, it is sometimes claimed that the difference between grades of 66%and 69% is rather more significant than the difference between grades of 62% and 65%, as the for-mer is closer than the latter to boundaries used for degree classification purposes (cf. Mellanby et al.,2000, p. 382). However, because the Sussex assessment system at the time of this study was totake averages over multiple courses, procedures were in place to regularly request that markers didnot make assumptions such as these but rather assigned percentage grades without ‘adjustment’.In addition to the final percentage assessments used to designate degree classification, more

specific academic success measures were obtained from percentage assessments for both Majorand School for each year of study.

3. Results

3.1. Data reduction

3.1.1. Tutorial report gradesPrincipal components analysis of the eight tutorial report grades obtained yielded a single factor

solution, with the resultant scale having a Cronbach’s alpha of 0.83.

3.1.2. Seminar absencesPrincipal components analysis of the eight seminar absence measures yielded a single factor

solution, with the resultant scale having a Cronbach’s alpha of 0.85.

3.1.3. Submission of non-assessed workThe four submission ratio measures had low intercorrelations and did not yield a stable factor

structure. Thus, the submission ratio measures were retained as separate measures, i.e. relating tothe Autumn and Spring courses in the first 2 years of study.

3.2. Analytic strategy

In common with other studies of this type, data analysis was conducted in two stages. First,correlations were run between all the variables of interest. Secondly, multiple regression analyses


sought to identify how well academic success could be predicted using a combination of theindividual difference measures at our disposal.

3.3. Correlational analyses

3.3.1. Correlations among key dispositional variablesTable 1 shows means and standard deviations for, and the correlations between, each of the

individual difference variables potentially predicting academic success. Most of the correlationsare more or less as expected, although there were a few notable exceptions.Openness to experience was positively correlated with Extraversion, Agreeableness, and verbal IQ.Conscientiousness was also positively correlated with Extraversion and Agreeableness, was not

significantly correlated with either form of intelligence, and was negatively correlated with Neu-roticism. Conscientiousness was significantly negatively related with seminar absences (but not,interestingly, with any of the indicators of submission of non-assessed work).As just noted, Extraversion was significantly associated with both Openness to experience and

Conscientiousness, and it was also significantly positively associated with Agreeableness andnegatively associated with Neuroticism.In addition to positively correlating with Openness to experience, Conscientiousness, and

Extraversion, Agreeableness was the only trait other than Openness to experience to correlate posi-tively with verbal intelligence. Agreeableness was also the only trait other than Conscientiousness tocorrelate negatively with seminar absences.

Table 1

Means, standard deviations and correlations between individual difference variables potentially predicting academicsuccess

Mean
S.D. O C E A N IQv IQs Absent Ratio 1 Ratio 2 Ratio 3
O
30.51 6.08
C
29.64 6.89 �0.07
E
29.40 5.61 0.18 0.18
A
30.34 6.02 0.22 0.13 0.25
N
23.74 8.34 0.08 �0.19 �0.28 �0.11
Iqv
10.01 4.26 0.16 �0.07 �0.00 0.17 �0.05
Iqs
16.12 5.32 0.01 �0.11 0.00 0.02 �0.09 0.45
Absent
0.12 0.10 �0.02 �0.16 0.07 �0.14 �0.13 �0.04 0.07
Ratio 1
0.96 0.12 �0.03 0.13 �0.11 �0.06 0.04 �0.02 �0.03 �0.31
Ratio 2
0.96 0.12 0.00 0.05 �0.03 �0.01 0.12 �0.12 �0.01 �0.35 0.25
Ratio 3
0.94 0.14 �0.12 0.09 0.00 0.10 0.00 �0.12 �0.13 �0.38 0.11 0.22
Ratio 4
0.93 0.16 �0.04 0.04 �0.14 �0.02 0.05 �0.02 0.06 �0.30 0.15 0.13 0.14
Italic values are significant at P (two-tailed) 40.01 (strict probability adopted to reduce occurrence of Type I Errors: cf.

Rothstein et al. (1994). O=Openness to experience, C=Conscientiousness, E=Extraversion, A=Agreeableness, N=Neuro-

ticism, IQv=Verbal AH5 score, IQs=Spatial AH5, Absent=Seminar absences, Ratio 1–Ratio 4=Ratio of non-assessed

work submitted for Year 1 Autumn term, Year 1 Spring term, Year 2 Autumn term, and Year 2 Spring term, respectively. All

the correlations in this table were repeated using Spearman rho statistics. All the significances shown in this table were repli-

cated, with four exceptions. Three of these exceptions were minor, with significance levels being at <0.05 instead of at40.01

(Extraversion with ‘Expected Degree’ and Neuroticism with ‘Intellectual Motive’) or at 40.01 instead of <0.05 (Agreeable-

ness with Neuroticism). The one remaining exception is that using Spearman’s rho, there was no significant relationship found

(even at P<0.05) between Agreeableness and ‘Absences’.


As noted, Neuroticism was negatively associated with both Conscientiousness andExtraversion.In addition to positively correlating with Openness to experience and Agreeableness, verbal

intelligence also correlated significantly with spatial intelligence.In addition to being negatively correlated with both Conscientiousness and Agreeableness,

seminar absenteeism was significantly associated with all four measures indicating failure to submitnon-assessed assignments.Perhaps the most striking of these results are those associated with Agreeableness. Of the five

traits measured, this was the only one to be positively associated with the verbal IQ measure andnegatively associated with seminar absenteeism. If one had to pick a single proxy for a combi-nation of intelligence and motivation in the current study, therefore, it would probably beAgreeableness.

3.3.2. Correlations among indications of academic successIn line with previous research (Osbourne, Leopold, & Ferrie, 1997; Saunders & Woodfield,

1999), Table 2 reveals A’ level points (or their equivalent) to be rather strongly associated withacademic success at Sussex University. In particular, the correlation between A’ level points andFinal Grade (in percentage) at university is r=0.32 (P<0.001). Nevertheless, the correlationsbetween A’ level points and more specific indicators of academic success at university are far fromconstant. Thus, despite their predictive success elsewhere (and in a sense ‘‘where it mattersmost’’), A’ level points are not significantly associated with Tutorial Report scores or with SchoolCourse assessment in Years 1 and 2.Tutorial Report scores are consistently and rather strongly correlated with all the other indicators

of undergraduate academic success.

Table 2Means, standard deviations and correlations between academic success variables

Mean
S.D. A’
Points

Tut

Grad

Maj

Yr 1

Sch

Yr 1

Maj

Yr 2

Sch

Yr 2

Maj

Yr 3

Sch

Yr 3

A’ Points
21.00 6.79
Tut Grad
2.29 0.62 �0.09
Maj Yr 1
58.70 11.86 0.17 �0.48
Sch Yr 1
58.33 13.18 0.03 �0.28 0.30
Maj Yr 2
57.98 9.56 0.26 �0.47 0.58 0.28
Sch Yr 2
60.14 11.38 0.15 �0.46 0.28 0.24 0.41
Maj Yr 3
58.04 9.86 0.22 �0.37 0.48 0.25 0.72 0.34
Sch Yr 3
60.69 5.24 0.25 �0.36 0.42 0.17 0.49 0.65 0.62
Final
58.49 8.82 0.32 �0.46 0.54 0.33 0.80 0.45 0.96 0.81
Italic values are significant at P (two-tailed)40.01 (strict probability adopted to reduce occurrence of Type I Errors). A’

Points=‘A level Points’ (or equivalent); Tut Grad=Tutorial Grades; Maj Yr 1–Maj Yr 3=Academic success on Major

component, Years 1–3, respectively; Sch 1–Sch 3=Academic success on School component, Years 1–3, respectively; Final=

Academic success as determined by final formal assessment percentage grade. Although the magnitude of the correlation

coefficients shown in this Table tended to be higher when calculated using Spearman’s rho, all but one of the significances and

non-significances shown in this table were replicated. The exception was that using Spearman’s rho, the correlation between

‘S10 and ‘S30 was significant at P40.01.


Strong correlations between the remaining measures of academic success are in large part to beexpected, especially between specific Year 2 and Year 3 grades and ‘Final Grade’, as the lattermeasure is derived from a combination of the former measures. Nevertheless, the pattern ofcorrelations remains interesting. First, specific Major grades tend to correlate more strongly withFinal Grade than do specific School grades, particularly in the first two years. Secondly, the Year1 Major grade is more strongly associated with subsequent Major grades than it is with sub-sequent School grades, while the Year 1 School grade correlates relatively moderately with allsubsequent grades, Major and School. Similarly, the Year 2 Major grade correlates quite stronglywith the subsequent Major and Final grades but not the Year 3 School grade, while the Year 2School grade correlates much more strongly with the Year 3 School grade than with any of theother subsequent grades. In summary, these patterns of results suggest that Year 1 and Year 2School grades may be rather distinct from both Major grades and/or Year 3 grades (and indeed,A’ level points). However, factor analysis of the six specific Year 1 to Year 3 grades resulted in asingle factor solution (accounting for nearly 53% of the original variance), albeit with the Year 1School course loading onto the resultant factor to a lesser extent than the other input variables(0.45 versus a minimum of 0.65).Thus, it would appear that our criterion variables are in fact all indicating academic success,

although the Year 2 and especially Year 1 School grades may be measuring relatively specificfacets of this overall construct.

3.3.3. Correlations between individual difference and academic success variablesTable 3 shows correlations between each of the individual difference measures and each of the

academic success variables. Straightaway, it can be seen that, supporting hypothesis H3 but

Table 3Correlations between individual difference variables and academic success variables

A’

Points

Tutor

Grade

Major

Yr 1

School

Yr 1

Major

Yr 2

School

Yr 2

Major

Yr 3

School

Yr 3

Final

% age

Openness
0.24 �0.07 0.11 0.14 0.17 0.24 0.24 0.09 0.26
Conscien
�0.07 �0.22 0.17 0.06 0.06 0.01 0.08 0.05 0.09
Extravert
0.12 �0.00 �0.04 0.04 0.00 0.02 0.04 �0.02 0.00
Agreeable
0.09 �0.11 0.04 0.05 0.11 0.14 0.15 0.15 0.14
Neurotic
0.09 �0.07 0.03 0.04 �0.00 0.07 0.02 0.07 0.03
IQ—Verbal
0.17 0.01 0.02 0.14 0.17 0.02 0.24 0.23 0.20
IQ—Spatial
0.07 0.04 0.06 0.06 0.09 0.04 0.05 0.10 0.11
Absences
0.02 0.59 �0.29 �0.23 �0.31 �0.35 �0.34 �0.24 �0.36
Ratio 1
�0.10 �0.33 0.13 0.07 0.04 0.14 0.09 0.06 0.09
Ratio 2
�0.10 �0.32 0.20 0.16 0.11 0.14 0.20 0.04 0.21
Ratio 3
�0.05 �0.29 0.09 0.04 0.08 0.08 0.26 0.06 0.17
Ratio 4
�0.01 �0.23 0.01 0.11 0.18 0.23 0.06 0.10 0.12
Italic values are significant at P (two-tailed)40.01 (strict probability adopted to reduce occurrence of Type I Errors). Although

most of the significances shown in this Table were replicated when correlation coefficients were obtained using Spearman’s rho,

there were several exceptions. The majority of these exceptions were minor, with significance levels being at <0.05 instead of

at40.01 (Ra2 and S1, Ra2 and Final, Ra3 and M3, Ra4 and Final) or at40.01 instead of <0.05 (Openness and S1, IQv and S1,

Intellectual Motivation and S3, Ra1 and M1). The two remaining exceptions occurred when Spearman’s rho indicated no sig-

nificant relationship found (even at P<0.05) between Ra2 and M3, and also between Ra3 and Final. See Table 2 for legend.


against hypothesis H5, neither Extraversion nor Neuroticism are significantly correlated with anyof the indicators of academic success.Conscientiousness is significantly associated only with success in obtaining good Tutorial

Grades and good results in the assessment of Year 1 Major courses (which do not contribute todegree classification). Against hypothesis H2, then, there is little evidence of a significant positiveassociation between Conscientiousness and undergraduate academic success.Hypothesis H4 predicted no significant relationship between Agreeableness and academic suc-

cess. Against this, a positive association between these two variables seems to strengthen as time atuniversity goes on, most importantly to the point where it correlates significantly with Final Grade.Of all the trait measures, and in support of hypothesis H1, Openness to experience bears the

most prominent association with academic success. Most notably, Openness to experience corre-lates to approximately the same moderate extent with both A’ level points and the Final Gradeachieved at university.Although existing evidence seemed strong enough to justify not formulating specific hypoth-

eses, it should be reported that verbal intelligence correlated positively with a broad range of aca-demic success measures. Most importantly, and like Openness to experience, verbal intelligencewas significantly associated with both Final Grade and A’ level points.Spatial IQ was not significantly associated with any of the indicators of academic achievement.Seminar absences were not significantly associated with A’ level points, but they were far and

away the best predictor of academic success of any of the criterion examined. With the soleexception with respect to the Year 3 School course, a lack of seminar absences provided thestrongest, and sometimes the only, significant predictor of academic success, whether indicatedvia Seminar Report or Final Grade. Hypothesis H6 therefore receives strong support when‘application’ is operationalised as seminar attendance.Although varying in consistency, the four measures of submitting non-assessed work proved to be

rather good predictors of academic success, each predicting Tutorial Grade at least as well as Con-scientiousness, as well as (in three out of four cases) predicting Final Grade about as well as Agree-ableness. Hypothesis H6 therefore receives further support: ‘application’ in the form of submittingnon-assessed work tends to predict undergraduate academic success.A final point to be noted about the data in Table 3 relates to the finding that Tutorial report

grades were very strongly associated with seminar attendance and fairly strongly with all fourmeasures of submission of non-assessed work, as well as with Conscientiousness. It appears, then,that these tutorial grades are primarily indicators of ‘application’ and effort rather than disposi-tional intellectual ability.

3.4. Regression analysis

A hierarchical multiple regression analysis attempted to predict Final Grade (as a percentage).1 Inaccordance with hypothesis H7, variables were entered in the following order: the two intelligence

1 For reasons of space, final grade (as a percentage) will be used as the sole indicator of academic success during theregression analysis. It may be remembered that Year 1 assessments do not contribute towards degree classification and

that the Final Grade percentage is derived from (and therefore highly correlated with) the more specific indicators ofacademic success available in Years 2 and 3.


measures on Step 1; the single seminar attendance and the four non-assessed work submission indi-cators of ‘application’ on Step 2; and, the five trait personality measures on Step 3.As may be seen from Table 4, Model 1 explained only 4% of the variance in Final Grade, with

verbal intelligence being the only significant predictor.Adding the ‘application’ variables in Step 2 increased the explained variance to 11%. The

additional explained variance came exclusively from the seminar absence measure, withoutaffecting the contribution of the verbal intelligence measure.Adding the personality variables in Step 3 increased the explained variance to 16%. The addi-

tional explained variance came exclusively from the Openness to experience measure, with theseminar absence ‘application’ measure and the verbal intelligence measure still making significantindependent contributions.As predicted by hypothesis H7, then, adding personality measures (in this case, particularly

Openness to experience) enabled us to explain significant variation in our key academic successscore (i.e. Final Grade), even once the effects of (verbal) intelligence and (seminar attendance)application were accounted for. Indeed, inspection of the respective ‘Betas’ reveals that Opennessto experience and ‘seminar attendance’ each account for approximately twice as much explainedvariance in Final Grade than does verbal intelligence.

3.5. Additional analysis

It is noteworthy that Agreeableness did not make a significant contribution to explained var-iance in the hierarchical regression equation predicting Final Grade (see Table 4). In part, thismay be due to shared explained variance with verbal IQ (see Table 1). In addition, it is possible that

Table 4

Regression analysis on undergraduate academic success (Final Grade, as a percentage)

Step 1
Beta t Step 2 Beta t Step 3 Beta t
IQv
0.19 3.54*** 0.18 3.44** 0.13 2.50*
IQs
0.03 0.49 0.05 0.93 0.08 1.50 Absences �0.25 �4.75*** �0.22 �4.32*** Ratio 1 �0.02 �0.51 �0.03 �0.55
Ratio 2
0.08 1.60 0.07 1.50 Ratio 3 0.04 0.74 0.06 1.15 Ratio 4 0.01 0.11 0.00 0.10
Openness
0.25 5.19*** Conscien 0.09 1.91 Extravert �0.05 �1.07 Agreeable 0.03 0.64
Neurotic
�0.00 �0.06
F change (2, 428)=8.91***
F (5, 423)=7.84*** F (5, 418)=6.43*** Multiple R=0.20 Multiple R=0.35 Multiple R=0.43
Adjusted Multiple R square=0.04
Adjusted Multiple R square=0.11 Adjusted Multiple R square=0.16
* P<0.05.** P<0.01.*** P<0.001.


the relationship between Agreeableness and academic success (see Table 3) is at least partiallymediated by the ‘application’ variable of seminar absences (see Table 1).Mediation analysis was carried out in accordance with recommendations made by Baron and

Kenny (1986). Each mediation analysis requires three regression analyses: (i) of the potentialmediator on the predictor, (ii) of the criterion on the predictor, and (iii) of the criterion on boththe predictor and the potential mediator. To establish mediation:

1. the predictor must affect the potential mediator in (i);
2. the predictor must affect the criterion in (ii); and, 3. the potential mediator must affect the criterion in (iii).
Mediation is established if these three conditions hold in the predicted direction and the effectof the (trait) predictor on the (academic success) criterion is less when accompanied by thepotential mediator (application) [i.e. in (iii)] than when not so accompanied [i.e. in (ii)].A first regression confirmed that Agreeableness significantly predicts seminar absences (see

Table 1); b=�0.14, t=�2.47, P<0.05. The second regression confirmed that Agreeableness sig-nificantly predicts Final Grade (see Table 3); b=0.14, t=2.82, P<0.005. The third regressionconfirmed that the contribution of Agreeableness to explaining Final Grade reduced to insignif-icance when accompanied by the seminar absence measure; b=0.08, t=1.63, P>0.1. Thus, itappears that the reason that Agreeableness did not make a significant independent contributionto explaining the variance of Final Grade in Step 3 of the earlier hierarchical regression analy-sis—despite earlier indications that Agreeableness is significantly associated with Final Grade—was because the effects of Agreeableness on Final Grade are wholly mediated by seminar atten-dance (which is why inclusion of the latter variable in the regression analysis meant that noindependent contribution to explained variance in Final Grade was made by Agreeableness).

4. Discussion

Intelligence and motivation are generally accepted as determinants of academic success. It hasbeen proposed recently that personality traits might also predict such success, although to dateempirical support for this proposition has been at best mixed. Having identified various possiblereasons for previous findings in this area, the current study attempted to investigate the extent towhich personality variables predicted undergraduate academic success, even once the effects ofintelligence and application (as an indicator of motivation-in-action) were taken into account.In line with predictions based on previous findings, Extraversion bore no significant linear

relationship with academic success. Also in line with such predictions, Openness to experiencewas positively correlated with undergraduate academic success. Contrary to expectations, neitherConscientiousness nor Neuroticism bore positive relationships with academic success, butAgreeableness did.As expected, both verbal intelligence and application whilst at university were significantly asso-

ciated with academic success. This was true both for zero order correlations and in terms of makingindependent contributions to explaining the variance of undergraduate ‘Final Grades’ when enteredinto regression equations.When trait measures were added to such regression equations, Openness to


experience explained additional unique variance but Agreeableness did not. Further analysis revealedthat the relationship between Agreeableness and undergraduate academic success was wholly medi-ated by the ‘application’ variable of seminar attendance. That is, Agreeableness seemed to be asso-ciated with academic success because Agreeable students went to seminars more often than didless Agreeable students and such application was rewarded in terms of improved Final Grades.Trait theory has often been criticised for being relatively atheoretical, at best providing parsi-

monious descriptions of phenomena but providing little in the way of explanations of why suchphenomena occur (e.g. Cooper, 1998; Pervin & John, 1997). The main significance of the findingsabove relating Openness to experience and Agreeableness to academic success is therefore notsimply identification of these relationships. In itself, such identification merely adds to the mixeddata already reviewed. Rather, the main importance of the findings presented here is that they givesome indication of the processes by which personality traits may influence undergraduate academicsuccess. For example, while it is interesting that Openness to experience was positively associatedwith Final Grade, it is more significant that this relationship remained even when controlling forintelligence and application variables. Apparently, being Open to experience provides academicbenefits beyond those provided by being clever and being motivated to turn up to classes. Previousresearch (e.g. Blickle, 1996; cf. Staudinger, Lopez, & Baltes, 1997; Staudinger, Maciel, Smith, &Baltes, 1998) suggests that this may be because Openness to experience facilitates the use oflearning strategies (e.g. critical evaluation, in-depth analysis, independent research to aid elusivecomprehension) that in turn affect academic success (Mumford & Gustafson, 1988).By way of contrast, Agreeableness in the present study aided academic success solely by fos-

tering application (rather than providing academic benefit independent of both intelligence andapplication). Application, specifically seminar attendance, was far and away the strongest andmost consistent predictor of academic success in the present study. Here, then, we see a potentialinteraction between a personality disposition and a social context variable in determining aca-demic success. It seems possible that relatively low-Agreeableness students would be less com-fortable than relatively high-Agreeableness students with the social interaction intrinsic toseminars. This would presumably contribute to a relative tolerance to missing seminars, with thesubsequent ill-effects on eventual academic performance. Future research might profitably inves-tigate whether the Agreeableness-academic success relationship varies as a function of learningenvironments, e.g., where there are no seminars or where seminar attendance is compulsory.2

De Raad and Schouwenburg (1996, p. 321) note that there has been little or no research on‘‘whether and how’’ the effects of personality on academic success might be ‘‘regulated or adjus-ted, complemented or compensated.’’ The present study sheds light on how such research mightbe attempted. Specifically, it seems likely that learning environments and assessment methodscould be effectively tailored to students’ Openness to experience and Agreeableness levels. Stu-dents relatively high in Openness to experience should thrive in educational settings promoting

2 Observant readers might object that Conscientiousness in the present study was associated with seminar atten-dance but not (generally) academic success. It can be noted, however, that zero-order correlation between Con-scientiousness and Final Grade approached significance (r=0.10, P=0.08) and that additional mediational analysissuggested that this relationship between Conscientiousness and Final Grade was mediated by seminar attendance. We

would argue that restriction of range problems may have had a more detrimental effect on measurement of Con-scientiousness than on Agreeableness among our sample.


and rewarding critical and original thought, while students relatively low in Openness to experi-ence (but matched in intelligence and motivation/application) should fare better in educationalsettings promoting and rewarding acquisition of received wisdom. Similarly, students relativelyhigh in Agreeableness should thrive when instruction and assessment occurs via collaborativesocial interaction, while those lower in Agreeableness should fare better in educational settingswhere students are less socially interdependent (or are even negatively interdependent, see Roth-stein et al., 1994). On-going analysis using the current data set is employing similar reasoning toground investigation of possible moderating effects of both subjects studied and assessmentmodes on trait-academic success relationships (cf. De Fruyt & Mervielde, 1996).To the extent that its findings are valid, the present study has considerable practical implica-

tions. First, it suggests that efforts to identify potentially successful students via personality pro-filing will be most effective when adequate account is taken both of what actually constitutes‘‘success’’ and of how such ‘‘success’’ is ascertained. For example, Conscientiousness was the besttrait predictor of Tutorial Grade in the present study, while Openness to experience was the besttrait predictor of Final Grade (with Agreeableness being the best trait predictor of both applica-tion and formal assessment success). Secondly, educators and students alike need to be aware ofthe implications of such potential mismatches between various indicators of ‘‘academic success’’.Many university tutors are aware of how common it is for students to complain that they havereceived very favourable feedback about how academically successful they are until they learnthat they have suddenly fared relatively poorly when it mattered most to them (e.g. during theirfirst major formal assessment). Again, the relatively distinct determinants of ‘Tutorial Grade’ and‘Final Grade’ illustrate how easily such an event could occur.3

Despite the possibilities outlined in the previous paragraph, perhaps the most important prac-tical implication of the current study is to recognise the limitations of employing a reified ‘indivi-dual differences’ approach to predicting academic success. In combination, all our trait,intelligence, and application measures explained only about 16% of the variance in Final Gradesamong our sample (cf. Paunonen & Ashton, 2001). This is roughly comparable to resultsobtained in earlier studies, even though every effort was made to overcome potential methodo-logical limitations associated with much previous research. Using the best available measures ofthe currently most popular individual difference constructs, our study was unable to account for84% of the variance of our main measure of academic success, i.e. Final Grades. Thus, it seemsthat individual differences in success within the current undergraduate system may be relativelyindependent of intelligence, application, and personality.4 This suggests that any attempt to useindividual difference measures to discriminate between the probable academic success of studentscurrently offered university places may itself be doomed to failure. Against this, personality

3 It has already been noted that Tutorial Grades seem to be primarily indicative of effort. Further regression analysisconfirmed that verbal intelligence and Openness to experience (each indicative of Final Grade success) made no additional

contribution to explained Tutorial Grade variance beyond that made by Conscientiousness and ‘seminar absences’. Thus, itis entirely possible that a motivated but not especially intelligent (or Open) student might obtain a disappointing FinalGrade if they had been unfortunate enough to infer their likely success from very positive Tutorial Grades.

4 This of course entails neither that intelligence, application, and personality are irrelevant to academic success, northat indicators of such constructs (e.g. A’ level results) should not be used to select university undergraduates. It simplymeans that relative differences in levels of such constructs among the undergraduates in our sample played a relatively

minor role in explaining differences in the Final Grades obtained.


measures seem to have a potentially promising role in ‘matching’ education environments andprocesses to students’ relative strengths.Although successfully avoiding many of the methodological problems associated with past

studies in the same domain, weaknesses of the current study should be assessed. First, whilst thetime-sequence of measures used limits some of the problems associated with correlational studies(e.g. direction of causation), it does not avoid them entirely. It is therefore hoped that all inter-pretations of findings reported will be both replicated and further investigated by experimentaland/or intervention studies, e.g. ones crossing individual differences with teaching and assessmentmodes. Secondly, whilst we are confident that our intelligence measure overcame restriction ofrange problems, we are not so sure that the same is true for all our measures. It would aid edu-cational research of the sort reported here if motivation and personality measures could bedeveloped specifically for populations of undergraduates. Thirdly, we are acutely aware that wehave only investigated a relatively small sub-set of available individual difference measures.Future research should attempt to compare the relative merits of measures we have used againstothers that show promise in predicting academic success, e.g., social support/loneliness andstress/coping (e.g. Halamandaris & Power, 1999; Mellanby et al., 2000); learning styles (Busato etal., 2000; Mellanby et al., 2000); causality orientations and academic experience (Wong, 2000);self-esteem, self-efficacy, and locus of control (Mellanby et al., 2000; Mwamwenda, 1996; Shereret al., 1982).

Acknowledgements

This paper results fron research financially supported by the British Academy and theUniversity of Sussex Teaching and Learning Development Fund.

References

Ackerman, P. L., & Heggestad, E. D. (1997). Intelligence, personality, and interests: evidence for overlapping traits.Psychological Bulletin, 121, 219–245.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychology research:conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182.

Blickle, G. (1996). Personality traits, learning strategies, and performance. European Journal of Personality, 10, 337–352.Busato, V. V., Prins, F. J., Elshout, J. J., & Hamaker, C. (2000). Intellectual ability, learning style, achievement moti-

vation and academic success of psychology students in higher education. Personality and Individual Differences, 29,1057–1068.

Cassidy, T., & Lynn, R. (1989). A multifactorial approach to achievement motivation: the development of a compre-

hensive measure. Journal of Occupational Psychology, 62, 301–312.Collis, J. M., & Messick, S. (Eds.). (2001). Intelligence and personality: bridging the gap in theory and measurement.Mahwah, NJ, US: Lawrence Erlbaum Associates.

Cooper, C. (1998). Individual differences. London: Arnold.Costa, P. T. Jr., & McCrae, R. R. (1989). NEO-PI/NEO-FFI Manual supplement. Odessa, FL, US: PsychologicalAssessment Resources.

Costa, P. T. Jr., & McCrae, R. R. (1992). Four ways five factors are basic. Personality and Individual Differences, 13,653–665.


De Fruyt, F., & Mervielde, I. (1996). Personality and interests as predictors of educational streaming and achievement.

European Journal of Personality, 10, 405–425.De Raad, B., & Schouwenburg, H. C. (1996). Personality in learning and education. European Journal of Personality,

10, 303–336.

Entwistle, N. J. (1972). Personality and academic attainment. British Journal of Educational Psychology, 42, 137–151.Furnham, A., & Mitchell, J. (1991). Personality, needs, social skills and academic performance: a longitudinal study.

Personality and Individual Differences, 12, 1067–1873.

Goff, M., & Ackerman, P. L. (1992). Personality-intelligence relations: assessment of typical intellectual engagement.Journal of Educational Psychology, 84, 537–552.

Halamandaris, K. F., & Power, K. G. (1999). Individual differences, social support and coping with examination stress:a study of the psychosocial and academic adjustment of first year home students. Personality and Individual Differ-

ences, 26, 665–685.Harris, D. (1940). Factors affecting college grades: a review of the literature, 1930–1937. Psychological Bulletin, 37,125–166.

Heaven, P. C. L., Mak, A., Barry, J., & Ciarrochi, J. (2002). Personality and family influences on adolescent attitudesto school and self-rated academic performance. Personality and Individual Differences, 32, 453–462.

Heim, A. W. (1968). AH5 Group Test of high-grade intelligence: Manual. Windsor: NFER Publishing.

Hirschberg, N., & Itkin, S. (1978). Graduate student success in psychology. American Psychologist, 33, 1083–1093.Kline, P., & Gale, A. (1971). Extroversion, neuroticism and performance in a psychology examination. British Journal

of Educational Psychology, 41, 90–93.

Matthews, G. (1997). Intelligence, personality and information-processing: An adaptive perspective. In J. Kingma, &W. Tomic (Eds.), Advances in cognition and educational practice: reflections on the concept of intelligence (pp. 175–200). Stamford, CT, US: JAI Press.

McClelland, D. C., Atkinson, J. W., Clark, R. A., & Lowell, F. L. (1953). The achievement motive. New York:

Appleton-Century-Crofts.Mellanby, J., Martin, M., & O’Doherty, J. (2000). The ‘gender gap’ in final examination results at Oxford University.

British Journal of Psychology, 91, 377–390.

Mumford, M. D., & Gustafson, S. B. (1988). Creativity syndrome: integration, application and innovation. Psycholo-gical Bulletin, 103, 27–43.

Mwamwenda, T. S. (1996). South African graduate students’ locus of control, gender differences, and academic per-

formance. Psychological Reports, 77, 629–631.Neisser, U., Boodoo, G., Bouchard, T. J. Jr., Boykin, A. W., Brody, N., Ceci, S. J., Halpern, D. F., Loehlin, J. C., Perloff,R., Sternberg, R. J., & Urbina, S. (1996). Intelligence: knowns and unknowns. American Psychologist, 51, 77–101.

Osbourne, M., Leopold, J., & Ferrie, A. (1997). Does Access work? The relative performance of Access Students at aScottish University. Higher Education, 33, 155–176.

Paunonen, S. V., & Ashton, M. C. (2001). Big Five Factors and Facets and the prediction of behaviour. Journal ofPersonality and Social Psychology, 81, 524–539.

Pervin, L. A., & John, O. P. (1997). Personality: theory & research (7th ed.). New York: Wiley.Pintrich, P. R., & Schunk, D. H. (1996). Motivation in education: theory, research and applications. Englewood Cliffs,NJ: Prentice Hall.

Roberts, M. J. (2002). The relationship between extraversion and ability. Personality and Individual Differences, 32,517–522.

Rolfhus, E. L., & Ackerman, P. L. (1999). Assessing individual differences in knowledge: Knowledge, intelligence, and

related traits. Journal of Educational Psychology, 91, 511–526.Rothstein, M. G., Paunonen, S. V., Rush, J. C., & King, A. (1994). Personality and cognitive ability predictors ofperformance in graduate Business School. Journal of Educational Psychology, 86, 516–530.

Saklofske, D. H., Zeidner, M. (Eds.). (1995). International handbook of personality and intelligence: perspectives on

individual differences. New York, NY, USA: Plenum Press.Saunders, P., & Woodfield, R. (1999). Explaining gender differences in achievement in Higher Education: Preliminaryresults from a British panel survey. In P. Fogelberg, J. Hearn, L. Husu, & T. Mankkinen (Eds.), Hard work in the

academy: research and interventions on gender inequalities (pp. 93–101). Helsinki: Helsinki University Press.


Schuerger, J. M., & Kuna, D. L. (1987). Adolescent personality and school performance: a follow-up study. Psychology

in the Schools, 24, 281–285.Sherer, M., Maddux, J. E., Mercandante, B., Prentice-Dunn, S., Jacobs, B., & Rogers, R. W. (1982). The Self-EfficacyScale: construction and validation. Psychological Reports, 51, 663–671.

Staudinger, U. M., Lopez, D. F., & Baltes, P. B. (1997). The psychometric location of wisdom-related performance:intelligence, personality, and more?. Personality and Social Psychology Bulletin, 23, 1200–1214.

Staudinger, U. M., Maciel, A. G., Smith, J., & Baltes, P. B. (1998). What predicts wisdom-related performance? A first

look at personality, intelligence, and facilitative experiential contexts. European Journal of Personality, 12, 1–17.Sternberg, R. J., Ruzgis, P. (Eds.). (1994). Personality and intelligence. Cambridge: Cambridge University Press.Stough, C., Brebner, J., Nettlebeck, T., Cooper, C. J., Bates, T., & Mangan, G. L. (1996). The relationship betweenintelligence, personality, and inspection time. British Journal of Psychology, 87, 255–268.

Wiggins, N., Blackburn, M., & Hackman, J. R. (1969). Prediction of first-year graduate success in psychology: peerratings. Journal of Educational Research, 63, 82–85.

Wolchover, J. (2000). New tests aim to dig out hidden university talent. Evening Standard, 12 April.

Wolfe, R. N., & Johnson, S. D. (1995). Personality as a predictor of college performance. Education and PsychologicalMeasurement, 55, 177–185.

Wong, M. M. (2000). The relations among causality orientations, academic experience, academic performance, and

academic commitment. Personality and Social Psychology Bulletin, 26, 315–326.Zeidner, M., & Matthews, G. (2000). Intelligence and personality. In R. J. Sternberg (Ed.), Handbook of intelligence(pp. 581–610). Cambridge: Cambridge University Press.


individual differences and undergraduate academic success,the roles of personality, intelligence,...

Documents