Little more than personality: Dispositional determinants of test anxiety (the Big Five, core self-evaluations, and self-assessed intelligence)

Transcription

1 Available online at Learning and Individual Differences 18 (2008) Little more than personality: Dispositional determinants of test anxiety (the Big Five, core self-evaluations, and self-assessed intelligence) Tomas Chamorro-Premuzic a,, Gorkan Ahmetoglu a, Adrian Furnham b a Goldsmiths University of London, United Kingdom b University College London Received 8 October 2006; received in revised form 3 August 2007; accepted 1 September 2007 Abstract This study attempted a hierarchical integration of several dispositional determinants of test anxiety (TA) [Sarason, I.G. (1984). Stress, anxiety and cognitive interference: Reactions to tests. Journal of Personality and Social Psychology, 46, ], namely the Big Five personality traits [Costa, P.T. Jr., & McCrae, R.R. (1992). Revised NEO personality inventory (NEO-PI-R) and NEO five-factor inventory (NEO-FFI): Professional manual. Odessa: Psychological Assessment Resources.], core self-evaluations (CSE) [Judge, T.A., Erez, A., Bono, J.E., & Thoreson, C.J. (2003). The Core Self-Evaluation Scale: Development of a measure. Personnel psychology, 56, ], and self-assessed intelligence (SAI) [Furnham, A. (2001). Self-estimates of intelligence: Culture and gender difference in self and other estimates of both general (g) and multiple intelligences. Personality and Individual Differences, 31, ] in a sample of 388 US and UK university students. Structural equation models showed that TA was largely a function of Neuroticism, and that CSE and SAI do not contribute to the prediction of TA over established personality traits. Furthermore, the relationship between CSE and TA was fully accounted for by personality traits, whereas SAI was not a significant predictor of TA. The results undermine the notion that self-beliefs affect TA and suggest that wider dispositions play a salient role determining individual differences in TA. Theoretical and applied implications are discussed, particularly in regards to educational settings Elsevier Inc. All rights reserved. Keywords: Test anxiety; Core self-evaluations; Self-assessed intelligence; Big Five; Personality 1. Introduction Every year, millions of students under-perform in school and university because of heightened test anxiety (TA) (Hill & Wigfield, 1984; Hopko, Ashcraft, Gute, Ruggiero, & Lewis, 1998; Zohar, 1998), which is defined as the set of phenomenological, physiological, and behavioral responses that accompany concern about possible negative consequences or failure on an exam or similar evaluative situation (Zeidner, 1998, p.17). Although TA is known to depend on situational variables, such as levels of motivation, task complexity, and the practical consequences of high or low performance (Humphreys & Revelle, 1984), it varies markedly from one individual to Corresponding author. Department of Psychology, Goldsmiths College, University of London, London, New Cross, SE14 6NW, United Kingdom. Tel.: ; fax: address: t.chamorro-premuzic@gold.ac.uk (T. Chamorro-Premuzic). another. Thus, some individuals will be relatively calm when it comes to completing a test, whilst others will generally perceive examinations as more dangerous or threatening ( ) and experience more intense levels of state anxiety when taking tests (Spielberger & Vagg, 1995, p.6). Studies on the self-defeating consequences of subjective cognitions (Ellsworth & Smith, 1988; Lazarus, 1991) suggest that poorer self-beliefs are a major cause of TA. However, psychologists use different labels to refer to what are arguably overlapping constructs, for instance: Self-regulation (Carver & Scheier, 1984), self-efficacy (Bandura, 1986, Smith, Arnkoff, & Wright, 1990; Zohar, 1998), locus of control (Braden, 1995; Feather & Volkmer, 1988), attribution for failure and success (Bandalos, Yates, & Thorndike-Christ, 1995), self-worth (Covington, 1992), emotional self-efficacy (Petrides & Furnham, 2000), self-concept (Bandalos et al., 1995), self-awareness (Fletcher & Baldry, 2000), meta-cognition (Stankov, 1999), self-handicap (Rhodewalt, 1990), self-evaluation (Morris & /$ - see front matter 2007 Elsevier Inc. All rights reserved. doi: /j.lindif

2 T. Chamorro-Premuzic et al. / Learning and Individual Differences 18 (2008) Liebert, 1969), self-motivation (Zeidner, 1995), self-concept (Rindermann & Neubauer, 2001), and self-confidence (Koivula, Hassmen & Fallby, 2002). To overcome this problem, and account for the overlap between different measures of self-belief, Judge, Erez, and Bono (1998) proposed the concept of core self-evaluations (CSE) and also developed a measure of this construct. The concept of CSE was defined as a broad personality trait reflecting the most general and fundamental beliefs individuals hold about themselves (Judge et al., 1998). Accordingly, individuals CSE comprise four components referring to individual differences in a) self-esteem or perception of one s worth, value, and importance, b) generalized self-efficacy or one s typical level of confidence about the likelihood of performing well, c) locus of control or one s perceived degree of control over life events and situations, and d) neuroticism/emotional stability or one s tendency to experience negative affect, increased levels of worry, and pessimistic beliefs. A large meta-analytic study (Judge, Erez, Thoreson, & Bono, 2002) confirmed that most variance among these components can be explained by a general factor, which was more accurate than the four components in predicting external criteria (Best, Downey, & Stapleton, 2005; Judge & Bono, 2001; Judge, Kluger, Locke & Durham, 1998), though research is yet to examine its relation to TA. Given the dispositional nature of CSE, it is particularly important to determine whether any links between CSE and TA could be explained by established personality traits, notably the Big Five factors of Neuroticism (N), Extraversion (E), Openness to Experience (O), Agreeableness (A), and Conscientiousness (C) (Costa & McCrae, 1992). Indeed, Judge et al.(2003) suggested that CSE may be partly dependent on N, E, and C, such that emotionally stable, extraverted, and conscientious individuals are generally more likely to hold positive selfbeliefs. In one of the few attempts to provide a hierarchical model for integrating dispositional (both personality and selfevaluative) determinants of AP, Chamorro-Premuzic and Furnham (2006a) proposed that self-expectancy constructs may largely be influenced by personality traits, not only N (negatively) but also E (positively) and O (positively). Accordingly, self-expectancy constructs may mediate (account for) the link between established personality factors and TA. However, the authors looked at AP measures (i.e., exams, continuous assessment, and supervised dissertations) rather than TA, and conceptualized self-evaluations merely in terms of selfassessed intelligence (SAI), that is, people's estimates of their cognitive abilities in relation to the overall population (see also Marsh, Trautwein, Lüdtke, Köller & Baumert, 2006 for similar hierarchical models of core versus surface characteristics in the prediction of AP). Whilst SAI has shown incremental validity predicting AP over and above personality and intelligence (Chamorro-Premuzic & Furnham, 2006b), its relationship to TA remains to be tested. Thus the present study set out to explore whether individual differences in self-beliefs, as operationalized through CSE and SAI, account for variance in TA independently of established traits, and whether CSE and SAI may explain the effects of personality traits on TA The hypothesized models The first model tested whether the relationship between the Big Five personality traits and TA was mediated by SAI and CSE. In line with previous research, paths from N, E, O, and A, were loaded on to SAI (see Chamorro-Premuzic, Moutafi & Furnham, 2005; Furnham & Buchanan, 2005), and paths from N, E, and C were loaded onto CSE (see Judge et al., 2003). Direct paths between the Big Five and TA were also allowed according to past findings (see Fitch, 2005; McIlroy & Bunting, 2002). SAI and CSE were both allowed to influence TA. This model was compared to a competing model. The main interest was directed at two variables, namely N and CSE. Because Judge et al. (1998) argued that CSE is a broad latent personality construct, a model, in which CSE and the Big Five were simply correlated, with no causal paths, was tested. The idea was to include both N and CSE in the same structure, where neither one factor is broader than the other (Judge et al., 2002). The directionality of the model is conceptual rather than causal and can be justified on the basis that the Big Five are less affected by situational variables than is TA. Furthermore, as poorer self-beliefs are assumed to be a major cause of TA, CSE are treated as both exogenous and endogenous, though whether CSE are as broad as the Big Five or merely a self-conception/ surface characteristic, akin to SAI, is a matter of theoretical speculation (see Asendorpf & van Aken, 2003). 2. Method 2.1. Participants Participants were 388 undergraduate students from British (250) and American (138) universities. There were 287 females and 101 males. Age ranged from 17 to 29 (M=19.7, SD=2.8) years. Most students were psychology undergraduates who volunteered to take part in this study in exchange of course credit units Measures Reaction to Tests (RTT, Sarason, 1984) This is a 40-item measure of TA composed of four anxiety components, namely worry (e.g. "I wonder how other people are doing";), irrelevant thinking (e.g. ";My mind wanders during tests";), tension (e.g. ";I feel distressed and uneasy before tests";), and bodily reactions (e.g. ";I get a headache before a test";). These components form two cognitive (worry irrelevant thinking) and two emotional (tension bodily reactions) dimensions of TA. Initially, four components were extracted as specified by Sarason (1984). However, these were highly inter-correlated and following a confirmatory factor analysis, it was decided to compute one higher order TA factor, which was retained for further analysis and treated as criterion. Studies have provided strong support for the psychometric adequacy of the RTTscale (Zimmer et al., 1992; Sarason, 1984).

3 260 T. Chamorro-Premuzic et al. / Learning and Individual Differences 18 (2008) Table 1 Correlations between the Big Five personality traits, SAI, CSE, and test anxiety N 2. E O A C SAI CSE TA n=388. N=Neuroticism, E=Extraversion, O=Openness, A=Agreeableness, C=Conscientiousness, CSE=core-self-evaluation, SAI=self-assessed intelligence, TA=test anxiety. pb.05, pb The NEO Five Factor Inventory Revised (NEO-FFI; Costa & McCrae, 1992) This 60-item, non-timed questionnaire, assesses the Big Five personality factors, namely Neuroticism, Extraversion, Openness to Experience, Agreeableness and Conscientiousness. Items involve questions about characteristic behaviors or reactions, which are answered on a five-point Likert scale, ranging from strongly disagree to strongly agree. The manual reports extensive data on the reliability and validity of this inventory (see Costa & McCrae, 1992) Self-assessed Intelligence SAI: (Furnham; 2001) This is a 10-item self-report measure of participants selfestimated intelligence, which contains: verbal, logical spatial, musical, body-kinetic, interpersonal, intrapersonal, existential, spiritual and naturalistic intelligence. The scale includes a bellcurve graph indicating numbers and labels for all ability scores (ranging from 55 mild retardation, to 145 gifted ). Following PCA (Eigenvalues and Scree Test), a single factor was computed which accounted for 38.3 % of the variance in the data The Core Self-evaluation Scale (CSES; Judge, Erez, Bono, & Thoreson, 2003) This is a 12-item inventory that measures a single factor, i.e. CSE. Items involve statements about typical thoughts/feelings ( Overall, I am satisfied with myself ) and behaviors ( I complete tasks successfully ), which are answered on a five-point Likert scale, ranging from strongly disagree to strongly agree. PCA showed that the items loaded onto a single factor, which accounted for 39.9% of the variance in the data. There has been good psychometric support for the CSES. The items reflect a onefactor structure, and the CSES correlates significantly with expected criteria (Judge et al., 2003) Procedure Questionnaires were administrated in a large lecture theatre and in the presence of five trained graduate students who ensured appropriate administration procedures. There were four groups of students, two in the US. and two in the UK. Order of questionnaires was randomized to control for order effects. Individual feedback was offered to each participant. 3. Analysis and results 3.1. Correlations Bivariate correlation coefficients (Pearson s r) for the central measures are shown in Table 1. Descriptive statistics (M and SD) and the reliability of the scales are reported in Table 2. Structural equation modeling using AMOS (Arbruckle, 1999) was applied to estimate a series of models. Latent variables were N, E, O, A, C, SAI, CSE, and TA. The observed variables for the Big Five were obtained by using a split-half method, where the observed variables are comprised of randomly determined halves of the items that load on a particular hypothesized factor (all 60 items were used). All 10 items of SAI and all twelve items of the CSE were used as indicators for their respective latent variables. For TA the observed variables were worry, test-irrelevant thinking, bodily reactions and tension. Several goodness-of-fit indices were used for the structural models. Bentler's (1990) comparative fit index can be interpreted as the improvement in fit of the hypothesized model over a baseline model, relative to the fit of the baseline model. The incremental index of fit was developed by Bollen (1989) to address the issues of parsimony and sample size, and thus taking degrees of freedom into account. The root mean square error of approximation (Steiger & Lind, 1980) takes into account the error of approximation in the population and evaluates how well the model would fit the population covariance matrix if it were available with unknown but optimally chosen parameter values. Brown and Cudeck's (1989) expected cross-validation index measures the discrepancy between the fitted covariance matrix in the analyzed sample, and the expected covariance matrix that would be Table 2 Means, standard deviations (SD), reliability, and item loadings of the Big Five, CSE, SAI, and test anxiety Means SD Cronbach α Item loadings Neuroticism (N) min:.79 max:.88 mean:.84 Extraversion (E) min:.69 max:.86 mean:.78 Openness (O) min:.60 max:.89 mean:.75 Agreeableness (A) min:.56 max:.91 mean:.74 Conscientiousness (C) min:.60 max:.85 mean:.73 Core self-evaluation (CSE) min:.36 max:.70 mean:.57 Self-assessed Intelligence (SAI) min:.32 max:.76 mean:.54 Test anxiety min:.40 max:.80 mean:.68 n=388. polarity: high scores on CSE refer to positive self-evaluations, whereas low scores refer to personal negativity. All other scales are also positive. Item loadings: completely standardized solution.

4 T. Chamorro-Premuzic et al. / Learning and Individual Differences 18 (2008) obtained in another sample of equivalent size. Finally, Akaike's (1987) information criterion, addresses the issue of parsimony in the assessment of model fit; as such, statistical goodness of fit and number of estimated parameters are taken into account Testing between structural equation models The fit statistics for the first hypothesized model (Fig. 1) were, χ 2 (545, N=388)=1499.8, p b.01; root mean square error of approximation (RMSEA) =.067; comparative fit index (CFI) =.98; incremental fit index (IFI) =.98; Akaike s information criterion (AIC) = ; expected cross-validation index (ECVI) =4.50. Though the RMSEA was somewhat above the desired value (b.06=good fit), this model showed adequate fit to the data. However, several paths in the model were found to have non-significant values. In further fitting efforts these paths were deleted one at a time, and the model was re-estimated each time. The first paths freed were between E and SAI (β =.02, pn.05), and E and CSE (β=.03, pn.05). In contrast to findings of previous studies these paths were non-significant. Another link that was freed, and was not in line with previous research, was that between C and TA (β=.03, pn.05), as also this was non-significant. Finally, in the hypothesized and modified models throughout, the link between SAI and TA remained nonsignificant (β=.03, pn.05). In addition, although the weight of the path between CSE and TA was β=.19, it was also found to be non-significant, therefore not supporting the hypothesis of the study. Non-significant parameters can be considered unimportant to the model, and in the interest of scientific parsimony, they should be deleted (Arbruckle, 1999). Dropping the two links between CSE and TA, and SAI and TA, did not affect the model fit but did increased parsimony (χ 2 (550, N=388)=1504.0, pb.01; RMSEA=.067; CFI=.98; IFI=.98; AIC = ; ECVI =.4.48), indicating that these links were Fig. 2. The modified model (completely standardized solutions; arrows with broken lines are non-significant paths). statistically not necessary in the study. The modified model is shown in Fig. 2. What is worth noting here is the weight of the path between N and CSE (β=.77, pn.01). This indicates that in this model, a great deal of the variance in CSE was accounted for by N, which might explain the non-significant path between CSE and TA. Thus, the only paths that loaded significantly on TA, were those of N (β=.72, pb.01) and E (β=.23, pb.01). It was concluded that the association between the Big Five and TA is not meditated by SAI and CSE and the hypothesized model of the study was not supported. The single best predictor of TAwas N, followed by E Alternative model In addition to the first hypothesized model, an alternative model was tested that involved competing conceptualizations, Fig. 1. The hypothesized model, linking the Big Five, CSE, SAI, and TA. N=Neuroticism, E=Extraversion, O=Openness, A=Agreeableness, C=Conscientiousness, CSE =core-self-evaluations, SAI =self-assessed intelligence, TA=test anxiety. Fig. 3. Model where CSE and Big Five traits are simply correlated with no causal links (completely standardized solutions; note: paths between N and CSE, and C and CSE, are correlations).

5 262 T. Chamorro-Premuzic et al. / Learning and Individual Differences 18 (2008) specifically, between the two latent variables, N and CSE. In this model, CSE was allowed to act as an exogenous broad personality factor, rather than a mediator in the relationship between the Big Five and TA. That is, though the direct paths between CSE, SAI and TA, as well as the Big Five and TA were retained, the causal paths between the Big Five and CSE were deleted (and replaced with correlational paths). Fig. 3 displays the results of the final model. The fit statistics of this model were, χ 2 (548, N=388) = , p b.01; RMSEA =.067; CFI =.98; IFI =.98; AIC = ; ECVI = In this model the fit values were almost the same (albeit with slightly worse parsimony), and the relationship between CSE and TA remained non-significant. Dropping this link, again, did not affect the model fit and increased parsimony slightly. The links between the Big Five and TA remained similar. The correlation between N and CSE was r= Discussion This study tested the extent to which individual differences in self-beliefs, as operationalized through CSE and SAI, account for variance in TA independently of established traits, and whether CSE and SAI may explain the effects of personality traits on TA. Although previous research had linked specific aspects of CSE (i.e., self-esteem, locus of control, self-efficacy, and neuroticism) to TA, no study had, to our knowledge, previously looked at CSE in academic contexts, let alone in regards to TA. Likewise, the extent to which individuals SAI may predict TA had not been examined in any previous study, although a recent paper showed that SAI had incremental validity (over personality and psychometric intelligence) in the prediction of AP (Chamorro-Premuzic & Furnham, 2006b). The present results showed that SAI and CSE did not mediate the relationship between the Big Five personality traits and TA. Rather, there was a strong direct path from N to TA, as well as modest but significant direct effects of E on TA. In fact, when these personality traits were taken into account, CSE no longer explained variance in TA. This indicates that established traits are better predictors of TA than are CSE, and that the relationship between CSE and TA can be fully explained by personality traits. In light of the vast number of past studies highlighting the self-fulfilling effects of self-expectancy constructs as a major cause of individual differences in TA (e.g., Bandalos et al., 1995; Bandura, 1986, Koivula et al, 2002; Smith et al, 1990; Zohar, 1998), the findings of this study are particularly striking. The present findings suggest that, when the main personality traits, particularly N, are taken into account, the effects of selfbeliefs on TA are substantially attenuated. Thus, future studies assessing the effects of self-beliefs on TA may wish to consider major dispositional factors, such as the Big Five. There were of course limitations to the present study, notably the use of a relatively small and highly selected university student sample, as well as the absence of AP data. Additionally, a general TA construct was used in the analysis, though this was based on statistical grounds (see also Zimmer, Hocevar, Bachelor, & Meinke, 1992). On the other hand, assessing state-rather than trait-ta, would have enabled us to combine experimental procedures with our correlational design, for instance by manipulating levels of TA via external variables (e.g., time pressure, reward, caffeine, etc.) and assess potential interactions between dispositional and experimental effects. That said, there are important implications to be drawn from the current results. Methodologically, they suggest that it is more effective to predict TA from established personality inventories, such as the Big Five (particularly the N scale), than from self-belief measures. Theoretically, they indicate that the unique contribution of self-belief variables to individual differences in TA is at least lower than one may have expected from the previous self-expectancy literature. Finally, from an applied point of view, it seems reasonable to recommend that any psychological or educational intervention directed at reducing the self-defeating cognitions of students with poor self-beliefs pays attention to those high on the N trait, as these individuals are clearly more vulnerable than others. References Akaike, H. (1987). Factor analysis and AIC. Psychometrica, 52, Arbruckle, J. L. (1999). AMOS 4.0 [Computer software] Chicago: Smallwaters. Asendorpf, J. B., & van Aken, M. A. G. (2003). Personality-relationship transaction in adolescence: Core versus surface personality characteristics. Journal of Personality, 71, Bandalos, D. L., Yates, K., & Thorndike-Christ, T. (1995). Effects of math selfconcept, perceived self-efficacy, and attributions for failure and success on test anxiety. Journal of Educational Psychology, 87, Bandura, A. (1986). Social foundations of thoughts and action: A social cognitive theory New Jersey: Prentice-Hall. Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, Best, R. G., Downey, R. G., & Stapleton, L. M. (2005). Core self-evaluations and job burnout: The test of alternative models. Journal of Occupational Health Psychology, 10, Bollen, K. A. (1989). Structural equations with latent variables New York: John Wiley. Braden, J. P. (1995). Intelligence and personality in school and educational psychology. In D. H. Saklofske & M. Zeidner (Eds.), International handbook of personality and intelligence: Perspectives on individual differences (pp ). New York, US: Plenum Press. Brown, M. W., & Cudeck, R. (1989). Single sample cross-validation indices for covariance structures. Multivariate Behavioral Research, 24, Carver, C. S., & Scheier, M. F. (1984). Self-focused attention in test anxiety: A general theory applied to a specific phenomenon. In H. M. Van der Ploeg, R. Schwarzer, & C. D. Spielberger (Eds.), Advances in test anxiety research, vol. 3. (pp. 3 20) Lisse, the Netherlands: Swets & Zeitlinger. Chamorro-Premuzic, T., & Furnham, A. (2006). Intellectual competence and the intelligent personality: A third way in differential psychology. Review of General Psychology, 10, Chamorro-Premuzic, T., & Furnham, A. (2006). Self-assessed intelligence and academic performance. Educational Psychology, 26, Chamorro-Premuzic, T., Moutafi, J., & Furnham, A. (2005). The relationship between personality traits, subjectively-assessed and fluid intelligence. Personality and Individual Differences, 38, Costa, P. T., Jr., & McCrae, R. R. (1992). Revised NEO personality inventory (NEO-PI-R) and NEO five-factor inventory (NEO-FFI): Professional manual Odessa: Psychological Assessment Resources. Covington, M. V. (1992). Making the grade: A self-worth perspective on motivation and school reform Cambridge, England: Cambridge University Press. Ellsworth, P. C., & Smith, C. A. (1988). From appraisal to emotion: Differences among unpleasant feelings. Motivation and Emotion, 12,

6 T. Chamorro-Premuzic et al. / Learning and Individual Differences 18 (2008) Feather, N. T., & Volkmer, R. E. (1988). Preference for situations involving effort, time pressure, and feedback in relation to type A behavior, locus of control, and test anxiety. Journal of Personality and Social Psychology, 55, Fitch, B. D. (2005). A test of the relationship between personality traits and test anxiety. Dissertation abstracts international section: Humanities and social sciences, 65, Fletcher, C., & Baldry, C. (2000). A study of individual differences and self awareness in the context of multi-source feedback. Journal of Occupational and Organizational Psychology, 73, Furnham, A. (2001). Self-estimates of intelligence: Culture and gender difference in self and other estimates of both general (g) and multiple intelligences. Personality and Individual Differences, 31, Furnham, A., & Buchanan, T. (2005). Personality, gender and self-perceived intelligence. Personality and individual differences, 39, Hopko, D. R., Ashcraft, M. H., Gute, J., Ruggiero, K. J., & Lewis, C. (1998). Mathematics anxiety and working memory: Support for the existence of a deficient inhibition mechanism. Journal of Anxiety Disorders, 12, Humphreys, M. S., & Revelle, W. (1984). Personality, motivation, and performance: A theory of the relationship between individual differences and information processing. Psychological Review, 91, Judge, T. A., & Bono, J. E. (2001). Relationship of core self-evaluations traits-selfesteem, generalized self-efficacy, locus of control, and emotional stabilitywith job satisfaction and job performance: A meta-analysis. Journal of Applied Psychology, 86, Judge, T. A., Erez, A., & Bono, J. E. (1998). The power of being positive: The relationship between positive self-concept and job performance. Human Performance, 11, Judge, T. A., Erez, A., Bono, J. E., & Thoreson, C. J. (2003). The core self-evaluation scale: Development of a measure. Personnel psychology, 56, Judge, T. A., Erez, A., Thoreson, C. J., & Bono, J. E. (2002). Are measures of self-esteem, neuroticism, locus of control, and generalized self-efficacy indicators of a common core construct? Journal of Personality and Social Psychology, 83, Judge, T. A., Kluger, A. N., Locke, E. A., & Durham, C. C. (1998). Dispositional effects on job and life satisfaction: The role of core evaluations. Journal of Applied Psychology, 83, Koivula, N., Hassmen, P., & Fallby, J. (2002). Self-esteem and perfectionism in elite athletes: Effects on competitive anxiety and self-confidence. Personality and Individual Differences, 32, Lazarus, R. S. (1991). Emotion and adaptation New York: oxford University Press. Liebert, R., & Morris, L. (1967). Cognitive and emotional components of test anxiety: A distinction and some initial data. Psychological Reports, 20, Marsh, H. W., Trautwein, U., Lüdtke, O., Köller, O., & Baumert, J. (2006). Integration of multidimensional self-concept and core personality constructs: Construct validation and relations to well-being and achievement. Journal of Personality, 74, McIlroy, D., & Bunting, B. (2002). Personality, behaviour, and academic achievement: Principles for educators to inculcate and students to model. Contemporary Educational Psychology, 27, Morris, L. W., & Liebert, R. M. (1969). Effects of anxiety on timed and untimed intelligence tests: another look. Journal of Consulting and Clinical Psychology, 33, Petrides, K. V., & Furnham, A. (2000). On the dimensional structure of emotional intelligence. Personality and Individual Differences, 29, Rindermann, H., & Neubauer, A. (2001). The influence of personality on three aspects of cognitive performance: Processing speed, intelligence and school performance. Personality and Individual Differences, 30, Rhodewalt, F. (1990). Self-handicappers: Individual differences in the preference for anticipatory, self-protective acts. In R. L. Higgins (Ed.), Self-handicapping: The paradox that isn't New York: Plenum Press. Sarason, I. G. (1984). Stress, anxiety and cognitive interference: Reactions to tests. Journal of Personality and Social Psychology, 46, Smith, R. J., Arnkoff, D. B., & Wright, T. L. (1990). Test anxiety and academic competence: A comparison of alternative models. Journal of Counseling Psychology, 37, Spielberger, C. D., & Vagg, P. R. (1995). Test anxiety: A transactional process. In C. D. Spielberger & P. Vagg (Eds.), Test anxiety: theory, assessment, and treatment (pp. 3 14). Washington, DC: Taylor & Francis. Stankov, L. (1999). Mining on the "no man's land" between intelligence and personality. In P. L. Ackerman & P.C. Kyllonen (Eds.), Learning and individual differences: Process, trait, and content determinants (pp ). Atlanta, GA: Georgia Institute of Technology. Steiger, J. H. (1980). On the empirical equivalence of regression components and common factors. British Journal of Mathematical and Statistical Psychology, 33, Wechsler, D. (1997). The Wechsler Adult Intelligence Scale Third Edition (WAIS-III) San Antonio, TX: Psychological Corporation. Zeidner, M. (1995). Personality trait correlates of intelligence. In D. Saklofske & M. Zeidner (Eds.), International handbook of personality and intelligence: Perspectives on individual differences (pp ). New York, US: Plenum Press. Zeidner, M. (1998). Test anxiety: The state of the art New York: Plenum. Zimmer, J., Hocevar, D., Bachelor, P., & Meinke, D. L. (1992). An analysis of the Sarason (1984) four factor conceptualization of test anxiety. In K. A.Hagvet & T. B. Johnsen (Eds.), Advances in test anxiety research, 7, Zohar, D. (1998). An additive model of test anxiety: Role of exam-specific expectations. Journal of Educational Psychology, 90,