Personality traits and personality disorders.
Currently, the most authoritative reference for diagnostic guidelines (operational criteria) in psychiatry is the American Psychiatric Association's (1994) Diagnostic and Statistical Manual, 4th ed. (DSM-IV). Clinical syndromes, such as schizophrenia, depression, anxiety disorders and so forth, are diagnosed on Axis I of the five axes used by the DSM-IV. Personality disorders are diagnosed on Axis II. The implication of being detailed on Axis II is that the personality disorders express their effect continuously from late adolescence, and do not relapse and remit like typical illness syndromes. Therefore, personality disorder definitions portray long-standing, usually maladaptive, features of the person's behaviour that importantly affect personal and social functioning in a wide range of contexts. The DSM-III-R (American Psychiatric Association 1987) scheme recognised 3 broad clusters of personality disorder, containing altogether 11 specific types. Cluster A - of 'odd or eccentric' types - includes paranoid, schizoid and schizotypal personality disorder diagnoses. Cluster B - of 'dramatic, emotional or erratic' types - includes antisocial, borderline, histrionic and narcissistic personality disorders. Cluster C gathers together 'anxious and fearful' types with personality disorder diagnoses of avoidant, dependent, obsessive-compulsive and passive-aggressive types. Because it is recognized that few patients with dysfunctional personalities fit neatly into any one type, the DSM-III-R rules allow more than one diagnosis to be made on Axis II. Because DSM-IV made relatively minor changes to DSM-III-R personality disorder diagnoses, it has been reckoned that the current personality disorder diagnostic schemes will influence psychiatric thinking for at least the next 15-20 years (Zimmerman, 1994).
In order to convey accurate information about a person, the scheme which is used to classify personality disorders must be both reliable and valid. Zimmerman (1994) published a comprehensive review of interview and self-rating scales which have been designed to assess DSM personality disorders. He found that the reliability (inter-rater and test-retest) of even the most detailed current instruments for assessing personality disorders is modest at best. However, these are better than unstandardized clinical evaluations of patients' personalities, in which the reliabilities are poor to fair at best. Further, Zimmerman found that three widely used instruments vary in their coverage of the DSM-III-R personality disorders, and their content with respect to individual personality disorder concepts varies considerably. Zimmerman's review made it clear that even the most detailed instruments for assessing personality disorders lack adequate reliability and vary considerably in the extent to which they index DSM-III-R categories. Therefore, there is a widely recognized need for much further development of instruments to assess personality disorders.
The above comments on the inadequacies of current assessment instruments for personality disorders beg a more basic question: are the DSM personality disorders a valid or complete description of personality dysfunction? The answer must be no, because: (a) the number and types of personality disorders vary from revision to revision of the DSM scheme (Widiger & Costa, 1994); (b) there is considerable overlap between some personality disorders and Axis I diagnoses (Widiger & Shea, 1991); (c) important features of personality dysfunction are omitted from the DSMIII-R scheme and the range of dysfunction is too narrow, including only severe disorders (Tyrer, 1988); and (d) there is much co-morbidity among personality disorder diagnoses, e.g. many patients meet the criteria for both avoidant and dependent personality disorders, and many patients with a diagnosis of histrionic personality disorder will also meet the criteria for borderline and narcissistic personality disorders (Widiger, Trull, Hurt, Clarkin & Frances, 1987).
Two issues are being explored which might help to provide clinical psychologists, psychiatrists and other mental health professionals with more valid indices of personality disorders. The first of these is to move from categories to dimensions in describing personality disorders. Prominent researchers in the field have suggested that the syndrome approach, which is well suited to describing illnesses, is poorly adapted for describing the subtleties of individual differences in personality, and that it should be replaced with a dimensional approach, which would provide a series of scales which could be used to give a profile of a person's personality (Blackburn, 1988; Cloninger, Svrakic & Przybeck, 1993; Livesley, 1995; Schroeder, Wormworth & Livesley, 1992; Tyrer, 1988; Widiger & Shea, 1991).
The second issue is: which personality dimensions should be measured? Because the structure of personality disorder dimensions does not appear to be different between general population and psychiatric samples (Cloninger et al., 1993; Livesley, Jackson & Schroeder, 1992; Livesley, Jang, Jackson & Vernon, 1993), recent research in personality disorders has turned to the structure of personality as discovered by psychologists studying normal personality (Costa & Widiger, 1994; Svrakic, Whitehead, Przybeck & Cloninger, 1993). In fact, a considerable consensus has been reached within the area of differential psychology to the effect that normal personality may be described by three to five broad dimensions, viz. Neuroticism (N), Extraversion (E), Psychoticism; or N, E, Agreeableness, Openness and Conscientiousness (Deary & Matthews, 1993; Goldberg, 1993; McCrae & Costa, 1997). Recent dimensional schemes that were constructed to assess personality disorders (e.g. Livesley, Jackson & Schroeder, 1989) are moderately associated with scores on normal personality dimensions (Schroeder, Wormworth & Livesley, 1992, 1994; Widiger & Trull, 1992). Therefore, the key focus for current personality disorder research has been to construct personality dimensions rather than categories, and to check the validity of the dimensions against validated personality dimensions from the psychological study of normal personality.
In this vein O'Boyle (1995; O'Boyle & Holzer, 1992) has studied the associations between Eysenck's three broad personality dimensions and the DSM-III-R personality disorders as measured by self report using the Personality Disorder Questionnaire-Revised (Hyler, Reider, Williams, Spitzer, Hendler & Lyons, 1988). In two separate studies, one using data from medical students (O'Boyle & Holzer, 1992) and the other from patients in a substance abuse programme (O'Boyle, 1995), he found little statistical evidence for the DSM-III-R's three-cluster model of personality disorder categories. Moreover, in both of O'Boyle's studies, joint factor analysis of Eysenck's dimensions and scores on the individual personality disorders revealed a three-factor structure. Factor 1 had high loadings from neuroticism and all of the personality disorders except schizoid and antisocial. This factor appeared to be related to personal distress. Factor 2 in one study and factor 3 in the other had high loadings from psychoticism and antisocial personality disorder. Factor 3 in one study and factor 2 in the other had high negative loadings for extraversion and high positive loadings from schizoid personality disorder. O'Boyle (1995) commented that his findings were:
consistent with the idea that personality disorder traits are variants of basic personality traits, and that personality disorder classification may be understood in terms of traditional personality dimensions ... Specifically, the present results support the idea that Eysenck's dimensional components of personality can account for the various personality disorders (p. 564).
Mulder & Joyce (1997) conducted a factor analysis of the Structured Clinical Interview for DSMs (SCID) personality disorder categories in 148 psychiatric outpatients, and related the factors to Eysenck's and Cloninger's dimensions of normal personality, using the Eysenck Personality Questionnaire (EPQ) and the Tridimensional Personality Questionnaire (TPQ), respectively. They found four higher order factors among the 12 personality disorder categories which they named antisocial (high loadings from conduct disorder, borderline, narcissistic, histrionic and paranoid), asocial (schizoid), asthenic (avoidant, dependent and self-defeating) and anankastic (obsessive-compulsive). Antisocial was positively correlated with TPQ novelty seeking and EPQ Psychoticism; asocial was negatively correlated with TPQ reward dependence; asthenic was positively correlated with TPQ harm avoidance and EPQ Neuroticism; and anankastic was significantly but only modestly associated with TPQ reward dependence (negatively) and persistence (positively).
The present study
The aims of the present study were to examine further the structure of personality disorder traits and the associations between the Eysenck personality scheme and the DSM personality disorders. The objectives of the study were to investigate (a) the internal consistency of self-report personality disorder scales using the DSM-III-R-derived Structured Clinical Interview for DSM-III-R (SCID), SCID-II Personality Disorders Questionnaire; (b) the second-order structure of personality disorders to test the DSM's three-domain/cluster model; (c) the association of personality disorders and Eysenck's personality dimensions.
We recruited 400 first-year undergraduate students at the University of Edinburgh. Estimated compliance rate was above 95 per cent. Because participants were approached individually (see below), almost all took part.
Eysenck Personality Questionnaire-Revised (EPQ - R); Eysenck, Eysenck & Barrett, 1985). This has 100 true/false response items and assesses the personality traits of Neuroticism, Extraversion and Psychoticism. It has a Lie scale to detect socially desirable responding.
Structured Clinical Interview for DSM-III-R Personality Disorders Questionnaire (SCID-II; Spitzer, Williams, Gibbon & First, 1990). This has 113 no/yes response items which assess 12 categories of personality disorder: avoidant, dependent, obsessive-compulsive, passive-aggressive, self-defeating, paranoid, schizotypal, schizoid, histrionic, narcissistic, borderline and antisocial. One question was excluded in order not to offend participants: Number 106, 'Did you ever force someone to have sex with you?'.
Each potential participant was approached individually by one of the investigators (AP) in his/her room at halls of residence and asked to take part in the study. The questionnaire response formats were explained and participants were left alone for about one hour to complete the questionnaires. Questionnaires were collected by the investigator at the end of the hour and checked for completeness. Ethical permission for the study was obtained from the Halls of Residence Committee.
Internal consistencies of SCID-II scores
Most of the 12 personality disorder categories, when scored as dimensions, had mean scores that were not close to floor levels (Table 1). The exceptions were antisocial and schizoid personality disorders. The internal consistencies of the 12 categories ranged from the highly acceptable - borderline and antisocial both had [Alpha] [greater than] .7 - to low or unacceptable - schizotypal, dependent and obsessive-compulsive were at or below .5 and schizoid was the lowest at .29.
[TABULAR DATA FOR TABLE 1 OMITTED]
[TABULAR DATA FOR TABLE 2 OMITTED]
Higher order personality disorder clusters
Scores on the 12 SCID personality disorder categories for each participant were subjected to principal components analysis with varimax rotation. Because of the a priori DSM structure of three personality disorder clusters, three components were stipulated. All of the 12 categories except schizoid had moderate to high loadings on the first principal component (Table 1), which accounted for 34.8 per cent of the total variance. The first three principal components accounted for 55.0 per cent of the total variance. Orthogonal rotation did not alter the position of the factors from the principal components solution. Loadings above .3 on the second component were schizoid (.74), histrionic (-.56), avoidant (.43) and obsessive-compulsive (.34). Loadings above .3 on the third factor were antisocial (.71), dependent (-.49) and schizoid (.32). Speculatively, this solution suggests a large first factor in which there is some form of persistent worry, a second factor related to social involvement and a third factor related to personal empathy. This solution does not replicate the DSM clusters of personality disorder.
The oblique factor solution of the 12 personality disorder category scores offers a better fit to the DSM clusters. Factor 1 has highest loadings (all positive and above .6) for all of the cluster B personality disorder categories (antisocial, histrionic, narcissistic and borderline) and passive-aggressive (Table 1). Factor 2 has highest loadings (all positive and above .5) for all of the cluster C personality disorder categories except passive-aggressive (i.e. avoidant, obsessive-compulsive and dependent) and self-defeating. The third factor has loadings above .5 for antisocial, schizoid (both positive) and dependent (negative). However, although factor 3 was independent of factors 1 and 2 (correlations of -.01 and .16, respectively, factors 1 and 2 correlated at .65. Therefore, clusters B and C emerge to a degree, but are highly correlated, and cluster A does not emerge as a factor.
SCID-II personality disorders and Eysenck's personality traits
Associations between the 12 SCID-II personality disorder category scores and scores on Eysenck's personality factors are shown in Table 2. Neuroticism had positive correlations of .4 and above with seven of the SCID-II personality
disorder scales; especially high correlations obtained with borderline (.64) and avoidant (.53) disorders. Neuroticism had positive correlations between .2 and .4 with three other disorder scales. Extraversion had near to zero correlations with all but avoidant (-.45) and histrionic (.41) personality disorders. Psychoticism had moderate correlations with antisocial (.45) and passive-aggressive (.39) personality disorders, and positive correlations above .2 with narcissistic and borderline disorders. A principal components analysis was carried out including 11 of the SCID-II Personality Disorders Scale factors and the Eysenck personality factors. Schizoid personality disorder was omitted because of its unacceptably low internal consistency; it was included in the previous analysis to explore the factorial space of all scales with reference to the DSM clusters. The scree slope [ILLUSTRATION FOR FIGURE 1 OMITTED] suggested four components should be extracted. The loadings on the first unrotated component and the four-factor oblique rotation solution are shown in Table 2. In each of the first three factors there were EPQ-R factors with very high loadings, and it is notable that the Neuroticism and Extraversion factors' loadings were limited to a single factor; i.e. they 'defined' the factors. Psychoticism's loadings were split between rotated factors 2 and 4, with a very high loading (.71) on factor 2. Neuroticism loaded at .81 on factor 1, which also had high positive ([greater than] .5) loadings for avoidant, dependent, passive-aggressive, self-defeating, paranoid, schizotypal and borderline personality scores. Psychoticism loaded at .71 (and the EPQ-R Lie scale -.72) on factor 2 which also had high positive loadings for antisocial (.80), narcissistic (.46) and passive-aggressive (.44) scores. Borderline scores had moderate loadings on this factor (.22). Extraversion loaded .88 on factor 3, which had a high positive loading for histrionic (.71), and high negative loadings for avoidant scores (-.52). The variable with the largest loading on factor 4 was obsessive-compulsive, with psychoticism loading at -.46 and narcissistic loading at .38.
To aid understanding of the multivariate associations described above, the loadings of EPQ-R factors and SCID-II categories are displayed in Figs. 2a to 2f. Orthogonal axes are used because the obliquely rotated factors had very small correlations and the orthogonal and oblique solutions were almost identical. Figure 2a shows that the quadrant formed by positive values of factors 1 and 2 holds most of the SCID variables. That is, there appear not to be personality disorders associated with low Neuroticism or low Psychoticism. Dependent personality disorder is the only variable with a moderate negative loading on factor 2. Narcissistic and passive-aggressive have positions intermediate between the two factors and obsessive-compulsive is the only personality disorder with a near zero loading on both. In Fig. 2b many SCID variables cluster around the high Neuroticism pole of factor 1 and histrionic and avoidant have intermediate positions in the high Neuroticism-high Extraversion and high Neuroticism-low Extraversion quadrants respectively. Narcissistic takes an intermediate position between the high Neuroticism pole of factor 1 and the high obsessive-compulsive pole of factor 4 [ILLUSTRATION FOR FIGURE 2C OMITTED]. Narcissistic is the only SCID variable not near the axes of the bivariate space described by factors 2 and 3, and 2 and 4 [ILLUSTRATION FOR FIGURES 2D AND E OMITTED].
The results thus far have been descriptive. In view of the growing consensus around a four-factor model of personality disorders and the four factors found in analysis of the scree slope from the present data, confirmatory factor analysis was used to test a four-factor model of what others have called the four As: asthenic, asocial, antisocial and anankastic (Mulder & Joyce, 1997). The EQS structural equation modelling program (Bentler, 1995) was used to test the fit of a four-factor model to the present data set. The method of maximum likelihood was used to analyse the covariance matrix based on data from 400 participants. Starting values for loadings on the four stipulated factors were based on the results of the exploratory factor analysis described above. The factors were allowed to correlate (i.e. the solution was allowed to be oblique). The Wald and Lagrange multiplier tests were used to suggest paths in the model that might be dropped or added, respectively, to improve the fit of the model.
The resultant model is shown in Table 3. The fit of this model will be described first and then its meaning. A good model should account for most of the covariance in the data in the most economical way. In the present model, with 72 degrees of freedom, the average of the off-diagonal standardized residuals was 0.029, indicating that most of the covariance had been accounted for. Acceptable models typically have values less than 0.04. The [[Chi].sup.2] for the model is 135.6 (72, p [less than] .001). The lower the [[Chi].sup.2] the better is the fit of the model and ideally one looks for a nonsignificant [[Chi].sup.2], which would indicate that the covariance not accounted for by the model was not significantly different from zero. However, with large numbers of participants, as were recruited in the present study, this is rarely achievable, and most investigators rely on Loehlin's (1987) criterion that an acceptable [[Chi].sup.2] value is one that is less than twice the degrees of freedom. On this criterion the present model has an acceptable fit. A number of fit statistics, which take values between 0 and 1, indicate the fit of the model, with the values [greater than or equal to] .9 indicating well-fitting models. The model in Table 3 had the following fit values: Bentler-Bonett normed fit index = .936; Bentler-Bonett non-normed fit index = .954; and comparative fit index = .969. By all of these indices the present model fits well. In addition, all of the paths in the model were significant when their parameter estimates were divided by their standard errors. In summary, the model in Table 3 has comprehensively good fit statistics.
[TABULAR DATA FOR TABLE 3 OMITTED]
The principal importance of the model shown in Table 3 is that the associations among 15 different variables (4 related to normal personality and 11 related to personality disorder) are adequately captured in 4 latent traits. With a few minor differences, they are the same as those found in the exploratory principal components analysis. The first factor's highest loading is Neuroticism and it has significant loadings from eight of the SCID-II personality disorder variables. The second factor has high loadings from Psychoticism and the Lie (negative) scale, and moderately high loadings from antisocial, narcissistic and passive-aggressive personality disorder variables. The third factor's highest loading is Extraversion and it has high loadings from avoidant and histrionic personality disorders. The fourth factor has high loadings from narcissistic and obsessive-compulsive personality disorders and from Psychoticism (negative). It was stated earlier that the factor intercorrelations were free parameters in the model, and the following correlations were found: factor 1-factor 2 = .33; factor 1-factor 4 = .61; and factor 2-factor 3 = .22. The only path not yet indicated was a significant negative association between the error term associated with Neuroticism and Extraversion (r = -.31). Such an association is common, and in the present analysis was not captured by the four latent traits or their intercorrelations.
In this sample of undergraduates, scores on all but one of the DSM-III-R personality disorder categories showed acceptable distributions and internal consistencies. The exception was schizoid personality disorder, though dependent and obsessive-compulsive variables also had alpha values less than .5. The DSM personality clusters emerged to a limited degree, but there was better agreement with the suggestion that there are four broad sources of personality disorder variance. Thus, Mulder & Joyce (1997) found antisocial, asthenic, asocial and anankastic factors among the DSM-based personality disorder categories from the SCID-II interview as applied to psychiatric out-patients. It should be noted that the schizoid (asocial) and obsessive-compulsive (anankastic) 'factors' found by Mulder & Joyce did not represent variance shared by more than one SCID subscale. Comparable four-factor solutions for personality disorder variance were arrived at by Tyrer & Alexander (1979) and Schroeder & Livesley (1991). Moreover, Presly & Walton (1973) reported social deviance, submissiveness, obsessional-schizoid and hysteria factors, accounting for about 70 per cent of personality disorder variance, from interviews with a mixed group of psychiatric patients. In the present study, the exploratory, joint principal components analysis of the DSM personality disorder cluster scores and the Eysenck Personality Questionnaire-Revised factors showed a four-factor solution in which asthenic antisocial, asocial (reversing the loadings) and anankastic were arguably the obliquely rotated factors. Confirmatory factor analysis provided an acceptable model that supported such a four-factor structure.
This and other studies which have examined normal and abnormal personality traits together suggest that there is much common ground between the two (O'Boyle 1995; O'Boyle & Holzer, 1992; Mulder & Joyce, 1997). Well-validated normal personality traits such as the EPQ-R factors might act as a reference structure for personality disorder. Neuroticism, especially, has a large part to play in personality disorders. Extraversion defined a factor of social interaction, which involved the disorders of Narcissism (high social involvement) and Avoidant (low social involvement). As Eysenck intended, Psychoticism accounts for a substantial proportion of the variance in antisocial personality disorder (Eysenck & Eysenck, 1976). It was interesting to note the parts played by Eysenck's Lie scale in the various analyses. In practice, the Lie scale is rarely used as such, and it is well known that it has a substantial negative correlation with Psychoticism. Should the Lie scale be treated as a personality factor? Though it tends to covary negatively with Psychoticism, it sometimes accounts for more variance. For example, its negative correlation with the narcissistic personality disorder variable was higher than any of Psychoticism's correlations, suggesting that people with high narcissism scores were frank in their self-reports.
One feature of interest in the results is that narcissistic personality disorder uniquely takes an intermediate position in all the bivariate factor plots shown in Fig. 2. Both Tables 2 and 3 show narcissistic as having large loadings both on the factor associated with the low agreeableness aspect of P and on that tentatively identified with the low conscientiousness aspect of P. The latter is characterized by P(-), obsessive-compulsive(+) and narcissistic(+). Previous work on the association of narcissism with other traits is partly congruent with these results. A joint study of the NEO Personality Inventory and the Narcissistic Personality Inventory (Ramanaiah, Detwiller & Byravan, 1994) showed narcissism to be associated with high E and low A scores, though no significant difference between C scores for narcissistic and non-narcissistic groups was found. Associations between narcissism and obsessive-compulsive traits have previously been observed (Thiel & Schussler, 1995).
While the low agreeableness aspect of Narcissistic is reasonably intuitive, the interpretation of its appearance on factor 4 in Tables 2 and 3 and the association with obsessive-compulsive is more problematic and suggests a degree of complexity underlying the DSM-III-R narcissistic construct. Studies of other narcissism inventories have indeed suggested that it is a complex construct requiring a multidimensional description. In particular, it has been hypothesized that some aspects of narcissism (associated with an unrealistically positive self-image) may be adaptive provided they are not present to an excessive degree, whereas other aspects are more pathological (Digiuseppe, Robin, Szesko & Primavera, 1995; Hickman, Watson & Morris, 1996; Watson & Biderman, 1994). It appears from our results that the brief DSM-III-R scale succeeds in capturing some of this complexity; this result may be attributable to the extensive clinical knowledge entering into the development of these particular items. Studies of the associations between the DSM-III-R and other narcissism scales are required to clarify these associations.
As others have found, obsessive-compulsive appears largely alone out of the personality disorder factors in factor 4, with a moderate positive loading from narcissistic personality disorders and a negative loading from Psychoticism. This suggests that Eysenck's (1992) view of Psychoticism as both a low agreeableness and low conscientiousness factor might be correct. However, the present analyses have separated these two aspects of Psychoticism, with the low agreeableness facet segregating with antisocial personality disorder in factor 2 and, arguably, the low conscientiousness facet segregating with obsessive-compulsive traits in factor 4. Thus, the five-factor model, with its explicit separation of agreeableness and conscientiousness, might have performed better here by aligning obsessive-compulsiveness with high conscientiousness (Costa & McCrae, 1995). In the study by Schroeder, Wormworth & Livesley (1992), conscientiousness from the NEO-PI-R and compulsiveness from Livesley's DAPP-BQ personality disorder scales have very high loadings on a factor with almost no other substantial loadings.
The present study provides further evidence for a limited number of personality disorder dimensions. There has been consensus around five normal personality factors, namely Neuroticism (N), Extraversion (E), conscientiousness (C), agreeableness (A) and openness (O; with variants sometimes called intellect or culture) (Deary & Matthews, 1993). However, whereas the first two of these attract universal agreement, some argue that A and C are primary factors of the superfactor of Psychoticism (Costa & McCrae, 1995). A similar account may be given of personality disorder factors now that several studies have examined their nature in different groups using different instruments. There seem to be clear factors of personal distress (asthenic) and differences in this factor are related to Neuroticism. Antisocial traits emerge as a clear factor, related to Eysenck's Psychoticism and, in other studies, to the negative pole of the agreeableness factor of the five-factor model (Schroeder, Wormworth & Livesley, 1992). Thereafter, studies regularly identify an obsessive-compulsive (anankastic) factor. Though this is modestly related to Psychoticism (and not to the Lie scale), it is our hypothesis that this would be related most closely to the conscientiousness factor of the five factor model. The present study then identifies low extraversion as a major contributor to the asocial factor.
Thus far, it is interesting that our data are concordant with those studies which have examined overlaps between normal and abnormal personality traits, and with those which have sought a small number of latent traits to account for the covariance among the DSM's personality disorder variables. Whether self-report scales or others' ratings are used, and whether the population studied is clinical or nonclinical, a similar, small number of personality factors emerges. However, what are the limitations of the multivariate studies conducted to date? Livesley (personal communication) has suggested that pathological dependence might be underemphasized in present personality disorder schemes. The importance of such a factor has been noted by others (Blackburn, 1988; Deary, Bedford & Fowkes, 1995). The failure to incorporate such a factor in either the Eysenck or the five-factor model may explain the relatively small amount of SCID-II dependent personality disorder variance accounted for in the solutions described in the present study. A further problem is the low internal consistency of the schizoid personality disorder items. It can be speculated that schizoid should load highly on the asocial factor, at the other end of the scale from Extraversion, but empirical confirmation awaits a better clinical scale.
In summary, the SCID-II questionnaire has a tractable structure that has substantial overlaps with some major dimensions of normal personality. In addition, the present results on undergraduates are in agreement with the structure of personality disorder found in clinical populations using other methods of assessment. This adds to recent research which has found quantitative but not qualitative differences in the personality disorder traits found in normal and clinical samples (Livesley, Jackson & Schroeder, 1992). Because the SCID items have come from decades of accumulated clinical wisdom, they should not be replaced lightly. However, their factorial structure can be improved by psychometric methods and their associations with established traits of normal personality can build a bridge between two areas of research that have previously been separated.
American Psychiatric Association. (1987). Diagnostic and Statistical Manual of Mental Disorders, 3rd ed., revised (DSM-III-R). Washington, DC: American Psychiatric Association.
American Psychiatric Association. (1994). Diagnostic and Statistical Manual of Mental Disorders, 4th ed. (DSM-IV). Washington, DC: American Psychiatric Association.
Bentler, P. M. (1995). EQS Structural Equations Program Manual Encino, CA: Multivariate Software.
Blackburn, R. (1988). On moral judgements and personality disorders: The myth of the psychopathic personality revisited. British Journal of Psychiatry, 153, 505-512.
Cloninger, C. R., Svrakic, D. M. & Przybeck, T. R. (1993). A psychobiological model of temperament and character. Archives of General Psychiatry, 50, 975-990.
Costa, P. T. & McCrae, R. R. (1995). Primary traits of Eysenck's P-E-N system: 3- and 5-factor solutions. Journal of Personality and Social Psychology, 69, 308-317.
Costa, P. T. & Widiger, T. A. (1994). Personality Disorders and the Five Factor Model of Personality. Washington, DC: American Psychiatric Association.
Deary, I. J., Bedford, A. & Fowkes, F. G. R. (1995). The Personality Deviance Scales: Their development, associations, factor structure and restructuring. Personality and Individual Differences, 19, 275-291.
Deary, I. J. & Matthews, G. (1993). Personality traits are alive and well. The Psychologist, 6, 299-311.
Digiuseppe, R., Robin, M., Szesko, P. R. & Primavera, I. H. (1995). Cluster-analysis of narcissistic personality disorders on the MCMI-II. Journal of Personality Disorders, 9, 304-317.
Eysenck, H. J. (1992). Four ways five factors are not basic. Personality and Individual Differences, 13, 667-673.
Eysenck, H. J. & Eysenck, S. B. G. (1976). Psychoticism as a Dimension of Personality. London: Hodder & Stoughton.
Eysenck, S. B. G., Eysenck, H. J. & Barrett, P. (1985). A revised version of the psychoticism scale. Personality and Individual Differences, 6, 21-29.
Goldberg, L. (1993). The structure of phenotypic personality traits. American Psychologist, 48, 26-34.
Hickman, S. E., Watson, P. J. & Morris, R. J. (1996). Optimism, pessimism and the complexity of narcissism. Personality and Individual Differences, 20, 521-525.
Hyler, S., Reider, R., Williams, J., Spitzer, R., Hendler, J. & Lyons, M. (1988). The personality diagnostic questionnaire: Development and preliminary results. Journal of Personality Disorders, 2, 229-237.
Livesley, W. J. (1995). Past achievement and future directions. In W. J. Livesley (Ed.), The DSM-IV Personality Disorders. New York: Guilford.
Livesley, W. J., Jackson, D.N. & Schroeder, M.L. (1989). A study of the factorial structure of personality pathology. Journal of Personality Disorders, 3, 292-306.
Livesley, W. J., Jackson, D. N. & Schroeder, M. L. (1992). Factorial structure of traits delineating personality disorders in clinical and general population samples. Journal of Abnormal Psychology, 101, 432-440.
Livesley, W. J., Jang, K. L., Jackson, D. N. & Vernon, P. A. (1993). Genetic and environmental contributions to dimensions of personality disorder. American Journal of Psychiatry, 150, 1826-1831.
Loehlin, J. C. (1987). Latent Variable Models: An Introduction to Factor, Path and Structural Analysis. Hillsdale, NJ: Erlbaum.
McCrae, R. R. & Costa, P.T. (1997). Personality trait structure as a human universal. American Psychologist, 52, 509-516.
Mulder, R. T. & Joyce, P. R. (1997). Temperament and the structure of personality disorder symptoms. Psychological Medicine, 27, 99-106.
O'Boyle, M. (1995). DSM-III-R and Eysenck personality measures among patients in a substance abuse programme. Personality and Individual Differences, 18, 561-565.
O'BoyLe, M. & Holzer, C. (1992). DSM-III-R personality disorders and Eysenck's personality dimensions. Personality and Individual Differences, 13, 1157-1159.
Presly, A. S. & Walton, H. J. (1973). Dimensions of abnormal personality. British Journal of Psychiatry, 159, 463-471.
Ramanaiah, N. V., Detwiller, F. J. & Byravan, A. (1994). Revised NEO Personality Inventory profiles of narcissistic and nonnarcissistic people. Psychological Reports, 75, 512-514.
Schroeder, M. L. & Livesley, W. J. (1991). An evaluation of DSM-III-R personality disorders. Acta Psychiatrica Scandinavica, 84, 512-519.
Schroeder, M. L., Wormworth, J. A. & Livesley, W. J. (1992). Dimensions of personality disorder and their relationships to the Big Five dimensions of personality. Psychological Assessment, 4, 47-53.
Schroeder, M. L., Wormworth, J. A. & Livesley, W. J. (1994). Dimensions of personality disorder and the five factor model of personality. In P. T. Costa & T. A. Widiger (Eds), Personality Disorders and the Five Factor Model of Personality. Washington, DC: American Psychological Association.
Spitzer, R. L., Williams, J. B., Gibbon, M. & First, M. B. (1990). Structured Clinical Interview for DSM-III-R Personality Disorders. Washington, DC: American Psychiatric Press.
Svrakic, D. M., Whitehead, C., Przybeck, T. R. & Cloninger, C. R. (1993). Differential diagnosis of personality disorders by the seven-factor model of temperament and character. Archives of General Psychiatry, 50, 991-999.
Thiel, A. & Schussler, G. (1995). Obsessive-Compulsive symptoms in narcissistic self-system disturbances - An exemplary study in anorexia and bulimia nervosa. Zeitschrift fur Psychosomatische Medizin und Psychoanalyse, 41, 60-76.
Tyrer, P. (1988). What's wrong with DSM-III personality disorders? Journal of Personality Disorders, 2, 281-291.
Tyrer, P. & Alexander, I. (1979). Classification of personality disorder. British Journal of Psychiatry, 135, 163-167.
Tyrer, P. & Stein, G. (1993). Personality Disorder Reviewed. London: Gaskell/Royal College of Psychiatrists.
Watson, P. J. & Biderman, M. D. (1994). Narcissistic Traits Scale: Validity evidence and sex differences in narcissism. Personality and Individual Differences, 16, 501-504.
Widiger, T. A. & Costa, P. T. (1994). Personality and personality disorders. Journal of Abnormal Psychology, 103, 78-91.
Widiger, T. A. & Shea, T. (1991). Differentiation of Axis I and Axis II disorders. Journal of Abnormal Psychology, 100, 399-406.
Widiger, T. A. & Trull, T. J. (1992). Personality and psychopathology: An application of the five factor model. Journal of Personality, 60, 363-393.
Widiger, T. A., Trull, T. J., Hurt, S. W., Clarkin, J. & Frances, A. (1987). A multidimensional scaling of the DSM-III personality disorders. Archives of General Psychiatry, 44, 557-563.
Zimmerman, M. (1994). Diagnosing personality disorders. Archives of General Psychiatry, 51, 225-245.
|Printer friendly Cite/link Email Feedback|
|Author:||Deary, Ian J.; Peter, Alistair; Austin, Elizabeth; Gibson, Gavin|
|Publication:||British Journal of Psychology|
|Date:||Nov 1, 1998|
|Previous Article:||Arousal, erroneous verbalizations and the illusion of control during a computer-generated gambling task.|
|Next Article:||The role of subjective concerns and characteristics of the moral issue in moral considerations.|