See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/7216486

The Brief Core Schema Scales (BCSS): Psychometric properties

and associations with paranoia and grandiosity in non-clinical
and psychosis samples

Article  in  Psychological Medicine · July 2006

DOI: 10.1017/S0033291706007355 · Source: PubMed

CITATIONS READS

342 4,010

11 authors, including:

Ben Smith Elizabeth Kuipers

University of London King's College London
13 PUBLICATIONS   1,151 CITATIONS    410 PUBLICATIONS   27,667 CITATIONS   

SEE PROFILE SEE PROFILE

Paul E Bebbington Hannah Twiddy

University College London The Walton Centre NHS Foundation Trust
588 PUBLICATIONS   44,097 CITATIONS    18 PUBLICATIONS   837 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Directed acyclic graphs in epidemiological research in psychology View project

Economics and psychosis View project

All content following this page was uploaded by Paul E Bebbington on 29 May 2014.

The user has requested enhancement of the downloaded file.

Psychological Medicine, 2006, 36, 749–759. f 2006 Cambridge University Press
doi:10.1017/S0033291706007355 First published online 27 March 2006 Printed in the United Kingdom

The Brief Core Schema Scales (BCSS) : psychometric

properties and associations with paranoia and
grandiosity in non-clinical and psychosis samples
D A V I D F O W L E R 1 *, D A N I E L F R E E M A N 2, B E N S M I T H 3 , E L I Z A B E T H K U I P E R S 2 ,
P A U L B E B B I N G T O N 3, H A N N A H B A S H F O R T H 4, S I A N C O K E R 1, J O A N N E
H O D G E K I N S 1, A L I S O N G R A C I E 2, G R A H A M D U N N 4 A N D P H I L I P P A G A R E T Y 2
1
University of East Anglia, Norwich, UK; 2 Institute of Psychiatry, King’s College London, London, UK ;
3
University College London, London, UK; 4 University of Manchester, Manchester, UK

ABSTRACT
Background. Traditional instruments that measure self-esteem may not relate directly to the
schema construct as outlined in recent cognitive models. The Brief Core Schema Scales (BCSS) aim
to provide a theoretically coherent self-report assessment of schemata concerning self and others in
psychosis. The scales assess four dimensions of self and other evaluation: negative-self, positive-self,
negative-other, positive-other.
Method. We analysed the psychometric properties of the BCSS using a sample of 754 students
recruited by email and 252 people with psychosis recruited as part of a trial of cognitive therapy.
We report the internal consistency, stability and the factor structure of the scale, and the association
of the BCSS with measures of self-esteem and with symptoms of paranoia and grandiosity.
Results. The BCSS have good psychometric properties and have more independence from mood
than the Rosenberg Self-Esteem Schedule. People with chronic psychosis reported extreme negative
evaluations of both self and others on these scales, but their levels of self-esteem and positive
evaluations of self and others were similar to the student sample.
Conclusions. Extreme negative evaluations of self and others appear to be characteristic of the
appraisals of people with chronic psychosis, and are associated with symptoms of grandiosity and
paranoia in the non-clinical population. The BCSS may provide a more useful measure of schemata
about self and others than traditional measures of self-esteem.

INTRODUCTION emotional responses (Gilbert, 1992 ; Brown et al.

1995). Many recent cognitive theories of both
The way in which people make judgements
psychotic and non-psychotic psychiatric dis-
or evaluate themselves and other people may
orders suggest that moment-to-moment evalu-
be a central process mediating their adaptation
ations of self and others may be linked to current
to the social world. Such evaluations probably
external events, current mood and cognitive
form part of a basic human response to social
state, and to memory processes in the form of
stresses and threats, and thus may be associated
schema that synthesize past reactions (Beck,
with the complex range of biological, cognitive,
and behavioural reactions that characterize 1976; Gilbert, 1992; Teasdale & Barnard,
1993). Such models can thus provide a coherent
psychological explanation for the way in which
* Address for correspondence : Professor David Fowler, School of past and current social stresses and emotional
Medicine, Health Policy and Practice, Elizabeth Fry Building,
University of East Anglia, Norwich NR4 7TJ, UK. reactions are mediated by ongoing appraisals.
(Email : D.fowler@uea.ac.uk) Such constructs were originally associated most
749
750 D. Fowler et al.

closely with theories of depression (Beck, 1976 ; also of positive evaluations of self and positive
Teasdale & Barnard, 1993 ; Clark et al. 1999), and negative evaluations of others. Different
but have also been incorporated into current aspects of self and other evaluation may be
theories of PTSD (Foa & Rothbaum, 1998 ; associated with different types of psychotic
Dalgleish, 1999 ; Ehlers & Clark, 2000; Brewin, symptom (Fowler et al. 1995; Fowler, 2000;
2003) and eating disorders (Fairburn et al. Garety et al. 2001). Several cognitive accounts
2003). Negative evaluations of self and others of paranoia emphasize that the evolution of
have also been incorporated into models of negative evaluations of others (e.g. evaluations
paranoia and persecutory delusions (Chadwick that others are untrustworthy, bad, devious)
et al. 1996; Bentall et al. 2001; Freeman et al. may be as important in understanding paranoia
2002 ; Bentall & Swarbrick, 2003), halluci- either alone or in combination with evaluations
nations (Close & Garety, 1998 ; Fowler, 2000 ; of self (Trower & Chadwick, 1995 ; Fowler,
Beck & Rector, 2003) and delusions (Fowler, 2000 ; Garety et al. 2001). However, grandiose
2000 ; Garety et al. 2001). There is still con- delusions may represent exaggerated positive
siderable debate about the exact nature and evaluations of self (Fowler et al. 1995). We
significance of the schema construct and dif- therefore designed the new instrument to assess
ferent cognitive theories may have competing positive and negative evaluations of self and
hypotheses. Further research in this important positive and negative evaluations of others. The
area would be advanced by the development of negative evaluation of self component of the
a more specific assessment tool. present scale derives from the Depressed States
In a recent study, Barrowclough et al. (2003), Checklist (Teasdale & Dent, 1987 ; Teasdale &
using a detailed interview assessment of self- Cox, 2001), and therefore has item content
esteem, showed specific and differential associ- which reflects an already well established con-
ations between negative personal evaluations cept of negative schematic models of self. The
(described as negative character deficits) and scale also has conceptual similarities to the
positive psychotic symptoms. Other types of existing Evaluative Beliefs Scale (Chadwick
evaluations of self were not associated with et al. 1999) ; however, while the negative self
positive psychotic symptoms. This study pro- dimension of both scales is similar, the negative-
vides support for the use of a measure of self and other scale of the Brief Core Schema Scales
other evaluation that is focused on a specific (BCSS) is different, as it represents appraised
aspect of self-appraisal, namely extreme nega- threat from others (e.g. others are untrust-
tive evaluations of self which may be regarded worthy, devious) rather than appraised hostility
as an assessment of negative self-schema. How- or dominance (e.g. other people are worthless,
ever, the self-esteem interview is very detailed, inferior) as assessed in the Evaluative Beliefs
and this precludes its use in large-scale clinical Scale. The present scale is also the only existing
studies. scale to assess positive evaluations of self
A variety of existing questionnaires have been and others in combination with negative evalu-
used to assess self-esteem and self-appraisals in ations of self and others. The resultant four-
psychosis (e.g. Rosenberg, 1965 ; Robson, 1988 ; dimensional structure is consistent theoretically
Young, 1990 ; Brown et al. 1995 ; Schmidt et al. with the widely used four-group attachment
1995 ; Freeman et al. 1998; Perris et al. 1998 ; styles framework proposed by Bartholomew &
Beck et al. 2001) ; however, few of these instru- Horowitz (1991). This may be an advantage in
ments provide a direct assessment of the nega- developing hypotheses linking schema in psy-
tive self-evaluation construct consistent with chosis with attachment theory in future work.
contemporary schema constructs as applied to In the present study we report basic psycho-
psychosis. Furthermore some, such as the metric properties of the BCSS in non-clinical
Rosenberg Self-Esteem Schedule (RSES), have and clinical samples, including internal consist-
outdated norms and poor psychometric ad- ency, stability, item factor analysis, and con-
equacy (Keith & Bracken, 1996). Our aim was current reliability with the RSES and subscales
to develop a scale that would provide a quick of Young’s Schema Questionnaire (YSQ). To
and easily administered measure of negative assess the validity of the scale we tested hypoth-
evaluations of self (negative self-schema) but eses implying differential associations between
 Psychometric properties of the Brief Core Schema Scales 751

self and other evaluations and the occurrence UK multi-centre randomized controlled trial
of different types of symptoms in the student of cognitive behaviour therapy and family
sample. In accordance with the predictions of intervention for psychosis, designed to answer
our cognitive model of psychosis (Fowler et al. questions both about outcome and the psycho-
1995 ; Fowler, 2000 ; Garety et al. 2001) we logical processes associated with psychosis.
hypothesized that paranoia would be associated Studies of psychological processes in psychosis
with a combination of negative beliefs about are incorporated into the baseline assessment of
self, negative beliefs about others, and anxiety, participating patients before randomization into
whereas grandiosity would be associated with the trial takes place. The trial was located in
positive evaluations of self and others and four NHS trusts in London and East Anglia :
not with negative evaluations of others. We the South London and Maudsley NHS Trust ;
also describe a comparison of norms in the Norfolk Mental Health NHS Trust ; Camden &
student sample compared with a sample of 252 Islington Mental Health and Social Care NHS
cases with chronic psychotic illness who were Trust ; and North East London Mental Health
recruited for a trial of cognitive behavioural NHS Trust. Recruitment was from specified
therapy. We hypothesized that the students and clinical teams, both in-patient and out-patient
cases with psychotic illness would show similar services, within each of these trusts. Patients
levels of self-esteem on the RSES and positive were recruited at the time of a relapse into
evaluations of self and others, but that the positive symptoms. The inclusion criteria were :
sample of people with psychotic illness would current diagnosis of non-affective psychosis
specifically show extreme negative evaluations (schizophrenia, schizoaffective psychosis, delu-
of self and others in comparison to the students. sional disorder ; ICD-10 F20) ; age 18–65 years ;
second or subsequent episode starting not
more than 3 months before individuals consent
METHOD
to enter the trial ; and rated at least 4 (moderate
Non-clinical sample severity) on the Positive and Negative Syn-
Eight hundred and fifty-six students were drome Scale (PANSS) on at least one positive
recruited from King’s College London and psychotic symptom at the first time of meeting.
University College London. Participants were
emailed the address of a website where they Measures
could complete the questionnaires online. Sixty-
five per cent were female, 35 % male, 70 % were The Brief Core Schema Scales
White, 17 % Asian, 2 % Afro-Caribbean, 1 % The BCSS have 24 items concerning beliefs
African and 10 % of other ethnic origin. Mean about the self and others that are assessed on a
age was 23.6 years (S.D.=6.5). In all, 754 five-point rating scale (0–4). Four scores are
students completed all items, which included obtained : negative-self (six items), positive-
version A and version B of the BCSS. A further self (six items), negative-others (six items)
257 students were recruited from University of and positive-others (six items) (see Appendix).
East Anglia and King’s College London, to The content of the negative-self component
undertake a study of test–retest reliability of the derived from self-devaluative words used by
Schema questionnaires online. These partici- Teasdale & Dent (1987) and subsequently by
pants repeated the schema questionnaires 3 Teasdale & Cox (2001) as globally negative self-
weeks apart (mean 21 days, S.D.=7 days). descriptors of personality. The other items were
generated by the first two authors of this paper
Psychosis sample on the basis of clinical experience with paranoia.
A sample of 252 people with psychosis was In the original version A (see Appendix) (used in
drawn from the cohort of patients recruited for both the clinical and non-clinical samples), the
the Psychological Prevention of Relapse in individual was asked to indicate in a dichot-
Psychosis (PRP) Trial (ISRCTN83557988). All omous No/Yes format whether they held each
those who completed the Schema questionnaire belief. Then, if they held the belief (i.e. answered
at the initial baseline assessment constitute yes), they were asked to indicate their degree
the sample for this study. The PRP Trial is a of belief conviction by circling a number from
752 D. Fowler et al.

1 to 4 (believe it slightly, believe it moderately, Higher scores indicate greater endorsement of a

believe it very much, believe it totally). schema.
In the non-clinical sample only, we assessed
two versions of the scale. In version B, the order Depression Anxiety Stress Scales (DASS ;
of the items was randomized and a simple 0–4 Lovibond & Lovibond, 1995)
Likert scale was used (there was no Yes/No We used the depression and anxiety scales of
response in version B). Scores on both scales this measure. Each scale consists of 14 items
were similarly distributed and the two versions with a 0–3 scale (0=did not apply to me at all,
of the scale also correlated very highly. For the 3=applied to me very much). Higher scores
positive-self, negative-self, positive-other, and indicate higher levels of emotional distress.
negative-other scales, the correlations were 0.75, Lovibond & Lovibond (1995) report results of
0.74, 0.70, and 0.61 respectively. Furthermore, a comparison of the DASS with the Beck
the pattern of results from the two scales in Depression (BDI) and Anxiety (BAI) inven-
terms of internal consistency, factor analysis, tories in a non-clinical sample (n=717). The
and associations with demographics and psy- DASS anxiety scale correlated at 0.81 with the
chopathology were very similar. In general, the BAI and the DASS depression scale correlated
associations between the BCSS and other at 0.74 with the BDI. The authors report
measures were slightly stronger in all analyses Cronbach’s alpha internal consistency coef-
for version A. We therefore only report the ficients of 0.91 and 0.81 for the DASS de-
results of the analysis of version A, as this pression and anxiety scales respectively. A
version had a clearer dimensional structure, and principal components analysis showed a good
was the version used in both the clinical and distinction between the depression and anxiety
non-clinical studies. The measure is very brief, factors of the DASS. The scale has also recently
taking only an average of 1 minute 25 seconds been shown to be reliable and valid in a
(S.D.=4 seconds) to complete. There was no large UK non-clinical population.
association between the time to complete the
measure and the scores. Paranoia Scale (PS ; Fenigstein & Vanable,
Rosenberg Self-Esteem Scale (Rosenberg, 1965) 1992)
The RSES is a 10-item, four-point scale (1–4) The 20-item self-report Paranoia Scale was
that assesses current levels of global self-esteem. developed to measure paranoia in college stu-
The total score is usually used. A high total dents, and includes items assessing both ideas of
score is indicative of low global self-esteem. persecution and reference. The Paranoia Scale is
The Rosenberg self-esteem scale is probably the the most widely used dimensional measure of
most widely used self-esteem measure, although paranoia and was derived from items in the
it has been criticized recently over its psycho- Minnesota Multiphasic Personality Inventory
metric properties (Keith & Bracken, 1996). (MMPI). Fenigstein & Vanable (1992) report
good internal consistency (based on a total
Young’s Schema Questionnaire – Short Version student sample of 581) and a Cronbach’s alpha
(YSQ-S; Young, 1998) of 0.84 and test re-test reliability (based on a
The YSQ-S is a 75-item self-report inventory 6-month re-test, n=107) of 0.70, indicating
designed to assess 15 types of core schematic good stability over time. The authors report
beliefs. The YSQ-S derives from the original good norms on their student sample, with a
YSQ (305 items). We did not administer mean score of 42.7 (range 20–100).
the whole YSQ-S because of its length. Instead
six key schemas that we predicted would The Peters et al. Delusions Inventory (PDI ;
relate to subscales of the BCSS were selected. Peters et al. 1999)
These were: mistrust/abuse, social isolation ; Two items assessing grandiosity were used from
defectiveness/shame ; failure ; self-sacrifice ; and this measure of delusional ideation. The two
entitlement. Each schema is assessed by five items were : (1) Do you ever feel as if you are or
items, each rated on a six-point Likert scale are destined to be someone very important ? and
from 1 (strongly disagree) to 6 (strongly agree). (2) Do you ever feel that you are a very special
 Psychometric properties of the Brief Core Schema Scales 753

or unusual person ? Each item was rated on the Table 1. Means and standard deviations for
following scale : 0 ‘never feel it ’, 1 ‘feel it rarely ’, all variables in the non-clinical sample
2 ‘feel it sometimes ’, 3 ‘feel it often ’, 4 ‘feel it
most of the time ’, 5 ‘feel it all the time ’. A total Min. Max. Mean S.D.

score from the two items was used, with higher Age (n=747) 16 65 23.70 6.66
scores indicating higher levels of grandiosity. RSES total (n=753) 10 39 20.85 6.04
DASS anxiety total (n=753) 0 41 7.05 7.28
The questionnaires were presented on the DASS depression total (n=753) 0 42 9.62 9.46
website in the following order : BCSS question- Paranoia scale total (n=753) 20 91 41.14 14.60
naire (0–4 Likert scale), RSES, YSQ subscales, PDI grandiosity total (n=753) 0 10 4.55 2.55
BCSS positive-self (n=754) 0 20 10.20 4.23
DASS, PS, and the PDI grandiosity items. BCSS negative-self (n=754) 0 23 3.55 3.55
BCSS positive-other (n=754) 0 22 10.43 4.51
Statistical analysis BCSS negative-other (n=754) 0 22 4.07 4.04

Analyses were conducted using SPSS for RSES, Rosenberg Self-Esteem Schedule ; DASS, Depression
Windows (version 11.0 ; SPSS, 2003). All vari- Anxiety Stress Scales ; PDI, Peters et al. Delusions Inventory ; BCSS,
ables were approximately normally distributed Brief Core Schema Scales.
except for negative self-evaluation, which was
positively skewed. Analyses were conducted
using both log-transformed negative self- within the multiple regression in order to take
evaluation scores and the original scores. As the account of the number of significance tests
transformation did not alter results appreciably, undertaken.
the results for the analyses using the original
scores are reported. Significance test results are RESULTS
quoted as two-tailed probabilities. Reliability
analysis was conducted by first assessing alpha Sample description
coefficients, and item whole correlations. A Table 1 provides the means and standard
principal components analysis with varimax deviations for variables in the non-clinical
rotation is reported for the BCSS. As the com- sample. There was a preponderance of females
ponents could be assumed to be correlated, both (65 %). The levels of depression, anxiety, para-
oblique and orthogonal rotations were used. noia and grandiosity were as expected for a non-
However, similar solutions were found, and the clinical group.
orthogonal solution using varimax rotation is The mean age of the psychosis group was 38
reported for ease of interpretation. Associations years (S.D.=10). There were more male (n=181 ;
between the BCSS and the other scales and 72%) than female (n=71; 28%) participants.
measures of psychopathology were examined, The group comprised the following ethnicities:
first using bivariate Pearson correlations. This White (n=187), Black-African (n=19), Black-
was followed by standard multiple regression Caribbean (n=19), Indian (n=5), other (n=
analyses predicting paranoia and grandiosity as 19). The diagnoses were schizophrenia (n=217),
a function of other emotions and schema. The schizoaffective disorder (n=35), and delusional
aim was to assess the degree to which the BCSS disorder (n=3). Of the sample, 88 % were single
explained variance in paranoia and grandiosity or divorced and 30 were married (12 %). The
independently of depression and self-esteem. mean length of illness was 11.7 years (S.D.=
Where independent variables have a bivariate 10.1), and 82% were unemployed, with small
correlation but lack significance in the re- numbers in full-time employment (6 %). Fifty-
gression, their relationship with the dependent five per cent of the sample reported persecutory
variable can be assumed to be made redundant delusions, 57 % auditory hallucinations, and
or mediated by the other variables. Assessments 17% grandiose delusions.
of significance and significance of semi-partial
correlations are therefore equivalent to forcing Internal consistency
entry of a series of other variables in the final The alpha coefficients for version A of the
step in stepwise multiple regression. We chose a positive- and negative-self-schema scales for
conservative significance level of p<0.001 to the non-clinical and clinical samples respectively
assess the significance of independent variables were : 0.78 (0.79) and 0.86 (0.84), and, for the
754 D. Fowler et al.

Table 2. Dimensionality of the Brief Core Schema Scales : initial eigenvalues from a principal
components analysis
Non-clinical sample (n=752) Chronic psychosis sample (n=252)

Component Total % of variance Cumulative % Total % of variance Cumulative %

Negative-other 7.2 29.9 29.9 6.1 25.5 25.5

Positive-other 3.2 13.4 43.3 3.8 15.8 41.3
Positive-self 2.0 8.5 51.7 2.6 10.8 52.2
Negative-self 1.3 5.3 57.0 1.4 5.9 58.0

Table 3. Component loading in the Brief Core Schema Scales

Component 1 : Component 2 : Component 3 : Component 4:
negative-others positive-others positive-self negative-self

Non- Chronic Non- Chronic Non- Chronic Non- Chronic

clinical psychosis clinical psychosis clinical psychosis clinical psychosis
BCSS item sample sample sample sample sample sample sample sample

I am unloved 0.21 — x0.10 x0.28 x0.16 0.28 0.54 0.58

I am worthless — — x0.15 — x0.30 — 0.69 0.81
I am weak 0.21 0.23 — — x0.24 x0.25 0.61 0.62
I am vulnerable 0.26 0.45 x0.15 — x0.11 — 57 0.42
I am bad 0.30 — — — x0.17 — 0.48 0.69
I am a failure — 0.28 — — — x0.37 0.67 0.68
I am respected — — 0.28 0.47 0.68 0.51 x0.22 —
I am valuable — — 0.19 0.3 0.78 0.75 x0.14 —
I am talented — — — — 0.79 0.75 — x0.2
I am successful — — 0.23 — 0.68 0.73 x0.27 —
I am good — — 0.16 0.28 0.57 0.53 x0.23 x0.37
I am interesting — — 0.14 — 0.77 0.8 — x0.21
Others are hostile 0.77 0.75 x0.14 — — — 0.11 —
Others are harsh 0.78 0.76 x0.12 — — — 0.14 —
Others are unforgiving 0.72 0.67 x0.14 — — — 0.11 0.3
Others are bad 0.78 0.69 — — — — — —
Others are devious 0.72 0.87 x0.21 — — — 0.17 —
Others are nasty 0.81 0.8 x0.12 — — — 0.11 —
Others are fair — — 0.76 0.72 0.12 — — —
Others are good x0.13 — 0.73 0.78 0.16 — — —
Others are trustworthy x0.19 — 0.78 0.71 0.15 — x0.13 —
Others are accepting x0.13 — 0.70 0.69 0.19 0.32 x0.10 —
Others are supportive x0.1 — 0.69 0.72 0.27 — x0.11 —
Others are truthful x0.14 x0.3 0.77 0.70 0.16 — x0.18 —

Factor items in bold.

positive- and negative-other schema scales : 0.88 four-component solution in Table 3. A four-
(0.84) and 0.88 (0.87) respectively. All item total component solution was indicated by the scree
correlations were highly significant (p<0.001). plot and eigenvalues. This solution accounted
The median item total correlation was 0.74 for 57% and 58 % of the variance in the
(0.76). non-clinical and clinical samples respectively.
The four-component solution corresponded to
Dimensionality negative-other, positive-other, positive-self, and
A principal components analysis with varimax negative-self subscales in both samples.
rotation was performed on the combined
scale of 24 items for both clinical and non- Concurrent and discriminant validity
clinical samples. Initial eigenvalues from the Table 4 shows bivariate Spearman’s correlations
analysis are shown in Table 2, and the rotated between BCSS subscales and the YSQ subscales
varimax component loadings matrix for a and RSES in the non-clinical sample. The
 Psychometric properties of the Brief Core Schema Scales 755

Table 4. Spearman’s bivariate correlations between Brief Core Schema subscales and the Young’s
Schema Questionnaire subscales and Rosenberg Self-Esteem Schedule in non-clinical sample
Positive-self Negative-self Positive-other Negative-other

YSQ mistrust/abuse x0.32 0.40 x0.51 0.49

YSQ social isolation x0.39 0.45 x0.41 0.28
YSQ defective/shame x0.53 0.44 x0.35 0.20
YSQ failure x0.24 0.15 x0.17 0.18
YSQ self-sacrifice 0.06, N.S. 0.06, N.S. x0.16 0.16
RSES total score x0.59 0.53 x0.39 0.23
Depression x0.45 0.63 x0.41 0.29
Anxiety x0.33 0.5 x0.34 0.30

YSQ, Young’s Schema Questionnaire ; RSES, Rosenberg Self-Esteem Schedule.

p<0.001 unless specified non-significant (N.S.).

BCSS had moderate to strong associations with Associations with paranoia and grandiosity
the RSES and with the defectiveness/shame, (non-clinical sample)
mistrust/abuse and social isolation Young sub- A standard multiple regression was undertaken,
scales. There were only very low associations with paranoia as the dependent variable and
between the BCSS and Young’s failure and depression, anxiety, the RSES and the BCSS
self-sacrifice scales. In the psychosis sample as the explanatory variables. The adjusted
the RSES and negative- self and positive-self R square was 0.51. R was highly significant
subscales had a moderate to strong associ- [F(7, 744)=110; p<0.0001]. Four variables
ation (r=0.64 and 0.65, p<0.001 respectively). contributed significantly to the prediction of
Correlations between the RSE and BCSS paranoia. These were, as predicted, in order of
negative-other and positive-other scales in the importance, negative-other (bivariate r=0.57 ;
clinical sample were weak (r=0.2 and x0.26, sr2=0.31 ; t=12.2 ; p<0.0001), positive-other
p<0.001, respectively). (bivariate r=x0.49; sr2=x0.2; t=x7.7 ; p<
0.0001), anxiety (bivariate r=0.48 ; sr2=0.13 ;
Stability t=5.2 ; p<0.0001) and negative-self (bivariate
The stability of the scales was assessed in the r=0.5 ; sr2=0.09 ; t=1.9; p<0.0001). This sug-
non-clinical sample by test-retest reliability gests that the negative-other scale explains
of scales taken by a sample of 257 students on approximately 9% of the variance in paranoia,
two occasions approximately 3 weeks apart independently of the other variables. Low
(mean 21 days, S.D.=7). Test-retest reliability positive-other scores may account for another
(Pearson’s r) for the negative-self, positive-self, 4% and anxiety another 2% respectively.
negative-other and positive-other scales re- Around 30% of the variability in paranoia was
spectively were r=0.84, 0.82, 0.7 and 0.72 (all explained by common variance between the
significant p<0.001). variables. Depression and the Rosenberg self-
esteem scores contributed no unique variance to
Associations with demographic variables paranoia, and apparent bivariate associations
There were no associations between the BCSS between depression (0.48) and the Rosenberg
and age, sex or years in education in the non- self-esteem scores (0.4) and paranoia were made
clinical sample. In the sample of people with redundant by or mediated by other variables.
psychosis there was no association between age, A standard multiple regression undertaken
sex or length of illness and the BCSS scales. with grandiosity as the dependent variable, and
However, the subgroup of 30 married people anxiety, depression, paranoia, the Rosenberg
with psychosis tended to have more extreme self-esteem schedule and the BCSS as explana-
negative self-evaluation than unmarried people tory variables provided an adjusted R square
(t=x2.9, df 207, p<0.005) and less positive of 0.3. R was highly significant (F=40.3 ; p<
self-evaluation than unmarried people (t=2.2, 0.001). Two variables contributed significantly
df 206, p<0.05). to grandiosity at the p<0.001 level. These were
756 D. Fowler et al.

positive-self (bivariate r=0.4; sr2=0.37 ; t=

total/non-
clinical
RSES

20.8
6.0
12.1; p<0.001) and paranoia (bivariate r=0.23;

753

10
39
16
21
25
29
sr2=0.22 ; t=7.2; p<0.001). Two other vari-
ables had weak trends, suggesting a small degree
of unique association with grandiosity. These
total/psychosis

were positive-other (bivariate r=x0.02 ; sr2=

sample
RSES

23.5
6.4 0x0.08 ; t=x2.7 ; p<0.06) and the RSES
10
40
19
23
28
31
260

(bivariate r=x0.2 ; sr2=x0.08 ; t=x2.4 ; p=

Comparisons between clinical and non-clinical samples on BCSS subscales and RSES

0.014). Fourteen per cent of the variance was

explained by positive-self, a further 5 % by
other/non-

paranoia, and less than 1% uniquely by the

Negative

clinical

0.75
4.0

positive-other scores and the Rosenberg. Ten

22

10
754

4
0

3
6

per cent of the variance was common. As pre-

dicted, neither negative-self nor negative-other
scores contributed to grandiosity.
other/psychosis
Negative

BCSS, Brief Core Schema Scales ; RSES, Rosenberg Self-Esteem Schedule.

sample

9.1
6.8

8.5
14.0

Comparison of scores in the non-clinical and

252

3
24

19

psychosis samples
Table 5 shows means, standard deviataions and
various centile scores for the psychosis and non-
other/non-

clinical samples. As expected there are clear

Positive

clinical

10.4
4.5

11.0
13.0

indications that the sample of people with

22

16
754

psychosis had extremely negative evaluations of

themselves and others. The mean score of 7.2 in
the sample of people with psychosis reflects the
other/psychosis

fact that moderate to strong endorsement of

Positive

sample

10.3
6.0

10.0
15.0

beliefs that ‘I am bad, weak, worthless’ etc. was

24

19
252

common. Endorsement of negative beliefs about

others (e.g. that others were ‘hostile ’, ‘harsh ’,
Non-clinical results in bold.

‘unforgiving ’, ‘nasty ’, etc.) was also common in

the psychosis sample. However, endorsement of
Negative
self/non-
clinical

3.5
3.5

3.0
5.0

such beliefs in the student sample was rare. As

23
754

predicted there were no significant differences

in positive evaluations of self and others or in
self-esteem as measured by the Rosenberg self-
self/psychosis
Negative

sample

esteem schedule. Scores in the non-clinical and

7.2
5.9

6.0
11.0
252

3
24

16

clinical samples in terms of positive evaluations

of self were similar [10.2 (S.D.=6.4) v. 9.4 (S.D.
=6.9)] respectively, and there was also no
difference in the positive evaluations of others
self/non-
Positive

clinical

scores [10.3 (S.D.=6) v. 10.2 (S.D.=6.2)] re-

10.2
4.2

11.0
13.0
754

8
20

15
Table 5.

spectively. These scores do not therefore indi-

cate a lack of positive views of self and others
amongst the psychosis group. Scores on the
self/psychosis

RSES between the psychosis sample (23.5, S.D.

Positive

sample

10.3
6.4

10.0
14.5

=6.4) and non-clinical samples (20.8, S.D.=6)

0

5
24

19
252

were also in a similar range.

25th centile
50th centile
75th centile
90th centile

DISCUSSION
Maximum
Minimum
Mean

Our aim in developing this scale was to create a

S.D.

quick and easy-to-use assessment of the type of

n
 Psychometric properties of the Brief Core Schema Scales 757

extreme positive and negative evaluations of differences between groups as the samples differ
self and others that we have observed clinically in many variables such as age, education and
to be typical of people with psychosis. The social circumstances, as well as symptom pres-
scales we have developed have good internal ence. However, the differential findings were as
consistency, and the principal components predicted. What was observed was that the
analysis suggested an underlying dimensional psychotic sample had similar levels of self-
structure that reflects independence between esteem and positive evaluations of self and
different dimensions of self and other evalu- others to the student sample, but showed very
ation. There were differential associations extreme negative evaluations of self and others.
between the BCSS and measures of paranoia A degree of independence between assessments
in the non-clinical sample, as required by of core schema and mood is suggested by
our cognitive model of psychotic symptoms associations between the BCSS and paranoia
(Fowler, 2000; Garety et al. 2001; Freeman and grandiosity remaining when assessments of
et al. 2002). As we predicted, negative-other mood were controlled in the regression analy-
evaluations were strongly and uniquely associ- sis. This may be an advantage in future studies.
ated with paranoia in the non-clinical sample, Further studies of the psychometric properties
in combination with negative-self scores and of the BCSS are needed. The preliminary
anxiety. Paranoia in the non-clinical sample studies reported here are promising but not
therefore occurs in a specific context of anxiety definitive. The dimensional structure appears
toned with evaluations of interpersonal threat robust, providing similar components in both
and personal vulnerability. The role of negative the non-clinical sample and the sample of
self-evaluation in association with paranoia acutely ill chronic patients. However, there is a
is consistent with the findings reported in a particular need to examine the performance of
companion paper (Smith et al. unpublished the BCSS in other clinical and matched non-
observations) from an analysis of the clinical clinical samples, and with recovered as well as
sample. Also as predicted, the positive-self score actively psychotic samples. Investigations of the
strongly and uniquely predicted grandiosity in stability of the tool in clinical groups are also
the non-clinical sample. required.
The BCSS deliberately assess an extreme The total time taken to complete the BCSS
form of negative evaluation of self (e.g. judging suggests that the judgement for each item is
oneself to be weak, worthless, useless, inad- made within seconds. While rapid, the associ-
equate) and others (e.g. judging others to be ations with psychopathology and the other
bad, devious, untrustworthy). Such appraisals, scales found in the present study suggests such
if held strongly, may be viewed as core with evaluations are not an epiphenomena, but
respect to their potentially catastrophic im- relate to judgements that have meaning and
plications for one’s place in a social context. importance. The lack of time to reflect may
Appraising oneself as inadequate whilst ap- imply that the reporting of self and other
praising others as devious and bad implies a evaluative thinking, as assessed by the BCSS, is
potentially very dangerous social position in likely to indicate a relatively automatic judge-
which one is both weak and threatened by ment about self and others that comes rapidly to
others. Consistent with Barrowclough and col- mind. This relative speed may be an advantage,
leagues (2003) we suggested that, rather than as it may tap the person’s ongoing moment-
lack of self-esteem (as measured by traditional to-moment evaluations of self in relation to
questionnaires), it may be specifically this type others used habitually, rather than a process
of core appraisal of danger and vulnerability in of conscious reflection that may be subject to
a social context that is typical of people with self-presentation bias. It may relate closely to
psychosis and may be linked to paranoia and the clinical concept of ‘hot cognition ’.
other extreme emotional and psychotic reac- In summary, the BCSS appear to be useful.
tions (Garety et al. 2001). The comparison of They have both good psychometric properties
a sample of students with a sample of people and construct validity for use in studies testing
with chronic relapsing psychosis allows few hypotheses relating to schema and self and
conclusions to be drawn regarding reasons for other evaluation constructs. Further studies are
758 D. Fowler et al.

required, particularly of stability over time, but with the biological, cognitive, behavioural and
also of how such appraisals may mediate the emotional responses that characterize psycho-
association of trauma and stressful life events pathology.

APPENDIX
The Brief Core Schema Scales : beliefs about self and others
This questionnaire lists beliefs that people can hold about themselves and other people.
Please indicate whether you hold each belief (NO or YES). If you hold the belief then please
indicate how strongly you hold it by circling a number (1–4). Try to judge the beliefs on how
you have generally, over time, viewed yourself and others. Do not spend too long on each
belief. There are no right or wrong answers and the first response to each belief is often the most
accurate.
Believe it Believe it Believe it Believe it
slightly moderately very much totally

MYSELF
I am unloved NO YES p 1 2 3 4
I am worthless NO YES p 1 2 3 4
I am weak NO YES p 1 2 3 4
I am vulnerable NO YES p 1 2 3 4
I am bad NO YES p 1 2 3 4
I am a failure NO YES p 1 2 3 4
I am respected NO YES p 1 2 3 4
I am valuable NO YES p 1 2 3 4
I am talented NO YES p 1 2 3 4
I am successful NO YES p 1 2 3 4
I am good NO YES p 1 2 3 4
I am interesting NO YES p 1 2 3 4
OTHER PEOPLE
Other people are hostile NO YES p 1 2 3 4
Other people are harsh NO YES p 1 2 3 4
Other people are unforgiving NO YES p 1 2 3 4
Other people are bad NO YES p 1 2 3 4
Other people are devious NO YES p 1 2 3 4
Other people are nasty NO YES p 1 2 3 4
Other people are fair NO YES p 1 2 3 4
Other people are good NO YES p 1 2 3 4
Other people are trustworthy NO YES p 1 2 3 4
Other people are accepting NO YES p 1 2 3 4
Other people are supportive NO YES p 1 2 3 4
Other people are truthful NO YES p 1 2 3 4

ACKNOWLEDGEMENTS REFERENCES
We thank all the participants who took part Barrowclough, C., Tarrier, N., Humphreys, L., Ward, J., Gregg, L. &
Andrews, B. (2003). Self-esteem in schizophrenia : relationships
in the studies from both the Universities and between self-evaluation, family attitudes and symptomology.
the NHS trusts, and in addition the NHS staff Journal of Abnormal Psychology 112, 92–99.
who helped with recruitment. This study was Bartholomew, K. & Horowitz, L. M. (1991). Attachment styles
among young adults : a test of a four-category model. Journal of
supported by a Wellcome Trust Programme Personality and Social Psychology 61, 226–244.
Grant : 062452. PRP is registered as ISRCTN : Beck, A. T. (1976). Cognitive Therapy and the Emotional Disorders.
83557988. IUP : New York.
Beck, A. T. & Rector, N. A. (2003). A cognitive model of halluci-
nations. Cognitive Therapy & Research 27, 19–52.
Beck, A. T., Brown, G. K., Steer, R. A., Kuyken, W. &
DECLARATION OF INTEREST Grisham, J. (2001). Psychometric properties of the Beck
Self-Esteem Scales. Behaviour Research & Therapy 39, 115–
None. 124.
 Psychometric properties of the Brief Core Schema Scales 759

Bentall, R. P., Corcoran, R., Howard, R., Blackwood, N. & psychosis IV : self-esteem and persecutory delusions. British
Kinderman, P. (2001). Persecutory delusions : a review and theor- Journal of Clinical Psychology 37, 415–430.
etical integration. Clinical Psychology Review 21, 1143–1192. Freeman, D., Garety, P. A., Kuipers, E., Fowler, D. & Bebbington,
Bentall, R. & Swarbrick, R. (2003). The best laid schemas of P. E. (2002). A cognitive model of persecutory delusions. British
paranoid patients : autonomy, sociotropy and need for closure. Journal of Clinical Psychology 41, 331–347.
Psychology & Psychotherapy: Theory, Research & Practice 76, Garety, P. A., Kuipers, E. K., Fowler, D., Freeman, D. & Bebbington,
163–171. P. E. (2001). A cognitive model of the positive symptoms of
Brewin, C. R. (2003). Posttraumatic Stress Disorder : Malady or psychosis. Psychological Medicine 31, 189–195.
Myth ? Yale University Press: New Haven, CT. Gilbert, P. (1992). Depression : The Evolution of Powerlessness.
Brown, G. P., Hammen, C. L., Craske, M. G. & Wickens, T. D. Erlbaum : London.
(1995). Dimensions of dysfunctional attitudes as vulnerabilities Keith, L. K. & Bracken, B. A. (1996). Self-concept instrumentation : a
to depressive symptoms. Journal of Abnormal Psychology 104, historical and evaluative review. In Handbook of Self-Concept :
431–435. Developmental, Social, and Clinical Considerations (ed. B. A.
Brown, G. W., Harris, T. O. & Hepworth, C. (1995). Loss, humili- Bracken), pp. 91–170. Oxford, UK.
ation and entrapment among women developing depression : a Lovibond, P. F. & Lovibond, S. H. (1995). The structure of negative
patient and non-patient comparison. Psychological Medicine 25, emotional states : comparison of the depression anxiety stress
7–21. scales (DASS) with the Beck depression and anxiety inventories.
Chadwick, P. D. J., Birchwood, M. J., & Trower, P. (1996). Cognitive Behaviour Research and Therapy 33, 335–343.
Therapy for Delusions, Voices and Paranoia. Wiley : Chichester. Perris, C., Fowler, D., Lund, G. & Skagerlind, P. (1998). The
Chadwick, P., Trower, P. & Dagnan, D. (1999). Measuring negative development and validation of the dysfunctional working
person evaluations : the Evaluative Beliefs Scale. Cognitive models of the self questionnaire (DWMS) : a scale to measure
Therapy & Research 23, 549–559. dysfunctional assumptions in schizophrenia and related con-
Clark, D. A., Beck, A. T. & Alford, B. A. (1999). Scientific ditions. Acta Psychiatrica Scandinavia 98, 219–223.
Foundations of Cognitive Theory and Therapy of Depression. Peters, E. R., Joseph, S. A. & Garety, P. A. (1999). Measurement
John Wiley : New York. of delusional ideation in the normal population : introducing the
Close, H. & Garety, P. A. (1998). Cognitive assessment of voices : PDI (Peters et al. Delusions Inventory). Schizophrenia Bulletin 25,
further developments in understanding the emotional impact of 553–576.
voices. British Journal of Clinical Psychology 37, 173–188. Robson, P. J. (1988). Self-esteem : a psychiatric view. British Journal
Dalgleish, T. (1999). Cognitive theories of post-traumatic stress dis- of Psychiatry 153, 6–15.
order. In Post-Traumatic Stress Disorder: Concepts and Therapy Rosenberg, M. (1965). Society and the Adolescent Self-Image.
(ed. W. Yule), pp. 197–212. John Wiley : New York. Princeton University Press : Princeton, USA.
Ehlers, A. & Clark, D. M. (2000). A cognitive model of post- Schmidt, N. B., Joiner, T. E., Young, J. E. & Telch, M. J. (1995).
traumatic stress disorder. Behaviour Research & Therapy 38, The schema questionnaire: investigation of psychometric prop-
319–345. erties and the hierarchical structure of a measure of maladaptive
Fairburn, C. G., Cooper, Z. & Shafran, R. (2003). Cognitive behav- schemas. Cognitive Therapy and Research 19, 295–321.
iour therapy for eating disorders : a ‘ transdiagnostic ’ theory and Teasdale, J. D. & Barnard, P. J. (1993). Affect, Cognition, and
treatment. Behaviour Research & Therapy 41, 509–528. Change : Re-modelling Depressive Thought. Lawrence Erlbaum
Fenigstein, A. & Vanable, P. A. (1992). Paranoia and self- Associates : Hove, UK.
consciousness. Journal of Personality and Social Psychology 62, Teasdale, J. D. & Cox, S. G. (2001). Dysphoria : self devaluative
129–138. and affective components in recovered depressed patients and
Foa, E. B. & Rothbaum, B. O. (1998). Treating the Trauma of never depressed controls. Psychological Medicine 31, 1311–1316.
Rape : Cognitive Behavioural Therapy for PTSD. Guildford : Teasdale, J. D. & Dent, J. (1987). Cognitive vulnerability to
New York. depression : an investigation of two hypotheses. British Journal of
Fowler, D. (2000). Psychological formulation of early episodes of Clinical Psychology 26 (Pt 2), 113–126.
psychosis : a cognitive model. In Early Intervention in Psychosis : A Trower, P. & Chadwick, P. D. J. (1995). Pathways to defense of the
Guide to Concepts, Evidence and Interventions (ed. M. Birchwood, self : a theory of two types of paranoia. Clinical Psychology :
D. Fowler and C. Jackson). John Wiley : Chichester. Science and Practice 2, 263–278.
Fowler, D., Garety, P. A. & Kuipers, E. K. (1995). Cognitive Young, J. E. (1990). Cognitive Therapy for Personality Disorders : A
Behaviour Therapy for Psychosis : Theory and Practice. John Schema-Focused Approach. Professional Resource Exchange :
Wiley : Chichester. Sarasota, FL.
Freeman, D., Garety, P. A., Fowler, D., Kuipers, E. K., Dunn, G., Young, J. E. (1998). The Young schema questionnaire short form
Bebbington, P. & Hadley, C. (1998). The London–East Anglia (http://home.sprynet.com/sprynet/schema/ysqs1.htm). Accessed
randomized controlled trial of cognitive-behaviour therapy for 2003.

View publication stats