P1: GDW/GCR/GIR/GMF

P2: GCV

Cognitive Therapy and Research [cotr]

pp528-cotr-375825

June 27, 2002

10:27

Style file version Jun 14th, 2002

C 2002), pp. 519530

Cognitive Therapy and Research, Vol. 26, No. 4, August 2002 (

The Schema QuestionnaireShort Form: Factor Analysis

and Relationship Between Schemas and Symptoms
Ken Welburn,1,3,4 Marjorie Coristine,2 Paul Dagg,2,3 Amanda Pontefract,3
and Shelley Jordan1,3

The original version of the Schema Questionnaire was developed by Young to measure
early maladaptive schemas. These maladaptive schemas are thought to be important in
the development and maintenance of psychiatric symptoms, such as anxiety and depression. Factor analytic research with this 205-item version of the Schema Questionnaire
has supported the schemas proposed by Young. The Schema QuestionnaireShort
Form (SQ-SF) was designed (J. E. Young, 1998) to measure 15 maladaptive schemas
and is a briefer (75 item) instrument. The present study examined the psychometric
properties of the SQ-SF with a sample of patients in a psychiatric day treatment program. The factor analysis supported the 15 schema subscales proposed by Young. These
15 subscales demonstrated good internal consistency. The present study also examined
the relationship between the SQ-SF subscales and psychiatric symptomatology. Results
provided support for the construct validity of the SQ-SF, suggesting the importance of
maladaptive schemas in the development and maintenance of psychiatric symptoms.
KEY WORDS: Brief Symptom Inventory; Schema Questionnaire; schema; factor analysis; questionnaire
validation; psychiatric symptoms.

The Schema Questionnaire (SQ) was developed by Young and Brown (1994)
to measure early maladaptive schemas (EMS). Schemas are underlying cognitive
structures that help to mediate and organize ones experience of the world (reality
filters) and are, therefore, salient in information processes such as selective attention. EMS are thought to reflect childhood experiences related to attachment and
approval/disapproval experiences. Schemas are, therefore, not irrational (in opposition to some models of cognitive therapy), but simply reflect this earlier learning and
make sense in that context. It has been posited that schemas are likely to be an integrative concept across differing therapeutic modalities such as cognitive, interpersonal,
1 Ottawa

Anxiety and Trauma Clinic, Ottawa, Ontario, Canada.

Ottawa Hospital, Ottawa, Ontario, Canada.
3 University of Ottawa, Ottawa, Ontario, Canada.
4 Correspondence should be directed to Dr Ken Welburn, Ottawa Anxiety and Trauma Clinic, Suite 202,
2277 Riverside Drive, Ottawa, Ontario, Canada K1H 7X6; e-mail: ottanx@igs.net.
2 Royal

519
C 2002 Plenum Publishing Corporation
0147-5916/02/0800-0519/0

P1: GDW/GCR/GIR/GMF

P2: GCV

Cognitive Therapy and Research [cotr]

520

pp528-cotr-375825

June 27, 2002

10:27

Style file version Jun 14th, 2002

Welburn, Coristine, Dagg, Pontefract, and Jordan

dynamic, and constructivist (Welburn, Dagg, Coristine, & Pontefract, 2000). Examples of these EMS are abandonment (the belief that those individuals who provide
support and connection cannot be relied upon), vulnerability to harm (unrealistic or
disproportionate fear that a catastrophe could strike at any time), and unrelenting
standards (the belief that one must perform at extremely high standards to avoid
rejection or criticism).
Theoretically, there is an association between certain schema and psychiatric
symptomatology. For example, the schemas for vulnerability to harm is more likely
to be related to feelings of anxiety rather than depression. Implications of schemas
are also noted in terms of information processing, with selective filtering of corroborating information and discounting of conflicting information (Beck, 1976; Beck,
Freeman, & Associates, 1991; McGinn & Young, 1996; Schmidt, Joiner, Young, &
Telch, 1995). The processes of scanning for proof and discounting conflicting information serve to perpetuate the schemas over time into ones adult life and relationships.
This self-perpetuating nature of EMS leaves the individual vulnerable to experiences
of depression and anxiety in situations that activate the schemas (see Segal, 1988 for
a further discussion of this issue). Therefore, the assessment of schemas has important clinical relevance. Once identified, schemas can be targeted with interventions
to correct cognitive distortions and reduce related symptomatology.
One of the difficulties in assessing schema is that they are thought to be underlying or unconscious structures whereas self-report measures are based on the respondents conscious awareness and self-perception (Muran, 1991; Muran, Samstag,
Segal, & Winston, 1998; Segal & Shaw, 1986). It can be argued, therefore, that schemas
can be best assessed when using non-self-report measures such as projective-like tests
or physiological indicators of information processing. However, even though schema
operate outside of conscious awareness, it is not unlikely that there would be some
awareness of the schema, particularly when they result in numerous negative experiences. For example, schemas for mistrust unconsciously influence and shape the
perception of ones interpersonal interactions, resulting in an automatic scanning
for evidence that others cannot be trusted and discounting of information to the
contrary. It would be unlikely that the individual would not be aware of not trusting others, particularly where the schema operates to the extent to create ongoing
negative interpersonal experiences. Furthermore, non-self-report measures are not
without their own limitations and difficulties. They may be more useful in highly
controlled research situations and they can be cumbersome and time consuming in
clinical practice. They also may require specialized training to administer thus restricting their use for many clinicians. There is clearly a need for a reliable and valid
self-report instrument that assesses maladaptive schema.
The SQ was developed based on the observations and reasoning of experienced
clinicians. This 205 item self-report inventory was originally designed to measure
16 primary maladaptive schemas (Young, 1994). The 16 subscales of the SQ have
demonstrated adequate test-retest reliability and internal consistency as well as convergent and discriminant validity. Schmidt et al. (1995) conducted a series of three
studies to assess the psychometric properties of the SQ. The results of the factor analyses with a large student sample (n = 1129) showed support for 13 of the 16 proposed
schemas. In a smaller sample of psychiatric patients (n = 187) the factor analysis

P1: GDW/GCR/GIR/GMF

P2: GCV

Cognitive Therapy and Research [cotr]

pp528-cotr-375825

June 27, 2002

10:27

Style file version Jun 14th, 2002

Schema Questionnaire

521

supported 15 of the 16 proposed schemas. A hierarchical factor analysis identified

three higher order factors: Disconnection, Overconnection, and Exaggerated Standards. Overall, the SQ was found to measure a number of relevant schemas for both
clinical and nonclinical populations and appears to be a useful clinical and research
tool. A recent study (Lee, Taylor, & Dunn, 1999) undertook a factor analysis of the
SQ with a larger clinical population and found the same 15 factors as (Schmidt et al.,
1995), with the addition of a fear of losing control factor.
The Schema QuestionnaireShort Form (SQ-SF) was developed in order to
have a shorter, but still clinically relevant assessment tool to measure maladaptive
schemas. The SQ-SF is comprised of a subset of 75 items from the original 205-item
SQ. These 75 items are thought to represent 15 early maladaptive schemas (Young,
1998). To date, no published research study has examined the factor structure of this
short form of the SQ. One study (Welburn et al., 2000) reported on schema change
as assessed by the SQ-SF as a measure of treatment outcome.
One purpose of this study was to examine the psychometric properties of the SQSF with a psychiatric day treatment population. Clients referred to the day treatment
program typically have had many crisis episodes associated with acute experiences
of anxiety and depression, suggesting that they would manifest a prevalence of early
maladaptive schemas. More specifically, the primary goal of the present research
was to examine the factor structure of the SQ-SF and determine if the results are
consistent with Schmidt et al.s results based on the original SQ. The second goal
was to examine the convergent and divergent validity of the SQ-SF subscales based
on their relationship to three relevant subscales of the BSI (depression, anxiety, and
paranoia subscales). According to Young (1994), schemas that are conceptually congruent with psychological symptoms should be significantly correlated with those
symptoms. Theoretically as well as from our clinical experience, we expected that
anxiety, depression, and paranoi would be most clearly linked to certain maladaptive schema. Specifically, we hypothesized that the vulnerability to harm schema
would be correlated with symptoms of anxiety, given the conceptual congruence of a
sense of apprehension and threat (Riskind, Williams, Gessner, Chrosniak, & Cortina,
2000). Similarily, we expected that the mistrust/abuse schema would be correlated
with symptoms of paranoia, given the common element of suspiciousness of others.
Lastly, we hypothesized that defectiveness, emotional deprivation, and abandonment schemas would be correlated with symptoms of depression, with congruence
in experiences of hopelessness, helplessness, and loss.
METHOD
Instruments
Schema Questionnaire ShortForm (SQ-SF)
The SQ-SF is a 75-item questionnaire that was designed to assess 15 early maladaptive cognitive schema: emotional deprivation, abandonment, mistrust/abuse,
social alienation, defectiveness, incompetence, dependency, vulnerability to harm,
enmeshment, subjugation of needs, self-sacrifice, emotional inhibition, unrelenting

P1: GDW/GCR/GIR/GMF

P2: GCV

Cognitive Therapy and Research [cotr]

522

pp528-cotr-375825

June 27, 2002

10:27

Style file version Jun 14th, 2002

Welburn, Coristine, Dagg, Pontefract, and Jordan

standards, entitlement, and insufficient self-control (Young, 1998). Each of the

75 items of the SQ-SF is rated on a 6-point scale (1 = completely untrue of me;
2 = mostly untrue of me; 3 = slightly more true than untrue; 4 = moderately true
of me; 5 = mostly true of me; 6 = describes me perfectly). Higher scores indicate a
greater presence of that maladaptive schema for the respondent.
The Brief Symptom Inventory (BSI)
The BSI (Derogatis, 1993) is a self-report psychiatric symptom scale and is a
shortened (53 items) version of the revised Symptom Checklist-90. The BSI has established norms of psychiatric distress for psychiatric outpatients. Each item of the
BSI is rated on a 5-point scale (0 = not at all; 1 = a little bit; 2 = moderately; 3 = quite
a bit; 4 = extremely). Higher scores for each subscale indicate a greater experience of
the symptom. The BSI measures current psychiatric symptomatology and has three
global measures of psychiatric distress (global severity index, positive symptom total, and positive symptom distress index) and nine primary symptom dimensions
including somatization (Som), obsessivecompulsiveness (OC), interpersonal sensitivity (IS), depression (Dep), anxiety (Anx), hostility (Hos), phobic anxiety (Phob),
paranoid ideation (Par), and psychoticism (Psy).
Participants and Procedure
The sample consisted of 203 consecutive referrals to a day treatment program of
a psychiatric hospital over a 2-year period. Patients admitted to day treatment attend
a 12-week intensive group program consisting of group psychotherapy, cognitive
behavioural group, assertiveness training, life skills, and health promotion. Hospital
psychiatrists referred patients from the walk-in emergency service, the crisis unit,
inpatient wards, and from general outpatient services. The most common reasons for
referral were depression, suicidal ideation, and being in a state of crisis. Referrals
related to symptoms of anxiety, PTSD, and personality disorders were also common. Exclusionary criteria from the group included active substance abuse or active
psychosis.
All participants filled out the SQ-SF at the time of the screening interview. Of the
203 referrals, 7 were excluded from the study due to a substantial number of missing
items on the SQ-SF. In the second year, participants also responded to the BSI at the
time of referral to day treatment, resulting in 135 participants who completed both
the SQ-SF and the BSI.
Of the 196 participants who completed the SQ-SF, 33% (n = 65) were male and
67% (n = 131) were female. The age range was from 18 to 63 years (M = 36.9; SD =
9.3). DSM-IV diagnoses (American Psychiatric Association, 1994) were assessed
through chart review. All psychiatric diagnoses that had been entered into the chart
by a psychiatrist were recorded. On review, 98% had received at least one Axis I
diagnosis and 46% had more than one Axis I diagnosis. Seventy-five percent had
received a diagnosis of mood disorder (46% depression, 29% dysthymia, 6% bipolar), 26% had an anxiety disorder (phobia, social anxiety, generalized anxiety, panic,
agoraphobia), 16% had a history of substance abuse, 14% had posttraumatic stress

P1: GDW/GCR/GIR/GMF

P2: GCV

Cognitive Therapy and Research [cotr]

pp528-cotr-375825

June 27, 2002

10:27

Schema Questionnaire

Style file version Jun 14th, 2002

523

disorder, 7% had adjustment disorder, 4% had a dissociative disorder, 3% had a

schizoaffective disorder, 3% had an eating disorder, and 1% had received a diagnosis of schizophrenia. Also, 36% had received at least one Axis II diagnosis and
3% had more than one personality disorder. These included borderline personality (14%), avoidant (8%), dependent (7%), passiveaggressive (4%), mixed (3%),
obsessivecompulsive (2%), antisocial (1%), histrionic (1%), and narcissistic (1%).
In all, 32% had both an Axis I and Axis II disorder. Fifty-one percent had been admitted to hospital for psychiatric reasons in the year prior to the interview, and, in that
same time frame, 40% of the patients had made at least one visit to the emergency
services of the psychiatric hospital.

RESULTS
Factor Analysis of the SQ-SF
An exploratory factor analysis was conducted using the principal components
method. A varimax rotation was used for interpretability of results (Stevens, 1986).
Participants missing less than 10% of the items on the SQ-SF were retained in the
analyses. Given that the pattern of missing data was random, for these participants
(n = 45) missing data were recoded to the item mean. The resulting sample size for
the factor analysis was 196, less than the suggested minimum of five individuals per
variable (Gorsuch, 1983). Although this warrants caution in the interpretation of the
results of the factor analysis, Gorsuch also suggests that a smaller sample size is less
problematic when the sample size is greater than 100 and the number of expected
variables per factor is low.
Based on the principal components methods, a 15-factor solution converged in
15 iterations, accounting for 73.1% of the variance. Using the Kaiser criterion (retention of components with eigenvalues greater than one; Stevens, 1986), all 15 factors
were retained. As recommended by Stevens (1986), the critical value for factor loadings was based on sample size. With a sample size of approximately 200, loadings
greater than .36 were considered significant. The resulting factor structure, including eigenvalues, percent of variance accounted for by each factor, internal reliability
estimates for each factor, and significant factor loadings are reported in Table I.
Sample SQ-SF items representing the two highest loading items on each factor are
detailed. Four of the 75 items cross-loaded on more than one factor. Item 50 loaded
on both Factor 6 (factor loading of .60) and Factor 2 (factor loading of .38). Based on
the factor loadings and the question content, this item fits best on Factor 6. Item 35
loaded on both Factor 11 (factor loading of .51) and Factor 2 (factor loading of .40).
Examination of question content suggests that this item fits best on Factor 11. Item 36
loads on both Factor 11 (factor loading of .47) and to Factor 15 (factor loading of .41).
We kept the item on Factor 15 given the similar loadings and theoretical congruence
with that factor. Item 46 cross-loaded to Factor 11 (factor loading of .37) and had an
insignificant loading on Factor 14 (factor loading of .30). Examination of item content suggested some overlap in content across these two factors. Based on Youngs
theoretical model, this item was left on Factor 14. Finally, Item 38 did not have a

P1: GDW/GCR/GIR/GMF

P2: GCV

Cognitive Therapy and Research [cotr]

pp528-cotr-375825

June 27, 2002

524

10:27

Style file version Jun 14th, 2002

Welburn, Coristine, Dagg, Pontefract, and Jordan

Table I. Fifteen Factors of the SQ-SF (n = 196)
Factors

Factor
loading

Factor 1 failure (eigenvalue = 20.63; % variance = 27.5, Cronbachs alpha = .93)

Q29 Im not as talented as most people are at their work
.86
Q28 Most other people are more capable than I am in areas of work and achievement
.82
Q30
.78
Q26
.77
Q27
.75
Factor 2 insufficient self-control (eigenvalue = 5.87; % variance = 7.8, Cronbachs alpha = .90)
Q72 If I cant reach a goal, I become easily frustrated and give up
.79
Q75 I have rarely been able to stick to my resolutions
.77
Q74
.75
Q73
.73
Q71
.69
Factor 3 abandonment (eigenvalue = 3.88; % variance = 5.2, Cronbachs alpha = .91)
Q6 I find myself clinging to people Im close to because Im afraid they will leave me
.79
Q7 I need other people so much I worry about losing them
.78
Q8
.78
Q9
.76
Q10
.66
Factor 4 mistrust/abuse (eigenvalue = 3.47; % variance = 4.6, Cronbachs alpha = .91)
Q14 I am quite suspicious of other peoples motives
.81
Q15 Im usually on the lookout for peoples ulterior motives
.79
Q12
.72
Q13
.69
Q11
.64
Factor 5 emotional deprivation (eigenvalue = 3.18; % variance = 4.3, Cronbachs alpha = .90)
Q1 Most of the time, I havent had someone to nurture me, share him/herself with me,
.85
or care deeply about everything that happens to me
Q2 In general, people have not been there to give me warmth, holding, and affection
.84
Q4
.77
Q5
.75
Q3
.69
Factor 6 self-sacrifice (eigenvalue = 2.92; % variance = 3.9, Cronbachs alpha = .87)
Q53 Im so busy doing for the people I care about that I have little time for myself
.83
Q51 I am the one who usually ends up taking care of the people I am close to
.76
Q55
.75
Q52
.73
Q54
.70
Factor 7 enmeshment (eigenvalue = 2.39; % variance = 3.2, Cronbachs alpha = .88)
Q42 My parent(s) and I tend to be over-involved in each others lives and problems
.84
Q44 I often feel as if my parent(s) are living though meI dont have a life of my own
.79
Q41
.76
Q43
.74
Q45
.63
Factor 8 social alienation (eigenvalue = 2.19; % variance = 2.9, Cronbachs alpha = .92)
Q16 I dont fit in
.77
Q18 I dont belong; Im a loner
.74
Q20
.71
Q17
.70
Q19
.70
Factor 9 emotional inhibition (eigenvalue = 2.12; % variance = 2.8, Cronbachs alpha = .87)
Q58 I find it hard to be warm and spontaneous
.82
Q57 I find it embarrassing to express my feelings to others
.81
Q59
.78
Q56
.73
Q60
.64
(Continued )

P1: GDW/GCR/GIR/GMF

P2: GCV

Cognitive Therapy and Research [cotr]

pp528-cotr-375825

June 27, 2002

10:27

Style file version Jun 14th, 2002

Schema Questionnaire

525
Table I. (Continued )
Factors

Factor
loading

Factor 10 unrelenting standards (eigenvalue = 1.71; % variance = 2.3, Cronbachs alpha = .84)
Q63 I must meet all my responsibilities
.80
Q62 I try to do my best; I cant settle for good enough
.76
Q65
.74
Q64
.73
Q61
.67
Factor 11 dependency (eigenvalue = 1.50; % variance = 2.0, Cronbachs alpha = .86)
Q32 I think of myself as a dependent person when it comes to everyday functioning
.67
Q31 I do not feel capable of getting by on my own in everyday life
.65
Q33
.55
Q34
.50
Q35
.51
Factor 12 defectiveness/shame (eigenvalue = 1.37; % variance = 1.8, Cronbachs alpha = .91)
Q24 I feel that Im not loveable
.72
Q25 I am too unacceptable in very basic ways to reveal myself to other people
.65
Q23
.62
Q21
.61
Q22
.60
Factor 13 entitlement (eigenvalue = 1.29; % variance = 1.7, Cronbachs alpha = .76)
Q69 I feel that I shouldnt have to follow the normal rules and conventions other
.77
people do
Q67 Im special and shouldnt have to accept many of the restrictions placed on other
.73
people
Q70
.70
Q68
.55
Q66
.39
Factor 14 subjugation of needs (eigenvalue = 1.23; % variance = 1.6, Cronbachs alpha = .88)
Q48 In relationships, I let the other person have the upper hand
.73
Q49 Ive always let others make choices for me, so I really dont know what I want for
.63
myself
Q50
.60
Q47
.52
Q46
.30 (ns)
Factor 15 vulnerability to harm (eigenvalue = 1.06; % variance = 1.4, Cronbachs alpha = .80)
Q40 I worry that Im developing a serious illness, even though nothing serious has been
.67
diagnosed by a physician
Q39 I worry that Ill lose all my money and become destitute
.65
Q37
.58
Q36
.41
Q38
.31 (ns)

significant loading on any factor. However, it fit best on Factor 15 (factor loading
of .31). In all following analyses the 15 subscales of the SQ-SF were scored using the
items as presented in Table I and consistent with original scoring of the instrument.
Internal reliability coefficients were calculated for each of the 15 subscales of
the SQ-SF. Cronbachs alpha coefficients ranged from .76 to .93, suggesting moderate
to good internal consistency for each of the 15 five-item subscales.
Relationship Between SQ-SF Subscales and the BSI
Correlational and regression analyses were conducted to explore the relationship between the subscales of the Schema Questionnaire and psychological distress

P1: GDW/GCR/GIR/GMF

P2: GCV

Cognitive Therapy and Research [cotr]

pp528-cotr-375825

June 27, 2002

526

10:27

Style file version Jun 14th, 2002

Welburn, Coristine, Dagg, Pontefract, and Jordan

Table II. Correlations of SQ-SF Subscales With BSI Subscales (n = 135)
BSI subscale
SQ-SF subscale

Som

OC

IS

Dep

Anx

Hos

Phob

Par

Psy

Failure
Insufficient self-control
Abandonment
Mistrust/abuse
Emotional deprivation
Self-sacrifice
Enmeshment
Social alienation
Emotional inhibition
Unrelenting standards
Dependency
Defectiveness/shame
Entitlement
Subjugation of needs
Vulnerability to harm

.34
.46
.28
.34

.42
.46
.35
.28

.49
.48
.47
.36

.39
.42
.55
.34

.48
.43
.55
.38

.50
.57
.39
.45

.48
.40
.37
.30

.32
.44
.45
.69

.09
.35
.28
.23
.22
.20
.38
.23
.30
.32
.48

.07
.31
.25
.32
.23
.22
.48
.27
.21
.41
.47

.11
.25
.42
.39
.24
.16
.45
.47
.26
.49
.38

.21
.27
.21
.37
.30
.31
.41
.47
.21
.36
.44

.12
.32
.37
.31
.26
.27
.45
.40
.28
.45
.57

.07
.35
.40
.36
.20
.23
.41
.36
.38
.42
.48

.05
.32
.36
.30
.23
.24
.51
.34
.24
.44
.52

.21
.40
.30
.31
.26
.17
.34
.40
.35
.48
.58

.46
.45
.51
.49
.26
.32
.32
.44
.35
.24
.51
.52
.34
.50
.59

Note. Som = somatization; OC = obsessivecompulsiveness; IS = interpersonal sensitivity; Dep =

depression; Anx = anxiety; Hos = hostility; Phob = phobic anxiety; Par = paranoid ideation; Psy =
psychoticism.
p < .001.

based on the BSI. Correlations of the 15 SQ-SF subscales and the BSI subscales are
presented in Table II. Given the large number of correlations, a more conservative
alpha level (.001) was used to determine significant correlations. Overall, most of the
subscales of the SQ-SF correlated significantly with the BSI subscales.
Three standard multiple regression analyses were conducted in order to examine the relationship between the subscales of the SQ-SF and three measures of
psychological distress (the depression, anxiety, and paranoid anxiety subscales of the
BSI). Results of the regression analysis with the anxiety subscale as the dependent
variable indicated that the SQ-SF subscales accounted for a significant proportion
of the variance (52%) in anxiety (R = .72, p < .0001). Five of the SQ-SF subscales
were significant individual predictors of anxiety, with abandonment accounting for
11.3% of the unique variance (t = 3.90, p < .01), vulnerability to harm accounting
for 10.5% of the unique variance (t = 3.74, p < .001), failure accounting for 5.2%
of the unique variance (t = 2.56, p < .05), self-sacrifice accounting for 3.5% of the
unique variance (t = 2.07, p < .05), and emotional inhibition accounting for 3.3%
of the unique variance (t = 2.03, p < .05).
The second regression analysis indicated that the SQ-SF subscales accounted
for a significant proportion of the variance (62%) in paranoia (R = .79, p < .0001).
Four of the SQ-SF subscales contributed significant unique variance to paranoia, with
mistrust/abuse accounting for 22.5% of the unique variance (t = 5.87, p < .0001),
vulnerability to harm accounting for 8.4% of the unique variance (t = 3.31, p <
.01), self-sacrifice accounting for 4.7% of the unique variance (t = 2.42, p < .05),
and insufficient self-control accounting for 3.4% of the unique variance (t = 2.04,
p < .05).
The third regression analysis indicated that the SQ-SF subscales accounted for
a significant proportion of the variance (47%) in depression (R = .69, p < .0001).

P1: GDW/GCR/GIR/GMF

P2: GCV

Cognitive Therapy and Research [cotr]

pp528-cotr-375825

June 27, 2002

10:27

Style file version Jun 14th, 2002

Schema Questionnaire

527

Table III. Univariate Analyses of Sex Differences on the SQ-SF (n = 196)

Men

Women

SQ-SF subscale

F(1, 194)

SD

SD

Self-sacrifice
Enmeshment
Failure
Abandonment
Defectiveness/shame
Subjugation of needs
Unrelenting standards
Mistrust/abuse
Vulnerability to harm
Emotional deprivation
Social alienation
Dependency
Entitlement
Emotional inhibition
Insufficient self-control

23.56
12.76
12.54
11.51
9.39

3.29
2.02
2.72
3.29
3.05
3.26
4.01
3.18
3.13
3.85
4.14
3.05
2.70
3.36
3.68

1.37
1.07
1.47
1.45
1.42
1.35
1.27
1.37
1.34
1.45
1.36
1.29
1.11
1.31
1.35

4.27
2.82
3.52
4.06
3.75
3.88
4.53
3.76
3.51
4.17
4.35
3.18
2.59
3.29
3.73

1.32
1.65
1.52
1.52
1.55
1.45
1.23
1.45
1.41
1.36
1.37
1.34
1.12
1.48
1.53

8.41
7.55
7.29
3.24
2.27
1.08
0.42
0.42
0.10
0.04

< .003.

Two of the SQ-SF subscales were significant individual predictors of depression. The
abandonment subscale accounted for 12.5% of the unique variance in depression
(t = 4.13, p < .001) and the insufficient self-control subscale accounted for 5.5% of
the unique variance in depression (t = 2.62, p < .01).
Scores for men and women on the Global Symptom Index (GSI; sum of items/
total number of responses) of the BSI were analyzed with a one-way ANOVA. The
sample included 45 men and 90 women. Results indicated that women (M = 2.28,
SD = 0.76) scored significantly higher on the GSI than did men (M = 1.88, SD =
0.71). The mean scores for both men and women were 1 SD above the norms for
psychiatric outpatients.
Differences between men and women on the 15 subscales of the SQ-SF were
conducted with a one-way between-subjects MANOVA. The sample for this analysis included 65 men and 131 women. Results indicated that men and women differed significantly on the combination of the 15 subscales of the SQ-SF, F(15, 180) =
3.88, p < .0001. Examination of univariate F ratios using a corrected alpha level of
.003 indicated that men and women differed significantly on five of the SQ-SF subscales (self-sacrifice, enmeshment, failure, abandonment, and defectiveness/shame),
with women scoring higher than men on all five of these subscales. Table III displays
univariate F ratios, group means, and standard deviations.
DISCUSSION
Overall, the results of the present study on the SQ-SF are consistent with previous findings (Lee et al., 1999; Schmidt et al., 1995) for the longer, 205-item Schema
Questionnaire. The factor analytic results provide strong support for the hypothesized internal structure of the questionnaire, resulting in the 15 factors. Furthermore,
alpha reliability coefficients indicate that the subscales of the SQ-SF have adequate
to very good internal consistency.

P1: GDW/GCR/GIR/GMF

P2: GCV

Cognitive Therapy and Research [cotr]

528

pp528-cotr-375825

June 27, 2002

10:27

Style file version Jun 14th, 2002

Welburn, Coristine, Dagg, Pontefract, and Jordan

This sample size is similar to that of Schmidt et al.s (Schmidt et al., 1995) psychiatric sample (although those authors had a much larger nonclinical sample), with
the present study having a slightly larger proportion of females (46% vs. 67% in the
current study). The sample in the present study also appears to have higher levels of
current psychological symptomatology and distress. For example, in the Schmidt et al.
(1995) sample, only 10% of the population had ever been hospitalised for psychiatric
reasons, whereas more than half of this sample had been admitted to hospital in the
past year alone. The present study extends the previous psychometric investigation
with a more severely ill psychiatric population.
The results from the factor analysis closely parallel the 15 subscales of the questionnaire. The high correspondence between the theoretical subscale structure and
the resulting empirical analysis provides good support for the construct validity of the
SQ-SF. In all, 70 of the 75 items loaded exactly with the theoretical structure of the
instrument, 4 had significant cross loadings with other subscales, and 1 item failed to
meet criteria for significance in loading on a factor. It is expected that there would
be at least some conceptual overlap for certain schema, such as dependency and
subjugation of needs. Those who feel highly dependent on others may well also feel
particularly at the mercy of others interpersonally and thus experience a sense of
subjugation of their own needs. Measures of internal consistency ranged from .76 to
.93, suggesting an adequate internal consistency of the 15 subscales.
In an examination of gender differences in negative schema, females were significantly higher on schemas of self-sacrifice, enmeshment, failure, abandonment,
and defectiveness. Schemas such as self-sacrifice may be culturally influenced and
manifest in gender differences where females learn to think of others needs as coming before their own. In a within subjects exploration, both men and women in
this sample gave the highest ratings on schema for unrelenting standards and social
alienation (all above 4 on a 6-point scale). Interestingly, females were lowest on entitlement schema whereas males were lowest on enmeshment. Consistent with high
self-sacrifice schema, females may also experience a lack of sense of entitlement for
having their own needs met. The relatively lower enmeshment ratings for males in
this sample may also reflect a culturally conditioning where men are expected to be
autonomous and independent. However, it remains to be seen if these results are
specific to a psychiatric population or if this pattern exists in a nonclinical population
as well.
The present study also examined the relationship of the SQ-SF subscales to
symptoms of anxiety, depression, and paranoia. The results support the construct validity of the SQ-SF and provide some insight into how cognitive schemas may relate
to specific psychological symptoms. Five of the maladaptive schemas were related
to anxiety. As hypothesized, the schema for vulnerability to harm was an important individual predictor of anxiety. However, abandonment, failure, self-sacrifice,
and emotional inhibition were also predictors of anxiety. The schemas most relevant to depression were abandonment (as predicted) and insufficient self-control.
Conceptually, items from the insufficient self-control schema overlap with symptoms of depression such as lack of energy and motivation. The link between the
abandonment schema and depression is consistent with the concept of abandonment depression in attachment theories. However, abandonment schema appear to

P1: GDW/GCR/GIR/GMF

P2: GCV

Cognitive Therapy and Research [cotr]

Schema Questionnaire

pp528-cotr-375825

June 27, 2002

10:27

Style file version Jun 14th, 2002

529

be highly salient not only in predicting depression, but also for anxiety, accounting for
the most variance in both cases and highlighting the clinical significance of that particular schema. In early attachment experiences, abandonment would clearly represent
a greater sense of hopelessness than emotional deprivation arising from neglect experiences. When a caretaker is present but neglectful, there is potentially the hope
of eventually or intermittently gaining the needed approval and affection through
efforts such as trying harder to be lovable. When the caretaker abandons the child,
the situation is more hopeless as the childs best efforts go without any reinforcement
from the absent adult. Hopelessness, helplessness, and loss arising from the abandonment experience would leave the individual prone to episodes of depression, consistent with our finding of abandonment schema and depression. The early experience
of abandonment may also destroy an illusion of interpersonal security and connectedness, resulting in an ongoing experience of anxiety. Even occasionally nurturing
caretaking may allow for the internalizing of some ability to self-soothe from an anxious state. Abandonment may represent a more primal and primitive developmental
disruption, resulting in less ability to modulate or reduce the experience of anxiety.
The schema for mistrust was the best predictor of paranoia, consistent with our
hypothesis and adding to the construct validity of the subscale. Vulnerability to harm,
self-sacrifice, and insufficient self-control were also significant predictors of paranoia.
Vulnerability to harm predicted anxiety and paranoi, but not depression and thus
provides evidence for the divergent validity of that subscale. Although we had not
hypothesized the association between vulnerability to harm and paranoia, there is a
conceptual overlap relating to the sense of the world as a dangerous and threatening
place. In paranoi, the scale assesses the sense of threat specific to the interpersonal
realm.
The correspondence of the subscales of the SQ-SF with subscales of the BSI does
strengthen the assertion that the questionnaire is likely to be a valuable clinical and
research tool. This instrument could be particularly useful in assessing how different
therapeutic interventions may effect specific symptoms by targeting underlying cognitive schema. Further research is needed to examine the SQ-SFs effectiveness in
this regard.
One important consideration is the relatively small sample size for the factor
analysis. For a questionnaire with 75 items, a sample size of 375 would be recommended. Results can be more easily influenced by the idiosyncratic responding of a
few individuals when using a smaller sample size. However, the results of the factor
analysis are consistent with the theoretical factor structure in spite of limitations of
sample size. The current study does provide initial support for the factor structure
of the SQ-SF although further research with a larger sample size and using a confirmatory factor analysis will be important. In addition, a larger sample size allow for
an analysis of the factor structure of the SQ-SF for men and women separately.
A second limitation of the present study is the reliance on self-report instruments. This methodology does not answer the question of whether the SQ-SF is
truly assessing underlying cognitive schema that may be not easily accessible to the
conscious awareness required in responding to the self-report questionnaire. The SQSF remains to be validated against other, non-self-report measures of information
processing.

P1: GDW/GCR/GIR/GMF

P2: GCV

Cognitive Therapy and Research [cotr]

530

pp528-cotr-375825

June 27, 2002

10:27

Style file version Jun 14th, 2002

Welburn, Coristine, Dagg, Pontefract, and Jordan

ACKNOWLEDGMENTS
This study was supported in part by the Associates of Psychiatry, Royal Ottawa
Hospital. The authors gratefully acknowledge the invaluable assistance of Jackie
Clark, Linda Fong, Rosemarie Lidstone, Carmela OKeefe, Dayle Raine, and Enid
Robins-Holm in carrying out this study. We also acknowledge the helpful comments
of the reviewers and Dr Jane Evans.

REFERENCES
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.).
Washington, DC: Author.
Beck, A. T. (1976). Cognitive therapy and the emotional disorders. New York: International Universities
Press.
Beck, A. T., Freeman, A., & Associates. (1991). Cognitive therapy of personality disorders. New York:
Guilford Press.
Derogatis, L. R. (1993). Brief Symptom Inventory (BSI): Administration, scoring and procedures manual
(3rd ed.). Minneapolis: National Computer Systems.
Gorsuch, R. L. (1983). Factor analysis (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum.
Lee, C. W., Taylor, G., & Dunn, J. (1999). Factor structure of the schema questionnaire in a large clinical
sample. Cognitive Therapy and Research, 23, 441451.
McGinn, L. K., & Young, J. E. (1996). Schema-focused therapy. In P. M. Salkauskis (Ed.), Frontiers of
cognitive therapy (pp. 182207). New York: Guilford Press.
Muran, J. C. (1991). A reformulation of the ABC model in cognitive psychotherapies: Implications for
assessment and theory. Clinical Psychology Review, 11, 399418.
Muran, J. C., Samstag, L. W., Segal, Z. V., & Winston, A. (1998). Interpersonal scenarios: An idiographic
measure of self-schemas. Psychotherapy Research, 8(3), 321333.
Riskind, J. H., Williams, N. L., Gessner, T. L., Chrosniak, L. D., & Cortina, J. M. (2000). The looming
maladaptive style: Anxiety, danger and schematic processing. Journal of Personality and Social Psychology, 79, 837852.
Schmidt, N. B., Joiner, T. E., Young, J. E., & Telch, M. J. (1995). The Schema Questionnaire: Investigation
of psychometric properties and the hierarchical structure of a measure of maladaptive schemas.
Cognitive Therapy and Research, 19, 295321.
Segal, Z. V. (1988). Appraisal of the self-schema construct in cognitive models of depression. Psychological
Bulletin, 103(2), 147162.
Segal, Z. V., & Shaw, B. F. (1986). Cognition in depression: A reappraisal of Coyne and Gotlibs critique.
Cognitive Therapy and Research, 10, 671693.
Stevens, J. (1986). Applied multivariate statistics for the social sciences. Hillsdale, NJ: Lawrence Erlbaum.
Welburn, K., Dagg, P., Coristine, M., & Pontefract, A. (2000). Schematic change as a result of an intensive
group-therapy day-treatment program. Psychotherapy, 37(2), 189195.
Young, J. E. (1998). Young Schema Questionnaire Short Form (1st ed.). New York: Cognitive Therapy
Center.
Young, J. E., & Brown, G. (1994). Young Schema Questionnaire. In J. E. Young (Ed.), Cognitive therapy for
personality disorders: A schema-focused approach. (Rev. ed., pp. 6376). Sarasota, FL: Professional
Resource Exchange.