BEHAVIOUR

RESEARCH AND
THERAPY
PERGAMON Behaviour Research and Therapy 37 (1999) 881±902
www.elsevier.com/locate/brat

The overvalued ideas scale: development, reliability and

validity in obsessive±compulsive disorderp
Fugen Neziroglu *, Dean McKay, Jose A. Yaryura-Tobias, Kevin P. Stevens,
John Todaro
Institute for Bio-Behavioral Therapy and Research, Department of BioPsychoSocial Research, 935 Northern Blvd,
Great Neck, NY 11021, USA
Received 16 April 1998; received in revised form 20 October 1998

Abstract

The presence of overvalued ideas in obsessive±compulsive disorder (OCD) has been theoretically
linked to poorer treatment outcome [Kozak, M. J. & Foa, E. B. (1994). Obsessions, overvalued ideas
and delusions in obsessive±compulsive disorder. Behaviour Research and Therapy, 32, 343±353]. To date,
no measures have been developed which quantitatively assess levels of overvalued ideas in obsessive±
compulsives. The present studies examined the psychometric properties of a scale developed to measure
this form of psychopathology, the Overvalued Ideas Scale (OVIS). In study 1, 102 patients diagnosed
with OCD were administered a battery of instruments including the OVIS at baseline and two weeks
later, prior to initiating treatment. Results indicate that the OVIS has adequate internal consistency
reliability (coecient a = 0.88 at baseline), test±retest reliability (r = 0.86) and interrater reliability
(r = 0.88). Moderate to high levels of convergent validity was found with measures of obsessive±
compulsive symptoms, a single item assessment of overvalued ideas and psychotic symptoms. Medium
levels of discriminant validity with measures of anxiety and depression was obtained in this study.
Individuals determined to have high OVI showed greater stability of this pathology than those with
lower OVI, suggesting that overvalued ideas are stable for extreme scorers. In study 2 a total of 40
patients participated who were diagnosed with OCD. The same battery of instruments was administered
as in study 1, as well as the Beck Depression Inventory and Beck Anxiety Inventories. Results were
similar to that obtained in study 1, including a relative lack of discriminant validity with self-report
measures of depression and anxiety. It is suggested that further research with the OVIS may show
predictive value in treatment outcome studies of OCD. # 1999 Elsevier Science Ltd. All rights reserved.

p
A preliminary report of this study was presented at the 66th meeting of the Eastern Psychological Association,
Boston, MA and the meeting of the World Congress for Behavioral and Cognitive Therapies, Copenhagen,
Denmark. Dean McKay is now at Fordham University, Department of Psychology, Bronx, NY. John Todaro is
now at University of Florida, Department of Psychology, Gainesville, FL.
* Corresponding author.

0005-7967/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 0 5 - 7 9 6 7 ( 9 8 ) 0 0 1 9 1 - 0
882 F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902

1. Introduction

Obsessive±compulsive disorder (OCD) is characterized by repetitive thoughts, images or

impulses which are viewed as unwanted (obsessions), and are often accompanied by ritualistic
behaviors designed to alleviate the obsessions (compulsions). Traditionally, individuals with
OCD view these obsessions and related behavioral activities as senseless, but experience the
symptoms as uncontrollable. The Diagnostic and Statistical Manual (DSM-IV; American
Psychiatric Association, 1994) classi®es OCD as an anxiety disorder, but the most recent
version includes the use of the identi®er `with poor insight'. This identi®er is intended to
demarcate individuals who su€er from OCD and also view the obsessions, or compulsions
accompanying them, as reasonable. This is contrary to the traditional view of OCD (i.e.
symptoms are senseless but have overwhelmingly compelling aspects, such as anxiety
reduction). The extent to which one has poor insight is said to predict treatment outcome, and
has been referred to in the literature as `overvalued ideation' (Foa, 1979; Kozak & Foa, 1994).
Overvalued ideation has appeared in the literature in the past, prior to Foa (1979) and
Kozak and Foa (1994). However, its role in psychopathology was somewhat di€erent. For
example, Wernicke (1900) believed that overvalued ideas were driven by a€ect, leading to
impaired judgement. In particular, overvalued ideas were considered by Wernicke to be
misperceptions linked to a€ective experience and were resistant to alteration. Further,
Wernicke stated that overvalued ideas are misattributions due to misperceptions whereby the
individual focuses narrowly on certain observations to the exclusion of others. Although he
does not speak about OCD in particular, certainly obsessions become overvalued when the
individual ®lters incoming information. In this process, some aspects of incoming information
are strongly believed as accurate and other aspects are ignored or dismissed. On the other
hand, Jaspers (1913) viewed overvalued ideas as the qualities of individuals with strong
convictions for social reform, and contrasted this with delusions, which he reserved for
individuals with functional impairments. Further, Jaspers suggested that overvalued ideas could
be challenged and were better attributed to personality features than pathology.
As described in Kozak and Foa (1994), overvalued ideation is currently conceptualized as
falling in the middle of a continuum between rational thought processes on the one end, and
delusions on the other. As part of the continuum, obsessions in the traditional sense are
pathologies of thought which are viewed as senseless. As individuals experience their obsessions
as more realistic, `insight' gets poorer and overvalued ideation increases until it becomes
delusional in quality. Indeed, others have cited evidence, either in case or controlled studies,
that some patients with OCD have symptoms that are delusional in nature (Insel & Akiskal,
1986; Lelliott & Marks, 1987; Eisen & Rasmussen, 1993). For example, Eisen and Rasmussen
(1993) found that of 475 patients with OCD, 14% had some type of psychotic symptom
(delusions), although only 6% of the total sample had their psychotic symptom directly related
to their OCD.
Recent research has shown that OCD shares greater commonalities with schizophrenia-
spectrum disorders than previously considered. For example, OCD patients were reported to
share similar perceptual and cognitive disturbances as patients with schizophrenia (Yaryura-
Tobias, Campisi, McKay & Neziroglu, 1995). Speci®cally, the results of this study indicated
that OCD patients experience visual (e.g. images), auditory (e.g. repetitive songs, melodies,
 F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902 883

sound hyperarousal) and tactile (e.g. sliminess, texture) disturbances. Close resemblance to the
above ®ndings is the series of studies by Enright (1996), where aberrations were observed in a
group of patients with OCD which were more similar to a group of patients with schizotypy or
schizophrenia and dissimilar to other anxiety disorder patients. Furthermore, the comparison
was substantiated on negative priming tasks (time required to state a color word placed on a
di€erent color background). This area of research requires greater examination as questions
have been raised as to the adequacy of determining that OCD is a separate entity apart from
the anxiety disorders. For example, Enright (1996) has suggested that OCD is more similar to
a schizotypy, with appreciable debate to follow (Salkovskis, 1996).
If overvalued ideas in general, and OCD with overvalued ideas in particular, are considered
linked to psychotic-like OCD or to schizophrenia spectrum disorders, then the impact of
aberrant symptomatology on outcome is important to consider. For example, Basoglu, Lax,
Kasvikis and Marks (1987) found that bizarreness of rituals was predictive of poorer treatment
outcome, although a later study using the same patient sample did not ®nd this to be true
(Lelliot, Noshirvawi, Basoglu, Marks & Monteiro, 1988). Patients with bizarreness and ®xity of
belief responded as well to treatment. Foa, Steketee, Grayson and Doppelt (1983) report that
those patients who report that their phobic reactions are realistic and feel that the compulsions
prevent actual disastrous consequences respond poorly to treatment. It has also been shown
recently that overvalued ideas could be modi®ed by exposure therapy, although the
multidimensional nature of this cognitive diculty is emphasized (Ito, de Araujo, Hemsley &
Marks, 1995)1. Speci®cally, it was found that patients who had frank recognition of the
senselessness of their beliefs also strongly believed that the compulsions were necessary.
Kozak and Foa (1994) suggest that overvalued ideas are di€erent from delusions. For both
the delusional and OCD patients with overvalued ideas the thought content is bothersome.
However, the former accepts and is less bothered by the behaviors performed as a response to
the thought whereas the latter is highly bothered by the behavioral response to the thought but
has a high degree of conviction that the feared outcome has a high probability of occurring.
This is best illustrated by an example of an individual who feels compelled to wash out of a
concern he or she will contract hepatitis. The delusional patient may believe that others are
continually bringing hepatitis into his/her presence and that washing is one reasonable response
to avoid becoming sick. The OCD patient strongly believes that he/she could contract hepatitis
by a number of benign sources (e.g. countertops) and would feel speci®cally compelled to
wash. Whereas both would be bothered by the misperceived presence of hepatitis, the patient
with OCD and strong belief conviction is bothered by the washing behavior but indicates that
the probability of contracting the illness is too high to take the risk of not washing.
At this point, the nature of overvalued ideation has been more clearly explicated, due in
large part to the recent work of Kozak and Foa (1994). However, there is still no instrument
that reliably measures overvalued ideas. There have been attempts at measuring overvalued
ideas in OCD, but these have been in the form of single item assessments (item 11 of the Yale±

1
It is important to note that these authors refer to overvalued ideation as belief ®xidity. Further, the data from
Ito et al. (1995) showed that some aspects of overvalued ideas improved in the absence of direct intervention.
Nevertheless, the balance of evidence supports the view that at least some aspects of this psychopathological con-
truct predict treatment outcome.
884 F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902

Brown Obsessive±Compulsive Scale (Y-BOCS)-clinician rating of insight; Goodman et al.,

1989), clinician based ratings of bizarreness and ®xidity that do not have reliability or validity
data (Lelliot et al., 1988), or clinician ratings on ®ve items designed for the DSM IV ®eld trial
(Foa et al., 1995). The latter instrument was designed to measure patients' con®dence that the
harmful consequence will happen; patients' recognition of the disparity of their beliefs from
conventional beliefs; patients? understanding of why they have unrealistic beliefs; ¯exibility in
changing mistaken beliefs and bizarreness of obsessive ideas. Internal item consistency was very
low (range 0.05±0.57) and therefore the single item, `belief in consequence' was used to address
whether the individual believes that his or her feared consequences are reasonable. Most
recently, The Brown Assessment of Beliefs Scale (BABS; Eisen et al., 1998) was developed as a
means of assessing overvalues ideas. This seven-item scale was developed as a measure to
``assess delusions across a wide range of psychiatric disorders'' (p. 102). Given the sampling of
participants and the domains assessed by this instrument, it is not clear that the measure has
adequate speci®city to OCD. This study utilized 20 patients diagnosed with OCD, 20
diagnosed with body dysmorphic disorder and 10 with mood disorders with psychotic features
in order to establish reliability and validity of the BABS. In fact, the authors conclude that the
BABS is essentially a ``valid measure of delusionality'' (p. 106) which is not necessarily
equivalent to our description of overvalued ideation.
The purpose of the present study is to describe the development of a clinician administered
rating scale for measuring overvalued ideas in an OCD sample. Although Kozak and Foa
(1994) suggest that overvalued ideas ¯uctuate in strength, strength of belief in obsessional ideas
have been implicated in therapy outcome (Foa, 1979; Foa et al., 1983). It stands to reason,
therefore, that in some patients overvalued ideas are more persistent and represent a stable
aspect of their obsessional problems. It was therefore hypothesized that overvalued ideas would
be stable in at least a proportion of the OCD patients sampled. The following two studies
describe the development of a structured clinician administered instrument for measuring
overvalued ideas. Study 1 is a preliminary investigation of the instrument and study 2 is an
update of the instrument employing structured prompts for the interviewer to allow naive users
to assess for the presence of overvalued ideas.

2. Study 1

2.1. Method

2.1.1. Participants
One hundred and two patients diagnosed with OCD participated. The diagnosis of OCD
was made independently by a psychologist and a psychiatrist. Data related to diagnoses was
gathered by clinical interviews, self-report and clinician rated scales in accordance with the
diagnostic criteria speci®ed in DSM-IV. OCD diagnoses were later con®rmed at weekly
multimember interdisciplinary sta€ meetings. Of the 102 patients, 57 were female and 45 were
male. The mean age of the sample was 24.3 (S.D. = 16.7), with the age ranging from 10 to 72
years. The average age of onset was 17.1 (S.D. = 14.1).
 F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902 885

2.1.2. Instruments

2.1.2.1. Overvalued ideas scale (OVIS)

Item development. The OVIS was developed by identifying a representative sample of items
associated with overvalued ideas. A major guide for developing items was the theoretical
formulation of overvalued ideation put forth by Kozak and Foa (1994). For example, items
assessed the extent to which patients endorsed the idea that their obsessions or associated
compulsions was reasonable, how others viewed the necessity of carrying out compulsions or
shared their beliefs and whether individuals felt that their symptoms were unusual. Once a
sample of items was identi®ed, the preliminary scale was given to professionals familiar with
OCD and rated to determine appropriateness for measuring overvalued ideas. The ®nal scale
was composed of nine items, each of which can be assessed for up to three beliefs associated
with OCD. Each belief provided by the patient is rated on the nine items. This was done since
many patients present for treatment with a heterogeneous array of symptoms that may or may
not be related to one another. Therefore, whereas one belief may be overvalued, another may
not. An initial pilot phase revealed that the OVIS is unreliable when administered as a self-
report instrument.
Item content and administration. The scale assessed the following domains of thought process
relative to presenting symptomatology: bizarreness (e.g. unusualness of obsessions such as the
fear of a tick imbedding itself into the head of one's penis), belief accuracy (e.g. whether this
imbedding can indeed occur), ®xidity (e.g. whether the belief persists), reasonableness (e.g. how
likely it is that imbedding will occur), e€ectiveness of compulsions (e.g. in reducing feared
outcome probability), pervasiveness of belief (e.g. among the public and other persons they
know), reasons others do not share the belief (e.g. others are unaware or misinformed) and two
items assessing stability of belief. The two items assessing belief stability were to address the
extent to which overvalued ideas remain unchanged over time, as hypothesized. The nine items
were assessed for up to three separate beliefs. Many participants o€ered fewer than three beliefs.
A total of 52 subjects o€ered one belief, 21 o€ered two beliefs and 29 o€ered three beliefs.
Scale scoring. Ratings of up to three obsessive beliefs were obtained for each patient
assessed. In order to ease understanding of scores, calculations were conducted such that
regardless of the number of beliefs presented by the patient, the scores would range from 0 to
10. This was achieved by ®rst adding the ratings for each of the nine items for each belief
(resulting in three total scores) and then dividing each total score by the number of items on
the scale (nine). This results in three average scores each ranging from 0 to 10. These three
scores were then added and divided by three, resulting in a total overvalued ideas score ranging
from 0 to 10. The present authors judged the average to be a representative ®gure given the
large sample size. For the purposes of data analysis, individual beliefs were rated for internal
consistency, to determine the extent to which symptoms rated were heterogeneous.

2.1.2.2. Yale±Brown obsessive±compulsive scale (Y-BOCS). The Y-BOCS (Goodman et al.,

1989) is a 10-item scale which assesses OCD on two dimensions: obsessions and compulsions.
Although the Y-BOCS was originally designed to yield a total score, which is a composite of
the obsessions and compulsions subscales, recent research has indicated that the total score
may not be factorially robust. Some have suggested that the subscales should be considered
886 F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902

separate (McKay, Danyko, Neziroglu & Yaryura-Tobias, 1995; McKay, Neziroglu, Yaryura-
Tobias & Stevens, 1998), while others suggest the single factor solution is satisfactory (Amir,
Foa & Coles, 1997). The subscales are comprised of ®ve items each, assessing the same
dimensions for obsessions and compulsions, as follows: frequency, interference, distress,
resistance and control. Each item is rated from 0 to 4, with higher numbers associated with
greater pathology. In addition to the items comprising the subscales, there are supplemental
items not contributing to the scores. For the present study, item 11 (judged to assess insight, or
overvalued ideas) was also administered.

2.1.2.3. Hamilton depression rating scale (HAM-D). The HAM-D (Hamilton, 1960) is a 17-
item clinician administered rating scale designed to assess vegetative and syndromal symptoms
of depression. Although it is not a diagnostic instrument (e.g. Williams, 1984), it is widely used
to assess the impact of depression in various populations. The measure provides a single total
score that can range from 0 to 52. There has been some recent issue in the reliability of this
measure since it has been informally modi®ed by a number of researchers (Gundy, Lunnen,
Lambert, Ashton & Tovey, 1994). For the purposes of this study, the original scale was used.
Internal consistency is adequate (0.76), interrater reliablity ranges from 0.8 to 0.9 and measures
of convergent validity range from 0.84 to 0.98 (Bech et al., 1975; Knesevich, Biggs, Clayton &
Ziegler, 1977; Rehm & O'Hara, 1985). Although no normative data exists, descriptive data
exists in the literature and is used as an approximate guide for clinical judgement (Bellack &
Hersen, 1988).

2.1.2.4. Hamilton anxiety rating scale (HAM-A). The HAM-A (Hamilton, 1959) is a 14-item
clinician rated scale which measures the impact and severity of anxiety in psychiatric
populations. The measure is chie¯y concerned with physiological aspects of anxiety, although
items also assess cognitive and motoric features. A single total score is obtained ranging from 0
to 56 and is sensitive to clinical change (Bellack & Hersen, 1988).

2.1.2.5. Structured clinical interview for DSM-III-R (SCID-P), psychotic screen. The SCID-P
(Spitzer, Williams, Gibbon & First, 1991) is a diagnostic instrument designed to identify the
presence of disorders enumerated in the third revised edition of the DSM (American
Psychiatric Association, 1987). At the time of this study, the edition of the SCID for DSM-IV
was not yet available. Only the psychotic screen portion of the SCID-P was administered in
this study. The presence of psychotic symptoms was assessed since prior literature examining
overvalued ideas has suggested that there is a quality of thought pathology in some OCD
patients which resembles delusions or hallucinations (e.g. Insel & Akiskal, 1986). Kappa
reliability for the psychotic screen portion of the SCID-P has been reported and ranges from
0.7 to 0.9, while interrater reliability as high as 1.0 has been reported (Segal, Hersen & Van
Hasselt, 1994). The screen identi®es auditory, visual, olfactory and tactile hallucinations and
delusional experiences.
 F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902 887

2.1.3. Procedure

2.1.3.1. Evaluator training. Evaluators were four of the authors of the present study (FN, DM,
KS, JAYT). In order to establish interrater consistency, each evaluator was trained to criterion
on interview scales based on the criteria set forth in the manual for the SCID-P and from
training sessions attended by all the authors. Once all evaluators were trained to criterion for
assessment and ratings with the instruments used in this study, patients were then evaluated at
baseline and at two weeks for stability of measures. A subset of patients were evaluated by
multiple raters (n = 15) during independent interviews to establish interrater reliability.
The extent of discriminant validity was assessed by examining the relationship between the
OVIS and the HAM-D and HAM-A. Although the research literature indicates that many
patients with OCD also su€er from anxiety and depression, it is suggested that overvalued
ideas are not directly related to a€ective states, as suggested in Kozak and Foa (1994).
Convergent validity, using the single items measure of overvalued ideas from the Y-BOCS
(referred to as insight), was administered as was the psychotic screen of the SCID-P. The
psychotic screen was examined since overvalued ideas are considered part of the continuum
from rational beliefs to delusions. Coecient kappa for the psychotic screen with this sample
was adequate (k = 0.82).

2.2. Results

The results are divided into the following sections. First, reliability data for the OVIS is
presented. This is divided into belief areas assessed (the scale measures up to three beliefs and
the associated overvalued ideas) and is provided for the total of each belief and the overall
scale aggregated across beliefs. Next, data relevant to the stability of the OVIS is provided.
Following is information on the interrater reliability. Convergent and discriminant validity
information is provided at the end.
Before discussing the reliability and validity of the OVIS, the range of scores obtained is
presented. For the OVIS, the obtained range was from 1.0 to 9.7 (M = 5.7, S.D. = 2.4). For
the Y-BOCS, the range was from 17 to 39 (M = 25.8, S.D. = 10.3). The range of scores for the
HAM-D was 2 to 23 (M = 4.7, S.D. = 6.2) and the HAM-A from 3 to 34 (M = 6.3,
S.D. = 6.1). Also worth noting, there were no signi®cant associations between age and the
major measures utilized, nor with age of onset and level of OVIS scores.

2.2.1. Internal consistency reliability of the OVIS

The total scale reliabilities for each belief was high, with scores of 0.88, 0.82 and 0.91
obtained for three beliefs areas assessed. The total internal consistency of the OVI, aggregated
across the three belief areas measured was 0.95. Although slightly lower, the internal
consistency reliabilities obtained at the two-week assessment were similar. Here, the individual
belief scores were 0.79, 0.87 and 0.76, with a total OVIS scale reliability of 0.91. Although
separate beliefs were rated to accommodate heterogeneous symptoms, it does not appear that
there was much variation in ratings as the total reliability was high.
888 F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902

2.2.2. Test±retest and interrater reliability of the OVIS

Test±retest reliability of the OVIS was established and determined to be high. When divided
for individual beliefs, the obtained correlation coecients were 0.86, 0.91 and 0.88. For the
total scale, the test±retest reliability was 0.93. In considering the interrater reliability, there was
adequate consistency of ratings between evaluators. Again, when examining interrater
reliability for individual belief scores, correlation coecients were 0.88, 0.92 and 0.90. For the
total scale aggregated across all three belief areas, the coecient was 0.95. The stability of the
OVIS over time, for the total score was signi®cant when examined using the intraclass
correlation (r = 0.74).

2.2.3. Convergent and discriminant validity

2.2.3.1. Convergent validity. Both psychotic screen data and symptomatic obsessive±compulsive
data were used to examinate convergent validity. In addition to this data, the single item
assessment of overvalued ideas previously available (item 11 of the Y-BOCS) would naturally
provide a useful metric of convergent validity. In this sample, there was adequate support for
convergent validity using these measures. In general, medium to large sized correlations were
obtained for these measures (Y-BOCS obsession and compulsion subscales, item 11, SCID-P
psychotic screen) with the OVIS. The range of correlations was from 0.44 to 0.83 (all
correlations signi®cant, p < 0.01).

2.2.3.2. Disciminant validity. Discriminant validity was assessed with correlations between the
OVIS and the HAM-A and HAM-D revealed a pattern of medium sized correlations using the
criteria of Cohen (1988). The correlation between the HAM-D and the OVIS at baseline was
0.47 ( p < 0.01). The correlation between the HAM-A and the OVIS at baseline was 0.53
( p < 0.01). Similar correlations were obtained at the two week administration (0.42 and 0.49,
respectively, p < 0.01). The data obtained for convergent and discriminant validity are
provided in Table 1.
Correlations between the Y-BOCS subscales and the OVIS, with the HAM-D and HAM-A
partialled out, were performed. In both instances, only the compulsions subscale was
signi®cantly related to OVIS (r = 0.64 and 0.55, respectively). All other associations were
nonsigni®cant.

2.2.4. Stability of OVI for extreme scores

Since there has been some question regarding the stability of OVI (Kozak & Foa, 1994),
patients scoring either a 3 or 4 on item 11 of the Y-BOCS were compared to those scoring a 1
or 2 on this item at baseline for the stability of their OVIS scores at two-week retest. A total of
15 participants met the criteria for high scorers, while 28 met the criteria for low scorers. This
analysis indicated that patients scoring higher for overvalued ideas at baseline with a single
item impression of OVI likewise had greater stability of OVIS scores (r-high scores = 0.94; r-
low scores = 0.71; z-di€erence = 1.99; p < 0.05). This is supportive of the idea that overvalued
ideas are held more stable for patients with higher scores than those with lower scores. A 2
(high, low OVIS) by 2 (time) ANOVA was conducted to further examine this relationship.
 F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902 889

Table 1
Convergent and discriminant validity assessment

Measure Study 1 OVIS Study 2 OVIS

Baseline
Y-BOCS item 11 0.67* 0.74*
Y-BOCS Obsessions 0.44* 0.52*
Y-BOCS Compulsion 0.83** 0.78*
Y-BOCS Total 0.61* 0.65*
SCID-P Psychotic Screen 0.58* 0.51*
Two-week interval
Y-BOCS item 11 0.75* 0.83*
Y-BOCS Obsessions 0.53* 0.60*
Y-BOCS Compulsions 0.80* 0.77*
Y-BOCS Total 0.66* 0.72*
Baseline
HAM-A 0.52* 0.56*
BAI 0.61*
HAM-D 0.46* 0.52*
BDI 0.55*
Two-week interval
HAM-A 0.44* 0.41
BAI 0.58*
HAM-D 0.48* 0.50*
BDI 0.39

Note: Y-BOCS represents the Yale-Brown Obsessive±Compulsive Scale, SCID-P the Structured Clinical Interview
for DSM-III-R, HAM-A the Hamilton Anxiety Rating Scale, HAM-D the Hamilton Depression Rating Scale, BAI
the Beck Anxiety Inventory and BDI the Beck Depression Inventory. Baseline OVIS ratings were correlated with
other measures at baseline and two-week OVIS ratings were correlated with the other ratings from the second week
assessment.
*
p < 0.05; **p < 0.01.

There was a signi®cant main e€ect of group membership (F(1, 41) = 7.51, p < 0.001) as well as
a signi®cant interaction (F(1, 41) = 4.97, p < 0.01), but not for time (F(1, 41) = 1.42, p > 0.05).

2.3. Discussion

This preliminary study (study 1) of the OVIS demonstrated that the scale possesses adequate
reliability. Further, very good convergent validity was established, but only limited discriminant
validity could be obtained. Because the preliminary scale was developed to determine that the
overvalued ideas could be reliably measured, the scale did not have extensive prompts for
interviewers. Instead, as part of the training, there was implicit prompts and agreement was
established regarding the probing of items. Study 2 was therefore conducted to establish that
should more formal prompts and anchor points be included, that the scale would possess at least
the same level of reliability and validity. In addition, it was believed that the probes and anchor
points for each item might make the scale more user friendly.
890 F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902

3. Study 2

3.1. Method

3.1.1. Participants
A total of 40 subjects participated in study 2. As in study 1, all subjects were diagnosed with
OCD based upon a separate clinical evaluation by a psychologist and psychiatrist, with
diagnoses con®rmed at an interdisciplinary sta€ meeting. There were 18 males and 22 females.
The mean age was 27.8 (S.D. = 8.9).

3.1.2. Instruments
All instruments were the same as in study 1, with the addition of the Beck Anxiety and Beck
Depression Inventories. These measures were added to broaden understanding of the
psychometric relationship of the OVIS to other commonly used measures. In addition, because
adequate discriminant validity was not achieved with the Hamilton scales, another two
measures of depression and anxiety were added to further examine this question.

3.1.2.1. Beck anxiety inventory (BAI). The BAI (Beck, Epstein, Brown & Steer, 1988) is a 21-
item self-report instrument designed to assess physiological and cognitive aspects of anxiety.
Internal reliability of this scale is good (a = 0.92), test±retest reliability is adequate (a = 0.75).
Convergent and discriminant validity correlations with the Hamilton Anxiety and Depression
Rating Scales, respectively, are somewhat lower (0.51 and 0.25).

3.1.2.2. Beck depression inventory (BDI).. The BDI (Beck, Ward, Mendelson, Mock &
Erbaugh, 1961) is a 21-item self-report instrument designed to assess symptomatic and
syndromal aspects of depression. Internal reliability is good (a = 0.86) as is convergent validity
(0.66).

3.1.2.3. Modi®ed overvalued ideas scale. The items for the OVIS remained the same as in study
1 with the exception of the inclusion of a 10th and 11th item to assess for one's degree of
resistance of the belief and the duration of belief stability. As noted, there was the addition of
anchor points for each item and prompts for each question, which allowed for a clearer
structure to the test. Also, whereas in study 1 patients were rated for three related or disparate
beliefs, the ®nal version examines the primary belief for each patient assessed. Anchor points
are provided to assist the clinician in arriving at a rating. For example, for question 2
(concerning reasonabless of belief) the following probes are provided: ``How reasonable is your
belief?; Is your belief justi®ed or rational?; Is the belief logical?'' The ®nal version of the scale
is provided in Appendix A.
Evaluator training was the same as in study 1. Evaluators for study 2 were three of the
present authors (FN, KS and JAYT). All patients were evaluated with the OVIS, HAM-A,
HAM-D, BDI, BAI, Y-BOCS (including item 11, measuring overvalued ideas) and the
psychotic screen portion of the SCID. Subjects were evaluated at an initial assessment period
and then two weeks later and prior to initiating treatment. A subset of this group (n = 10)
 F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902 891

were evaluated at the initial time point by two independent raters to estimate interrater
reliability.

3.2. Results

The reliability and validity of the OVIS for study 2 is presented below. First, however, the
range of scores obtained is presented. For the OVIS, the obtained range was from 1.6 to 8.7
(M = 4.6, S.D. = 3.1). For the Y-BOCS, the range was from 21 to 36 (M = 26.7, S.D. = 8.8).
The range of scores for the HAM-D was 2 to 20 (M = 3.9, S.D. = 5.1) and the HAM-A from
4 to 21 (M = 7.1, S.D. = 5.7). Finally, the range of scores for the BDI was 3 to 17 (M = 5.9,
S.D. = 4.4) and the BAI was 8 to 18 (M = 11.2, S.D. = 3.5).
The internal consistency reliability for the total scale was high (r = 0.85) at the initial
evaluation and at two-week retest (r = 0.81). Test±retest reliability of the OVIS was also
established (r = 0.80), as was stability with the intraclass correlation (r = 0.77). Finally,
interrater reliability on the subset of patients evaluated independently at the initial evaluation
was adequate, using the intraclass correlation coecient (r = 0.81).

3.2.1. Convergent validity

Both psychotic screen and symptomatic obsessive compulsive data were used to examine
convergent validity. In addition to this data, the single item assessment of overvalued ideas
previously available (item 11 of the Y-BOCS) would naturally provide a useful metric of
convergent validity. Given this, the correlations obtained with these measures were typically
medium to large in size (using Cohen's de®nition of e€ect size) for all these measures (Y-BOCS
obsession and compulsion subscales, item 11, SCID-P psychotic screen). The range of
correlations was from 0.50 to 0.83 (all correlations signi®cant, p < 0.05).

3.2.2. Discriminant validity

Discriminant validity was assessed by the relation between the OVIS and the measures for
depression (HAM-D, BDI) and anxiety (HAM-A, BAI). The correlations between the OVIS
and each depression measure was relatively large at baseline (correlation with HAM-D,
r = 0.52; correlation with BDI, r = 0.55, p < 0.05). For anxiety measures, there were likewise
large correlations (with HAM-A and BAI, r = 0.56 and r = 0.61, respectively). These results
are also shown in Table 1.
Correlations between the Y-BOCS subscales and the OVIS revealed signi®cant correlations
when the HAM-D was partialled out. Speci®cally, the relation with the obsession (r = 0.49),
compulsions (r = 0.58) and the total scale (r = 0.53) were all signi®cant. When the same
relationships were examined with the HAM-A partialled out, only the compulsions subscale
was signi®cant (r = 0.61) whereas the obsessions (r = 0.27) and total scale (r = 0.36) were not.

3.3. Discussion

The ®ndings from study 2 were similar to that obtained for study 1. Speci®cally, the ®nal
version of the OVIS, using a structured interview format with probes and anchor points
demonstrated moderate high reliability and convergent validity. As in the preliminary version
892 F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902

of the OVIS, discriminant validity was poor as established with the additional use of self-report
measures of depression and anxiety (i.e. the Beck Inventories). This is, however, not unusual in
measures of obsessive±compulsive symptoms as has been noted for many major self-report and
interview measures of OCD (Taylor, 1995). In addition, perhaps the poor discriminant validity
may be due to overvalued ideas being related to a€ective states as suggested by Wernicke
(1900). The stability of extreme scores was not examined in study 2 given the sample size.

4. General discussion

These studies illustrate the development, reliability and validity of a scale designed to assess
overvalued ideas in a group of obsessive±compulsive patients. This is a group which has been
identi®ed in the research as having various levels of overvalued ideas associated with their
psychopathology, but to date no psychometrically sound instrument has been developed for its
assessment. The only previously available instruments were single item measures (item 11 of
the Y-BOCS), a clinician based rating of bizarreness and ®xidity with no reliability or validity
measurements available (Lelliot et al., 1988), or dichotomously assessed overvalued ideas from
a series of clinical criteria (as in the recent DSM-IV ®eld trials; Foa et al., 1995). or the more
recent BABS that the authors not is a measure of delusions (Eisen et al., 1998). Overvalued
ideas hold a special importance in understanding and treating OCD since it has been
implicated as a prognostic indicator (Foa, 1979). However, the conceptualization of overvalued
ideas was not clearly explicated until recently (Kozak & Foa, 1994) and was variously referred
to as obsessive±compulsive psychosis (Insel & Akiskal, 1986), ®xed ideas (as in ®eld trials
clinician rating scale; Foa et al., 1995) or poor insight (as in item 11 of the Y-BOCS). Indeed,
overvalued ideas are still referred to as `poor insight' by the optional identi®er in the current
edition of the DSM.
The OVIS was developed in an e€ort to quantify the extent to which one has overvalued
ideas regarding their obsessions and compulsions. Since Kozak and Foa (1994) suggest that
overvalued ideas range from rational identi®cation of symptomatic behavior at one extreme to
delusional belief in symptoms at the other, it seems that dichotomous assessments would fall
short of capturing the nature of this pathology. The extent of overvalued ideation regarding
symptoms could also provide greater information regarding prognosis and relapse in obsessive±
compulsive patients. The OVIS was found to have very good reliability and stability, as well as
interrater consistency.
Convergent validity with the OVIS was found to be good, with large size correlations
obtained with measures theoretically related to overvalued ideas. Again, assuming that
overvalued ideas form part of a continuum from rationality to delusions, measures at both
extremes were used, along with a single item measure on the YBOCS identi®ed as measuring
overvalued ideas. Discriminant validity, on the other hand, was poor. Instead, there were
similar correlations found with measures considered discriminative in nature (HAM-D, HAM-
A, BDI, BAI). As it has been stated earlier it is not uncommon for other measures of obsessive
compulsive symptoms to lack discriminant validity with unrelated assessments of
psychopathology (Taylor, 1995) and also overvalued ideas may be related to a€ective states.
 F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902 893

Although convergent validity was obtained for the OVIS in relation to both the obsessions
and compulsions subscales of the Y-BOCS, another noteworthy aspect of this ®nding is that
overvalued ideas increased as obsessive±compulsive symptoms increased when measured with
the Y-BOCS. Since overvalued ideas have been characterized as phenomena in which the
patient suspends rationality regarding the nature of feared situations, one would expect lower
ratings of obsessions and compulsions as overvalued ideas increase. It is suggested, therefore,
that although patients having overvalued ideas evaluate their symptoms as necessary, they also
view the symptoms as problematic irrespective of rationality. This is consistent with previous
descriptions of overvalued ideation.
Stability of OVIS scores were examined by a dichotomous rating of a single item from the
Y-BOCS and examining stability of test±retest reliability in study 1. This analysis suggests that
the distribution of test±retest scores for lower and higher scoring patients on overvalued ideas
is heteroscedastic. Moreover, it implies that individuals with greater belief ®xidity hold those
beliefs with greater stability than individuals at the lower ranges. This ®nding requires greater
investigation as the comparison was based on a clinical rating from a single item scale. Indeed,
such tests may be limited due to restricted range of scores and future tests may examine the
stability with other measures of associated ®xidity of belief, or measures of psychosis. The
stability of overvalued ideas was not evaluated in study 2.
Of note in the present analysis is the ®nding that OVIS scores were correlated moderately
(0.44 and 0.51) with the Y-BOCS obsessions scale and high (0.83 and 0.81) with the Y-BOCS
compulsions scale. This may be attributed to some diculties which patients have in
distinguishing obsessions from the responses engaged in to alleviate resultant anxiety. For
example, others have noted that patients have diculty responding to the item regarding
resistance to obsessions when assessed with the Y-BOCS (Woody, Steketee & Chambless,
1995). Likewise, Steketee, Frost and Bogart (1996) found that the obsessions subscale of the Y-
BOCS had generally poorer test±retest reliability than the compulsions subscale. They attribute
the diculties to the resistant item, but diculties with the subscale remained when that item
was deleted. The reliable assessment of obsessions, in particular as they relate to overvalued
ideas, requires further investigation. In addition, another explanation for this ®nding may be
that the more compelled a patient is to perform a compulsion and the longer he or she engages
in these behaviors, the stronger is the strength of the belief. It may be that although a patient
may spend a lot of time obsessing, he or she may not necessarily believe in the content of the
obsession but extensively performing a compulsion may be the result of an overvalued idea.
Perhaps it is for this reason that the compulsion subscale correlated higher with the OVIS
measure than the obsessions subscale. This needs further exploration.
Overall, it appears that the OVIS as constructed in the present studies is a reliable measure
with adequate test±retest stability and interrater consistency. Adequate support was found for
convergent validity, but support for discriminant validity was not obtained. Validity
considerations require further investigation, using either the same instruments for replication or
other instruments theoretically unrelated to overvalued ideation. Of note, it would be informative
to examine the relation between OVIS scores and self-report measures of obsessions and
compulsions. Because overvalued ideas are theoretically considered prognostic in the treatment
of OCD, it is hoped that by establishing a systematic, quantitative assessment tool that empirical
examinations of this feature of pathology in therapy outcome may be examined.
894 F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902

Appendix A
F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902 895
896 F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902
F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902 897
898 F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902
F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902 899
900 F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902
 F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902 901

References

American Psychiatric Association (1987). Diagnostic and statistical manual of mental disorders (revised 3th ed.). Washington, DC: APA.
American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: APA.
Amir, N., Foa, E. B., & Coles, M. (1997). Factor structure of the Yale±Brown Obsessive±Compulsive Scale. Psychological Assessment,
9, 312±316.
Basoglu, M., Lax, T., Kasvikis, Y., & Marks, I. M. (1987). Predictors of improvement in obsessive±compulsive disorder. Journal of
Anxiety Disorders, 2, 299±317.
Bech, P., Gram, L. F., Dein, E., Jacobson, O., Vitger, J., & Bolwig, T. G. (1975). Quantitative ratings of depressive states. Acta
Psychiatrica Scandinavica, 51, 161±170.
Beck, A. T., Epstein, N., Brown, G., & Steer, R. A. (1988). An inventory for measuring clinical anxiety: psychometric properties.
Journal of Consulting and Clinical Psychology, 56, 397±893.
Beck, A. T., Ward, C. H., Mendelson, M., Mock, J. E., & Erbaugh, J. K. (1961). An inventory for measuring depression. Archives of
General Psychiatry, 4, 561±571.
Bellack, A. S., & Hersen, M. (1988). Behavioral assessment (3rd ed.). NY: Pergamon.
Cohen, J. (1988). Statistical power analysis for the social sciences (2nd ed.). Hillsdale, NJ: Erlbaum.
Eisen, J. L., & Rasmussen, S. A. (1993). Obsessive±compulsive disorder with psychotic features. Journal of Clinical Psychiatry, 54, 373±
379.
Eisen, J. L., Phillips, K. A., Baer, L., Beer, D. A., Atala, K. D., & Rasmussen, S. A. (1998). The Brown Assessment of Beliefs Scale:
reliability and validity. American Journal of Psychiatry, 155(1), 102±108.
Enright, S. J. (1996). Obsessive±compulsive disorder: anxiety disorder or schizotype? In R. Rapee (Ed.), Current controversies in the
anxiety disorders (pp. 161±190). NY: Guilford.
Foa, E. B. (1979). Failure in treating obsessive±compulsives. Behaviour Research and Therapy, 17, 169±176.
Foa, E. B., Kozak, M. J., Goodman, W. K., Hollander, E., Jenike, M. A., & Rasmussen, S. A. (1995). DSM-IV ®eld trial: obsessive±
compulsive disorder. American Journal of Psychiatry, 152, 90±96.
Foa, E. B., Steketee, G., Grayson, J. B., & Doppelt, H. G. (1983). Treatment of obsessive±compulsives: when do we fail? In E. Foa &
P. M. G. Emmelkamp (Eds.), Failures in behavior therapy (pp. 10±57). NY: Wiley.
Goodman, W. K., Price, L. H., Rasmussen, S. A., Mazure, C., Fleischman, R. L., Hill, C. L., Heninger, G. R., & Charney, D. S.
(1989). The Yale±Brown Obsessive±Compulsive Scale. I. Development, use and reliability. Archives of General Psychiatry, 46,
1006±1011.
Gundy, C. T., Lunnen, K. M., Lambert, M. J., Ashton, J. E., & Tovey, D. R. (1994). The Hamilton Rating Scale for Depression: one
scale or many?. Clinical Psychology: Science and Practice, 1, 197±205.
Hamilton, M. (1959). The assessment of anxiety states by rating. British Journal of Medical Psychology, 32, 50±55.
Hamilton, M. (1960). A rating scale for depression. Journal of Neurology, Neurosurgery and Psychiatry, 23, 56±62.
Insel, T. R., & Akiskal, H. S. (1986). Obsessive±compulsive disorder with psychotic features: a phenomenological analysis. American
Journal of Psychiatry, 143, 1527±1533.
Ito, L. M., de Araujo, L. A., Hemsley, D. R., & Marks, I. M. (1995). Beliefs and resistance in obsessive±compulsive disorder: obser-
vations from a controlled trial. Journal of Anxiety Disorders, 9, 269±281.
Jaspers, K. (1913). Psicopathologia general (Allgemeine psychopatologie) (R. O. Saubidet, Trans.). Bueno Aires, Argentina: Beta
Publishers.
Knesevich, J. W., Biggs, J. T., Clayton, P. J., & Ziegler, V. E. (1977). Validity of the Hamilton Rating Scale for Depression. British
Journal of Psychiatry, 131, 49±52.
Kozak, M. J., & Foa, E. B. (1994). Obsessions, overvalued ideas and delusions in obsessive±compulsive disorder. Behaviour Research
and Therapy, 32, 343±353.
Lelliott, P., & Marks, I. (1987). Management of obsessive±compulsive rituals associated with delusions, hallucinations and depression:
a case report. Behavioural Psychotherapy, 15, 77±87.
Lelliot, P. T., Noshirvani, H. F., Basoglu, M., Marks, I. M., & Monteiro, W. O. (1988). Obsessive compulsive beliefs and treatment
outcome. Psychological Medicine, 18, 697±702.
McKay, D., Danyko, S. J., Neziroglu, F., & Yaryura-Tobias, J. A. (1995). Factor structure of the Yale±Brown Obsessive±Compulsive
Scale: a two-dimensional measure. Behaviour Research and Therapy, 33, 865±869.
McKay, D., Neziroglu, F., Yaryura-Tobias, J. A., & Stevens, K. (1998). The Yale±Brown Obsessive±Compulsive Scale: factor analytic
®ndings. Psychological assessment, in press.
Rehm, L. P., & O'Hara, M. W. (1985). Item characteristics of the Hamilton Rating Scale for Depression. Journal of Psychiatric
Research, 19, 31±41.
Salkovskis, P. M. (1996). Understanding of obsessive±compulsive disorder is not improved by rede®ning it as something else. In R. M.
Rapee (Ed.), Current controversies in the anxiety disorders (pp. 191±200).
902 F. Neziroglu et al. / Behaviour Research and Therapy 37 (1999) 881±902

Segal, D. L., Hersen, M., & Van Hasselt, V. B. (1994). Reliability of the Structured Clinical Interview for DSM-III-R: an evaluative
review. Comprehensive Psychiatry, 35, 316±327.
Spitzer, R. L., Williams, J. B. W., Gibbon, M., & First, M. (1991). Manual for the structured clinical interview for DSM-III-R with
psychotic screen. Washington, DC: American Psychiatric Press.
Steketee, G., Frost, R., & Bogart, K. (1996). The Yale±Brown Obsessive±Compulsive Scale: interview versus self-report. Behaviour
Research and Therapy, 34, 675±684.
Taylor, S. (1995). Assessment of obsessions and compulsions: reliability, validity and sensitivity to treatment e€ects. Clinical
Psychology Review, 15, 261±296.
Wernicke, C. (1900). Grundisse der Psychiatrie (Foundations of psychiatry). Leipzig: Verlag.
Williams, J. M. G. (1984). The psychological treatment of depression. NY: Free Press.
Woody, S. R., Steketee, G., & Chambless, D. L. (1995). Reliability and validity of the Yale±Brown Obsessive±Compulsive Scale.
Behaviour Research and Therapy, 33, 597±605.
Yaryura-Tobias, J. A., Campisi, T., McKay, D., & Neziroglu, F. A. (1995). Overlapping features of obsessive±compulsive disorder and
schizophrenia. Neurology, Psychiatry and Brain Research, 3, 143±148.