Content uploaded by Harlan R. Juster
Author content
All content in this area was uploaded by Harlan R. Juster
Content may be subject to copyright.
Psychological Assessment Copyright 1997 by the American Psychological Association, Inc.
1997, Vol. 9, No. 1, 21-27 1040-3590/97/$3.00
Validation of the Social Interaction Anxiety Scale
and the Social Phobia Scale Across the Anxiety Disorders
Elissa J. Brown, Julia Turovsky, Richard G. Heimberg, Harlan R. Juster,
Timothy A. Brown, and David H. Barlow
The University at Albany, State University of New York
The psychometric adequacy of the Social Interaction Anxiety Scale (SIAS; R. P. Mattick & J. C.
Clark, 1989), a measure of social interaction anxiety, and the Social Phobia Scale (SPS; R. E Mattick &
J. C. Clarke, 1989), a measure of anxiety while being observed by others, was evaluated in anxious
patients and normal controls. Social phobia patients scored higher on both scales and were more likely
to be identified as having social phobia than other anxious patients (except for agoraphobic patients
on the SPS) or controls. Clinician-rated severity of social phobia was moderately related to SIAS and
SPS scores. Additional diagnoses of mood or panic disorder did not affect SIAS or SPS scores among
social phobia patients, but an additional diagnosis of generalized anxiety disorder was associated with
higher SIAS scores. Number of reported feared social interaction situations was more highly correlated
with scores on the SIAS, whereas number of reported feared performance situations was more highly
correlated with scores on the SPS. These scales appear to be useful in screening, designing individual-
ized treatments, and evaluating the outcomes of treatments for social phobia.
Social phobia was first introduced into the psychiatric nomen-
clature in the third edition of the
Diagnostic and Statistical
Manual of Mental Disorders ( DSM-III;
American Psychiatric
Association, 1980) and has since remained a part of the diagnos-
tic lexicon
(DSM-III-R,
American Psychiatric Association,
1987;
DSM-IV,,
American Psychiatric Association, 1994). It is
defined in
DSM-1V
(p. 416) as "a marked and persistent fear
of one or more social or performance situations in which the
person is exposed to unfamiliar people or to possible scrutiny
by others. The individual fears that he or she will act in a
way (or show anxiety symptoms) that will be humiliating or
embarrassing." Although the diagnosis of social phobia has
been established for over a decade, only recently have research-
ers begun to investigate its assessment and treatment (Heimberg,
Liebowitz, Hope, & Schneier, 1995; Stein, 1995). To accurately
diagnose social phobia and evaluate treatment efficacy, assess-
ment devices that effectively measure its various elements are
needed (Cox & Swinson, 1995; McNeil, Ries, & Turk, 1995).
Although there are several self-report measures of social anxi-
Elissa J. Brown, Julia Turovsky, Richard G. Heimberg, Harlan R.
Juster, Timothy A. Brown, and David H. Barlow, Center for Stress and
Anxiety Disorders, Department of Psychology, the University at Albany,
State University of New York. Timothy A. Brown and David H. Barlow
are now located at the Anxiety and Related Disorders Clinic at Boston
University.
Portions of this article were presented at the annual meeting of the
Association for Advancement of Behavior Therapy, Atlanta, Georgia,
November 1993. This research was supported by Grants 44119 and
39096 from the National Institute of Mental Health.
Correspondence concerning this article should be addressed to Rich-
ard G. Heimberg, who is now at the Social Phobia Program, Department
of Psychology, Temple University, Weiss Hall, 1701 North 13th Street,
Philadelphia, Pennsylvania 19122-6085. Electronic mail may be sent via
Internet to rheimber @nimbus.ocis.temple.edu.
21
ety, few have assessed differences in the types of situations
feared or avoided. These devices include the Social Phobia sub-
scale of the Fear Questionnaire (Marks & Mathews, 1979), the
Social Avoidance and Distress Scale (Watson & Friend, 1969),
the Fear of Negative Evaluation Scale (Watson & Friend, 1969),
and the Social Phobia and Anxiety Inventory (Turner, Beidel,
Dancu, & Stanley, 1989). The limitations of the first three scales
in the assessment of social phobia have been discussed else-
where (Heimberg, Mueller, Holt, Hope, & Liebowitz, 1992).
Although the Social Phobia and Anxiety Inventory assesses a
broad range of social situations and has substantial data support-
ing its reliability and validity among patients with social phobia,
it does not provide separate scores for different types of anxiety-
provoking situations. Assessment of anxiety responses to differ-
ent classes of situations should have utility for the planning of
individualized treatment interventions for patients with social
phobia. A set of scales developed by Mattick and Clarke (1989)
addresses this concern.
Several investigators have examined the types of situations
feared or avoided by patients with social phobia (e.g., Holt,
Heimberg, Hope, & Liebowitz, 1992; Turner, Beidel, Dancu, &
Keys, 1986). Liebowitz (1987) proposed two broad categories
of feared situations: those involving social interactions (e.g.,
initiating and maintaining conversations) and those in which the
person may be observed by others (e.g., formal public speaking,
eating or drinking in public). A similar distinction has been
discussed by Leary (1983b), who described such interactions
as either contingent or noncontingent. In contingent interactions
(e.g., conversations), people adjust their own behavior ac-
cording to their perceptions of the other person. In noncontingent
interactions (e.g., formal public speaking), people are less likely
to adjust their behavior, possibly because they receive little feed-
back from their audience. Mattick and Clarke (1989) also con-
ceptualized social anxiety as occurring in two similar types of
22 BROWN ET AL.
situations: those in which the person interacts with others and
those in which the person may be observed or scrutinized by
others. The Social Interaction Anxiety Scale (SIAS; Mattick &
Clarke, 1989) assesses
social interactional anxiety,
defined as
extreme distress when initiating and maintaining conversations
with friends, strangers, or potential mates. The companion Social
Phobia Scale (SPS; Mattick & Clarke, 1989) assesses
anxiety
when anticipating being observed or actually being observed
by other people
and when undertaking certain activities in the
presence of others (e.g., public speaking, eating, or writing).
Although the report on the initial development of the SIAS
and SPS remains unpublished, these scales increasingly have
been used in the evaluation of treatments for social phobia, and
each has been shown to be sensitive to the effects of cognitive-
behavioral treatments (Mattick & Peters, 1988; Mattick, Pe-
ters, & Clarke, 1989; see reviews by Cox & Swinson, 1995,
and McNeil et al., 1995; both scales are reprinted in Cox &
Swinson, 1995). In the development of the SIAS and SPS,
Mattick and Clarke (1989) generated a pool of 164 items from
existing inventories and from interviews with social phobic pa-
tients. This initial pool was reduced to 75 items with reliably
coded relevance to fears of social interaction or scrutiny by
others, which were then administered to samples of 243 patients
with a
DSM-III
diagnosis of social phobia, 481 college stu-
dents, 315 community volunteers, and small samples of patients
with agoraphobia or simple phobia. Examination of item-total
correlations resulted in the deletion of additional items and
selection of the final set of 20 scrutiny items (SPS) and 20
social interaction items (SIAS). With this development strategy,
the SIAS and SPS may be best considered as subscales of one
larger measure.
The SIAS and SPS appear to be psychometrically sound.
Mattick and Clarke (1989) reported Cronbach's alphas for each
scale for patients with social phobia, college students, commu-
nity volunteers, agoraphobics, and simple phobics that ranged
from .88-.93 for the SIAS and .89-.94 for the SPS. Test-
retest correlation coefficients exceeded .90 for both scales after
intervals of 1 and 3 months. Similar figures for internal consis-
tency (SIAS = .85-.90; SPS = .87-.93) were reported by
Heimberg et al. (1992) in a study of 66 patients with social
phobia, 50 community volunteers, and 53 undergraduate stu-
dents. In that study, 2-week test-retest reliability in the student
sample was .86 for the SIAS but only .66 for the SPS.
Concurrent validity of the SIAS and SPS has also been exam-
ined among patients with social phobia. Mattick and Clarke
(1989) found that both the SIAS and SPS were positively corre-
lated with scores on the Fear of Negative Evaluation Scale, the
Social Avoidance and Distress Scale, the Social Phobia subscale
of the Fear Questionnaire, and the Interaction Anxiety and Audi-
ence Anxiety Scales (Leary, 1983a). Ries et al. (1996) reported
that both scales were positively correlated with scores on the
Social Phobia and Anxiety Inventory. Heimberg et al. (1992)
also addressed convergent and divergent validity, reporting that
the SPS was more highly correlated with other measures of
performance fear, whereas the SIAS was more highly correlated
with other measures of social interactional anxiety. However,
these findings may not generalize to patients' reports of distress
in specific situations. Correlations between the SIAS and SPS
and degree of anxiety reported in actual situations involving
performance or social interaction should be examined to further
assess convergent and divergent validity.
Discriminant validity of the two scales has been examined in
a number of studies. Patients with social phobia scored higher
than undergraduates and community controls on both the SIAS
and SPS in the studies by Mattick and Clarke (1989) and Heim-
berg et al. (1992). Patients who met criteria for the generalized
subtype of social phobia reported greater social interactional
anxiety (SIAS) than did patients who did not meet these criteria
in four studies (E. J. Brown, Heimberg, & Juster, 1995; Heimb-
erg et al., 1992; Holt, Heimberg, & Hope, 1992; Ries et al.,
1996), although this difference was evident for the SPS in only
one study (E. J. Brown et al., 1995). In Mattick and Clarke
(1989), patients with social phobia also achieved higher scores
on the SIAS and SPS than did patients with agoraphobia or
simple phobia. Rapee, Brown, Antony, and Barlow (1992) also
reported that patients with social phobia scored higher on the
SIAS than did patients with various other anxiety disorders
and a group with no mental disorder. Despite these promising
findings, no study has adequately examined scores on both the
SIAS and SPS across the range of anxiety disorders. Neither
has the impact of social phobia as a secondary diagnosis on the
SIAS and SPS scores of patients with other principal anxiety
disorders been examined.
In this study, we extended previous research on the SIAS and
SPS in a number of ways. We investigated (a) the ability of
these measures to discriminate patients with social phobia from
other anxiety-disordered patients and to identify patients with
social phobia from a mixed anxious sample, (b) the relationship
between SIAS and SPS scores and severity of social phobia,
(c) the impact of additional diagnoses of anxiety or mood disor-
ders on scale scores among patients with social phobia, and (d)
the impact of secondary social phobia among patients with other
anxiety disorders. To examine whether the two scales assess fear
of different types of situations, we also compared the relation-
ship of SIAS and SPS scale scores to number of social interac-
tional and performance situations feared.
Panic attacks may occur in any of the anxiety disorders, in-
cluding social phobia. Additionally, social situations are a com-
mon sphere of worry in generalized anxiety disorder. Thus, we
examined the relationship between the SIAS and SPS and self-
report measures of fear of anxiety symptoms and worry to
evaluate the relationship between social interactional and obser-
vational fears and symptoms of panic disorder and generalized
anxiety disorder. We also investigated whether the number of
panic symptoms, commonly reported in performance situations,
was related to elevated scores on the SPS but not the SIAS.
Method
Participants
The participants were 104 women and 61 men who met
DSM-III-
R criteria for any of the anxiety disorders (other than posttraumatic
stress disorder) on the basis of structured interview. Fourteen women
and 7 men who were acquaintances (but not relatives) of the patients
with anxiety disorders also participated and were paid $60 for their
participation in this and other studies. Anxious participants from each
diagnostic group were randomly selected for this study from a larger
group of 753 patients with anxiety disorders who had received services
ASSESSMENT OF SOCIAL PHOBIA 23
at the Center for Stress and Anxiety Disorders between April 1988 and
January 1991. Diagnoses were determined by the Anxiety Disorders
Interview Schedule--Revised (ADIS-R; DiNardo & Barlow, 1988).
When two or more diagnoses were deemed to be present, the "principal"
diagnosis was the one receiving the highest rating of distress and interfer-
ence with life functioning. Patients' principal diagnoses were as follows:
social phobia (n = 50), panic disorder (n = 29), panic disorder with
agoraphobia (n = 27), generalized anxiety disorder (n = 20), simple
phobia (n = 19), and obsessive-compulsive disorder (n = 20). Persons
in the control group participated in the ADIS-R interview but received
no diagnoses of current or past mental disorder.
The proportion of men and women differed across groups, X 2 (6, N
= 186 = 15.33, p < .02. A greater percentage of patients with general-
ized anxiety disorder (60%) and social phobia (50%) were male than
was the case in any of the other groups (% male, range = 19-35%).
However, gender was unrelated to scores on either the SIAS, t(184) =
-1.51, p > .13, or the SPS, t(184) = -0.28,
ns.
Age also differed
across groups, F(6,179) = 4.42, p < .001. Duncan's multiple range
test (p < .05) indicated that patients with generalized anxiety disorder
were significantly older (M = 42.95,
SD
= 11.79) than all participants
other than patients with simple phobia (M = 38.89,
SD
= 12.59).
Patients with simple phobia were significantly older than patients with
panic disorder (M = 32.20,
SD
= 6.03) or obsessive-compulsive disor-
der (M = 30.90,
SD
= 6.77) as well as the participants in the normal
group (M = 31.48,
SD
= 9.50), none of whom differed from each
other. Patients with social phobia (M = 35.00,
SD
= 9.31) and panic
disorder with agoraphobia (M = 35.11,
SD
= I0.01 ) were intermediate
in age, differing only from the patients with generalized anxiety disorder.
However, like gender, age was unrelated to scores on the SIAS (r = .02,
ns)
or the SPS (r = -.10,
ns).
Because of their lack of relationship
to the SIAS and SPS, age and gender are not considered further in the
data analyses.
Measures
ADIS-R.
In addition to
DSM-111-R
diagnoses and their respective
ratings of severity (range of 0-8), the ADIS-R assesses fear and avoid-
ance in 11 situations often feared by patients with social phobia (rated
by the participant on a 0-4 scale: 0 =
no fear or avoidance, 4 =
very severe fear and avoidance).
These situations were categorized
as primarily interactional in nature (e.g., initiating or maintaining a
conversation, interacting with an authority figure) or primarily involving
performance or observation by others (e.g., formal public speaking).
Categorization of situations as interactional versus performance-oriented
was accomplished by a panel of nine clinical psychologists and doctoral
students experienced in the assessment and treatment of social phobia
as part of work previously conducted for the
DSM-IV
task force sub-
workgroup on social phobia (Schneier et al., in press). Disagreements
were rare and were resolved by Richard G. Heimberg. The number of
DSM-III-R
panic attack symptoms experienced in social situations was
also tallied. The ADIS-R has been shown to have high rates of interrater
reliability for the diagnosis of anxiety disorders (e.g., k = .79 for social
phobia; DiNardo, Moras, Barlow, Rapee, & Brown, 1993).
Questionnaires.
In the larger study conducted at the Center, partici-
pants completed questionnaires after the ADIS-R interview. Those ex-
amined in this study included the SIAS, SPS, Anxiety Sensitivity Index
(Peterson & Reiss, 1987), and Penn State Worry Questionnaire (Meyer,
Miller, Metzger, & Borkovec, 1990).
The SIAS consists of 20 items that are rated from 0
(not at all
characteristic or true of me)
to 4
(extremely characteristic or true of
me).
Items are self-statements describing one's representative reaction
to situations that involve social interaction in dyads or groups. The SIAS
is scored by summing the ratings (after reversing the 3 positively worded
items). Total scores range from 0 to 80, with higher scores representing
higher levels of social interaction anxiety.
The SPS also contains 20 items that are rated on a similar 0-4 scale.
Items pertain to situations or themes that involve being observed by
others (e.g., public speaking, eating or writing in public, etc.). All items
are negatively worded. As with the SIAS, scores range from 0 to 80,
with higher scores representing greater anxiety about being observed.
The Anxiety Sensitivity Index was used to evaluate the hypothesis
that the SPS would correlate more highly than the SIAS with a self-
report measure of fear of anxiety. This hypothesis followed from the
SPS's focus on anxiety while being observed by others and the findings
of Heimberg, Hope, Dodge, and Becker (1990) and Levin et al. (1993)
that heart rate was higher among public-speaking phobics than patients
with generalized social phobia during a behavioral test. Perhaps patients
with circumscribed fears are more sensitive to anxiety cues because
their reactions to performance situations are predictable and unique to
these situations in contrast to the ubiquitous anxiety experienced by
patients with generalized social phobia. Additionally, patients with dis-
crete observational fears may be more sensitive to physiological cues
for fear that these symptoms might become visible to others and lead
to negative evaluation.
The Anxiety Sensitivity Index contains 16 items measuring fear of
the symptoms of anxiety (e.g., "It scares me when I feel shaky").
Items are rated on a 5-point scale
(very little
to
very much), and the
score is the simple sum of the 16 items (ranging from 0-64). Reliability
and validity of the Anxiety Sensitivity Index have been established with
anxious patients in a number of studies (e.g., Peterson & Reiss, 1987).
The Penn State Worry Questionnaire was included to assess whether
the degree of worry associated with feared social situations is greater
for those experiencing anxiety while interacting with others compared
with anxiety occurring in performance-oriented situations. Given the
overlapping symptomatology of social phobia and generalized anxiety
disorder (Rapee, Sanderson, & Barlow, 1988) and the probability that
interactional anxiety affects more situations than performance anxiety,
a measure of worry (the principal component of generalized anxiety
disorder) may correlate more highly with interactional anxiety (SIAS)
than observational anxiety (SPS).
The Penn State Worry Questionnaire contains 16 items assessing the
specific trait of worry, and it has been used to measure the severity of
generalized anxiety disorder (e.g., "I'm always worrying about some-
thing" ). Items are rated on a 5-point scale
(not at all typical of me
to
very typical of me).
Summed scores range from 16 to 80, with higher
scores representing a greater degree of worry. The Penn State Worry
Questionnaire has been shown to have adequate internal consistency and
convergent and divergent validity with anxious patients (T. A, Brown,
Antony, & Barlow, 1992).
Results
Differences Between Social Phobics, Patients with
Other Anxiety Disorders, and Normal Controls
The different groups were compared on the SIAS and SPS
using one-way analyses of variance (ANOVAs). Significant dif-
ferences were found for both the SIAS, F(6,179) = 13.1, p <
.0001, and SPS, F(6,179) = 11.9,p < .0001. Duncan's multiple
range tests (p < .05) revealed that patients with social phobia
scored higher on the SIAS than did patients with any other
anxiety disorder or normals (Table 1 ). Patients with panic disor-
der and agoraphobia (PDA) scored higher than patients with
simple phobia, but there were no other differences among anx-
ious patients on the SIAS. Normals achieved lower scores on
the SIAS than did all other groups. On the SPS, patients with
social phobia scored higher than did all other groups except
those with PDA. (This finding is examined further in the section
24
BROWN ET AL.
Table 1
Comparison of Diagnostic Groups on the Social Interaction Anxiety Scale
(SIAS) and Social Phobia Scale (SPS)
Social Simple
phobia PDA OCD PD GAD phobia Normal
Scale (n = 50) (n = 27) (n = 20) (n = 29) (n = 20) (n = 19) (n = 21) F(6, 179) p
SIAS
M 50.7a 40.2b 34.6b.c 33.9b.~ 32.7b.c 27.1~ 14.3d 13.1 .0001
SD
17.0 16.8 20.1 19.3 16.3 17.0 11.0
SPS
M 36.9a 33.6~.b 26.8b.~ 23.%d 18.0c.d 14.7d.o 6.3° ll.9 .0001
SD
17.5 20.6 21.8 14.8 12.6 14.3 4.9
Note.
PDA = panic disorder with agoraphobia; OCD = obsessive-compulsive disorder; PD = panic
disorder without agoraphobia; GAD = generalized anxiety disorder. Group means with different subscripts
are significantly different at p < .05 (Duncan's multiple range test).
The Effect of Comorbid Diagnoses. )
Patients with PDA, in turn,
scored higher than patients with panic disorder without agora-
phobia, patients with generalized anxiety disorder, patients with
simple phobia, and normals. Patients with obsessive-compul-
sive disorder also achieved higher scores than patients with sim-
ple phobia and normals. Normals scored lower on the SPS than
did all other groups except patients with simple phobia. In a
comparison between patients with and without a secondary diag-
nosis of social phobia (in addition to a principal diagnosis of
another anxiety disorder), patients with a secondary diagnosis
of social phobia scored higher on the SIAS, t(l13) = 5.44, p
< .0001, and SPS, t(l13) = 4.98, p < .0001, than did those
without a secondary diagnosis of social phobia.
Classification of Patients as Cases of Social Phobia
In addition to the ANOVAs, a "caseness" strategy was used
in which a person was identified as having social phobia if he
or she scored one standard deviation above the mean of Heim-
berg et al.'s (1992) community sample on the SIAS (->34) or
SPS (->24). The percentages of each of the anxiety disorder
groups (without an additional diagnosis of social phobia) and
the normal control group identified by the two measures (sepa-
rately and jointly) as cases of social phobia are presented in
Table 2. The percentage of patients with a principal diagnosis
of social phobia identified as cases was significantly higher than
the percentage of each of the other groups for both the SIAS
and SPS.
Sensitivity,
or the percentage of actual cases of social
phobia correctly identified [ true positives/(true positives + false
negatives) × 100], was 86% for the SIAS and 76% for the
SPS.
Specificity,
or the number of patients without social phobia
correctly identified [true negatives/(false positives + true nega-
tives) × 100], was 70% for the SIAS and 72% for the SPS.
The overall efficiency of the test (i.e., hit rate) was 75% for
the SIAS and 73% for the SPS. When a criterion score on both
scales was required for classification as a case, the pattern was
similar but more conservative. A slightly smaller percentage
(72%) of patients with social phobia were correctly identified.
However, a larger percentage (80%) of patients without social
phobia were correctly classified as noncases. The overall hit
rate increased to 77%. (These analyses were repeated, including
patients with a principal diagnosis of an anxiety disorder other
than social phobia but with an additional diagnosis of social
phobia. These patients were classified as positive for social pho-
bia in the reanalyses. The results were equivalent.) ~
The Effects of Comorbid Diagnoses
SIAS and SPS scores of subgroups of patients with social
phobia with and without additional diagnoses were compared
using independent sample t tests. The additional diagnoses that
were examined included mood disorders (major depressive dis-
order, dysthymia, or depressive disorder not otherwise speci-
fied), generalized anxiety disorder, and panic disorder. No dif-
ferences were found between patients with social phobia with
and without additional diagnoses of mood disorder or panic
disorder (with or without agoraphobia) on either the SIAS and
SPS or generalized anxiety disorder on the SPS. The mean SIAS
score for patients with social phobia with generalized anxiety
disorder (M = 60.1,
SD
= 12.5) was higher than the score for
patients with social phobia without generalized anxiety disorder
(M = 48.9,
SD
= 17.2), t(48) = 2.18, p < .05.
In addition to the caseness analyses, we also examined the classifica-
tion of patients by using discriminant function analysis (DFA), as sug-
gested by an anonymous reviewer. Several DFAs were conducted, but
we present only the analysis examining the classification of patients into
groups using both the SIAS and SPS as predictors of group membership,
the analysis that resulted in the highest correct classification rate. Results
of other DFAs are available from Richard G. Heimberg. Patients with
secondary diagnoses of social phobia were not included in these analy-
ses, which attempted to classify patients into one of three groups: (a)
patients with a principal diagnosis of social phobia, (b) patients with a
principal diagnosis of another anxiety disorder other than simple phobia,
and (c) patients with simple phobia and normal controls. Two discrimi-
nant functions were calculated, with a resultant X2(4, N = 154) =
76.25, p < .0001. The first function accounted for nearly 99% of the
variance. The SIAS and SPS contributed approximately equally to the
prediction of group membership, and the overall correct classification
rate (hit rate) was 61%. Prior probability of correct classification into
the three groups based only on sample size was 32%, 41%, and 26%,
respectively. The use of the SIAS and SPS increased these figures to
70%, 64%, and 45%.
ASSESSMENT OF SOCIAL PHOBIA
Table 2
Percentage of Each Diagnostic Group Classified as Cases Using the Social Interaction
Anxiety Scale (SIAS) and Social Phobia Scale (SPS)
Social Simple
phobia PDA OCD PD GAD phobia Normal
Scale (n = 50) (n = 16) (n = 13) (n = 20) (n = 15) (n = 19) (n = 21)
SIAS 86%a 50%b 23%c.d.e 35%b,c.d 20%~e 42%b,c 10%~
SPS 76%a 50%b 38%b,c 45%b 20%¢ 21%¢ 0%d
SIAS and SPS 72%a 44%b 23%b,c 35%b 7%¢,d 16%¢ 0%d
Note.
PDA = panic disorder with agoraphobia; OCD --- obsessive-compulsive disorder; PD = panic
disorder without agoraphobia; GAD = generalized anxiety disorder. A
case
is defined as a participant who
scored 1
SD
above the mean of the Heimberg et al. (1992) community sample (SIAS >- 34; SPS ~ 24).
Percentages with different subscripts are significantly different at p < .05 (test for significance of difference
between two proportions). Participants with an additional diagnosis of social phobia were excluded from
this analysis.
25
Because patients with PDA were not distinguished from those
with social phobia on the SPS in the ANOVA of anxiety disorder
groups, we further examined the scores of patients with PDA
on the SPS. Patients with a principal diagnosis of panic disorder
and agoraphobia but no additional diagnosis of social phobia
and patients with a principal diagnosis of panic disorder and
agoraphobia and an additional diagnosis of social phobia were
compared on the SPS (and the SIAS). The mean SPS score for
patients with panic disorder, agoraphobia, and social phobia (M
= 43.9,
SD
= 18.5) was substantially higher than the score for
patients with panic disorder and agoraphobia but no additional
diagnosis of social phobia (M = 26.5,
SD
= 19.4), t(25) =
2.33, p < .03. Similarly, patients with panic disorder, agorapho-
bia, and social phobia scored substantially higher (M = 48.2,
SD
= 19.5) than did patients with panic disorder and agorapho-
bia but no additional diagnosis of social phobia on the SIAS
(M = 34.7,
SD
= 12.4), t(25) = 2.21, p < .04.
Correlational Analyses
The SIAS and SPS were correlated with the Anxiety Sensitiv-
ity Index, Penn State Worry Questionnaire, and each other (see
Table 3). T tests for the significance of differences between
dependent correlations revealed that the SPS correlated more
highly with the Anxiety Sensitivity Index than did the SIAS.
No significant differences were found for the Penn State Worry
Questionnaire.
Three measures taken from the social phobia section of the
ADIS-R (Clinician Severity Rating for patients who had a
Table 3
Correlations Between Measures of Anxiety Sensitivity and
Worry and the Social Interaction Anxiety Scale (SIAS)
and Social Phobia Scale (SPS)
Scale SIAS SPS t ( 181 ) p
SPS .72 --
Anxiety Sensitivity Index .41 .55 3.05 .005
Penn State Worry Questionnaire .43 .45 0.47
ns
Note.
N = 184. All correlations were significant (p < .001).
principal or additional diagnosis of social phobia, number of
feared social situations, and number of panic symptoms en-
dorsed during social situations) were also correlated with scores
on the SIAS and SPS. The correlation between the SIAS and
the Clinician Severity Rating was .50 (p < .001 ), and the corre-
lation between the SPS and the Clinician Severity Rating was
.40 (p < .01 ). The SIAS and SPS were also correlated with the
number of interactional situations and performance situations
receiving an anxiety rating of 2 or more on the 0-4 scale (Table
4). T tests for the significance of differences between dependent
correlations revealed that the number of interactional situations
endorsed correlated more highly with the SIAS than the SPS,
whereas the number of performance situations endorsed was
more highly correlated with the SPS than the SIAS. Similarly,
the SIAS correlated more highly with number of interactional
situations than with number of performance situations, whereas
the SPS correlated more highly with the number of performance
situations than the number of interactional situations. Neither
the SIAS nor SPS correlated significantly with the number of
panic symptoms endorsed during social situations.
Discussion
The SIAS and SPS appear to be promising measures of social
phobia. These measures reliably discriminated patients with so-
cial phobia from patients with other anxiety disorders and per-
sons with no mental disorder. Scores on the SIAS were higher
for patients with social phobia than for all other groups. Similar
Table 4
Correlations Between Number of Feared Social Situations
(per the ADIS-R) and the Social Interaction Anxiety Scale
(SIAS) and Social Phobia Scale (SPS)
Situation SIAS SPS t(70) p
Social interaction .53** .18 3.49 .005
Performance .28* .53** 2.49 .01
Note.
N = 73. ADIS-R = Anxiety Disorders Interview Schedule--
Revised.
*p < .01. **p < .001.
26 BROWN ET AL.
findings were obtained for the SPS, although patients with social
phobia could not be distinguished from those with panic disorder
and agoraphobia. Scores on these two self-report scales were
also correlated with a clinician-administered index of the sever-
ity of social phobia. Although the magnitude of these correla-
tions may be best described as moderate, they were observed
despite the differences in methodology between questionnaire
and interview assessments.
The finding that patients with social phobia and those with
panic disorder and agoraphobia were not distinguishable on the
SPS suggests that the SPS may be less sensitive to the difference
between these groups than is desirable. However, comparisons
of SIAS and SPS scores between the 11 patients with panic
disorder and agoraphobia who had an additional diagnosis of
social phobia and the 16 patients with panic disorder and agora-
phobia but no additional diagnosis of social phobia revealed
that patients with social phobia scored higher on both scales
than did those patients without social phobia. Thus, the failure
of the SPS to separate patients with social phobia from patients
with panic disorder and agoraphobia may be a result of the
social anxieties of a substantial proportion of the PDA group
(41%). In fact, patients with PDA without an additional diagno-
sis of social phobia were not as likely as those with social phobia
to be selected as cases using only the SPS.
The results of the caseness analyses further demonstrate the
sensitivity and specificity of both scales to social phobia. Note
that these analyses may be conservative because less than one
third (n = 50) of the relevant sample (N = 154) had a principal
diagnosis of social phobia, patients with a secondary diagnosis
of social phobia were excluded, and the overall hit rate should
be expected to drop as the base rate declines. 2 Of the patients
with a principal diagnosis of social phobia, 86% scored a stan-
dard deviation or more above the mean of the Heimberg et al.
(1992) community sample on the SIAS, and 76% did so on the
SPS. The percentages of patients correctly classified as cases
by the S1AS and SPS were substantially higher than percentages
of patients in all other diagnostic groups misclassified by these
measures. The conjoint use of both scales correctly classified a
slightly smaller percentage of patients with social phobia, but it
was less likely to misclassify other patients as social phobic,
resulting in an overall higher hit rate and suggesting that the
SIAS and SPS have incremental validity in the assessment of
social phobia.
We examined the ability of the SIAS and SPS to correctly
classify anxiety disorder patients and normal controls using dis-
criminant function analysis (DFA) as well as the caseness ap-
proach. The DFA improved classification substantially beyond
base rates in the sample but produced a slightly less effective
classification than the caseness approach. Because (a) the case-
ness approach is much easier for clinicians to apply, (b) the hit
rate using both scales was 16% higher than the best outcome
with the DFA (77% vs. 61%), and (c) it has now been demon-
strated to separate social phobics from other anxiety patients
and normals in two separate samples (the present one and the
one utilized by Heimberg et al., 1992), we believe the caseness
approach to have substantial clinical utility. Although not ade-
quate for purposes of diagnosis, an overall hit rate of 77% may
be quite appropriate for use as a screening measure or as a cue
for more in-depth assessment.
Analyses comparing patients with social phobia with and
without additional diagnoses produced evidence of the specific-
ity of the SIAS and SPS to social phobia, but that evidence was
somewhat mixed. In support, there were no differences between
patients with social phobia with and without an additional diag-
nosis of mood disorder or panic disorder on either the SIAS or
SPS, nor was there a difference between patients with social
phobia with and without generalized anxiety disorder on the
SPS. However, patients with social phobia and generalized anxi-
ety disorder scored higher on the SIAS than did those with
social phobia without generalized anxiety disorder. The latter
finding is consistent with the high endorsement of social anxiety
symptoms among patients with generalized anxiety disorder
(Rapee et al., 1988) and the frequent assignment of social pho-
bia as an additional diagnosis to these patients (T. A. Brown &
Barlow, 1992). Scores on the SIAS may be the result of patients'
social phobia rather than their general anxiety, but it is unlikely
that the SIAS can discriminate sufficiently well between the
fears of the person with social phobia and the worries of the
patient with generalized anxiety disorder about interactions with
others.
To further evaluate the SIAS and SPS, the influence of panic
symptomatology was investigated. The number of panic symp-
toms endorsed during feared social situations was not related
to SIAS and SPS scores. The SPS did correlate more highly
than the SIAS with scores on the Anxiety Sensitivity Index, a
measure of fear of anxiety and panic symptoms. This finding
extends Heimberg et al.'s (1990) and Levin et al.'s (1993)
report of the relationship between fears in a specific perfor-
mance situation (i.e., public speaking) and increased heart rate,
suggesting a possible relation between panic symptomatology
and fear of performance in specific situations involving observa-
tion by others. Further investigation of this hypothesis is
warranted.
The SIAS and SPS appear to measure different yet related
constructs. The measures correlated with one another and with
the Anxiety Sensitivity Index, the Penn State Worry Question-
naire, and interviewer-rated severity of social phobia. Neverthe-
less, the number of interactional situations rated as moderately,
severely, or very severely feared correlated more highly with
the SIAS, whereas the number of performance situations so
rated correlated more highly with the SPS. Although both are
highly relevant to social phobia, these differences support Mat-
tick and Clarke's (1989) distinction between social interactional
anxiety and scrutiny fears.
Although questions remain about the relation of panic symp-
tomatology and generalized anxiety disorder to SIAS and SPS
scores among patients with social phobia, this study provides
further evidence of the psychometric adequacy of these two
scales and extends previous research on their validity (Heimberg
et al., 1992; Mattick & Clarke, 1989; Rapee et al., 1992). The
ability of these scales to distinguish among anxiety disorder
diagnoses and correctly classify patients with social phobia ex-
emplifies their utility as tools for the assessment of social
phobia.
2 We thank associate editor Stephen N. Haynes for pointing this out
to us.
ASSESSMENT OF SOCIAL PHOBIA
27
References
American Psychiatric Association. (1980). Diagnostic and statistical
manual of mental disorders (3rd ed.). Washington, DC: Author.
American Psychiatric Association. (1987). Diagnostic and statistical
manual of mental disorders (3rd ed., rev.). Washington, DC: Author.
American Psychiatric Association. (1994). Diagnostic and statistical
manual of mental disorders (4th ed.). Washington, DC: Author.
Brown, E, J., Heimberg, R. G., & Juster, H. R. (1995). Social phobia
subtype and avoidant personality disorder: Effect on severity of social
phobia, impairment, and outcome of cognitive-behavioral treatment.
Behavior Therapy, 26, 467-486.
Brown, T. A., Antony, M. M., & Barlow, D. H. (1992). Psychometric
properties of the Penn State Worry Questionnaire in a clinical anxiety
disorders sample. Behaviour Research and Therapy, 30, 33-37.
Brown, T.A., & Barlow, D.H. (1992). Comorbidity among anxiety
disorders: Implications for treatment and DSM-IV. Journal of Con-
sulting and Clinical Psychology, 60, 835-844.
Cox, B. J., & Swinson, R. P. (1995). Assessment and measurement. In
M. B. Stein (Ed.), Social phobia: Clinical and research perspectives
(pp. 261-291 ). Washington, DC: American Psychiatric Press.
DiNardo, P. A., & Barlow, D. H. (1988). Anxiety Disorders Interview
Schedule--Revised (ADIS-R). Albany, NY: Graywind.
DiNardo, P. A., Moras, K., Barlow, D. H., Rapee, R. M., & Brown, T. A.
(1993), Reliability of DSM-III-R anxiety disorder categories using
the Anxiety Disorders Interview Schedule--Revised (ADIS-R). Ar-
chives of General Psychiatry, 50, 251-256.
Heimberg, R. G., Hope, D. A., Dodge, C. S., & Becker, R. E. (1990).
DSM-III-R subtypes of social phobia: Comparison of generalized
social phobics and public speaking phobics. Journal of Nervous and
Mental Disease, 178, 172-179.
Heimberg, R.G., Liebowitz, M.R., Hope, D.A., & Schneier, E R.
(1995). Social phobia: Diagnosis, assessment, and treatment. New
York: Guilford Press.
Heimberg, R. G., Mueller, G. P., Holt, C. S., Hope, D. A., & Liebowitz,
M. R. (1992). Assessment of anxiety in social interaction and being
observed by others: The Social Interaction Anxiety Scale and the
Social Phobia Scale. Behavior Therapy, 23, 53-73.
Holt, C. S., Heimberg, R. G., & Hope, D. A. (1992). Avoidant personal-
ity disorder and the generalized subtype of social phobia. Journal of
Abnormal Psychology, 101, 318-325.
Holt, C. S., Heimberg, R. G., Hope, D. A., & Liebowitz, M. R. (1992).
Situational domains of social phobia. Journal of Anxiety Disorders,
6, 63-77.
Leary, M. R. (1983a). Social anxiousness: The construct and its mea-
surement. Journal of Personality Assessment, 47, 66-75.
Leary, M. R. (1983b). Understanding social anxiety: Social, personal-
ity, and clinical perspectives. Beverly Hills, CA: Sage.
Levin, A. P., Saoud, J. B., Strauman, T., Gorman, J. M., Fyer, A. J., Craw-
ford, R., & Liebowitz, M. R. (1993). Responses of "generalized"
and "discrete" social phobics during public speaking. Journal of
Anxiety Disorders, 7, 207-221.
Liebowitz, M. R. (1987). Social phobia. Modern Problems in Pharma-
copsychiatry, 22, 141-173.
Marks, I. M., & Mathews, A. M. (1979). Brief standard rating for phobic
patients. Behaviour Research and Therapy, 17, 263-267.
Mattick, R. P., & Clarke, J. C. (1989). Development and validation of
measures of social phobia scrutiny fear and social interaction anxiety.
Unpublished manuscript.
Mattick, R. P., & Peters, L. (1988). Treatment of severe social phobia:
Effects of guided exposure with and without cognitive restructuring.
Journal of Consulting and Clinical Psychology, 56, 251-260.
Mattick, R. P., Peters, L., & Clarke, J. C. (1989). Exposure and cognitive
restructuring for social phobia: A controlled study. Behavior Therapy,
20, 3-23.
McNeil, D. W., Ries, B. J., & Turk, C. L. (1995). Behavioral assessment:
Self-report, physiology, and overt behavior. In R. G. Heimberg, M. R.
Liebowitz, D. A. Hope, & E R. Schneier (Eds.), Socialphobia: Diag-
nosis, assessment, and treatment (pp. 202-231 ). New York: Guilford
Press.
Meyer, T. J., Miller, M. L., Metzger, R. L., & Borkovec, T. D. (1990).
Development and validation of the Penn State Worry Questionnaire.
Behaviour Research and Therapy, 2& 487-495.
Peterson, R. A., & Reiss, S. (1987). Anxiety Sensitivity Index manual.
Palos Heights, IL: International Diagnostic Systems.
Rapee, R. M., Brown, T. A., Antony, M. M., & Barlow, D.H. (1992).
Response to hyperventilation and inhalation of 5.5% carbon dioxide-
enriched air across the DSM-III-R anxiety disorders. Journal of
Abnormal Psychology, 101, 538-552.
Rapee, R. M., Sanderson, W. C., & Barlow, D. H. (1988). Social phobia
features across the DSM-III-R anxiety disorders. Journal of Psycho-
pathology and Behavioral Assessment, 10, 287-300.
Ries, B. J., McNeil, D. W., Boone, M. L., Turk, C. L., Carter, L. E., &
Heimberg, R. G. (1996). Assessment of contemporary social phobia
verbal report instruments. Manuscript submitted for publication.
Schneier, E R., Liebowitz, M. R., Beidel, D. C., Garfinkel, R., Heimberg,
R. G., Hornig, C. D., Johnson, J., Juster, H., Law, K., Mannuzza, S.,
Mattia, J. I., Oberlander, E., Orsillo, S., Turner, S. M., & Weissman,
M.M. (in press). MacArthur data reanalysis for DSM-IV: Social
phobia. In T. A. Widiger, A. H, Frances, H. A. Pincus, M. J. First, R.
Ross, & W. Davis (Eds.), DSM-IV source book. Washington, DC:
American Psychiatric Press.
Stein, M. B. (1995). Social phobia: Clinical and research perspectives.
Washington, DC: American Psychiatric Press.
Turner, S. M., Beidel, D. C., Dancu, C. V., & Keys, D. J. (1986). Psycho-
pathology of social phobia and comparison to avoidant personality
disorder. Journal of Abnormal Psychology, 95, 389-394.
"lhrner, S. M., Beidel, D. C., Dancu, C. V., & Stanley, M. A. (1989). An
empirically derived inventory to measure social fears and anxiety:
The Social Phobia and Anxiety Inventory. Psychological Assessment:
A Journal of Consulting and Clinical Psychology, 1, 35-40.
Watson, D., & Friend, R. (1969). Measurement of social-evaluative
anxiety. Journal of Consulting and Clinical Psychology, 33, 448-
457.
Received February 6, 1996
Revision received August 8, 1996
Accepted August 15, 1996 •
