Anxiety Disorders

20 (2006) 543–561

Dimensionality of somatic complaints: Factor

structure and psychometric properties of the
Self-Rating Anxiety Scale
Bunmi O. Olatunji a,*, Brett J. Deacon b, Jonathan S.
Abramowitz c, David F. Tolin d
a
Department of Psychiatry, Massachusetts General Hospital/Harvard Medical School,
15 Parkman Street, ACC 812, Boston, MA 02114, USA
b
Department of Psychology, University of Wyoming, Department 3415,
1000 E. University Avenue, Laramie, WY 82071, USA
c
Department of Psychiatry & Psychology, Mayo Clinic,
200 First Street SW, Rochester, MN, USA
d
Anxiety Disorders Center, The Institute of Living, 200 Retreat Avenue,
Hartford, CT 06106, USA
Received 21 January 2005; received in revised form 27 July 2005; accepted 19 August 2005

Abstract

Somatic complaints are often key features of anxiety pathology. Although most
measures of anxiety symptoms capture somatic complaints to some degree, the Self-
Rating Anxiety Scale (SAS) was developed primarily as a measure of somatic symptoms
associated with anxiety responding. We evaluated the psychometric properties and factor
structure of the SAS in two large undergraduate samples who completed the SAS and
measures of anxiety and depression. Exploratory factor analysis revealed four lower-order
SAS factors in both samples: (1) anxiety and panic; (2) vestibular sensations; (3) somatic
control; and, (4) gastrointestinal/muscular sensations. The SAS demonstrated good
reliability in both samples, and the correlations between the SAS factors and other anxiety

* Corresponding author. Tel.: +1 617 724 5600; fax: +1 617 726 8907.
E-mail address: [email protected] (B.O. Olatunji).

0887-6185/$ – see front matter # 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.janxdis.2005.08.002
544 B.O. Olatunji et al. / Anxiety Disorders 20 (2006) 543–561

variables provide supportive evidence for convergent validity, though evidence for dis-
criminant validity was limited. The strengths and limitations of the SAS are offered as well
as the implications of our findings for the nature and assessment of somatic complaints in
anxiety disorders.
# 2005 Elsevier Inc. All rights reserved.

Keywords: Self-Rating Anxiety Scale; Factor analysis; Anxiety; Somatic complaints

Somatic symptoms are the leading cause of outpatient medical visits and also
the predominant reason why patients with common mental disorders present in
primary care (Kellner, 1990; Kroenke, 2003). Theoretical models suggest that
somatic complaints may represent a core feature of anxiety pathology (e.g.,
Goldberg, 1996; Lang, 1971). Somatic complaints may manifest as cardiophobia,
the repeated complaint of chest pain, heart palpitations, and other sensations
related to having a heart attack (Eifert, 1992). The fear and catastrophic
misinterpretations of somatic sensations may place individuals at risk for the
development of anxiety-related conditions (Clark, 1986; Ehlers, 1991; Reiss &
McNally, 1985), particularly panic disorder. Indeed, studies have demonstrated a
strong, positive relationship between fear of bodily sensations and panic disorder
(e.g., Apfledorf, Shear, Leon, & Portera, 1994; McNally & Lorenz, 1987) and
patients with panic disorder also endorse more somatic anxiety symptoms than do
controls (Hoehn-Saric, McLeod, Funderburk, & Kowalski, 2004).
Somatic complaints have also been implicated in other anxiety disorders (e.g.,
Koksal, Power, & Sharp, 1991). For instance, studies have shown a strong
association between somatic symptoms and posttraumatic stress disorder (PTSD)
independently of anxiety, depression, injury severity, and medical comorbidity
(Van Ommeren et al., 2002; Zatzick, Russio, & Katson, 2003). Patients with
generalized anxiety disorder (GAD) also score higher on somatic anxiety
symptoms than controls (Hoehn-Saric et al., 2004) and studies have shown a
unique relation between muscle tension and pathological worry observed in GAD
(Joormann & Stober, 1999). Social phobia is also often accompanied by somatic
symptoms, such as trembling, blushing, and sweating (Mersch, Hilderbrand,
Lavy, Wessel, & Van Hout, 1992) as well as concerns that others will notice one’s
anxiety-related somatic symptoms (e.g., Taylor, Koch, & McNally, 1992).
Hypochondriasis, the excessive worry about one’s health, is yet another example
of an anxiety problem in which somatic sensations are prominent (Abramowitz,
Schwartz, & Whiteside, 2002; Taylor & Asmundson, 2004).
In recognition of the importance of somatic complaints in anxiety disorders,
many self-report measures of anxiety incorporating items assessing somatic
concerns have been developed. For example, many items of the Beck Anxiety
Inventory (BAI; Beck, Epstein, Brown, & Steer, 1988) inquire about numbness or
tingling and wobbliness in the legs. The BAI was developed to better discriminate
anxiety from depression, and as a result it consists primarily of somatic items.
 B.O. Olatunji et al. / Anxiety Disorders 20 (2006) 543–561 545

However, the BAI has been criticized because its composition of somatic
complaints may be specific to panic disorder rather than anxiety disorders in
general (Cox, Cohen, Direnfeld, & Swinson, 1996). For example, Cox et al.
(1996) found that among panickers, the BAI and Panic Attack Questionnaire
(Norton, Dorward, & Cox, 1986) items loaded together onto the same factors,
indicating that the panic patients completed the BAI as if it were a measure of
panic symptoms. These authors argued that the BAI is too ‘‘panic-centric.’’
Another concern is that the number of items incorporated to assess somatic
complaints in other anxiety measures are often low, which may influence
reliability (Koksal & Power, 1990; Maranell, 1974) and adequacy in capturing the
potentially heterogeneous nature of somatic complaints.
Somatic symptoms found in anxiety disorders are manifested in many
systems (e.g., Zung, 1971) including musculoskeletal (e.g., muscle tension,
trembling), cardiovascular (e.g., palpitations, tachycardia), respiratory (e.g.,
constriction in chest, dyspnea), gastrointestinal (e.g., nausea, diarrhea),
genitourinary (e.g., frequency of micturition, urgency), and skin (e.g., flushing,
sweating). Thus, rather than being a one-dimensional construct, anxiety-related
somatic complaints appear to consist of several factors (Liu, Clark, & Eaton,
1997; Taylor & Cox, 1998). For instance, Landy and Stern (1971) found that in
a nonclinical population, somatic perceptions consist of four factors; cardiac
response, sweating, vasoconstriction, and frequent urination. In a sample of
patients with a clinical diagnosis of anxiety, depression, hysteria, or
hypochondriasis, four factors corresponding to the head, chest, abdomen,
and fatigue were identified (Mumford et al., 1991). Consideration of the
dimensionality of somatic complaints may have important implications for our
understanding of the etiology of anxiety disorders. Appeal to specific somatic
complaints could also be useful in differentiating anxiety from depression. For
example, it has been shown that the unclear boundaries between GAD and
major depression may be reduced by emphasizing muscle tension and de-
emphasizing concentration difficulties in the diagnosis of GAD (Joormann &
Stober, 1999). Furthermore, if somatic complaints are truly multidimensional,
relationships between global measures of somatic concerns and measures of
specific anxiety symptoms may be misleading. For example, the relation
between overall somatic concerns and response to CO2 challenge may be less
pronounced than the relation between the somatic concerns specific to the
cardiovascular and respiratory systems and fearful responses to CO2 (e.g.,
Schmidt, 1999).
In an attempt to provide necessary content coverage of a variety of somatic
complaints associated with anxiety symptoms, Zung (1971) developed the 20-
item Self-Rating Anxiety Scale (SAS). The initial psychometric evaluation of
the measure revealed adequate split-half reliability (r = .71). A subsequent
evaluation reported adequate internal consistency in normal (a = .69) and
outpatient samples (a = .81; Jegede, 1977). Michelson and Mavissakalian
(1983) also found good test–retest reliability in a clinical sample of
546 B.O. Olatunji et al. / Anxiety Disorders 20 (2006) 543–561

agoraphobics over a period ranging from 1 to 16 weeks (r’s = .81 to .84).

Despite some research on the psychometric properties of the SAS, to our
knowledge no published study has evaluated its factor structure. The SAS may
be a better instrument than existing anxiety measures for assessing the somatic
symptoms of anxiety given its broader content sampling of such symptoms. The
SAS may also have utility in studies specifically examining the negative impact
(i.e., excessive health concerns) of somatic anxiety symptoms in anxiety
disorders. However, this possibility cannot be adequately evaluated at present
due to the limited research on the psychometric properties of the SAS. To
address this limitation, we examined the factor structure and psychometric
properties of the SAS in two large, nonclinical samples. We also provide an
item-level analysis of the SAS. Consistent with the notion that somatic
complaints may be multidimensional (e.g., Liu et al., 1997; Taylor & Cox,
1998), it was predicted that the SAS would yield replicable lower-order somatic
factors that may load on a single higher-order factor. Finally, it was
hypothesized that the SAS and its lower-order factors would demonstrate a
pattern of theoretically consistent relationships with measures of anxiety and
depression.

1. Study 1

1.1. Method

1.1.1. Participants
The sample consisted of 552 college students recruited from introductory
psychology courses at University of North Carolina at Chapel Hill. The sample
was 75.2% female with a mean age of 19.0. Four hundred and thirteen participants
(74.8%) identified themselves as White/Caucasian, followed by 80 Black/African
Americans (14.5%), 30 Asians or Pacific Islanders (5.4%), and 29 participants
(5.3%) of other, multiple, or unreported ethnicities.

1.1.2. Measures
1.1.2.1. Self-Rating Anxiety Scale (SAS). The SAS (Zung, 1971) is a 20-item
measure developed to assess the frequency of anxiety symptoms based on
diagnostic conceptualizations. It consists primarily of somatic symptoms. The
respondent indicates how often he or she has experienced each symptom on a
4-point Likert scale consisting of ‘‘none or a little of the time’’ (coded as 1),
‘‘some of the time’’ (coded as 2), ‘‘good part of the time’’ (coded as 3), and ‘‘most
or all of the time’’ (coded as 4). Items 5, 9, 13, 17, and 19 are reversed scored and
total scores on the SAS range from 0 to 80.

1.1.2.2. Agoraphobic Cognitions Questionnaire (ACQ). The ACQ (Chambless,

Caputo, Bright, & Gallagher, 1984) measures frequency of 14 different fearful
 B.O. Olatunji et al. / Anxiety Disorders 20 (2006) 543–561 547

cognitions associated with panic attacks and agoraphobia. The ACQ was
constructed to measure the cognitive aspect of ‘‘fear of fear.’’ The ASC can
generate two subscales reflecting loss of control and physical concerns.
Chambless et al. (1984) reported that the ACQ has adequate test–retest reliability
(r = .86) and internal consistency (a = .80). The ACQ was included in the present
study to assess the convergent validity of the SAS. Given its relevance to physical
concerns, the ACQ was included in the present study to assess the convergent
validity of the SAS.

1.1.2.3. Body Vigilance Scale (BVS). The BVS (Schmidt, Lerew, & Trakowski,
1997) measures the tendency to attend to or be vigilant to panic-related body
sensations. Schmidt et al. conceptualized body vigilance as a natural consequence
of learning to fear body sensations through the experience of unexpected panic
attacks. The BVS has demonstrated good internal consistency (a = .82) and
adequate test–retest reliability (Schmidt et al., 1997). Given its emphases on
bodily sensations, the BVS was also included in the present study to assess the
convergent validity of the SAS.

1.1.2.4. Fear of Negative Evaluation Scale (FNE). The FNE (Watson & Friend,
1969) was used to measure participants’ social anxiety. The FNE is a 30-item true/
false scale that assesses expectation and distress related to negative evaluation
from others. The scale has demonstrated good internal consistency (a = .94 to .96)
and test–retest reliability (Oei, Kenna, & Evans, 1991; Watson & Friend, 1969).
The FNE was included in the present study to assess the discriminant validity of
the SAS.

1.1.2.5. Center for Epidemiological Studies-Depression Scale (CES-D). The

CES-D (Radloff, 1977) is a 20-item measure that assesses the frequency of
depressive symptoms experienced during the past week. The CES-D has
demonstrated good internal consistency in both general and clinical populations
(a’s = .85 and .90, respectively; Radloff, 1977) and correlates strongly with the
Beck Depression Inventory (r = .87; Santor, Zuroff, Ramsay, Cervantes, &
Palacios, 1995). The CES-D was also included in the present study to assess the
discriminant validity of the SAS.

1.1.3. Procedure
Participants completed the psychometric assessment on a website created
for the study and received course credit for their participation. Informed consent
was obtained electronically via clicking a web link as proxy for signature.
After completing the measures, participants’ data was submitted electronically
to a database that was read into a statistical software package (SPSS) for
data analysis. Participants were informed that their responses would be
kept confidential and that they were free to withdraw from the study at any
time.
548 B.O. Olatunji et al. / Anxiety Disorders 20 (2006) 543–561

Table 1
Self-Rating Anxiety Scale: item means and standard deviations, obliquely rotated factor loadings, and
communalities for the four-factor solution from Study 1
SAS item M S.D. SAS factor
I II III IV h2
1. I feel more nervous and anxious than usual 1.67 .66 .80
.14 .08
.01 .62
2. I feel afraid for no reason at all 1.25 .50 .78 .08 .05
.16 .61
3. I get upset easily or feel panicky 1.61 .71 .73 .06 .15
.01 .63
4. I feel like I’m falling apart and going to pieces 1.48 .68 .51 .25 .17 .00 .46
20. I have nightmares 1.43 .61 .42 .00
.16 .25 .29
18. My face gets hot and blushes 1.68 .74 .41
.07
.10 .29 .30
10. I can feel my heart beating fast 1.44 .60 .40 .34
.01 .03 .36
12. I have fainting spells or feel like it 1.11 .36
.08 .81
.06 .02 .62
11. I am bothered by dizzy spells 1.20 .47
.07 .76
.08 .12 .64
6. My arms and legs shake and tremble 1.15 .44 .17 .53 .13
.01 .42
14. I get feelings of numbness and tingling 1.22 .50 .00 .35 .17 .10 .21
in my fingers and toes
17. My hands are usually dry and warm 2.64 1.07
.09 .06 .66
.02 .43
19. I fall asleep easily and get a good night’s rest 2.35 .92 .03
.08 .62 .23 .47
9. I feel calm and can sit still easily 2.22 .92 .13
.20 .60 .21 .48
5. I feel that everything is all right and 2.42 .87 .22 .00 .57
.06 .41
nothing bad will happen
13. I can breathe in and out easily 1.55 .91
.02 .20 .51
.09 .33
15. I am bothered by stomachaches or indigestion 1.54 .74 .01 .13 .01 .69 .55
16. I have to empty my bladder often 1.78 .84
.07
.08 .05 .61 .35
8. I feel weak and get tired easily 1.66 .76 .11 .20 .09 .55 .48
7. I am bothered by headaches neck and back pain 1.70 .84 .02 .27 .05 .51 .41
% Variance of rotated factors 24.63 7.80 6.83 6.07
Note. Factor loadings j.30j are listed in boldface type. The first five eigenvalues were 4.92, 1.56, 1.36,
1.21, and 1.06.

1.2. Results

1.2.1. Reliability and item-level analyses

The mean SAS total score was 33.09 (S.D. = 6.88). SAS total scores for
women (M = 33.95, S.D. = 6.97) were higher than those for men (M = 30.47,
S.D. = 5.89), t(550) = 5.25, P < .01. Given that the scale consisted of 20 items,
these mean SAS total scores indicate that participants tended to indicate between
‘‘None or a little of the time’’ or ‘‘Some of the time’’ agreement with the scale
items. Means and standard deviations for the SAS items are presented in Table 1.
Mean scores on 16 out of 20 items were below 2.0 (i.e., ‘‘Some of the time’’
agreement with the item), suggesting that the content of most SAS items was
generally outside of the experience of most participants. The SAS demonstrated
adequate internal consistency (a = .81). Based on the criterion of .30 as an
acceptable corrected item-total correlation (Nunnally & Bernstein, 1994), all 20
items performed adequately (range: .34 to .65).
 B.O. Olatunji et al. / Anxiety Disorders 20 (2006) 543–561 549

1.2.2. Factor structure of the SAS

To our knowledge, no published study has reported on the factor structure of
the SAS. Accordingly, we elected to use exploratory factor analysis in the first
study. We chose principal components analysis (PCA) as the primary method
because factor scores from principal-axis factor analysis (PAF) are indeterminate
(Schönemann & Wang, 1972). Factors were rotated using an oblique (Oblimin)
transformation in both cases because we expected the lower-order factors to show
low-to-moderate intercorrelations. The number of factors to retain was
determined by parallel analysis, a statistical procedure for determining the
break in the scree plot (Horn, 1965). This method is one of the most accurate
techniques for determining the number of factors to retain across varying sample
conditions (Zwick & Velicer, 1986). Based on the recommendations of Longman,
Cota, Holden, and Fekken (1989), parallel analyses were conducted twice, once
using the mean eigenvalues and once using the 95th percentile eigenvalues.
Although five factors had eigenvalues greater than 1.0, parallel analysis
indicated a four-factor solution for both the mean and 95th percentile eigenvalues.
Accordingly, four factors were extracted. Table 1 displays the eigenvalues, pattern
matrices (loadings), communalities, and percentage of variance for the four
rotated factors. The four-factor solution accounted for 45.3% of the item variance.
The magnitude of the communalities suggests that the factors accounted for a
moderately large portion of the variance in most items. Table 1 also shows that the
first factor accounted for a substantial portion of the variance in SAS item scores
(24.6%), whereas the remaining three factors explained smaller portions of the
item variance (between 7.8 and 5.3% each).
Factor I had seven items with salient (.30) loadings and assessed complaints
related to anxiety and panic (e.g., ‘‘I get upset and panicky’’). Accordingly, this
factor was labeled ‘‘anxiety and panic.’’ Factor II had five items with salient
loadings and was labeled ‘‘vestibular sensations.’’ Factor III also contained five
items with salient loadings and assessed adaptive somatic functioning and the
perception of calm (e.g., ‘‘I can breathe in and out easily’’). Accordingly, this
factor was labeled ‘‘somatic control.’’ The fourth factor consisted of four items
with salient loadings. Most items on Factor IV pertained to gastrointestinal (e.g.,
‘‘I am bothered by stomachaches or indigestion’’) and muscular complaints (e.g.,
‘‘I am bothered by headaches, neck and back pains’’). Accordingly, Factor IV was
labeled ‘‘gastrointestinal/muscular sensations.’’
Adequacy of the four-factor solution was examined through consideration of
simple structure (Thurstone, 1947), the criteria for stability suggested by
Guadagnoli and Velicer (1988), and by examining the internal consistency of each
factor. As shown by the pattern matrices in Table 1, the four-factor solution
appears to have adequate simple structure. Each factor consisted of an adequate
number of items with salient loadings (4–6) and PCA resulted in only item 10 (‘‘I
can feel my heart beating fast’’) presenting as a complex item (i.e., items with
salient loadings, >0.30, on more than one factor). However, item 10’s second
highest loading was not higher than .40, suggesting that no items on the SAS were
550 B.O. Olatunji et al. / Anxiety Disorders 20 (2006) 543–561

salient markers for more than one factor. Guadagnoli and Velicer (1988)
recommended that to be considered stable, factors should have (a) four or more
loadings above .60, (b) 10 or more items with loadings above .40 and a sample
size greater than 150, or (c) a sample size of greater than 300 for factors with only
a few loadings. Based on criterion C, the four factors appear to be stable.

1.2.3. The higher-order structure of the SAS

The higher-order factor structure of the SAS was examined by conducting a
PCA on the four obliquely rotated factor scores obtained (e.g., Taylor & Cox,
1998). The eigenvalues were 1.69, 0.88, 0.77, and 0.64, and thus a single higher-
order factor was extracted. The higher-order factor accounted for 42.3% of the
variance, and each lower-order factor loaded greater than .59 on this factor. Thus,
the results appear to support a hierarchical solution for the SAS in which the four
lower-order factors load on a single higher-order arousal factor.

1.2.4. Correlates of the SAS and its factors

Table 2 presents correlations between the SAS, the lower-order SAS factors, and
the ACQ, BVS, FNE, and CES-D. The SAS lower-order factors were strongly
correlated with SAS total scores (range: .60 to .81). The SAS ‘‘anxiety and panic’’
factor was most strongly associated with the remaining three factors (range: .42 to
.46), while other comparisons between SAS factors yielded low-to-moderate
correlations. SAS total scores were moderately to highly correlated with measures
of fearful cognitions, body vigilance, negative evaluation, and depression (range:
.40 to .64). SAS total scores were most strongly associated with the CES-D, a
measure of depression (r = .64, P < .001). Excluding depression, the SAS ‘‘anxiety
and panic’’ factor was more strongly associated with panic and agoraphobic-related

Table 2
Pearson correlations between the Self-Rating Anxiety Scale factors and related measures from Study 1
Scale SAS total score SAS factor scores Mean S.D.
I II III IV
SAS total score – 33.09 6.88
SAS Factor I .81 – 10.55 2.88
SAS Factor II .60 .42 – 4.67 1.25
SAS Factor III .75 .40 .29 – 11.17 2.96
SAS Factor IV .71 .46 .39 .32 – 6.67 2.15
ACQ .45 .47 .25 .28 .29 22.82 6.32
BVS .40 .34 .29 .29 .28 22.14 20.44
FNE .42 .45 .16 .31 .24 13.99 7.90
CES-D .64 .59 .31 .49 .40 14.83 9.71
Note. All r’s are significant, P < .001. SAS factor labels assigned in the present study: Factor I: anxiety
and panic; Factor II: vestibular sensations; Factor III: somatic control; Factor IV: gastrointestinal/
muscular sensations. ACQ: Agoraphobic Cognitions Questionnaire; BVS: Body Vigilance Scale;
FNE: Fear of Negative Evaluation Scale; and CES-D: Center for Epidemiological Studies-Depression
Scale.
 B.O. Olatunji et al. / Anxiety Disorders 20 (2006) 543–561 551

cognitions (ACQ) than the other factors. The SAS ‘‘vestibular sensations’’ factor
was more related to tendency to attend to panic and agoraphobic-related cognitions
and panic-related body sensations (BVS) than the fear of negative evaluation. The
‘‘somatic control’’ and ‘‘gastrointestinal/muscular sensations’’ factors appear to
have low-to-moderate correlations with each criterion variable at about the same
strength. Means and standard deviations (S.D.) of the SAS lower-order factors and
measures of anxiety and depression are also listed in Table 2.

1.3. Discussion

Findings from Study 1 suggest that the SAS is composed of four lower-order
factors assessing: (1) anxiety and panic; (2) vestibular sensations; (3) somatic
control; and (4) gastrointestinal/muscular sensations. The four lower-order SAS
factors also appear to load on a single higher-order arousal factor. The four factors
were generally stable and demonstrated significant correlations with theoretically
related variables. The four lower-order SAS factors were also generally more
strongly associated with the CES-D, a measure of depression (range: .31 to .59).
However, the SAS total score demonstrated the most robust correlation with the
other scales suggesting that the factors may be less potent by themselves. The
SAS factor structure obtained in Study 1 has important implications for the
validity and utility of the SAS in nonclinical samples. However, given that this is
the first study to report on the factor structure of the SAS, replication of these
findings in an independent sample would bolster confidence in their reliability.
Accordingly, we elected to repeat our examination of the SAS in a second sample.

2. Study 2

2.1. Method

The study questionnaires were administered to a second sample of

undergraduate students recruited from introductory psychology courses at
University of North Carolina at Chapel Hill. This sample consisted of 443
participants, including 332 women (74.9%), with a mean age of 19.0. The sample
was 77% White/Caucasian (n = 341), followed by 56 Black/African Americans
(12.6), 21 Asians or Pacific Islanders (4.7%), and 25 participants (5.6%) of other,
multiple, or unreported ethnicities. The measures and procedures were identical to
those used in Study 1.

2.2. Results

2.2.1. Reliability and item-level analyses

The mean SAS total score was 32.66 (S.D. = 6.86). SAS total scores for
women (M = 33.23, S.D. = 7.24) were higher than those for men (M = 30.97,
552 B.O. Olatunji et al. / Anxiety Disorders 20 (2006) 543–561

Table 3
Self-Rating Anxiety Scale: item means and standard deviations, obliquely rotated factor loadings, and
communalities for the four-factor solution from Study 2
SAS item M S.D. SAS factor
I II III IV h2
2. I feel afraid for no reason at all 1.23 .69 .76 .05 .04 .06 .58
3. I get upset easily or feel panicky 1.58 .71 .71 .08 .04 .02 .53
1. I feel more nervous and anxious than usual 1.64 .69 .70 .17 .10 .09 .57
4. I feel like I’m falling apart and going to pieces 1.38 .61 .58 .06 .06
.08 .42
18. My face gets hot and blushes 1.56 .71 .54 .15
.07
.11 .31
9. I feel calm and can sit still easily 2.17 .98 .14 .69
.07
.16 .57
5. I feel that everything is all right and 2.44 .88 .23 .61 .02 .20 .44
nothing bad will happen
17. My hands are usually dry and warm 2.64 1.08
.04 .60 .03 .01 .36
13. I can breathe in and out easily 1.44 .85
.10 .60 .12
.04 .40
19. I fall asleep easily and get a good night’s rest 2.28 .94
.03 .54
.08
.34 .42
12. I have fainting spells or feel like it 1.12 .40
.08 .05 .82 .18 .62
11. I am bothered by dizzy spells 1.21 .49
.08 .03 .79
.10 .66
10. I can feel my heart beating fast 1.47 .64 .18
.06 .59
.05 .46
14. I get feelings of numbness and tingling in 1.28 .54 .07
.02 .42
.14 .25
my fingers and toes
6. My arms and legs shake and tremble 1.15 .41 .18 .01 .33
.23 .30
7. I am bothered by headaches neck and back pain 1.73 .83
.13 .01 .07
.76 .56
8. I feel weak and get tired easily 1.63 .81 .03 .07 .11
.71 .60
15. I am bothered by stomachaches or indigestion 1.54 .74 .18 .09 .11
.46 .36
20. I have nightmares 1.41 .59 .08 .15
.01
.44 .27
16. I have to empty my bladder often 1.77 .88 .27
.11 .01
.30 .21
% Variance of rotated factors 23.69 7.98 7.04 6.11
Note. Factor loadings j.30j are listed in boldface type. The first five eigenvalues were 4.74, 1.59, 1.40,
1.22, and 1.11.

S.D. = 5.22), t(441) = 3.03, P < .01. Given that the scale consisted of 20 items,
these mean SAS total scores indicate that participants in the second study also
tended to indicate between ‘‘None or a little of the time’’ or ‘‘Some of the time’’
agreement with the scale items. Means and standard deviations for the SAS items
are presented in Table 3. Mean scores on 16 out of 20 items were below 2.0 (i.e.,
‘‘Some of the time’’ agreement with the item), suggesting that the content of most
SAS items was generally outside of the experience of most participants. The SAS
demonstrated adequate internal consistency (a = .81). Based on the criterion of
.30 as an acceptable corrected item-total correlation (Nunnally & Bernstein,
1994), all 20 items performed adequately (range: .34 to .59).

2.2.2. Factor structure of the SAS

Exploratory factor analysis was used to reexamine the factor structure of the
SAS. Although confirmatory factor analysis (CFA) is sometimes used in similar
situations, an exploratory approach appears to be a more appropriate analytic
 B.O. Olatunji et al. / Anxiety Disorders 20 (2006) 543–561 553

strategy as only our Study 1 has investigated the factor structure of the SAS to
date. Thus, additional exploratory research on the SAS’s factor structure is needed
before researchers attempt to confirm its latent structure. As in Study 1, the lower-
order factor structure of the SAS was examined using PCA with Oblimin rotation.
Parallel analysis indicated a four-factor solution for the mean and 95th percentile
eigenvalues. However, a three-factor solution was indicated for the 95th
percentile eigenvalues. This somewhat discrepant finding is likely due to the
smaller N in the second sample. We specified a three-factor solution but the
resulting structure did not yield an interpretable pattern of item-factor loadings or
sufficient stability. Accordingly, for purposes of interpretability and consistency
with our Study 1 we elected to extract four factors.
Table 3 displays the item means and standard deviations, factor loadings, and
communalities for the four-factor SAS solution from Study 2. As can be seen,
these results are highly comparable with those from Study 1 (see Table 1). The
solution accounted for 44.8% of the SAS item variance. Consistent with Study 1,
the pattern of loadings in Table 3 suggests the following factor labels: ‘‘anxiety
and panic’’ (Factor I, five items), ‘‘somatic control’’ (Factor II, five items),
‘‘vestibular sensations’’ (Factor III, five items), and ‘‘gastrointestinal/muscular
sensations’’ (Factor IV, five items).

2.2.3. Replicability of the four-factor SAS solution

To examine the replicability of the four-factor SAS solution, coefficients of
congruence (Gorsuch, 1983) were computed between the factor loadings from
PCA in the present study and those reported in Study 1. These data are presented
in Table 4. The first factor from the present study, labeled ‘‘anxiety and panic,’’
was highly comparable with the ‘‘anxiety and panic’’ factor from Study 1
(coefficient of congruence = .92). The second factor from the present study,
labeled ‘‘somatic control,’’ was most similar to Factor III, ‘‘somatic control’’ from
Study 1 (coefficient of congruence = .93). The third factor from the present study,

Table 4
Coefficients of congruence between Self-Rated Anxiety Scale (SAS) factors from Study 1 and Study 2
Factor from Study 1 Factor from Study 2
I II III IV
I .92 .19 .19 .13
II .04 .03 .93 .24
III .18 .93 .03 .16
IV .14 .07 .15 .87
Note. Coefficients between SAS Factor IV from Study 2 and other variables, although negative in sign,
reflect positive relationships; these are reported here as positive correlations for ease of interpretability.
Coefficients of congruence (Gorsuch, 1983) were derived using loadings from the factor pattern
matrix. SAS factor labels assigned in Study I: Factor I: anxiety and panic; Factor II: vestibular
sensations; Factor III: somatic control; and Factor IV: gastrointestinal/muscular sensations. Corre-
sponding factors in Study 2 are listed in boldface type.
554 B.O. Olatunji et al. / Anxiety Disorders 20 (2006) 543–561

labeled ‘‘vestibular sensations,’’ was most similar to Factor II ‘‘vestibular

sensations’’ from Study 1 (coefficient of congruence = .93). The fourth factor
from the present study, labeled ‘‘gastrointestinal/muscular sensations,’’ was most
similar to the ‘‘gastrointestinal/muscular sensations’’ factor from Study 1
(coefficient of congruence = .87). Coefficients of congruence between non-
corresponding factors were uniformly weak (range: .03 to .24).

2.2.4. Reexamination of the higher-order SAS structure

To reexamine the hierarchical structure of the SAS factor scores on the four
lower-order factors obtained in the initial analysis were factor analyzed using
PCA. A single factor was extracted in PCA (eigenvalues = 1.76, 0.88, 0.71, and
0.64) that accounted for 43.6% of the variance. This replicates the findings of
Study 1 that suggests that the four lower-order SAS factors appear to load on a
single higher-order arousal factor.

2.2.5. Correlates of the SAS and its factors

Table 5 presents correlations between the SAS, the lower-order SAS factors,
and the ACQ, BVS, FNE, and CES-D. The SAS lower-order factors were strongly
correlated with SAS total scores (range: .67 to .76). The SAS ‘‘anxiety and panic’’
factor was most strongly associated with the vestibular sensations and the
gastrointestinal/muscular sensations factor (r’s = .41 and .44, respectively), while
the somatic control factor was most strongly associated with the gastrointestinal/
muscular sensations factor (r = .34). Though highly correlated with measures of
fearful cognitions, body vigilance, and negative evaluation (range: .42 to .49), the
SAS total scores were most strongly associated with depression (r = .61,

Table 5
Pearson correlations between the Self-Rating Anxiety Scale factors and related measures from Study 2
Scale SAS total score SAS factor scores Mean S.D.
I II III IV
SAS total score – 32.66 6.86
SAS Factor I .72 – 7.39 2.24
SAS Factor II .74 .30 – 10.97 3.05
SAS Factor III .67 .41 .31 – 6.21 1.64
SAS Factor IV .76 .44 .34 .45 – 8.07 2.44
ACQ .49 .48 .28 .33 .34 22.46 6.03
BVS .40 .30 .25 .33 .31 21.38 19.22
FNE .42 .43 .33 .19 .23 13.96 7.90
CES-D .61 .57 .42 .35 .43 14.54 9.82
Note. All r’s are significant, P < .001. SAS factor labels assigned in the present study: Factor I: anxiety
and panic; Factor II: somatic control; Factor III: vestibular sensations; Factor IV: gastrointestinal/
muscular sensations. ACQ: Agoraphobic Cognitions Questionnaire; BVS: Body Vigilance Scale;
FNE: Fear of Negative Evaluation Scale; CES-D: Center for Epidemiological Studies-Depression
Scale.
 B.O. Olatunji et al. / Anxiety Disorders 20 (2006) 543–561 555

P < .001) when calculating the difference between independent correlation

coefficients. Of the measures of anxiety, the SAS ‘‘anxiety and panic’’ factor was
more strongly associated with panic and agoraphobic-related cognitions (ACQ)
than the other factors. The SAS ‘‘vestibular sensations’’ and ‘‘gastrointestinal/
muscular sensations’’ factors was more related to tendency to attend to panic and
agoraphobic-related cognitions and panic-related body sensations (BVS) than the
fear of negative evaluation. The ‘‘somatic control’’ factor appears to be low-to-
moderately correlated with each criterion variable at about the same strength.
Means and standard deviations (S.D.) of the SAS lower-order factors and
measures of anxiety and depression are also listed in Table 5.

2.3. Discussion

The findings from Study 2 suggest that the SAS is composed of four lower-
order factors assessing: (1) anxiety and panic; (2) somatic control; (3) vestibular
sensations; and (4) gastrointestinal/muscular sensations. These factors were
generally stable, and demonstrated significant correlations with theoretically
related variables. Consistent with Study 1, the four lower-order SAS factors
appear to load on a single higher-order arousal factor. The four lower-order SAS
factors were also generally more strongly associated with the CES-D, a measure
of depression (range: .35 to .57).

3. General discussion

The present study represents the first factor analytic study of the SAS. We
evaluated the factor structure and psychometric properties of this self-report
instrument in two large, independent nonclinical samples. Studies 1 and 2
converged to indicate that the SAS is internally consistent, with all items
correlating moderately to highly with the total score. In accord with previous
findings (e.g., Silverstein, 2002), significant gender differences were detected in
both studies on SAS total scores such that females reported more somatic
complaints than males (females also reported more somatic complaints than
males on each of the four factors in Study 1 and Study 2, P’s < .05). Item analysis
revealed that participants tended to endorse either ‘‘None or a little of the time’’ or
‘‘Some of the time’’ agreement with the vast majority of SAS items. These
findings suggest that the somatic complaints as assessed by the SAS items are
relatively far removed from the experience of most nonclinical participants. Thus,
it is possible that the SAS is not an optimal measure of somatic complaints in
nonclinical populations. Item analysis revealed that items assessing somatic
sensations related to musculoskeletal (i.e., ‘‘My arms and legs shake and
tremble’’) and respiratory (i.e., ‘‘I have fainting spells or feel like it’’) systems
were endorsed less highly than were other items. Limited variability in these items
would call into question the generalizability of the results to clinical populations.
556 B.O. Olatunji et al. / Anxiety Disorders 20 (2006) 543–561

However, future research may help determine if musculoskeletal and respiratory

complaints are less normative and are more evident in specific anxiety disorders
(i.e., GAD, panic disorder).
Prior research on the fear and experience of somatic complaints suggests that
somatic concerns are multidimensional in nature. For instance, factor analytic
studies have found fear of somatic sensations consisting of three factors; fear of
arousal-related symptoms, fear of gastrointestinal symptoms, and fear of cardiac
symptoms (e.g., Olatunji et al., 2005). Furthermore, Liu et al. (1997) found a
strong primary factor of somatic experiences consisting of gastrointestinal
complaints, pain, and throat problems and a second factor consisted of
cardiopulmonary symptoms and pain in the extremities. The four symptom
categories (gastrointestinal and cardiopulmonary complaints, throat problems,
and pain) also loaded strongly on the unrotated first factor. Similarly, results from
a series of exploratory factor analyses in the present study revealed that the SAS
consists of four lower-order factors, all of which load on a single higher-order
factor. These lower-order factors were assigned the following labels: (1) ‘‘anxiety
and panic’’; (2) ‘‘vestibular sensations’’; (3) ‘‘somatic control’’; and (4)
‘‘gastrointestinal/muscular sensations.’’ These results converge with prior
findings suggesting that somatic complaints are multidimensional and that they
may be hierarchically organized.
Although, each of the four lower-order factors were generally stable, factor
analysis did yield one complex item (i.e., an item with salient loadings on more
than one factor) in Study 1 (item 10) and one complex item in Study 2 (item 19).
In addition, we found two instances in which items had primary loadings on
different factors in Study 1 and Study 2. For example, in Study 1, item 10 (‘‘I
can feel my heart beating fast’’) loaded primarily on the anxiety and panic
factor (Factor 1), whereas it loaded primarily on the vestibular sensations factor
(Factor III) in Study 2. In Study 1, item 20 (‘‘I have nightmares’’) loaded
primarily on the anxiety and panic factor (Factor 1), whereas it loaded primarily
on the gastrointestinal/muscular sensations factor (Factor IV) in Study 2.
Although item 10 does appear to have a face valid rationale for loading on either
the anxiety and panic factor or the vestibular sensations factor, future research
appears warranted to examine the somatic distinctiveness of complaints of
nightmares.
An interesting factor that emerged consistently in the present study was
‘‘somatic control’’ (‘‘I can breathe in and out easily’’). Prior studies have shown
that control is associated with lower levels of physical and psychological
symptoms of stress (e.g., Spector, 1986) and more recent findings have shown that
limited control is significantly associated with somatic complaints (e.g., Gebhardt
& Brosschot, 2002). The present findings appear to converge with prior research
emphasizing the importance of control or perceived control in the phenomen-
ology of somatic complaints (as indicated in the methods section, the items for the
‘‘somatic control’’ factor is reversed scored. Thus, the scores on this factor are
actually negatively correlated with the other scales in Tables 2 and 5).
 B.O. Olatunji et al. / Anxiety Disorders 20 (2006) 543–561 557

An important goal of the present study was to attempt to replicate the SAS
factor structure. Coefficients of congruence (Gorsuch, 1983) indicated that the
SAS factors were highly replicable across both studies. Specifically, the ‘‘anxiety
and panic’’ factor reported in Study 1 was most congruent with the ‘‘anxiety and
panic’’ factor reported in Study 2. The ‘‘vestibular sensations’’ factor reported in
Study 1 was also most congruent with the ‘‘vestibular sensations’’ factor reported
in Study 2. A similar pattern of congruence between Study 1 and Study 2 was
found for the ‘‘somatic control’’ and ‘‘gastrointestinal/muscular sensations’’
factors (see Table 4). Importantly, the factors displayed very little congruence
with other factors. Results from these analyses support the replicability and
distinctiveness of SAS factors and suggest that researchers using the SAS in
nonclinical samples are likely to obtain similar results.
In the present study, the SAS demonstrated adequate convergent validity.
Specifically, the SAS and its factors were significantly correlated with measures
of anxiety. However, there was evidence of limited divergent validity as the SAS
and its factors appear to be low-to-moderately correlated with measures of panic
and agoraphobic cognitions, body vigilance, and the fear of negative evaluation at
about the same strength. The SAS and its factors also displayed statistically
significant correlations with depression, thus further questioning its divergent
validity.1 The relation between the SAS and its factors and depression may be
expected given that prior research suggests that somatic symptoms (i.e., back
pain, feelings of heaviness/lightness in parts of the body, periods of bodily
weakness, fatigue, and tension) are significantly correlated with depression (e.g.,
Goldberg, 1996). Prior research has also shown that somatic complaints
significantly contribute to the onset (e.g., Barkow et al., 2004) and course (e.g.,
Wilson, Widmer, Cadoret, & Judiesch, 1983) of depression. However,
examination of individual SAS items, e.g. (‘‘I feel afraid for no reason at
all’’; ‘‘I fall asleep easily and get a good night’s rest’’) and CES-D items (e.g., ‘‘I
felt fearful’’; ‘‘My sleep was restless’’) suggests that the heightened relation
between these measures may be partially attributable to item-content overlap. The
item-content overlap between the SAS factors and the CES-D may also reflect the
conceptual overlap between anxiety and depression (e.g., Clark & Watson, 1991;
Joiner, 1996). The heightened relation between the SAS and the CES-D may also
be attributable to content overlap in measurement properties as the two scales
primarily assess frequency of depressive (CES-D: ‘‘rarely or none of the time’’ to
most or all of the time’’) and somatic (SAS: ‘‘none or a little of the time’’ to ‘‘most
or all of the time’’) complaints.
The SAS appears to be a useful measure of somatic anxiety. These findings
may have some value for retrospective analyses of somatic symptoms in archival
data sets pertaining to anxiety-related pathology when the SAS has been used as a

1
The items that composed each SAS factor were also included in the computation of the SAS total
score. Without this potential inflation factor, the CES-D might correlate more highly with the SAS
total score than the individual SAS factors.
558 B.O. Olatunji et al. / Anxiety Disorders 20 (2006) 543–561

unitary construct. The SAS may also be more ideal than alternative measures
given its broader content sampling of somatic symptoms. However, future studies
confirming our psychometric examination of the SAS will be necessary. For
instance, future research is needed to examine whether the factor structure of the
SAS varies across and within different samples. Indeed, prior research suggests
that the factor structure and factor stability of somatic complaints may differ for
males and females (e.g., Liu et al., 1997). Examination of the factor structure of
the SAS in different cultures is also warranted as prior research suggests that the
structure of somatic symptoms may differ across cultures (e.g., Mumford et al.,
1991). A limitation of the present research was our use of an undergraduate
sample. Extensive research examining the factor structure of the SAS in diverse
community and clinical samples may provide useful information on the
generalizability of the present findings. For instance, clinical samples may be
more likely to reveal different somatic factors (i.e., anxiety and panic and
vestibular sensations may form one factor in panic patients). Similarly, the SAS
may have utility as an outcome indicator in the treatment of specific anxiety-
related disorders. With the hierarchical scaling of the measure, perhaps different
factors will be more sensitive to treatment (e.g., anxiety and panic in panic
disorder; gastrointestinal/muscular sensations in irritable bowel syndrome).
However, future studies will need to reexamine the factor structure and construct
validity across diverse samples before confident inferences can be drawn
regarding the utility of a hierarchical scaling of the SAS.

Acknowledgments

The authors thank Carol M. Woods for her assistance with data collection.
Preparation of this manuscript was supported by NIMH NRSA grant
1F31MH067519-1A1 awarded to Bunmi O. Olatunji.

References

Abramowitz, J. S., Schwartz, S. A., & Whiteside, S. P. (2002). A contemporary conceptual model of
hypochondriasis. Mayo Clinic Proceedings, 77, 1323–1330.
Apfledorf, W. J., Shear, M. K., Leon, A. C., & Portera, L. (1994). A brief screen for panic disorder.
Journal of Anxiety Disorders, 8, 71–78.
Barkow, K., Heun, R., Üstün, T. B., Berger, M., Bermejo, I., Gaebel, W., et al. (2004). Identification of
somatic and anxiety symptoms which contribute to the detection of depression in primary health
care. European Psychiatry, 19, 250–257.
Beck, A. T., Epstein, N., Brown, G., & Steer RA, (1988). An inventory for measuring clinical anxiety:
psychometric properties. Journal of Consultation and Clinical Psychology, 56, 893–897.
Chambless, D. L., Caputo, G. C., Bright, P., & Gallagher, R. (1984). Assessment of fear of fear in
agoraphobics: the body sensations questionnaire and the Agoraphobic Cognitions Questionnaire.
Journal of Consulting and Clinical Psychology, 52, 1090–1097.
Clark, D. M. (1986). A cognitive approach to panic. Behaviour Research and Therapy, 24, 461–470.
 B.O. Olatunji et al. / Anxiety Disorders 20 (2006) 543–561 559

Clark, L. A., & Watson, D. (1991). Tripartite model of anxiety and depression: psychometric evidence
and psychometric implications. Journal of Abnormal Psychology, 100, 316–336.
Cox, B. J., Cohen, E., Direnfeld, D. M., & Swinson, R. P. (1996). Does the Beck Anxiety Inventory
measure anything beyond panic attack symptoms? Behaviour Research and Therapy, 34, 949–
954.
Ehlers, A. (1991). Cognitive factors in panic attacks: symptom probability and sensitivity. Journal of
Cognitive Psychotherapy, 5, 157–173.
Eifert, G. H. (1992). Cardiophobia: a paradigmatic behavioural model of heart-focused anxiety and
non-anginal chest pain. Behaviour Research and Therapy, 30, 329–345.
Gebhardt, W. A., & Brosschot, J. F. (2002). Desirability of control: psychometric properties and
relationships with locus of control, personality, coping, and mental and somatic complaints in three
Dutch samples. European Journal of Personality, 16, 423–438.
Goldberg, D. (1996). A dimensional model for common mental disorders. British Journal of
Psychiatry, 168, 44–49.
Gorsuch, R. L. (1983). Factor analysis. Hillsdale, NJ: Erlbaum.
Guadagnoli, E., & Velicer, W. F. (1988). Relation of sample size to the stability of component patterns.
Psychological Bulletin, 103, 265–275.
Hoehn-Saric, R., McLeod, D. R., Funderburk, F., & Kowalski, P. (2004). Somatic symptoms and
physiological responses in generalized anxiety disorder and panic disorder. Archives of General
Psychiatry, 61, 913–921.
Horn, J. L. (1965). A rationale and test for the number of factors in factor analysis. Psychometrika, 30,
179–185.
Jegede, R. O. (1977). Psychometric attributes of the Self-Rating Anxiety Scale. Psychological Reports,
40(303), 306.
Joiner, T. E. (1996). A confirmatory factor-analytic investigation of the tripartite model of depression
and anxiety in college students. Cognitive Therapy and Research, 20, 521–539.
Joormann, J., & Stober, J. (1999). Somatic symptoms of generalized anxiety disorder from the DSM-
IV: associations with pathological worry and depression symptoms in a nonclinical sample.
Journal of Anxiety Disorders, 13, 491–503.
Kellner, R. (1990). Somatization: theories and research. The Journal of Nervous and Mental Disease,
178, 150–160.
Koksal, F., & Power, K. G. (1990). Four systems anxiety questionnaire (FSAQ): a self-report measure
of somatic, cognitive, behavioral and feeling components. Journal of Personality Assessment, 54,
534–544.
Koksal, F., Power, K. G., & Sharp, D. M. (1991). Profiles of DSM III anxiety disorders on the somatic,
cognitive, behavioural and feeling components of the four systems anxiety questionnaire.
Personality and Individual Differences, 12, 643–651.
Kroenke, K. (2003). Patients presenting with somatic complaints: epidemiology, psychiatric co-
morbidity and management. International Journal of Methods in Psychiatric Research, 12, 34–43.
Landy, F. J., & Stern, R. M. (1971). Factor analysis of a somatic perception questionnaire. Journal of
Psychosomatic Research, 15, 179–181.
Lang, P. J. (1971). The application of psychophysiological methods. In: S. L. Garfield, & A. E. Bergin
(Eds.), Handbook of psychotherapy and behaviour change. New York: Wiley.
Liu, G., Clark, M. R., & Eaton, W. W. (1997). Structural factor analyses for medically unexplained
somatic symptoms of somatization disorder in the Epidemiologic Catchment Area Study.
Psychological Medicine, 27, 617–626.
Longman, R. S., Cota, A. A., Holden, R. R., & Fekken, G. C. (1989). A regression equation for the
parallel analysis criterion in principal components analysis: mean and 95th percentile eigenvalues.
Multivariate Behavioral Research, 24, 59–69.
Maranell, G. M. (1974). Scaling: a source book for behavioural sciences. Chicago: Aldine.
McNally, R. J., & Lorenz, M. (1987). Anxiety sensitivity in agoraphobics. Journal of Behavior
Therapy and Experimental Psychiatry, 18, 3–11.
560 B.O. Olatunji et al. / Anxiety Disorders 20 (2006) 543–561

Mersch, P. P. A., Hilderbrand, M., Lavy, E. H., Wessel, I., & Van Hout, W. J. P. J. (1992). Somatic
symptoms in social phobia: a treatment method based on rational emotive therapy and paradoxical
interventions. Journal of Behavior Therapy and Experimental Psychiatry, 23, 199–211.
Michelson, L., & Mavissakalian, M. (1983). Temporal stability of self-report measures in agoraphobia
research. Behaviour Research and Therapy, 21, 695–698.
Mumford, D. B., Bavington, J. T., Bhatnagar, K. S., Hussain, Y., Mirza, S., & Naraghi, M. M. (1991).
The Bradford Somatic Inventory: a multi-ethnic inventory of somatic symptoms reported by
anxious and depressed patients in Britain and the Indo-Pakistan subcontinent. British Journal of
Psychiatry, 158, 379–386.
Norton, G. R., Dorward, J., & Cox, B. J. (1986). Factors associated with panic attacks in nonclinical
participants. Behavior Therapy, 17, 239–252.
Nunnally, J., & Bernstein, I. (1994). Psychometric theory. New York: McGraw-Hill.
Oei, T. P., Kenna, D., & Evans, L. (1991). The reliability, validity, and utility of the SAD and FNE
scales for anxiety disorder patients. Personality and Individual Differences, 12, 111–116.
Olatunji, B. O., Sawchuk, C. N., Deacon, B. J., Tolin, D. F., Williams, N. L., Lilienfeld, S. O., et al.
(2005). The Anxiety Sensitivity Profile revisited: Factor structure and psychometric properties in
two non-clinical samples. Journal of Anxiety Disorders, 19, 603–625.
Radloff, L. S. (1977). The CES-D Scale: a self-report depression scale for research in the general
population. Applied Psychological Measurement, 1, 385–401.
Reiss, S., & McNally, R. J. (1985). Expectancy model of fear. In: S. Reiss, & R. R. Bootzin (Eds.),
Theoretical issues in behavior therapy (pp. 107–121). San Diego, CA: Academic Press.
Santor, D. A., Zuroff, D. C., Ramsay, J. O., Cervantes, P., & Palacios, J. (1995). Examining scale
discriminibility in the BDI and CES-D as a function of depressive severity. Psychological
Assessment, 7, 131–139.
Schmidt, N. B. (1999). Examination of differential anxiety sensitivities in panic disorder: a test of
anxiety sensitivity subdomains predicting fearful responding to a 35% CO2 challenge. Cognitive
Research and Therapy, 23, 3–19.
Schmidt, N. B., Lerew, D. R., & Trakowski, J. H. (1997). Body vigilance in panic disorder:
evaluating attention to bodily perturbations. Journal of Consulting and Clinical Psychology, 65,
214–220.
Schönemann, P. H., & Wang, M. M. (1972). Some new results on factor indeterminacy. Psychometrika,
37, 61–91.
Silverstein, B. (2002). Gender differences in the prevalence of somatic versus pure depression: a
replication. American Journal of Psychiatry, 159, 1051–1052.
Spector, P. E. (1986). Perceived control by employees: a meta-analysis of studies concerning autonomy
and participation at work. Human Relations, 39, 1005–1016.
Taylor, S., & Asmundson, G. J. G. (2004). Treating health anxiety: a cognitive–behavioral approach.
New York, NY: Guilford Press.
Taylor, S., & Cox, B. J. (1998). Anxiety sensitivity: multiple dimensions and hierarchic structure.
Behaviour Research and Therapy, 36, 37–51.
Taylor, S., Koch, W. J., & McNally, R. J. (1992). How does anxiety sensitivity vary across the anxiety
disorders? Journal of Anxiety Disorders, 6, 249–259.
Thurstone, L. L. (1947). Multiple factor analysis. Chicago, IL: University of Chicago Press.
Van Ommeren, M., Sharma, B., Sharma, G. K., Komproe, I., Cardena, E., & de Jong, J. T. V. M. (2002).
The relationship between somatic and PTSD symptoms among Bhutanese refugee torture
survivors: examination of comorbidity with anxiety and depression. Journal of Traumatic Stress,
15, 415–421.
Watson, D., & Friend, R. (1969). Measurement of social-evaluative anxiety. Journal of Consulting and
Clinical Psychology, 33, 448–457.
Wilson, D. R., Widmer, R. B., Cadoret, R. J., & Judiesch, K. (1983). Somatic symptoms: a major
feature of depression in a family practice. Journal of Affective Disorders, 5, 199–207.
 B.O. Olatunji et al. / Anxiety Disorders 20 (2006) 543–561 561

Zatzick, D., Russio, J. E., & Katson, W. (2003). Somatic, posttraumatic stress, and depressive
symptoms among injured patients treated in trauma surgery. Psychosomatics Journal of Con-
sultation Liaison Psychiatry, 44, 479–484.
Zung, W. W. K. (1971). A rating instrument for anxiety disorders. Psychosomatics, 12, 371–379.
Zwick, W. R., & Velicer, W. F. (1986). Comparisons of five rules for determining the number of
components to retain. Psychological Bulletin, 99, 432–442.