Content uploaded by Sherry H Stewart
Author content
All content in this area was uploaded by Sherry H Stewart
Content may be subject to copyright.
Anxiety Sensitivity and Negative Interpretation Biases:
Their Shared and Unique Associations with Anxiety
Symptoms
Janine V. Olthuis & Sherry H. Stewart & Margo C. Watt &
Brigitte C. Sabourin & Edmund Keogh
Published online: 14 April 2012
#
Springer Science+Business Media, LLC 2012
Abstract Anxiety sensitivity (AS) is a psychological risk
factor for anxiety disorders. Negative interpretation biases
are a maladaptive form of information-processing also as-
sociated with anxiety disorders. The present study explored
whether AS and negative interpretation biases make inde-
pendent contributions to variance in panic and generalized
anxiety symptoms and whether particular interpretation bias
domains (e.g., of ambiguous arousal sensations) have spe-
cific associations with panic and/or generalized anxiety
symptoms. Eighty-nine female undergraduates (44 low
AS; 45 high AS) completed measures of AS, interpretation
biases, and panic and generalized anxiety symptoms.
Findings showed that AS and negative interpretation biases
both significantly added to the prediction of anxiety symp-
toms. Negative interpretations of ambiguous arousal sensa-
tions were uniquely associated with panic symptoms, while
negative interpretations of ambiguous general and social
events were uniquely associated with generalized anxiety
symptoms. Findings support the conceptual validity of AS
and negative interpretation biases and their unique and
shared contributions to anxiety symptoms.
Keywords Anxiety
.
Anxiety sensitivity
.
Interpretation
biases
.
Women
Several prominent cognitive theories of anxiety (e.g., Beck
et al. 1985; Clark 1986; Reiss 1991) emphasize the role of
maladaptive beliefs and biased information-proce ssing in
the development and maintenance of anxiety symptoms
and disorders. For instance, Clark’s theory of panic (1986 )
suggests that individuals predisposed to maladaptive beliefs
about physical sensations tend to misinterpre t increases in
physiological arousal (e.g., heart racing) as evidence for
impending catastrophe (e.g., heart attack) which leads to
further increases in arousal, culminating in a panic attack.
Weems and Watts (20 05) suggest that these maladaptive
forms of information processing may include attentional
biases, wherein individua ls selectively attend to threatening
or negative information; memory biases, in which negative
or threatening information is more likely to be encoded and
remembered; and interpretation biases, wherein individuals
orient toward threatening and away from safe aspects of
situations, behaviour, and other stimuli. Theories such as
these suggest that individuals who demonstrate malad aptive
ways of processing and interpreting stimuli may be at in-
creased risk for developing anxiety relative to those display-
ing a more normative pattern of thinking.
As well as theory, there is also good empirical evidence
that negative thinking is important in anxiety. Negative
interpretation biases – a tendency or predisposition to inter-
pret ambiguity or neutral information in a threatening man-
ner (Weems et al. 2007) – have been widely associated with
anxiety. For instance, research on anxiety disorders suggests
J. V. Olthuis (*)
:
S. H. Stewart
:
M. C. Watt
:
B. C. Sabourin
Department of Psychology, Dalhousie University,
1355 Oxford Street, PO Box 15000, Halifax, NS,
Canada B3H 4R2
e-mail: janine.olthuis@dal.ca
S. H. Stewart
Department of Psychiatry, Dalhousie University,
Halifax, NS, Canada
M. C. Watt
Department of Psychology, Saint Francis Xavier University,
Antigonish, NS, Canada
E. Keogh
Department of Psychology, University of Bath,
Bath, UK
J Psychopathol Behav Assess (2012) 34:332–342
DOI 10.1007/s10862-012-9286-5
that individuals with panic disorder (PD) are more likely to
negatively interpret arousal sensations and endorse stronger
beliefs in these interpretations than non-anxious controls
(Austin and Kiropoulos 2008; Clark et al. 1997). Similarly,
those with generalized anxiety disorder (GAD) are more
likely to interpret ambiguous general and social events
negatively as compared to non-anxious controls (Clark et
al. 1997). Individuals with high social anxiety are also more
likely to interpret ambiguous social events in a threatening
and negative fashion compared to those with low social
anxiety (Kanai et al. 2010). Recent research has shown that
directly modifying cognitive biases and promoting a more
positive interpretation bias, using a technique known as
Cognitive Bias Modification, can reduce anxiety symptoms
(Brosan et al. 2011).
Research has also examined whether there are stable
vulnerability factors that may predispose people towards
panic. For example, anxiety sensitivity (AS) is an enduring
fear of anxiety-related sensations (e.g., increased heart rate)
arising from concerns that they will have serious physical,
psychological, or social consequences (e.g., heart attack;
Reiss and McNally 1985). In contrast to those with high
AS, individuals with low AS tend to regard these sensations
as unpleasant but harmless (McNally 1999). As an individ-
ual difference factor (i.e., a trait-like variable), AS has been
implicated in many anxiety disorders, including PD, GAD,
social phobia, and hypochondriasis (e.g., Naragon-Gainey
2010; Norton and Asmundson 2004; Schmidt et al. 2006).
Research shows that individuals with these anxiety disor-
ders show higher levels of AS than the general population
(for review, see Naragon-Gainey 2010). Furthermore,
results of both cross-sectional (e.g., Cox et al. 2001) and
longitudinal studies (e.g., Hayward et al. 2000) have shown
AS to predict panic, suggesting that AS is more than simply
a consequ ence of panic, but instead may actually serve as a
vulnerability factor that contributes towards its develop-
ment. Moreover, targeting AS has been associated with
reductions in anxiety symptoms in several treatment studies
(Smits et al. 2008).
Empirical research has examined the associations be-
tween these distinct but related predisposing characteristics
– cognitive biases and AS – in contributing to anxiety
collectively; however, questions remain about the unique
contribution of each in explaining anxiety symptoms
(Weems et al. 2007). In particular, research investigating
the incremental contribution of AS and interpretation biases
in explaining anxiety symptoms is quite limit ed. As such,
the present study sought to fill this gap by focusing on the
shared and unique associations of AS and negative interpre-
tation biases in predicting anxiety symptoms among young
women.
At first glance, AS and interpretation biases may appear to
be similar concepts, raising question as to their distinctiveness.
Indeed, re sear chers (e.g., Weems et al. 2007)havebeen
known to describe AS and interpretation biases as somewhat
the same construct. Given the empirical and theoretical asso-
ciations between AS and anxiety, high AS individuals and
those with anxiety disorders likely share similar information-
processing biases (Teachman 2005), such as negative inter-
pretation biases (Vancleef and Peters 2008). As noted by
Allport (1937), however, “personality is something and per-
sonality does something” (p.48). Applied to this case, one
might say that AS is the trait and negative interpretation biases
are one of the cognitive processes by which this trait predis-
poses an individual to anxiety symptoms. In a similar vein,
Cox (1996) suggested that AS could be considered a disposi-
tion, while misinterpretations would be considered a state,
manifesting only in the moment of interpretation. It is impor-
tant to note, however, that interpretation biases of the sort
under investigation in the present study are only one cognitive
process by which AS confers risk for anxiety symptoms.
There are other cognitive processes (e.g., judgment biases,
attentional biases, memory biases) as well as other affective,
physiological, behavioural, and social processes that may also
help explain the relation between AS and anxiety symptoms.
For instance, in the area of cognitive processes alone, there are
attentional and memory biases that have been documented
among high compared to low AS individuals (e.g., Noel et al.
2012; Stewart et al. 1998). The multidimensional nature by
which AS confers risk for anxiety symptoms helps explain the
distinctiveness of AS and interpretation biases and allows for
an association between AS and anxiety symptoms outside of
that which might be explained by interpretation biases.
Given their empirical and conceptual similarities, re-
search has investigated the relation between AS and nega-
tive interpretation biases. Research has found negative
interpretation biases in high AS individuals with PD, as well
as in high AS individuals who are panic-free, suggesting the
biases cannot be accounted for by experience with panic
alone (Richards et al. 2001; Teachman 2005). Keogh and
Cochrane (2002) found that high (vs. low) AS individuals
were more likely to negatively interpret arousal sensations
(e.g., racing heart) and general (e.g., rece iving a letter
marked ‘urgent’ in the mail) and social (e.g., being ignored
in the street by an acquaintance ) events, though not non
arousal-related body symptoms (e.g., pain in leg). The high
AS group also believed more strongly in these negative
interpretations. Similarly, Vancleef and Peters (2008) found
positive associations between AS and strength of belief in
negative interpretations and Keogh et al. (2004) found that
higher AS was associated with increased negative interpre-
tation biases in acute pain patients.
Taken together, evidence suggests that AS and negative
interpretation biases are associated with each other as well
as with anxiety symptoms. It remains unclear, however, if
AS and negative interpretation biases, when considered
J Psychopathol Behav Assess (2012) 34:332–342 333
together, have distinct associations with, or predict unique
variance in, anxiety symptoms. In other words, could the
association of AS with anxiety be entirely explained through
the association of more general negative interpretation
biases with anxiety or do the two independently predict
anxiety? To date, only Weems and colleagues (2007) have
considered this questio n as part of a larger study of the
association of cognitive errors, AS, a nd anxiety control
beliefs with anxiety and depressive symptoms among youth.
Using hierarchical regression, Weems et al. (2007) found
that anxiety control beliefs, AS, and cognitive errors (in-
cluding personalizing, selective abstraction, catastrophizing,
and overgeneralizing) each added significantly to the pre-
diction of anxiety symptoms beyond the other indices.
Whereas Weems and colleagues describe AS and interpre-
tation biases as falling under the same conceptual umbrella,
their findings that AS and interpretation bias measures
might make independent contributions to variance in anxi-
ety symptoms argues against them only tapping the same
construct. Given the conceptualization of interpretation
biases as one of the processes by which the trait of AS
confers risk for anxiety (Cox 1996), it seems important to
further distinguish these two concepts and recognize th e
multiple pathways by which AS confers risk. Thus, we
aimed to further investigate the unique and shared contribu-
tions of AS and interpretation biases to further our under-
standing of their conceptual distinction and overlap.
In addition, Weems and colleagues’ study was conducted
with youth, making it unclear if the findings would extend
to adults. Prior research has been inconsistent in suggesting
a role for age in moderating the associations between cog-
nitive errors, AS, and anxiety symptoms (e.g., Weems et al.
2001, 2007). We sought to further investigate this relation in
an adult population in the curre nt study. Moreover, we
aimed to extend the Weems et al. (2007) investigation to
explore the unique and shared relations of AS and interpre-
tation biases to different types of anxiety symptoms (e.g.,
panic symp toms and generalized anxiety symptoms) as well
as the relation of particular domains of negative interpreta-
tion biases (e.g., of arousal-related body sensations, of social
events, etc.) with these different types of anxiety symptoms.
Study Aims and Hypotheses
The primary aims of the current study were two-fold. First, we
aimed to investigate the conceptual validity of the constructs
of AS and negative interpretation biases. Specifically, we
examined whether AS and negative interpretation biases
(i.e., the likelihood of endorsing catastrophic consequences
of ambiguity and/or the strength of belief in negative inter-
pretations) have distinct as well as shared associations with
anxiety symptoms. In the present study we examined two
types of anxiety symptoms: panic-related symptoms and gen-
eralized anxiety-type symptoms. The panic-related symptoms
closely match the distinguishing symptoms of PD including
trembling, shakiness, shortness of breath, heart racing, dizzi-
ness, etcetera. The generalized anxiety-type symptoms reflect
those commonly endorsed by individuals with GAD including
irritability, difficulty relaxing, nervous tension, agitation, et-
cetera. It is important to note that the symptoms measured do
not exactly mirror DSM-IV-TR (APA 2000)diagnosticcrite-
ria for PD or GAD; however, due to their similarity to the
diagnostic criteria and in the interest of simplicity, we refer to
these collections of symptoms as ‘panic symptoms’ and ‘gen-
eralized anxiety symptoms’ in this manuscript.
We selected panic and generalized anxiety symptoms as the
focus of our analyses for two primary reasons. First, they both
show important associations with AS: elevated levels of AS
have been found amongst those with GAD (Rodriguez et al.
2004; Viana and Rabian 2008) while AS has been shown to
predict the development of panic attacks (Schmidt et al. 1997,
2008). Second, panic and generalized anxiety map onto two
key dimensions of anxiety-related psychopathology: fear and
anxiety. As a classic anxiety-based disorder, GAD is future-
oriented and focused on bodily symptoms of physical tension
and cognitive and affective symptoms of apprehension about
the future (Barlow et al. 2009). Conversely, as a classic fear-
based disorder, PD is an immediate alarm reaction to danger
with strong sympathetic nervous system arousal (Barlow et al.
2009). The differences between the roots and symptom pro-
files of these disorders suggest that they may have unique
associations with risk factors including AS and negative in-
terpretation biases that are worth investigating.
The second aim of our study was to extend the work of
Weems et al. (2007) by exploring the specificity of the
association between negative interpretation biases and anx-
iety symptoms within this context. While Weems and col-
leagues examined the components of interpretation biases
by type (e.g., catastrophizing, overgeneralizing, person aliz-
ing, and selective abstraction), we investigated specific in-
terpretation biases by content domain, placing the present
study in line with the majority of studies on interpretation
biases and anxiety in the literature. We examined whether
particular domains of interpretation biases – speci fical ly,
interpretation biases of ambiguous arousal-related body sen-
sations (e.g., interpreting feeling lightheaded and weak as a
sign one is going to faint), other body sensations (e.g.,
interpreting back pain as a sign something is wrong with
one’s spine), social events (e.g., interpreting an acquaintance
passing on the street without saying hello as a sign that they
think one is not worth talking to), or general events (e.g.,
interpreting a letter marked ‘urgent’ in the mail as news that
someone one knows has died), have specific associations
with either panic and/or generalized anxiety symptoms. This
allows us to further investigate the relation between specific
334 J Psychopathol Behav Assess (2012) 34:332–342
interpretation bias domains and fear- and anxiety-based
anxiety symptoms while accounting for AS, and vice versa.
Women w ere the focus of this investigation given prior
research indicating that women report higher AS levels
than men (Stewart et al. 1997) and that the relation
between AS and interpretation bias may be stronger among
women than men (Keogh and Cochrane 2002; Keogh et al.
2004).
Basedonpreviousresearch(Weemsetal.2007), we
hypothesized that AS and negative interpretation biases
(both the likelihood of endorsing catastrophic consequences
of ambiguity and/or the strength of belief in negative inter-
pretations) would each predict unique as well as sha red
variance in anxiety sym ptoms. We als o predicted som e
specificity with respect to the associations between particu-
lar domains of negative interpretation biases and certain
types of anxiety symptoms. We hypothesize d that neg ative
interpretations of arousal sensations would be associated
with panic symptoms while negative interpretations of
general events would be associated with generalized
anxiety symptoms. In both cases, we expected these
interpretation biases to be associated with panic or gen-
eralized anxiety symptoms above and beyond their as-
sociationwithAS.
Method
Participants
Participants were screened using the Anxiety Sensitivity
Index (ASI; Peterson and Reiss 1992) completed as part of
a mass screening at three universities. This pre-treatment
screening was conducted for a larger treatment outcome
study of a cognitive-behavioural approach to the treatment
of high AS for college women (see Sabourin et al. 2008 or
Watt and Stewart 2008, for a detailed description of the
approach). To qualify for both the larger study and the
present study, individuals had to score at least one standard
deviation above (high AS) or below (low AS) the mean ASI
screening score for women as reported in past research with
a similar population (i.e., 17.9±8.7; Watt et al. 2006). These
cut-offs optimize statistical power by maximizing the dif-
ference between AS groups while maintaining an adequate
group size. The final sample consisted of 89 female under-
graduates (M age0 18.9 years; range0 17–34 years) selected
into high (n0 45; M ASI0 35.4, SD0 7.9) and low (n0 44; M
ASI0 8.1, SD0 1.9) AS groups.
Measures & Procedure
Participants completed the questionnaires detailed below as
part of a pre-treatment assessment.
Anxiety Sensitivity Index (ASI; Peterson and Reiss 1992)
The ASI measures the amount of fear an individual experi-
ences around anxiety-related body sensations. Participants
indicate the extent to which they agree or disagree with each
item (e.g., “It scares me when my heart beats rapidly”)ona
5-point Likert scale. The ASI has good internal consistency,
test-retest reliability, and construct and criterion validity
(Reiss et al. 2008).
Body Sensations Interpretation Questionnaire (BSIQ; Clark
et al. 1997) Interpretation biases were measured using the
Body Sensations Interpretation Questionnaire (BSIQ). The
BSIQ measures catastrophic i nterpretations of ambiguity
across four doma ins : arou sal se nsa tions ( e.g .,
“You feel
lightheaded and weak. Why?”), general events (e.g., “A
letter marked ‘urgent’ arrives. What is in the letter?”), social
events (e.g., “An old acquaintance passes you in the street
without acknowledging you. Why?” ), and other body sen-
sations unrelated to panic (e.g., “You have a pain in the
small of your back. Why?”). Participants are presen ted with
ambiguous scenarios and rank order three explanations pro-
vided (one negative/catastrophic and two neutral) according
to which they think is the most likely reason for the situation
(‘likelihood ranking’ score). Participants then rate the extent
to which they believe each of these explanations (00 not at
all likely to 80 extremely likely; ‘belief rating’ score). Mean
likelihood ranki ng and belief rating scores are calculated for
each subscale reflecting how likely individuals are to inter-
pret ambiguity catastrophically and the extent to which they
believe these interpretations, respectively (Clark et al.
1997). The BSIQ has good internal consistency (α’s0
0.71–0.86; Keogh et al. 2004; Vancleef and Peters 2008)
and good content and construct validity (Clark et al. 1997).
Depression Anxiety Stress Scales (DASS; Lo vibond and
Lovibond 1993) Current symptoms of panic and general-
ized anxiety were evaluated using the Anxiety and Stress
subscales of the DASS, respec tively. Individuals indicate the
extent to which a particular negative emotional state has
applied to them over the past week using a 4-point Likert
scale. The Anxiety subsca le measures symptoms such as
autonomic arousal and fearfulness (e.g., “I experienced
trembling [e.g., in the hands]”) often associated with panic,
while the Stress subscale measures generalized anxiety
symptoms like tension and irritability (e.g., “I found myself
getting agitated”; Brown et al. 1997). The DASS has good
internal consistency and convergent, discriminant, and
structural validity (Brown et al. 1997; Lovibond and
Lovibond 1995). Individuals with PD score significantly
higher on the Anxiety subscale than those with other anxiety
disorders, while those with GAD score significantly higher
on the Stress subscale than those with other anxiety disor-
ders (except OCD; Brown et al. 1997).
J Psychopathol Behav Assess (2012) 34:332–342 335
Results
Bivariate correlations are reported in Table 1. Because AS
group was a dichotomous variable, point biserial correla-
tions were used to assess its association with other study
variables. While several correlations among study variables
suggest overlap, they would not be considered redundant, as
multicollinearity and redundancy of variables are a concern
only when r>0.90 (Tabachnick and Fidell 2001). Notably,
the small-to-medium correlations (r0 0.31 to 0.42) between
AS group and BSIQ scores lend support to the conceptual-
ization of AS and interpretation biases as conceptually dis-
tinct constructs — the former (AS) being the trait and the
latter (interpretation biases) being one process by which
the trait theoretically confers risk.
A series of hierarchical regression analyses were con-
ducted to examine the incremental contributions of AS and
negative interpretation biases in predicting panic or gener-
alized anxiety symptoms. As all participants were female
and predomi nant ly col lege a ged, ag e and s ex w ere n ot
included in the analyses. All analyses were first completed
using the BSIQ belief rating subscales (i.e., the stre ngth of
belief in negative interpretations of ambiguity) and subse-
quently replicated using the BSIQ likelihood ranking sub-
scales (i.e., the likelihoo d of endor sing catastrophic
consequences of ambiguity).
First, two regression s were conducted to test the unique
contributions of AS and negative interpretation biases to
panic symptoms (see Table 2). In the first, AS was entered
into the initial step of the regression and the four BSIQ
rating subscales (i.e., negative interpretations of ambiguous
arousal-related body sensations, other body sensations,
social events, and general events) were added as predictors
into the second step. The DASS-Anxiety subscale served as
the criterion variable. In the second, the four BSIQ rating
subscales were entered into the first step of the regression
and AS was entered as the predictor in the last step. Again,
the DASS-Anxiety subscale served as the criterion variable.
For the first regression equation, after controlling for AS
group, BSIQ scores together accounted for an additional
16 % of the variance in panic symptoms, ΔR
2
0 0.16 (p<
0.01). For the second regression equation, after controlling
for the BSIQ subscales, AS group accounted for an addi-
tional 10 % of the variance in panic symptoms, ΔR
2
0 0.10
(p<0.01). In the final model (equivalent for both equations),
AS and negative interpretation biases of arousal sensations
specifically (i.e., BSIQ arousal-related body sensations sub-
scale scores), accounted for significant unique variance in
panic symptoms. None of the remaining BSIQ subscales
(i.e., other body sensations, social events, or general events)
emerged as significant predictors in the final equation.
Overall, AS and negative interpretation biases accounted
for 40 % of the variance in panic symptoms, R
2
0 0.40
(p<0.01). Of this explained variance in panic symptoms,
25 % was contributed uniquely by AS group, 40 % was
contributed unique ly by interpretive biases, and the remain-
ing 35 % was contributed by what AS and interpretive bias
hold in common (see Fig. 1).
Next, two regression models were conducted to test the
distinct contributions of AS and negat ive interpretation
biases to generalized anxiety symptoms (see Table 3). In
both cases, the DASS-Stress subscale served as the criterion
variable; the analyses were otherwise performed as above.
In the first regression equation, after accounting for AS
Table 1 Bivariate correlations between AS Group, BSIQ subscales, and DASS subscales
1234567891011
1. AS Group – 0.51** 0.50** 0.42** 0.34** 0.42** 0.39** 0.38** 0.31* 0.39** 0.37**
DASS Subscale Scores
2. DASS-Stress 0.85** 0.54** 0.57** 0.61** 0.40** 0.49** 0.50** 0.56** 0.36**
3. DASS-Anxiety 0.61** 0.55** 0.51** 0.42** 0.53** 0.44** 0.47** 0.35**
BSIQ Likelihood Ranking Scores
4. Arousal sensations 0.72** 0.57** 0.70** 0.57** 0.37** 0.39** 0.39**
5. General events 0.65** 0.67** 0.47** 0.54** 0.47** 0.40**
6. Social events 0.38** 0.47** 0.44** 0.69** 0.36**
7. Other body sensations 0.49** 0.35** 0.25* 0.57**
BSIQ Belief Rating Scores
8. Arousal sensations 0.79** 0.72** 0.83**
9. General events 0.72** 0.78**
10. Social events 0.58**
11. Other body sensations –
AS anxiety sensitivity; BSIQ body sensations interpretation questionnaire; DASS depression anxiety stress scales. N’s range from 79–87 due to
missing cases. **p≤0.001, *p<0.05
336 J Psychopathol Behav Assess (2012) 34:332–342
group, BSIQ scores accounted for an additional 20 % of the
variance in generalized anxiety symptoms, ΔR
2
0 0.20 (p<
0.001). In the second regression equation, after controlling
for the BSIQ subscales, AS group accounted for an addi-
tional 10 % of the variance in generalized anxiety symp-
toms, ΔR
2
0 0.10 (p<0.01). In the final model (equivalent for
both regression equations), AS and negative interpretation
biases of social events and of general events (i.e., BSIQ
socia l a nd general events subscale scores) accounted for
unique variance in generalized anxiety symptoms. Neither
of the remaining BSIQ subscales (i.e., arousal sensations,
other body sensations) emerged as significant predictors in
the final equation. Overall, AS and negative interpretation
biases accounted for 45 % of the variance in generalized
anxiety symptoms, R
2
0 0.45 (p<0.01). Of this explained
variance in generalized anxiety symptoms, 22 % was con-
tributed uniquely by AS group, 45 % was contributed
uniquely by interpretive biases, and the remaining 33 %
was contributed by what AS and interpretive bias hold in
common (see Fig. 2).
All of the above analyses were subsequently re-run using
the BSIQ like lihood rankings subscale scores in place of the
BSIQ belief ratings subsca les in the regression models. In
these analyses, results were replicated, with AS and negative
interpretation biases (of arousal sensations in the case of
DASS-Anxiety and of general and social events in the case
Table 2 Anxiety sensitivity
group and negative interpreta-
tion biases as predictors of
DASS-Anxiety
AS anxiety sensitivity; BSIQ
body sensations interpretation
questionnaire; DASS depression
anxiety stress scales
β tp R
2
ΔR
2
Anxiety Sensitivity & Negative Interpretation Biases Predicting DASS-Anxiety: Equation 1
Step 1 0.24 0.24***
AS Group 0.49 4.88 0.000
Anxiety Sensitivity & Negative Interpretation Biases Predicting DASS-Anxiety:
Equation 2
Step 1 0.30 0.30***
BSIQ Arousal Sensations Belief Ratings 0.50 2.32 0.023
BSIQ General Events Belief Ratings 0.12 0.57 0.571
BSIQ Social Events Belief Ratings 0.17 1.07 0.289
BSIQ Other Body Sensations Belief Ratings −0.26 −1.38 0.173
Anxiety Sensitivity & Negative Interpretation Biases Predicting DASS-Anxiety:
Equations 1 & 2
Step 2 Model 1: 0.40 0.16**
Model 2: 0.40 0.10**
AS Group 0.35 3.42 0.001
BSIQ Arousal Sensations Belief Ratings 0.46 2.28 0.025
BSIQ General Events Belief Ratings 0.16 0.84 0.405
BSIQ Social Events Belief Ratings 0.07 0.43 0.668
BSIQ Other Body Sensations Belief Ratings −0.32 −1.78 0.080
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
Cumulative Variance Contributing to DASS-
Anxiety
Unique Variance: AS
Unique Variance:
Interpretation Biases
Shared Variance: AS
& Interpretation
Biases
Fig. 1 Unique and shared
variance contributed by AS and
interpretation biases in
predicting DASS-Anxiety
scores
J Psychopathol Behav Assess (2012) 34:332–342 337
of DASS-Stress) making independent and significant con-
tributions to both panic and generalized anxiety-related
symptoms. The only exception to this replication was that
the BSIQ general events subscale only emerged as a mar-
ginal predictor (p0 0.068) in the model in which the DASS-
Stress subscale was entered as the criterion variable. Of the
explained variance in panic symptoms, 10 % was contribut-
ed uniquely by AS group, 46 % was contributed uniquely by
interpretive biases, and the remaining 44 % was contributed
by shared variance. Of the explained variance in generalized
anxiety symptoms, 10 % was contributed uniquely by AS
group, 50 % was contrib uted uniquely by interpretive
biases, and the remaining 40 % was contributed by shared
variance. Tables with the results of the abo ve analyses using
BSIQ likelihood ranking subscales in place of BSIQ belief
rating scales are available from the corresponding author
upon request.
Discussion
The present study investigated the unique and shared con-
tributions of AS and negative interpretation biases to vari-
ance in panic and generalized anxiety symptoms. While past
Table 3 Anxiety sensitivity
group and negative interpreta-
tion biases as predictors of
DASS-Stress
AS anxiety sensitivity; BSIQ
body sensations interpretation
questionnaire; DASS depression
anxiety stress scales
β tp R
2
ΔR
2
Anxiety Sensitivity & Negative Interpretation Biases Predicting DASS-Stress: Equation 1
Step 1 0.25 0.25***
AS Group 0.50 5.01 0.000
Anxiety Sensitivity & Negative Interpretation Biases Predicting DASS-Stress:
Equation 2
Step 1 0.35 0.35***
BSIQ Arousal Sensations Belief Ratings 0.09 0.42 0.677
BSIQ General Events Belief Ratings 0.31 1.60 0.115
BSIQ Social Events Belief Ratings 0.42 2.73 0.080
BSIQ Other Body Sensations Belief Ratings −0.23 −1.31 0.200
Anxiety Sensitivity & Negative Interpretation Biases Predicting DASS-Stress:
Equations 1 & 2
Step 2 Model 1: 0.45 0.20***
Model 2: 0.45 0.10**
AS Group 0.34 3.45 0.001
BSIQ Arousal Sensations Belief Ratings 0.05 0.25 0.805
BSIQ General Events Belief Ratings 0.38 2.08 0.041
BSIQ Social Events Belief Ratings 0.31 2.08 0.041
BSIQ Other Body Sensations Belief Ratings −0.31 −1.89 0.063
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
Cumulative Variance Contributing to DASS-
Stress
Unique Variance: AS
Unique Variance:
Interpretation Biases
Shared Variance: AS
& Interpretation
Biases
Fig. 2 Unique and shared
variance contributed by AS and
interpretation biases in
predicting DASS-Stress scores
338 J Psychopathol Behav Assess (2012) 34:332–342
research has investigated the individual contribution of AS
and interpretation biases to anxiety, little research has of yet
examined the incremental contribution of AS and interpre-
tations biases in predicting anxiety symptoms, and no stud-
ies to our knowledge have examined the unique vs. shared
contributions of these correlated but conceptually distinct
constructs to anxiety symptom prediction. The present study
looked to replicate and extend the work of Weems and
colleagues (2007) in several imp ortant ways. First, we in-
vestigated interpretation biases in particular with the unique
goal of understanding the conceptual distinction between
AS and interpretation biases. Second, we exten ded this
research to adult women. Third, we assessed specific
domains of interpretation biases (e.g., of arousal sensations,
other bodily sensations, social events, and general events)
and their unique associations with certa in types of an xiety
symptoms which represented the important distinction be-
tween fear-based (panic) and anxiety-based (generalized
anxiety) symptomatology.
As hypothesized, negative interpretation biases and AS
each made distinct, independent, as well as shared, contri-
butions to explaining variance in both panic and generalized
anxiety symptoms. This is generally in line with past re-
search (Weems et al. 2007) indicating that several factors
thought to predispose an individual to developing anxiety
symptoms, including AS, anxiety control beliefs, and cog-
nitive errors, each dem onstrate unique associations with
anxiety symptoms.
Weems and colleagues (2007) defined cognitive errors as
a range of interpretation biases that could be assessed using
the Children’s Negative Cognitive Errors Q uestionnaire
(CNCEQ; Leitenberg et al. 1986). Rather than identify
interpretation biases based on their conten t domains (e.g.,
of social events, of general events, etc.), the CNCEQ iden-
tifies interpretation biases by type: catastrophizing, over-
generalizing, personalizing, and selective abstraction.
Thus, our work extends Weems et al.’s(2007) findings by
examining the unique associations between different content
domain-based interpretation biases with types of anxiety
symptoms, placing our research in line with the majority
of studies on interpretation biases and anxiety in the litera-
ture. Moreover, we investigated this question among a sam-
ple of young adults, while Weems et al. (2007) tested this
question with a sample of youth, showing that these results
extend across both developmental stages.
The present findings lend credence to the suggestion that
AS and negative interpretati on biases are unique constructs
despite their conceptual and empirical similarities. While
bivariate correlations imply that AS and interpretation
biases are indeed closely related, with the high (vs. low)
AS group showing a stronger tendency to interpret ambigu-
ity catastrophically, their independent contributions to anx-
iety symptoms in the regression analyses suggest that they
are not redundant. Support for this postulation can also be
found in other work in which researchers have investigated
the interrelations among AS, interpretation biases, and psy-
chological problems such as anxiety (Vancleef and Peters
2008) and pain (Keogh et al. 2004).
The empirical distinction between AS and negative inter-
pretation biases may be due in part to the fact that both
constructs are multidime nsional in nature. As illustrated in
the present investigation, an individual’s negative interpre-
tation biases are not necessarily generalized; rather, they
may be specific to amb iguous arousa l-related sensations
(e.g., heart racing), other ambiguous body sensations (e.g.,
pain in leg), ambiguous social events (e.g., being ‘ignored’
by an acquaintance in the street), and/or ambiguous general
events (e.g., receiving a letter in the mail marked ‘urgent’;
Clark et al. 1997). Similarly, AS is composed of several
different components including physical concerns marked
by worries about the physical health consequences of arous-
al sensations, social concerns encompassing fears that others
will notice and judge one’s physical symptoms of anxiety,
and psychological co ncerns focused on worries that arousal
sensations may be indicators of mental catastrophe (Stewart
et al. 1997). While there is some overlap in content areas
between these components of AS and the negative interpre-
tationbiasesassessedontheBSIQ,theremaybesome
intricacies that are captured by one and not the other con-
struct. For instance, negative interpretations of social events
may not be restricted to social situations in which one
worries that others will notice their anxiety sensations but
may also include a broader range of social situations in
which the individual experiences fears of being evaluat-
ed negatively by others (e.g., worries about colleagues,
friends, or strangers finding one irritating or boring).
Unfortunately, due to the dichotomous nature of our
AS variable we could not include AS subscales (Stewart et
al. 1997) in our analyses; this is an important area for future
research.
The present findings suggest some specificity with re-
spect to interpretation biases and their association with
different types of anxiety symptoms. The specificity of these
associations is impressive in light of the strong intercorrela-
tions among study variables. Nega tive interp retati ons of
ambiguous arousal sensations in particular, showed a strong
association with panic-related symptoms, even after ac-
counting for the contribution of AS group to panic symp-
toms. Consistent with Clark’s(1986) theory, a tendency to
interpret arousal sensations catastrophically may contribute
to greater panic symptomatology. Our patt e rn of re sul ts
showing unique associations of interpretive biases for am-
biguous arousal sensations with panic symptoms is consis-
tent with previous findings in the clinical literature showing
that individuals with PD are more likely to negatively inter-
pret arousal sensations and endorse stronger beliefs in these
J Psychopathol Behav Assess (2012) 34:332–342 339
interpretations than non-anxious controls (Austin and
Kiropoulos 2008; Clark et al. 1997).
In contrast, negative interpretations of social events pre-
dicted unique variance in generalized anxiety sy mptoms
beyond that accounted for by AS group. Because AS can
be focused on fears of publicly-observable anxiety symp-
toms (e.g., perspiration) due to beliefs that these symptoms
will lead to embarrassment (Stewart et al. 1997) it is notable
that the negative interpretation bias of social events showed
a distinct association with generalized anxiety. Social con-
cerns may be particularly salient for undergraduates, who
are confronting novel social situations regularly and are in a
period of emerging identity (Stewart and Mandrusiak 2007).
Encountering frequent ambiguous social events which are
interpreted catastrophically might increase generalized anx-
iety symptoms such as worry and tension.
Additionally, strength of belief in negative interpretations
of general events showed a distinct association with gener-
alized anxiety symptoms above and beyond AS group. This
relation was only marginally significant when considering
the likelihood of endorsing ca tastrophic consequences of
ambiguity. The heightened fear and/or distress caused by a
tendency to interpret general events (e.g., a fami ly member
arriving home late) in a negative catastrophic manner may
lead to the elevated irritability, tension, and worry charac-
teristic of generalized anxiety. It is unclear why this relation
emerged significantly only for the belief rating and not the
likelihood ranking scale. It may b e that the belief rating
scale allows for greater flexibility in endorsing interpretation
biases, whereas respondents may feel constrained by the
forced ordering of the like lihood ranking scale. However,
it remains unclear why this discrepancy was not evident on
the other BSIQ scales. Nonetheless, our pattern of results
showing unique associations of interpretive biases for both
ambiguous general and social events with generalized anx-
iety symptoms is consistent with previous findings in the
clinical literature that those with GAD are more likely to
interpret ambiguous general and social events negatively as
compared to non-anxious controls (Clark et al. 1997).
Finally, negative interpretations of body sensations unre-
lated to panic did not show significant associations with
either panic or generalized anxiety symptoms. Negative
interpretations of other body sensations may be more rele-
vant in explaining health anxiety (e.g., Watt et al. 2008). In
accordance, Vancleef and Peters (2008) found that excessive
fear and worry about possible injury or illness predicted
interpretation biases only on the other body sensations sub-
scale of the BSI Q. This is an area for future research.
The present findings should be considered in light of
several limitations. First, this study used self-report meas-
ures of interpretation biases, possibly introducing response
bias (Fazio and Olson 2003). Future work should assess
interpretation biases using implicit association paradigms
(e.g., Lefaivre et al. 2006). Second, this study was cross-
sectional, preventing inferences about causality. In other
words, the study design prevents any inferences as to wheth-
er AS and negative interpretation biases predict (i.e., pre-
cede) the development of panic and/or generalized anxiety
symptoms. Other studies have tackled this question (e.g.,
Hayward et al. 2000); however, future longitudinal work
should continue to elucidate the contribution of specific
maladaptive beliefs and information-processing biases in
predicting the development and/or maintenance of specific
types of anxiety symptoms.
Third, participants were restricted to women, and given
sex differences in AS (Stewart et al. 1997), the findings may
not generalize to men. Given that the association between
interpretation biases and AS may be stronger for women
(Keogh et al. 2004), our focus on women is a justified first
step. Future research could consider whether AS and nega-
tive interpretation biases have a different association with
anxiety symptoms in men versus women. Fourth, labelling
of the DASS Stress and Anxiety subscales as representing
generalized anxiety symptoms and panic symptoms, respec -
tively, should be interpreted with caution. While resear ch
has suggested a strong link between scores on the DASS-
Stress subscale and GAD and scores on the DASS-Anxiety
subscale and PD (Brown et al. 1997), the symptoms on each
of these scales can also be found in multiple anxiety disor-
ders and thus could be considered transdiagno stic in nature.
Furthermore, the use of the DASS-Stress scale as an index
of generalized anxiety symptoms, specifically, is somewhat
limited. While research has suggested that those with GAD
score higher on the DASS-Stress subscale than those with
other anxiety disorders (except OCD; Brown et al. 1997),
those with major depressive disorder also tend to score high
on the DASS-Stress subscale. This raises quest ions as to
whether the DASS-Stress scale might serve as a measure of
more general distress. Future wo rk should select more
symptom specific assessment measures for the types of
anxiety symptoms under investigation.
Finally, given research suggesting that AS may be tax-
onic in nature (Bernstein et al. 2007), there may be a more
psychometrically-sound method for creating high and low
AS groups (i.e., taxon vs. “complement” class). Despite this,
the selection of AS groups in the present study resulted in a
high AS group whose mean ASI score (M0 35.4, SD0 7.9)
was not appreci ably differ ent th an the mea n ASI score
reported for the AS taxon (M0 39.2, SD0 6.2; Bernstein et
al. 2007).
Despite these limitations, the current study supports the
unique and shared contributions of both AS and negative
interpretation biases in predicting anxiety symptoms among
young adult women. Future research should consider the
meaning of the shared contributions of AS and interpreta-
tion biases in contributing to anxiety symptoms. We
340 J Psychopathol Behav Assess (2012) 34:332–342
hypothesize that it might be due to their shared features; for
instance, c atastro phizin g is a ke y chara cteri stic of b oth
individuals with high AS and individuals demonstrating a
negative interpretation bias. This study also suggests that
certain domains of negative interpretation biases may have
unique associations with particular types of anxiety symp-
toms (e.g., fear-based vs. anxiety-bas ed symptoms).
Taken together, if the present findings are also found in
clinical populations, then this may have implications for
treatment and prevention of anxiety disorders. Specifically,
the unique contribution of AS and negative interpretation
biases to anxiety suggests that it may be important to ad-
dress both factors in treatment. Such an approach is current-
ly practiced in a number of existing cognitive behavioural
treatment protocols. The present study strongly supports this
practice of targeting both the disposition that predisposes
one to experience anxiety (i.e., AS) as well as the processes
by which the disposition confers risk (e.g., interpretation
biases) in treatment. Additionally, while the disposition
(i.e., AS) may be general to both pan ic and generalized
anxiety symptoms, in line with current practice, the present
study strongly emphasizes the impor tance of a treatment
focus on targeting specific types of interpretation biases in
accordance with the presenting symptoms. For instance, a
focus on interpretation biases of arousal-re lated body sensa-
tions would be appropriate when treating panic symp toms
while interpretation biases of general and social events
would be more appropriate to target when treating general-
ized anxiety. Future intervention studies should examine
these variables as mechanisms of change to identify the
unique and/or conjoint influence of reducing AS (Watt and
Stewart 2008) and/or modifying neg ative interpretation
biases (Brosan et al. 2011) on anxiety symptom outcomes.
Moreover, because the associations between AS, negative
interpretation biases, and anxiety symptoms have been iden-
tified among a non-clinical sample, the present findings
support the continued use of prevention programs that target
both AS and interpretation biases in preventing clinically
significant anxiety.
References
Allport, G. W. (1937). Personality: A psychological interpretation.
New York: Holt, Rinehart, & Winston.
American Psychiatric Association. (2000). Diagnostic and statistical
manual of mental disorders, Fourth Edition, Text Revision (DSM-
IV-TR). Washington, DC: Author.
Austin, D., & Kiropoulos, L. (2008). An internet-based investigation of
the catastrophic misinterpretation model of panic disorder.
Journal of Anxiety Disorders, 22, 233–242.
Barlow, D. H., Durand, V. M., & Stewart, S. H. (2009). Abnormal
psychology: An integrativ e approach (2nd ed.). Toronto, ON:
Nelson Education Ltd.
Beck, A. T., Emery, G., & Greenberg, R. L. (1985). Anxiety disorders
and phobias: A cognitive perspective. New York: Basic Books.
Bernstein, A., Zvolensky, M. J., Norton, P. J., Schmidt, N. B., Taylor,
S., Forsyth, J. P., et al. (2007). Taxometric and factor analytic
models of anxiety sensitivity: In tegrating approaches to latent
structural research. Psychological Assessment, 19,74–87.
Brosan, L., Hoppitt, L., Shelfer, L., Sillence, A., & Mackintosh, B.
(2011). Cognitive bias modification for attention and interpreta-
tion reduce trait and state anxiety in anxious patients referred to an
out-patient service: Res ults f rom a pilot study. Journal of
Behavior Therapy and Experimental Psychiatry, 42, 258–264.
Brown, T. A., Chorpita, B. F., Korotitsch, W., & Barlow, D. H. (1997).
Psychometric properties of the Depression Anxiety Stress Scales
(DASS) in clinical samples. Behaviour Research and Therapy, 35,
79–89.
Clark, D. M. (1986). A cognitive approach to panic. Behavi our
Research and Therapy, 24, 461–470.
Clark, D. M., Salkovskis, P. M., Öst, L., Breitholtz, E., Koehler, K. A.,
Westling, B. E., et al. (1997). Misinterpretation of body sensations
in panic disorder. Journal of Consulting and Clinical Psychology,
65, 203–213.
Cox, B. J. (1996). The nature and assessment of catastrophic thoughts
in panic disorder. Behaviour Research and Therapy, 34, 363–374.
Cox, B. J., Enns, M. W., Walker, J. R., Kjernisted, K., & Pidlubny, S.
R. (2001). Psychological vulnerabilities in patients with major
depressive vs. panic disorder. Behavior Research and Therapy,
39, 567–573.
Fazio, R. H., & Olson, M. A. (2003). Implicit measures in social
cognition research: Their me aning and use. Annual Review of
Psychology, 54, 297–327.
Hayward, C., Killen, J. D., Kraemer, H. C., & Taylor, C. B. (2000).
Predictors of panic attacks in adolescents. Journal of American
Academy of Child and Adolescent Psychiatry, 39, 207–214.
Kanai, Y., Sasagaway, S., Chen, J., Shimada, H., & Sakano, Y. (2010).
Interpretation bias for ambiguous social behaviour among indi-
viduals with high and low levels of social anx iety. Cognitive
Therapy and Research, 34, 229–240.
Keogh, E., & Cochrane, M. (2002). Anxiety sensitivity, cognitive
biases, and the experience of pain. The Journal of Pain, 3, 320–
329.
Keogh, E., Hamid, R., Hamid, S., & Ellery, D. (2004). Investigating
the effect of anxiety sensitivity, gender, and negative interpreta-
tive bias on the perception of chest pain. Pain, 111, 209–217.
Lefaivre, M., Watt, M. C., Stewart, S. H., & Wright, K. D. (2006).
Implicit associations between anxiety-related symptoms and cat-
astrophic c onsequences in high anxiety sensitive individuals.
Cognition and Emotion, 20, 195–308.
Leitenberg, H., Yost, L. W., & Carroll-Wilson, M. (1986). Negative
cognitive errors in children: Questionnaire development, norma-
tive data, and comparisons between children with and without
self-reported symptoms of depression, low self-esteem, and eval-
uation anxiety. Journal of Consulting and Clinical Psychology,
54, 528–536.
Lovibond, S. H., & Lovibond, P. F. (1993). Manual for the Depression
Anxiety Stress Scales (DASS) (2nd ed.). Sydney, Australia:
Psychology Foundation.
Lovibond, P. F., & Lovibond, S. H. (1995). The structure of negative
emotional states: Comparison of the Depression Anxiety Stress
Scales (DASS) with the Beck Depression and Anxiety
Inventories. Behaviour Research and Therapy, 33 , 335–342.
McNally, R. J. (1999). Theoretical approaches to the fear of anxiety. In
S. Taylor (Ed.), Anxiety sensitivity: Theory, research and treat-
ment of the fear of anxiety (pp. 3–16). Mahwah, NJ: Erlbaum.
Naragon-Gainey, K. (2010). Meta-analysis of the relations of anxiety
sensitivity to the depressive and anxiety disorders. Psychological
Bulletin, 136, 128–150.
J Psychopathol Behav Assess (2012) 34:332–342 341
Noel, M., Taylor, T. L., Quinlan, C. K., & Stewart, S. H. (2012). The
impact of attention style on directed forgetting among high anx-
iety sensitive individuals. Cognitive Therapy and Research.
Norton, P. J., & Asmundson, G. J. G. (2004). Anxiety sensitivity, fear,
and avoidance behaviour in headache pain. Pain, 111, 218–223.
Peterson, R. A., & Reiss, S. (1992). Anxiety sensitivity index manual
(2nd ed.). Worthington, OH: International Diagnostic Services.
Reiss, S. (1991). Expectancy theory of fear, anxiety, and panic.
Clinical Psychology Review, 11, 141–153.
Reiss, S., & McNally, R. J. (1985). The expectancy model of fear. In S.
Reiss & R. R. Bootzin (Eds.), Theoretical issues in behaviour
therapy (pp. 107–122). New York: Academic.
Reiss, S., Peterson, R., Taylor, S., Schmidt, N., & Weems, C. F. (2008).
Anxiety sensitivity index consolidated user manual: ASI, ASI-3,
and CASI (3rd ed.). Worthington, OH: International Diagnostic
Services Publishing Corporation.
Richards, J. C., Austin, D. W., & Alvarenga, M. E. (2001).
Interpretation of ambiguous interoceptive stimuli in panic disor-
der and nonclinical panic. Cognitive Therapy and Research, 25,
235–246.
Rodriguez, B. F., Bruce, S. E., Pagano, M. E., Spencer, M. A., &
Keller, M. B. (2004). Factor structure and stability of the Anxiety
Sensitivity Index in a longitudinal study of anxiety disorder
patients. Behaviour Research and Therapy, 42,79–91.
Sabourin, B. C., Stewart, S. H., Sherry, S. H., Watt, M. C., Wald, J., &
Grant, V. V. (2008). Physical exercise as interoceptive exposure
within a brief cognitive-behavioral treat ment for anxie ty-
sensitive women. Journal of Cognitive Psychotherapy, 22,
303–320.
Schmidt, N. B., Lerew, D. R., & Jackson, R. J. (1997). The role of
anxiety sensitivity in the pathogenesis of panic: Prospective eval-
uation of spontaneous panic attacks during acute stress. Journal of
Abnormal Psychology, 106, 355–364.
Schmidt, N. B., Zvolensky, M. J., & Maner, J. K. (2006). Anxiety
sensitivity: Prospective prediction of panic attacks and Axis I
pathology. Journal of Psychiatric Research, 40, 691–699.
Schmidt, N. B., Mitchell, M. A., & Richey, J. A. (2008). Anxiety
sensitivity as an incremental predictor of later anxiety symp-
toms and syndromes. Comprehensive Psychiatry, 49, 407–
412.
Smits, J. A. J., Berry, A. C., Tart, C. D., & Powers, M. B. (2008). The
efficacy of cognitive-behavioural interventions for reducing anx-
iety sensitivity: A meta-analytic review. Behaviour Research and
Therapy, 46, 1047–1054.
Stewart, D. W., & Mandrusiak, M. (2007). Social phobia in college
students: A developmental perspective. Journal of College
Student Psychotherapy, 22,65–76.
Stewart, S. H., Taylor, S., & Baker, J. M. (1997). Gender differences in
dimensions of anxiety sensitivity. Journal of Anxiety Disorders,
11, 179–200.
Stewart, S. H., Conrod, P. J., Gignac, M. L., & Pihl, R. O. (1998).
Selective processing biases in anxiety-sensitive men and women.
Cognition and Emotion, 12, 105–133.
Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics
(4th ed.). Needham Heights, MA: Allyn and Bacon.
Teachman, B. A. (2005). Information processing and anxiety sensitiv-
ity: Cognitive vulnerability to panic reflected in interpretation and
memory biases. Cognitive Therapy and Research, 29, 479–499.
Vancleef, L. M. G., & Peters, M. L. (2008). Examining content spec-
ificity of negative interpretation biases with the Body Sensations
Interpretation Questionnaire (BSIQ). Journal of Anxiety
Disorders, 22, 401–415.
Viana, A. G., & Rabian, B. (2008). Perceived attachment: Relations to
anxiety sensitivity, worry, and GAD symptoms. Behaviour
Research and Therapy, 46, 737–747.
Watt, M. C., & Stewart, S. H. (2008). Overcoming the fear of fear:
How to reduce anxiety sensitivity. Oakland, CA: New Harbinger
Publications.
Watt, M. C., Stewart, S. H., Lefaivre, M., & Uman, L. S. (2006). A
brief cognitive-behavioral approach to reducing anxiety sensitiv-
ity decreases pain-related anxiety. Cognitive Behaviour Therapy,
35, 248–256.
Watt, M. C., O’Connor, R. M., Stewart, S. H., Moon, E. C., & Terry, L.
(2008). Specificity of childhood learning experiences in relation
to anxiety sensitivity and injury/illness sensitivity: Implications
for health anxiety and pain. Journal of Cognitive Psychotherapy:
An International Quarterly, 22, 128–142.
Weems, C. F., & Watts, S. E. (2005). Cognitive models of childhood
anxiety. In C. M. Velotis (Ed.), Anxiety disorder research (pp.
205–232). Hauppauge, NY: Nova Science Publishers, Inc.
Weems, C. F., Berman, S. L., Silverman, W. K., & Saavedra, L. S.
(2001). Cognitive errors in youth with anxiety disorder s: The
linkages between negative cognitive errors and anxious symp-
toms. Cognitive Therapy and Research, 25, 559–575.
Weems, C. F., Costa, N. M., Watts, S. E., Taylor, L. K., & Cannon, M.
F. (2007). Cognitive errors, anxiet y sensitivity, and anxiety con-
trol beliefs: Their unique and specific associations with childhood
anxiety symptoms. Behaviour Modification, 31, 174–201.
342 J Psychopathol Behav Assess (2012) 34:332–342