Journal of Anxiety Disorders 22 (2008) 429–440

Anxiety sensitivity and affect regulatory strategies: Individual and

interactive risk factors for anxiety-related symptoms
Todd B. Kashdan a,*, Michael J. Zvolensky b, Alison C. McLeish c
a
Department of Psychology, MS 3F5, George Mason University, Fairfax, VA 22030, United States
b
The University of Vermont, United States
c
University of Mississippi Medical Center, United States
Received 2 January 2007; received in revised form 13 March 2007; accepted 15 March 2007

Abstract
Studies have shown that anxiety sensitivity (AS) is a risk factor in the development of pathological anxiety. Recent theoretical
models emphasize the additional importance of how people handle their anxious experiences. The present study examined whether
high AS and being fixated on the control and regulation of unwanted anxious feelings or being unable to properly modulate affect as
needed lead to particularly problematic outcomes. We examined the interactive influence of AS and affect regulatory strategies on
the frequency and intensity of anxiety symptoms. Questionnaires were completed by 248 young adults in the community. Results
showed a general pattern with anxiety symptoms being the most severe when high AS was paired with affect regulatory difficulties.
Of participants high in AS, anxious arousal and worry were heightened in the presence of less acceptance of emotional distress;
anxious arousal, worry, and agoraphobic cognitions were heightened when fewer resources were available to properly modulate
affect; and agoraphobic cognitions were heightened in the presence of high emotion expressiveness. As evidence of construct
specificity, an alternative model with anhedonic depressive symptoms as a main effect and interaction effect (with regulatory
strategies) failed to predict anxiety symptoms. However, anxiety sensitivity and less acceptance of emotional distress were
associated with greater anhedonia. Results are discussed in the context of how and when affect regulatory behavior shifts individuals
from normative anxiety to pathology.
# 2007 Elsevier Ltd. All rights reserved.

Keywords: Anxiety sensitivity; Emotion regulation; Worry; Acceptance; Approach coping; Self-regulation

Researchers have made great strides in conceptua- and its disorders is anxiety sensitivity (fear of anxiety
lizing and measuring emotion vulnerabilities that and its sensations; McNally, 2002). Indeed, extant
contribute to the development of human suffering research suggests that anxiety sensitivity may be
generally and anxiety psychopathology specifically. A usefully conceptualized as a variable risk factor for
vulnerability factor that has received a great deal of anxiety problems (Taylor, 1999; Zvolensky, Schmidt,
scholarly attention in contemporary work on anxiety Bernstein, & Keough, 2006). This cognitive factor
increases the risk for the subsequent development of
anxiety symptoms, unexpected panic attacks, as well as
anxiety psychopathology (e.g., panic disorder; Ehlers,
* Corresponding author. Tel.: +1 703 993 9486;
fax: +1 703 993 1359.
1995; Hayward, Killen, Kraemer, & Taylor, 2000;
E-mail address: tkashdan@gmu.edu (T.B. Kashdan). Maller & Reiss, 1992; Schmidt, Lerew, & Jackson,
URL: http://mason.gmu.edu/tkashdan 1997, 1999; Schmidt, Zvolensky, & Maner, 2006).

0887-6185/$ – see front matter # 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.janxdis.2007.03.011
430 T.B. Kashdan et al. / Journal of Anxiety Disorders 22 (2008) 429–440

Other work suggests that anxiety sensitivity is uniquely thereby gaining a more objective perception of the level
related to escape and avoidance behavior (Stewart, of personal threat, rather than reacting to it in an
Peterson, & Pihl, 1995; Zvolensky et al., 2004; excessively anxiety-relevant manner (e.g., catastrophiz-
Zvolensky & Forsyth, 2002). These data collectively ing). This type of perspective is generally predicted by
indicate that for individuals with higher levels of theoretical models and intervention strategies that
anxiety sensitivity, their outcome expectations can attempt to modify anxiety and other emotional
serve to increase (1) fears about experiencing anxiety disturbances by changing one’s response to such
(anticipatory stage), (2) anxious arousal, and (3) the anxiety states and life events (Hayes, 2002; Hayes
subsequent use of cognitive or behavioral strategies to et al., 2006; Orsillo & Roemer, 2005; Roemer & Orsillo,
escape internal sensations. 2002).
To date, scientific activity on anxiety sensitivity has Though the empirical database is limited in overall
principally been focused on ‘‘main effect’’ types of scope, there is some evidence to support explorations of
questions. Though this type of approach is a useful the interactive effects of affect regulatory variables and
starting point given the developmental nature of the anxiety sensitivity. Using a laboratory approach,
literature, the manner in which anxiety sensitivity may anxiety sensitivity interacted with emotional suppres-
interplay with other processes relevant to anxiety sion to delay affective recovery from an anxiety-
psychopathology is less well documented. Such neglect relevant laboratory stressor (Feldner, Zvolensky,
is unfortunate, as there is a growing recognition that Stickle, Bonn-Miller, & Leen-Feldner, 2006). Such
how individuals regulate emotional experiences, parti- findings support other work suggesting emotional
cularly whether they accept or avoid emotional suppression may be problematic for anxiety sensitive
experiences, is critical in understanding how anxious females (Eifert & Heffner, 2003) and persons with panic
and fearful responding is maintained and exacerbated disorder (Campbell-Sills, Barlow, Brown, & Hofmann,
(Gross, 1999; Kashdan & Steger, 2006; Zvolensky, 2006). In a more recent study, Vujanovic, Zvolensky,
Feldner, Leen-Feldner, & Yartz, 2005). Numerous Bernstein, Feldner, and McLeish (in press) found that
scholars have, in fact, suggested that the role of individuals high in anxiety sensitivity reported fewer
emotional vulnerabilities like anxiety sensitivity may be anxiety symptoms when they also were high, but not
more complex than linear main effect models (Hayes & low, in mindful attention (defined as attention to, and
Feldman, 2004; Zvolensky, Feldner et al., 2005). This awareness of, what is occurring in the present; Brown &
general perspective sits on the backdrop of the Ryan, 2003). Though such data is generally in accord
recognition that affect regulation processes may play with self-regulatory models of anxiety (Kashdan, 2007;
a formative role in the etiology and maintenance of Mennin, 2005; Wells, 2000), empirical data that more
anxiety and its disorders (Mennin, 2005). Specifically, comprehensively document such matters with affect
some scholars have suggested that whether or not regulatory variables are not currently available. Build-
anxiety sensitivity ultimately ‘‘leads’’ to anxiety ing from past theory and research, it would be clinically
psychopathology may depend, at least in part, on important and theoretically useful to examine whether
how people monitor and manage the physical, when people scoring high in anxiety sensitivity are
cognitive, and behavioral manifestations of anxiety unwilling to accept and experience the inevitable and
and the situations that elicit such states (Eifert & natural occurrence of anxiety-related states, they are
Forsyth, 2005; Hayes, Wilson, Gifford, Follette, & more apt to show greater vulnerability to excessive and
Strosahl, 1996). impairing anxiety-related symptoms (Hayes, Luoma,
Interestingly, to the extent an individual can Bond, Masuda, & Lillis, 2006).
‘‘tolerate’’ negative affect states and cognitions (with- As another affect regulatory variable, there is reason
out the need to change or escape them), it is not fully to believe that the willingness to openly express
clear that the presence of high anxiety sensitivity would emotions is an important determinant of whether
be as problematic (Brown, Lejuez, Kahler, Strong, & anxiety sensitivity is related to greater levels of anxiety
Zvolensky, 2005). For example, insofar as people high symptoms. While emotional expression is often an
in anxiety sensitivity emotionally accept aversive effective coping strategy to cope with stress and thrive
anxious states or thoughts, they may be able to forestall in meaningful life domains (e.g., interpersonal relation-
escalation of problematic anxiety experiences. Speci- ships), these benefits may be compromised in people
fically, affect regulatory variables such as emotional with particular vulnerabilities. That is, the costs and
acceptance may, theoretically, permit emotionally benefits of strategically expressing emotions may differ
vulnerable people to attend to the current situation, as a function of people’s dispositional tendencies.
 T.B. Kashdan et al. / Journal of Anxiety Disorders 22 (2008) 429–440 431

Theory and research suggest two competing models on emotional states would demonstrate particularly ele-
how emotional expression can operate in people with vated anxiety-related symptoms. In essence, this
high anxiety sensitivity. In the first model, it can be regulatory strategy is a combination of an unwillingness
hypothesized that high anxiety sensitivity, in conjunc- to accept emotional states and subsequent beliefs that
tion with high emotional expression, leads to emotion little that can be done to change these undesirable states.
disturbances. This hypothesis is driven by the perspec- Third, we tested competing models of whether more or
tive that the open, behavioral expression of emotions less intentional emotional expression in the context of
can be maladaptive in people who are burdened by high anxiety sensitivity would lead to the greatest
frequent and intense emotional distress and negative anxiety-related symptoms. Finally, to test the specificity
feelings and thoughts about these reactions. For of predictions on anxiety-related symptoms, the effects
example, the frequent expression of intense negative of anxiety sensitivity and affect regulatory variables
emotions and thoughts can exhaust social supports such were examined on anhedonic depressive symptoms.
that excessive needs for reassurance and care-taking There has been minimal research on the relevance of
eventually elicit rejection (Kashdan, Volkmann, Breen, anxiety sensitivity and particular affect regulatory
& Han, in press; Kennedy-Moore & Watson, 2001). strategies on negative mood states other than anxiety.
These negative social consequences can serve to Here, we theorized that the mechanisms and interactive
intensity already elevated negative internal reactions. models under study would be conceptually specific to
In an alternative model, it can be hypothesized that high anxiety-related states.
anxiety sensitivity, in conjunction with minimal
emotional expression, can lead to emotional distur- 1. Method
bances. This perspective derives from research suggest-
ing there is a rebound effect for people attempting to 1.1. Participants
strategically inhibit or hide overt signs of internal,
negative emotional states (see Gross, 1999; Hayes et al., The sample consisted of 248 young adults (136
2006 for reviews). That is, people trying to hide or females; Mage = 22.41 years, S.D. = 7.94) recruited
conceal thoughts, feelings, and images show a para- through the general community in Vermont via
doxical increase in the emotional material they are advertising using flyers displayed in a local well-
trying to avoid. More emotionally vulnerable people traveled marketplace, local restaurants, bars, and
may demonstrate greater rebound effects following university-based bulletin boards. The racial composi-
attempts to inhibit their emotional expression. tion of the studied sample reflected that of the local
The global aim of the present investigation was to population (State of Vermont Department of Health,
evaluate the interactive nature between anxiety sensi- 2000): approximately 93.1% of the sample was
tivity and theoretically relevant affect-related regulation Caucasian, 2.4% African-American, 1.2% Hispanic,
variables. People with high anxiety sensitivity were 1.6% Asian American, and 1.6% other. Approximately
expected to vary in their willingness to experience and 5.2% of the sample had at least a 4-year college
flexibly adapt to negative emotions in the service of education, 45.6% had some college education, 46% had
situational demands and personal goals. The degree of a high school degree or the equivalent, 2.8% did not
anxiety-related distress associated with high anxiety have a high school education, and 0.4% did not respond
sensitivity was proposed to be amplified when people to this item. Participants were excluded from the study
over-relied on regulatory strategies that increased the if they displayed limited mental competency or the
difficulty of modulating emotions. Anxiety-related inability to give informed, written consent. Mental
outcomes were operationalized by high anxious arousal, competency was assessed by insuring that the partici-
excessive and uncontrollable worry, and excessive pant was oriented to time and place.
agoraphobic cognitions. Each of these constructs is a
central diagnostic feature of numerous types of anxiety 1.2. Measures
psychopathology. First, it was hypothesized that
individuals high in anxiety sensitivity endorsing less 1.2.1. Anxiety sensitivity index (ASI)
rather than more ability to accept emotional reactions The ASI (Reiss, Peterson, Gursky, & McNally, 1986)
would demonstrate the most anxiety-related symptoms. is a 16-item measure that asks respondents to rate on a
Second, it was hypothesized that individuals high in 5-point Likert scale (0 = very little to 4 = very much) the
anxiety sensitivity who also believed that they could not degree to which they fear negative consequences
effectively access strategies to cope with unwanted stemming from anxiety symptoms. Responses to each
432 T.B. Kashdan et al. / Journal of Anxiety Disorders 22 (2008) 429–440

item are summed to provide a total score from 0 to 64. 1.2.4. Mood and Anxiety Symptom Questionnaire
Previous research indicates that the ASI is made up of (MASQ)
one higher-order factor (ASI Total Score) and three The MASQ is a comprehensive measure of affective
lower-order factors: Physical, Psychological, and Social symptoms with well-established psychometric proper-
Concerns (Rodriguez, Bruce, Pagano, Spencer, & ties (see Watson et al., 1995, for details). Participants
Keller, 2004; Stewart, Taylor, & Baker, 1997; Zinbarg, indicate how much they have experienced each
Barlow, & Brown, 1997). The ASI shows adequate test– symptom from 1 (not at all) to 5 (extremely). The
retest reliability (r = .75 for 2 weeks), criterion validity Anxious Arousal scale (MASQ-AA) is a 17-item scale
(e.g., individuals with agoraphobia score higher than that measures the symptoms of somatic tension and
those with other anxiety disorders and those with no arousal (e.g., ‘‘felt dizzy’’). The Anhedonic Depression
disorder), and is distinct from trait anxiety (Reiss et al., scale (MASQ-AD) is a 22-item scale that measures a
1986). In the present investigation, we utilized the total loss of interest in life (e.g., ‘‘felt nothing was
ASI score, as it represents the global-order anxiety enjoyable’’) and reverse-keyed items measuring posi-
sensitivity factor and therefore takes into consideration tive affect. As in past work (Zvolensky, Kotov,
different types of fears, including fears of panic-related Antipova, & Schmidt, 2005), only the MASQ-AA
somatic, cognitive, and social cues. and MASQ-AD subscales were used in the present
investigation, as opposed to the general distress
1.2.2. Penn State Worry Questionnaire (PSWQ) depressive symptoms and general distress anxiety
The PSWQ is a 16-item measure of pathological symptoms subscales. The MASQ-AA and MASQ-AD
worry that assesses three areas of worry: generality, subscales provide empirically sound and specific
excessiveness, and uncontrollability (Meyer, Miller, composites for ‘‘pure’’ anxiety and ‘‘pure’’ depression
Metzger, & Borkovec, 1990). Items are rated on a 5- symptoms, respectively (Watson et al., 1995). The alpha
point Likert scale from (1) not at all typical to (5) very for the anxious arousal scale in the present sample was
typical. Total scores range from 16 to 80 with higher .91 and the anhedonic depressive subscale was .89.
scores indicating greater levels of worry. The PSWQ is
able to distinguish individuals with generalized anxiety 1.2.5. Difficulties in Emotion Regulation Scale
disorder from other anxiety disorders (e.g., social (DERS)
anxiety disorder and posttraumatic stress disorder; The DERS is a 36-item measure of six dimensions of
Brown, Antony, & Barlow, 1992; Fresco, Mennin, affect regulation (Gratz & Roemer, 2004). Items are rated
Heimberg, & Turk, 2003; Meyer et al., 1990) and has on a 5-point Likert scale from (1) almost never applies to
high internal consistency and good test-rested reliability me to (5) almost always applies to me. The dimensions of
(Meyer et al., 1990). affect regulation assessed include non-acceptance of
emotional responses (‘‘When I’m upset, I feel guilty for
1.2.3. Agoraphobic Cognitions Questionnaire feeling that way’’), limited access to emotion regulation
(ACQ) strategies (‘‘When I’m upset, I believe that I will remain
The ACQ is a 14-item scale measuring the that way for a long time’’), difficulties in engaging in goal
frequency of catastrophic thoughts during the experi- directed behavior (‘‘When I’m upset, I have difficulty
ence of anxiety and fear (Chambless, Caputo, Bright, concentrating’’), impulse control difficulties (‘‘When
& Gallagher, 1984). Items are rated on a 5-point Likert I’m upset, I lose control over my behaviors’’), lack of
scale from (1) thought never occurs to (5) thought emotional awareness (‘‘I care about what I am feeling’’—
always occurs. The ACQ is comprised of two factors: reverse scored), and lack of emotional clarity (‘‘I have
social/behavioral concerns and physical concerns. The difficulty making sense out of my feelings’’). Higher
ACQ has been shown to have high internal consistency scores on these dimensions are indicative of greater
(Cronbach alpha = .87), moderate test–retest reliabil- difficulties. The DERS has high internal consistency
ity (r = .67 for 1 month) and sensitivity to changes due (overall a = .93, a > .80 for each subscale; Gratz &
to treatment (Chambless et al., 1984). The ACQ can Roemer, 2004). Only the non-acceptance and limited
also discriminate clinical from non-clinical groups, access to (effective) emotion regulation strategies
especially individuals with anxiety disorders (Chamb- subscales were used in the present investigation. These
less & Gracely, 1989). As in past research (Zvolensky, subscales contain six items and eight items, respectively.
Bonn-Miller et al., 2006), the ACQ total score was The other scales assess important emotion related
used to index anxiety-related cognitions (a criterion constructs but are not face-valid measures of regulatory
variable). strategy use per se.
 T.B. Kashdan et al. / Journal of Anxiety Disorders 22 (2008) 429–440 433

1.2.6. Emotional Approach Coping Questionnaire 2. Results

(EACQ)
The EACQ is an 8-item questionnaire in which 2.1. Preliminary analyses
respondents indicate, on a 4-point Likert-type scale (1 = I
usually don’t do this at all to 4 = I usually do this a lot), Means, standard deviations, and internal consistency
their tendency to approach their emotions in response to coefficients for all scales are reported in Table 1. All
stressful or difficult situations (Stanton, Kirk, Cameron, scales had acceptable psychometric properties.
& Danoff-Burg, 2000). It measures two factors of
emotional approach coping: emotional processing and 2.1.1. Anxiety sensitivity and affect regulatory
emotional expression. For this study, our interest was in strategies as predictors of anxiety-related symptoms
the emotional expression subscale. Sample items from We conducted a series of hierarchical regression
the emotional expression (EE) subscale include ‘‘I let my models to examine whether non-acceptance of emotional
feelings come out freely’’ and ‘‘I feel free to express my distress (DERS subscale), limited access to effective
emotions.’’ The emotional expression subscale shows regulatory strategies (DERS subscale) or emotional
good internal consistency (a = .82) and test–retest expression (EACQ subscale) moderated the effects of
reliability (r = .72; Stanton et al., 2000). anxiety sensitivity on anxiety outcomes. Anxious arousal
(MASQ subscale), worry (PSWQ-total score), and
1.3. Procedure agoraphobic cognitions (ACQ-total score) served as
dependent measures of anxiety-related symptoms. After
Participants responding to community-based adver- entering main effects, the relevant Anxiety Sensitivi-
tisements for the study were scheduled for an individual ty  Regulatory Strategy interaction was entered. Pre-
appointment by a trained research assistant. At this dictor and outcome variables were centered and
appointment, participants first were presented with a lay significant interaction effects were explored with simple
summary description of the study and then gave verbal slope analyses (see Aiken & West, 1991).
and written consent. Participants then completed the
following self-report measures: Anxiety Sensitivity 2.1.2. Non-acceptance of emotional distress models
Index, Mood and Anxiety Symptom Questionnaire, Both anxiety sensitivity and non-acceptance of
Penn State Worry Questionnaire, Agoraphobic Cogni- emotional distress were related to greater anxious
tions Questionnaire, Emotional Approach Coping arousal, worry, and agoraphobic cognitions (see
Questionnaire, and Difficulties in Emotion Regulation Table 2). We also found support for non-acceptance of
Scale. These measures were presented in the order listed emotional distress as a moderator of the effects of anxiety
above for all participants and required approximately sensitivity on anxious arousal and worry. With the simple
45 min to complete. Upon completion of the study, effects plotted in Fig. 1, for high anxiety sensitive
participants were debriefed regarding the aims of the individuals, greater non-acceptance was related to
study and compensated $25 for their efforts. greater anxious arousal and worry, respectively.

Table 1
Means, standard deviations, and internal consistency coefficients for, and zero-order relations between all variables
1 2 3 4 5 6 7 8 M S.D. a
1. ASI – .52** .68** .06 .62** .56** .51** .50** 20.1 13.1 .93
2. DERS-NA – – .63** .26** .50** .46** .40** .43** 11.5 5.49 .92
3. DERS-S – – – .11 .59** .65** .46** .62** 15.2 6.57 .88
4. EACQ-EE – – – – .02 .09 .06 .34** 10.4 3.46 .90
5. MASQ-AA – – – – – .44** .47** .32** 25.4 8.2 .85
6. PSWQ – – – – – – .42** .51** 31.5 16.3 .95
7. ACQ – – – – – – – .35** 1.5 0.51 .81
8. MASQ-AD – – – – – – – – 54.7 14.1 .90
Note: A double asterisk indicates correlation is significant at .01 level; all p-values were two-tailed. ASI: Anxiety Sensitivity Index (Reiss et al.,
1986); DERS-NA: Difficulties in Emotion Regulation Scale-Non-acceptance (Gratz & Roemer, 2004); DERS-S: Difficulties in Emotion Regulation
Scale-Limited access to emotion regulation strategies (Gratz & Roemer, 2004); EACQ-EE: Emotional Approach Coping Questionnaire-Emotional
Expression subscale (Stanton et al., 2000); PSWQ: Penn State Worry Questionnaire (Meyer et al., 1990); MASQ-AA: Mood and Anxiety Symptom
Questionnaire- Anxious Arousal subscale (Watson et al., 1995); ACQ: Agoraphobic Cognitions Questionnaire (Chambless et al., 1984); MASQ-AD:
Mood and Anxiety Symptom Questionnaire-Anhedonic Depression subscale (Watson et al., 1995).
434 T.B. Kashdan et al. / Journal of Anxiety Disorders 22 (2008) 429–440

Table 2
Hierarchical regression models of anxiety sensitivity and the non-acceptance of emotional distress on anxiety-related symptoms
Step b S.E.b b t DR2 DF
Criterion variable: anxious arousal
1 AS .48 .05 .49 8.49*** .42 89.44***
Non-acceptance .24 .06 .25 4.39***
2 AS  non-acceptance .14 .04 .19 3.62*** .03 13.10***
Criterion variable: worry
1 AS .45 .06 .44 7.30*** .36 66.22***
Non-acceptance .23 .06 .23 3.80***
2 AS  non-acceptance .09 .04 .13 2.17* .01 4.69*
Criterion variable: agoraphobic cognitions
1 AS .40 .07 .40 6.17*** .28 46.27***
Non-acceptance .20 .06 .20 3.09**
2 AS  non-acceptance .07 .04 .09 1.48 .01 2.18
Ns = 242–243; + p < .10; *p < .05; **p < .01; ***p < .001.

2.1.3. Access to regulation strategies models Fig. 2, for high anxiety sensitive individuals, less access
Even after accounting for anxiety sensitivity, limited to strategies was related to greater anxious arousal,
access to effective regulation strategies were related to worry, and agoraphobic cognitions, respectively.
greater anxious arousal, worry, and agoraphobic
cognitions (see Table 3). We also found support for 2.1.4. Emotional expression models
limited access to strategies as a moderator of each Emotional expression was not related to anxiety
anxiety outcome. With the simple effects plotted in outcomes after accounting for anxiety sensitivity

Fig. 1. Anxiety-related symptoms as a function of anxiety sensitivity and non-acceptance of emotional distress. Notes: High and low anxiety
sensitivity, and more and less non-acceptance, was defined as at least +1 and 1 standard deviations from the mean, respectively.
 T.B. Kashdan et al. / Journal of Anxiety Disorders 22 (2008) 429–440 435

Fig. 2. Anxiety-related symptoms as a function of anxiety sensitivity and access to effective emotion regulation strategies. Notes: High and low
anxiety sensitivity, and more and less limited access to strategies, was defined as at least +1 and 1 standard deviations from the mean, respectively.

(see Table 4). We found support for emotional 2.2. Specificity of models to anxiety-related
expression as a moderator of agoraphobic cognitions. symptoms
With the simple effects plotted in Fig. 3, for high
anxiety sensitive individuals, greater emotional expres- To examine the specificity of prior models to
sion was related to greater agoraphobic cognitions anxiety-related symptoms, we examined anhedonic
whereas for less anxiety sensitive individuals, greater depressive symptoms as the criterion. As simultaneous
emotional expression was related to less agoraphobic predictors in separate models, anxiety sensitivity and
cognitions. Thus, whether high emotional expression each affect regulatory strategy (greater non-acceptance,
had costs or benefits was dependent on anxiety less access to strategies, less emotional expression)
sensitivity levels. were related to greater anhedonic symptoms, R2D = .30,
436 T.B. Kashdan et al. / Journal of Anxiety Disorders 22 (2008) 429–440

Fig. 3. Anxiety-related symptoms as a function of anxiety sensitivity and emotion expression. Notes: High and low anxiety sensitivity, and high and
low emotional expression, was defined as at least +1 and 1 standard deviations from the mean, respectively.

Table 3
Hierarchical regression models of anxiety sensitivity and limited access to emotion regulation strategies on anxiety-related symptoms
Step b S.E.b b t DR 2 DF
Criterion variable: anxious arousal
1 AS .40 .07 .42 5.88*** .45 80.74***
Strategies .29 .07 .31 4.30***
2 AS  strategies .09 .04 .13 2.09* .01 4.38*
Criterion variable: worry
1 AS .30 .07 .29 4.14** .47 87.90***
Strategies .47 .07 .46 6.46***
2 AS  strategies .09 .04 .13 2.10* .01 4.39*
Criterion variable: agoraphobic cognitions
1 AS .32 .09 .31 3.68*** .26 34.10***
Strategies .25 .09 .25 2.94**
2 AS  strategies .11 .05 .16 2.17* .02 4.70*
Ns = 198–199; + p < .10; *p < .05; **p < .01; ***p < .001.

Table 4
Hierarchical regression models of anxiety sensitivity and emotional expression on anxiety-related symptoms
Step b S.E.b b t DR 2 DF
Criterion variable: anxious arousal
1 AS .59 .05 .61 11.87*** .37 70.61***
Emotion express .06 .05 .06 1.20
2 AS  emotion express .06 .05 .07 1.29 .00 1.67
Criterion variable: worry
1 AS .57 .05 .56 10.54*** .32 56.82***
Emotion express .05 .05 .05 .99
2 AS  emotion express .01 .05 .01 .15 .00 .02
Criterion variable: agoraphobic cognitions
1 AS .51 .06 .50 9.07*** .26 41.63***
Emotion express .03 .06 .03 .51
2 AS  emotion express .15 .05 .16 2.92** .03 8.50**
Ns = 242–244; + p < .10; *p < .05; **p < .01; ***p < .001.
 T.B. Kashdan et al. / Journal of Anxiety Disorders 22 (2008) 429–440 437

.41, and .35, ts for anxiety sensitivity = 6.51, 2.51, and The related constructs of being unwilling to accept
9.27, and ts for regulatory strategies = 3.58, 6.60, and emotional distress and believing that these states cannot
6.05, respectively (all ps < .05). However, there was be effectively tolerated or regulated appear to be
no support for any interactive effects between anxiety particularly important in understanding vulnerability
sensitivity and regulatory strategies on anhedonic processes for anxiety symptoms. The present data
symptoms ( ps = .68, .95, and .88, respectively). suggest that anxiety sensitivity may be especially
relevant to greater levels of anxiety symptoms among
3. Discussion individuals with reflexive, non-accepting approaches to
internal feelings, thoughts, and physiological arousal.
There is consistent evidence that anxiety sensitivity is Although these interactive models were specific to
related to excessive anxiety-related symptoms. Scholars predicting anxiety-related symptoms and not depres-
have increasingly suggested that whether or not anxiety sive-related symptoms, anxiety sensitivity, the non-
sensitivity ultimately ‘‘leads’’ to anxiety psychopathol- acceptance of emotional distress, and limited access to
ogy may depend, at least in part, on how people monitor regulatory strategies were each positively related to
and manage the physical, cognitive, and behavioral anhedonia. Thus, explanatory specificity was apparent
manifestations of anxiety and the situations that elicit for the interactive, but not main effect level of analysis
them (Eifert & Forsyth, 2005; Hayes et al., 1996). The of the studied variables. Further work is needed on if,
purpose of the present investigation was to concurrently and how, these constructs operate in the development
evaluate the interactive nature between anxiety sensi- and/or maintenance of depressive conditions and other
tivity and theoretically relevant affect regulatory vari- clinical conditions (e.g., sexual dysfunctions).
ables in regard to prototypical anxiety symptoms. The related constructs of non-acceptance of emo-
Anxiety sensitivity was related to greater anxious tional distress and limited access to effective regulatory
arousal and worry in the presence of an unwillingness to strategies were shown to be more important predictors
accept and experience emotional distress. Using a of anxiety and depressive symptoms than emotional
similar but more comprehensive regulatory strategy, expressiveness. These acceptance-based constructs are
anxiety sensitivity was shown to be related to greater defined by reflexive negative evaluations of normative
anxious arousal, worry, and agoraphobic cognitions negative emotional reactions, beliefs that it is unhelpful
when people also believed they could not do little to to be in direct contact with these undesirable emotions,
effectively cope with unwanted and distressing emo- and deliberate attempts to control or monitor these
tional states. Additionally, anxiety sensitivity was states due to concerns about their harmful conse-
related to greater agoraphobic cognitions in the quences. These affect regulatory strategies may
presence of greater emotional expressiveness. Strik- theoretically serve to enhance the negative effects of
ingly, in the absence of these self-regulatory processes, anxiety sensitivity in a variety of ways. Past research
people with high anxiety sensitivity did not show any suggests self-regulation consumes limited resources
elevations in anxiety-related symptoms. Thus, people such as what can be attended to at any given moment
with high anxiety sensitivity do not appear to be a and physical and mental stamina during situational
homogenous group, as associated risks appear to be at tasks and challenges (Muraven & Baumeister, 2000).
least partially dependent on the habitual use of Prolonged, inflexible non-acceptance of emotional
particular affect regulatory strategies. That being said, responses can consume attention, vitality and other
the main effects of anxiety sensitivity and non- resources, leaving fewer resources to cope and thrive in
acceptance of emotional distress were robust predictors everyday life. The over-exertion and depletion of these
of greater anxiety-related symptoms. The magnitude of resources are expected to be particularly pronounced for
our Anxiety Sensitivity  Regulatory Strategy interac- people with emotion vulnerabilities such as anxiety
tions were small (R2D ranged from .01 to .03). However, sensitivity. Specifically, for individuals who fear the
these interaction effects are in the typical range for negative consequences of anxiety states, it may be
psychological studies (e.g., 1–3% of variance particularly problematic to believe that nothing can be
explained; Aiken & West, 1991). More importantly, done with unpleasant and bothersome emotional
an examination of simple effects show that the experiences (limited access to strategies DERS sub-
magnitude of incremental variance do not adequately scale). Although the present research design cannot
account for how these moderational relations elucidate explicate the nature of this explanation or how these
the heterogeneous outcomes of people high in anxiety processes unfold over time, the results provide
sensitivity. preliminary, albeit needed, evidence for interactive
438 T.B. Kashdan et al. / Journal of Anxiety Disorders 22 (2008) 429–440

mechanisms. Future use of prospective paradigms could sectional nature of the present research design, it is not
usefully build from the present study. possible to make causal statements. An important next
There was some mixed support that the benefits of step is the use of prospective methodologies or to
intentionally expressing emotions to cope with stress are experimentally manipulate certain regulatory strategies
compromised in the presence of high anxiety sensitivity. in the laboratory and test singular and interactive effects
For less anxiety sensitive people, greater emotional to theoretically relevant stressors (e.g., bodily sensa-
expression was related to less agoraphobic cognitions but tions). Third, the current findings were based on a
for more anxiety sensitive people this mode of expression relatively homogenous community sample. It will be
was associated with more anxiety symptoms. However, important to examine the current models in select
this model was not evident for anxious arousal or worry. clinical samples and ethnically diverse individuals.
Also, emotional expression was not directly related to Fourth, our outcome measures were limited to ‘‘pure’’
these symptoms (‘‘main effects’’). This lack of suppor- self-report indices of anxious arousal, worry, and
tive evidence may be related to the ‘‘complex nature’’ of agoraphobic cognitions. Thus, generalizations about
emotional expression. In general, expressing emotions our findings should be conducted cautiously. There is
openly, as opposed to trying to conceal and hide them, is merit in examining the current findings with more
adaptive. Being more expressive of emotions allows for broad-band measures of anxious responding and how
greater authenticity or congruence between self-percep- everyday roles and activities are affected. In addition,
tions and internal experiences with behavior (John & extensions of the current study can be derived by
Gross, 2004; Rogers, 1951). However, uninhibited sampling people who differ in anxiety sensitivity and
emotional expression may be problematic when people examining the degree to which regulatory strategies
are burdened by frequent and intense negative feelings, operate successfully in different situations. Transla-
thoughts, images, and sensations (Joiner, 2000). For tional and clinical studies need to operationalize
example, for relationship partners, it may be burdensome situations and account for variance attributable to
to encounter and support this chronic level of emotional person–situation interactions. For example, the degree
distress and can eventually elicit rejection and the erosion to which it is important for a person to be competent in a
of social support (see Kashdan et al., in press for situation should affect the utility of any particular
supportive data; Kennedy-Moore & Watson, 2001). regulatory strategy. The interplay between emotion
Despite inconsistent findings, there was stronger support vulnerabilities and affect regulatory strategies on
for a model in which high but not less emotional distress and impairment is expected to be more complex
expression was associated with the most anxiety than our current model (Zvolensky, Schmidt et al.,
symptoms in the context of high anxiety sensitivity. 2006). However, our current findings are a necessary
Given such data, it may be that tests of more specific types first step in moving beyond linear ‘‘main effect’’ models
of emotional expression are needed; that is, predictions that often omit important sources of variance which can
derived from an explicit understanding of the social lead to misinterpretations of risk and resilience.
context in which emotional expression is completed. Overall, our data suggest that dispositional anxiety
Additionally, whether or not people are high in anxiety sensitivity may be particularly problematic in the
sensitivity, psychological health may be best defined by presence of certain affect regulatory strategies. The
the flexible ability to express, suppress, or amplify the primary questions behind this line of research continue
visibility of emotions as desired or required by situational to be how and why people differ in their generation of
demands. There is merit in using research designs that distress, which people are most susceptible to pain and
can adequately test these more complex models. suffering, and which processes are the most important to
Outside of the above noted issues, the current study target for successful intervention. Emotion and affect
has a number of other interpretative caveats that warrant regulation offer promise in expanding the way in which
consideration. First, the present investigation relied on anxiety-related pathology is understood, how it devel-
self-report instruments and all of the limitations of this ops, and how to treat it.
strategy are relevant. As an example, one of our primary
moderator variables was emotion expression and higher
self-reported emotion expression may be confounded Acknowledgments
with people’s attendance to and willingness to disclose
anxious arousal (our primary outcome). Future work This paper was supported by a National Institute of
should incorporate multimethod approaches to indexing Mental Health grant (MH-73937) awarded to Todd B.
the variables of interest. Second, due to the cross- Kashdan and National Institute on Drug Abuse research
 T.B. Kashdan et al. / Journal of Anxiety Disorders 22 (2008) 429–440 439

grants (1 R01 DA018734-01A1, R03 DA16307-01, and Hayward, C., Killen, J. D., Kraemer, H. C., & Taylor, C. B. (2000).
1 R21 DA016227-01) awarded to Dr. Zvolensky. Predictors of panic attacks in adolescents. Journal of the American
Academy of Child and Adolescent Psychiatry, 39, 207–214.
John, O. P., & Gross, J. J. (2004). Health and unhealthy emotion
References regulation: personality processes, individual differences, and life
span development. Journal of Personality, 72, 1301–1333.
Aiken, L. S., & West, S. G. (1991). Multiple regression: testing and Joiner, T. E. (2000). A test of the hopelessness theory of depression in
youth psychiatric inpatients. Journal of Clinical Child Psychology,
interpreting interactions. London: Sage.
Brown, K. W., & Ryan, R. M. (2003). The benefit of being present: 29, 167–176.
mindfulness and its role in psychological well-being. Journal of Kashdan, T. B. (2007). Social anxiety spectrum and diminished
Personality and Social Psychology, 84, 822–848. positive experiences: theoretical synthesis and meta-analysis.
Clinical Psychology Review, 27, 348–365.
Campbell-Sills, L., Barlow, D. H., Brown, T. A., & Hofmann, S. G.
(2006). Effects of suppression and acceptance on emotional Kashdan, T. B., & Steger, M. F. (2006). Expanding the topography of
responses of individuals with anxiety and mood disorders. Beha- social anxiety: an experience-sampling assessment of positive
emoitons, positive events, and emotion supression. Psychological
viour Research and Therapy, 44, 1251–1263.
Chambless, D., Caputo, G., Bright, P., & Gallagher, R. (1984). Science, 17, 120–128.
Assessment of fear of fear in agoraphobics: the body sensations Kashdan, T. B., Volkmann, J., Breen, W. E., & Han, S. (in press).
questionnaire and the agoraphobic cognitions questionnaire. Jour- Social anxiety and romantic relationships: the costs and benefits of
negative emotion expression are context-dependent. Journal of
nal of Consulting and Clinical Psychology, 52, 1090–1097.
Chambless, D. L., & Gracely, E. J. (1989). Fear of fear and anxiety Anxiety Disorders, doi:10.1016/j.janxdis.2006.08.007.
disorders. Cognitive Therapy and Research, 13, 9–20. Kennedy-Moore, E., & Watson, J. C. (2001). How and when does
Ehlers, A. (1995). A 1-year prospective study of panic attacks: clinical emotional expression help? Review of General Psychology, 5,
187–212.
course and factors associated with maintenance. Journal of Abnor-
mal Psychology, 104, 164–172. Maller, R. G., & Reiss, S. (1992). Anxiety sensitivity in 1984 and
Eifert, G. H., & Forsyth, J. P. (2005). Acceptance and Commitment panic attacks in 1987. Journal of Anxiety Disorders, 6, 241–247.
McNally, R. J. (2002). Anxiety sensitivity and panic disorder. Biolo-
Therapy for anxiety disorders: a practitioner’s treatment guide
gical Psychiatry, 52, 938–946.
using mindfulness, acceptance, and values based behavior change
strategies. Oakland, CA: New Harbinger. Mennin, D. S. (2005). Emotion and the acceptance based approaches
Eifert, G. H., & Heffner, M. (2003). The effects of acceptance versus to the anxiety disordors. In: S. M. Orsillo & L. Roemer (Eds.),
Acceptance and mindfulness-based approaches to anxiety: con-
control contexts on avoidance of panic-related symptoms. Journal
of Behavior Therapy & Experimental Psychiatry, 34, 293–312. ceptualization and treatment. New York: Springer.
Feldner, M. T., Zvolensky, M. J., Stickle, T. R., Bonn-Miller, M. O., & Meyer, T. J., Miller, M. L., Metzger, R. L., & Borkovec, T. D. (1990).
Leen-Feldner, E. W. (2006). Anxiety sensitivity—physical con- Development and validation of the Penn State Worry Question-
naire. Behaviour Research and Therapy, 28, 487–495.
cerns as a moderator of the emotional consequences of emotion
suppression during biological challenge: an experimental test Muraven, M., & Baumeister, R. F. (2000). Self-regulation and deple-
using individual growth curve analysis. Behaviour Research tion of limited resources: does self-control resemble a muscle?
and Therapy, 44, 249–272. Psychological Bulletin, 126, 247–259.
Orsillo, S. M., & Roemer, L. (2005). Acceptance and mindfulness-
Fresco, D. M., Mennin, D. S., Heimberg, R. G., & Turk, C. L. (2003).
Using the Penn State Worry Questionnaire to identify individuals based approaches to anxiety: conceptualization and treatment.
with generalized anxiety disorder: a receiver operating character- New York: Springer.
Reiss, S., Peterson, R. A., Gursky, D. M., & McNally, R. J. (1986).
istic analysis. Journal of Behavior Therapy and Experimental
Psychiatry, 34, 283–291. Anxiety sensitivity, anxiety frequency and the prediction of fear-
Gratz, K. L., & Roemer, L. (2004). Multidimensional assessment of fulness. Behaviour Research and Therapy, 24, 1–8.
emotion regulation and dysregulation: development, factor struc- Rodriguez, B. F., Bruce, S. E., Pagano, M. E., Spencer, M. A., &
Keller, M. B. (2004). Factor structure and stability of the Anxiety
ture, and initial validation of the Difficulties in Emotion Regula-
tion Scale. Journal of Psychopathology and Behavioral Sensitivity Index in a longitudinal study of anxiety disorder
Assessment, 26, 41–54. patients. Behaviour Research and Therapy, 42, 79–91.
Gross, J. J. (1999). Emotion regulation: past, present, future. Cogni- Roemer, L., & Orsillo, S. M. (2002). Expanding our conceptualization
of and treatment for Generalized Anxiety Disorder: integrating
tion & Emotion, 13, 551–573.
Hayes, A. M., & Feldman, G. (2004). Clarifying the construct of mindfulness/acceptance-based approaches with existing cogni-
mindfulness in the context of emotion regulation and the process tive-behavioral models. Clinical Psychology: Science and Prac-
tice, 9, 54–68.
of change in therapy. Clinical Psychology: Science and Practice,
11, 255–262. Rogers, C. R. (1951). Perceptual reorganization in client-centered
Hayes, S. C. (2002). Acceptance, mindfulness, and science. Clinical therapy. In: R. R. Blake & G. V. Ramsey (Eds.), Perception:
Psychology: Science and Practice, 9, 101–106. an approach to personality (pp. 307–327). Oxford, England:
Ronald.
Hayes, S. C., Luoma, J., Bond, F., Masuda, A., & Lillis, J. (2006).
Acceptance and Commitment Therapy: model, processes, and Schmidt, N. B., Lerew, D. R., & Jackson, R. J. (1997). The role of
outcomes. Behaviour Research and Therapy, 44, 1–25. anxiety sensitivity in the pathogenesis of panic: prospective
Hayes, S. C., Wilson, K. G., Gifford, E. V., Follette, V. M., & Strosahl, evaluation of spontaneous panic attacks during acute stress.
Journal of Abnormal Psychology, 106, 355–364.
K. (1996). Experiential avoidance and behavioral disorders: a
functional dimensional approach to diagnosis and treatment. Schmidt, N. B., Lerew, D. R., & Jackson, R. J. (1999). Prospective
Journal of Consulting and Clinical Psychology, 64, 1152–1168. evaluation of anxiety sensitivity in the pathogenesis of panic:
440 T.B. Kashdan et al. / Journal of Anxiety Disorders 22 (2008) 429–440

replication and extension. Journal of Abnormal Psychology, 108, Zinbarg, R. E., Barlow, D. H., & Brown, T. A. (1997). Hierarchical
532–537. structure and general factor saturation of the Anxiety Sensitivity
Schmidt, N. B., Zvolensky, M. J., & Maner, J. K. (2006). Anxiety Index: evidence and implications. Psychological Assessment, 9,
sensitivity: prospective prediction of panic attacks and Axis I 277–284.
pathology. Journal of Psychiatric Research, 40, 691–699. Zvolensky, M. J., Bonn-Miller, M. O., Bernstein, A., McLeish, A. C.,
Stanton, A. L., Kirk, S. B., Cameron, C. L., & Danoff-Burg, S. (2000). Feldner, M. T., & Leen-Feldner, E. W. (2006a). Anxiety sensitivity
Coping through emotional approach: scale construction and vali- interacts with marijuana use in the prediction of anxiety symptoms
dation. Journal of Personality and Social Psychology, 78, 1150– and panic-related catastrophic thinking among daily tobacco
1169. users. Behaviour Research and Therapy, 44, 907–924.
Stewart, S. H., Peterson, J. B., & Pihl, R. O. (1995). Anxiety sensitivity Zvolensky, M. J., Feldner, M. T., Leen-Feldner, E. W., Bonn-Miller,
and self-reported alcohol consumption rates in university women. M. O., McLeish, A. C., & Gregor, K. (2004). Evaluating the role of
Journal of Anxiety Disorders, 9, 283–292. anxiety sensitivity in smoking outcome expectancies among reg-
Stewart, S. H., Taylor, S., & Baker, J. M. (1997). Gender differences in ular smokers. Cognitive Therapy and Research, 28, 473–486.
dimensions of anxiety sensitivity. Journal of Anxiety Disorders, Zvolensky, M. J., Feldner, M. T., Leen-Feldner, E. W., & Yartz, A. R.
11, 179–200. (2005a). Exploring basic processes underlying acceptance and
Taylor, S. (1999). Anxiety sensitivity: theory, research, and treatment mindfulness. In: S. M. Orsillo & L. Roemer (Eds.), Acceptance
of the fear of anxiety. Mahwah, NJ: Lawrence Erlbaum. and mindfulness-based approaches to anxiety: conceptualization
Vujanovic, A. A., Zvolensky, M. J., Bernstein, A., Feldner, M. T., & and treatment (pp. 325–357). New York: Springer.
McLeish, A. C. (in press). A test of the interactive effects of Zvolensky, M. J., & Forsyth, J. P. (2002). Anxiety sensitivity dimen-
anxiety sensitivity and mindful attention in the prediction of sions in the prediction of body vigilance and emotional avoidance.
anxious arousal, agoraphobic cognitions, and body vigilance. Cognitive Therapy and Research, 26, 449–460.
Behaviour Research and Therapy. Zvolensky, M. J., Kotov, R., Antipova, A. V., & Schmidt, N. B.
Watson, D., Weber, K., Assenheimer, J. S., Clark, L. A., Strauss, M. E., (2005b). Diathesis stress model for panic-related distress: a test in
& McCormick, R. A. (1995). Testing a tripartite model. I. Eval- a Russian epidemiological sample. Behaviour Research and
uating the convergent and discriminant validity of anxiety and Therapy, 43, 521–532.
depression symptom scales. Journal of Abnormal Psychology, Zvolensky, M. J., Schmidt, N. B., Bernstein, A., & Keough, M. E.
104, 3–14. (2006b). Risk factor research and prevention programs for anxiety
Wells, A. (2000). Emotional disorders and metacognition: innovative disorders: a translational research framework. Behaviour Research
cognitive therapy. Chichester, UK: Wiley. and Therapy, 44, 1219–1239.