ARTICLE IN PRESS

Behaviour Research and Therapy 46 (2008) 2838

www.elsevier.com/locate/brat

Intolerance of uncertainty and perceived threat

Keith Bredemeier, Howard Berenbaum
Department of Psychology, University of Illinois at Urbana-Champaign, 603 E. Daniel Street, Champaign, IL 61820, USA
Received 13 June 2007; received in revised form 25 September 2007; accepted 25 September 2007

Abstract

The present study examined the relations between dimensions of intolerance of uncertainty (i.e., desire for predictability
and uncertainty paralysis) and perceptions of threat (i.e., perceptions of the probabilities and costs of future undesirable
outcomes) in a sample of 239 college students. Uncertainty paralysis was positively associated with both perceived
probabilities and perceived costs for negative outcomes. Desire for predictability was positively associated with perceived
costs for negative outcomes but was not associated with perceived probabilities for negative outcomes. When probability
estimates for low base rate outcomes were examined separately, desire for predictability was negatively associated with
perceived probabilities. Finally, perceived threat partially mediated the relations between dimensions of intolerance of
uncertainty and worry. The results of the present study suggest mechanisms through which dimensions of intolerance of
uncertainty might lead to excessive worry.
Published by Elsevier Ltd.

Keywords: Intolerance of uncertainty; Perceived threat; Probability estimates; Cost estimates

Introduction

A signicant body of research has shown that intolerance of uncertainty is associated with excessive worry
in both clinical and nonclinical samples (e.g., Dugas, Freeston, & Ladouceur, 1997; Dugas, Gagnon,
Ladouceur, & Freeston, 1998). In fact, research suggests that this association is specic to worry and is not
accounted for by shared variance with other relevant constructs, such as perfectionism and intolerance of
ambiguity (Buhr & Dugas, 2006; Dugas, Gosselin, & Ladouceur, 2001; Dugas, Marchand, & Ladouceur,
2005; Dugas, Schwartz, & Francis, 2004). Furthermore, there is growing evidence that this relationship is
causal. More specically, treatment research has shown that changes in intolerance of uncertainty generally
precede changes in levels of worry (Dugas, Langlois, Rheaume, & Ladouceur, 1998). Also, de Bruin, Rassin,
and Muris (2006) found that individual differences in intolerance of uncertainty predict state levels of worry
following tasks intended to elicit feelings of uncertainty. Finally, experimental research has shown that efforts
to manipulate intolerance of uncertainty lead to increases in worry (Ladouceur, Gosselin, & Dugas, 2000).
Although a great deal of research has documented a link between intolerance of uncertainty and worry,
surprisingly little research has been conducted to elucidate the specic mechanisms through which intolerance

Corresponding author. Tel.: +1 217 333 9624; fax: +1 217 244 5876.
E-mail address: kbredem2@uiuc.edu (K. Bredemeier).

0005-7967/$ - see front matter Published by Elsevier Ltd.

doi:10.1016/j.brat.2007.09.006
 ARTICLE IN PRESS
K. Bredemeier, H. Berenbaum / Behaviour Research and Therapy 46 (2008) 2838 29

of uncertainty leads to excessive worry. Dugas, Buhr, and Ladouceur (2004) have speculated that intolerance
of uncertainty may lead to excessive worry in two different ways: (a) by leading to increased levels of positive
beliefs about worry, negative problem orientation, and cognitive avoidance (all of which are positively
associated with worry) and (b) by increasing perceived threat. In terms of perceived threat, Dugas, Buhr et al.
(2004) have posited that high levels of intolerance for uncertainty may cause individuals to (a) overestimate the
likelihood of negative outcomes (i.e., increase probability estimates) and (b) overestimate the consequences of
negative outcomes (i.e., increase cost estimates). Given the growing evidence that both estimated probabilities
of negative outcomes and estimated costs of negative outcomes are positively associated with levels of worry
(Berenbaum, Thompson, & Pomerantz, 2007; Butler & Mathews, 1983; MacLeod, Williams, & Bekerian,
1991), it is important to examine whether intolerance of uncertainty is in fact associated with perceptions of
threat. After all, Borkovec, Alcaine, and Behar (2004) have argued that it is the perception of threat that
initiates the anxiety process (p. 81).
Several studies (Berenbaum Bredemeier, & Thompson, in press; Buhr & Dugas, 2002; Carleton, Norton, &
Asmundson, 2007) have reported factor analyses of the Intolerance of Uncertainty Scale (IUS; Buhr & Dugas,
2002), the most commonly used measure of intolerance of uncertainty. The results of these studies suggest that
intolerance of uncertainty is a multidimensional construct. More specically, Berenbaum et al. (in press)
described four factors, as did Buhr and Dugas (2002). Carleton, Norton et al. (2007) utilized a combination of
exploratory and conrmatory factor analysis to describe two factors in their proposed abbreviated version of
the IUS. Importantly, two common factors emerged in all three studies. The dimension Berenbaum et al.
(in press) labeled desire for predictability closely aligned with the third factor described by Buhr and Dugas
(2002). In addition, the seven items Berenbaum et al. (in press) used to compute the desire for predictability
subscale score were the same seven items included in the Prospective Anxiety subscale of Carleton et al.s
abbreviated IUS. This dimension involves a strong preference to know what will happen in the future and is
positively associated with conscientiousness and worry. The dimension Berenbaum et al. (in press) labeled
uncertainty paralysis closely aligned with the rst factor described by Buhr and Dugas (2002). The
uncertainty paralysis subscale also overlapped some, though not perfectly, with the Inhibitory Anxiety
subscale of Carleton et al.s abbreviated IUS. This dimension involves a tendency to be frozen into inaction by
uncertainty and is negatively associated with both decisiveness and extraversion. Other subscales derived by
Berenbaum et al. (in press) did not signicantly overlap with factors/dimensions found in previous research.
In summary, mounting evidence supports the idea that the original IUS has at least two stable and replicable
dimensions (i.e., desire for predictability and uncertainty paralysis). Furthermore, Berenbaum et al. (in press)
presented evidence to support the discriminant validity of these two dimensions/subscales. For this reason, we
chose to use the desire for predictability and uncertainty paralysis subscales described by Berenbaum et al.
(in press) in the present study.
The purpose of the present study was to examine the relations between dimensions of intolerance of
uncertainty and perceived threat. In contrast to Dugas and colleagues, we hypothesized a complex relationship
between desire for predictability and probability estimates. Past research has shown that desire for
predictability is a core component of the construct of intolerance for uncertainty and has the strongest
associations with the core components of a similar construct known as need for cognitive closure (Berenbaum
et al., in press). In turn, research suggests that individuals who are high in need for cognitive closure process
and utilize information differently than do individuals who are low in need for cognitive closure. More
specically, Kruglanski and Webster (1996) have argued that people who are high in need for cognitive closure
display motivated closing of the mind. In other words, these individuals exhibit two general tendencies:
urgency (i.e., an inclination to attain closure or certainty as soon as possible) and permanence (i.e., an
inclination to maintain closure or certainty for as long as possible). In fact, there is some evidence to suggest
that similar tendencies are observed in individuals who are prone to worry. For example, Bensi and Giusberti
(2007) have shown that individuals who are high in trait anxiety have an implicit goal of uncertainty reduction.
More specically, they found that individuals who are high in trait anxiety gather less evidence prior to
making a decision. To the extent that individuals with high levels of desire for predictability are likely to
exhibit motivated closing of the mind, they may be motivated to be more certain in their probability
estimates (i.e., have estimates closer to 0% or 100%). More specically, when a particular outcome seems
unlikely to happen, individuals who are high in desire for predictability may report being more certain that it
 ARTICLE IN PRESS
30 K. Bredemeier, H. Berenbaum / Behaviour Research and Therapy 46 (2008) 2838

will not happen. On the other hand, when a particular outcome seems likely to happen, individuals who are
high in desire for predictability may report being more certain that it will happen. In other words, we
hypothesized that desire for predictability would be negatively associated with probability estimates for
outcomes which are unlikely to occur (i.e., low base rate outcomes) and positively associated with probability
estimates for outcomes which are likely to occur (i.e., high base rate outcomes), but not associated with
probability estimates in general.
Uncertainty paralysis, on the other hand, is associated with both depression and anxiety (Berenbaum et al.,
in press), both of which have been found to be positively associated with probability estimates for negative
outcomes (e.g., Miranda & Mennin, 2007). Furthermore, uncertainty paralysis is negatively associated with
decisiveness, suggesting that individuals who are high in uncertainty paralysis may exhibit a tendency to be
frozen into inaction (Berenbaum et al., in press). In turn, this tendency might compromise an individuals
ability to prevent negative outcomes. Thus, we hypothesized that uncertainty paralysis would be positively
associated with probability estimates for negative outcomes.
There were two reasons to suspect that elevated levels of intolerance of uncertainty would be associated with
cost estimates for negative outcomes. First, Dugas, Buhr et al. (2004) have argued that intolerance for
uncertainty may contribute to overestimation of the consequences of negative outcomes. Second, past research
has found that both intolerance of uncertainty and cost estimates are positively associated with worry
(Berenbaum, Thompson, & Bredemeier, 2007; Butler & Mathews, 1983; Dugas et al., 1997). Whereas we
expected desire for predictability and uncertainty paralysis to be differentially associated with probability
estimates, we had no reason to hypothesize that they would be differentially associated with cost estimates.
Finally, we hypothesized that perceptions of threat would mediate the link between intolerance of
uncertainty and worry. However, we did not expect the sole means by which intolerance of uncertainty
contributes to worry to be via perceived threat. The reason for this is that we believe that intolerance of
uncertainty also contributes to worry in other ways, such as by moderating the impact of perceived threat on
worry (Berenbaum, Thompson, & Bredemeier, 2007). It is also possible that intolerance of uncertainty
contributes to worry independently of perceived threat. In fact, intolerance of uncertainty has been dened as
the excessive tendency of an individual to consider it unacceptable that a negative event may occur, however
small the probability of its occurrence (Dugas et al., 2001, p. 552). Dugas, Buhr et al. (2004) have argued that
individuals who are high in intolerance of uncertainty nd uncertain situations inherently distressing, and may
even prefer a certain negative outcome over an uncertain one. In line with this argument, Koerner and Dugas
(in press) found that negative attributions of ambiguous situations mediate the relationship between
intolerance of uncertainty and worry. Consequently, we hypothesized that perceived threat would partially
mediate the relationship between dimensions of intolerance of uncertainty and worry (i.e., it would contribute
to worry both directly and via its impact on perceived threat).
In summary, the present study tested the following hypotheses: (a) desire for predictability and uncertainty
paralysis would both be associated with increased cost estimates for negative outcomes; (b) uncertainty
paralysis would be associated with increased probability estimates for negative outcomes; (c) desire for
predictability would be associated with decreased probability estimates for low base rate negative outcomes
(but not associated with probability estimates for negative outcomes in general); and (d) perceived threat
would partially mediate the relationship between dimensions of intolerance of uncertainty and worry.

Method

Participants and procedure

Participants1 were 239 university students2 (58.8% female) between the ages of 18 and 23 (M 19.0;
SD .9). Of those participants reporting their race/ethnicity, the majority (75.5%) reported being European
American, 9.7% reported being Asian American, 5.1% reported being Latina/o, and 4.6% reported being
1
The participants in the present study were the same individuals who participated in Berenbaum et al. (in press) and Berenbaum,
Thompson, and Bredemeier (2007, Study 2).
2
Due to missing data, the number of participants included in any given analysis ranged from 228 to 239.
 ARTICLE IN PRESS
K. Bredemeier, H. Berenbaum / Behaviour Research and Therapy 46 (2008) 2838 31

African American. Participants were tested in groups of 10 or fewer individuals. As part of their participation
in a research project focusing on perceptions of threat, participants completed the questionnaires described
below. Participants received partial credit toward a research participation requirement in exchange for their
participation.

Measures

Intolerance of uncertainty
Individual differences in intolerance of uncertainty were measured using the English version of the IUS
(Buhr & Dugas, 2002). The IUS is composed of 27 items such as Uncertainty makes me uneasy, anxious, or
stressed, and When its time to act, uncertainty paralyses me. In our analyses, we focused on the desire for
predictability and uncertainty paralysis3 subscales described in Berenbaum et al. (in press). The desire for
predictability subscale contains seven items indicating a strong preference to know what will happen in the
future (e.g., I always want to know what the future has in store for me, A small unforeseen event can spoil
everything, even with the best of planning). The uncertainty paralysis subscale contains seven items
indicating an inclination to be frozen into inaction in the face of uncertainty (e.g., When it is time to act,
uncertain paralyzes me, When I am uncertain I cant go forward). These subscales have demonstrated both
convergent and discriminant validity (Berenbaum et al., in press). In the present sample, as of the desire for
predictability and uncertainty paralysis subscales were .84 and .83, respectively.

Perceived threat
To measure perceived threat, participants were asked to indicate how likely (0 not at all likely;
6 almost certain) it was that 39 negative outcomes (e.g., your health deteriorating, making a mistake at
work) were to occur, and how bad (0 not at all bad; 6 horric) it would be if each of these outcomes
were to occur. This strategy of measuring perceived threat has been used in past research (Butler & Mathews,
1983; MacLeod et al., 1991). The undesirable outcomes included in this instrument included several from each
of the major sources of worry identied in past research (e.g., interpersonal, achievement, nancial, and
health). The average of the 39 how likely scores was used as a measure of participants probability estimates,
and the average of the 39 how bad scores was used as a measure of participants cost estimates. In this
sample, the as for probability and cost estimates were .95 and .94, respectively.
In addition, we obtained an idiographic measure of participants probability and cost estimates.
Participants were rst asked to provide written descriptions of the ve negative outcomes they worried
about the most. After doing so, they were asked to indicate (a) how likely they thought it was for each of the
outcomes to actually occur (1 extremely unlikely; 7 extremely likely) and (b) how upset they would be if
each of the outcomes were to occur (1 not at all upset; 7 extremely upset). The average of the ve how
likely scores was used as a measure of participants probability estimates, and the average of the ve how
bad scores was used as a measure of participants cost estimates. In the present sample, the as for the
probability and cost estimates were .67 and .65, respectively. As expected, the two measures of probability
estimates were signicantly correlated (r .36, po.01), as were the two measures of cost estimates (r .46,
po.01). Consequently, we averaged across the standardized probability scores to compute a composite
probability score, and across the standardized cost scores to compute a composite cost score.4
To test the hypothesis that desire for predictability would be negatively associated with probability
estimates for low base rate negative outcomes, we computed the mean probability estimate for those 10 items
on the nomothetic measure that had low probability estimates when averaging across all participants.5 We
operationally dened low probability as an average probability score less than 2.0. Examples of outcome
3
Results were virtually identical when we used the Inhibitory Anxiety subscale described by Carleton, Norton et al. (2007) rather than
the uncertainty paralysis subscale.
4
Results were virtually identical when scores on the idiographic and nomothetic measures were examined separately.
5
We also planned to select high probability items in order to test the hypothesis that desire for predictability would be positively
associated with probability estimates for high base rate outcomes. However, none of the items on the nomothetic measure had a sample
mean greater than or equal to 4.0.
 ARTICLE IN PRESS
32 K. Bredemeier, H. Berenbaum / Behaviour Research and Therapy 46 (2008) 2838

items, which met this criterion, include Your house/apartment being robbed, Being in a natural disaster
(e.g., tornado, earthquake, etc.), and Becoming seriously ill. a for this new scale was .83 in the sample.

Worry
Worry was measured using both the Penn State Worry Questionnaire (PSWQ; Meyer, Miller, Metzger, &
Borkovec, 1990) and the Worry Domains Questionnaire (WDQ; Tallis, Eysenck, & Mathews, 1992). Past
research has indicated that the PSWQ has excellent testretest reliability and good convergent and
discriminant validity (Meyer et al., 1990; Nitschke, Heller, Imig, McDonald, & Miller, 2001). In the present
sample, a for the PSWQ was .94. Individuals with high PSWQ scores were not rare; 52 participants had scores
greater than 62 (which represents a proposed cutoff for pathological worry in unselected student samples;
Behar, Alcaine, Zuellig, & Borkovec, 2003). Like the PSWQ, the WDQ has been found to have good
psychometric properties (Stober, 1998). We prepared ve new items regarding worry about school, since such
worries were expected to be common in a sample of college students; these ve items were written to resemble
the work domain items in the original WDQ. Because the validity of the socio-political domain items is
considered questionable, it was not included in this study. Thus, the WDQ was composed of 30 items
(e.g., I worry that I will lose close friends). In the present sample, a for the WDQ was .94. As expected, the
PSWQ and WDQ were signicantly correlated (r .58, po.01); consequently we averaged across the
standardized PSWQ and WDQ scores to compute a composite worry score.6

Results

Descriptive statistics for all of these variables, and correlations among them, can be found in Table 1. As
described earlier, we hypothesized that uncertainty paralysis would be positively associated with probability
estimates for negative outcomes. We also hypothesized that desire for predictability would not be associated
with probability estimates for all negative outcomes. To test these hypotheses, we conducted a multiple
regression analysis in which we simultaneously entered desire for predictability and uncertainty paralysis as
predictors, and used composite probability estimates as the criterion variable. The results were consistent with
our hypotheses and are displayed in Table 2. As can be seen in Table 2, higher levels of uncertainty paralysis
were signicantly associated with higher composite probability estimates, whereas desire for predictability was
not associated with composite probability estimates.
To test our hypothesis that desire for predictability would be negatively associated with probability
estimates for low base rate outcomes, we conducted another multiple regression analysis in which we
simultaneously entered desire for predictability and uncertainty paralysis as predictors, this time using
probability estimates for low base rate outcomes as our criterion variable. The results were once again
consistent with our hypotheses and are also displayed in Table 2. As can be seen in Table 2, desire for
predictability was signicantly negatively associated with probability estimates for low base rate outcomes.
Uncertainty paralysis was still associated with probability estimates for low base rate outcomes in the same
direction as they were when composite probability estimates scores were used as the criterion variable.
To test whether dimensions of intolerance for uncertainty would be associated with cost estimates, we
conducted a third multiple regression analysis in which we simultaneously entered desire for predictability and
uncertainty paralysis as predictors, and used composite cost estimates for negative outcomes as the criterion
variable. The results of this analysis are also displayed in Table 2. As can be seen in Table 2, both desire for
predictability and uncertainty paralysis were positively associated with cost estimates.
Having found that desire for predictability was associated with cost estimates, and that uncertainty
paralysis was associated with both probability and cost estimates, we proceeded to examine the degree to
which the relations between dimensions of intolerance of uncertainty and worry were mediated by perceived
threat. Based on the recommendations of Preacher and Hayes (2004, under review), we used bootstrapping
procedures7 (rather than traditional Sobel tests) to test (a) a simple mediation model examining direct and
6
Results were virtually identical when scores on the PSWQ and WDQ were examined separately.
7
Additional information about procedures recommended by Preacher and Hayes, along with SPSS and SAS syntax for using these
procedures, can be found at the following website: http://www.psych.ku.edu/preacher/.
 ARTICLE IN PRESS
K. Bredemeier, H. Berenbaum / Behaviour Research and Therapy 46 (2008) 2838 33

Table 1
Descriptive statistics and zero-order correlations for all variables

IU IU Nomothetic Idiographic Low base rate Idiographic Nomothetic PSWQ WDQ Mean SD
DFP UP probability probability probability cost cost

IU desire for 2.49 .81

predictability
IU uncertainty .53 2.21 .79
paralysis
Nomothetic .10 .32 2.45 .82
probability
Idiographic .12 .19 .35 3.52 1.11
probability
Low base rate -.03 .16 .79 .16 1.67 .80
probability
Nomothetic cost .32 .35 .12 .09 .13 3.84 .76
Idiographic cost .22 .23 .10 .18 .08 .46 5.77 .95
PSWQ .46 .50 .20 .25 .11 .43 .42 3.10 .91
WDQ .44 .61 .44 .41 .19 .38 .39 .58 2.54 .77

Table 2
Dimensions of intolerance of uncertainty predicting probability and cost estimates

Composite probability Low base rate probability Composite cost

ba R2 ba R2 ba R2

Desire for predictability .04 .15 .17

Uncertainty paralysis .34 .24 .23
Model .10 .04 .15
a
Standardized.
 po.05.
po.01.

indirect effects of desire for predictability on worry, treating cost estimates as a mediator8 and uncertainty
paralysis as a covariate and (b) a multiple mediation model examining direct and indirect effects of uncertainty
paralysis on worry, treating both probability and cost estimates as mediators, and desire for predictability as a
covariate. The advantage of this approach is that tests of indirect effects are nonparametric. Also, this
approach allows investigators to test multiple mediation models (i.e., models in which multiple hypothesized
mediators are tested simultaneously). Using these procedures, indirect effects are estimated based upon data
from the sample, and condence intervals are estimated by resampling from the data in an iterative fashion. If
these condence intervals do not contain zero, one can reject the null hypotheses that there is no indirect effect
of the predictor variable(s) on the criterion variable via the mediating variable(s) being tested. These
procedures also simultaneously estimate direct effects of the predictor variable(s) on the criterion variable
(i.e., effects that are not accounted for by indirect effects via hypothesized mediating variables). In cases in
which there is a signicant indirect effect, a nonsignicant direct effect suggests full mediation, whereas a
signicant direct effect suggests partial mediation.
The results from these procedures were consistent with our hypotheses and are displayed in Table 3. As can
be seen in Table 3, zero was not included in any of the bias corrected condence intervals for indirect effects,
suggesting that relations between dimensions of intolerance of uncertainty and worry were mediated by

8
We did not test whether probability estimates mediate the relationship between desire for predictability and worry since desire for
predictability and probability estimates were not positively correlated.
 ARTICLE IN PRESS
34 K. Bredemeier, H. Berenbaum / Behaviour Research and Therapy 46 (2008) 2838

Table 3
Tests of direct and indirect (mediated via perceived threat) associations between intolerance of uncertainty and worry

Mediator Indirect effecta 95% CIb Direct effect

Desire for predictability Cost estimates .06 .01.12 .19**

Uncertainty Probability estimates .10 .05.17 .38**
Paralysis Cost estimates .08 .03.14

*po.05; **po.01.
a
Sample estimate.
b
Bias corrected and accelerated; 5000 resamples.

perceived threat. More specically, the relation between desire for predictability and worry was mediated by
cost estimates, whereas the relation between uncertainty paralysis and worry was mediated by both probability
and cost estimates. However, given that the bs for the remaining contributions (i.e., direct effects) were
statistically signicant in each case, it seems that perceived threat did not fully mediate the relations between
dimensions of intolerance of uncertainty and worry.

Discussion

As hypothesized, the results of this study demonstrate that dimensions of intolerance of uncertainty are
differentially associated with perceived threat. More specically, we found that uncertainty paralysis was
positively associated with both probability and cost estimates. In contrast, although desire for predictability
was positively associated with cost estimates, it was associated with probability estimates in a more complex
way. We found that desire for predictability was not associated with probability estimates for negative
outcomes in general but was negatively associated with probability estimates for low base rate negative
outcomes. The results of the present study provide further evidence for the discriminant validity of IUS
subscales, and thus support the potential utility of using subscale scores rather than total scores when
examining the relation between intolerance of uncertainty and worry.
The results of the present study were consistent with our hypotheses about the relations between dimensions
of intolerance of uncertainty and probability estimates for negative outcomes. More specically, we
hypothesized that desire for predictability would not be associated with probability estimates in general but
would be negatively associated with probability estimates for low base rate negative outcomes. The fact that
our data supported this hypothesis lends support to the notion that desire for predictability can be
conceptualized as a motivational construct. More specically, individuals who are high in desire for
predictability may be motivated to feel more certain about what will happen in the future, and this will be
reected in their probability estimates for future outcomes. When individuals who are high in desire for
predictability are inclined to believe that a particular outcome will not occur but are still not certain, they may
be motivated to judge the outcome as less likely to occur in order to feel more certain about the future.
Though the current study did not address the mechanism(s) through which this occurs, we speculate that this
motivation may manifest in the form of (a) needing less information in order to feel certain about the future
and/or (b) selectively seeking out and/or attending to information which supports ones original inclination
about whether a given outcome will occur. The results of the present study also supported our hypothesis that
uncertainty paralysis would be positively associated with probability estimates for negative outcomes, which
lends support to the notion that individuals who are prone to freeze in the face of uncertainty overestimate the
likelihood of negative outcomes. One possible explanation for this nding is that uncertainty paralysis
prevents individuals from taking action in order to prevent these outcomes from occurring, though again this
hypothesis remains to be tested in future research.
We also found that intolerance of uncertainty was positively associated with cost estimates for negative
outcomes. More specically, we found that both desire for predictability and uncertainty paralysis were
 ARTICLE IN PRESS
K. Bredemeier, H. Berenbaum / Behaviour Research and Therapy 46 (2008) 2838 35

positively associated with cost estimates. One possible explanation for these ndings is that both desire for
predictability and uncertainty paralysis may contribute to catastrophizing (Carleton, Sharpe, & Asmundson,
2007), which in turn has been found to be associated with excessive worry (e.g., Davey & Levy, 1998; Vasey &
Borkovec, 1992). More specically, individuals who are high in intolerance of uncertainty may be more
motivated to think about the likelihood of potential consequences of negative outcomes, even when the
likelihood of that outcome is low. It is also possible that intolerance of uncertainty may be positively
associated with cost estimates for negative outcomes because individuals who are high in intolerance of
uncertainty tend to interpret ambiguous outcomes as negative or threatening (Dugas et al., 2005). In line with
this argument, Rassin and Muris (2005) have found that indecisiveness (which evidence suggests may be highly
related to uncertainty paralysis; e.g., Berenbaum et al., in press) is associated with worst case scenario
reasoning. Finally, the fact that desire for predictability was associated with cost estimates is not surprising,
given that desire for predictability is the dimension of intolerance of uncertainty, which has been found to be
most specic to worry (Berenbaum et al., in press), whereas high cost estimates seem to be specic to worry
(compared with other forms of psychological distress; e.g., Berenbaum, Thompson, & Bredemeier, 2007).
The results of the present study also supported our hypothesis that perceived threat would partially mediate
the links between dimensions of intolerance of uncertainty and worry. This suggests that one important
mechanism through which intolerance of uncertainty might contribute to worry is by leading to greater
perceived threat. However, the fact that perceived threat did not fully mediate suggests that intolerance of
uncertainty contributes to worry in other ways as well. Though there is some research examining other ways in
which intolerance of uncertainty may contribute to worry (e.g., Berenbaum, Thompson, & Bredemeier, 2007;
Koerner & Dugas, in press), it is clear that more research is needed on this topic.
The results of the present study may have important implications for the treatment of anxiety disorders.
Dugas and Robichaud (2007) have introduced a cognitivebehavioral treatment for generalized anxiety
disorder, which specically targets intolerance of uncertainty. More specically, this treatment targets
intolerance of uncertainty by (a) helping clients recognize that uncertainty is unavoidable, (b) helping clients to
recognize how intolerance of uncertainty affects their behavior in their everyday lives, and (c) requiring clients
to expose themselves to uncertainty-inducing situations. Existing research suggests that this treatment package
is efcacious (e.g., Ladouceur, Dugas et al., 2000).
The results of the present study lend support to the notion that targeting intolerance of uncertainty in
therapy may be an effective way to reduce perceived threat (i.e., overestimations of both probability and costs
for future negative outcomes). Furthermore, the results of the present study (as well as previous research
examining the dimensionality of the construct of intolerance of uncertainty) lead us to posit that targeting the
different dimensions of intolerance of uncertainty (i.e., uncertainty paralysis and desire for predictability)
separately will enhance the efcacy of interventions aimed at reducing intolerance of uncertainty. More
specically, we propose that therapists could target high levels of desire for predictability by trying to foster
acceptance of feelings of uncertainty. The relation between acceptance and change has recently become an
important area of study in the development of interventions for psychological disorders (e.g., Hayes,
Jacobsen, Follette, & Dougher, 1994). In fact, Roemer and Orsillo (2002) have introduced an acceptance-
based treatment for generalized anxiety disorder, which uses mindfulness techniques to cultivate acceptance,
and preliminary research suggests that this treatment is efcacious (Roemer & Orsillo, 2007). However, this
treatment targets acceptance of internal experiences in a general sense, and does not specically attempt to
foster acceptance of uncertainty. Thus, in addition to therapists helping clients to recognize that uncertainty is
unavoidable, therapists could help clients gain the necessary skills to be aware of feelings of uncertainty when
they arise (as well as the negative thoughts and/or emotions that they habitually generate) and to cultivate an
intentional attitude of acceptance toward these feelings.
On the other hand, therapists could target high levels of uncertainty paralysis by promoting decisiveness and
action rather than indecision and/or avoidance. More specically, this might involve activities to help clients
to acquire the necessary skills to act under conditions of uncertainty (e.g., problem-solving skills, assertiveness
training), to identify and correct cognitive distortions which may contribute to indecision and/or avoidance,
and to practice utilizing these skills in situations involving uncertainty. Thus, in addition to therapists
encouraging clients to expose themselves to uncertainty-inducing situations, therapists could also help clients
gain the necessary skills to take action in uncertain situations.
 ARTICLE IN PRESS
36 K. Bredemeier, H. Berenbaum / Behaviour Research and Therapy 46 (2008) 2838

The present study has some limitations that should be addressed in future research. First, since this study
did not involve a clinical sample, it is difcult to say whether these results would generalize to individuals with
anxiety disorders. However, given that results of past research suggests that normal and pathological worrying
differ quantitatively rather than qualitatively (Ruscio, Borkovec, & Ruscio, 2001), we suspect that they
probably will. Nevertheless, the ndings of the present study should be replicated using clinical samples.
Second, given the cross-sectional, correlational nature of the present study, it is impossible to draw
conclusions about whether relations between dimensions of intolerance of uncertainty and perceived threat are
causal. In fact, it is very possible that perceived threat may cause individuals to experience high levels of desire
for predictability and/or uncertainty paralysis. It is also possible that the associations we found are driven by
other variables associated with perceived threat that were not included. Thus, longitudinal and/or
experimental research is needed to test whether intolerance of uncertainty causes individuals to overestimate
the probability and/or costs of negative outcomes. Third, since none of the items in our nomothetic measure of
perceived threat could be classied as high base rate negative outcomes, we were unable to test our hypothesis
that desire for predictability would be positively associated with probability estimates for high base rate
outcomes. Thus, additional research is needed to test the hypothesis that individuals who are high in desire for
predictability overestimate the probability of high base rate outcomes. If future research supports this
hypothesis, such ndings could provide interesting insights into the nature of the outcomes about which
individuals who are high in desire for predictability might worry about most. More specically, this might
suggest that individuals who are high in desire for predictability may worry more about high base rate negative
outcomes (since they overestimate the likelihood that they will occur), but may be less inclined to worry about
low base rate negative outcomes (since they underestimate the likelihood that they will occur).
In conclusion, we found that intolerance of uncertainty is associated with perceived threat, and that
perceived threat partially mediates the relationship between intolerance of uncertainty and worry. This nding
provides preliminary support for the hypothesis set forth by Dugas, Buhr et al. (2004) that intolerance of
uncertainty may contribute to worry by increasing perceived threat. Nevertheless, this relation appears to be
more complex than Dugas and colleagues have speculated. More specically, dimensions of intolerance of
uncertainty (e.g., desire for predictability and uncertainty paralysis) were found to be differentially associated
with probability and cost estimates for future negative outcomes. These ndings provide further evidence that
the construct of intolerance of uncertainty is multidimensional. Though it is clear that more research is needed
to elucidate the mechanisms through which intolerance of uncertainty leads to excessive worry, the present
study provides important clues as to how this might occur.

Acknowledgments

We would like to thank Sarah Sass, Edelyn Verona, and Naomi Sadeh for their helpful comments on earlier
drafts of this paper. We would also like to thank Renee Thompson and Barrett Kern for their assistance with
data preparation.

References

Behar, E., Alcaine, O., Zuellig, A. R., & Borkovec, T. D. (2003). Screening for generalized anxiety disorder using the Penn State Worry
Questionnaire: A receiver operating characteristic analysis. Journal of Behavior Therapy and Experimental Psychiatry, 34, 2543.
Bensi, L., & Giusberti, F. (2007). Trait anxiety and reasoning under uncertainty. Personality and Individual Differences, 43, 827838.
Berenbaum, H., Bredemeier, K., & Thompson, R. J. (in press). Intolerance of uncertainty: Exploring its dimensionality and associations
with need for cognitive closure, psychopathology, and personality. Journal of Anxiety Disorders.
Berenbaum, H., Thompson, R. J., & Bredemeier, K. (2007). Perceived threat: Exploring its association with worry and its hypothesized
antecedents. Behaviour Research and Therapy, 45, 24732482.
Berenbaum, H., Thompson, R. J., & Pomerantz, E. (2007). The relation between worrying and concerns: The importance of perceived
probability and cost. Behaviour Research and Therapy, 45, 301311.
Borkovec, T. D., Alcaine, O. M., & Behar, E. (2004). Avoidance theory of worry and generalized anxiety disorder. In R. G. Heimberg,
C. L. Turk, & D. S. Mennin (Eds.), Generalized anxiety disorder: Advances in research and practice (pp. 77108). New York: Guilford.
Buhr, K., & Dugas, M. J. (2002). The intolerance of uncertainty scale: Psychometric properties of the English version. Behaviour Research
and Therapy, 40, 931945.
 ARTICLE IN PRESS
K. Bredemeier, H. Berenbaum / Behaviour Research and Therapy 46 (2008) 2838 37

Buhr, K., & Dugas, M. J. (2006). Investigating the construct validity of intolerance of uncertainty and its unique relationship with worry.
Journal of Anxiety Disorders, 20, 222236.
Butler, G., & Mathews, A. (1983). Cognitive processes in anxiety. Advances in Behavior Research and Therapy, 5, 5162.
Carleton, R. N., Norton, P. J., & Asmundson, G. J. G. (2007). Fearing the unknown: A short version of the Intolerance of Uncertainty
Scale. Journal of Anxiety Disorders, 21, 105117.
Carleton, R. N., Sharpe, & Asmundson, G. J. G. (2007). Anxiety sensitivity and intolerance of uncertainty: Requisites of the fundamental
fears? Behaviour Research and Therapy, 45, 23072316.
Davey, G. C. L., & Levy, S. (1998). Catastrophic worrying: Personal inadequacy and a perseverative iterative style as features of the
catastrophizing process. Journal of Abnormal Psychology, 107, 576586.
De Bruin, G. O., Rassin, E., & Muris, P. (2006). Worrying in the lab: Does intolerance of uncertainty have predictive value? Behaviour
Change, 23, 138147.
Dugas, M. J., Buhr, K., & Ladouceur, R. (2004). The role of intolerance of uncertainty in etiology and maintenance. In R. G. Heimberg,
C. L. Turk, & D. S. Mennin (Eds.), Generalized anxiety disorder: Advances in research and practice (pp. 143163). New York: Guilford.
Dugas, M. J., Freeston, M. H., & Ladouceur, R. (1997). Intolerance of uncertainty and problem orientation in worry. Cognitive Therapy
and Research, 21, 593606.
Dugas, M. J., Gagnon, F., Ladouceur, R., & Freeston, H. (1998). Generalized anxiety disorder: A preliminary test of a conceptual model.
Behaviour Therapy and Research, 36, 215226.
Dugas, M. J., Gosselin, P., & Ladouceur, R. (2001). Intolerance of uncertainty and worry: Investigating narrow specicity in a nonclinical
sample. Cognitive Therapy and Research, 25, 551558.
Dugas, M. J., Hedayati, M., Karavidas, A., Buhr, K., Francis, K., & Phillips, N. A. (2005). Intolerance of uncertainty and information
processing: Evidence of biased recall and interpretations. Cognitive Therapy and Research, 29, 5770.
Dugas, M. J., Langlois, F., Rheaume, J., & Ladouceur, R. (1998, November). Intolerance of uncertainty and worry: Investigating
causality. In J. Stober (Chair), Presented at the Symposium of the annual meeting of the Association for Advancement of Behavior
Therapy on Worry: New findings in applied and clinical research, Washington, DC.
Dugas, M. J., Marchand, A., & Ladouceur, R. (2005). Further validation of a cognitivebehavioral model of generalized anxiety disorder:
Diagnostic and symptom specicity. Journal of Anxiety Disorders, 19, 329343.
Dugas, M. J., & Robichaud, M. (2007). Cognitive behavioral treatment for generalized anxiety disorder: From science to practice. New
York: Routledge.
Dugas, M. J., Schwartz, A., & Francis, K. (2004). Intolerance of uncertainty, worry, and depression. Cognitive Therapy and Research, 28,
835842.
Hayes, S. C., Jacobsen, N. S., Follette, V. M., & Dougher, N. H. (Eds.). (1994). Acceptance and change: Content and context in
psychotherapy. Reno, NV: Context Press.
Koerner, N., & Dugas, M.J. (in press). An investigation of appraisals in individuals vulnerable to excessive worry: The role of intolerance
of uncertainty. Cognitive Therapy and Research.
Kruglanski, A. W., & Webster, D. M. (1996). Motivated closing of the mind: Seizing and freezing. Psychological Review, 103,
263283.
Ladouceur, R., Dugas, M. J., Freeston, M. H., Leger, E., Gagnon, F., & Thibodeau, N. (2000). Efcacy of a cognitivebehavioral
treatment for generalized anxiety disorder: Evaluation in a controlled clinical trial. Journal of Consulting and Clinical Psychology, 68,
957964.
Ladouceur, R., Gosselin, P., & Dugas, M. J. (2000). Experimental manipulation of intolerance of uncertainty: A study of a theoretical
model of worry. Behaviour Research and Therapy, 38, 933941.
MacLeod, A. K., Williams, J. M. G., & Bekerian, D. A. (1991). Worry is reasonable: The role of explanations in pessimism about future
personal events. Journal of Abnormal Psychology, 100, 478486.
Meyer, T. J., Miller, M. L., Metzger, R. L., & Borkovec, T. D. (1990). Development and validation of the Penn State Worry
Questionnaire. Behaviour Research and Therapy, 28, 487495.
Miranda, R., & Mennin, D. S. (2007). Depression, generalized anxiety disorder, and certainty in pessimistic predictions about the future.
Cognitive Therapy and Research, 31, 7182.
Nitschke, J. B., Heller, W., Imig, J. C., McDonald, R. P., & Miller, G. A. (2001). Distinguishing dimensions of anxiety and depression.
Cognitive Therapy and Research, 25, 122.
Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behavior
Research Methods, Instruments, & Computers, 36, 717731.
Preacher, K. J., & Hayes, A. F. (under review). Asymptotic and resampling strategies for assessing and comparing indirect effects in simple
and multiple mediator models.
Rassin, E., & Muris, P. (2005). Indecisiveness and the interpretation of ambiguous situations. Personality and Individual Differences, 39,
12851291.
Roemer, L., & Orsillo, S. M. (2002). Expanding our conceptualization of and treatment for generalized anxiety disorder: Integrating
mindfulness/acceptance-based approaches with existing cognitivebehavioral models. Clinical Psychology: Science and Practice, 9,
5468.
Roemer, L., & Orsillo, S. M. (2007). An open trial of an acceptance-based behavior therapy for generalized anxiety disorder. Behavior
Therapy, 38, 7285.
Ruscio, A. M., Borkovec, T. D., & Ruscio, J. (2001). A taxometric investigation of the latent structure of worry. Journal of Abnormal
Psychology, 110, 413422.
 ARTICLE IN PRESS
38 K. Bredemeier, H. Berenbaum / Behaviour Research and Therapy 46 (2008) 2838

Stober, J. (1998). Worry, problem elaboration, and suppression of imagery: The role of concreteness. Behaviour Research and Therapy, 36,
751756.
Tallis, F., Eysenck, M. W., & Mathews, A. (1992). A questionnaire for the measurement of nonpathological worry. Personality and
Individual Differences, 13, 161168.
Vasey, M. W., & Borkovec, T. D. (1992). A catastrophizing assessment of worrisome thoughts. Cognitive Therapy and Research, 16,
505520.