See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/7133147

Fearing the unknown: A short version of the

Intolerance of Uncertainty Scale

Article in Journal of Anxiety Disorders · February 2007

DOI: 10.1016/j.janxdis.2006.03.014 · Source: PubMed

CITATIONS READS

253 2,111

3 authors:

R. Nicholas Carleton Peter J Norton

University of Regina University of Houston
124 PUBLICATIONS 2,728 CITATIONS 131 PUBLICATIONS 4,644 CITATIONS

SEE PROFILE SEE PROFILE

Gordon J G Asmundson
University of Regina
367 PUBLICATIONS 14,434 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Concussion View project

All content following this page was uploaded by R. Nicholas Carleton on 20 October 2017.

The user has requested enhancement of the downloaded file.

 Journal of Anxiety Disorders 21 (2007) 105–117

Fearing the unknown: A short version of the

Intolerance of Uncertainty Scale
R. Nicholas Carleton a,*, M.A. Peter J. Norton b,
Gordon J.G. Asmundson a
a
Anxiety and Illness Behaviours Laboratory, University of Regina, Regina, Sask., Canada S4S 0A2
b
Department of Psychology, University of Houston, Houston, TX, United States
Received 6 October 2005; received in revised form 9 March 2006; accepted 27 March 2006

Abstract
Intolerance of uncertainty is the tendency of an individual to consider the possibility of a negative event
occurring unacceptable, irrespective of the probability of occurrence. It is a key component of worry, state
anxiety, and related anxiety pathologies. The 27-item Intolerance of Uncertainty Scale (IUS) was developed
to measure intolerance of uncertainty. Previous psychometric analyses of the IUS have suggested both four-
and five-factor models. High inter-item correlations, factor instability, and previous theoretical research
support the development of a reduced measure. The present study used two undergraduate samples and
evaluated a psychometrically stable 12-item two-factor version of the IUS. The reduced measure (IUS-12)
retained exemplary internal consistency, while correlating extremely well with the original IUS and related
measures of anxiety and worry. The IUS-12 also demonstrated a stable two-factor structure, representing
both anxious and avoidance components of intolerance of uncertainty. Directions for future research and
potential applications for assessment are discussed.
# 2006 Elsevier Ltd All rights reserved.

Keywords: Intolerance of uncertainty; IUS; Anxiety; Avoidance; Worry

1. Introduction

Fear is a protective response to a current, identifiable threat (e.g., being attacked by a harmful
animal) and is typically accompanied by a fight or flight response having a strong physiological
reaction (e.g., blood pressure increase, muscle tension, startle response) (Barlow, 2002; McNeil
& Vowles, 2004). In contrast, anxiety is a response to a potential threat that may or may not occur
at some point in the future (e.g., a harmful animal might attack, somewhere, sometime) and is

* Corresponding author. Tel.: +1 306 585 4455; fax: +1 306 585 4784.
E-mail address: carletor@uregina.ca (R.N. Carleton).

0887-6185/$ – see front matter # 2006 Elsevier Ltd All rights reserved.
doi:10.1016/j.janxdis.2006.03.014
106 R.N. Carleton et al. / Journal of Anxiety Disorders 21 (2007) 105–117

typically accompanied by an attenuated version of the physiological reaction to fear

(Asmundson, Norton, & Vlaeyen, 2004; Barlow, 2002; McNeil & Vowles, 2004). Worrying
can be interpreted as ruminating about the possibility of a negative occurrence (e.g., another
attack), thereby maintaining a heightened level of anxiety (Dugas, Gosselin, & Ladoucer, 2001).
These ruminations are validated and supported because the possibility of re-exposure to the
fearful stimulus is not zero (e.g., the harmful animal, or similar ones, continue to exist in the same
environment as the worrier) (Dugas et al., 2001). Therefore, how tolerant a person is of
uncertainty surrounding the possibility of re-exposure will impact how often they worry
(Laugesen, Dugas, & Bukowski, 2003) and therein their anxiety (Greco & Roger, 2001, 2003).
Intolerance of uncertainty has been identified as a discriminating individual difference
characteristic involved in excessive worry (Laugesen et al., 2003), state anxiety (Greco & Roger,
2001), and to have strong positive associations with anxiety pathologies such as Generalized
Anxiety Disorder (GAD), Obsessive Compulsive Disorder (OCD), and Panic Disorder (Dugas,
Gagnon, Ladouceur, & Freeston, 1998; Dugas et al., 2001; Tolin, Abramowitz, Brigidi, & Foa,
2003). It is intolerance of the notion that negative events may occur and there is no definitive way
of predicting such events. Indeed, people who are intolerant of uncertainty are likely to interpret
all ambiguous information as threatening (Heydayati, Dugas, Buhr, & Francis, 2003),
contributing to significant somatic stress reactions (e.g., increased heart rate and blood pressure;
Greco & Roger, 2001, 2003). Furthermore, high intolerance of uncertainty may impair problem-
solving skills, leading to inaction and avoidance of ambiguous situations (Dugas, Freeston, &
Ladoucer, 1997). The Intolerance of Uncertainty Scale (IUS; Freeston, Rhéaume, Letarte, Dugas,
& Ladouceur, 1994) is one of several scales developed to measure intolerance of uncertainty,
which is conceptually linked to ambiguity (Greco & Roger, 2001). Popular alterative measures
are either longer than the IUS (e.g., Uncertainty Response Scale, Greco & Roger, 2001) or suffer
from poor internal reliability and convergent validity (e.g., Intolerance of Ambiguity Scale,
Budner, 1962; Tolerance of Ambiguity Scale, Kirton, 1981).
The original IUS was developed (in French) to assess reactions to ambiguous situations,
uncertainty, and future events (Freeston et al., 1994). The items were derived from a larger list of
statements that had irrelevant or redundant items removed using rational relatedness followed by
discriminant and correlational validations. Assuming the possibility that intolerance of
uncertainty might be a multidimensional construct, an exploratory factor analysis was performed
with an unspecified rotation and Cattell’s (1966) scree test to determine the number of factors to
retain. The final scale consisted of 27 items representing five different factors; however, one item
of the 27 was hyperplanar (i.e., did not load significantly on any factor) and four items cross-
loaded (i.e., loaded onto more than one factor). The five factors were the Unacceptability and
Avoidance of Uncertainty (nine items), Negative Social Evaluation Caused by Uncertainty (nine
items), Uncertainty-Related Frustration (three items), Uncertainty Causes Stress (four items),
and Uncertainty Preventing Action (two items; Freeston et al., 1994).1 Each item was rated on a
Likert scale from 1 (not at all characteristic of me) to 5 (entirely characteristic of me).
Using an undergraduate university sample, Freeston et al. (1994) demonstrated the IUS has
excellent internal consistency (a = .91) and good test–retest reliability (r = .74) at five weeks.
Convergent validity was indicated via correlations with measures of, or associated with, worry.
Specifically, the total IUS score correlated moderately (Westgard, 1999) with French versions of

1
The total number of items exceeds 27 because the reported factors include items that cross-loaded saliently on
multiple factors. This also applies to the summary of the Buhr and Dugas (2002) study that follows.
 R.N. Carleton et al. / Journal of Anxiety Disorders 21 (2007) 105–117 107

the Beck Anxiety Inventory (r = .57; Beck, Epstein, Brown, & Steer, 1988), the Beck Depression
Inventory (r = .52; Beck, Ward, Mendelson, Mock, & Erbaugh, 1961), the Worry Domains
Questionnaire (Tallis, Eysenck, & Mathews, 1992), and the Penn State Worry Questionnaire
(r = .63, Meyer, Miller, Metzger, & Borkovec, 1990). Further, the IUS accounted for additional
variance in the Worry Domains Questionnaire and Penn State Worry Questionnaire over and
above variance shared with anxiety and depression.
While Freeston et al. (1994) found support for a five-factor model, Buhr and Dugas (2002)
found stronger support for a four-factor structure in a back-translated English version of the IUS.2
A principal components analysis was performed with an undergraduate sample using Promax
(oblique) rotation, the Kaiser rule (eigenvalues > 1; Kaiser, 1961), and Cattell’s (1966) scree test.
Results suggested a four-factor solution was best, accounting for 56.8% of the variance and
having excellent internal consistency (a = .94); however, six items continued to cross-load on
multiple factors. The four factors were Uncertainty Leading to Inability to Act (10 items),
Uncertainty Being Stressful and Upsetting (12 items), Unexpected Events are Negative and
Should be Avoided (7 items), and Uncertainty Being Unfair (5 items).
Norton (2005) performed a subsequent assessment of the IUS factor structure, evaluating
reliability and validity across various racial groups in an undergraduate sample. Again, principle
factor analysis with Promax (oblique) rotation was used to delineate the factor structure. Factor
retention was based on the Kaiser rule (Kaiser, 1961), Cattell’s scree (1966) test, and the
interpretability of the factor structure. This resulted in extremely divergent five- and six-factor
solutions having extensive multi-vocal item loadings and poor interpretability. Nevertheless, the
internal consistency within each racial group was very high (a > .93). Based on these findings,
and because of the high alpha coefficients reported with the IUS to date (Buhr & Dugas, 2002;
Freeston et al., 1994), Norton suggested item-removal might improve factor structure of the IUS
without substantively impacting scale reliability.
The earlier psychometric analyses of the IUS are limited in several ways. First, all relied on
eigenvalues >1 and Cattel’s scree test, both of which can be too liberal in their derivations of number
of factors to retain (Tabachnick & Fidell, 2001). Second, despite high item-total correlations and
alpha coefficients, a unitary factor structure has yet to be evaluated. Third, as suggested by Norton
(2005), the continued instability of the factor structure and lack of simple structure, in spite of large
sample sizes, suggests redundancy and unrelatedness within the items. The intent of this study was
two-fold. First, to perform confirmatory factor analyses (CFA) comparing unitary, four-, and five-
factor structures. Second, to evaluate Norton’s (2005) suggestion that a shortened IUS might be a
more stable and efficient measure for evaluating intolerance of uncertainty.

2. Method

2.1. Participants

This study required two independent data sets for evaluation. The initial CFAs were performed
on data collected at the University of Regina. Participants included 254 university students [61
men ages 19–37 (M = 23.3; S.D. = 4.2) and 193 women ages 18–50 (M = 23.0; S.D. = 5.8)] from
the faculties of Psychology and Kinesiology and Health Studies who volunteered following a

2
The back-translation reduced the likelihood of connotative differences that might have otherwise resulted from an
unchecked one-way translation.
108 R.N. Carleton et al. / Journal of Anxiety Disorders 21 (2007) 105–117

brief presentation during one of their classes. All of the participants from the sample had either
some university/technical school training (87.4%), or had completed an undergraduate degree or
technical school diploma (12.6%). Racial demographics for the Regina sample were not
collected; however, based on the population demographics they can be assumed primarily
Caucasian (85%), First Nations (8%), or Asian (5%). Eligible participants received class
credit as compensation for participation. Participants provided responses to a questionnaire
battery including the IUS. Descriptive statistics for the full IUS from the Regina sample,
including the corrected-total item correlations, are presented in Table 1.
The second data set comprising 818 undergraduate participants [227 men ages 17–50
(M = 20.7; S.D. = 3.9), 576 women ages 17–51 (M = 20.6; S.D. = 4.1), 7 people who chose not to
report their sex, 8 people chose not to report their age] was obtained at the University of Houston
as part of a larger study to cross-validate results from the first analysis. The sample represented all
years of undergraduate education (43.4% Freshmen, 21.1% Sophomore, 18.4% Junior, 15.9%
Senior, 0.8% Post-baccalaureate). Racial demographics were collected and participants self

Table 1
Descriptive statistics for the IUS (Regina sample, n = 254)
Item number M S.D. Skewness Kurtosis CITC
1. Uncertainty stops me from having a firm opinion. 2.42 0.98 0.47 0.11 .52
2. Being uncertain means that a person is disorganized. 1.67 0.79 1.05 0.77 .56
3. Uncertainty makes life intolerable. 1.78 0.95 1.21 0.90 .59
4. It is unfair not having any guarantees in life. 1.77 0.85 0.85 0.09 .57
5. My mind can’t be relaxed if I don’t know what will 1.77 0.96 1.12 0.60 .68
happen tomorrow.
6. Uncertainty makes me uneasy, anxious, or stressed. 2.16 1.06 0.69 0.17 .70
7. Unforeseen events upset me greatly. 1.81 0.90 1.20 1.48 .68
8. It frustrates me not having all the information I need. 2.70 1.03 0.28 0.45 .59
9. Uncertainty keeps me from living a full life. 1.56 0.89 1.79 2.92 .73
10. One should always look ahead so as to avoid surprises. 2.22 1.01 0.56 0.15 .65
11. A small unforeseen event can spoil everything, even with 1.89 0.97 0.97 0.45 .61
the best of planning.
12. When it’s time to act, uncertainty paralyses me. 1.50 0.74 1.58 2.68 .70
13. Being uncertain means that I am not first rate. 1.48 0.78 1.75 2.93 .73
14. When I am uncertain, I can’t go forward. 1.48 0.76 1.69 2.73 .70
15. When I am uncertain I can’t function very well. 1.61 0.82 1.37 1.58 .74
16. Unlike me, others always seem to know where they are 1.90 1.06 0.95 0.08 .51
going with their lives.
17. Uncertainty makes me vulnerable, unhappy, or sad. 1.68 0.89 1.32 1.51 .76
18. I always want to know what the future has in store for me. 2.34 1.06 0.66 0.07 .65
19. I can’t stand being taken by surprise. 1.76 0.94 1.20 0.93 .70
20. The smallest doubt can stop me from acting. 1.77 0.96 1.19 0.70 .67
21. I should be able to organize everything in advance. 2.24 1.09 0.65 0.28 .66
22. Being uncertain means that I lack confidence. 1.81 0.99 1.22 0.97 .72
23. I think it is unfair that other people seem sure 1.61 0.88 1.34 0.99 .51
about their future.
24. Uncertainty keeps me from sleeping soundly. 1.76 0.97 1.20 0.75 .69
25. I must get away from all uncertain situations. 1.42 0.75 1.86 2.85 .72
26. The ambiguities in life stress me. 1.86 0.95 1.05 0.48 .73
27. I can’t stand being undecided about my future 2.31 1.19 0.52 0.79 .63
Total 50.28 17.22 1.10 1.05 1.00
CITC: corrected item-total correlations.
 R.N. Carleton et al. / Journal of Anxiety Disorders 21 (2007) 105–117 109

reported as being Caucasian (27%), Hispanic/Latino (20%), African-American (17%), Southeast

Asians (17%), East Indian (8%), or other (11%). Norton (2005) reported on IUS internal validity,
factor structure (using exploratory factor analysis), and convergent validity across racial groups
using a subset of this sample (n = 540). The IUS was found to be invariant across racial groups
and the analyses performed here with this larger sample are novel relative to those in the subset.
Participants from Houston provided responses to a questionnaire battery including the IUS and
established measures of worry and anxiety.

2.2. Measures

The Beck Anxiety Inventory (BAI; Beck et al., 1961, 1988). The BAI is a 21-item self-report
measure designed as a general measure of anxiety symptom severity (Beck et al., 1988). The
psychometric properties of BAI for non-clinical populations have been well established (see
Creamer, Foran, & Bell, 1995).
The Beck Depression Inventory-II (BDI; Beck, Steer, & Brown, 1996). The BDI-II is a
21-item measure intended to assess the somatic, cognitive and affective domains of depression.
The BDI-II has well-established reliability and validity (Dozois, Dobson, & Ahnberg, 1998).
The Intolerance of Uncertainty Scale (Buhr & Dugas, 2002; Freeston et al., 1994). This 27-
item scale assesses reactions to uncertainty, ambiguous situations, and the future. Items are
scored on a Likert scale ranging from 1 (not at all characteristic of me) to 5 (entirely characteristic
of me), yielding possible scores from 27 to 135. Despite the reported multifactor structures, the
IUS is most commonly summed as a total scale score (Antony, Orsillo, & Roemer, 2001).
The Generalized Anxiety Disorder Questionnaire-IV (GADQ-IV; Newman et al., 2002). The
GADQ-IVis a nine-item self-report questionnaire intended to assess criteria for generalized anxiety
disorder. Specifically, the GADQ-IV assesses the presence of abnormally high, distressing,
uncontrollable worry and associated symptoms. Sensitivity, specificity, reliability, and both
convergent and discriminant validity of the GADQ-IV has been established (Newman et al., 2002).
The Penn State Worry Questionnaire (PSWQ; Meyer et al., 1990). The PSWQ is a 16-item
self-report measure assessing the frequency and intensity of worry, which is the principal
characteristic of generalized anxiety disorder. The strong, stable psychometric properties of
PSWQ have been successfully demonstrated in non-clinical samples (Meyer et al., 1990).

3. Results

3.1. Original IUS summary scores

Using the data from the Regina sample, scores on the full IUS ranged from 27 to 119
(M = 50.28, S.D. = 17.22). No significant differences based on sex were observed for the total
score (women: M = 49.99, S.D. = 17.16; men, M = 51.19, S.D. = 17.51), t(252) = 0.474,
P = .636, r2 < .01. There is as of yet no reason to assume substantive sex-based differences
on the IUS; therefore, sex was not considered as a covariate. In line with Norton’s (2005) findings
and supporting item reduction, the internal consistency for the full IUS was very high (a = .96).

3.2. Confirmatory factor analysis

Confirmatory factor analyses were conducted to assess the degree to which unitary, four-, and
five-factor structures of the IUS fit the Regina data set. These analyses were performed using
110 R.N. Carleton et al. / Journal of Anxiety Disorders 21 (2007) 105–117

LISREL 8.3 (Jöreskog & Sörbom, 1999), and followed the procedures outlined by Jöreskog and
Sörbom (1996). The raw data were used as input, along with a maximum likelihood estimation
procedure with the covariance matrix. Item loadings for the four- and five-factor structures were
based on the recommendations from Buhr and Dugas (2002) and Freeston et al. (1994),
respectively. Following recommendations by Hu and Bentler (1999), model fit was determined
using several indices: (1) x2/df ratio (values should be <2.0), (2) Comparative Fit Index (CFI;
values should be >0.95), (3) Root Mean Square Error of Approximation (RMSEA; values should
be <0.06), (4) the Standardised Root Mean Square Residual (SRMR; values should be <0.08)
(see also Browne & Cudeck, 1993; Marsh, Balla, & McDonald, 1988), and (5) Expected Cross-
Validation Index (ECVI; lower values indicate increasingly better fit; Browne & Cudeck, 1989,
1993). Because x2 statistics are inflated in larger samples, emphasis is placed on the latter fit
indices (Hu & Bentler, 1999).
The initial CFA results of the tested structures (i.e., unitary, four, and five factors) with the
Regina sample did not indicate adequate fit with the data. Further, comparisons of the three
structures were inconclusive because of similarity between the fit indices of the solutions (see
Table 2). However, the unitary solution presented the worst fit, suggesting it to be least likely. All
factor loadings, residual terms, and latent factor variances were statistically significant (P < .01).
High numbers of cross-loading items in the four- and five-factor solutions (six in the four-factor
solutions, nine in the five-factor solutions), along with a high Cronbach’s a (a = .96), supported
Buhr and Dugas’ (2002) suggestion that five factors may be inappropriate and Norton’s (2005)
suggestion that a more efficient version that still meets minimum criteria for reliability is
possible.
Interestingly, both the Buhr and Dugas (2002; four-factor) model and the Freeston et al. (1994;
five-factor) model had one factor for which the loaded items were shared between all but one of
the factors in the other model. Each of these independent factors correlated very highly with other
factors in its corresponding parent model. Unacceptability and avoidance of uncertainty
correlated highly (between .82 and .93) with all but one of the other factors in the five-factor (i.e.,
Freeston et al., 1994) model (i.e., uncertainty related to frustration; r = .63). Comparably, in the
four-factor (i.e., Buhr & Dugas, 2002) model, the uncertainty leading to inability to act factor
correlated highly with all of the other factors (between .81 and .92). While the correlations lent

Table 2
Confirmatory factor analyses fit indices for of the IUS versions
x2/df CFI SRMR RMSEA RMSEA; CI ECVI ECVI; CI
a
27-item , 1-factor 2.97 0.84 0.06 0.10 0.09; 0.11 4.84 4.45; 5.26
27-itema, 4-factor 2.59 0.88 0.06 0.08 0.08; 0.09 3.96 3.62; 4.32
27-itema, 5-factor 2.43 0.89 0.07 0.08 0.07; 0.09 3.67 3.36; 4.02
17-itema, 2-factor 2.22 0.94 0.05 0.07 0.06; 0.08 1.31 1.14; 1.51
12-itema, 2-factor 1.90 0.97 0.04 0.06 0.04; 0.08 0.58 0.49; 0.71
12-itema, 1-factor 3.84 0.90 0.06 0.12 0.11; 0.14 1.22 1.03; 1.43
12-itemb, 2-factor 4.34 0.96 0.04 0.07 0.06; 0.07 0.35 0.30; 0.42
12-itemb, 1-factor 7.47 0.92 0.05 0.10 0.09; 0.11 0.64 0.56; 0.73
CFI, Comparative Fit Index; RMSEA, Root Mean Square Error of Approximation; SRMR, Standardized Root Mean
Square Residual; ECVI, Expected Cross-Validation Index. Higher CFI values indicate better fit, whereas lower values on
all other indices indicate better fit. RMSEA CI: 90% confidence interval for RMSEA (low; high). ECVI CI: 90%
confidence interval for ECVI (low; high).
a
Regina sample, n = 254.
b
Houston sample, n = 818.
 R.N. Carleton et al. / Journal of Anxiety Disorders 21 (2007) 105–117 111

support to the idea of item redundancy, the item-independence of the two factors (i.e.,
unacceptability and avoidance of uncertainty; uncertainty leading to inability to act) suggested
they contained items that might best be used in the reduced measure. We hypothesized that these
item-groupings (the two factors, one from each model, which do not share any items) would
remain stable if assessed concurrently. A CFA was performed using the two independent factors
(one from each of the four- and five-factor models). This resulted in a superior model fit using
only 17 of the original 27 items (Table 2). A subsequent correlation analysis revealed a significant
and high correlation (r = .98) between the total scores for the 17- and 27-item models.
In an effort to assess further item-redundancy and relatedness, a reliability analysis was
performed on the new, 17-item compilation. The analysis revealed that Cronbach’s a remained
high (a = .94) and most inter-item correlations were low to moderate (between .27 and .65).
Nevertheless, the correlations between items 14 and 15 (r = .74; original IUS item numbers,
Table 1), and items 9 and 13 (r = .68; original IUS item numbers, Table 1), were thought high
enough to warrant a more detailed evaluation. Given the semantic similarity between the two
questions in each pair, it was thought only one of each would be sufficient in a shortened IUS.
Therefore, item 14 (When I am uncertain, I can’t go forward) was removed because item 15
(When I am uncertain I can’t function very well) had a higher factor loading (.60 vs. .66,
respectively) and better face validity. Similarly, item 13 (Being uncertain means I am not first
rate) was dropped rather than item 9 (Uncertainty keeps me from living a full life), because item
13 had a lower factor loading (.62) than item 9 (.70), which also had better face validity.
Following independent review by each of the three authors of this study, three remaining items
stood out as not being rationally related to their parent factors. Item 22 (Being uncertain means
that I lack confidence) appears to have more to do with self-esteem than with uncertainty leading
to an inability to act. Also, items 27 (I can’t stand being undecided about my future) and 1
(Uncertainty stops me from having a firm opinion) seemed to have more to do with indecision.
Therefore, all three items were removed from further analysis. The internal consistency for the
remaining 12 items was excellent (a = .91), as it was for each of the two independent factors
originally called unacceptability and avoidance of uncertainty (now having seven items, a = .85)
and uncertainty related to frustration (now having five items, a = .85).
A subsequent CFA was performed using the Regina sample and the revised 12-item version of
the IUS. The presupposed two-factor structure was compared to a unitary structure and to the
previous 17- and 27-item models. All fit indices for the two-factor, 12-item solution (see Table 2)
were superior relative to the other solutions. Furthermore, the total score of the 12-item scale
correlated highly (r = .96) with the total score of the 27-item scale.

3.3. Cross validation

Using the Houston sample data, a CFA was performed evaluating one- and two-factor
solutions for the 12 item IUS. All of the six fit indices (Table 2) indicated the 12-item two-factor
model provided the best fit to the data even though the correlation between the two factors was
high (r = .73). Similar to the Regina sample, the total score of the 12-item scale was highly
correlated (r = .96) with the total score of the 27-item scale. Effectively, the total score from the
12-item IUS accounted for (92.8%) of the variance in the original 27-item scale, supporting the
comparability of the 12- and 27-item IUS.
Further supporting the use of the 12-item IUS was an analysis of convergent validity similar to
that performed by Norton (2005). Total scores for the 27-item IUS and the 12-item IUS were
correlated with measures related in previous studies (i.e., GADQ-IV, PSWQ, BAI). The
112 R.N. Carleton et al. / Journal of Anxiety Disorders 21 (2007) 105–117

Table 3
Convergent validity for the IUS-12 (n = 818)
Pearson correlations Independent variancea
IUS-27 IUS-12 IUS-27 IUS-12
BDI .63 .56 – –
BAI .62 .57 – –
PSWQ .57 .54 .09 .08
GADQ .64 .61 .05 .05
IUS-27, original 27-item IUS; IUS-12, revised 12-item IUS; BDI, Beck Depression Inventory-II; BAI, Beck Anxiety
Inventory; GADQ-IV, Generalized Anxiety Disorder Questionnaire-IV; PSWQ, Penn State Worry Questionnaire.
a
Variance accounted for by the IUS measure independent of anxiety or depression as measured by the BAI and BDI.

correlations between both versions of the IUS (12- and 27-items) from the same sample and
each related measure were significant (P < .01; Table 3). The magnitudes of the correlations
were then compared between the 27-item IUS and 12-item IUS using the Fisher test of Z
transformed independent-sample correlations. None of the pairs of correlations differed
significantly (all Z’s < .76), indicating little or no loss of construct validity as a result of
item-reduction.
In addition to convergent validity, Buhr and Dugas (2002), Freeston et al. (1994), and Norton
(2005) all demonstrated that the 27-item IUS was able to account for variance beyond that which
could be attributed to neuroticism or anxiety using hierarchical regression. Regressing the
GADQ-IV and the PSWQ, they each loaded first the BAI and the BDI, and then the IUS. Results
indicated that, in almost every case, the IUS significantly predicted GADQ-IV and PSWQ scores
beyond that already accounted for by the BAI and the BDI. Therefore, a similar regression
analysis was performed with the 12-item IUS to ensure item-removal had not reduced potentially
important construct independence. The percentage of variance accounted for by the 12-item IUS
was nearly identical to that of the 27-item IUS (Table 3). As per Norton (2005) the 12-item IUS
was also able to account for a significant portion of variance in both the GADQ-IVand the PSWQ
after accounting for variance shared with anxiety and depression.
Given these successful demonstrations of the psychometric properties of the 12-item IUS its
descriptive statistics and factor loadings are presented in Table 4. Factor loadings are presented
for psychometric evaluation only and should not be used to compute item weights (Wainer,
1976). IUS-12 scores should be based on a simple sum of items, with the total score being used
for evaluating a general intolerance of uncertainty. Content for the seven-item factor involves fear
and anxiety based on future events; accordingly, it might best be described as Prospective
Anxiety. In contrast, the five-item factor describes uncertainty inhibiting action or experience;
therefore, it might be best described as Inhibitory Anxiety.

4. Discussion

This study had two main goals. The first goal was to perform the first CFA of the English
version of the IUS, comparing the unitary, four-, and five-factor models. The results of the CFA
indicated that none of these models appropriately fit the data. Furthermore, the comparison did
not result in any one model being superior. This pattern of findings did, however, lend support
to the notion of revising the measure as recommended by Norton (2005), which was the second
goal of this study. Using one factor from each of the earlier multi-factorial models a 17-item
reduced scale was created. The 17-item scale, while reasonably statistically sound, was further
Table 4

R.N. Carleton et al. / Journal of Anxiety Disorders 21 (2007) 105–117

Descriptive statistics for the IUS-12 (n = 818)
Item IUS Factor loading M S.D. Skew Kurtosis
1. Unforeseen events upset me greatly. 7 .75 1.94 1.11 1.08 0.31
2. It frustrates me not having all the information I need. 8 .69 2.89 1.18 0.12 0.85
3. One should always look ahead so as to avoid surprises. 10 .64 2.44 1.20 0.49 0.71
4. A small, unforeseen event can spoil everything, even with the best of planning. 11 .67 2.21 1.22 0.77 0.42
5. I always want to know what the future has in store for me. 18 .66 2.75 1.26 0.30 0.88
6. I can’t stand being taken by surprise. 19 .71 1.85 1.03 1.17 0.78
7. I should be able to organize everything in advance. 21 .59 2.60 1.22 0.35 0.84
Prospective Anxiety 16.68 6.00 0.61 0.10
8. Uncertainty keeps me from living a full life. 9 .75 1.84 1.10 1.24 0.68
9. When it’s time to act, uncertainty paralyses me. 12 .75 1.75 1.04 1.35 1.04
10. When I am uncertain I can’t function very well. 15 .79 1.94 1.01 0.92 0.22
11. The smallest doubt can stop me from acting. 20 .65 1.95 1.05 1.24 0.31
12. I must get away from all uncertain situations. 25 .74 1.70 1.03 1.51 1.58
Inhibitory Anxiety 9.17 4.15 1.18 1.04
Total IUS-12 25.85 9.45 0.84 0.31

113
114 R.N. Carleton et al. / Journal of Anxiety Disorders 21 (2007) 105–117

reduced via an iterative process of inter-item correlations, factor loadings, and rational theory.
The reduced scale (IUS-12) maintained exemplary internal consistency, was highly correlated
with the 27-item version, and based on principle of parsimony was accepted over the 17-item
version.
Following the reduction process, a CFA using an independent sample for replication and
cross-validation of the postulated 12-item, two-factor structure of the IUS-12 was conducted. The
results supported the postulated model, and the two factors were named Prospective Anxiety and
Inhibitory Anxiety. The internal consistency of each of the factors was acceptable and the two
remained moderately correlated, as expected when measuring two aspects of one latent variable.
Collectively these findings provide convincing support for use of the IUS-12 as an efficient,
psychometrically-sound tool for measuring intolerance of uncertainty. This shorter IUS should
also encourage a broader evaluation of intolerance of uncertainty in both clinical and non-clinical
populations.
Intolerance of uncertainty, being fundamental to worry (Laugesen et al., 2003), state anxiety
(Greco & Roger, 2001), and related anxiety pathologies (Tolin et al., 2003), may play an as-of-yet
unexplored but substantial role in several anxiety disorders. The clinical and theoretical utility of
the statistically significant variance beyond depression and anxiety accounted for by the IUS-12
in the measures of worry remains to be assessed. Although the additional variance was small,
future models of anxiety constructs can test whether intolerance of uncertainty is a core
component of anxiety disorders in general, similar to the growing research investigating various
fears as vulnerability factors in diathesis-stress models of psychopathology. For example, recent
research has suggested that a variety of cognitive vulnerability factors contribute to the
development of pathologies such as PTSD and chronic pain (e.g., Asmundson, 1999;
Asmundson, Coons, Taylor, & Katz, 2002; Keane & Barlow, 2002).
Anxiety sensitivity – the tendency to catastrophically misinterpret physical sensations
related to anxiety (e.g., rapid heart rate, shaky hands) as being threatening (Peterson & Reiss,
1992) – is a cognitive vulnerability that may have a significant relationship with intolerance
of uncertainty. Ambiguity or uncertainty itself can be considered threatening (Epstein, 1972),
provoking anxiety, and further enhancing the perceived threat of a stimulus or situation (e.g.,
Hadjistavropoulos, Craig, & Hadjistavropoulos, 1997; Heydayati et al., 2003; Hock &
Krohne, 2004; Mogg et al., 1994). Uncertainty regarding a stimulus opens a person to a
variety of potential interpretations and responses (Lazarus, 1991), including catastrophic
appraisal. A person with high anxiety sensitivity may be uncertain whether a heart palpitation
is a nondescript anomaly or a precursor to a heart attack. If the person was certain the
palpitation was not threatening, that certainty might alleviate or completely ameliorate
associated anxiety and fear. Therefore, intolerance of uncertainty and anxiety sensitivity may
have a hierarchical or mediational relationship; however, this remains to be explored. The
IUS-12 facilitates such explorations by eliminating difficulties with the length and unstable
psychometric properties of the original IUS (Buhr & Dugas, 2002; Freeston et al., 1994;
Norton, 2005).
In spite of the positive implications this work has for anxiety research, there are some
limitations that provide avenues for future evaluation of the IUS-12. First, it could be argued that
the positive keying of all items in both the IUS and the IUS-12 precludes the opportunity to
provide scales that are balanced in their keying direction. However, with fewer items there is less
opportunity for an unintentional response bias that could be corrected by negatively keying some
items. Regarding intentional response biases, it was thought premature to suggest both item
reduction and re-keying simultaneously, particularly given the evidence that reverse-wording can
 R.N. Carleton et al. / Journal of Anxiety Disorders 21 (2007) 105–117 115

create undesirable methodological factors (see Marsh, 1996; Rodebaugh et al., 2004; Spector,
Van Katwyk, Brannick, & Chen, 1997). Second, the persistent use of undergraduate samples in
the design and revision of the IUS makes establishing psychometric validity for clinical
populations an important next step for research into intolerance of uncertainty. Until then, the use
of the IUS-12 in clinical populations warrants caution. Third, discriminant and clinical validity of
the IUS-12 remain to be tested, as do psychometric properties for different racial groups and
languages for which it might be translated.
Future research should include additional measures of anxiety disorders and anxiety-related
constructs (e.g., GAD, OCD, panic disorder). Dugas et al. (2001) began such work by associating
panic sensations and anxiety sensitivity with intolerance of uncertainty using the original IUS.
Given the reasonable prospect that intolerance of uncertainty may have overlap with anxiety
sensitivity, it may also overlap with other related fears such as illness/injury sensitivity and fear of
negative evaluation (Carleton, Asmundson, & Taylor, 2005; Taylor, 1993), which could be
included in future investigations. The theoretical association between an inability to cope with
ambiguity or uncertainty and the development or maintenance of an anxiety disorder may be
demonstrable with the IUS-12. This is an avenue of future investigation that may result in further
delineation of fundamental dimensions and underlying mechanisms of the anxiety disorders;
thereby improving assessment and treatment efforts for these potentially chronic and debilitating
conditions.

References

Antony, M. M., Orsillo, S. M., & Roemer, L. (2001). Practitioner’s guide to empirically based measures of anxiety. AABT
clinical assessment series. Dordrecht, Netherlands: Kluwer Academic Publishers.
Asmundson, G. J. G. (1999). Anxiety sensitivity and chronic pain: empirical findings, clinical implications, and future
directions. In: S. Taylor (Ed.), Anxiety sensitivity, theory, research and treatment of the fear of anxiety (pp. 268–269).
Mahwah, NJ: Erlbaum.
Asmundson, G. J. G., Coons, M. J., Taylor, S., & Katz, J. (2002). PTSD and the experience of pain: research and
clinical implications of shared vulnerability and mutual maintenance models. Canadian Journal of Psychiatry, 47,
930–937.
Asmundson, G. J. G., Norton, P. J., & Vlaeyen, J. W. S. (2004). Understanding and treating fear of pain. Oxford: Oxford
University Press.
Barlow, D. H. (2002). Anxiety and its disorders. New York, NY: Guilford Press.
Beck, A. T., Epstein, N., Brown, G., & Steer, R. A. (1988). An inventory for measuring clinical anxiety: psychometric
properties. Journal of Consulting & Clinical Psychology, 56, 893–897.
Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Beck depression inventory manual (2nd ed.). San Antonio, TX:
Psychological Corporation.
Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An inventory for measuring depression.
Archives of General Psychiatry, 4, 561–571.
Browne, M. W., & Cudeck, R. (1993). Alternate ways of assessing model fit. In: K. A. Bollen, & J. S. Long (Eds.), Testing
structural equation models (pp. 136–162). Newbury Park, CA: Sage.
Budner, S. (1962). Intolerance of ambiguity as a personality variable. Journal of Personality, 30, 29–50.
Buhr, K., & Dugas, M. J. (2002). The Intolerance of Uncertainty Scale: psychometric properties of the English version.
Behaviour Research and Therapy, 40, 931–946.
Carleton, R. N., Asmundson, G. J. G., & Taylor, S. (2005). Fear of physical harm: factor structure and psychometric
properties of the Injury/Illness Sensitivity Index. Journal of Psychopathology and Behavioral Assessment, 27, 235–
241.
Cattell, R. B. (1966). The scree test for number of factors. Multivariate Behavioural Research, 1, 245–276.
Creamer, M., Foran, J., & Bell, R. (1995). The Beck Anxiety Inventory in a non-clinical sample. Behaviour Research and
Therapy, 33, 477–485.
Dozois, D. J. A., Dobson, K. S., & Ahnberg, J. L. (1998). A psychometric evaluation of the Beck Depression Inventory-II.
Psychological Assessment, 10, 83–89.
116 R.N. Carleton et al. / Journal of Anxiety Disorders 21 (2007) 105–117

Dugas, M. J., Freeston, M. H., & Ladouceur, R. (1997). Intolerance of uncertainty and problem orientation in worry.
Cognitive Therapy & Research, 21, 593–606.
Dugas, M. J., Gagnon, F., Ladouceur, R., & Freeston, M. H. (1998). Generalized anxiety disorder: a preliminary test of a
conceptual model. Behaviour Research and Therapy, 36, 215–226.
Dugas, M. J., Gosselin, P., & Ladouceur, R. (2001). Intolerance of uncertainty and worry: investigating specificity in a
nonclinical sample. Cognitive Therapy and Research, 25, 551–558.
Epstein, S. (1972). The nature of anxiety with emphasis upon its relationship to expectancy. In C. D. Spielberger (Ed.)
Anxiety: current trends in theory and research (Vol. 2, pp. 291–337). New York: Academic Press.
Freeston, M., Rhéaume, J., Letarte, H., Dugas, M. J., & Ladouceur, R. (1994). Why do people worry? Personality &
Individual Differences, 17, 791–802.
Greco, V., & Roger, D. (2001). Coping with uncertainty: the construction and validation of a new measure. Personality &
Individual Differences, 31, 519–534.
Greco, V., & Roger, D. (2003). Uncertainty, stress, and health. Personality & Individual Differences, 34, 1057–
1068.
Hadjistavropoulos, H. D., Craig, K. D., & Hadjistavropoulos, T. (1997). Cognitive and behavioral responses to illness
information: the role of health anxiety. Behaviour Research and Therapy, 36, 149–164.
Heydayati, M., Dugas, M. J., Buhr, K., & Francis, K. (2003, November). The relationship between intolerance of
uncertainty and the interpretation of ambiguous and unambiguous information. Poster presented at the Annual
Convention of the Association for Advancement of Behaviour Therapy, Boston, MA.
Hock, M., & Krohne, H. W. (2004). Coping with threat and memory. Emotion, 4, 65–86.
Hu, L., & Bentler, P. M. (1999). Fit indices in covariance structure modeling: sensitivity to underparameterized model
mis-specification. Psychological Methods, 3, 424–453.
Jöreskog, K., & Sörbom, D. (1996). LISREL 8: user’s reference guide. Chicago, IL: Scientific Software International, Inc.
Jöreskog, K., & Sörbom, D. (1999). LISREL 8.30 and PRELIS 2.30. Scientific Software International, Inc.
Kaiser, H. F. (1961). A note on Guttman’s lower bound for the number of common factors. Multivariate Behavioral
Research, 1, 249–276.
Keane, T. M., & Barlow, D. H. (2002). Posttraumatic stress disorder. In: D. H. Barlow (Ed.), Anxiety and its disorders (pp.
418–453). New York: Guilford.
Kirton, M. J. (1981). A reanalysis of two scales of tolerance of ambiguity. Journal of Personality Assessment, 45, 407–
414.
Lazarus, R. S. (1991). Emotion and adaptation. New York: Oxford University Press.
Laugesen, N., Dugas, M. J., & Bukowski, W. M. (2003). Understanding adolescent worry: the application of a cognitive
model. Journal of Abnormal Child Psychology, 31, 55–64.
Marsh, H. W. (1996). Positive and negative global self-esteem: a substantively meaningful distinction or artifactors?
Journal of Personality and Social Psychology, 70, 810–819.
Marsh, H. W., Balla, J. R., & McDonald, R. P. (1988). Goodness-of-fit indexes in confirmatory analysis: the effect of
sample size. Psychological Bulletin, 103, 391–410.
McNeil, D. W., & Vowles, K. E. (2004). In: G. J. G. Asmundson, P. J. Norton, & J. W. S. Vlaeyen (Eds.), Understanding
and treating fear of pain. Oxford: Oxford University Press.
Meyer, T. J., Miller, M. L., Metzger, R. L., & Borkovec, T. D. (1990). Development and validation of the Penn State Worry
Questionnaire. Behaviour Research and Therapy, 28, 487–495.
Mogg, K., Bradley, B. P., Miller, T., Potts, H., Glenwright, J., & Kentish, J. (1994). Interpretation of homophones related to
threat: anxiety or response bias effects. Cognitive Therapy and Research, 18, 461–477.
Newman, M. G., Zuellig, A. R., Kachin, K. E., Constantino, M. J., Przeworski, A., Erickson, T., et al. (2002). Preliminary
reliability and validity of the generalized anxiety disorder questionnaire-IV: a revised self-report diagnostic measure
of generalized anxiety disorder. Behavior Therapy, 33, 215–233.
Norton, P. J. (2005). A psychometric analysis of the Intolerance of Uncertainty Scale among four racial groups. Journal of
Anxiety Disorders, 6, 699–707.
Peterson, R. A., & Reiss, S. (1992). Anxiety Sensitivity Index manual (2nd ed.). Worthington, OH: International
Diagnostic Systems.
Rodebaugh, T. L., Woods, C. M., Thissen, D. M., Heimberg, R. G., Chambless, D. L., & Rapee, R. M. (2004). More
information from fewer questions: the factor structure and item properties of the original and brief fear of negative
evaluation scale. Psychological Assessment, 2, 169–181.
Spector, P. E., Van Katwyk, P. T., Brannick, M. T., & Chen, P. Y. (1997). When two factors don’t reflect two constructs:
how item characteristics can produce artifactual factors. Journal of Management, 23, 659–677.
Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics (4th ed.). New York: Harper and Row.
 R.N. Carleton et al. / Journal of Anxiety Disorders 21 (2007) 105–117 117

Tallis, F., Eysenck, M., & Mathews, A. (1992). A questionnaire for the measurement of nonpathological worry.
Personality and Individual Differences, 13, 161–168.
Taylor, S. (1993). The structure of fundamental fears. Journal Behavior Therapy and Experimental Psychiatry, 24,
289–299.
Tolin, D. F., Abramowitz, J. S., Brigidi, B. D., & Foa, E. B. (2003). Intolerance of uncertainty in obsessive-compulsive
disorder. Journal of Anxiety Disorders, 17, 233–242.
Wainer, H. (1976). Estimating coefficients in linear models: it don’t make no nevermind. Psychological Bulletin, 83,
312–317.
Westgard, J. O. (1999). Basic method validation. Madison, WI: Westgard QC, Inc.

View publication stats