Journal of Anxiety Disorders 21 (2007) 493509

The Driving Cognitions Questionnaire: Development

and preliminary psychometric properties
Anke Ehlers a,*, Joanne E. Taylor b, Thomas Ehring a,1,
Stefan G. Hofmann c, Frank P. Deane d, Walton T. Roth e,f,
John V. Podd b
a
Department of Psychology, Institute of Psychiatry, Kings College London, London, UK
b
School of Psychology, Massey University, New Zealand
c
Department of Psychology, Boston University, Boston, MA, USA
d
Department of Psychology, University of Wollongong, Australia
e
Department of Veterans Affairs Health Care System, Palo Alto, CA, USA
f
Stanford University School of Medicine, Stanford, CA, USA
Received 15 March 2006; received in revised form 5 July 2006; accepted 1 August 2006

Abstract
Recent research has suggested that fear of driving is common in the general population. People may have
various concerns when driving, and instruments for the assessment of these concerns are lacking. The present
paper describes the development and preliminary evaluation of the Driving Cognitions Questionnaire (DCQ).
The DCQ is a 20-item scale that measures three areas of driving-related concernspanic-related, accident-
related, and social concerns. In three separate samples from different countries (n = 69, 100, and 78), the scale
showed good internal consistency and substantial correlations with measures of the severity of driving fear. It
discriminated well between people with and without driving phobia. It also showed convergent validity with
other measures. The questionnaire shows promise for use in research and clinical practice.
# 2006 Elsevier Ltd. All rights reserved.

Keywords: Driving phobia; Cognition; Panic disorder; Social phobia; Posttraumatic stress disorder; Travel phobia

Fear of driving is common in the general population (Ehlers, Hofmann, Herda, & Roth, 1994;
Mathew, Weinman, Semchuk, & Levin, 1982; Munjack, 1984; Taylor, 2002; Taylor & Deane,
2000; Taylor, Deane, & Podd, 2000). In clinical settings, fear of driving may either present as the

* Correspondence to: Department of Psychology P077, Institute of Psychiatry, De Crespigny Park, London SE5 8AF,
UK. Tel.: +44 20 7848 5033; fax: +44 20 7848 0763.
E-mail address: a.ehlers@iop.kcl.ac.uk (A. Ehlers).
1
Present address: Department of Psychology, University of Bielefeld, Germany.

0887-6185/$ see front matter # 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.janxdis.2006.08.002
494 A. Ehlers et al. / Journal of Anxiety Disorders 21 (2007) 493509

main problem (i.e., as a specific phobia) or as part of another anxiety disorder (e.g., agoraphobia).
It may also develop as a response to a traumatic road crash or motor vehicle accident (MVA;
Blanchard & Hickling, 1997). Whereas the phenomenology of driving fears shares many features
among these subgroups of people (Blanchard & Hickling, 1997; Ehlers et al., 1994; Hofmann,
1992), the concerns that lead people to be afraid of driving may be quite different. For example,
whereas some people may mainly be concerned about having an accident and about the
consequences of an accident, others may have concerns about being suddenly incapacitated by
anxiety while driving, similar to the concerns of patients with panic disorder. Furthermore,
driving fear can be thought of as performance-related fear that is specific to the driving situation,
thus overlapping in relevant concerns with social phobia.
Several reliable and valid measures of cognition have been developed to assess the typical
concerns of people with panic disorder (e.g., Chambless, Caputo, Bright, & Gallagher, 1984),
post-traumatic stress disorder (e.g., Foa, Ehlers, Clark, Tolin, & Orsillo, 1999), and social phobia
(e.g., Turner, Johnson, Beidel, Heiser, & Lydiard, 2003). None of these specifically addresses
driving-related concerns. A specific and economical measure that captures the range of concerns
in people with driving phobia is needed to inform cognitive-behavioral treatments of driving
phobia, considering that evidence supports the role of cognitive mediation in fear reduction,
regardless of whether treatment is primarily behaviorally or cognitively oriented (Craske &
Rowe, 1997, p. 259).
The present paper describes the development and preliminary validation of the Driving
Cognitions Questionnaire (DCQ). The DCQ was developed in three related studies in different
countries. Study 1 was conducted by Ehlers, Hofmann and Roth (Ehlers et al., 1994; Hofmann,
1992) at Stanford University, California, USA. Study 2 was completed by Taylor, Deane, and
Podd at Massey University in Palmerston North, New Zealand. Study 3 was conducted by Ehring
and Ehlers at the Institute of Psychiatry in London, UK.

1. Study 1: scale development and initial psychometric properties

The goals of Study 1 were:

1. To develop a brief Driving Cognitions Questionnaire (DCQ) that represents typical

concerns of patients with driving phobia.
2. To conduct a preliminary test of the reliability and validity of the scale. Convergent and
discriminant validity were assessed by testing (a) whether the DCQ discriminated between
individuals with and without driving phobia, (b) whether the DCQ correlated with the
severity of driving phobia, and (c) whether the DCQ correlated with measures of anxiety
and depression, and measures of cognitions and behaviors typical of patients with panic
disorder and agoraphobia (who often show significant fear and avoidance of driving). We
expected the DCQ to show moderate correlations with the latter measures (e.g., Ninan &
Berger, 2001).

1.1. Method

1.1.1. Participants
Participants were 42 patients with driving phobia (35 women and 7 men) and 27 control
participants (23 women and 4 men) who did not report fear of driving, all of whom were recruited
from the study by Ehlers et al. (1994) through media advertisements. Those in the phobic group
 A. Ehlers et al. / Journal of Anxiety Disorders 21 (2007) 493509 495

(and none in the control group) met DSM-III-R criteria for specific phobia (driving). The majority
(81%) also reported panic attacks. The groups did not differ in age, t(65) = 0.48, p = .63, but there
was a marginally significant difference in years of education, t(63) = 1.89, p = .06. Mean age was
49 years for patients (S.D. = 12.80) and 50 years for controls (S.D. = 13.00). Patients had a mean of
15.80 years of education (S.D. = 3.50) and controls a mean of 14.20 years (S.D. = 3.40).

1.1.2. Generation of the item pool

The initial item pool of 49 items was generated by AE on the basis of structured clinical
interviews with patients who were afraid of driving. The item pool consisted of several a priori
categories of items:

1. Panic-related concerns. These were concerns shown by patients with panic disorder or
agoraphobia when driving (e.g., My heart will stop beating; I will be stranded).
2. Accident-related concerns. These fell into two groups of concerns: (2a) concerns that an
accident could happen (e.g., I will die in an accident, I will cause an accident), and (2b)
concerns about the consequences of an accident (e.g., If I have an accident, it will cause
financial problems, If I have an accident, I will not get to the hospital fast enough).
3. Concerns about other adverse events when driving (e.g., The engine will break down; I
will be attacked if the car breaks down).
4. Social concerns (e.g., People I care about will criticize me; People will think I am a bad
driver).

Each item on the DCQ was rated according to how often each thought (i.e., item) occurs while
driving, using a 5-point Likert scale from 0 (Never) to 4 (Always). The answer format was
modeled on the Agoraphobic Cognitions Questionnaire (Chambless et al., 1984). The total score
is the sum of the items (although the overall item mean was reported by Hofmann, 1992, and
Ehlers et al., 1994).

1.1.3. Other measures

Driving Situations Questionnaire (DSQ). The DSQ (Ehlers et al., 1994) is a self-report
measure of the severity of driving phobia. Respondents rate their discomfort or fear regarding 42
driving situations on a scale from 0 (No discomfort) to 4 (Extreme discomfort), and how much
they avoid the situations on a scale from 0 (Never avoid) to 4 (Always avoid). Ten situations
concern driving in residential areas (e.g., left turn, changing lanes), 10 situations concern driving
on busy urban thoroughfares, 13 situations concern driving on freeways (motorways), and 9 other
driving situations include tunnels, bridges, or steep roads. The total score summarizing
participants discomfort/anxiety ratings (DSQ-anxiety) and avoidance ratings (DSQ-avoidance)
when driving alone was used for Study 1. The internal consistencies for the scales in the present
sample were a = .98 and a = .97, respectively.
Agoraphobic Cognitions Questionnaire (ACQ) and Body Sensations Questionnaire (BSQ).
The ACQ and BCQ are companion scales developed by Chambless et al. (1984) for measuring
the construct of fear of fear. The ACQ is a 14-item measure of cognitions about the consequences
of anxiety, while the BSQ is a 17-item scale assessing the fear or concern about the somatic
symptoms associated with anxiety. The scales have demonstrated good psychometric properties,
including internal consistency (a = .80), testre-test reliability (r = .74), sensitivity to changes
with treatment, and discrimination between agoraphobic patients and non-patient controls
(Chambless et al., 1984).
496 A. Ehlers et al. / Journal of Anxiety Disorders 21 (2007) 493509

Mobility Inventory for Agoraphobia (MI). The MI (Chambless, Caputo, Jasin, Gracely, &
Williams, 1985) is a 27-item self-report measure of agoraphobic avoidance behavior and
frequency of panic attacks. Respondents rate to what extent they avoid a range of situations, (a)
when they are alone and (b) when accompanied, using a scale from 1 (Never avoid) to 5 (Always
avoid). The MI has adequate testre-test reliability (r = .62.90) and individual item reliability
(r = 0.500.90; Chambless et al., 1985). It has good convergent and construct validity (Cox,
Swinson, Kuch, & Reichman, 1993; Kwon, Evans, & Oei, 1990) and has been shown to
discriminate those with agoraphobia from those with other anxiety disorders (Craske, Rachman,
& Tallman, 1986). Study 1 used the ratings for when the person is alone.
State-Trait Anxiety Inventory, trait version (STAI-T). The trait version of the STAI (Form Y;
Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983) is a standardized 20-item self-report
measure of general anxiety. The STAI has been shown to have good reliability and validity. Test
re-test reliability for the trait anxiety scale is a = .81. Internal consistency coefficients range from
.83 to .92 (Spielberger et al., 1983).
Beck Depression Inventory (BDI). The BDI (Beck, Ward, Mendelsohn, Mock, & Erbaugh,
1961) is a 21-item self-report measure of depression that has been shown in previous research to
have good reliability and validity. Internal consistency for the BDI ranges from .58 to .93, and
testre-test reliability estimates range from .69 to .90 (Beck, Steer, & Garbin, 1988). The BDI
correlates highly (r = .96) with clinician ratings of depression (Beckham & Leber, 1995).

1.1.4. Data analysis

Data for all studies were analyzed using SPSS 13.0.1 for Windows (SPSS Inc., 2004). All
correlations and t tests are two-tailed unless otherwise specified.

1.2. Results

1.2.1. Item selection for the DCQ

The item selection followed a step-wise procedure. First, 15 items that did not show a
significant difference between patients with driving phobia and controls at p < .01 were
eliminated from the item pool. This mainly affected items of categories 2b and 3 (i.e., concerns
about the consequences of an accident and about other adverse events when driving). Second, the
remaining 34 items were subjected to a Principal Component Analysis with Varimax rotation.
The scree test suggested a three-factor solution, regardless of whether the whole sample (patients
and controls) or patients only were considered. The three factors explained 63.70% of the
variance. The first factor could be interpreted as panic-related concerns (16 items) and the second
factor could be interpreted as accident-related concerns (10 items). The third factor could be
interpreted as social concerns, or concerns about being embarrassed when driving (8 items).
Third, we calculated the correlations of the items with scores on the DSQ in the patient group
and eliminated items with the lowest correlations to reduce the number of items for each factor,
and to make the subscales similar in length (six items of factor 1, two items of factor 2, and one
item of factor 3 with the lowest correlations deleted). Fourth, we eliminated some items with low
mean scores in the patient group (two items of factor 1, and one item for each of factors 2 and 3).
Fifth, we calculated the internal consistencies of the factor scales and eliminated items that
lowered the Cronbachs a for that scale (one item of factor 1).
The final questionnaire contained 20 items (see Appendices A and B). Seven of the items
represented panic-related concerns, seven accident-related concerns, and six social concerns.
The remaining items were analyzed with a Principal Axis Factor analysis to test whether the
 A. Ehlers et al. / Journal of Anxiety Disorders 21 (2007) 493509 497

Table 1
Principal factor analyses of final items (factor loadings on original factors and all above .40)
Item Study 1 Study 2
P A S P A S
Factor number 1 2 3 3 2 1
Panic concerns
Will tremble and not be able to steer .51 .36 .45
Will be trapped .79 .47
Will be unable to catch breath .79 .93
Heart will stop beating .87 .75
Will be stranded .84 .54
Will not be able to think clearly .56 .46 .43 .43
Will be unable to move .75 .44
Accident-related concerns
Will not react fast enough .85 .23 .50
Will injure someone .73 .61
Will be injured .75 .41 .76
Will die in an accident .68 .78
People riding with me will be hurt .56 .53 .57
I will cause an accident .71 .32 .71
I cannot control other cars .51 .61
Social concerns
People will criticize me .65 .75
Will lose control and act stupidly .47 .53 .55 .45
People will think I am a bad driver .54 .69 .79
I will hold up traffic/people will be angry .51 .36 .78
People will laugh at me .82 .71
Other people will notice I am anxious .82 .62
Note: P, panic concerns; A, accident-related concerns; S, social concerns.

original factor structure replicated. This was the case, as shown in Table 1. The scree plot again
suggested a three-factor solution, with the first three factors explaining 46.4, 11.6 and 7.5% of the
variance, respectively. The factors could be interpreted as panic-related concerns, accident-
related concerns and social concerns. All items showed substantial loadings on the original
factor. With the exception of two items (I will hold up traffic and people will be angry and
People riding with me will be hurt), all items also showed the highest factor loading on the
original factor.

1.2.2. Psychometric properties of the DCQ

Internal consistency. The internal consistencies for the DCQ factor scales in the patient
sample were a = .93 for panic-related concerns, a = .92 for accident-related concerns, and
a = .89 for social concerns. The scales showed substantial inter-correlations, as seen in Table 2.
Cronbachs a for the total scale was a = .96.
Validation. Table 2 also shows the inter-correlations of the DCQ scales with the severity of
driving phobia, as measured by the DSQ, in the patient group. As expected, the correlations were
high. Table 3 shows the group differences on the DCQ scales. Patients had significantly higher
scores on all scales. A logistic regression analysis showed that the DCQ total scale significantly
predicted diagnostic status, Wald (1, 69) = 12.39, p < .001, x2 = 35.95, Nagelkerke R2 = .55.
498 A. Ehlers et al. / Journal of Anxiety Disorders 21 (2007) 493509

Table 2
Intercorrelations between the Driving Cognitions Questionnaire Scales and the Severity of Driving Phobia
Driving Cognitions Questionnaire Driving Anxiety and Avoidance
DCQ-Panic DCQ-Accident DCQ-Social DSQ-Anxiety DSQ-Avoidance
Study 1 (n = 42, driving phobia)
DCQ-Total .87 *** .88 *** .83 *** .70*** .60 ***
DCQ-Panic .63 *** .56 *** .64*** .65 ***
DCQ-Accident .64 *** .59*** .43 **
DCQ-Social .55*** .43 ***
Study 2 (n = 50, driving fear)
DCQ-Total .76 *** .78 *** .79 *** .58***
DCQ-Panic .40 ** .44 *** .40**
DCQ-Accident .37 ** .49***
DCQ-Social .44***

DCQ-Panic DCQ-Accident DCQ-Social TPQ-Anxiety TPQ-Avoidance

Study 3 (n = 78, traffic accident survivors)
DCQ-Total .96 *** .93 *** .91*** .59 *** .65***
DCQ-Panic .85 *** .82*** .56 *** .62***
DCQ-Accident .74*** .58 *** .56***
DCQ-Social .53 *** .65***
Note: DCQ, Driving Cognitions Questionnaire; DSQ, Driving Situations Questionnaire; TPQ, Travel Phobia Ques-
tionnaire.
**
p < .01.
***
p < .001.

The total percentage of correct classifications was 78%, with 79% of the patient group and 78%
of the control group correctly classified.
As shown in Table 4, the DCQ scales also correlated with trait anxiety and depression in the
patient group. The DCQ total score also correlated with cognitive and behavioral measures
designed for panic disorder and agoraphobia, the ACQ, BSQ, and MI. This was mainly due to the
correlations of the DCQ-Panic subscale with these measures.
This raises the question of whether the correlation of the DCQ with the severity of driving
phobia is due to its correlation with the ACQ and BSQ. This was not the case. The partial
correlations of the DCQ total score with DSQ-Anxiety and DSQ-Avoidance were r = .65 and
r = .56, respectively, when ACQ and BSQ scores were partialled out.
The DCQ scales were unrelated to age, sex, marital status, or education.

2. Study 2: replication of psychometric properties and further validation

The goals of Study 2 were:

1. To further investigate the internal consistency as well as convergent and discriminant

validity of the DCQ. As with Study 1, we were interested in assessing how well the DCQ
discriminated people with driving fear from nonanxious controls. We were also interested in
the correlations of the DCQ with measures of the severity of driving fear and other measures
of phobic avoidance, anxiety and depression.
2. To examine the factor structure of the DCQ with an independent sample of people with
driving fear.
 A. Ehlers et al. / Journal of Anxiety Disorders 21 (2007) 493509 499

Table 3
Group differences on the Driving Cognitions Questionnaire, means (S.D.)
Patients with driving phobia (n = 42) Control group (n = 27) t*
Study 1
DCQ-Total 23.26 (17.61) 4.78 (4.93) 6.42 (df = 50.43)
DCQ-Panic 6.95 (7.61) 1.03 (1.43) 4.96 (df = 45.40)
DCQ-Accident 9.52 (6.95) 2.85 (2.60) 5.64 (df = 56.53)
DCQ-Social 6.79 (5.89) 0.89 (1.58) 6.15 (df = 49.68)

Driving fear group (n = 50) Control group (n = 50) t

Study 2
DCQ-Total 27.68 (13.77) 6.30 (5.25) 10.26 (df = 62.98)
DCQ-Panic 5.42 (5.09) 1.20 (2.00) 5.46 (df = 63.77)
DCQ-Accident 11.34 (6.40) 4.10 (2.77) 7.34 (df = 66.78)
DCQ-Social 10.92 (6.22) 1.00 (1.34) 11.01 (df = 53.53)

Accident survivors with travel Accident survivors without t

phobia (n = 40) phobia (n = 38)
Study 3
DCQ-Total 27.85 (16.95) 10.61 (10.58) 5.42 (df = 65.88)
DCQ-Panic 7.98 (6.45) 2.00 (3.86) 4.99 (df = 64.27)
DCQ-Accident 12.69 (6.44) 6.08 (4.42) 5.31 (df = 69.29)
DCQ-Social 7.25 (5.75) 2.52 (3.26) 4.49 (df = 62.32)
Note: DCQ, Driving Cognitions Questionnaire. All t-tests were significant at p < .001.

2.1. Method

2.1.1. Participants
Participants were 50 New Zealand volunteers who responded to media advertising that asked
for people who had some degree of driving fear. The sample was comprised solely of females as
no males volunteered. Their mean age was 43.60 years, S.D. = 14.99. A control comparison
group of 50 females who reported no driving fear was recruited. There were no significant
differences between the fearful and control groups for age, t(98) = 0.74, p = .46 (control group,
M = 41.36, S.D. = 14.95) and years of driving experience, t(98) = 0.94, p = .35 (fearful group,
M = 20.36, S.D. = 14.38; control group, M = 22.94, S.D. = 12.98). All participants held a current
drivers license. Exclusion criteria for the control group included the absence of current
psychological disorders that met Diagnostic and Statistical Manual of Mental Disorders criteria
(fourth edition, DSM-IV; American Psychiatric Association, 1994), as assessed by a
computerized diagnostic interview (Composite International Diagnostic Interview, CIDI Auto
2.1; World Health Organization, 1997). Although not recruited from a clinical sample, 23 (46%)
of the fearful group met DSM-IV criteria for an anxiety disorder (and there were a number of
cases with multiple anxiety disorder diagnoses). Overall, the driving fear sample in Study 2 was
considered less severe diagnostically than the phobic sample in Study 1.

2.1.2. Measures
Driving Cognitions Questionnaire (DCQ). Participants completed the item pool of the DCQ,
and Study 2 focuses on the 20-item version developed in Study 1.
500 A. Ehlers et al. / Journal of Anxiety Disorders 21 (2007) 493509

Table 4
Correlations between the DCQ scales and other measures
Cognition measures Avoidance measures Symptom measures
ACQ BSQ MI-Alone STAI-T BDI
Study 1 (n = 42, driving phobia)
DCQ-Total .68*** .52*** .43** .57 *** .48**
DCQ-Panic .77*** .60*** .55*** .49 *** .41**
DCQ-Accident .41** .26 .19 .49 *** .45**
DCQ-Social .54** .47** .33* .47 ** .35*

Avoidance measures Symptom measures

FQ-Total FQ-Social FQ-Ag STAI-T BDI-II
Study 2 (n = 50, driving fear)
DCQ-Total .38 ** .28 * .38** .37 ** .34*
DCQ-Panic .45 ** .19 .55** .40 ** .24
DCQ-Accident .30 .17 .27 .27 .35*
DCQ-Social .17 .29 * .11 .21 .19

Cognition measures Avoidance measures Symptom measures

TPBQ PTCI IPSI SBQ-t PDS BDI
Study 3 (n = 78, traffic accident survivors)
DCQ-Total .47 *** .42 *** .45 *** .48 *** .57*** .53***
DCQ-Panic .44 ** .41 *** .46 *** .52 *** .57*** .50***
DCQ-Accident .42 ** .35 ** .36 *** .36 ** .50*** .46***
DCQ-Social .43 ** .42 *** .44 *** .46 *** .54*** .53***
Note: DCQ, Driving Cognitions Questionnaire; ACQ, Agoraphobic Cognitions Questionnaire; BSQ, Body Sensations
Questionnaire; MI, Mobility Inventory; STAI-T, State-Trait Anxiety InventoryTrait Scale; BDI, Beck Depression
Inventory; FQ-Total, Total score on the Fear Questionnaire; FQ-Ag, Fear QuestionnaireAgoraphobia subscale; FQ-
Social, Fear QuestionnaireSocial phobia subscale; TPBQ, Travel Phobia Beliefs Questionnaire; PTCI, Posttraumatic
Cognitions Questionnaire; IPSI, Interpretations of PTSD Symptoms Inventory; PDS, Posttraumatic Diagnostic Scale;
SBQ-t, Safety Behavior Questionnaire (Travel).
*
p < .05.
**
p < .01.
***
p < .001.

Driving Situations Questionnaire (DSQ). The DSQ (Ehlers et al., 1994) as described above
was used, with some modifications (see Taylor, 2002). The total score summarizing participants
discomfort/anxiety ratings (DSQ-anxiety) when driving alone was used for Study 2 (avoidance
ratings were not included because of the use of the Fear Questionnaire, a measure of general
avoidance behavior).
State-Trait Anxiety Inventory, trait version (STAI-T). As with Study 1, Study 2 used the STAI-
T (Spielberger et al., 1983).
Fear Questionnaire (FQ). The FQ (Marks & Mathews, 1979) is a brief, standard self-rating
form that has been widely used in research on anxiety disorders and has well-documented
psychometric properties (e.g., Lee & Oei, 1994; Marks & Mathews, 1979; Oei, Moylan, & Evans,
1991). Respondents rated the extent to which they avoided various phobic situations. Consistent
with the recommendations of Moylan and Oei (1992), only the main FQ items were used in Study
2. The FQ provides a total phobia score as well as scores for the agoraphobia, social phobia, and
bloodinjury phobia subscales.
 A. Ehlers et al. / Journal of Anxiety Disorders 21 (2007) 493509 501

Beck Depression Inventory-Second Edition (BDI-II). The BDI has been described above, and
Study 2 used the second edition of this measure (Beck, Steer, & Brown, 1996).

2.2. Results

2.2.1. Psychometric properties of the DCQ

Internal consistency. The internal consistency for the fearful group was a = .88 for the total
scale, a = .78 for panic-related concerns, a = .82 for accident-related concerns, and a = .86 for
social concerns. The subscales showed moderate intercorrelations (Table 2).
Validation. The 20 DCQ items were subjected to a principal axis factor analysis
with varimax rotation, using the scores for the fearful group. The scree test suggested a three-
factor solution. The three factors explained 28.7, 12.4 and 10.2% of the variance, respectively.
The first factor could be interpreted as social concerns, the second factor as accident-related
concerns, and the third factor as panic-related concerns. With a few exceptions, the
items loaded highest on the factor that they were originally assigned to in Study 1
(Table 1).
As shown in Table 2, the DCQ scales correlated moderately with the severity of driving fear, as
measured by the DSQ, in the fearful group.
Table 3 shows the group differences on the DCQ scales. Participants with driving fears had a
higher total DCQ score than the control group. The pattern of group differences on the factor
scales were also similar to the results from Study 1, although the mean score for the social
concerns factor scale was higher in the New Zealand driving fear group, with relatively lower
scores on the panic- and accident-related factor scales.
Table 4 presents the intercorrelations between the DCQ scales and the other avoidance and
symptom measures used in Study 2 for the fearful sample. As with Study 1, the total DCQ
score correlated moderately with trait anxiety and depression. The DCQ total score also
correlated moderately with general avoidance behavior (FQ) as well as with agoraphobic and
social phobic avoidance. The correlation with the agoraphobia subscale of the FQ was mainly
due to the contribution of the panic factor scale, and similarly the correlation with the social
phobia subscale of the FQ was mainly due to the contribution of the social factor scale of the
DCQ.
The DCQ scales were unrelated to age, marital status, or education.

3. Study 3: replication of psychometric properties and further validation with traffic

accident survivors

The aim of the third study was to replicate the psychometric properties of the DCQ in a
sample of road traffic accident survivors. Goals were to determine whether the DCQ (a)
discriminated between accident survivors with and without travel phobia, (b) correlated with
measures of travel-related fear and avoidance, and (c) correlated with symptoms of
posttraumatic stress disorder and depression. We also investigated (d) whether the DCQ
correlated with measures of cognitions and safety behaviors typical of patients with
posttraumatic stress disorder (who often show significant fear and avoidance of travel
after traffic accidents). Safety behaviors are excessive precautions that are thought to be
motivated by maladaptive appraisals such as the ones measured in the DCQ (Ehlers & Clark,
2000). We therefore expected the DCQ to show moderate correlations with these
measures.
502 A. Ehlers et al. / Journal of Anxiety Disorders 21 (2007) 493509

3.1. Method

3.1.1. Participants
Participants were recruited from two samples of road traffic accident survivors who had
participated in earlier studies (Ehring, Ehlers, & Glucksman, 2006, in preparation). In the
original studies, 55 participants (out of 247) fulfilled diagnostic criteria for travel phobia as
assessed by the Structured Clinical Interview for the DSM-IV (First, Spitzer, Gibbon, &
Williams, 1996) and were therefore selected for participation in this study. Fifty-five participants
without travel phobia and matched by age and sex were drawn from the same samples as a control
group. All 110 individuals were contacted via letter and asked to fill in a pack of questionnaires,
containing the DCQ and other measures described below. Participants were offered 15 as a
reimbursement for their time. The sample for Study 3 comprised 78 participants who returned the
questionnaire (response rate: 70.9%; n = 40 with travel phobia).
All participants had experienced a road traffic accident on average 2 years prior to
participation in this study (M = 27.7 months, S.D. = 8.4, minimum = 13.7, maximum = 43.7).
Participants had experienced their accident either as a car driver or as a passenger, motorcyclist,
or bicyclist. Participants with versus those without travel phobia did not differ in demographic
variables: age, M = 36.0 years in both groups, t(76) = 0.29, p = .77; sex (58% vs. 65% female,
x2(1, N = 78) = 0.52, p = .64); years of education, t(73) = 0.75, p = .46; ethnic background, 65%
versus 76% Caucasian, x2(2, N = 78) = 4.17, p = .12, or type of accident experienced, x2(3,
N = 78) = 0.35, p = .95. However, there was a marginally significant difference in self-rated
injury severity in that participants with travel phobia rated their injuries as more severe than those
without the disorder, t(73) = 1.95, p = .06.
Of the 40 participants with travel phobia, 5 participants also fulfilled criteria for posttraumatic
stress disorder (PTSD) and 14 participants fulfilled criteria for both PTSD and a current major
depressive episode. Only one participant in the group without travel phobia met criteria for PTSD
and no participant in this group fulfilled DSM-IV criteria for major depression.

3.1.2. Measures and procedure

Participants filled in the 20-item Driving Cognitions Questionnaire and the Travel Phobia
Questionnaire (described below) for the purposes of this study. Some other measures that were
part of the original studies were included in the analyses: Three scales represented other cognitive
measures that were expected to correlate with the DCQ. One scale measured safety behaviors
(i.e., excessive precautions) that people may develop after a traffic accident. Two scales measured
symptoms of posttraumatic stress disorder and depression.
Travel Phobia Questionnaire (TPQ). The TPQ (Ehring et al., 2006) was used to assess fear and
avoidance of travel. The TPQ closely follows DSM-IV criteria for specific phobia. The main part
of the questionnaire consists of 12 items enquiring about fear experienced in travel situations
(e.g., I am very afraid of driving/traveling as a passenger/riding a bike), items related to the
DSM-IV criteria of insight (I am more afraid of driving/traveling as a passenger/riding a bike
than I should be) and interference (My fear of driving/traveling as a passenger/riding a bike
interferes with my life, e.g., work, relationships, free time activities), and the degree of avoidance
related to travel. In earlier studies, the scale has shown high internal consistency (a = .94) as well
as good re-test reliability over an interval of two weeks (n1 = 32, r = .76; n2 = 104, r = .85). In
addition, participants with a SCID diagnosis of travel phobia showed significantly higher scores
on the TPQ (M = 42.83, S.D. = 9.66) than participants without the disorder (M = 25.37,
S.D. = 9.96); t(98) = 8.18, p < .001), supporting the validity of the measure.
 A. Ehlers et al. / Journal of Anxiety Disorders 21 (2007) 493509 503

Travel Phobia Beliefs Questionnaire (TPBQ). The TPBQ was developed by Ehring et al.
(2006) to assess dysfunctional beliefs about travel in road traffic accident survivors. The first part
of the questionnaire measuring concerns related to travel was included in the present analyses.
The scale consists of 10 items (a = .87) measuring concerns related to another accident
happening (e.g., I will get injured in an accident) as well as panic-related concerns (I am going to
pass out) and social concerns (I will become hysterical). An initial evaluation of the measure has
shown good psychometric properties (Ehring et al., 2006).
Posttraumatic Cognitions Inventory (PTCI). The PTCI is a 33-item scale comprising negative
cognitions that are characteristic for patients with PTSD. The measure has demonstrated good
internal consistency (a = .97), re-test reliability (r = .74.86) and concurrent validity with other
cognition measures (Foa et al., 1999). It discriminates well between traumatized people with and
without PTSD.
Interpretation of PTSD Symptoms Inventory (IPSI). This questionnaire assesses the extent to
which participants interpret posttraumatic symptoms in a negative way. This scale was developed
by Dunmore, Ehlers, and Clark (1999, 2001) and has been shown to have good reliability
(a = .84) and predictive validity in assault survivors (11 items such as My reactions since the
event show I must be losing my mind).
Safety Behaviors Questionnaire-travel (SBQ-t). The 14-item travel subscale of the SBQ was
developed in a series of studies (Ehlers et al., 2003; Ehring et al., 2006) and assesses excessive
precautions related to travel (14 items, e.g. I keep checking the position of other traffic; a = .87).
It has demonstrated good internal consistency (a = .87) and predictive validity.
Posttraumatic Diagnostic Scale (PDS). The PDS (Foa, Cashman, Jaycox, & Perry, 1997) is a
standardized and validated self-report measure of PTSD symptom severity. Participants are asked
to rate each of the 17 PTSD symptoms according to the DSM-IV on a scale from 0 (Not at all or
only one time) to 3 (5 or more times a week/almost always). The PDS has widely been used in
samples of trauma survivors and has demonstrated high reliability (a = .92, retest reliability
r = .83) and validity (82% agreement with a SCID diagnosis of PTSD).
Beck Depression Inventory (BDI). As in Studies 1 and 2, symptom levels of depression were
assessed using the BDI (Beck et al., 1961).

3.2. Results

3.2.1. Internal consistency

The internal consistencies were a = .95 for the total scale of the DCQ, a = .91 for panic-
related concerns, a = .87 for accident-related concerns, and a = .86 for social concerns. The
subscales showed high intercorrelations.

3.3. Validation

As shown in Table 2, subscales of the DCQ were significantly correlated with fear and avoidance
of travel (TPQ). Participants with a diagnosis of travel phobia showed significantly higher scores on
all DCQ scales than participants without the disorder (see Table 3). A logistic regression analysis
showed that the DCQ total scale significantly predicted diagnostic status, Wald (1, 78) = 14.79,
p < .001, x2 = 26.94, Nagelkerke R2 = .39. The total percentage of correct classifications was 76%,
with 70% of the travel phobia group and 82% of the control group correctly classified.
Table 4 shows correlations between the DCQ scales and cognition measures, safety
behaviors and symptom measures. The total scale and the subscales correlated significantly
504 A. Ehlers et al. / Journal of Anxiety Disorders 21 (2007) 493509

with travel-related concerns (TPBQ) and measures of negative posttraumatic cognitions

(PTCI, IPSI). This raises the question of whether the correlation of the DCQ with the severity
of travel phobia is due to its correlation with the PTCI or IPSI. This was not the case. The
partial correlations of the DCQ total score with TPBQ-Anxiety and TPBQ-Avoidance were
r = .56 and r = .58, respectively, when PTCI and IPSI scores were partialled out. As expected,
the DCQ also correlated with excessive precautions related to travel (SBQ-t). All DCQ scales
showed significant correlations with symptom levels of PTSD and depression.

4. Discussion

This paper described the development and initial psychometric evaluation of the Driving
Cognitions Questionnaire (DCQ). The scale measures a range of cognitions that may play a role
in driving phobia. The 20-item measure showed excellent internal consistencies in three samples
in different countries. In all studies, the DCQ discriminated well between participants with and
without driving/travel phobia. In Studies 1 and 3, the DCQ classified 78 and 76% of the
participants correctly into the phobic and nonphobic diagnostic groups, as determined by the
SCID. In further support of its validity, the DCQ further showed substantial correlations with
measures of the severity of driving or travel phobia in all studies (correlations between r = .58
and r = .70).
As expected, the DCQ also correlated with measures of agoraphobic avoidance (MI in
Study 1; FQ-Ag in Study 2), and these correlations were lower (correlations between r = .38
and .43) than those between the DCQ and driving phobia measures. The panic-related concern
subscale showed the highest correlations with agoraphobic avoidance. Furthermore, the
social-concerns subscale correlated with the social phobia subscale of the FQ. Thus, the DCQ
and its subscales showed the expected relationships with measures of specific and general
avoidance.
In support of convergent validity, the DCQ showed moderate correlations in Studies 1 and
3 with other measures that have been developed to assess typical cognitions of patients with
panic disorder or PTSD (ACQ, BSQ, TPBQ, PTCI, and IPSI). Importantly, however, the
correlations of the DCQ with the severity of driving/travel phobia remained significant when
the other cognition scales were partialled out, suggesting that the DCQ explains variance over
and above what can be explained with other cognition measures that are not specific to
driving.
Factor analyses of the DCQ suggested a three-factor structure of panic-related, accident-
related, and social concerns. These analyses remain preliminary as sample sizes were relatively
small (Hair, Anderson, Tatham, & Black, 1998). Overall, the relative importance of each domain
of concerns may vary across samples. Panic-related concerns were prominent and emerged as
the strongest factor in Study 1, which recruited a clinical sample of patients with driving phobia,
a majority of whom also had panic attacks. Social concerns comprised the strongest factor in
Study 2, which included less disabled participants, 54% of whom did not have symptoms severe
enough to meet diagnostic criteria for driving phobia. Emergence of social concerns as an
important factor in non-clinical driving fear is consistent with previous research (Taylor &
Deane, 2000) and may be linked to the increasing attention to road rage in the media and
professional psychological literature (e.g., Galovski & Blanchard, 2004; Gaulton, 1997;
Wright, Gaulton, & Miller, 1997). The potential links between social concerns in driving fear
and worry about road rage requires future research. Finally, means suggested that in accident
survivors, accident-related concerns were elevated compared to the other two samples. Overall,
 A. Ehlers et al. / Journal of Anxiety Disorders 21 (2007) 493509 505

the difference in pattern of results appears to reflect the composition of the different samples in a
meaningful way.
The studies had several limitations. First, sample sizes were modest. The analyses, especially
the factor analyses, should thus be repeated using larger samples of people with driving phobia.
Second, the subscales were highly correlated, and future research may indicate that the
questionnaire can be shortened further without compromising reliability and validity. Third, in
Study 3 the DCQ was given on a separate occasion from the other questionnaires. The reported
correlations may thus reflect underestimates of the true correlations. Fourth, in patients with
agoraphobia or PTSD avoidance of driving may be part of their widespread avoidance of a
multitude of situations. It will need to be determined whether the DCQ adds to the explanation of
driving fear over and above what can be explained by other disorder-specific cognition scales in
these patient groups. Our sample sizes were not large enough to address this question. However,
our finding that the correlations between the DCQ and the severity of driving/travel phobia
remained significant when panic- and PTSD-related cognitions were partialled out suggests that
the DCQ may be useful in assessing a patients particular problems with driving regardless of
their other concerns and avoidance.
In conclusion, the DCQ shows promise as a measure of cognitions that may motivate people to
avoid driving. In a clinical context, the DCQ may be useful in assessing which cognitions are
most important for a particular patient. The focus of treatment would differ depending on
whether the patient mainly reports panic-related, accident, or social concerns. In a research
context, the DCQ may be of interest in assessing the mediating or moderating effects of
cognitions in treatment outcome studies.

Acknowledgements

The studies were supported by the Wellcome Trust, the German Academic Exchange Service,
and the Department of Veterans Affairs. We thank those who volunteered to participate in the
studies reported in this article.
506 A. Ehlers et al. / Journal of Anxiety Disorders 21 (2007) 493509

Appendix A
 A. Ehlers et al. / Journal of Anxiety Disorders 21 (2007) 493509 507

Appendix B

References

American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). WA: American
Psychiatric Association.
Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Beck depression inventory-second edition (BDI-II) manual. San
Antonio, USA: The Psychological Corporation.
Beck, A. T., Steer, R. A., & Garbin, M. G. (1988). Psychometric properties of the Beck Depression Inventory: twenty-five
years of evaluation. Clinical Psychology Review, 8, 77100.
Beck, A. T., Ward, C. H., Mendelsohn, M., Mock, J., & Erbaugh, J. (1961). An inventory for measuring depression.
Archives of General Psychiatry, 41, 561571.
Beckham, E. E., & Leber, W. R. (Eds.). (1995). The handbook of depression (2nd ed.). New York: Guilford Press.
Blanchard, E. B., & Hickling, E. J. (1997). After the crash: assessment and treatment of motor vehicle accident survivors.
WA: American Psychological Association.
Chambless, D. L., Caputo, G. C., Bright, P., & Gallagher, R. (1984). Assessment of fear of fear in agoraphobics: the Body
Sensations Questionnaire and the Agoraphobic Cognitions Questionnaire. Journal of Consulting and Clinical
Psychology, 52, 10901097.
Chambless, D. L., Caputo, G. C., Jasin, S. E., Gracely, E. J., & Williams, C. (1985). The Mobility Inventory for
agoraphobia. Behaviour Research and Therapy, 23, 3544.
508 A. Ehlers et al. / Journal of Anxiety Disorders 21 (2007) 493509

Cox, B. J., Swinson, R. P., Kuch, K., & Reichman, J. T. (1993). Dimensions of agoraphobia assessed by the Mobility
Inventory. Behaviour Research and Therapy, 31, 427431.
Craske, M. G., Rachman, S. J., & Tallman, K. (1986). Mobility, cognitions, and panic. Journal of Psychopathology and
Behavioral Assessment, 8, 199210.
Craske, M. G., & Rowe, M. K. (1997). A comparison of behavioral and cognitive treatments of phobias. In: G. C. L.
Davey (Ed.), Phobias: a handbook of theory, research and treatment (pp. 247280). Chichester, UK: Wiley.
Dunmore, E., Clark, D. M., & Ehlers, A. (1999). Cognitive factors involved in the onset and maintenance of posttraumatic
stress disorder (PTSD) after physical and sexual assault. Behaviour Research and Therapy, 37, 809829.
Dunmore, E., Clark, D. M., & Ehlers, A. (2001). A prospective study of the role of cognitive factors in persistent
posttraumatic stress disorder after physical or sexual assault. Behaviour Research and Therapy, 39, 10631984.
Ehlers, A., & Clark, D. M. (2000). A cognitive model of posttraumatic stress disorder. Behaviour Research and Therapy,
38, 319345.
Ehlers, A., Clark, D. M., Hackmann, A., McManus, F., Fennell, M., Herbert, C., et al. (2003). A randomized controlled
trial of cognitive therapy, a self-help booklet, and repeated assessments as early interventions for posttraumatic stress
disorder. Archives of General Psychiatry, 60, 10241032.
Ehlers, A., Hofmann, S. G., Herda, C. A., & Roth, W. T. (1994). Clinical characteristics of driving phobia. Journal of
Anxiety Disorders, 8, 323339.
Ehring, T., Ehlers, A., & Glucksman, E. (2006). Contribution of cognitive factors to the prediction of posttraumatic stress
disorder, phobia and depression after road traffic accidents. Behaviour Research and Therapy, 44, 16991716.
Ehring, T., Ehlers, A., & Glucksman, E. (in preparation). A prospective longitudinal study of posttraumatic stress disorder,
phobia, and depression after road traffic accidents.
First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (1996). Structured clinical interview for DSM-IV axis I
disorders. Washington, DC: American Psychiatric Press.
Foa, E. B., Cashman, L., Jaycox, L., & Perry, K. (1997). The validation of a self-report measure of posttraumatic stress
disorder: the Posttraumatic Diagnostic Scale. Psychological Assessment, 9, 445451.
Foa, E. B., Ehlers, A., Clark, D. M., Tolin, D. F., & Orsillo, S. M. (1999). The Posttraumatic Cognitions Inventory (PTCI):
development and validation. Psychological Assessment, 11, 303314.
Galovski, T. E., & Blanchard, E. B. (2004). Road rage: a domain for psychological intervention? Aggression and Violent
Behavior, 9, 105127.
Gaulton, R. E. (1997). Road rage: a secondary study. Wellington, New Zealand: New Zealand Police.
Hair, J. F., Jr., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate data analysis (5th ed.). NJ: Prentice-Hall.
Hofmann, S.G. (1992). A critique on the DSM-III-R classification of situational phobias: studies on speech phobia and
driving phobia. Unpublished doctoral dissertation. Germany: University of Marburg.
Kwon, S.-M., Evans, L., & Oei, T. P. S. (1990). Factor structure of the Mobility Inventory for agoraphobia: a validational
study with Australian samples of agoraphobic patients. Journal of Psychopathology and Behavioral Assessment, 12,
365374.
Lee, H. B., & Oei, T. P. S. (1994). Factor structure, validity, and reliability of the Fear Questionnaire in a Hong Kong
Chinese population. Journal of Psychopathology and Behavioral Assessment, 16, 189199.
Marks, I. M., & Mathews, A. M. (1979). Brief standard self-rating for phobic patients. Behaviour Research and Therapy,
17, 263267.
Mathew, R. J., Weinman, M. L., Semchuk, K. M., & Levin, B. L. (1982). Driving phobia in the city of Houston: a pilot
study. American Journal of Psychiatry, 139, 10491051.
Moylan, A., & Oei, T. P. S. (1992). Is the Fear Questionnaire (FQ) a useful instrument for patients with anxiety disorders?
Behaviour Change, 9, 3849.
Munjack, D. J. (1984). The onset of driving phobias. Journal of Behavior Therapy and Experimental Psychiatry, 15, 305
308.
Ninan, P. T., & Berger, J. (2001). Symptomatic and syndromal anxiety and depression. Depression and Anxiety, 14, 7985.
Oei, T. P. S., Moylan, A., & Evans, L. (1991). Validity and clinical utility of the Fear Questionnaire for anxiety-disorder
patients. Psychological Assessment: A Journal of Consulting and Clinical Psychology, 3, 391397.
Spielberger, C. D., Gorsuch, R. L., Lushene, R. E., Vagg, P. R., & Jacobs, G. A. (1983). Manual for the State-Trait Anxiety
Inventory. Palo Alto, CA: Consulting Psychologists Press.
SPSS Inc. (2004). SPSS for Windows (Release 13.0.1). Chicago: SPSS Inc.
Taylor, J.E. (2002). Understanding driving-related fear. Unpublished doctoral dissertation. Palmerston North, New
Zealand: Massey University.
Taylor, J. E., & Deane, F. P. (2000). Comparison and characteristics of motor vehicle accident (MVA) and non-MVA
driving fears. Journal of Anxiety Disorders, 14, 281298.
 A. Ehlers et al. / Journal of Anxiety Disorders 21 (2007) 493509 509

Taylor, J. E., Deane, F. P., & Podd, J. V. (2000). Determining the focus of driving fears. Journal of Anxiety Disorders, 14,
453470.
Turner, S. M., Johnson, M. R., Beidel, D. C., Heiser, N. A., & Lydiard, R. B. (2003). The Social Thoughts and Beliefs
Scale: a new inventory for assessing cognitions in social phobia. Psychological Assessment, 15, 384391.
World Health Organization (1997). CIDI-Auto 2.1: administrators guide and reference. Sydney, Australia: Training and
Reference Centre for WHO CIDI.
Wright, P. G., Gaulton, R. E., & Miller, I. (1997). Road rage: an exploratory study. Wellington, New Zealand: New
Zealand Police.