Journal of Consulting and Clinical Psychology Copyright 2000 by the American Psychological Association, Inc.

2000, Vol. 68, No. 6, 1020-1026 0022-006X/00/S5.00 DOI: 10.1037//0022-006X.68.6.1020

A Controlled Study of Virtual Reality Exposure Therapy

for the Fear of Flying

Barbara Olasov Rothbaum Larry Hodges

Emory University School of Medicine College of Computing, Georgia Institute of Technology

Samantha Smith Jeong Hwan Lee

Emory University School of Medicine and Virtually Better, Inc. College of Computing, Georgia Institute of Technology
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Larry Price
This document is copyrighted by the American Psychological Association or one of its allied publishers.

The Psychological Corporation

Fear of flying (FOF) affects an estimated 10-25% of the population. Patients with FOF (N = 49) were
randomly assigned to virtual reality exposure (VRE) therapy, standard exposure (SE) therapy, or a
wait-list (WL) control. Treatment consisted of 8 sessions over 6 weeks, with 4 sessions of anxiety
management training followed by either exposure to a virtual airplane (VRE) or exposure to an actual
airplane at the airport (SE). A posttreatment flight on a commercial airline measured participants'
willingness to fly and anxiety during flight immediately after treatment. The results indicated that VRE
and SE were both superior to WL, with no differences between VRE and SE. The gains observed in
treatment were maintained at a 6-month follow up. By 6 months posttreatment, 93% of VRE participants
and 93% of SE participants had flown. VRE therapy and SE therapy for treatment of FOF were
unequivocally supported in this controlled study.

Fear of flying (FOF) is a significant problem, affecting an ing, implosion, and relaxation treatments (Beckham, Vrana, May,
estimated 10% to 25% of the population (Agras, Sylvester, & Gustafson, & Smith, 1990; Haug, Brenne, Johnson, Berntzen,
Oliveau, 1969; Deran & Whitaker, 1980), or approximately 25 Gotestam, & Hugdahl, 1987; Howard, Murphy, & Clarke, 1983),
million adults in the United States (Deran & Whitaker, 1980). In but many have been criticized for not including a posttreatment
addition, approximately 20% of those who do fly depend on flight to evaluate outcome (Haug et al., 1987). The difficulty and
alcohol or sedatives during flights (Greist & Greist, 1981). Avoid- expense of using actual airplanes and flights for exposure have
ance of flying causes sufferers serious vocational and social daunted many researchers and therapists despite the prevalence
consequences. and impact of FOF. Some FOF programs exist in large metropol-
Several FOF programs have been described and tested, includ- itan cities, often sponsored by airlines, but these programs have not
ing stress inoculation training, systematic desensitization, flood- been subjected to rigorous evaluation.
Virtual reality (VR) offers a new human-computer interaction
paradigm in which users are no longer simply external observers of
images on a computer screen but are active participants within a
Barbara Olasov Rothbaum, Department of Psychiatry and Behavioral computer-generated three-dimensional virtual world. Virtual envi-
Sciences, Emory University School of Medicine; Larry Hodges and Jeong
ronments differ from traditional displays in that computer graphics
Hwan Lee, College of Computing, Georgia Institute of Technology; Sa-
mantha Smith, Department of Psychiatry and Behavioral Sciences, Emory
and various display and input technologies are integrated to give
University School of Medicine, and Virtually Better, Inc., Atlanta, Geor- the user a sense of presence or immersion in the virtual environ-
gia; Larry Price, The Psychological Corporation, San Antonio, Texas. ment. The most common approach to the creation of a virtual
This study was supported by National Institute of Mental Health Grant environment is to outfit the user in a head-mounted display.
1-R41-MH58493-01. We thank Edna Foa for consultation during this Head-mounted displays consist of separate display screens for
research. each eye, along with some type of display optics, stereo earphones,
Barbara Olasov Rothbaum and Larry Hodges received research funding and a head-tracking device. The user is presented with a computer-
and were entitled to sales royalty from Virtually Better, Inc., which generated view of a virtual world that changes in a natural way
developed products related to the research described in this article. In with head and body motion.
addition, the investigators served as consultants to and owned equity in
What distinguishes VR from a mere multimedia system or an
Virtually Better, Inc. The terms of this arrangement have been reviewed
and approved by Emory University and the Georgia Institute of Technol-
interactive computer graphics display is a sense of presence. A
ogy in accordance with its conflict of interest policies. sense of presence is also essential to conducting exposure therapy.
Correspondence concerning this study should be addressed to Barbara Researchers of emotional processing theory as it is applied to
Olasov Rothbaum, Department of Psychiatry and Behavioral Sciences, The anxiety disorders (Foa & Kozak, 1986; Foa, Steketee, & Roth-
Emory Clinic, 1365 Clifton Road, Atlanta, Georgia 30322. baum, 1989) purport that fear memories can be construed as
1020
 VIRTUAL REALITY EXPOSURE THERAPY 1021

structures that contain information regarding stimuli, responses, et al., 1983). Eligible patients on psychoactive medication must have been
and meaning. Therapy is aimed at facilitating emotional process- on that dose and medication for at least 3 months and had to agree to
ing. It has been proposed that, for this to occur, the fear structure remain on that dose throughout the project. Participants with a history of
must be activated and modified. Exposure therapy, in which the mania, schizophrenia, or other psychoses; with prominent suicidal ideation;
patient is intentionally confronted with the feared stimuli in a or with current alcohol or drug abuse or dependence were excluded.
Analyses of variance (for age, years of education, and income) and
therapeutic manner, activates the fear structure through confron- chi-square analyses (for gender, race, marital status, and primary diagnosis)
tation with the feared stimuli, which elicits the fearful responses. revealed no significant differences between groups on any demographic
The processes of habituation and extinction in which the feared characteristic or pretreatment variable (p > .05). Participants ranged in age
stimuli cease to elicit anxiety aid modification of the fear structure, from 24 to 69, with a mean age of 40.5. The sample was relatively wealthy
making its meaning less threatening. Any method capable of and well educated, with a mean level of 16.5 years of education and with
activating the fear structure and modifying it would be predicted to most participants (64%, n = 29) reporting household incomes of $50,000
improve symptoms of anxiety. Thus, VR exposure therapy (VRE) or more. The majority of participants were female (71%, n = 32), Cauca-
has been proposed as a new medium for exposure therapy. sian (82%, n = 37), and married (71%, n = 32).
Rothbaum et al. (1995) conducted the first controlled study The majority of participants received a primary diagnosis of specific
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

phobia, situational type (flying; 93%, n = 42). The remaining participants

This document is copyrighted by the American Psychological Association or one of its allied publishers.

applying VR to the treatment of a psychological disorder: VRE

received a primary diagnosis of panic disorder with agoraphobia (7%, n =
was incorporated in the treatment of acrophobia. Participants were 3). Most participants received only one current diagnosis (67%, n = 30);
repeatedly exposed to virtual foot bridges of varying heights and however, 24% (n = 11) received two diagnoses and 9% (n = 4) received
stability, outdoor balconies of varying heights, and a glass elevator three diagnoses. Specific phobia of heights was the most common second-
that ascended 50 floors. VRE was effective in significantly reduc- ary or tertiary diagnosis (occurring in 53% of those with more man one
ing fear of and improving attitudes toward heights, whereas no diagnosis, n = 8). Consistent with diagnoses of specific phobia and
change was noted in the control group. Seven out of the ten VRE agoraphobia, at screening all participants reported either avoiding flying
treatment completers exposed themselves to height situations in entirely or enduring flying with great discomfort. Interrater reliability of
real life during treatment although they were not instructed to do. diagnostic judgments was calculated for the three clinician raters. Mean
Physical symptoms of anxiety described by the participants while overall diagnostic agreement was 92% (range = 87-96%).
Four individuals (3 VRE and 1 SE) withdrew prematurely from the
in virtual height situations included sweating, abdominal discom-
study: Two dropped out after the initial formal assessment but prior to the
fort usually described as "butterflies," loss of balance or light- first session, 1 dropped out after one session because of scheduling prob-
headedness, heart palpitations, pacing, tremulousness or shaking, lems and uncertainty about the benefits from treatment, and the 4th was
feeling nervous or scared, weakness in the knees, tightness in the withdrawn from the study by the investigators after two sessions because
chest, and feeling tense (Hodges et al., 1995). VRE has also been of protocol violations (change in psychotropic medication). Analyses of
effective in two case studies of patients with EOF (Rothbaum, variance indicated that the dropouts as a group were not significantly
Hodges, Watson, Kessler, & Opdyke, 1996; Smith, Rothbaum, & different from completers with two exceptions: They reported higher levels
Hodges, 1999) and in a case study treating a person with spider of state anxiety (p < .05) than had completers at pretreatment and reported
phobia (Carlin, Hoffman, & Weghorst, 1997). lower income levels (p < .05). Demographic characteristics of completers
VRE is potentially an efficient and cost-effective treatment of can be found in Table 1.
EOF. The current study sought to determine the relative efficacy of
VRE and standard exposure (SE) compared with wait list (WL) Measures
control in the treatment of FOF. It was predicted that (a) VRE
would be more effective than WL control in reducing participants' The Questionnaire on Attitudes Toward Flying (QAF; Howard et al.,
FOF and avoidance behavior and (b) VRE and SE would be 1983) assesses history of FOF, previous treatment, and attitudes toward
equally effective in reducing participants' FOF and avoidance flying. It includes a 36-item questionnaire rating the level of fear on an
behavior. 11-point scale ranging from 0 (no fear) to 10 (extreme fear) in different
flying situations. The possible range of scores is 0 to 360. Test-retest
reliability was .92, and split-half reliability was .99. The QAF Fear item is
Method a subscale of this measure and asks the participant to rate, using a
Likert-type scale ranging from 0 to 10, his or her present FOF. Reports of
Forty-nine patients with FOF were randomly assigned to one of three the total number of anxiety symptoms the participant typically experienced
groups: VRE, SE, or WL. Forty-five patients completed, resulting in 15 when flying (Fear of Flying Interview [FOFI]; McNally & Louro, 1992)
completers per group. In addition to standard paper and pencil measures were significantly correlated with both the QAF total score (r = .32, p <
pre- and posttreatment and at follow-ups of 6- and 12-months posttreat- .05) and the QAF Fear item (r = .52, p < .001). The FOFI assesses
ment, a posttreatment test flight was conducted to assess participants' whether the fear of flying is a symptom of agoraphobia or a simple phobia.
anxiety and avoidance for an actual airplane flight. All assessments were It has been found to discriminate between agoraphobic and simple phobic
conducted by an independent assessor. patients with a fear of flying.
The Structured Clinical Interview for the DSM-JV (SCID; First, Spitzer,
Participants Gibbon, & Williams, 1995) was given to screen Axis I disorders as well as
establish comorbid diagnoses.
To participate in this project, all participants had to meet current Diag- The Clinical Global Improvement (CGI) Scale is a global measure of
nostic and Statistical Manual of Mental Disorders (4th ed.; American change in severity of symptoms. The scale ranges from 1 (very much
Psychiatric Association, 1994) criteria for either specific phobia, situa- improved) to 4 (no change) to 7 (very much worse). It has been used
tional type, panic disorder with agoraphobia, or agoraphobia without a extensively in clinical trials (Guy, 1976).
history of panic disorder with flying as the feared stimulus. Only partici- The Fear of Hying Inventory (FFI; Scott, 1987) is a 33-item scale
pants who had flown before were eligible (Beckham et al., 1990; Howard measuring intensity of FOF. Items are rated on a 9-point scale ranging
 1022 ROTHBAUM, HODGES, SMITH, LEE, AND PRICE

Table 1 were administered. The posttreatment assessment was conducted individ-

Demographic Characteristics of Sample ually after 6 weeks for all participants. WL participants were assessed at
the same pre- and posttreatment points in time as the treatment participants
Variable M SD but did not have any treatment in between. They were scheduled for their
flights at the same point in time as the treated participants. For ethical
Age (years) 40.55 10.64 reasons, WL participants were offered the treatment of their choice (VRE
Education (years) 16.55 2.22 or SE) following the flight. Six- and 12-month follow-up assessments were
Income
conducted using the same measures as at the posttreatment assessment. All
Over $50,000 64 29
$30,000-50,000 22 10 patients were asked to participate in a posttreatment or post-WL behavioral
$20,000-30,000 9 4 avoidance test consisting of an actual round-trip flight accompanied by the
$10,000-20,000 4 2 therapist. No one was forced or coerced, however. Group flights of ap-
Race proximately 5 participants and one therapist each were scheduled on Delta
Caucasian 82 37 Airlines between Atlanta and Houston, approximately 1.5 hours per flight.
Hispanic 7 3 Patients were asked to pay $164 each for their flights to (a) increase
Caucasian/Hispanic 4 2 motivation, (b) offset study costs, (c) ensure that people didn't sign up for
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

African American 4 2 a free flight, and (d) more closely match the real world. Delta agreed to
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Native American 2 1
provide full refunds if flights were not taken. Group flights were used to
Marital status
Married 71 32 offset study costs. It was prohibitively expensive for the therapist to
Single 16 7 accompany 45 patients on individual flights.
Divorced 9 4
Living with partner 4 2
Gender Treatment
Female 71 32
Male 29 13 Participants were treated for eight individual sessions over 6 weeks.
Primary diagnosis VRE and SE groups received identical treatment for Sessions 1 to 4 lasting
Specific phobia of flying 93 42 approximately 1 hr each. The first session for both VRE and SE lasted
Panic disorder with agoraphobia 7 3 approximately 90 min and was spent in information gathering, treatment
Other diagnoses
planning, and explaining the rationale to the patient. Brief breathing
Specific phobia of heights 18 8
Social phobia 11 5 retraining was taught to manage physical signs of anxiety such as increased
Major depression 4 2 heart rate. Cognitive restructuring to learn how to challenge irrational
Dysthymia 4 2 thoughts was taught in the second treatment session. Typical irrational
Claustrophobia 2 1 thoughts included "This plane is going to crash" and "I will panic on this
Specific phobia of snakes 2 1 plane and [embarrass myself/go crazy/die/have a heart attack]." Thought-
stopping to counter ruminative thinking was taught in Session 3. Session 4
reviewed these three anxiety management techniques, spent more time on
cognitive restructuring, and taught hyperventilation exposure if the patient
from 0 (not at all) to 8 (very severely disturbing). Test-retest reliability reported a history of panic attacks. Exposure was conducted in Sessions 5
for 15 WL patients was .92, and it has been sensitive to change with to 8.
treatment. Reports of the total number of anxiety symptoms the participant VRE of sitting in an airplane, experiencing take-offs and landings in an
typically experienced when flying (Fear of Hying Interview) were signif- airplane, and flying in both calm and stormy weather were provided twice
icantly correlated with FFI score (r = .45, p < .01). weekly for 2 weeks in Virtually Better, Inc.'s office, according to a
treatment manual developed by the authors (Rothbaum & Hodges, 1997).
Apparatus The patient chair is equipped with a woofer under the seat to add noise and
vibrations at appropriate times in the flight. Recorded sounds of flight
The computer system used in the current study consisted of a 300 MHz attendants, takeoffs, landings, and weather effects on the outside of the
Pentium II processor with 128 MB memory, a SCSI disk drive, and a Fire airplane were included. The therapist made appropriate comments and
GL 1000 video card. A Virtual Research VR6 (Virtual Research Systems, encouraged continued exposure until anxiety decreased. The patient was
Santa Clara, CA) head-mounted display with stereo earphones transmitted allowed to progress at his or her own pace in the preset order of the
the VR image to the participant. The participant was seated in a Thunder- hierarchy of exposures: sitting on plane, engines off; sitting on plane,
seat™ (Thunderseat, Los Angeles, CA), a specially designed seat with an engines on; taxiing; take-off; smooth flight; landing; and thunderstorm and
embedded 100-watt sub woofer and an attached airplane seatbelt. The turbulent flight.
virtual airplane software for FOF was created by Virtually Better, Inc. SE was conducted at the airport. Because of the time required for travel
(Atlanta, GA; www.virtuallybetter.com). VR-generated scenes placed par- to and from the airport, parking, and in vivo exposure, Sessions 5 and 6
ticipants in a passenger seat by the window on a commercial airplane. As were combined into one extended session and were spent at the airport
they moved their heads to the left, they were able to see out the left-hand exposing patients to the preflight stimuli (e.g., ticketing, trains, parked
window. As they moved their heads to the right, the empty seats to the right planes, and waiting area). Sessions 7 and 8, also combined into one
side of the airplane and the right side window came into view. extended session, were spent on a stationary airplane to allow habituation
to airplane stimuli and time for imaginal exposure (i.e., imagining takeoffs,
Procedure cruising, landing, etc.).
Exposure components could not be made 100% equal between VRE and
If appropriate following phone screening, an initial evaluation was SE because of real world constraints, so they were arranged to naturalis-
scheduled to evaluate patients as to inclusion and exclusion criteria and tically match what would occur before an actual flight. In VRE, this was
explain the procedures of the project in detail. The patients signed a the above exposure hierarchy. In SE, this was going to the airport and
consent form at this time and the SCID was administered. Following entry sitting on a stationary airplane. The time spent in exposure and with the
into the study, the pretreatment assessment was conducted and all measures therapist was equal.
 VIRTUAL REALITY EXPOSURE THERAPY 1023

Results conditions at pre- and posttreatment on the FFI, QAF, and QAF
Fear.
Statistical Analysis
We examined the FFI and QAF data descriptively by treatment Repeated Measures Analysis Results for QAF Fear
and control groups for normality by using skewness and kurtosis
coefficients (z tests of greater or less than 1.96) and the Shapiro- The overall F test for the QAF Fear Pretest-Posttest X Group
Wilks test where indicated. Homogeneity of variance was exam- analysis was F(4, 66) = 10.06, p < .001, with an effect size of .38
ined across the treatment and control (factor) groups by the (de- standard deviation units. Tests of simple effects yielded a signif-
pendent) variables by using the Levene test (a = .05) for icant reduction in QAF Fear measures for the VRE and SE groups
univariate homogeniety of variance. Additionally, the Box test when compared with the WL group. Specifically, significant mean
(a — .05) was performed to determine the equality of the covari- differences were found between the WL and VRE groups, F(l,
ance matrices between treatment groups in the multivariate re- 25) = 25.00, p < .001, and between the WL and SE groups, F(l,
peated measures analysis models. The requisite assumptions for 25) = 16.43, p < .001, with effect sizes of .70 and .62,
respectively.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

using multivariate normal parametric statistical techniques were

This document is copyrighted by the American Psychological Association or one of its allied publishers.

met in almost all analyses. In instances in which nonnormality,

heterogeneity of variance, or both were present in the data, we Repeated Measures Analysis Results for FFI
performed transformations to use parametric statistical techniques.
After data screening for general linear model assumptions, three The overall F test for the FFI Pretest-Posttest X Group analysis
K-group, split-plot multivariate repeated measures analyses, across was F(4, 72) = 3.27, p < .05, with an effect size of .15 standard
two measures (pretest and posttest), were performed on the depen- deviation units. Tests of simple effects demonstrated a significant
dent variables FFI, QAF, and QAF Fear by WL control, VRE, and reduction in FFI scores for the VRE group compared with the WL
SB. If significant, one-tailed analyses of variance were conducted control group. Specifically, a significant difference was noted
on individual variables (Cohen, 1988; Maxwell & Delaney, 1990; between WL and VRE groups, F(l, 25) = 3.62, p < .05, with an
Stevens, 1996; SPSS, 1998). effect size of .23 standard deviation units.

Repeated Measures Analysis Results for QAF Repeated Measures Analyses for the FFI, QAF, and QAF
Fear for 6-Month Follow-Up Data
The overall F test for the QAF Pretest-Posttest X Group anal-
ysis was F(4, 72) = 5.54, p < .01, with an effect size of 0.24 To assess the degree to which participants in each active therapy
standard deviation units. Tests of simple effects detected a signif- maintained or did not maintain treatment gains over the course of
icant reduction in QAF measures for both treatment groups when the 6-month follow-up period, additional analyses were performed.
compared with the WL group. Significant mean differences were Only VRE and SE groups were compared in these analyses as the
detected between the WL and SE groups, F(l, 25) = 6.64, p < .01, WL participants received treatment following their wait list period.
with an effect size of .37 and between the WL and VRE groups, For ethical reasons, WL participants were allowed to choose their
F(l, 25) = 9.75, p < .01, with an effect size of .45. Table 2 treatment, and thus were not randomly assigned and are not in-
contains means and standard deviations for participants in the three cluded in 6-month follow-up analyses here. Of the 15 WL partic-

Table 2
Means, Standard Deviations, and One-Way ANOVA Results for Effects of Treatment Type at
Pre- and Posttreatment and 6-Month Follow-Up on Fear of Flying

VRE (n = 15) SE (n = 15) WL (n = 15) ANOVA

Variable M SD M SD M SD F(l, 25) Tj2 Comparisons

FFI
Pretreatment 105.85 35.91 133.30 42.00 119.52 50.51
Posttreatment 86.14 37.40 87.53 42.30 118.80 61.00 3.60* 0.21 VRE< WL
6-month follow-up 75.35 27.08 62.23 32.31
QAF
Pretreatment 194.92 55.97 237.09 42.00 210.69 79.26
Posttreatment 121.39 51.62 132.46 67.91 192.07 77.64 9.80** 0.51 VRE< WL
6-month follow-up 107.28 41.67 103.15 51.26 6.61** 0.40 SE< WL
QAF Fear item
Pretreatment 7.57 2.21 8.38 2.21 7.40 2.52
Posttreatment 4.14 1.30 4.61 2.00 7.40 2.21 25.00*** 0.70 VRE< WL
6-month follow-up 3.93 1.32 3.70 2.12 16.43*** 0.62 SE<WL

Note. Tj2 = effect size. At 6-month follow-up, VRE and SE ns = 14 and 13, respectively. ANOVA = analysis
of variance; VRE = virtual reality exposure; SE = standard exposure; WL = wait list; FFI = Fear of Flying
Inventory; QAF = Questionnaire on Attitudes Toward Flying.
* p < . 0 5 . **p<.01. ***/>< .001.
 1024 ROTHBAUM, HODGES, SMITH, LEE, AND PRICE

ipants, 14 chose VRE and 1 chose SE. A repeated measures Table 3

analysis was conducted using the SE and VRE as the between- Means, Standard Deviations, and One-Way ANOVA for Effects
groups factor and the posttreatment FFI score to 6-month of Treatment Type at Posttreatment and 6-Month Follow-Up
follow-up FFI score as the within-subjects factor. No significant on Fear of Flying Measures
differences were detected on the FFI between the posttreatment
score and the 6-month follow-up score, indicating that as a group, VRE (n = 14) SE (n = 13)
treated participants maintained treatment gains over the 6-month Variable M SD M SD F(\, 25)
follow-up period. Paired t tests further revealed significant mean
differences between the post-FFI and 6-month follow-up FFI score FFI
for the SE group. The associated effect size for this finding was .40 Posttreatment 86.15 37.40 87.53 42.30
6-month follow-up 75.35 27.08 62.23 32.31 1.42*
standard deviation units, indicating a further reduction in FFI QAF
scores from the posttest FFI to 6-month follow-up FFI for the SE Posttreatment 121.39 51.62 132.46 68.01
group. For the VRE group, no significant differences were found 6-month follow-up 107.28 41.67 103.15 51.26 0.72*
between the posttreatment and 6-month follow-up FFI scores, QAF Fear item
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Posttreatment 4.14 1.29 4.61 2.00

This document is copyrighted by the American Psychological Association or one of its allied publishers.

providing evidence for a sustained treatment effect from posttest to

6-month follow-up 3.93 1.32 3.70 2.05 1.65*
6-month follow-up.
QAF scores were examined in the same manner. A repeated Note. ANOVA = analysis of variance; VRE = virtual reality exposure;
measures analysis was conducted using SE and VRE as the be- SE = standard exposure; FFI = Fear of Flying Inventory; QAF = Ques-
tween groups factor and the mean posttreatment QAF score to tionnaire on Attitudes Toward Flying.
*p< .05.
mean 6-month follow-up QAF score as the within-subjects factor.
No significant differences were detected between the VRE and SE
groups across the within-subjects QAF factor. Additionally, paired
t tests conducted separately by group revealed no significant mean (SUDs; range = 0-100, in which 0 = no anxiety and 100 =
differences between the post QAF and 6-month follow-up QAF maximum anxiety, i.e., anxiety) ratings taken during the actual
measures for either the VRE or the SE group, indicating that flights indicated no differences between participants who received
participants in both treatment groups maintained the gains made VRE and those who received SE. The average SUDs rating on the
during the treatment period on this measure. flight from Atlanta to Houston was 33.19 (SD =15.6) for VRE
Finally, the QAF Fear item was examined using the same and 33.88 (SD = 16.3) for SE. The average SUDS rating on the
procedures. A repeated measures analysis was conducted using the return flight from Houston to Atlanta was 28.73 (SD = 13.3) for
SE and VRE as the between-groups factor and the posttreatment VRE and 29.77 (SD = 18.3) for SE.
QAF Fear to 6-month follow-up QAF Fear as the within-subjects At the 6-month follow-up evaluation, 11 of 14 (79%) VRE
factor. No significant differences were detected between the VRE participants and 9 of 13 (69%) SE participants had flown since (but
and SE groups across the within-subjects QAF Fear factor. Addi- not including) the graduation flight. Overall, by the 6-month
tionally, paired t tests conducted separately by group revealed no follow-up, 14 of 15 (93%) VRE participants and 14 of 15 (93%)
significant mean differences between the posttest QAF Fear and SE participants had flown since the termination of treatment, either
6-month follow-up QAF Fear measures for either the VRE or SE on the graduation flight or otherwise.
group, again indicating that the participants in both treatment
groups maintained treatment gains over the follow-up period on Posttest Rating of Improvement Variables
this measure. Table 3 provides the means, standard deviations, and
The Kruskal-Wallis chi-square test was performed on a posttest
F ratios for the repeated measures analyses using posttest and
rating of self-improvement to determine whether there were dif-
6-month follow-up measures.
ferences between VRE and SE therapy groups on patient global
ratings of improvement. On the CGI, a significant difference
Kruskal-Wallis Chi-Square Test for Posttreatment Flight between the WL and VRE groups, ^(1, N = 29) = 23.31, p <
.001, and between the WL and SE groups, ^(l, N = 30) = 25.20,
The Kruskal-Wallis chi square was performed to determine if
p < .001, was detected such that participants who received either
there were differences between VRE, SE, and WL with respect to
VRE or SE improved significantly more than did WL participants,
the likelihood of persons in each group taking a flight at the end of
with no significant differences between VRE and SE.
the treatment or wait list period. The overall result of the test
detected significant differences between the groups, ^(2, N =
45) = 11.93, p < .01. After analysis of group comparisons, the Posttreatment Client Satisfaction Ratings
WL and VRE groups were found to differ significantly, ^(1, N = There were no significant differences between VRE and SE
30) = 7.52, p < .01, as did the WL and SE groups, )f(l, N = participants on ratings of client satisfaction, which were very
30) = 11.24, p < .01. In both cases, VRE and SE groups were positive.
more likely to take their actual graduation flight. Eight of 15 VRE
participants (53%), 10 of 15 SE participants (67%), and 1 of 15
Discussion
WL participants (7%) took their posttreatment graduation flights.
Logistic regression analysis revealed that after treatment, the SE In this controlled trial comparing VRE therapy, SE therapy, and
and VRE groups were approximately 3.5 times more likely to take a WL control for treatment of FOF, VRE and SE were shown to be
a flight than the WL control group. Subjective Units of Discomfort equally effective both in decreases in symptoms as measured by
 VIRTUAL REALITY EXPOSURE THERAPY 1025

standardized questionnaires and by the number of participants to accident or fear of driving due to panic disorder, agoraphobia, or
actually fly on a real airplane following treatment. Regarding specific phobia.
written measures and the number of participants to fly on a real Other limitations to this study include the fact that the posttreat-
airplane following treatment, VRE therapy was statistically indis- ment test flights were conducted in small groups rather than
tinguishable from SE therapy, in which participants were exposed individually. This was done to reduce costs, but is clearly not the
to an actual airplane. Anxiety ratings during the actual flight manner in which the majority of patients would fly in real life.
indicated that VRE-treated patients were as comfortable as SE- Also, because of the dearth of standardized measures assessing
treated patients. The participants rated themselves equally as im- POP, the measures, other than the SCID and the Fear of Flying
proved with VRE as SE and more so than following the WL, and Interview, were based on patients' self-report and were completed
their satisfaction with treatment ratings were no different for VRE by the patient. Although it was well controlled, the completer
and SE. When allowed to choose which treatment they would sample size (45) was small.
receive following WL, the overwhelming majority of participants The strengths of the current study center on the methodological
chose VRE, indicating a clear preference. The gains observed in rigor used. Participants were randomly assigned to treatment con-
dition. Two active treatments were compared, with SE therapy
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

treatment were maintained at a 6-month follow-up.

This document is copyrighted by the American Psychological Association or one of its allied publishers.

There may be advantages to conducting exposure therapy in VR being the current standard in the field and thus the measuring stick
rather than in vivo. Although in this study the time spent in SE and for a new treatment. Treatments were well delineated, standard-
VRE were kept the same, in clinical settings, SE can be very costly ized, brief, and easily replicated. Standard measures were used.
because it usually requires leaving the therapist's office and using Clear eligibility requirements were incorporated, including a DSM
extended sessions. Many insurance companies will not pay for diagnosis. The same therapist delivered both treatments, thus elim-
extended sessions. VRE can accomplish the same exposure within inating potential therapist effects. A major strength of the study
the standard therapy hour (usually 45-50 min) and within the was the inclusion of an actual airplane flight following treatment to
confines of the therapist's office. Many stimuli for SE are expen- measure avoidance behavior and in vivo anxiety. Maintenance of
treatment gains was measured in follow-up assessments following
sive and time-consuming to arrange, such as a real airplane at the
treatment.
airport and a flight. SE is bound by real-world limitations; for
example, only one take-off and landing per flight. VRE allows the
therapist to manipulate situations to best suit the patient, for References
example, to repeatedly fly the virtual airplane within one session.
Agras, S., Sylvester, D., & Oliveau, D. (1969). The epidemiology of
SE exposes the patient to increased risk and violation of patient
common fears and phobias. Comprehensive Psychiatry, 10, 151-156.
confidentiality, whereas VRE maintains the safety and privacy of American Psychiatric Association. (1994). Diagnostic and statistical man-
the therapist's office. ual of mental disorders (4th ed.). Washington, DC: Author.
There are obstacles and disadvantages to VR as well. Some Beckham, J. C., Vrana, S. R., May, J. G., Gustafson, D. ]., & Smith, G. R.
patients may not be able to overcome the fact that the exposure is (1990). Emotional processing and fear measurement synchrony as indi-
not to the "real" stimulus, and we do not yet have information on cators of treatment outcome in fear of flying. Journal of Behavior
what type of person constitutes a good candidate for VR versus in Therapy and Experimental Psychiatry, 21, 153-162.
vivo exposure. The obvious goal of therapy is to increase patients' Carlin, A. S., Hoffman, H. G., & Weghorst, S. (1997). Virtual reality and
comfort in the real feared environments, so in vivo exposure is tactile augmentation in the treatment of spider phobia: A case study.
Behaviour Research and Therapy, 35, 153-158.
always recommended as the end goal. Although the VR does allow
Cohen, J. (1988). Statistical power analysis for the behavioral sciences.
the therapist to control some aspects such as take-off, landing, and Hillsdale, NJ: Erlbaum.
turbulence, many aspects are not under therapist control. If partic- Deran, R., & Whitaker, K. (1980). Fear of flying: Impact on the U.S. air
ular exposures are required that are not already included in the travel industry (Document #BCS-00009-RO/OM). Boeing Company.
virtual environments, it would be very difficult for the therapist to First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (1995).
add these components. The seat does not move, and several users Structured Clinical Interview for DSM-IV (SCID). New York: New
have commented that they miss the G-force and seat positions of York State Psychiatric Institute, Biometrics Research.
a real airplane. Although hydraulic seats exist that would more Foa, E. B., & Kozak, M. J. (1986). Emotional processing of fear: Exposure
to corrective information. Psychological Bulletin, 99, 20-35.
closely match this experience, they are expensive and tempera-
Foa, E. B, Steketee, G., & Rothbaum, B. O. (1989). Behavioral/cognitive
mental, and it was decided not to include them in this VR setup to conceptualizations of post-traumatic stress disorder. Behavior Therapy,
keep costs relatively low. Of course, with any computer applica- 20, 155-176.
tion, there are occasional computer glitches or difficulties that may Greist, J. H., & Greist, G. L. (1981). Fearless flying: A passenger guide to
interfere with a smooth exposure. The therapist may find treatment modern airplane travel. Chicago: Nelson Hall.
difficult or frustrating if a computer glitch occurs, although these Guy, W. (1976). ECDEV Assessment Manual for Psychotherapy—Revised
have been rare. (DHEW Publication No. ADM 76-338; pp. 217-222). Bethesda, MD:
In addition, the cost of using VR for exposure therapy may be National Institute of Mental Health.
daunting to some clinicians. The same hardware setup can be used, Haug, T., Brenne, L., Johnson, B. H., Berntzen, D., Gotestam, K., &
Hugdahl, K. (1987). A three-systems analysis of fear of flying: A
with the development of additional environment models and soft-
comparison of a consonant vs. a non-consonant treatment method.
ware, for the treatment of numerous psychiatric disorders, most Behaviour Research and Therapy, 25, 187-194.
notably the anxiety disorders. Many VR environments have mul- Hodges, L. F., Rothbaum, B. O., Kooper, R., Opdyke, D., Meyer, T.,
tiple applications. For example, a driving application may be used North, M., de Graff, J. J., & Williford, J. (1995). Virtual environments
for posttraumatic stress disorder as a result of a motor vehicle for exposure therapy. IEEE Computer, 7, 27-34.
 1026 ROTHBAUM, HODGES, SMITH, LEE, AND PRICE

Howard, W. A., Murphy, S. M., & Clarke, J. C. (1983). The nature and D. (1996). Virtual reality exposure therapy in the treatment of fear of
treatment of fear of flying: A controlled investigation. Behavior Ther- flying: A case report. Behaviour Research and Therapy, 34, 477-481.
apy, 14, 557-567. Scott, W. (1987). A fear of flying inventory. In P. Kellar & S. Hayman
Maxwell, S., & Delaney, H. (1990). Designing experiments and analyzing (Eds.), Innovations of clinical practice (Vol. 7). Professional Resource
data. Pacific Grove, CA: Brooks/Cole. Exchange.
McNally, R. J., & Loura, C. E. (1992). Fear of flying in agoraphobia and Smith, S., Rothbaum, B. O., & Hodges, L. F. (1999). Treatment of fear of
simple phobia: Distinguishing features. Journal of Anxiety Disorders, 6, flying using virtual reality exposure therapy: A single case study. The
319-324. Behavior Therapist, 22, 154-158.
SPSS. (1998). SPSS advanced statistics (Version 9.0) [Computer soft-
Rothbaum, B. O., & Hodges, L. F. (1997). Virtually better® therapist
ware]. Chicago: Author.
manual for fear of flying. Atlanta, GA: Virtually Better Inc..
Stevens, J. (1996). Applied multivariate statistics for the social sciences.
Rothbaum, B. O., Hodges, L., Kooper, R., Opdyke, D., Williford, J., & Mawah, NJ: Erlbaum.
North, M. M. (1995). Effectiveness of virtual reality graded exposure in
the treatment of acrophobia. American Journal of Psychiatry, 152, Received September 14, 1999
626-628. Revision received May 3, 2000
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Rothbaum, B. O., Hodges, L., Watson, B. A., Kessler, G. D., & Opdyke, Accepted May 4, 2000 •
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Members of Underrepresented Groups:

Reviewers for Journal Manuscripts Wanted

If you are interested in reviewing manuscripts for APA journals, the APA Publications
and Communications Board would like to invite your participation. Manuscript re-
viewers are vital to the publications process. As a reviewer, you would gain valuable
experience in publishing. The P&C Board is particularly interested in encouraging
members of underrepresented groups to participate more in this process.

If you are interested in reviewing manuscripts, please write to Demarie Jackson at the
address below. Please note the following important points:

• To be selected as a reviewer, you must have published articles in peer-reviewed

journals. The experience of publishing provides a reviewer with the basis for
preparing a thorough, objective review.
• To be selected, it is critical to be a regular reader of the five to six empirical jour-
nals that are most central to the area or journal for which you would like to review.
Current knowledge of recently published research provides a reviewer with the
knowledge base to evaluate a new submission within the context of existing re-
search.
• To select the appropriate reviewers for each manuscript, the editor needs detailed
information. Please include with your letter your vita. In your letter, please iden-
tify which APA journal(s) you are interested in, and describe your area of exper-
tise. Be as specific as possible. For example, "social psychology" is not suffi-
cient—you would need to specify "social cognition" or "attitude change" as well.
• Reviewing a manuscript takes time (1-4 hours per manuscript reviewed). If you
are selected to review a manuscript, be prepared to invest the necessary time to
evaluate the manuscript thoroughly.

Write to Demarie Jackson, Journals Office, American Psychological Association, 750

First Street, NE, Washington, DC 20002-4242.