A Randomized, Controlled Clinical Trial
of In Virtuo and In Vivo Exposure for Spider Phobia
David Michaliszyn, M.Ps.,
Andre´ Marchand, Ph.D.,
Ste´ phane Bouchard, Ph.D.,
Marc-Olivier Martel, B.Sc.,
and Joannie Poirier-Bisson, B.Sc.
The present study compared the efﬁcacy of virtual reality (VR) in virtuo exposure and in vivo exposure in the
treatment of spider phobia. Two treatment conditions were compared to a waiting-list condition. A 3-month
follow-up evaluation was conducted in order to assess the durability of the treatment effects. Participants were
randomly assigned to the treatment groups. A total of 16 participants received the in virtuo treatment, and 16
received the in vivo treatment. The waiting-list condition included 11 participants. Participants received eight 1.5-
hour treatment sessions. Efﬁcacy was measured with the Fear of Spiders Questionnaire, the Spider Beliefs
Questionnaire (SBQ-F), and a Behavioral Avoidance Test (BAT). In addition, a clinician administered the
Structured Interview for DSM-IV to assess DSM-IV’s criteria for speciﬁc phobia and severity. Clinical and
statistically signiﬁcant improvements were found for both groups. Differences in treatment groups were found
on one of ﬁve measures of fear: greater improvement on the SBQ-F beliefs subscale was associated with in vivo
The efﬁcacy of in vivo exposure in the treatment of
speciﬁc phobia has been empirically demonstrated, and
is currently considered the treatment of choice for this prob-
Research in this area has investigated the effectiveness
of different methods of exposure, and found that in vivo ex-
posure is generally more effective than imaginal exposure
treating speciﬁc phobias. Results from some studies that have
explored the optimal duration of treatment suggest that one
3-hour treatment session is as effective as shorter, multiple-
However, Rowe and Craske
better results after 1 month with multiple exposure sessions or
four sessions a week than with a 1-day 4-hour session. These
results suggest that gains can be maximized by spacing out
exposure over several days. It is additionally recommended
that the client be exposed to varied stimuli, such as different
types of spiders. The use of varied stimuli maximizes the
maintenance of gains and reduces the possibility that fears
will reemerge. Exposure in multiple contexts also protects
Virtual reality (VR) is a new tool of interest
for exposure, offering the possibility of optimal exposure
parameters in the treatment of spider phobia. However, it still
requires more empirical validation.
The effectiveness of VR exposure therapy (VRET), or
in virtuo exposure, has gained empirical support. The litera-
ture in this area indicates a 25% refusal and drop-out rate
(combined) for individuals offered conventional exposure
In the same direction, a survey has found that 76%
of respondents would prefer in virtuo exposure over tradi-
tional exposure therapy, and the refusal rate for in vivo expo-
sure (27%) was higher than the refusal rate for in virtuo
Further development of in virtuo exposure could be
an effective treatment for people who consult for phobias.
Moreover, this type of treatment could reduce the stigma
associated with traditional therapy and the reluctance often
associated with exposure.
In virtuo exposure has a number of advantages over con-
ventional therapy, such as: (a) greater control over phobo-
genic stimuli and thus greater accuracy in inducing anxiety,
Department of Psychology, Universite
´al, Montreal, Quebec, Canada.
Department of Psychology, Universite
´al, Centre de recherche Fernand-Se
´guin, Montreal, Quebec, Canada.
Department of Psychology, Universite
´bec en Outaouais, Gatineau, Canada.
Department of Psychology, McGill University, Montreal, Quebec, Canada.
CYBERPSYCHOLOGY,BEHAVIOR, AND SOCIAL NETWORKING
Volume 13, Number 6, 2010
ªMary Ann Liebert, Inc.
and the ability for the therapist to repeat exposure at will; (b)
limited unexpected events during exposure; (c) exposure to
fears that can be difﬁcult to reproduce in vivo (e.g., fear of
ﬂying, fear of storms) and reduction of costs (e.g., taking the
plane); (d) remaining in the clinician’s ofﬁce during exposure
facilitates conﬁdentiality; and (e) decreased maintenance and
associated costs required for animals (hygiene, food, etc.)
used for exposure.
In addition, research in this area suggests that in virtuo ex-
posure can be effective in the treatment of arachnophobia.
Garcia-Palacios et al.
found signiﬁcant improvement in 23
patients treated with in virtuo exposure, in comparison to a
waiting-list condition. Using a Stroop task and heart-rate
gains with 28 participants on a cognitive and physiological
level using in virtuo exposure. Further evidence of the effec-
tiveness of this type of treatment comes from Carlin et al.
Hoffman et al.
Both of these studies used ‘‘tactile augmen-
tation,’’ which consists of holding an artiﬁcial spider (taran-
tula) while visually perceiving a virtual spider. These studies
demonstrated efﬁcacy of tactile augmentation. Hoffman et al.
demonstrated that participants in the tactile augmentation
condition showed the greatest progress on behavioral mea-
sures. This ﬁnding is limited by the fact that only 8 of the 36
participants were phobic at a clinical level.
The most important issue in considering the use of in virtuo
exposure is comparing the efﬁcacy of this form of therapy
with that of conventional therapy. This has been addressed
and most studies converge toward the ﬁnding that in virtuo
exposure is as effective as in vivo exposure for acrophobia
(fear of heights) and fear of ﬂying.
This same conclusion
cannot be made yet for arachnophobia, because the existing
studies of in virtuo exposure for this problem report com-
parisons of pre–post measures and=or comparisons to a
waiting list but do not directly compare a in virtuo exposure
group to a traditional exposure group. In addition, few
studies have included follow-up evaluations to verify treat-
ment durability for arachnophobia.
The ﬁrst goal of this study is to evaluate the efﬁcacy of
in virtuo exposure therapy and simultaneously compare it to
conventional in vivo exposure therapy. The second goal is to
measure the maintenance of gains in a 3-month follow-up
Participants were French speakers recruited in the Mon-
treal area by several methods, including: (a) oral presenta-
tions and ﬂyers in university courses, (b) advertisements in
local newspapers, and (c) individuals who requested treat-
ment for fear of spiders through referrals and acquaintances.
The eligibility criteria for participation in the study were:
(a) Fulﬁllment of DSM-IV
diagnostic criteria for Speciﬁc
Phobia Animal Type (spiders), as evaluated by trained
doctoral students. Criteria E and D varied between
clinical and subclinical levels, given that people could
avoid places where they might come in contact with
spiders without necessarily affecting everyday func-
tioning, but still react with irrational fear in the pres-
ence of spiders.
(b) A minimum of 1-year duration of the phobia.
(c) An inability to touch a vivarium with a tarantula in it
prior to treatment, as evaluated during the Behavioral
Avoidance Test (BAT).
(d) A score in the clinical range on both the Fear of
and Spider Beliefs Ques-
(e) No other psychiatric problem in need of immediate
treatment (participants were screened for anxiety dis-
orders and depression; participants with comorbid
spider phobia and primary anxiety disorder were ex-
(f) No current alcohol or drug dependence or medication.
(g) No severe physical illness.
Participants were on average 29.1 years old (SD ¼7.99; range
18–51), with 16.53 years of education (SD ¼2.37; range 12–
22). A total of 64% of the sample were in a romantic rela-
tionship and 36% had a child. Further, 64% of participants
were currently employed, and only one participant was male.
Of the 32 participants included (16 in virtuo,16in vivo), 28 met
the full DSM-IV criteria for speciﬁc phobia (15 in virtuo,13
in vivo). Four participants had a partial diagnosis of speciﬁc
phobia but scored within the phobic range on the question-
naire measures and on the BAT.
The virtual environments (VEs) were generated using a
personal computer with a Pentium
4 3.0 GHz 1.00 GB of
RAM cpu, 256MB of graphics memory, and a wide 256-bit
memory interface ATI Radeon 9800 XT video card. The en-
vironments were displayed on monoscopic I-glasses
(i-O Display Systems) head-mounted
display (HMD) with a resolution of 800600 pixels. The
HMD was draped with a 30 cm40 cm black cloth to block
out ambient light. The HMD was also equipped with an IS-
tracker (3 dof) that sensed the movement of the
participants’ heads. Together, the HMD and tracker provided
a view that followed the participants’ head movements as
they tilted, panned, and swiveled their heads to scan the VE.
The participants used a handheld wireless Gyration mouse to
control their forward and backward movements in the VE.
Ambient sounds were played on the PC’s stereo speakers and
through the HMD. The arachnophobia environments were
modiﬁed computer-game environments based on the Max
Payne video game and downloaded from the Universite
´bec en Outaouais (UQO) Cyberpsychology laboratory
Web site (www.uqo.ca=cyberpsy). The computer-graphic
artists used the Max Payne platform to customize the envi-
ronments and populate them with animated spiders of dif-
ferent shapes and sizes.
Participant evaluation. A battery of questionnaires and a
structured clinical interview were systematically adminis-
tered to participants in order to obtain and document the
following information: age, gender, marital status, general
health, level of education, presence intensity, duration of
arachnophobia, and presence of comorbid disorders. Mea-
sures were taken pretreatment, during treatment, posttreat-
ment, and at a 3-month follow-up.
690 MICHALISZYN ET AL.
Measures of fear. The Structured Clinical Interview for
was used to conﬁrm the diagnoses of
spider phobia (principal diagnosis) and comorbid anxiety
disorders (secondary diagnosis), as deﬁned by DSM-IV.
SCID-I has good concurrent validity with clinician judgment
(k¼0.69) and inter-rater reliability has been reported at 0.77
The Questionnaire sur la Peur des Araigne
French translation of the Fear of Spiders Questionnaire)
chosen for a subjective measure of the efﬁcacy of in virtuo
treatment. This questionnaire is reported to have excellent
split-half reliability and internal consistency, and good test–
retest consistency. In addition, the FSQ-F’s good convergent
validity is demonstrated by highly signiﬁcant correlations
with the BAT (r¼0.65). Good construct validity is demon-
strated by the test’s ability to discriminate phobics from non-
phobics, as measured by a BAT.
The FSQ-F is composed
of 18 items about fear and avoidance of spiders to be rated on
a scale from 1 to 7 (1 ¼does not apply to me;7¼very much
applies to me).
The Questionnaire des Croyances a
`propos des Araigne
(SBQ-F) is a French translation of the Spider Beliefs Ques-
This instrument is composed of 78 items divided
into two subscales (SBQ-F beliefs, SBQ-F behaviors) that re-
spectively address beliefs about spiders and beliefs about
own behavior in the presence of spiders. Arntz et al.
ported internal consistency of 0.94 for this measure. Test–
retest reliability values between 0.59 and 0.84 have been
The BAT is a popular objective measure of clinical progress
in overcoming phobias. A large spider (female rosy-haired
tarantula, approximately 10 centimeters long including front
legs and cephalothorax) was placed in a vivarium without a
lid. The cage was placed on a table at the far end of a room
and a chair was placed 3 meters from the vivarium. The pa-
tient was instructed to enter the room, sit down in the chair,
and then get up and walk as close to the cage as possible.
Participants were advised that the BAT was an objective
measure of their fear of spiders and not part of the therapy.
During the test, the experimenter stayed behind the client, in
an effort to minimize any potential impact of his presence.
When the participants were as near as possible to the spider,
the distance in meters between participant and spider was
measured. The measure of distance was converted to a be-
havioral score that ranged from 0 to 11, where 0 ¼anything
less than sitting down on the chair; 1 ¼50 centimeters; 2 ¼100
centimeters; 3 ¼150 centimeters; 4 ¼200 centimeters; 5 ¼250
centimeters; 6 ¼300 centimeters; 7 ¼staring into the open
vivarium for 5 seconds; 8 ¼touching the vivarium on side
farthest from the spider for at least 5 seconds; 9 ¼touching the
vivarium on side closest to the spider for at least 5 seconds;
10 ¼inserting one hand into the vivarium and putting one
ﬁnger on the ground on the side farthest from the spider for at
least 5 seconds; 11 ¼placing one hand on the branch in the
middle of the vivarium for at least 5 seconds.
An end-state functioning index was developed to assess
clinically signiﬁcant improvement using ﬁve measures and
their respective cut-off points to rate clinical success. The in-
(a) BAT (10), SBQ-F beliefs (23.15), SBQ-F behaviors (21.5),
FSQ-F (65.3), and SCID-I evaluation of speciﬁc phobia.
The cut-off score for treatment success on the BAT
(scores 0–11) was determined by the sample distribu-
tion and set at 10 or higher (inserting one hand in the
vivarium and putting one ﬁnger on the ground on the
side farthest from the spider for at least 5 seconds)
(b) For the SBQ-F beliefs, SBQ-F behaviors, and FSQ-F
scales, the established cut-off point for clinically sig-
niﬁcant change was entry into the non-phobic range.
This stricter cut-off point was chosen for its relevance
within our data (stricter than 2 SD from the phobic
(c) An absence of speciﬁc phobia diagnosis on the SCID-I
was required to meet the criteria of clinical success.
Clinical success was awarded 1 point and failure was
awarded 0 point.
The scores on these ﬁve measures were added together to
produce a score out of ﬁve.
Control measures. The Inventaire de la De
the French translation of the Beck Depression
measured the presence and intensity of 21
symptoms of depression. The French version of this ques-
tionnaire was validated by Bourque and Baudette,
was used in this to match the treated participant’s language
and population. Temporal stability over 4 months was 0.62
and internal coherence varied between 0.90 and 0.92.
´valuation du The
French translation of the Therapist Evaluation Scale.
questionnaire includes 25 items that measure the participant’s
perception of his or her therapist.
The Questionnaire sur la Perception du Traitement pour
´ciﬁques (QPTPS) is a French adaptation for spe-
ciﬁc phobia of the Questionnaire on Treatment Perception
It consists of ﬁve questions that measure
the participant’s perception of treatment credibility. The test–
retest reliability is 0.90 ( p<0.05).
Virtual-reality questionnaires. The Questionnaire sur
´tat de Pre
´sence (PQ-F) (French translation of the Pre-
) was administered following the
exposure session. Each of the PQ-F’s 19 items are rated on
total score and ﬁve subscale scores. The ﬁve subscales are:
Realism (similarity between the VE and the equivalent
natural environment), Affordance to Act (ability to actively
explore and manipulate the VE), Interface Quality (delays
or awkwardness related to the software or apparatus),
Affordance to Examine (ability to approach virtual objects
and to examine them from different angles), and Self-
Evaluation of Performance (feeling of competence for per-
forming tasks in the VE). The Presence Questionnaire has
good reliability (Cronbach’s a¼0.81).
The Questionnaire sur les Cybermalaises (SSQ-F; French
translation of the Simulator Sickness Questionnaire
also administered after the exposure session. The SSQ-F has a
4-point scale to rate 16 symptoms of simulator sickness, such
as nausea, eye fatigue, and vertigo. The SSQ-F produces a
total score and three subscale scores: Nausea,Ocular-Motor
Problems, and Disorientation. Although the SSQ-F is presently
in the validation process, it is already commonly used in VR
COMPARING IN VIRTUO AND IN VIVO EXPOSURE 691
A total of 32 participants with a diagnosis of speciﬁc
phobia took part in this randomized controlled clinical trial.
Participants were randomly assigned to one of three condi-
tions: waiting list, in virtuo exposure, or in vivo exposure.
Participants in the waiting-list group waited for 8 weeks be-
fore being reevaluated and randomly assigned to one of the
two treatment groups.
The treatment followed a standardized exposure protocol
from the Cyberpsychology Laboratory at UQO. It was ad-
ministered by ﬁve doctoral students in psychology, super-
vised by a senior psychologist. Treatments were equal in both
groups (in vivo and in virtuo) in terms of number of sessions,
exposure time, and techniques in order to allow for valid
comparisons. Treatment consisted of psychoeducation about
spiders, gradual exposure, and cognitive restructuring. Eight
90-minute treatment sessions were planned. The ﬁrst session
included evaluation, information about treatment, and psy-
choeducation about phobias and spiders. The next six ses-
sions consisted of gradual exposure tasks and cognitive
restructuring. Each exposure session lasted approximately 1.5
hours, with pauses every 20 to 30 minutes to prevent cyber-
sickness in VR and to rest in the in virtuo group. The ﬁnal
session focused on relapse prevention. Treatment completion
criteria were as follows: for participants in the in virtuo group,
completing treatment meant going through all three levels of
the VR program and confronting a large black-widow spider.
In addition, participants had to report low levels of anxiety
throughout the program. Participants in the in vivo group
were considered to have completed treatment when they
were able to manipulate two types of live spiders (Tegenaria
domestic and Pholcus) in their hands. Three participants (two
in vivo and one in virtuo) achieved these goals prematurely
and therapy ended after seven sessions for them. All other
participants received eight sessions. Exposure proceeded ac-
cording to a list of tasks assigned by the therapist and rated
by the participants with a Subjective Units of Distress (SUDs)
rating. The tasks were designed to elicit a SUDs rating be-
tween 30 and 70 (out of 100).
Signiﬁcance was set at p¼0.05 for this study. Posttreat-
ment analyses included data from the waiting-list group after
they were distributed among the in vivo and in virtuo groups.
Independent ttests revealed no signiﬁcant differences in
demographic and clinical variables between the waiting-list
condition and the treatment condition at pretreatment,
p<0.05. Further ttests revealed no signiﬁcant differences in
demographic and clinical variables between the in virtuo
group and the in vivo group at pretest ( p<0.05).
Pretreatment and posttreatment tests
A one-way between-groups multivariate analysis of vari-
ance (22 MANOVA) was conducted to compare the waiting-
list condition to the combined treatment groups. The depen-
dant variables were FSQ-F, SBQ-F beliefs, and SBQ-F behav-
iors. No signiﬁcant differences were found on any clinical
variables for the waiting-list group between both pretreat-
ment evaluations, F(3, 10) ¼0.233, p<0.871, indicating no
improvement prior to treatment. A statistically signiﬁcant
difference was found between the waiting-list group at pretest
and the treatment group at posttest, F(3, 44) ¼3.65, p¼0.02.
Scores on the PQ-F showed elevated total presence
(M¼84.21, SD ¼14.35). Cybersickness scores on the SSQ-F
were slightly higher than the questionnaire norms (M¼19.41,
SD ¼14.79), but lower than SSQ-F scores in a phobic sample
reported by Robillard et al.
We therefore consider our
in virtuo treatment to have been conducted according to the
Posttreatment and follow-up tests
Of the 36 participants in this study, 32 completed the
posttest evaluation, and 26 were evaluated at a 3-month
follow-up. Incomplete evaluations at posttest were explained
by psychosocial stressors, and the participants with incom-
plete evaluations were distributed evenly between the treat-
ment groups. Three participants were switched from the
in virtuo group to the in vivo group due to a lack of reactivity
to the virtual spiders. This created an even distribution of
participants, with 16 in the in virtuo group and 16 in the in
vivo group. At follow-up, seven non-completers were counted
in the in vivo condition, and three in the in virtuo group. The
non-completion in the in virtuo condition was explained by
cybersickness related to a medical condition and three of the
absences in the in vivo condition were due to being out of the
country. The remaining four participants removed them-
selves for personal reasons. Independent ttests revealed no
signiﬁcant differences between the missing cases and the
completers in age, gender, level of education, and employ-
ment status. Signiﬁcant differences were found with regard to
children and civil status. None of the missing participants
were married or had children. No signiﬁcant differences in
clinical variables were found between missing participants
and completers ( p<0.05).
A repeated-measures MANOVA was conducted on the
SBQ-F (beliefs and behaviors subscales) and FSQ-F, revealing
a signiﬁcant time effect, and demonstrating treatment efﬁcacy
over time (Table 1). Both in virtuo and in vivo exposure pro-
duced signiﬁcant improvement at posttest and follow-up.
However, a signiﬁcant time-by-treatment interaction was
found on the SBQ-F beliefs subscale ( p<0.05) at follow-up.
In vivo exposure scores decreased slightly over time, whereas
in virtuo exposure scores did not change signiﬁcantly (Table 2).
A similar non-signiﬁcant trend was observed for the SBQ-F
behaviors subscale. There were no signiﬁcant differences be-
tween groups for interaction effects or treatment effects on the
To address attrition in this study, an intent-to-treat analysis
(repeated-measures MANOVA) was conducted on the SBQ-F
692 MICHALISZYN ET AL.
(beliefs and behaviors subscales) and FSQ-F. This analysis
revealed a signiﬁcant time effect, demonstrating treatment
efﬁcacy over time, F(6, 28) ¼17.12, p¼0.00. Both in virtuo
and in vivo exposure produced signiﬁcant improvement at
posttest and follow-up. There was no signiﬁcant time-by-
treatment interaction for the SBQ-F beliefs subscale, p<0.05.
The BAT posttreatment scores were asymmetrical and non-
transformable. This variable was therefore modiﬁed to rep-
resent participants’ improvement by (a) subtracting pretest
scores from posttest scores and (b) subtracting pretest scores
from follow-up scores. The new change variable had a normal
distribution. A ttest revealed a signiﬁcant change at posttest,
M¼6.73, SD ¼2.96; t(31) ¼12.86, p¼0.00, and at follow-up,
M¼6.87, SD ¼3.18; t(25) ¼11.01, p¼0.00. A two-way
between-groups ANOVA was conducted to compare treat-
ment groups on the new change variable. There was no
signiﬁcant difference between in vivo and in virtuo treatment
groups, F(1, 24) ¼2.55, p¼0.12.
The posttreatment and follow-up results on the SCID-I
were asymmetrical and non-transformable. This variable was
therefore dichotomized, and non-parametric tests were em-
ployed for the analysis. Friedman’s test revealed a signiﬁcant
difference between pretest and posttest results, as well as
between pretest and follow-up measures ( p¼0.00), thereby
demonstrating treatment efﬁcacy. Fisher’s exact test was used
to compare the treatment outcome on the SCID-I. No signif-
icant differences were found between the in virtuo and in vivo
(p¼0.226) groups at posttest or at follow-up ( p¼0.238). It
was also observed that, at pretest, the in vivo group included
14 participants with a full diagnosis of speciﬁc phobia and
two participants with a partial diagnosis (criteria D and E
subclinical). At posttest, none of the in vivo participants still
had a diagnosis of speciﬁc phobia. The in virtuo group in-
cluded 15 participants with a diagnosis of speciﬁc phobia and
one participant with a partial diagnosis at pretest. At posttest,
only two participants had a remaining partial diagnosis and
one participant still had a full diagnosis of speciﬁc phobia.
Table 1. Treatment Outcome and Treatment Interaction for In Vivo
and In Virtuo Exposure for Spider Phobia
Time effect Treatment interaction
DF p Z
FSQ-F 70.12 0.000 0.753 1 0.814 0.445 0.034 0.177
SBQ-F beliefs 39.48 0.000 0.632 1 4.96 0.012 0.177 0.778
SBQ-F behaviors 39.92 0.000 0.634 1 0.309 0.714 0.013 0.094
FSQ-F, Fear of Spiders Questionnaire; SBQ-F beliefs, Spider Beliefs Questionnaire, beliefs subscale; SBQ-F behaviors, Spider Beliefs
Questionnaire, behaviors subscale.
Table 2. Mean and Standard Deviation for Outcome Measures at Pretreatment,
Posttreatment, and 3-Month Follow-Up
In vivo In virtuo
Pretreatment 3.17 2.55 3.56 2.89
Posttreatment 10.47 1.67 9.25 2.72
Follow-up 9.86 2.15 9.73 2.43
Pretreatment 103.28 13.13 104.61 9.59
Posttreatment 47.88 14.07 54.37 22.46
Follow-up 47.81 32.25 56.67 23.99
Pretreatment 47.73 14.31 41.17 15.58
Posttreatment 16.50 17.83 18.47 20.26
Follow-up 9.71 9.02 16.92 10.81
Pretreatment 45.20 15.45 45.53 18.49
Posttreatment 11.32 17.38 16.39 21.40
Follow-up 9.76 8.29 13.54 15.95
Pretreatment 1.83 0.38 1.94 0.24
Posttreatment 0.00 0.00 0.25 0.58
Follow-up 0.17 0.58 0.33 0.72
BAT, Behavioral Avoidance Test; FSQ-F, Fear of Spiders Questionnaire; SBQ-F beliefs, Spider Beliefs Questionnaire, beliefs subscale; SBQ-F
behaviors, Spider Beliefs Questionnaire, behaviors subscale; SCID-I, Structured Clinical Interview for DSM-IV.
COMPARING IN VIRTUO AND IN VIVO EXPOSURE 693
End-state functioning index
The posttreatment and follow-up scores on the end-state
functioning index were asymmetrical and non-transformable.
This variable was therefore dichotomized, and non-parametric
tests were employed for the analysis. The Friedman test re-
vealed signiﬁcant differences in scores over the three time
periods ( p¼0.00), indicating treatment efﬁcacy. Differences
between the in vivo and in virtuo groups were evaluated with a
Fisher’s exact test. No signiﬁcant differences were found at
posttreatment ( p¼0.76)oratfollow-up(p¼0.62).
Participants in both in virtuo and in vivo exposure therapy
for spider phobia demonstrated signiﬁcant improvement on
objective and subjective measures of fear after eight 90-minute
treatment sessions. No signiﬁcant differences between the
groups were found at posttest or at follow-up on the fol-
lowing measures of fear: FSQ-F, SBQ-F behaviors subscales,
and the BAT. At the posttest evaluation, several participants
in the in virtuo group indicated that they were curious to
know how they would react to a spider in their natural en-
vironment. At the follow-up period, some had ﬁnally en-
countered a spider and were appreciative of their reaction
toward it. Signiﬁcant differences between groups were also
found for the SBQ-F beliefs subscale at follow-up. In vivo
scores decreased on this measure from posttest to follow-up,
whereas in virtuo scores were maintained over time. The
greater treatment gains in the in vivo group may be attribut-
able to the direct contact with spiders and learning about the
behaviors and reactions of a live spider.
This randomized controlled study found slight differences
between in vivo and in virtuo treatments for fear of spiders.
The effect size between treatment groups at follow-up on the
SBQ-F beliefs subscale was 6.9%.
Three participants in the in virtuo group displayed no re-
action to the virtual spiders and were reassigned to the other
group. This possible limitation of VR should be considered in
treatment planning. The primary limitations of in virtuo ex-
posure published to date have been associated with lack of a
sense of presence and cybersickness. Further studies should
be required to evaluate levels of presence and cybersickness
in participants who react unfavorably to VEs. In addition,
future research could explore other variables that may predict
treatment success. In the meantime, in vivo exposure is re-
commended when possible. The rate of attrition in the present
study was higher in the in vivo group than in the in virtuo
group at follow up. Speciﬁcally, at post-test, 2 non completers
were counted in both groups, while 7 and 3 non completers
were counted at follow up for the in vivo group and the
in virtuo group, respectively. The majority of drop-outs
occurred after treatment was completed. Therefore this does
not support Garcia-Palacios et al.’s
suggestion that in virtuo
exposure may be less threatening than in vivo exposure. The
drop-out rate in this study (27%) is therefore attributed to
The strength of this study is the use of randomized dis-
tribution and a control group to directly compare in vivo and
in virtuo exposure. Authors of future research on spider
phobia may wish to consider the following suggestions for
improving study design. First, spider phobia in Quebec is
likely to differ from spider phobia in other areas. Spiders in
Quebec may bite and cause a local reaction, but they are not
deadly or cause for medical concern. To assess functionality
in participants in Quebec properly, a BAT task of touching or
manipulating a local spider should be considered as an ad-
junct to the traditional BAT task with a tarantula in both
exposure methods. Second, in their study of the use of VR in
spider phobia, Garcia Palacios et al.
included anxiety mea-
sures such as the State–Trait Anxiety Inventory (STAI) in the
pre, post, and follow-up evaluations, and included the SUDs
during the BAT. This more rigorous evaluation could have
been used in the present study to provide more accurate
measures of treatment outcome on the BAT. Third, the
posttest results on the BAT in the present study demonstrated
a ceiling effect. The inclusion of a more complex BAT task,
such as manipulating a local spider or touching the live ta-
rantula with a pencil or straw, could prevent the ceiling effect
and provide a more accurate measure of participants’ im-
provement. Fourth, although this idea was not explored in
this study, in vivo exposure following in virtuo exposure could
consolidate treatment gains Finally, another option would be
to include tactile augmentation to the in virtuo exposure as
did Hoffman et al.,
allowing greater presence and improved
This study did not include physiological measures such as
heart rate and skin conductance although they have been
demonstrated empirically to be modiﬁed by in virtuo expo-
sure. Wiederhold and Wiederhold
found that arousal in the
VR environments varied between individuals on measures of
skin resistance and cardiac response, which could affect
treatment efﬁciency. Also, Co
in virtuo exposure modiﬁed cardiac response. In another
found that changes in perceived
self-efﬁcacy and dysfunctional beliefs were the best pre-
dictors of change in general outcome and cardiac response.
These measures should be included in future researches to
measure therapeutic gains.
In conclusion, both in vivo and in virtuo exposure are
efﬁcient methods of treating spider phobia. A slight ad-
vantage of in vivo exposure over in virtuo exposure was
found, as revealed by signiﬁcant and continued gains on
the SBQ-F beliefs subscale after posttest in the in vivo
group. The effectiveness of VR therapy could likely be en-
hanced considerably by using tactile augmentation
complementary in vivo exposure. If equivalent efﬁciency is
achieved, in virtuo exposure has compelling advantages
over in vivo exposure and could make it the treatment of
This study is the doctoral thesis of David Michaliszyn,
Ph.D. (candidate). Andre
´Marchand, Ph.D., acted as his thesis
advisor. The study was conducted at the Laboratoire de
Cyberpsychologie of l’Universite
(UQAM) where the ﬁrst author is completing his Ph.D. Spe-
cial thanks go to Ste
´phane Bouchard from the Laboratoire de
Cyberpsychologie of l’Universite
´bec en Outaouais
(UQO) (www.uqo.ca=cyberpsy) for his invaluable help dur-
ing all the stages of this study. Also, we wish to give special
thanks to Me
´lissa Martin and Joannie Poirier-Bisson for their
contributions. This study was funded by the Canada Foun-
dation for Innovation.
694 MICHALISZYN ET AL.
No competing ﬁnancial interests exist.
1. Emmelkamp PMG, Bowman TKO, Scholing A. (1995) An-
xiety disorders. A practitioner’s guide. Chichester, UK: John
2. Emmelkamp PMG. (1982) Phobic and obsessive–compulsive dis-
orders: Theory, research, and practice. New York: Plenum Press.
3. Hellstrom K, O
¨st LG. One session therapist directed expo-
sure vs. two forms of manual directed self exposure in the
treatment of spider phobia. Behaviour Research & Therapy
¨st LG, Ferebee I, Furmark T. One session group therapy of
spider phobia: Direct versus indirect treatments. Behaviour
Research & Therapy 1997; 35:721–32.
5. Rowe MK, Craske MG. Effects of expanding spaced vs
massed exposure schedule on fear reduction and return of
fear. Behaviour Research & Therapy, 1998; 36:707–17.
6. Gunther LM, Denniston JC, Miller RR. Conducting exposure
treatment in multiple contexts can prevent relapse. Beha-
viour Research & Therapy 1998; 36:75–91.
7. Garcia-Palacios A, Botella C, Hoffman H, et al. Comparing
acceptance and refusal rates of virtual reality exposure vs.
in vivo exposure by patients with speciﬁc phobias. Cyber-
Psychology & Behavior 2007; 10:722–4.
8. Rothbaum BO, Hodges L, Smith S, et al. A controlled study
of virtual reality exposure therapy for the fear of ﬂying.
Journal of Consulting & Clinical Psychology 2000; 68:1020–6.
9. Carlin AS, Hoffman HG, Weghorst S. Virtual reality and
tactile augmentation in the treatment of spider phobia: A
case report. Behaviour Research & Therapy 1997; 35:153–8.
10. Garcia-Palacios A, Hoffman HG, Carlin A, et al. Virtual re-
ality in the treatment of spider phobia: A controlled study.
Behaviour Research & Therapy 2002; 40:983–93.
´S, Bouchard S. Documenting the efﬁcacy of virtual re-
ality exposure with psychophysiological and information
processing measures. Applied Psychophysiology & Bio-
feedback 2005; 30:217–32.
12. Hoffman HG, Garcia-Palacios A, Carlin A, et al. Interfaces
that heal: Coupling real and virtual objects to treat spider
phobia. International Journal of Human–Computer Interac-
tion 2003; 16:283–300.
13. Emmelkamp PMG, Krijn M, Hulsbosch AM, et al. Virtual
reality treatment in acrophobia: A comparison with in vivo
exposure. Behaviour Research & Therapy 2002; 40:509–16.
14. Rothbaum BO, Hodges L, Anderson LP, et al. Twelve-month
follow-up of virtual reality and standard exposure therapies
for the fear of ﬂying. Journal of Consulting & Clinical Psy-
chology 2002; 70:428–32.
15. American Psychiatric Association. (1994) Diagnostic and sta-
tistical manual of mental disorders. 4th ed. Washington DC:
16. Szymanski J, O’Donoghue W. Fear of Spiders questionnaire.
Journal of Behavior Therapy & Experimental Psychiatry
17. First MB, Spitzer R, Gibbon M, et al. (1996) Structured clinical
interview for DSM-IV axis-I disorders—Patient version. New
York: Biometrics Research Department, New York State
18. Summerfeldt LJ, Anthony MM. (2002) Structured and
semistructured diagnostic interviews. In Antony MM, Bar-
low DH, eds. Handbook of assessment and treatment planning
for psychological disorders. New York: Guilford Press, pp.
19. Muris P, Merckelbach H. A comparison of two spider fear
questionnaires. Behaviour Research & Therapy 1996; 27:
20. Arntz A, Lavy E, Van den Berg G, et al. Negative beliefs of
spider phobics: A psychometric evaluation of the Spider
Phobia Beliefs Questionnaire. Advances in Behaviour Re-
search & Therapy 1993; 15:257–77.
21. Cottraux J, Bouvard M, Le
´geron P. (1985) Me
´thodes et e
´valuation des comportements. Paris: E
22. Beck AT, Rush AJ, Shaw BF, et al. (1979) Cognitive therapy of
depression. New York: Guilford Press.
23. Bourque P, Beaudette D. E
´trique du ques-
tionnaire de de
´pression de Beck aupre
`s d’un e
´tudiants universitaires francophones. Canadian Journal
of Behavioral Science=Revue Canadienne des Sciences du
Comportement 1982; 14:211–18.
24. Freeston MH, Dugas MJ, Le
´ger E, et al. (1995) E
´valuation du the
´bec, Canada: E
´cole de psy-
25. Williams KE, Chambles LD. The relationship between ther-
apist characteristics and outcome of in vivo exposure treat-
ment for agoraphobia. Behavior Therapy 1990; 21:111–16.
26. Borkoveck TD, Nau SD. Credibility of analogue therapy
rationales. Journal of Behavior Therapy & Experimental
Psychiatry 1972; 3:257–60.
27. Witmer BG, Singer MJ. Measuring presence in virtual en-
vironments: A presence questionnaire. Presence 1998; 7:
28. Kennedy RS, Lane NE, Berbaum KS, et al. Simulator sickness
questionnaire: An enhanced method for quantifying simu-
lator sickness. International Journal of Aviation Psychology
29. Robillard G, Bouchard S, Fournier T, et al. Anxiety and
presence during VR immersion: A comparative study of the
reactions of phobic and non-phobic participants in thera-
peutic virtual environments derived from computer games.
CyberPsychology & Behavior 2003; 6:467–76.
30. Wiederhold BK, Wiederhold MD. Lessons learned from 600
virtual reality sessions. CyberPsychology & Behavior 2000;
´S, Bouchard S. Cognitive mechanisms underlying vir-
tual reality exposure. CyberPsychology & Behavior 2009;
Address correspondence to:
Department of Psychology
Case Postale 8888
´bec H3C 3P8
COMPARING IN VIRTUO AND IN VIVO EXPOSURE 695