Isr J Relat Sci - Vol. 48 - No. 2 (2011) Helene S. Wallach et al. Virtual Reality Exposure versus for Treatment of Public Speaking Anxiety: A Pilot Study

Helene S. Wallach, PhD, 1 Marilyn P. Safir, PhD,1 and Margalit Bar-Zvi, MD2

1 Department of Psychology, University of Haifa, Haifa, Israel 2 Out-Patient Psychiatry, Ziv Hospital, Safed, Israel

The gold standard treatment for phobias is Cognitive ABSTRACT Behavior (CBT) which combines both behav- Objectives: To determine the utility of Virtual Reality ioral and cognitive elements (1-4). Although numerous (VRE) in comparison with Cognitive studies have been conducted to examine the necessity Therapy (CT) and with Cognitive-Behavior Therapy (CBT). of employing both cognitive and behavioral elements in therapy, it remains unclear whether it is sufficient to Method: Subjects suffering from public speaking anxiety deliver (CT) alone, behavior therapy (PSA) were randomly allocated to VRE and CT, and (BT) alone or if both need to be employed (5-8). In received 12 therapy sessions, employing standardized recent years, Virtual Reality Exposure (VRE) has been treatment manuals. Outcome (questionnaires, observer utilized for the BT (exposure) component of CBT. and self ratings of a behavioral task) was compared to Therefore, for this pilot study, we compared VRE to CT results of subjects in a previous study CBT and Wait in an attempt to examine the relative efficacy of each List Controls who were not significantly different on component independently, and in comparison with the demographic data. combined treatment. Results: CT was not superior to VRE on cognitive measures, but was superior to VRE on one behavioral Social Phobias measure (LSAS fear). VRE was superior to CT on one Social phobia is defined as a fear of and desire to avoid behavioral measure (fear reduction on a behavioral a situation in which the individual may come under task). No differences were found between either TC , or scrutiny by others and fears he/she may act in ways that VRE, and CBT and all were superior to WL. may be humiliating or embarrassing (9). Social phobia affects school performance, ability to create social net- Limitations: Subject group was small and homogeneous. works and intimate relationships as well as work perfor- It appeared advisable to increase number of therapy mance, is the most common anxiety disorder, and the sessions. third most common psychiatric disorder (10, 11). In the Conclusions: VRE and CT proved to be equally effective present study we focused on a non-generalized social to CBT in reducing PSA relative to a control group, with phobia - public speaking anxiety. Public speaking anxi- minimal differential effects between them. Therefore, ety (PSA) is the most common social phobia, afflicting employing either one may be satisfactory and sufficient. 40% of all those who suffer from social phobia (11).

Virtual Reality Exposure (VRE) for Social Phobia Exposure is presumed to extinguish maladaptive con- ditioning (phobic object = fear) through anxiety reduc- tion, and to shape new conditioned responses (neutral or adaptive responses to the formally phobic object).

Address for Correspondence: Helene S. Wallach PhD, Department of Psychology, University of Haifa, Haifa, 31905, Israel [email protected]

91 Virtual Reality Exposure versus Cognitive Restructuring

This occurs as the client experiences both a reduction compare VRE to both CT and to the gold standard in anxiety during the exposure (habituation) and the treatment for PSA, namely CBT. absence of the catastrophic event he/she anticipated (extinction) (1, 12, 13). VR is defined as a situation in which sensory information is generated by a com- Methods and Experimental Design puter rather than by the natural environment (e.g., 14). Participants During the exposure stage in VR, the client puts on a Twenty subjects who suffered from PSA were recruited helmet which is connected to a computer. The helmet for this study. Candidates suffering from a psychotic provides both visual and audio input. The therapist disorder, drug or alcohol abuse, or epilepsy, or who were employs a special computer program enabling her/ in for this problem, or were taking psy- him to change various elements in the virtual environ- chotropic medication were excluded. Epilepsy may be ment, thus providing the client with gradual exposure aggravated by the use of VR, and therefore this served to aversive stimuli. The therapist views exactly what the as an exclusion criteria. Subjects were young (average client views in the helmet, on the computer screen. The age 28), were mostly Jewish (85%) and single (75%) client’s head movements change the environment, just (Table 1). In addition, their results were compared with as they would in the real world, hence increasing the data from a recent study that employed CBT (28 par- sense of immersion. The therapist monitors the client’s ticipants) for PSA using similar protocols and a wait list Subjective Units of Discomfort (SUD) and thus con- control (WL) (30 participants) condition (16). Subjects trols the fear level of the environment. Thus, VRE is in the previous study were also primarily Jewish (89%, superior to conventional exposure for people who have 90%), young (average age 28, 25) and single (75%, 93%) difficulty imagining situations vividly, for those who and not significantly different on these variables from avoid remaining in the imagined fearful situation, and the subjects in the present study (Table 1). for those who are unable to control their imagination and flood themselves with higher levels of anxiety than Table 1. Demographic Variables are produced by the hierarchy. In addition, the drop-out Family rate from VRE has been found to be much lower than situation Age Sex Single Ethnicity the drop-out rate from traditional BT (15, 16). N M Sd M F Other Maj.a Min.b VRE 10 27.50 7.85 5 5 9 1 8 2 Cognitive Therapy for Social Phobia CT 10 28.11 1.69 4 6 6 4 9 1 Cognitive factors are especially obvious in social phobia CBT 28 28.18 7.97 5 23 21 7 25 3 (6, 17, 18). Socially anxious individuals over-estimate WL 30 25.29 2.62 9 21 28 2 27 3 the threat of public criticism, scrutiny or embarrassment (17). They also have negative and distorted images of how Note. aMaj=Majority; bMin=Minority others view them (19, 20). In addition, clients exagger- ate the “social cost” of their performance (21). This leads Measures to the perception of social situations as dangerous, to an 1. Liebowitz Social Anxiety Scale (LSAS) (26) includes increase in anxiety relevant physical sensations when 24 questions. For each question, the participant rates encountering feared situations, to hyper vigilance to the the amount of fear experienced (answered on a 4-point environment and to bodily sensations and to avoidance Likert scale from 0 – “not at all” to 3 – “very much”) of feared social situations or the use of “safety” behav- and the amount of avoidance experienced (rated on a iors. Therefore, faulty cognitions develop, maintain and 4-point Likert scale from 0 – “never, 0%” to 3 – “usually, increase social phobia (18, 20). However, therapy that 68%-100%”). This scale has high reliability (Cronbach α challenges these cognitions was found to be beneficial .90-.95). It also has good validity (high correlations with and crucial in the treatment of PSA (2, 22, 23). other anxiety measures, e.g., Social Interaction Anxiety Scale – fear scale .72, avoidance scale .68; Social Phobia Cognitive-Behavior Therapy for Phobias Scale – fear scale .64, avoidance scale .59). The combination of Cognitive and Behavior 2. Self Statements During Public Speaking (SSPS) (27). (CBT) has been reported to be the treatment of choice This 10-item measure assesses fearful thoughts dur- for phobias (e.g., 23-25). Thus this pilot attempted to ing public speaking. It consists of two subscales: posi-

92 Helene S. Wallach et al.

tive and negative self-statements. This scale has high our subjects met the formal criterion for PSA. However, reliability (.78-0.80 & .80-.86, respectively). It also they volunteered for this study, reporting that they suf- has good validity (discriminates between social pho- fered from PSA, and would like treatment. Additionally bics and the general population; correlates with other their pre-therapy scores were similar to those of indi- anxiety scales: e.g., Personal Report of Confidence as viduals suffering from social phobia in previous stud- a Speaker .67; Fear of Negative Evaluation .49; Social ies. For example, mean LSAS Fear scores ranged from Phobia and Anxiety Inventory .48. For the positive 28.18-34.80 in this study which is higher than mean scale, higher scores indicate less anxiety. scores found in a previous studies (18.92 and 24.9), 3. Fear of Negative Evaluation (FNE) (28) includes 30 mean LSAS Avoidance scores ranged from 24.46-27.97 yes/no questions. The questions relate to cognitive in this study which is also higher than 18.20 and 16.7 aspects of the phobic experience – fear of criticism found in a previous studies (29, 30). Mean FNE scores or negative evaluation. FNE has high reliability (.78- ranged from 20.07-21.80 in this study which is similar .94), and good validity (e.g., correlations with Taylor’s to mean FNE scores (23.14-25.04) found in previous Manifest Anxiety -.60, social and evaluative parts of studies (29, 31). Mean SSPS positive scores ranged the Endler-Hunt S-R -.47). from 10.18-13.90 in this study similar to the score of 4. Behavioral Task: Upon completion of treatment, par- 9.4 found in a previous study, while mean SSPS negative ticipants presented a 10-minute talk on a topic they scores ranged from 11.64-13.10 in this study compared choose (standing, and without notes) in front of a to mean of 15.80 in a previous study (32). live audience composed of four or five staff mem- Subjects in both treatment groups (VRE and CT) in bers who served as observers, and were unaware of this study and in the CBT treatment group in the pre- treatment group placement. Participants were rated vious study received 12 individual one-hour treatment on 10 anxiety indicators. Nine indicators were rated sessions administered according to protocols designed on a 5-point Likert type scale from “very much” to for both studies. Therapists (graduate psychology stu- “not at all” (e.g., eye contact, stuttering), and one was dents in the final stages of their clinical training) were a global fear assessment rated on a 0-10 scale from given extensive training in application of the treatment 0 – “not at all anxious” to 10 – “very anxious.” Prior protocols, and supervised throughout the study by to rating the subjects, all observers underwent a brief senior Clinical Psychologists to ensure that they were training session in order to insure high inter-rater actually following treatment protocols. The WLC group reliability. In addition, subjects rated their anxiety at filled out questionnaires at the beginning and end of a various points (waiting outside the room, giving the 12-week wait period, after which they were randomly lecture, etc.) on 5-point Likert type scales. assigned to CBT (first study). All subjects performed the behavioral task at the end Procedure of treatment or wait period. Subjects were recruited through advertisements on the university website, in campus newspapers and strate- Treatment Protocols: gically placed flyers on campus. Those interested in Cognitive treatment protocol: The cognitive aspects of participating were interviewed by a research assistant, the treatment outlined by Heimberg and Becker (22) signed an informed consent form, completed pre-treat- were utilized in this research. They included: presen- ment questionnaires, and were randomly assigned by tation and discussion of the cognitive model of social the research assistant by order of arrival to one of two phobia and the rational for cognitive treatment, training groups: VRE or CT. Random assignment was performed in cognitive restructuring (identification of automatic before scoring of the questionnaires was undertaken. thoughts, identification of thinking errors, learning Data compiled from subjects that participated in a pre- to correct thinking errors and to replace automatic vious study was employed for the CBT treatment and thoughts with rational responses). Cognitive restructur- the WL groups (16). Recruitment and exclusion criteria ing was practiced during sessions through role play and were similar for that study. in homework assignments as well. No behavioral home- Due to technical, logistic and budgetary constraints work was given and role-play in the sessions was limited no formal diagnosis of PSA or of psychiatric co-mor- to cognitive disputation without behavioral exposure. bidity was undertaken. Therefore, we cannot be certain Virtual reality exposure treatment components:

93 Virtual Reality Exposure versus Cognitive Restructuring

The behavioral aspects of the treatment outlined by groups were compared on demographic variables to the Heimberg and Becker (22) were also utilized in this WL group from the former study (Table 1). Here again, research. They included: presentation and discussion of no significant differences were found on majority/ the behavioral model of social phobia and the rationale minority group, family status, or percentages of males for behavioral treatment, graded in-session behavioral and females. A significant difference was found on age exposure in virtual environments using SUDS ratings, using the Wilcoxon two sample test, Z=2.04, p≤0.05. and graded homework assignment. No cognitive home- The WL group participants were younger (25.29) than work assignments were given, and no discussion of cog- the treatment group participants (27.79). nitions occurred during in session exposures. VR was employed for the exposure. VR components Pre-therapy measures consisted of software and hardware (computer, head- We compared the two treatment groups from this study, mounted display – HMD). The software package was CT and VRE to the WL control group from the previous purchased from Virtually Better, Inc. It provides scenes study on pre-treatment clinical measures. None of the in which the subject is required to read from text, which comparisons were significant: LSAS fear F(2,50)=.48, appears on a podium in the virtual world, in front of a n.s.; LSAS avoidance F(2,50)=.16, n.s.; SSPS positive large audience in various situations (audience clapping, F(2,50)=1.17, n.s.; SSPS negative F(2,50)=.02, n.s.; FNE asking questions, appearing hostile, etc.). The subjects F(2,50)=.23, n.s. Therefore, the groups did not differ provided the text, but the scenes were controlled by the significantly prior to treatment. therapist, according to the hierarchy which was developed for each subject prior to exposure. The computer used Differential utility of VRE in comparison with CT was Intel Pentium 4 with 1 GB Ram and 120 GB hard We compared CT to VRE on both cognitive (FNE, disk with Windows XP operating system, NVIDEA 6600 SSPS positive, SSPS negative), and behavioral (self and 256MB PCI-E graphics card and Integrated SoundMAX observer ratings of global fear, LSAS ratings of fear Cadenza +Internal Mono Speaker audio card. The HMD and avoidance) measures (Tables 2 & 3). As our pri- was VFX3D from Interactive Imaging Systems, Inc. This mary goal was to differentiate between CT and VRE, HMD has a 3D stereoscopic Smart Visor, 35 degree field of and not in their interaction, comparisons were made view, fixed focus at 11 feet, 100% stereo overlap, requires using difference scores employing alpha correction for no IPD and can be worn over eyeglasses. It has two 0.7 multiple analyses. In addition, we calculated effect sizes inch color LCD displays with 360,000 pixel resolution, for the difference between these two groups using the virtual orientation system tracking with pitch and roll Mann-Whitney statistic. No significant differences were sensitivity 70 degrees and yaw sensitivity 360 degrees. found between CT and VRE on FNE, SSPS positive, SSPS negative, therefore, CT was not superior to VRE on cognitive measures. Effect sizes for these compari- Results sons were 0.665 for FNE, 0.615 for SSPS positive, 0.575 Demographic Variables for SSPS negative. For comparison, effect size for FNE The two treatment groups composed of subjects solic- in a previous study was 0.68 (29). Nor was a signifi- ited for this pilot study (CT, VRE) were compared on cant difference found between CT and VRE on LSAS demographic variables (Table 1) using Wilcoxon Two avoidance. Effect size for this comparison was 0.620. Sample Test for age, and χ2 for group status, family sta- However, a significant difference was found between tus and gender. No significant differences were found them on the LSAS fear, t(18)=-2.88, p≤0.01, after per- on majority/minority group status, family status, or per- forming alpha correction: p≤0.049). Effect size for this centages of males and females. A significant difference comparison was 0.845. For comparison, effect size in a was found on age (p≤0.05). The CT group participants previous study was 0.12-0.34 for LSAS fear and 0.32- were slightly older (28.11) than VRE group participants 0.51 for LSAS avoidance (29). Contrary to predictions, (27.5). The three treatment groups were then com- the CT group showed larger reductions in fear than pared (CT, VRE, and CBT) on demographic variables the VRE group. Comparing the treatment groups on (Table 1). No significant differences were found on age, self ratings of global fear (Table 3), only one measure majority/minority group, family status, or percentages (level of fear while standing in front of the audience) of males and females. In addition, the two treatment was significant, S=80, p≤0.05. As predicted the VRE

94 Helene S. Wallach et al.

Table 2. Pre- and Post-Anxiety Measures

FNE LSAS Fear LSAS Avoidance SSPS Positive SSPS Negative M Sd M Sd M Sd M Sd M Sd VRE Pre 21.80 3.58 30.80 12.24 25.30 10.15 13.80 4.47 12.50 4.99 Post 18.90 5.02 28.4 12.03 18.70 10.08 17.20 3.36 9.40 4.20 Change 2.90 4.48 2.40** 6.06 6.60*** 3.86 -3.40* 4.97 3.10 5.78 CT Pre 20.50 4.95 34.80 11.18 27.30 14.14 10.90 6.19 12.60 4.62 Post 19.50 4.48 23.60 8.67 16.80 10.04 15.00 5.42 11.30 4.24 Change 1.00 6.18 11.20** 7.54 10.50*** 8.32 -4.10* 3.63 1.30 4.90 CBT Pre 20.07 6.89 28.18 11.17 24.46 13.65 10.18 5.57 11.64 6.49 Post 16.36 7.86 19.39 8.31 14.86 8.34 16.68 6.22 9.89 6.24 Change 3.71 9.00 8.79 13.81 9.60 13.29 -6.50 6.90 1.75 6.98 WL Pre 21.70 5.70 30.93 11.29 27.97 12.10 11.10 5.20 13.10 5.33 Post 20.97 6.48 29.83 12.94 29.13 11.47 12.10 5.83 13.07 6.61 Change 0.73 4.03 1.10* 8.89 -1.16*** 9.16 -1.00* 5.22 0.03* 4.66

Note. * p ≤ 0.05; **p≤0.01; *** p≤0.0005

Comparing CT and VRE to WL Table 3. Ratings of Behavioral Task (lecture) As the two treatment groups did not differ significantly SG1a SG2 SG3 SG4 SG5 SGt Mean Sd Mean Sd Mean Sd Mean Sd Mean Sd Mean Sd on most of the measures and the groups were so small, we combined the two treatment groups from this pilot VRE 6.10 2.42 4.40 1.17 5.90 2.73 7.40* 1.07 7.80 1.03 6.32 0.78 study (VRE, CT) and compared them to the WL par- CT 7.00 2.98 4.40 2.41 7.10 3.35 8.40* 1.84 7.30 1.77 6.84 1.79 ticipants. We found significant differences on three of Note. aSG-self grading (1-preparing for the lecture, 2-listening to instructions, 3-waiting for turn, 4-standing in front of the audience, the five questionnaires (Table 2). LSAS fear, t (48)=2.31, 5-giving the talk, t-total) p≤0.05, however, this difference disappeared after alpha * p ≤ 0.05 correction p≤0.12. The combined treatment group improved significantly more (reducing their fear by 6.80) group reported feeling less anxious than the CT group. than the WL group (who reduced their fear by 1.10). The However, no significant differences were found when effect size of this comparison was 0.67. LSAS avoidance, they were compared on observer ratings. t(48)=4.08, p≤0.0005, after alpha correction p≤0.001. The treatment group also significantly reduced their avoid- Are CT and VRE Effective? ance (8.55) more than the WL group (-1.16). Effect size Examining changes from pre to post therapy in each for this comparison was 1.21. SSPS positive, t(48)=-1.96, treatment group p≤0.05, however, this difference disappeared following For each therapy group, post-therapy scores were com- alpha correction p≤0.25. The treatment group improved pared to pre-therapy scores using t-tests. The VRE group more (increased by 3.75) than the WL group (increased improved significantly on LSAS avoidance t(9)=5.40, by -1.00). Effect size for this comparison was 0.58. No p≤0.0005, after alpha correction, p≤0.005, and on SSPS significant results were found on the FNE, t(48)=0.92, positive t(9)=2.16, p≤0.05, but not after alpha correc- n.s. (effect size 0.26) or on the SSPS negative, t(48)=-1.52, tion, p≤0.264. The changes on LSAS fear t(9)=1,25; SSPS n.s. (effect size 0.43). negative t(9)=1.70; and FNE t(9)=2.05 were not signifi- cant. The CT group improved significantly on LSAS Is CBT superior to CT and to VRE? fear t(9)=4.70, p≤0.001, after alpha correction p≤0.005; Comparing CBT to CT and to VRE LSAS avoidance t(9)=3.99, p≤0.005, after alpha correc- No significant differences were found between the three tion p≤0.01; and SSPS positive t(9)=3.57, p≤0.01, after treatment groups (VRE and CT from this pilot study, alpha correction p≤0.05. The changes on SSPS negative CBT from the previous study) (Table 2). Using the t(9)=0.84 and on FNE t(9)=0.51 were not significant. Kruskal-Wallis test, LSAS fear, χ2(2)=5.21, n.s.; LSAS avoidance, χ2(2)=0.64, n.s.; SSPS positive, χ2(2)=1.24,

95 Virtual Reality Exposure versus Cognitive Restructuring n.s.; SSPS negative, χ2(2)=0.42, n.s.; FNE, χ2(2)=1.56, n.s. female and unmarried, as well as extremely compliant. Therefore, CBT was not superior to either CT or VRE. It is necessary to repeat this study with a larger, more heterogeneous sample. We accepted self identified PSA participants and did not formally diagnose public speak- Discussion ing anxiety, nor did we examine for psychiatric co-mor- VR has only been employed in therapy for public speak- bidity. This may also have affected our results. Therefore, ing anxiety in a few studies (14, 32, 33). Although VR we recommended employing clinical interviews to derive was found to reduce anxiety in these studies, random formal diagnoses for future research participants. Lastly, assignment with a control group and comparative sta- although we closely supervised our graduate student tistical analyses were not used. One randomized study therapists, we did not have the means to tape therapist using VR with public speaking anxiety (16) found that performance in order to ensure treatment fidelity. It is VR was equally effective as CBT without VR, and that important in future research to do this. both were superior to WL control group. The present pilot study was designed to compare the effectiveness Acknowledgements of Virtual Reality Exposure (i.e., exposure without the The authors wish to thank Sharon Dvir and Nadav Harris, who served as the cognitive component) in comparison with Cognitive therapists, and Hadas Kleider, who assisted in running the project. This research project was made possible by the aid of an excellence research Therapy (i.e., the cognitive component without expo- grant by the Research Authority of the University of Haifa. sure) and with Cognitive-Behavior Therapy (i.e., both the cognitive and the exposure components) for indi- References viduals who suffer from public speaking anxiety. 1. Craske MG. Anxiety disorders: psychological approaches to theory and Both CT and VRE significantly reduced anxiety from treatment. Boulder, Colorado: Westview, 1999. pre to post treatment. CT was not superior to VRE on 2. Beck JS. Cognitive therapy: Basics and beyond. New York: Guilford, cognitive measures, and VRE was superior to CT on 1995. 3. Feske U, Chambless DL. Cognitive behavioral versus exposure only only one aspect of self observation (fear while standing treatment for social phobia: A meta-analysis. Behav Ther 1995; 26: 695- in front of the audience), but CT was superior to VRE 720. on LSAS fear reduction, which is a behavioral measure. 4. Hofmann SG. Cognitive mediation of treatment change in social No other differences were found between CT and VRE. phobia. J Consult Clin Psychol 2004; 72: 393-399. 5. Butler AC, Chapman JE, Forman EM, Beck AT. The empirical status of In addition, no differences were found between either cognitive-behavioral therapy: A review of meta-analyses. Clin Psychol CT or VRE and CBT. All three active treatments (CT, Rev 2006; 26: 17-31. VRE and CBT) were superior to the WL. 6. Oakley ME, Padesky CA. Cognitive therapy for anxiety disorders. In: Hersen M, Eisler RM, Miller P, editors. Progress in Behavior Previous studies (treating various phobias other than Modification (Vol. 25). Thousand Oaks, Cal.: Sage, 1990: pp. 11-46. public speaking anxiety) failed to find significant last- 7. Abramowitz JS, Franklin ME, Foa EB. Empirical status of cognitive- ing differences between CT and BT. Similarly, using a behavioral therapy for obsessive-compulsive disorder: A meta-analytic new technology, namely Virtual Reality, for the expo- review. Rom J Cogn Behav Psychother 2002; 2: 89-104. 8. Fama J, Wilhelm S. Reply to Kozak and Coles: Expanding the sure rendered similar findings whereby using either CT conceptualization of Cognitive Therapy and its therapeutic potential. or VRE alone is satisfactory, and probably sufficient. In: Abramowitz JS, Houts AC, editors. Concepts and controversies in Obsessive-Compulsive disorder. New York: Springer Science + Business In accordance with the literature, both CT and VRE Media, 2005: pp. 305-310. changed maladaptive cognitions and reduced fear lev- 9. American Psychiatric . Diagnostic and statistical manual of els, even though operating through different modalities. mental disorders (4th ed)(Text rev.). Washington, DC: Author, 2000. Although our conclusions parallel those of previous 10. Pollack MH. Social anxiety disorder: Designing a pharmacologic treatment strategy. J Clin Psychiatry 1999; 60: 20-26. studies (e.g., 1, 3), it is premature to conclude that VRE 11. Ruscio AM, Brown TA, Chiu WT, Sareen J, Stein MB, Kessler RC. and CT do not differ in their impact on specific aspects Social fears and social phobia in the USA: Results from the National of the phobia. We found large, but insignificant differ- Comorbidity Survey Replication. Psychol Med 2008; 38: 15-28. 12. Wolitzky-Taylor KB, Horowitz JD, Powers MB, Telch MJ. Psychological ences between them. Because of our small sample size, approaches in the treatment of specific phobias: A meta-analysis. Clin statistical analyses were not powerful. Therefore, we Psychol Rev 2008; 28: 1021-1037. plan to replicate this study utilizing larger sample sizes 13. Goldberg J, Zwibel A, Safir MP, Merbaum M. Mediating factors in the modification of smoking behavior. J Behav Ther Exp Psychiatry 1983; in order to further investigate this important question. 14: 325‑330. In addition, it should be noted that our participants 14. Wiederhold BK, Wiederhold MD. Virtual Reality therapy for anxiety cannot be considered representative – they were young, disorders: Advances in evaluation and treatment. Washington, DC:

96 Helene S. Wallach et al.

American Psychological Association, 2005. Model, methods and outcome. J Clin Psychiatry 1999; 60: 14-19. 15. Garcia-Palacios A, Botella C, Hoffman H, Fabregat S. Comparing 25. Smits JAJ, Hofmann SG. A meta-analytic review of the effects of acceptance and refusal rates of Virtual Reality exposure vs. In Vivo psychotherapy control conditions for anxiety disorders. Psychol Med exposure by patients with specific phobias. Cyberpsychol Behav 2007; 2008; 38: 1-11. 10: 722-724. 26. Fresco DM, Coles ME, Heimberg RG, et al. The Liebowitz Social Anxiety 16. Wallach HS, Safir MP, Bar-Zvi M. Virtual Reality Cognitive Behavior Scale: A comparison of the psychometric properties of self-report and Therapy for public speaking anxiety: A randomized clinical trial. Behav clinician-administered formats. Psychol Med 2001; 31: 1025-1035. Modif 2009; 33: 314-338. 27. Hofmann SG, DiBartolo PM. An instrument to assess self-statements 17. Stopa L, Clark DM. Social phobia and interpretation of social events. during public speaking: Scale development and preliminary Behav Res Ther 2000; 38: 273-283. psychometric properties. Behav Ther 2000; 31: 499-515. 18. Heimberg RG, Becker RE. Cognitive-Behavioral group therapy for 28. Watson D, Friend R. Measurement of social-evaluative anxiety. J Social Phobia: Basic mechanisms and clinical strategies. New York: Consult Clin Psychol 1969; 33: 448-457. Guilford, 2002. 29. Cox BJ, Ross L, Swinson RP, Direnfeld DM. A comparison of social 19. Freeman A, Pretzer J, Fleming B, Simon KM. Clinical applications of phobia outcome measures in cognitive-behavioral group therapy. Behav Cognitive Therapy (2nd ed.). New York: Kluwer Academic, 2004. Modif 1998; 22; 285-297. 20. Wells A, Clark DM. Social phobia: A cognitive approach. In: Davey 30. Klinger E, Bouchard S, Legeron P, Roy S, Lauer F, Chemin I, Nugues GCL, editor. Phobias – a handbook of theory, research and treatment. P. Virtual reality therapy versus cognitive behavior therapy for social phobia: A preliminary controlled study. Cyberpsychology Behav 2005; New York: John Wiley, 1997: pp. 3-26. 8; 76-88. 21. Foa EB, Franklin ME, Perry KJ, Herbert JD. Cognitive biases in 31. Clark DM, Ehlers A, Hackmann A, McManus F, Fennell M. Cognitive generalized social phobia. J Abnorm Psychol 1996; 105: 433–439. therapy versus exposure and applied relaxation in social phobia: A 22. Beck AT, Emery G. Anxiety disorders and phobias: A cognitive randomized controlled trial. J Consult Clin Psychol 2006; 74; 568-578. perspective. New York: Basic Books, 1985. 32. Anderson PL, Zimand E, Hodges LF, Rothbaum BO. Cognitive 23. Hope DA, Heimberg RG. Social phobia and social anxiety. In: Barlow behavioral therapy for public-speaking anxiety using virtual reality for DH, editor. Clinical handbook of psychological disorders (2nd ed). New exposure. Depress Anxiety 2005; 22; 156-158. York: Guilford, 1993: pp. 99-136. 33. Harris SR, Kemmerling Rl, North MM. Brief virtual reality therapy for 24. Otto MW. Cognitive-behavioral therapy for social anxiety disorder: public speaking anxiety. Cyberpsychol Behav 2002; 5; 543-550.

97