CyberPsychology & BehaviorVolume 1, Number 1, 1998Mary Ann Liebert, Inc.

Virtual Reality in Psychological Assessment:The Body Image Virtual Reality Scale

GIUSEPPE RIVA, Ph.D.

Virtual environments (VEs) are attracting much attention in clinical psychology, especially inthe treatment of phobias. However, a possible new application of virtual reality (VR) in psy-chology is as assessment tool: VEs can be considered as a highly sophisticated form of adap-tive testing. In fact, the key characteristic of VEs is the high level of control of the interactionwith the tool without the constraints usually found in computer systems. Both the syntheticenvironment itself and the manner in which this environment is modified by the user's re-

sponses can be tailored to the needs of each client and/or therapeutic application. This arti-cle describes the context of current psychological assessment and underlines possible ad-vantages of a VR-based assessment tool. It also details the characteristics of the Body ImageVirtual Reality Scale, an assessment tool designed to assess cognitive and affective compo-nents of body image. It consists of a nonimmersive 3D graphical interface through which thepatient is able to choose among seven figures that vary in size from underweight to over-

weight.

Virtual environments (VEs) offer a new hu-man-computer interaction paradigm in

which users are no longer simply external ob-servers of images on a computer screen but areactive participants within a computer-gener-ated 3D virtual world. Virtual reality (VR) can

add, delete, or emphasize details to better helpclinicians perform basic functions. Theseunique features can provide the patient withspecialized, safer treatment techniques forproblems that previously were expensive or

impossible to treat in traditional training and

Applied Technology for Psychology Lab, Istituto Aux-ologico Italiano, 28044 Verbania, Italy.

This research is part of the Virtual Reality Enivorn-ments for Psycho-Neuro-Physiological Assessment andRehabilitation—VREPAR project, a European Commis-sion funded research project (DGXIII—Telematics forHealth Care—HC 1053).

therapy. For these reasons, VEs have recentlyattracted much attention in clinical psychology.One of the main advantages of a VE is that itcan be used in a medical facility, thus avoidingthe need to venture into public situations. Infact, in many applications VEs are used in or-

der to simulate the real world and to ensure

that the researcher has full control of all the pa-rameters implied. Many stimuli for exposureare difficult to arrange or control, and when ex-

posure is conducted outside of the therapist'soffice, it becomes more expensive in terms oftime and money. The ability to conduct expo-sures of virtual airplanes for flying phobies or

virtual highways for driving phobies, for ex-

ample, without leaving the therapist's office,would make better treatment available to more

sufferers at a lower cost.However, a promising new use of VEs is in

psychological assessment. This article de-

37

38 RIVA

scribes the context of current psychological as-

sessment and underlines the possible advan-tages of VR-based assessment tools. Also dis-cussed are the characteristics of the Body ImageVirtual Reality Scale (BIVRS), an assessmenttool designed to assess cognitive and affectivecomponents of body image.

VR AS ASSESSMENT TOOL INPSYCHOLOGY

Assessment in modern psychologyAssessment in the modern era, particularly

by clinical, counseling, and research psycholo-gists, rests heavily on tools whose origins ex-

tend back a half century or more.1 For instance,the series of Wechsler scales, the MinnesotaMultiphasic Personality Inventory (MMPI),and the Thematic Appercetio Test (TAT) hadtheir beginnings in the late 1930s. Comparedwith these venerable instruments, widely ac-

cepted behavioral assessment is a relativelynew development.

So, the practice of psychological assessment,along with a number of the conceptualizationson which it rests, now differs dramatically fromthe situation that prevailed when the tenets ofthe measurement tradition provided the basis forthe use of psychological diagnostic instruments.

As Tallent1 (p. 19) underlined "Today's psy-chological assessors are concerned with new

and still developing concepts of therapy, withpsycho-pharmacology, and with the Diagnos-tic and Statistical Manuals and their implica-tions for treatment." In this sense, the practiceof psychological assessment must be flexibleand open-ended attentive to the opportunitiesoffered by unanticipated or developing situa-tions. The watchword is adaptation—to what-ever the situation might be in the ongoing giveand take of the assessment process.2

A key issue in this situation is the distinctionbetween psychological assessment and psycholog-ical testing. Commonly, but erroneously,3 theexpressions psychological assessment and psycho-logical testing are used synonymously or inter-changeably. Sloves, Docherty, and Schneider4made the following distinction between assess-

ment and testing:

Psychological testing [italics added] is de-fined as a set of skills, tactics, and strate-gies subsumed under the heading of psy-chological methods. In this view, methodsrepresent the technical skills used as a

means of carrying out a psychological as-

sessment. Psychological assessment [italicsadded] is systems and problem oriented,dynamic, and conceptual; whereas psy-chological testing is methods and mea-

surement oriented, descriptive and tech-nical.

The authors then point out the consequenceof not attending to this distinction, and theysuggest a remedy:

The failure by many practitioners andtrainers in professional psychology to dis-tinguish between assessment and testinghas led to a tendency for the profession tofocus its attention on the mechanistic andtechnical aspects of test administrationand to ignore or slight the conceptual ba-sis of the assessment process.

Bardon and Bennett5 proposed as a solutionto this problem that practice and training, notjust in assessment but in all areas of psycho-logical services, shift away from their current

emphasis on knowledge and technical exper-tise and toward a conceptual approach to pro-fessional psychology that trains psychologiststo think like psychologists.6Computers and psychological assessment

According to this new paradigm, the use ofcomputers and in particular of VEs, can offernew powerful tools to psychologists. However,the use of computers in psychological testingis not a novelty: It was initiated well over a

quarter century ago.7 Technically, computertesting has its origin in physicalism and psy-chometrics, and the computer applied to psy-chological testing may be considered a psy-chometric machine.8 The basic thesis is that testscores may be empirically linked to contest be-haviors of test takers through the use of algo-rithms (a partly Greek term honoring the 9th-century Arab mathematician Al-Khuwarizmi),

THE BODY IMAGE VIRTUAL REALITY SCALE 39

which are mechanical rules for making deci-sions. For example, on the basis of empiricalcorrelates, when MMPI Scale 6 is above a Tscore of 75, we may expect that the test takerwill show disturbed thinking, have delusionsof persecution and/or grandeur, ideas of ref-erence, feel mistreated, picked on, be angry andresentful, harbor grudges, and rely heavily on

projection as a defense mechanism.1 On the ba-sis of such established relations between scores

and symptom pictures, characteristics that are

commonly found with particular score eleva-tions and patterns may be fashioned into state-ments, stored in a statement library, and calledback whenever a test taker registers the scores

that have been shown to relate to these state-ments empirically.

In practice, however, the statements thateventually issue from the computer are not de-rived entirely by blind adherence to thisscheme. The algorithms also incorporate the ex-

perience of their author and are not free of the-oretical bias, clinical flavoring, intuition, andpersonally held interpretation. So, early beliefsthat the computer would eliminate the need forskilled diagnostic clinicians have not material-ized. Errors, inconsistencies, and misleadingstatements are always a possibility. When com-

puter-derived information is to be employedby a person who does not have sufficient psy-chological background to use the material re-

sponsibly, it falls upon a psychologist to ex-

plain to that person the computer inter-pretation: There must be a clinician betweenthe computer and the client.9

Advantages of VR-based assessment tools

The main problem of current computer-basedassessment is the transformation of the processof psychological assessment into psychologicaltesting. As Tallent1 (p. 25) pointed out:

The reaching of conclusions through theuse of psychometrics often is mislabeledas assessment, as, for example in com-

puter assessment. . . . [Computer tests]do not provide automatic answers to realproblems. . . . What test results mean inany given case is a human judgment.

However, the rate of growth of computertesting is remarkable.10 Computer programsare available for administering, scoring, profil-ing, interpretation, and report writing for oldtests and for new instruments designed specif-ically for computer analysis. Creative varia-tions have appeared. In adaptive testing,11 forexample, items presented to the test taker are

contingent on his or her earlier responses, sim-ilar to Binet testing, where tests at a given agelevel are administered only if at least one sub-set has been passed at the immediately loweryear level.

VR can be considered as a highly sophisti-cated form of adaptive testing. In fact, the keycharacteristic of VR is the high level of controlof the interaction with the tool without the con-straints usually found in computer systems. VRis highly flexible and programmable. It enablesone to present a wide variety of controlledstimuli and to measure and monitor a wide va-

riety of responses made by the user. Both thesynthetic environment itself and the manner inwhich this environment is modified by theuser's responses can be tailored to the needs ofeach client and/or therapeutic application.12Moreover, VR is highly immersive and can

cause the participant to feel "present" in thevirtual rather than the real environment. It isalso possible for the psychologist to accompanythe user into the synthesized world.

In greater detail, there are three importantaspects of VR systems that can offer new pos-sibilities to psychological assessment:

1. How they are controlled: Present alternatecomputer access systems accept only one or

at most two modes of input at a time. Thecomputer can be controlled by single modessuch as pressing keys on a keyboard, point-ing to an on-screen keyboard with a headpointer, or hitting a switch when the com-

puter presents the desired choice, but pres-ent computers do not recognize facial ex-

pressions, idiosyncratic gestures, or monitoractions from several body parts at a time.Most computer interfaces accept only pre-cise, discrete input. Thus, many commu-nicative acts are ignored, and the subtlenessand richness of the human communicative

40 RIVA

gesture are lost. This results in slow, en-

ergy-intensive computer interfaces. VR sys-tems open the input channel: The potentialis there to monitor movements or actionsfrom any body part or many body parts atthe same time. All properties of the move-ment can be captured, not just contact of a

body part with an effector. In the VE, theseactions or signals can be processed in a

number of ways. They can be translatedinto other actions that have more effect on

the world being controlled; for example,virtual objects could be pushed by blowing,pulled by sipping, and grasped by jaw clo-sure.

2. Feedback: Because VR systems display feed-back in multiple modes, feedback andprompts can be translated into alternatesenses for users with sensory impairments.The environment could be reduced in sizeto get the larger or overall perspective (with-out the "looking through a straw effect"usually experienced when using screen

readers or tactile displays). Sounds could betranslated into vibrations or into a registerthat is easier to pick up. Environmentalnoises can be selectively filtered out. For theindividual, multimodal feedback ensures

that the visual channel is not overloaded. Vi-sion is the primary feedback channel of pres-ent-day computers; frequently the messageis further distorted and alienated by repre-sentation through text. It is very difficult torepresent force, resistance, density, temper-ature, pitch, and so on through vision alone.VR presents information in alternate waysand in more than one way.

3. What is controlled: The final advantage iswhat is controlled. Until the last decade,computers were used to control numbersand text by entering numbers and text us-

ing a keyboard. Recent direct manipulationinterfaces have allowed the manipulation oficonic representations of text files or 2Dgraphic representations of objects throughpointing devices such as mice. The objectiveof direct manipulation environments was to

provide an interface that more directly mim-ics the manipulation of objects in the realworld. The latest step in that trend, VR sys-tems, allows the manipulation of multisen-

sory representations of entire environmentsby natural actions and gestures.The focus of this article now shifts to the

BIVRS, an assessment tool designed to assess

cognitive and affective components of bodyimage. BIVRS, which tries to exploit some ofthe advantages of VR, is a clear improvementover current drawing-based body imagescales; even with some limits, mainly due tocurrent technology, it can be considered as thefirst step toward a new approach to computertesting that is more oriented to psychologicalassessment.

BIVRS: A VR-BASEDASSESSMENT TOOL

Assessment of body imageThe construction of measurement proce-

dures for the assessment of body image hasproliferated in recent years.13 Generally, re-

searchers and clinicians have focused on two

aspects of body image: a perceptual compo-nent, commonly referred to as "size perceptionaccuracy," and a subjective component, whichentails aspects such as body size and weightand physical appearance.14

There are two broad categories of proceduresused for the assessment of size perception ac-

curacy:3 body-site and whole-image proce-dures.

Body-site estimation procedures require thatsubjects match the width of the distance be-tween two points to their own estimation of thewidth of a specific body site. For instance, Sladeand Russell15 constructed the movable callipertechnique, which consists of a horizontal barwith two lights mounted onto a track. The sub-ject adjusts the width between the two lights tomatch her or his estimate of the width of a

given body site. The comparison of estimationswith actual body widths, measured with bodycallipers, is used to derive a percentage of over-

or underestimation. For these and other size es-timation procedures, an assessment of the sub-ject's actual width (measured with body cal-lipers) is compared with the subject's estimate,and a ratio of over- or underestimation of size

THE BODY IMAGE VIRTUAL REALITY SCALE 41

is computed. Generally, the great majority ofsubjects overestimate all body sites, and some

data suggest that the waist is overestimated tothe greatest degree.16 Because the estimates ofthe sites are highly correlated, some researcherssum across sites, giving a generic index of over-estimation. It may be advisable, given the ex-

perimental or clinical purpose of the assess-

ment, however, to evaluate each estimation siteindividually.

The whole-image adjustment methods con-stitute a second major category of size esti-mation procedures. With these procedures,the individual is confronted with a real-lifeimage, presented via videotape, photographicimage, or mirror feedback. The experiment isable to modify the representation to make itobjectively smaller or larger than reality. Themeasure of perceptual inaccuracy is the de-gree of discrepancy between the actual real-life image and the one selected by the subject.The schematic figures or silhouettes of differ-ent body sizes are the most widely used mea-

sure for the assessment of subjective compo-nents of body image disturbance.13'17'18 Withthis methodology, subjects are asked to choosethe figures that they think reflect their currentand their ideal body sizes. The discrepancy be-tween these two measures is taken as an in-dication of level of dissatisfaction. A recenttechnical improvement of the figurai or

schematic rating procedure involves the pre-sentation of body schemes on a computerscreen.19

With this method, subjects can adjust thesizes of nine body sites to arrive at the exactimage representation that they believe firsttheir own dimensions. Again, a measure ofgeneric satisfaction with the body can be ob-tained by asking subjects to create an ideal tocompare with their selection of their own cur-rent image. A computer-based test was alsopresented by Schlundt and Bell.20 They de-veloped a microcomputer program for assess-

ing cognitive and affective components ofbody image called the Body Image TestingSystems. The program, which is written inTurbo Pascal language for IBM PC, generatesfrontal-view and side-view silhouettes of a hu-man body. Subjects can make the body sil-houette image grow smaller or larger for nine

independent body regions via the computercontrol system.

The virtual reality modeling languageThe virtual reality modeling language

(VRML) is a "language for describing multi-participant interactive simulations—virtualworlds networked via the global Internet andhyper linked with the World Wide Web." (p.11) 21 All aspects of virtual world display, in-teraction, and internetworking can be specifiedusing VRML.

The first version of VRML (1.0) allowed forthe creation of virtual worlds with limited in-teractive behavior. These worlds can containobjects that have hyperlinks to other worlds,HTML documents, or other valid multimediainternet mail extentions types. The second ver-sion of VRML (2.0), available now, allows theuser to exhibit richer behaviors, including ani-mations, motion physics, and real-time multi-user interaction.

The first step in viewing a VRML documentis retrieving the document itself. The documentrequest comes from a Web browser—either aVRML browser or an HTML browser. Userssend their request to the Web browser, and theWeb browser sends the request on to its in-tended recipient. The Web server that receivesthe request for a VRML document attempts tofulfill the request with a reply. This reply goesback to the VRML browser.21 Once the docu-ment has been received by the VRLM browser,it is read and understood by it creating visiblerepresentations of the objects described in thedocument. Each VRML scene has a "point ofview," which is called a camera—You see thescene through the eye of the camera. It is alsopossible to predefine view points. All browsersfeature some interface for navigation, so thatyou can move the scene's camera throughoutthe world. A VRML world can be distributed—That is, it can be spread across the Web in manydifferent places. In the same way that an In-ternet Web page can be composed of text fromone place and images from another, a VRMLworld can specify that some of its scene comesfrom this place and other objects come from thatplace.

This means that VRML files often load in

42 RIVA

stages; first the basic scene description isloaded, and then—if this refers to nested (scenewithin a scene) descriptions—the browserloads these after the basic scene has beenloaded. Computer speeds are never quite asfast as we would like, and modems are not ca-

pable of meeting the demands we make uponthem. For this reason, there is almost alwayssome delay involved in loading a VRMLworld—It rarely appears immediately or all atonce.

VRML has the ability to show you where ob-jects will appear before they have been down-loaded. Before the object appears, it is shownas an empty box of the correct dimension(called a bounding box), which is replaced by theactual object once it reads in. Called lazy load-ing, it allows the VRML browser to take its time(when it has no other choice, that is), loadingthe scene from several different places whilestill giving you an accurate indication of whatthe scene will look like when it is fully loaded.21

The research projectThe previous considerations have led to the

design of the following research protocol.Objectives. The main aim of this research is

the development of VR-based body image as-sessment technique: BIVRS. BIVRS is a soft-ware program consisting of a nonimmersive 3Dgraphical interface through which the patientis able to choose among seven figures that varyin size from underweight to overweight. Sub-jects are asked to choose the figures that theythink reflect their current and their ideal bodysizes. The discrepancy between these two mea-sures is an indication of their level of dissatis-faction.

The software was developed in two archi-tectures—the first (a) runs on a single user

desktop computer equipped with a VR devel-opment software, such as VREAM or Super-scape, and the second (b) is split into a server

(bl) that is accessible via the Internet and ac-

tually runs in the same VE as in (a) and a VRMLclient (b2) chosen among the ones available forfree in many Internet sites, so that anyone canaccess the application.

The reasons why we propose a BIVRS arevarious.

1. Even though it is by now possible to choosebetween a wide range of different tests forthe assessment of body image, we are stillfar from a culture-free form, because re-search is usually carried out in just one ortwo institutions and in perfect isolation fromthe rest of the world. BIVRS, being designedto run on any local desktop computer andon the Internet in VRML format, would soon

provide a powerful tool to quickly stan-dardize its results (e.g., by an immediatefeedback given on-line right after the as-sessment session). This way, we couldrapidly raise an international multiculturaldatabase that would be capable of furtherdata splitting when needed.

2. VR can add the third dimension to the bodysize silhouettes presented in the test, therebyimproving its effectiveness. Using 3D, it iseasier for the subject to perceive the differ-ences between the silhouettes, especially forspecific body areas (breasts, stomach, hips,and thighs).

3. VR is highly immersive, and it improves theeffectiveness of the test by focusing the at-tention of the user to the testing environ-ment only.

4. The extremely low cost of the system is veryattractive, especially when compared withthe costs of either a traditional assessmentor a computer-assisted assessment devel-oped to run on machines other than smallPCs.

Population. Initially, BIVRS will be submit-ted to a sample of Italian normal (200 subjects)and clinical subjects (30 obese, 30 bulimic, and30 anorectic subjects). Other subjects from dif-ferent countries will then be added to the orig-inal sample. The subject submitted to BIVRSwill also be submitted to other body-image self-report scales in order to investigate the corre-lation among them.

System design and implementation. BIVRS was

developed using a Pentium-based PC (166MHz, 32 mega-RAM, graphic engine: MatroxMillenium 4Mb WRam) and a Power PC-basedMacintosh (PPC 604,160 MHz, 32 mega-RAM).

The development system. We developed thetwo sets of seven silhouettes using the FractalDesign Poser software for Macintosh. Poser isa 3D modeling software through which it is

THE BODY IMAGE VIRTUAL REALITY SCALE 43

possible to build virtual objects easily, particu-larly objects representing human bodies. Itspurpose is then to provide, given some basicdata about the dimensions of specific bodysites, a ready-to-use object to be included in anyVE. The two sets of figures were first devel-oped in a wire-framed mode to obtain preciselygraduated increments between adjacent sizes.Using this model, it was possible to createseven female and seven male schematic figuresthat ranged from underweight to overweight.Both the female and the male sets were thenrendered and pretested. The final sets, com-

posed of more than 10,000 polygons, were thenexported as .DXF files and converted in theVRML standard using the WCTV2POV.EXEprogram. This program is a freeware file con-

verter, developed by Keith Rule for the Win-dows environment, that converts 3D DXF mod-els into VRML 1.0 files. The final VRML fileswere then tested using Netscape 3.1 for Win-dows 95.

Motion input system. We have consideredtwo different input systems, based on the spe-cific module running: the single-user stationapplication and the VRML client-server appli-cation.

Single-user station module. The data glove-type motion input device is very commonlyused in VEs because of its ability to sense manydegrees of freedom simultaneously. The prob-lem with such devices, however, is that the op-erator is also frequently confused because ofthe difficulty in correctly using it, especiallywhen there is a time delay contained in thefeedback loop.

To provide an easy way of motion in BIVRS,we used an infrared two-button joystick-typeinput device: Pressing the upper button, the op-erator moves forward; pressing the lower but-ton, the operator moves backward. The direc-tion of the movement is given by the rotationof the operator's head.

VRML module. As for the VRML module,there is no other choice available than using thehabitual keyboard as an input device. The im-portance of spreading, in any available way,the assessment system all over the Net to stan-dardize the results quickly and neatly has al-ready been discussed. We believe that as In-ternet clients grow more and more sophisti-

cated, and as technology becomes more easilyavailable, it should be possible in a couple ofyears to support better input devices on theVRML version too, thus making it unnecessaryto keep the two modules separated.

ConclusionsVR can be considered as a highly sophisti-

cated form of adaptive testing. In fact, the keycharacteristic of VR is the high level of controlof the interaction with the tool without the con-

straints usually found in computer systems. VRis highly flexible and programmable. It enablesone to present a wide variety of controlledstimuli and to measure and monitor a wide va-

riety of controlled stimuli and to measure andmonitor a wide variety of responses made bythe user.

A first step in the diffusion of VR-based as-

sessment tools is the development of BIVRS.This scale, which tries to exploit some of theadvantages of VR, is a clear improvement overcurrent drawing-based body-image scales. Theimportance of a VR-based body-image scale re-

lies on the possibility to test rapidly in betterand different ways one's perceived body im-age. It also gives researchers a chance to createa transcultural database on body-image datawith ease.

This article has discussed the importance ofVR for the possibility of adding the third di-mension to the body-size silhouettes presentedin the test: Using 3D can improve the effec-tiveness of the test because it is easier for thesubject to perceive the differences between thesilhouettes, especially for specific body areas

(breasts, stomach, hips, and thighs). It has alsobeen noted that such a system should becomevery important for the standardization of body-image assessment data, because of the ex-

tremely high diffusion of the Internet in severaldifferent countries.

A stand-alone version has also been set up,which could run on any low-cost PC. In thissystem, it is possible to provide a better inputdevice, based on an infrared joystick and on a

low-cost head-mounted display. This solutioncan be exported to the VRML version too, as

soon as such input devices are supported onthe Net.

44 RIVA

REFERENCES

1. Tallent, N. (1992). The practice of psychological assess-ment. Englewood Cliffs, NJ: Prentice-Hall.

2. Allen, J.G. (1981). The clinical psychologist as a diag-nostic consultant. Bulletin of the Menninger Clinic, 45,247-258.

3. Wechsler, D. (1958). The measurement and appraisal ofadult intelligence (4th ed.). Baltimore: Williams &Wilkins.

4. Sloves, R.E., Docherty, E.M., Ir., & Schneider, K.C.(1979). A scientific problem-solving model of psy-chological assessment. Professional Psychology, 10,28-35.

5. Bardon, J.L., & Bennet, V.D.C. (1967). Preparation forprofessional psychology: An example from a schoolpsychology training program. American Psychologist,22, 652-656.

6. Berg, M.R. (1985). The feedback process in diagnosticpsychological testing. Bulletin of the Menninger Clinic,49, 52-69.

7. Fowler, R.D. (1985). Landmarks in computer-assistedpsychological assessment. Journal of Consulting andClinical Psychology, 53, 748-759.

8. Tallent, N. (1987). Computer generated psychologicalreports: A look at the modern psychometric machine.Journal of Personal Assessment, 51, 95-108.

9. Fowler, R.D., & Butcher, J.N. (1986). Critique ofMatarazzo's views on computerized testing: All signsand no meaning. American Psychologist, 41, 94-96.

10. Butcher, J.N. (1987). Preface. In J.N. Butcher (Ed.),Computerized psychological assessment: A practitioner'sguide. New York: Basic Books.

11. Weiss, DJ. (1985). Adaptive testing by computer. Jour-nal of Consulting and Clinical Psychology, 53, 774-779.

12. Glantz, K., Durlach, N.I., Barnett, R.C., & Aviles, W.A.(1997). Virtual reality (VR) and psychotherapy: Op-portunities and challenges. Presence, 6(1), 87-105.

13. Thompson, J.K. (1990). Body image disturbance: Assess-ment and treatment. Elmsford, NY: Pergamon.

14. Cash, T.F., & Brown, T.A. (1987). Body image inanorexia nervosa and bulimia nervosa: A review ofthe literature. Behavior Modification, 11, 487-521.

15. Slade, P.D., & Russell, G.F.M. (1973). Awareness ofbody dimension in anorexia nervosa: Cross-sectionaland longitudinal studies. Psychological Medicine, 3,188-189.

16. Thompson, J.K., & Spana, R.E. (1988). The adjustablelight beam method for the assessment of size estima-tion accuracy: Description, psychometrics, and nor-

mative data. International Journal of Eating Disorders, 7,521-526.

17. Fallón, A.E., & Rozin, P. (1985). Sex differences in per-ceptions of desirable body shape. Journal of AbnormalPsychology, 94,102-105.

18. Thompson, J.K., & Paltiz, K. (1988). Multiple aspectsand correlates of body figure ratings: A replicationand extention of Fallón and Rozin (1985). InternationalJournal of Eating Disorders, 7, 813-818.

19. Dickson-Parnell, B., Iones, M., & Braddy, D. (1987).Assessment of body image perceptions using a com-

puter program. Behavior Research Methods, Instru-ments, & Computers, 19, 353-354.

20. Schlundt, D.G., & Bell, C. (June, 1988). BITS: A micro-computer program for assessing cognitive and affectivecomponents of body image. Paper presented at the an-

nual meeting of the Association of the Advancementof Behavior Therapy, New York.

21. Pesce, M. (1995). VRML: Browsing and building cyber-space. Indianapolis: New Riders Publishing.

Address reprint requests to:Giuseppe Riva, Ph.D.

Applied Technology for Psychology LabIstituto Auxologico Italiano

P.O. Box 128044 Verbania, Italy

E-mail: auxo.psylab@auxologico.inet.it